ARTICLE IN BRIEF
Using a technique called next-generation sequencing for capturing ataxia mutations, investigators reported finding 13 mutations — nine of them novel — in eight genes, including a new ataxia disorder recently described.
Ataxia, like several other neurological disorders, results from a multitude of mutations in numerous genes, producing symptoms that often overlap with other conditions including spasticity, dystonia, and highly variable mitochondrial disorders. As a result, ataxia patients often endure years of invasive testing, including muscle biopsy and lumbar puncture, without achieving a clear genetic diagnosis.
A new approach, however, focuses less on interpreting symptoms and more on detecting their cause — the mutations themselves.
Researchers at the University of Oxford have completed a pilot study in which they used next-generation sequencing (NGS) — massive parallel sequencing in which millions of fragments of DNA can be sequenced simultaneously — to capture 58 known ataxia genes in 50 patients who had already been tested for spinocerebellar ataxia 1–3, 6, 7, and Friedrich's ataxia, and undergone other biochemical, genetic, and invasive tests.
The patients all had the core clinical features of cerebellar ataxia, some had additional clinical features such as plasticity, peripheral neuropathy, learning difficulties, and epilepsy.
In the October issue of the journal Brain, they reported that they found 13 mutations — nine of them novel — in eight genes, including a new ataxia disorder recently described in PLoS Genetics. They identified the ataxia mutation in 18 percent of the cases overall, ranging from 8.3 percent in patients with adult onset progressive ataxia, to 40 percent in those with childhood or adolescent onset. Among those with adolescent onset and a family history of ataxia, the detection rate reached 75 percent. The time to diagnosis for these patients ranged from 3–35 years, with a mean of 18.1 years.
“We suspected that if we tested a highly varied group of (ataxia) patients, we would make a molecular diagnosis in some, proving that NGS is valuable for clinical diagnostics,” said lead author Andrea H. Németh, MBBS, DPhil, of the Nuffield department of clinical neurosciences at the University of Oxford. “We also suspected that some patients would not have mutations in known genes, and in these patients we could then work out the best way of finding new ataxia genes.”
The most common ataxias involve trinucleotide repeat expansions, and genetic testing for these is relatively simple and inexpensive. Genetic testing for point mutations, however, is so expensive that most patients do not have it done. NGS, by screening multiple genes at once, enables a more comprehensive testing at a much lower cost.
“Mutations in different genes can cause similar clinical phenotypes,” Dr. Németh explained. “This makes specific selection of one gene for testing quite difficult in a specific family or patient. However, humans are genetically variable, and when interpreting NGS data, it is essential to work out which variations are disease-causing, and which are not. We used bioinformatics to help us predict which genetic variations were likely to be pathogenic, and then confirmed these findings in the laboratory, demonstrating that our bioinformatics assessments worked well.”
The researchers noted that they failed to achieve a molecular diagnosis in all patients, possibly due to clinical overlap between patients with other neurological features such as spasticity, neuropathy, and retinitis pigmentosa, which can be caused by mutations in other genes such as those associated with mitochondrial disorders and neurodegenerative metabolic conditions.
The researchers chose ataxia because the genetics of the disorder are reasonably well understood, but genetic testing for them is not widely available. “We know about many of the genes in which mutations cause ataxia, and in these genes we are usually able to work out which genetic variants are pathogenic, which then enables us to make a molecular diagnosis,” Dr. Németh said.
NGS has been introduced for ataxias and a number of other disorders at the Oxford University Hospitals Medical Genetics Laboratories, Dr. Németh said, and is available both within and outside the UK.
For the past 18 months, clinicians at the University of Minnesota Medical School have been using NGS to diagnose the genetic underpinnings of ataxia and about 130 other disease conditions, said Bharat Thyagarajan, MD, PhD, assistant professor in the department of laboratory medicine and pathology, who was not involved in the current study.
