by Jennifer Couzin-Frankel in Science Magazine
published January 14, 2011
In the fall of 2008, Stephen Kingsmore, a longtime gene hunter, was approached by two biotech entrepreneurs. One of them, Craig Benson, had just learned that his 5-year-old daughter had juvenile Batten disease, a rare, fatal, inherited, neurological disorder. The pair had a question for Kingsmore: Could he develop a cheap, reliable genetic test for Batten and other equally horrible diseases, available to all parents to prevent the conception or birth of affected children? Their goal was simple: Do everything possible to eradicate these diseases, because, knowing now which genes cause them, we can.
At the time this kind of screening, called carrier testing, was relatively uncommon. Both parents need to carry the same mutated gene for their child to develop a disease like Batten, and many of these recessive diseases are vanishingly rare. The number of affected children born each year can be in the single digits. Given that, it hasn’t made fiscal sense to offer tests for dozens of diseases to everyone when so few couples will be carriers of any given one. In communities in which certain mutated genes pop up more often, such as Ashkenazi Jews, carrier testing has been common for years and has drastically reduced the number of babies born with diseases like Tay-Sachs.
But DNA sequencing technology was moving fast and costs were dropping. What the two men proposed might now be doable, Kingsmore thought. He took on the project.
Two years later, Kingsmore, bioinformaticist Callum Bell, and their colleagues describe in a paper published online this week by Science Translational Medicine(http://stm.sciencemag.org/content/3/65/65ra4.full) what appears to be the broadest application of “next-generation sequencing” to a medical problem. They used technology that, in Kingsmore’s words, “sprays the genome with sequences” to look for mutations in the genes behind 448 childhood recessive diseases. The experience has been career-changing for Kingsmore: This week, he moves from the National Center for Genome Resources in Santa Fe, which conducts basic genetics research, to Children’s Mercy Hospital in Kansas City, Missouri. There, with support from the hospital and the Beyond Batten Disease Foundation formed by Benson and his wife, Charlotte, he will work to make the test clinically available, he hopes for $500, before the end of the year. The test will be sold by the foundation, which will use some of the proceeds for research into Batten disease and support for families living with it.
This carrier test is different from others now offered, including one for more than 100 diseases sold through physicians around the United States by the California company Counsyl. Those tests all hunt for previously identified genetic mutations for various diseases, working from a list that cobbles together what’s been described in the scientific literature. This captures many carriers, but not all of them. “For some diseases, the mutations on these panels may only account for 20% of mutations” that can cause disease, says Wendy Chung, who directs the clinical genetics program at Columbia University. This means that someone could be told they’re not a carrier when in fact they are.
Next-generation sequencing changes that. Instead of starting with the mutations we know about, it sequences the same DNA again and again to reduce the likelihood of error, and then researchers look for any mutation in a gene involved in one of these rare diseases. The technology is being applied, still experimentally, across medicine, for instance, to diagnose uncommon diseases and to better understand cancer. But the first broad, real-world application will likely be carrier testing of prospective parents, because the medical challenge is straightforward and the technology is nearly ready.
There are still kinks. Among the trickiest is determining whether gene variants that no one has seen before could, when paired with a mutation on the same gene from the other parent, cause disease. We all harbor gene variants that are harmless. Distinguishing those from the pathological is “going to be quite difficult,” says Lawrence Brody, chief of the genome technology branch at the National Human Genome Research Institute in Bethesda, Maryland.
Brody should know: For the past 14 years, he’s run a database of mutations in BRCA1 and BRCA2, genes involved in breast and ovarian cancer. About 10% of people tested for BRCA genes have a variant of uncertain significance. Brody and many others have been trying to determine which of these actually raise cancer risks.
When it comes to Kingsmore’s test, “how many people are going to be confident enough in discovering a new mutation that they’d be willing to terminate a pregnancy?” asks Stephen Quake, who studies biophysics and genomics at Stanford University in Palo Alto, California. Although one goal in all carrier screening is to encourage couples to screen prior to conception, that’s currently rare.
In their test, Kingsmore and colleagues screened 104 unrelated individuals; on average people carried mutations for about three of the 448 diseases. The group built computer software to analyze the DNA sequences and, among other things, determine whether they matched mutations already published. They’ve since expanded the test to cover 570 diseases and are testing it on hundreds more people.
Kingsmore admits to a catch-22 when it comes to assessing whether a new variant is a problem: The best shot at doing so comes from carrier sequencing of many, many people, just as Brody is doing with BRCA1 and –2. But this means that “initially this test will not have perfect knowledge of all diseases.” He predicts it will be about a decade before that changes, and he isn’t sure what, if anything, physicians and prospective parents should be told about variants of uncertain significance before then.
One benefit of next-generation sequencing is that it’s far more accurate than what came before it. When double-checking the mutations that showed up in Kingsmore’s small sample against published work, about a fifth of those were wrong. For example, he cites a paper published about 20 years ago on a mutation in a very rare disease, Lesch-Nyhan syndrome, which reported a massive DNA deletion. In fact, the deletion Kingsmore’s team found in one person (which was predicted, based on bioinformatics analysis, to have the biological effect described in that older paper) was just four DNA bases long. “They never intended that initial paper to be the definitive paper 20 years ago,” says Kingsmore. But as with work in so many rare diseases, studies are sparse and data often hard to come by. He and others hope that sequencing on a much bigger scale will change that, with time.
http://www.sciencemag.org/content/331/6014/130.full?sid=4d79291b-339b-42a4-82a6-5ba5c4a35e20