Dangerous Transformations

The first-ever full-genome analysis of one cancer patient becomes a priceless legacy.

Understanding the dynamic gene mutations that eventually took her life will mean better diagnoses and treatments in the future.



Each colored dot in the top image, obtained from the Illumina Genome Analyzer II sequencer, represents a cluster of DNA fragments with the same sequence. By adding a fluorescent dye-labeled nucleotide during sequencing, each of the four bases in DNA (A,C,G,T) has a specific color to identify it. In the bottom image, a technician prepares for a sequencing run.

Download the Decoding a Cancer Patient's Genes graphic.

"We didn't know what we would find, but we felt that the answers to why this patient had AML had to be in her DNA."
Timothy J. Ley, MD

For the first time, scientists at the School of Medicine have unraveled the DNA of a cancer patient and traced her disease, acute myelogenous leukemia (AML), to its genetic roots. A large research team at The Genome Center and the Siteman Cancer Center sequenced the genome of the patient — a woman in her 50s who ultimately died of her disease — and the genome of her leukemia cells to identify genetic changes unique to her cancer.

Funded in part by a $1 million gift from Alvin J. Siteman, the pioneering work, reported in the journal Nature, sets the stage for using a more comprehensive, genome-wide approach to unravel the genetic basis of other cancers. The team is now sequencing the genomes of additional patients with AML, and they also are planning to expand the whole-genome approach to breast and lung cancers.

"Our work demonstrates the power of sequencing entire genomes to discover novel cancer-related mutations," says Richard K. Wilson, PhD, director of The Genome Center. "A genome-wide understanding of cancer, which is now possible with faster, less expensive DNA sequencing technology, is the foundation for developing more effective ways to diagnose and treat cancer."

At the root of every cancer lies a genetic glitch, a slipup that begins with one mutation and ends with a tumor. Researchers looking for cancer-causing mutations have spent decades scouring suspect genes for clues to the disease. But the task of unraveling cancer in an individual patient has remained out of reach — until now.

Washington University researchers discovered just 10 genetic mutations in the AML patient's tumor DNA that may have caused her disease or promoted its progression. Eight of those were rare and occurred in genes never before linked to AML.

They also showed that virtually every cell in the tumor sample had nine of the mutations, and that the single genetic alteration that occurred less frequently was likely the last to be acquired.

GENE TEAM Richard K. Wilson, PhD, Timothy J. Ley, MD, and Elaine R. Mardis, PhD, stand amid a powerful computational cluster housed in The Genome Center's new, 16,000-square-foot data center.

Like most cancers, AML — a cancer of blood-forming cells in the bone marrow — arises from mutations that amass in people's DNA over the course of their lives. However, little is known about the precise nature of those changes and how they disrupt biological pathways to cause the uncontrolled cell growth that is the hallmark of cancer.

Previous efforts to decode individual human genomes have looked at DNA variations that may be relevant for disease risk. What's striking about the new research is that the scientists were able to sift through the 3 billion pairs of chemical bases that make up the human genome to pluck out the mutations that contributed to the patient's cancer.

"Until now, no one had sequenced a patient's genome to find all the mutations that are unique to that person's disease," says Timothy J. Ley, MD, the Alan A. and Edith L. Wolff Professor of Medicine, who led the project. "We didn't know what we would find, but we felt that the answers to why this patient had AML had to be in her DNA."

To date, scientists involved in large-scale genetic studies of cancer have not done a full side-by-side comparison of the genomes of normal cells and tumor cells from the same patient. Rather, earlier studies have involved the sequencing of genes with known or suspected relationships to cancer, a method that likely misses key mutations.

The new research has been hailed as a true landmark in cancer research by other scientists, including geneticist Francis S. Collins, MD, PhD, former director of the National Human Genome Research Institute. "In the past, cancer researchers have been 'looking under the lamppost' to find the causes of malignancy, but the Washington University team has lit up the whole street," says Collins. "This achievement ushers in a new era of comprehensive understanding of the fundamental nature of cancer, and offers great promise for the development of powerful new approaches to diagnosis, prevention and treatment."

An estimated 13,000 cases of AML will be diagnosed in the United States this year, and some 8,800 people will die of the disease. AML occurs most often among those age 60 or older and becomes more difficult to treat as patients age. According to the American Cancer Society, the five-year survival rate for AML is just 21 percent.

The bleak statistics are all too real to Ley. "It is very discouraging that more progress has not been made for patients with AML," he says. "We treat these patients today in much the same way that we did 20 years ago. If we're ever going to understand cancer, we need to understand the whole genomes of cancer patients. We felt that with new genome sequencing technology, now was the time to take a whole-genome approach."

The researchers sequenced the patient's full genome — DNA from both sets of chromosomes — using genetic material obtained from a normal, healthy skin sample. This gave them a reference DNA sequence to which they could compare genetic alterations in the patient's tumor cells, taken from a bone marrow sample. Both samples were obtained before the patient received cancer treatment, which can further damage DNA.

The scientists then looked for differences — points of single DNA base changes — in the patient's tumor genome compared with the genome of her skin sample. Of the nearly 2.7 million single nucleotide variants in the patient's tumor genome, almost 98 percent also were detected in DNA from the patient's skin sample. This narrowed the number of suspicious variants to about 60,000.

Using sophisticated software and analytical tools, some of which the researchers developed specifically for this project, they identified the 10 mutations (including the two previously known genetic mutations that are common to her leukemia subtype but do not directly cause the disease) by looking for single base DNA changes that altered the instructions for making proteins.

Of the eight novel mutations discovered, three were found in genes that normally act to suppress tumor growth. Four other mutated genes appear to be involved in molecular pathways that promote cancer growth. In particular, one mutation was found in a gene family that also is expressed in embryonic stem cells and may be involved with cell self-renewal, a process researchers think may be an essential feature of leukemia cells.

Another gene alteration appears to affect the transport of drugs into the cell, and may have contributed to the patient's chemotherapy resistance.

"We're still analyzing the patient's non-coding DNA and expect to find a number of additional relevant mutations in this portion of the genome," says Elaine R. Mardis, PhD, co-lead author of the study and co-director of The Genome Center. "But the role of these non-coding mutations will be more of a challenge to elucidate because we do not yet fully understand the function of this part of the genome."

The team also determined that the eight novel mutations in the patient's tumor genome did not occur in the DNA of tumor samples from 187 additional AML patients.

"This tells us there is a tremendous amount of genetic diversity in cancer, even in this one disease," Wilson says. "There are probably many, many ways to mutate a small number of genes to get the same result, and we're only looking at the tip of the iceberg in terms of identifying the combinations of genetic mutations that can lead to AML."

Based on their current understanding of cancer, the researchers suspect that the mutations occurred sequentially. The first mutation gave the cell a slight tendency toward cancer, and then one by one, the other genetic alterations were acquired, with each contributing something to the cancer. One mutation, in the FLT3 gene, was not present in all of the tumor cells, and they suspect that it was the last one to occur.

"The final mutation may represent a tipping point that causes the cancer cells to become more dangerous," Ley says.