“We have tested about 325 patients — 37 with ataxia,” Dr. Thyagarajan said. “We have a 568-gene panel that covers a lot of neurodegenerative diseases and congenital blindness disorders, pediatric metabolic disorders, familial cancers. The power of the technology is, you can sequence a lot of the genome very cheaply. Previously, there were a lot of ataxia genes known, but each mutation in each individual gene was too rare to make it viable for a clinical test. Now you can put all genes together and offer it as one test. That's where the economies of this come in.”
In traditional genetic sequencing, each nucleotide is tagged with a dye, which cannot be removed after the chemical reaction has taken place. “With this new technology, you can read the dye intensity and then remove the dye and continue adding bases,” Dr. Thyagarajan said. “We went from being able to sequence a few thousands bases to hundreds of thousands of bases at the same time.” The investigators detected mutations in roughly 25 percent of the cases tested so far, he added.
The researchers at Oxford, for example, were able to search for variants or mutations in 118 genes, and they identified over 5,000, Dr. Thyagarajan observed. “Then they run these mutations through these computer programs that predict if a mutation will be pathologic or not,” he said. “The program considers such things as, is this mutation in a gene evolutionarily conserved across species? If it is, we know it is important, and any change in its position probably is pathogenic. Or if there are changes in the protein structure, that's probably important.”
NGS has trouble sequencing certain regions of the genome, however, and it is not good for detecting repeat regions, according to Dr. Thyagarajan. Still, it is transforming traditional diagnostic testing, which relies heavily on the judgment of the clinician who guesses what the inheritance pattern of the disorder might be, and then orders tests.
“The major reason why clinicians made educated guesses was to reduce the cost of testing,” he said. “Each test for a gene costs around $1,000-$1,500, so you want to limit the number of genes you test. Our ataxia panel has about 227 genes, and it costs the same whether you test for one mutation or 500. We have discovered mutations in genes we wouldn't have suspected clinically. This also works for congenital blindness and hearing loss syndromes. You have a good 60–100 genes that have been shown to be involved in the pathogenesis of these diseases. Now you can test for them all.”
And NGS will become even less expensive and more convenient as companies develop standardized kits. “As these bioinformatic products become more accessible, many smaller labs will be able to use this technology,” he said. “Right now it's restricted to big reference commercial labs or academic institutions. We have desktop sequencers that do smaller amounts of data, but plenty for clinical use. I can see cost of clinical testing bottoming out at around $1,000, which is not terribly expensive for the kind of conditions being tested.”
While NGS is very useful for disorders with known genetic causes, such as hereditary ataxia, it will be less useful for disorders such as autism and schizophrenia, which have multiple unknown genetic causes, Kenneth H. Fischbeck, MD, chief of the Neurogenetics Branch at the National Institute of Neurological Disorders and Stroke (NINDS), told Neurology Today.
“The use of NGS is currently limited by the difficulty in establishing the pathogenicity of sequence variants that are found,” he added. “Because of this difficulty in establishing that a variant is pathogenic, NGS use is still mostly limited to research labs. As the databases of established pathogenic variants grow through ongoing research, NGS should become more clinically useful, and it may eventually become the diagnostic procedure of choice for patients with hereditary diseases.”
But within a decade, NGS will be supplanted by whole-genome sequencing, said Nicholas Katsanis, PhD, director of the Center for Human Disease Modeling and Jean and George W. Brumley distinguished professor of pediatrics, and professor of cell biology at Duke University.
“The (NGS) panels are just an intermediate measure while the cost of whole genome comes down,” he said. “Then we'll have the advantage of having done whole genomes for hundreds of thousands, if not millions, of people, so we will have a very rich understanding of the amounts and types of variations that exist in different genes, and the pathologies.”
If integrated with clinical records of individual patients, such an accumulation of data could produce reliable computational prediction models, according to Dr. Katsanis.
“We'll have some sort of probabilistic measurement based on the genome of individuals,” he said. “So for Mendelian disorders that have alleles that are very highly deterministic, we should be able to make a very accurate diagnosis at the first sign of pathology. Blending functional testing of alleles with population-based observations will be the key to understanding susceptibility, causality, and outcomes. At the moment, computational interpretation is weak, but it will get better as we accrue more data.”