A potential goldmine of cancer-related pharmacogenomic data may be on the horizon as the US National Cancer Institute and the US National Human Genome Research Institute this week kicked off a pilot effort to determine the feasibility of mapping the genomic changes involved in all types of human cancer — a project that NCI and NHGRI have dubbed the Cancer Genome Atlas.
"I think pharmacogenomics has driven a lot of the interest in pursuing a project like this, so hats off to the pharmacogenomicists," Anna Barker, NCI deputy director for Advanced Technologies and Strategic Partnerships, told Pharmacogenomics Reporter this week. Determining why some tumors respond to certain drugs is "one of the major interests from this project, and could be one of the early wins from the project — I think clinicians are going to be very smart in looking at this data and figuring those kinds of things out," she said.
The pilot project and the upcoming Atlas promise to be a boon for diagnostics firms, pharmaceutical companies, and genomics tool shops in the form of genome sequence data as well as grant and research opportunities. The three-year pilot project is funded by $100 million, with each of the two institutes contributing half. In that time it will investigate between two and five tumor types and serve as a testbed for the genomics, bioinformatics, and other technologies needed for the later Atlas.
"The intention here is to identify two or three tumor types, and then to extensively characterize the genomic state of hundreds of tumors of each of those types," NHGRI Director Francis Collins said this week in an interview.
Determining why some tumors respond to certain drugs is "one of the major interests from this project, and could be one of the early wins from the project — I think clinicians are going to be very smart in looking at this data and figuring those kinds of things out."
The project should deliver a great deal of data related to important, cancer-related biological pathways, such as those involving epidermal growth factor receptors, "If you expect that we may [make] additional findings that relate to something like EGFR, and the way in which that has given us insights into why certain cancers respond to Iressa and Tarceva, I think that's very likely to be the outcome — at least we hope so," Collins said.
Barker said that the difficulty of obtaining homogeneous samples useful for sequencing and genomic analysis will be a primary criterion for the selection of the pilot's model tumor types. "We're trying to sift through all the tumors in storage in the country that are somewhat homogeneous — hematologic tumors probably being the major focus, but some other solid tumors," she said.
Pilot planners are trying to pare down the number of tumor candidates for analysis, but the list currently includes tumor types important to the field of pharmacogenomics, said Barker. "I think it's going to take another couple of months to sort this out," she said. Lung cancer "is on the plate." While "it would not be our first choice in term of homogeneity," she said, "it might be a good choice in terms of the pervasiveness of the cancer, and the extent to which it's a burden to our country."
Barker added that "there are other tumors that are benefiting from pharmacogenomics that are smaller tumors as well."
When characterizing tumor types known to vary in drug response due to genetic or epigenetic changes, the pilot will "absolutely" investigate and compare responding and non-responding types, said Barker. "That's a key question for us — at least in cancer, we are particularly concerned about understanding why that's true," she said.
Data produced by the pilot should be immediately available in the public domain. "We anticipate large, integrated databases that would be publicly available" supported by the NCI's Cancer Biomedical Informatics Grid, as well as the US National Library of Medicine's National Center for Biotechnology Information, said Barker. "We have really not a precedent for integrating this span of information on this scale," she said.
In the effort to define the road ahead for the later Atlas, the pilot project could reveal much about new drug targets. "We're going to uncover previously unrecognized genes that are involved in carcinogenesis, and those will, in some instances, quite nicely stratify what has previously been considered more monomorphic into some subset [that] may very well predict something about therapeutic response," said Collins.
The two institutes plan for the pilot to drive the improvement of new technology to allow researchers to learn just what to expect from a larger project, in terms of biological discoveries and tool speed and cost efficiency, said Collins. "Obviously, there are many unknowns there, including what the state of the budget is going to be to accommodate that," he said.
The project will likely use "standard technologies that are pretty broadly used today … and potentially some of the new epigenomic technologies," said Barker. "We're just going to wait and see, in terms of how [grant applicants] are going to address the questions — what we want them to do is propose robust technologies that can support high throughput so that we can put genes in the pipeline for the [Genome Sequencing Centers]," she said. "For each of [the pilot's needs], if you're looking at copy number or you're looking at some of the other genomic changes, there are two or three technologies in use across the board for most of these."
In a previous interview with Pharmacogenomics Reporter, Barker said the pilot would also consider novel sequencing technologies.
Some quarters of the research community have criticized the idea of a human cancer genome project for various reasons, including an inadequate appreciation of the difficulties such a project might face. Lee Hood, president of the Institute for Systems Biology, said in July that the premise of the project "is a naïve approach because the signal-to-noise issues ... from sequencing cancer genomes on six or seven dimensions are absolutely enormous."
Collins dismisses much of the criticism. "Some of that skepticism sounds rather familiar," he told Pharmacogenomics Reporter. "Some of us recall similar comments back in 1988 and 1989 about the Human Genome Project — things like, 'Oh, the technology's not up to the task, and it's going to take money from other things; you don't know how to do it anyway; and you won't recruit any smart people because it'll be too boring' — none of that turned out to be true that time, and I'm quite confident it won't be true here either."
— Chris Womack ([email protected])