thumbnail image

Illustration 1. Left: Drs. Richard Wilson, Timothy Ley, and Elaine Mardis in the Washington University Genome Institute's Data Center.

Download figure to PowerPoint

In 2008, for the first time, scientists at the Genome Institute of Washington University in St. Louis, Missouri, sequenced the complete DNA of a patient with acute myelogenous leukemia (AML), leading to important clues about mutations that can contribute to the disease.

That groundbreaking work was the beginning of a more comprehensive approach to using whole-genome sequencing to better understand the genetics of cancer. Sequencing involves spelling out the precise order of the 6 billion chemical letters (a long string of As, Cs, Gs, and Ts) that comprise a molecule of DNA. The order of the letters varies with each person, and researchers try to learn which variations are healthy and which can lead to cancer and other genetic diseases. Prior to sequencing the genome of the patient with AML, earlier studies involved the sequencing of genes that had a known or possible connection to cancer, a process that is not as thorough and may miss key mutations.

Since that first whole-genome sequencing effort, not only have the genomes of many more cancer patients been sequenced and more mutations discovered, but the efficiency of the technology has improved considerably. “The cost of that project was about $1.5 to $2 million, and a lot of that was technology and software development,” says Richard Wilson, PhD, director of Washington University's Genome Institute. “It took us about 18 months to produce the data and make sense of it. Now we could do the same experiment for about $10,000 in less than a month.”

The ultimate goal of scientists is to be able to complete that analysis in less than a day, which Dr. Wilson thinks will be achievable in approximately 1 year with the help of new sequencing instruments. Researchers hope to accelerate the process of sequencing a cancer patient's DNA and RNA in cancer and normal tissue and then look for the mutations that may have caused cells to become cancerous. The hope is that these discoveries eventually will lead to the development of new targeted therapies for mutations in a variety of cancers.

Dr. Wilson envisions a time, perhaps within the next 5 to 7 years, when a small sequencing device in a physician's office can turn the process around in 1 day, revealing important information concerning mutations that could indicate specific therapies. “We have to expect that there will be new and better targeted drugs and also that there are a lot of drugs on the shelf that you can better utilize once you know what kind of mutations are present in the tumors,” he says.

A Wealth of Findings

The Genome Institute is 1 of only 3 large federally funded genome centers in the United States. The other 2 are Baylor University's Human Genome Sequencing Center (HGSC) and the Eli and Edythe L. Broad Institute of Harvard and MIT. Among other sources, all receive funding from the National Human Genome Research Institute (NHGRI), which is part of the National Institutes of Health.

When their NHGRI grants were re-funded in 2011, all 3 centers experienced a funding cut of approximately 24% and were forced to do some downsizing. Washington University in particular had to lay off 54 of its 327 employees due to a decrease in funding from $37.6 million to approximately $28 million. The fact that the cost of genome sequencing continues to fall has helped the institute continue to pursue many of its projects.

Meanwhile, leaders at Baylor University in Waco, Texas, have been able to cope with their funding cuts by increasing funding from other sources, including the state-funded Cancer Prevention Research Institute of Texas. “It's a very exciting time to see the pace of discovery accelerating,” says David Wheeler, PhD, director of cancer genomics and assistant director of the HGSC at Baylor University. “New genes are being discovered that previously weren't known to be involved in cancer, and we are able to discover new subtypes of cancer.”

Among the wealth of findings made in recent years by the national network of research and technology teams that comprise the National Cancer Institute/NHGRI-funded Cancer Genome Atlas (TCGA) project are genomic analyses of the deadly brain tumor glioblastoma multiforme (the original data were published using Sanger sequencing; it has since been repeated with nextgeneration sequencing in 500 patients rather than the original 91), the most common form of ovarian cancer, and lung adenocarcinoma. Manuscripts also have been submitted regarding colorectal cancer and breast cancer, while articles are being written on endometrial and clear cell kidney cancer as well, says Dr. Wheeler.

A sampling of some of the findings includes:

  • Scientists from the Broad Institute and Dana-Farber Cancer Institute sequenced the whole genome of 25 metastatic melanoma tumors, confirming the role of chronic sun exposure and discovering new genetic changes that are important in tumor formation. In an article published in Nature on May 9th, scientists noted that the rates of genetic mutations rose along with chronic sun exposure in patients.1 The analysis also demonstrated the existence of many structural rearrangements in the tumor type, including both BRAF and NRAS mutations, which are known, as well as PREX2. The latter gene has been found to block a tumor suppressor pathway in breast cancer. In this study, it was found to be altered in 44% of patients.

