Based on exon resequencing data from a single cancer patient, researchers at Johns Hopkins University have discovered a new gene that, when mutated, predisposes its carrier to hereditary pancreatic cancer.
Maybe more important than the gene itself, which only explains a fraction of hereditary pancreatic cancer, is the sequencing-based approach the researchers used to pinpoint it, which they believe will be useful for finding new disease genes — especially once the cost of sequencing drops further.
"In the future, this will be an extraordinarily cost-effective way of finding genes," said Jim Eshleman, an associate professor of pathology and oncology at the Sol Goldman Pancreatic Cancer Research Center at the Johns Hopkins Medical Institutions, one of the researchers involved in the study, which was published online in Science last week.
According to Eshleman, this is the first time scientists have identified a gene responsible for a hereditary disease based on sequencing all human exons.
For their study, the scientist analyzed exome data from a germline control DNA sample of a single pancreatic cancer patient. The data was generated by Sanger sequencing of PCR-amplicons as part of a larger project that involved sequencing the exomes of 24 pancreatic cancer samples and matched normal controls and was published in Science last September (see In Sequence 9/9/2008).
The patient had a hereditary form of pancreatic cancer — which accounts for about 10 percent of that cancer type — based on the fact that his sister had also developed the disease.
Among the more than 15,000 germline variants the researchers found in the patient's exon-coding sequence, they focused on those that rendered one allele of a gene inactive, for example through a truncating mutation. They further narrowed the list by requiring that the second copy, which is normal in the germline, was inactivated in the patient's cancer.
Three genes showed such mutations, but for two of them, such mutations are "relatively common" in healthy individuals, according to the scientists.
They then sequenced the third gene, PALB2, in 96 additional familial pancreatic cancer samples and found truncating mutations in three of them. Other scientists did not find such mutations of the PALB2 gene in more than 1,000 healthy individuals from a previous study.
Some, but not all, of the patients with a PALB2 stop mutation also had a history of breast cancer. Also, germline mutations in the gene — the function of which is not well understood — had previously been associated with breast cancer and Fanconi anemia.
To translate their findings on PALB2 into clinical use, the Hopkins researchers are now tentatively planning to develop a test for the gene, which could take several months and cost "tens of thousands" of dollars, according to Eshleman. "It's a matter of cost and priorities," he said.
According to Alison Klein, director of the National Familial Pancreas Tumor Registry at Johns Hopkins and another study author, the new gene explains about 3 percent of hereditary pancreatic cancer, making it the second most common pancreatic cancer susceptibility gene after BRCA2, which accounts for up to 10 percent of the familial form of the disease.
However, 85 percent of hereditary pancreatic cancer cannot be explained by any genes today, she said, and the exon-sequencing approach might help researchers find additional susceptibility genes for this and other cancers, as well as for hereditary single-gene diseases.
[ pagebreak ]
Traditionally, researchers have found such disease susceptibility genes by studying large families with several affected members, and tracking genetic markers associated with the disease as they get passed on between generations.
However, "that's a very expensive approach, and as the cost of sequencing go down, which is well predicted that it's going to do, this may be a better approach," Eshleman said.
Another problem with pancreatic cancer in particular is that the disease is rapidly fatal.
"Getting DNA from just two affected members of the family, let alone three or four or five, which is what you really need for linkage studies, is extremely challenging," according to Klein. "By the time the second or third case develops in the family, the first person who developed the disease is typically deceased."
"We have been hampered by the lack of these large families, and a lack of our ability to use traditional approaches," she said. "This is really an approach that can be applied to pancreatic cancer as well as many other diseases."
Sanger-based sequencing and analysis for the exome of the patient in this study cost on the order of $150,000 to $200,000 — a price point that currently prevents the researchers from analyzing another eight familial pancreatic cancer and germline samples that they would like to study. "We would do it this way tomorrow if we had the money," Eshleman said.
As a lower-cost alternative, he and his colleagues have been exploring whether to use the Illumina Genome Analyzer or the Applied Biosystems SOLiD, both of which they have in-house at Johns Hopkins. They have also been evaluating undisclosed exon capture methods for use with these second-generation sequencing technologies. "We are looking at them, but we haven't made a decision yet," Eshleman said.
In addition, they are considering Complete Genomics' whole-genome sequencing service, although Eshleman noted that interpreting data from the entire genome is currently too complex for their purposes. "We are just at the stage where we think we can interpret the coding portion of it," he said.
Nevertheless, while it might seem "strange to sequence the entire genome, and then literally throw way 99 percent of it," he said, "that may be the most cost-effective way of doing it going forward."