Futreal's Deeper and Wider Approach to Cancer Genomes
By Allison Proffitt
November 11, 2010 | SINGAPORE—“I’ve got a list of five genes that have come out of systematic sequencing efforts,” said Andy Futreal. “It’s a small list, but I think the glass is more than half full . . . but I’m just a naturally sunny kind of guy!”
Futreal, director of the Cancer Genome Project and senior investigator at The Wellcome Trust Sanger Institute in the UK, believes there is plenty of cause for optimism when it comes to sequencing cancer genomes. Speaking at the Frontiers in Cancer Science 2010* meeting in Singapore, Futreal catalogued the cancer advances that have resulted from the past ten years of genome sequencing, and proposed future directions for the field.
Sequence data is already revealing information about the process of mutagenesis that occurs in cancers, Futreal said. “Patterns of mutations, the types of base changes, the context those base changes occur in—they can give you information about past exposure, they can tell you about alterations in DNA maintenance repair . . . and eventually all of this sum total goes to define the somatic genetic architecture and we’d like to know that for each and every phenotype.”
Calling cancer “a disease of the genome,” Futreal advocated a deeper and wider approach to the genome to identify genes that are drivers of mutation and to understand the different classifications of cancer at the gene pathway level.
“If you see a gene that’s mutated and you believe it’s a cancer gene in 2% of your favorite tumor type, there needs to be a next step where that gene is taken through as many cancer types as possible to see if there’s a spike somewhere else,” he said. “There are a large number of genes that need to be sequenced in a large number of cancers to really define this entity we’re talking about.”
With ten years of sequencing behind us, Futreal made his predictions for using sequencing in the future of cancer genomics.
Whole exome sequencing is already being used to identify new cancer genes. As an illustrative example, Futreal cited a breast cancer study of about 75 cases. “About half of the cases don’t have mutations in any of the known cancer genes. Again, this notion of heterogeneity rears its head.” The range of mutation burden is also quite large. The average number of mutations in the exome of a breast cancer patient is 68, but some patients may only have ten coding mutations while others harbor more than 300.
“Lots of people are doing exomes and I think it’s going to be an interesting time,” Futreal said. “I think cancer genes are going to fall out of the sky for a little while, but I think understanding the role they play is going to be a little more challenging.”
Predictive Sequencing
Understanding rearrangements within the genome can be particularly challenging, but Futreal predicts that rearrangements could be key markers and predictors of treatment success.
Hematologists already use markers in the blood to predict relapse and treatment efficacy. Futreal believes this could be used for solid tumors as well. As a proof of concept, Futreal’s team identified the rearrangements in a solid, primary tumor and designed a quantitative PCR assay to identify them. They then tested plasma for DNA with these unique rearrangements dumped into the blood from dead cells.
“You could actually use these rearrangements to hopefully try to monitor tumor response to therapy, identify relapse before it becomes clinically evident, or choose the intensity of therapy based on risk stratification and levels of rearrangement in the blood,” he said. Rearrangements could even serve as a surrogate marker of cell kill in early phase clinical trials.
In one study, tumor DNA rearrangements were sequenced from a patient with osteosarcoma, and by periodically testing the plasma for that patients’ specific tumor rearrangements, soft tissue metastases were presaged by four or five months based on a surge in the levels of tumor-specific rearrangements in the blood.
But as the cost of sequencing plummets, and doing a whole cancer genome is “probably quite feasible,” Futreal is looking towards “whole-genome shotgun, deep coverage sequencing of entire cancer genomes.” Futreal believes that sequencing the whole genome is essential. “The notion that there are no non-coding mutations that are drivers is probably not true, so we need to understand how to begin to think about creating an analytical framework for identifying those,” he said. “You’re not just looking under the lamppost now.”
The whole genome “essentially buys you everything,” said Futreal. For example, in melanoma there are 32,000 somatic mutations in that genome; small cell lung cancer has 23,000. “Along the way you get rearrangements, you get copy number variations, you get information on what strand the mutation occurred in, you get evidence of transcription-coupled repair.”
That scope of the problem led the Sanger Institute and others to create the International Cancer Genome Consortium, to obtain a comprehensive description of genomic, transcriptomic and epigenomic changes in 500 cases of 50 different tumor types and/or subtypes within five to seven years. Twelve countries are represented so far, and Futreal says that the consortium has enjoyed strong government buy in.
“These [efforts] are beginning to bear fruit,” Futreal said. “It’s incredibly rich, deep data.”
* Frontiers in Cancer Science 2010; 8-10 November, 2010, organized by the Cancer Science Institute of Singapore, Duke-NUS Graduate Medical School, Genome Institute of Singapore, and the Institute of Molecular and Cell Biology.