Neurogenetics Flashcards
intro to genetic testing for kids with neuro conditions
large proportion of children with neurological disorders have monogenic conditions, where an alteration, or variant, in a single gene or short chromosome segment is responsible for the disorder. Consequently, genetic tests are increasingly being incorporated into the diagnostic work-up of paediatric neurology patients
Taking a detailed history, including at least 3 generations, and thorough examination of both the child and their parents is essential to gain a full picture of the clinical presentation, the phenotype. Thorough and accurate description of the phenotype of the child and their parents is a pre-requisite for the correct interpretation of genetic variants
Incorporating genetic tests into the diagnostic pathway of many neurodevelopmental conditions has been shown to accelerate diagnosis in a cost-effective manner and can usually be performed on blood, so is frequently less invasive for a child than conventional testing, which may involve procedures such as lumbar puncture or tissue biopsy
however place for clinical testing even in suspected genetic aetiology as with current methodology, a substantial proportion (in many studies at least 50%) of children in whom a genetic diagnosis is strongly suspected will have negative results in all genetic tests. This is because we do not yet fully understand the genetic architecture of neurodevelopmental conditions
currently no single genetic test that will identify all genetic conditions. Even the most powerful test of agnostic trio whole genome sequencing has substantial diagnostic limitations and will miss a proportion of genetic diagnoses that can be made with other genetic testing strategies. The clinical team must therefore choose the most appropriate genetic test(s) given a child’s presentation and what the possible diagnoses are. Working within a multi-disciplinary team involving a paediatric neurologist, clinical geneticist, neuroradiologist and, for some patients, a biochemist, histopathologist and therapy teams is critical for efficient diagnosis / subsequent high-quality care
neurogenetics and consent
Often children do not have capacity to consent, and consent is taken from the parents, acting in the child’s best interests. However, wherever possible the child should be involved in this process. The nature of the proposed tests should be discussed, including what the test will detect and what it cannot rule out. Consent should be taken by the most appropriate clinician, and for many genetic tests this will remain a specialist such as a clinical geneticist or a paediatric neurologist, but nowadays even general paeds trainees may consent for microarray
Firstly, it is important to explore a family’s views about pursuing a genetic diagnosis. For example:
“We know that for children who are behind in their development, sometimes their difficulties are caused by genetic changes – changes in our DNA ‘instruction manual’. We can look for these by doing certain tests. Some families feel strongly about genetic tests, have you any thoughts about this? ….. If we identify a genetic cause for your son’s problems, this may give us more information about what the future might look like for your child. Having a genetic diagnosis might help us to better look after his health. We may be able to put you in touch with support groups for children with similar challenges, which some people find helpful. Depending on what we find, we may be able to give you more accurate advice regarding whether other family members might be affected, including future children. However, it is your choice, and many families feel that they need time to think about this.”
Secondly, it is important to explain what the test checks for and what it cannot exclude:
“A microarray checks for pieces of missing or duplicated DNA in our “instruction manual”. We know that sometimes this can cause children to struggle with developing certain skills. However, if the test is normal, there might still be a genetic cause. This is because the test does not check for all genetic changes. For example, there may still be small sentences missing, or spelling mistakes in the genetic information. We may therefore suggest doing further tests in the future. “
neurogenetics and VUS
a genetic test can result in a true positive, a false positive, a true negative or a false negative. Because of the substantial dangers of using false positive or false negative genetic test results in clinical care, variants that cannot be interpreted with confidence as being pathogenic or benign are reported as variants of uncertain significance (VUS) - this is those with a probability of being pathogenic of between 10-90%
don’t use these to guide clinical decision making
it is important to warn families during the consenting process that we may identify variants where we are unable to give them a clear answer as to whether they are the cause of their child’s difficulties. It is important to explain that you will probably need to test the parents in order to interpret the genetic status of the child. With the exception of trio whole-genome or whole-exome testing – where the parents’ genomes or exomes are tested alongside that of the child – this is usually limited to testing for specific candidate variants.