  • An analysis of genomic changes in ovarian cancer tumors from 500 patients was published in the June 30, 2011 issue of Nature.2 Among the findings was confirmation that mutations in a single gene, tumor protein p53 (TP53), are present in greater than 90% of all such cancers. The gene encodes a tumor suppressor protein. When it is mutated, it contributes to the uncontrolled growth of ovarian cancer cells. Researchers also learned how sets of genes are expressed in a way that can predict patient survival. They identified patterns for 108 genes associated with poor survival and 85 genes associated with better survival.


The new question being considered is whether cancer should be treated according to mutations rather than based on tissue type, Dr. Wheeler says. He notes that the TCGA is working on an effort called the Pan-Cancer Project, in which scientists will analyze mutation profiles for approximately one dozen different cancers.

It may well turn out that, eventually, cancers across different tissue types will be categorized based on their mutation profiles, he says, adding that TCGA data are enabling scientists to explore the commonality in mutation profiles across cancers. “There are just hints that this is true, and if you asked a clinician they'd think you were from Mars,” he says.

Scientists are cataloging, at the molecular level, all the possible changes that can occur in DNA, from simple base substitutions to small insertions and deletions, to largescale rearrangements, amplifications, and chromosome deletions. These changes to the structure of DNA affect the expression of genes to reprogram the cell to exhibit the deadly characteristics of uncontrolled growth and metastasis associated with cancer, according to Dr. Wheeler. “It's across the whole range of molecular and genomic analyses that we can see similarities from one cancer to another,” he adds.

Approximately 20 targeted therapies currently exist for actionable mutations. TCGA will guide the development in very strategic ways to make targeted strategies more effective because scientists are gaining a much better understanding of how specific pathways are affected.

More Work To Be Done

Meanwhile, researchers at Washington University's Genome Institute already are using some of their data to generate the rapid return of results for oncologists. They recently sequenced the genome of a patient with pancreatic cancer and found a mutation in a gene for which there is a targeted therapy, bevacizumab, used to treat other cancers such as those of the brain, lung, and colon. It was used off-label at the patient's own expense, and the tumor did respond, although it was a very latestage cancer.

Clinicians at the university's medical school have opened a variety of studies in conjunction with the institute and are banking samples for genome sequencing. In some cases, when current therapy is not working, patients' genomes are sequenced to determine whether anything can be found that might help.

“Often, the patients who come to us are very late in the game,” Dr. Wilson says. “They are told that it may not help them but it will help future generations.” In approximately 50% of the few dozen patients who have been sequenced, scientists have found something that is actionable, he says.

Another patient, who is a member of the institute's research team, was diagnosed with acute lymphocytic leukemia, which is treatable in children but can be deadly in adults. After undergoing a bone marrow transplantation and twice developing disease recurrence, physicians assumed he would not survive past last Christmas. Researchers sequenced his genome and found some important deletions. They also found that 1 gene, Fmslike tyrosine kinase 3 (FLT3) (which has been implicated in many types of leukemia) was overexpressed by approximately 800-fold. It clearly was driving his tumor. Physicians treated him off-label with sunitinib, which inhibits activated FLT3.

“The results were quick and dramatic,” Dr. Wilson says. “He was in remission in about 10 days, and he got his second bone marrow transplant. I saw him at a meeting this week, and he's looking good and feeling much better.”

Washington University also is beginning its third year in a partnership with St. Jude Children's Research Hospital to decode the genomes of more than 600 childhood cancer patients. They are investigating various leukemias and solid tumors that are common in children. The $65 million project involves sequencing the entire genomes of both normal and cancer cells from each patient. Scientists already have found markers that could indicate early on in the diagnosis which children have a more aggressive type of cancer than others.

Dr. Wilson says that some clinicians already are eager to ask pathologists to develop tests for particular mutations, while others continue to treat patients with 25-year-old protocols.

“We still have a lot of people we talk to who say, ‘We don't know why you're doing this—it will never become part of mainstream treatment,’ but we already see an impact,” he says. “We still need to sequence more patients and more cancer types to really build that database and understand which mutations are common and how they work in concert. There's lots more work for the 3 NHGRI centers.”


  • 1
    Berger MF, Hodis E, Heffernan TP, et al. Melanoma genome sequencing reveals frequent PREX2 mutations [published online ahead of print May 9, 2012]. Nature. doi:10.1038/nature11071.
  • 2
    Cancer Genome Atlas Research Network. Integrated genomic analyses of ovarian carcinoma. Nature. 2011; 474: 609-615.