For example: In an infant with seizures, we find a VUS in the gene SCN1A. Loss-of-function variants in this gene cause Dravet syndrome. Neither parent has seizures, nor had them as a child. We would sequence, “look at”, this region of the SCN1A gene in both parents. If one parent carried the change this makes it more likely that the VUS may in fact be benign. However, the genetic penetrance and expressivity of the condition needs to be taken into account. Variable penetrance has been described in SCN1A. Therefore, careful assessment of the parents, and expert analysis of variant classification in the context of the family members’ phenotypes is essential to accurate variant interpretation
An example of how to explain VUS in clinic: “our understanding of the human genetic code is not complete, and therefore sometimes we find things where we are unsure what they mean. For some genetic changes, we may not be able to give you a clear answer as to whether it’s the cause of your son’s difficulties. We call these variants of uncertain significance. We may need to take blood tests from yourselves, the parents, to try to interpret his results. In time, we may have more information to be able to interpret these results
you also need to discuss chance of incidental findings, unexpected prognostic information, and findings that may have implications for other family members
As current clinical practice is to not undertake predictive testing in a child for any later-onset condition until they have the capacity to consent, only information regarding childhood-onset conditions are currently returned
Particularly when testing young children, it is important that the family are aware that a genetic test may reveal unexpected prognostic information. For example, genetic testing performed on a neonate may return a diagnosis that carries a significant risk of learning difficulties. This can cause significant distress, particularly if unexpected.
For example: A neonate with a structural cardiac anomaly had a microarray which identified 22q11 deletion (DiGeorge syndrome) implying a high likelihood of intellectual disability. His parents were looking for an explanation for the cardiac anomalies and were not expecting a result that would indicate a high probability of learning difficulties.
“We are doing this test to see if we can find a reason why your child has some issues with their heart. However, by doing this test, we sometimes find additional information that we weren’t looking for – for example, we may discover that your son may have an increased risk of learning or behavioural difficulties in the future. This test may also identify information that is relevant for others in your family, such as your siblings, nieces and nephews.”
For some families, such prognostic information may be very unwelcome. For others, there may be a stigma associated with genetic diagnoses. Where consent is declined, this must be respected. However, if a genetic diagnosis has the potential to substantially alter clinical management this should be carefully explained to ensure that the family’s decision is well-informed and taken in the child’s best interest
choosing a genetic test - when specific diagnosis strongly suspected
often appropriate to order a specific test for that condition. By ordering specific tests, the risk of identifying VUS and incidental findings is reduced. For example, Trisomy 21 has a characteristic facial appearance and examination findings. When Trisomy 21 is suspected a rapid quantitative PCR test will allow molecular counting of the number of copies of chromosome 21 in each cell (Table 2). Typically results are available in a few days. However, if the results are negative, exclusion of mosaic trisomy (present in just a subset of cells) by karyotype analysis and/or more extensive genetic testing may be indicated.
choosing a genetic test - when the clinical phenotype is non-specific, yet the suspicion of a genetic cause remains high, and there are multiple possible genetic diagnoses
For example, epileptic encephalopathy presenting in an infant may be caused by variants in a large number of genes, and often there is a significant degree of overlap between the clinical presentations. In such cases, the most efficient approach is to choose an investigation that will interrogate a large number of candidate genes concurrently, such as a gene panel for infantile epileptic encephalopathy, or whole-exome or whole-genome sequencing
intro to interpretation of genetic tests
We all carry 4-5 million genetic variants (sites where our genome differs from the reference sequence). Therefore, it can be challenging to identify whether an observed genetic change is responsible for a child’s difficulties, or whether it is part of normal inter-individual variation. Misdiagnoses and missed diagnoses can have severe consequences for families
The American College of Human Genetics have published clear and objective guidelines for the systematic classification of genetic variants
It is important to bear in mind that there is an up to 10% chance that a variant classified as “likely pathogenic” is in fact benign or not the major causative variant. It is therefore almost always appropriate to conduct independent confirmatory testing, for example biochemical assays, or biomarker testing where these tests are available. e.g. If we found a variant in the gene for phenylketonuria, we must check the level of phenylalanine- which should be raised if the variant is disease causing. Likewise, it is almost always appropriate to conduct segregation analysis – to test the child’s parents to establish whether a variant has been inherited from them, or occurred de novo, for the first time, in that child
For example: a gene panel for early infantile epileptic encephalopathy is performed on a 5 month old baby. She has had pharmacoresistant seizures since 2 months of age, and no response to trial of pyridoxine. In addition, she has progressive microcephaly, hypotonia and developmental delay. In her sequence, a single nucleotide variant is identified in the gene CDKL5 that results in a premature stop codon, a stop-gained variant. This is predicted to prevent the proper function of CDKL5. This variant has not previously been reported before in healthy individuals. Her parents do not carry this variant. Other patients have been reported with similar variants in CDKL5 who have features similar difficulties to our patient. Following MDT discussion, the clinical genetics team feel that this variant is likely to be pathogenic (the cause). As the parents do not carry this variant, they are advised that there is a low probability of another child being affected. However, because there is a possibility that one parent may carry a sub-population of eggs or sperm that have this variant (germline mosaicism), they are advised that they would be offered pre-conception counselling to discuss their options in a future pregnancy, which could include predictive testing of a future pregnancy
domains considered in variant interpretation
- Is the variant in a gene associated with disordered development? Care must be taken to critically evaluate the literature reporting associations between genes and disease; the Developmental Disorder Genes to Phenotype (DDG2P) database is an invaluable clinical resource; This clinically curated list of genes reported to be associated with developmental disorders is categorized by the degree of certainty that the gene causes a developmental disorder, the likely mechanism through which genetic variants are thought to act, such as loss-of-function, or activating; and the allelic status associated with disease (monoallelic or biallelic)
- Does the associated disorder from step 1 fit the phenotype of the child? It is important to bear in mind that young children may not demonstrate the full phenotypic expression of a disorder
- Has the variant been observed:
a. In another patient or patients with a developmental disorder? In which case, what is the strength of the reported association, and does it fit well with the patient we are considering?
b. In healthy people? Variants which are observed in healthy adult members of the general population in population genetic databases such as the Genome Aggregation Database (GnomAD) are highly unlikely to be responsible for a monoallelic rare, severe developmental disorder
c. In the patient’s parents? - Does the genetic variant alter the function of the protein, and does this fit with the mechanism through which pathological variants are known to act, such as loss-of-function or activating variants? Sequence variants which are predicted to severely disrupt a protein-coding sequence, hence truncating the protein product of the gene, for example by the introduction of a stop codon or a frame-shift are called loss-of-function (LoF) variants. They are more likely to cause disease than variants which do not alter the amino acid sequence - synonymous variants; Missense variants change the amino acid sequence, and establishing whether this alters the function of the protein can be difficult. In some cases it will be feasible to do experiments to directly test this. Where this is not possible, computational algorithms are used that attempt to predict the effect that a genetic variant will have on the function of the encoded protein. However, the results of these algorithms are sometimes contradictory. This is why the use of genetic databases such as GnomAD, DECIPHER and ClinVar and the analysis of the segregation pattern of the variant within the child’s family, is essential to interpret the functional impact of the variant.
It is essential to refer all children with a suspected genetic disorder to a clinical geneticist
data sharing in paediatric genomics
because our understanding of the genome is incomplete, we need to compare an individual’s genetic sequence to that of many others in order to best understand it, but we need to balance these clinical benefits with the potential harms of sharing highly identifying data - field uses proportionate approach
broad types of tests
Genetic tests can be divided into those which assess structural variants and those which test at the individual base level for sequence variants. It is also important to bear in mind that there are some conditions such as imprinted syndromes, tri-nucleotide repeat conditions and mitochondrial disorders that must be tested for specifically, as they may be missed by commonly used tests for structural or sequence variants
testing for structural variants
CNVs refer to deletions or duplications of parts of chromosomes, and are estimated to account for about 14% of neurodevelopmental disorders. Currently, CNVs are usually tested for clinically by genomic microarray
However, targeted FiSH and MLPA are often used to investigate specific small CNVs, for example, when Spinal Muscular Atrophy is suspected to test for the deletion of SMN1 exon 7
Large CNVs (>5-10Mb) will be identified by karyotype analysis (Table 2). However, not all CNVs cause disease. Many show variable penetrance and expressivity, requiring expert interpretation. There are several factors that affect whether a CNV is likely to be pathogenic. In general, large and gene-rich CNVs are more likely to be pathogenic, and deletions tend to be more damaging than duplications
Where a trisomy or a relatively common CNV syndrome (such as 22q11.2 deletion syndrome) is clinically suspected, specific tests that count the number of copies per cell are indicated, as they will return a faster result than microarray analysis. For trisomies, usually this is done by rapid quantitative PCR, while for CNVs like 22q11.2 deletion syndrome, fluorescent in situ hybridisation (FiSH) will often be used. However, these techniques may not detect mosaic cases. Therefore, where clinical suspicion remains but results suggest a normal genetic structure, additional testing is indicated
Karyotype analysis characterises the physical structure of all chromosomes at a lower resolution than a microarray. However, it is useful for identifying underlying translocations, and is particularly important in cases of aneuploidy (to exclude mosaicism or a Robertsonian translocation) or terminal deletions/duplications, where one parent may carry a balanced translocation. Parents with a balanced translocation will have the correct amount of genetic information, but it is arranged in an unusual way that may predispose to a high recurrence risk in future pregnancies. Microarray analysis may not identify such a translocation.
For example, translocation of the long arm of chromosome 21 to the long arm of chromosome 14 results in a balanced Robertsonian translocation. The individual carrying this unusual chromosome is unlikely to have any symptoms, however they have an increased risk of having a child with Down’s syndrome.
testing for sequence variants
- Single gene sequencing
When a disorder has a distinct, clinically recognisable phenotype, it may be appropriate to seek a genetic diagnosis by Sanger sequencing of the associated gene. However, in situations where one is considering sequencing two or more genes, an exome or a gene panel is likely to be more economical. - Gene panel testing
Gene panel tests are now usually conducted by sequencing the whole exome, but limiting the analysis of the data to a list of genes, a virtual gene panel, known to be associated with a given clinical presentation (e.g. neuronal migration disorders). This will reduce the identification of VUS. The genes included on the panel may vary depending on the provider. However, as an exome is run before analysis is restricted to the gene panel, it may be possible to analyse additional genes if clinically indicated. - Whole-exome sequencing
The protein-coding portions of the genome, the exome, are isolated and sequenced. It is now cheaper to sequence an exome than to sequence more than a few genes using Sanger sequencing. However, there are other potential costs associated with genetic testing at this scale. Only approximately 30% of all genes are currently associated with a clinical disease. The increased amount of data generated makes analysis substantially more complex and increases the risk of false positives, false negatives and VUS. For this reason, whole-exome and whole-genome sequencing is usually undertaken in a ‘trio’ context, involving sequencing of the child together with both of their parents in order to aid variant interpretation, and to limit the identification of VUS - Whole-genome sequencing
In contrast to whole-exome sequencing where only the protein-encoding sequence is captured and analysed, whole-genome sequencing sequences the whole genome. It is therefore currently a substantially more expensive test than whole-exome sequencing, although the cost is likely to fall over the coming years.
special tests
Where imprinted syndromes, tri-nucleotide repeat conditions and mitochondrial disorders are included in the differential diagnosis, specific tests must be considered to avoid missing them
A small number of genes, about 100 in humans, is subject to genomic imprinting. This means that the two copies of the gene do not behave the same. Imprinted genes are expressed in a parental-origin dependent manner. For some genes the maternally-inherited copy is expressed, for other genes the paternally-inherited copy is expressed. Imprinted genes are not distributed at random through the genome, but grouped together in clusters; The parental-origin specific expression of the imprinted genes in a cluster is controlled by imprinting control regions which are differentially labelled with epigenetic marks such as DNA methylation on the maternally and paternally inherited chromosomes; Disruption of these epigenetic marks at imprinting control regions can interfere with the parental-origin specific expression of these genes, thus When imprinted syndromes are suspected it is important to test for these by analysing methylation status at the relevant imprinting control region; For example: a neonate has difficulties with feeding and is noted to be very hypotonic. Clinically, Prader Willi syndrome is suspected. The methylation level at the imprinting control region of the 15q11-13 imprinted cluster is assessed. The normal methylation level is ~50%, as this region is methylated on the maternally-inherited copy and unmethylated on the paternally-inherited copy. In this patient, the methylation level is ~100%, suggesting a diagnosis of Prader-Willi syndrome. A microarray confirms deletion of the paternally-inherited (normally unmethylated) region, confirming the diagnosis. The parents are tested and, as expected, they do not carry this deletion. Therefore, the risk that another child of theirs will be affected is therefore low
Disorders caused by the expansion of trinucleotide repeat regions such as Fragile X syndrome, Friedrich’s ataxia and myotonic dystrophy type 1, must be tested for specifically using triplet primed PCR tests. Current analysis methods of whole-exome and whole-genome sequencing techniques do not detect this type of repeat expansion, due to difficulty in mapping the precise length of the repeat region
Identification of variants in mitochondrial DNA will require direct sequencing of the mitochondrial genome. It is important to check whether a panel/exome/genome analysis will include the mitochondrial DNA. If the causative variant is found in the nuclear DNA, testing on any tissue, such as blood, will identify it. However, not every mitochondrion carries the same genetic sequence. Furthermore, it is well recognised that different tissues sometimes carry varying proportions of mitochondria with particular genetic variants. This phenomenon is called heteroplasmy. Therefore, if there is a strong clinical suspicion of a mitochondrial disorder, and appropriate genetic testing of blood is negative, it is important to consider repeating the tests using another disease-relevant accessible tissue, such as muscle
