Gold in those hills

    St. Jude leads the way in the rush to harness the human genome and save lives.

    In January 1848, a construction worker in Northern California spied a lone gold nugget glinting at the bottom of a stream. As news of his discovery spread, thousands of prospectors boarded sailing ships and hitched mules to wagons, converging on the region in a furious race for riches. That tiny fleck of gold transformed the American West.

    One hundred fifty-nine years later, a discovery in Memphis, Tennessee, promises to have a similar effect on the field of cancer research. A team of researchers found new mutations that contribute to acute lymphoblastic leukemia (ALL), the most common type of childhood cancer. The strategy the scientists used to make that discovery has started a “gold rush” worldwide, because it shows researchers how they can identify unsuspected mutations in adult cancers, as well.

    Prospecting for a cure

    During the past four decades, researchers at St. Jude Children’s Research Hospital have made incredible progress in eradicating such childhood killers as ALL. Researchers and clinicians have worked in tandem to figure out how to fine-tune drug combinations and understand the genetic lesions or abnormalities that spawn leukemia. As a result, the ALL survival rate has skyrocketed from a terrifying 4 percent in 1962 to about 94 percent today. While that improvement is reason for celebration, the survival rate is not high enough. The ultimate goal? One hundred percent.

    A pathologist by training, James Downing, MD, has spent the past 20 years trying to understand how genetic lesions can cause a cell to become leukemic and how that information can be used to improve diagnosis and treatment. Downing is particularly interested in how that process occurs and how it contributes to leukemia.

    “We thought that if we could better understand how each of the genetic lesions leads to leukemia, we could figure out which ones are going to be the Achilles heels that we could develop more specific therapy against,” explains Downing, the hospital’s scientific director.

    Although researchers worldwide have studied the issue, nobody knows all of the genetic changes that lead to leukemia. If scientists could find those genetic lesions and catalog the important ones, new treatments could be created and more lives could be saved.

    Until recently, scientists lacked the tools for such a project. Then the human genome project provided a kind of blueprint of what normal genes look like in humans. Scientists also developed new technology to aid in the search.

    In 2005, the timing and conditions were right for St. Jude to conduct a study to pinpoint the lesions that lead to leukemia. The hospital had the technology and a vast store of leukemia samples from St. Jude patients. “We thought that we could apply that technology and gain insights into the lesions that were present in leukemic cells that were not present in patients’ normal cells,” Downing explains. “We would then be able to take that information and start identifying the number of lesions in existence.”

    Tapping the resources

    It was an ambitious proposal, but St. Jude never balks at a challenge. The study would be the largest of its kind thus far. Thanks to the hospital’s Hartwell Center for Bioinformatics and Biotechnology, researchers would be able to conduct a gene-by-gene comparison of DNA taken from both leukemic and normal cells.

    The team used postage-stamp–sized chips, called microarrays, which contain hundreds of thousands of DNA probes. Using computers, investigators can figure out whether a gene is missing part of its DNA or identify increases in the number of specific genes.

    “Nobody else was positioned to do this study,” Downing says. “Nobody else had the collection of tumor samples readily available; nobody else had the well-annotated samples where we know the clinical outcome and many other features of the molecular pathology of the leukemia.”

    Nay-sayers along the way

    Not everyone in the scientific community had confidence in St. Jude and its approach to the problem. Opinions were divided. Many scientists believed that there were so many lesions that St. Jude would never be able to identify them, and that the project would be a colossal waste of time. Other people thought the study would not turn up any lesions.

    “We were convinced that there were going to be interesting lesions,” Downing says. “We knew that it was going to take a lot of hard work and a lot of thought and a lot of well-designed studies to figure out which ones contribute and which ones don’t, but we were in a position to do that.

    “We are positioned to do it because of who we are—we are St. Jude Children’s Research Hospital,” he says. “First we had the samples, which are well-annotated and well-stored. But even more importantly, we had the human and capital resources. We have the Hartwell Center for Biotechnology and Bioinformatics. We have Biostatistics. We have a Hematopoietic Malignancy Program that is a team of people who can collaborate on a project like this. And we have the Pharmaceutical Sciences department that has, through years of effort, acquired normal samples on every patient.”

    And so the project began.

    Panning for data

    In the following months, researchers acquired nearly a half-billion data points from samples that had been obtained from 242 patients and stored in the hospital’s tumor banks.

    “That was an unthinkable amount of data,” Downing says. “A year before that, a couple of hundred thousand data points would have been a lot. But now we were gathering more than 1.2 million data points on every patient, and we were looking at 242 patients, with many repeated tests.”

    Nobody—not even the companies that developed the chips—had fully figured out how to analyze all that data. No one in the world could tell St. Jude how to combine data from the chips, analyze it, normalize it or visualize it. No one had yet figured out how to develop a program that would statistically pick out likely lesions from the millions of data points on every patient. So teams across St. Jude worked together on the problem.

    “Every refinement in analyzing the data helped us see more within it, and so very quickly we were able to show that the data was exquisitely sensitive, highly reproducible and was able to identify many of the lesions that we knew existed,” Downing says.

    A glimmer of gold

    Then the team found a new lesion on one chromosome. “It was a deletion of a gene, and that gene didn’t mean anything to us. There are 20,000 genes, so it was just one of the 20,000,” Downing says.

    On closer scrutiny, the investigators realized that the gene, called EBF, was required for a normal B lymphocyte to mature into a mature B-cell. B lymphocytes are the cells from which acute lymphoblastic leukemia arises. The researchers knew that about 100 genes control B-cell differentiation. So they decided to look at all of those genes.

    The next gene they looked at was a gene called PAX5.

    When Downing walked into his lab one day, postdoctoral fellow Charles Mullighan, MD, PhD, was staring at his computer screen, transfixed. “He was literally white,” Downing recalls.

    “You are not going to believe what I just found,” Mullighan said. “Thirty percent of childhood B-lineage ALL have a deletion in PAX5.”

    Nobody had seen that before. Eventually, the team discovered that 40 percent of patients with ALL had deletions or mutations in either PAX5, EBF or Ikaros, three genes that control the differentiation of blood stem cells into mature B cells. When these cells do not mature, leukemic cells continue to grow, eventually killing the patient.

    Researchers are now considering ways to develop therapy based on this information, a task that St. Jude is poised to do through its new Chemical Biology and Therapeutics department.

    The rush is on

    This study showed the world that undiscovered lesions most likely exist for all cancers. And St. Jude showed the medical community how to find them. “So part of the excitement in the field is, ‘Wow! There are lesions we don’t know about, and here is the way to find them, and everybody should roll up their sleeves and do it,’” Downing says.

    “The other part is that this is the first big paper to say, ‘There are all kinds of lesions; here’s the way to validate them, here’s the way to think about them, and here are some tricks to analyze the data. Eventually there is even going to be better technology, but those will just be refinements, and here is the way to go forward.’ It shows us where to look so that we can begin investigating new therapies.” 

    Developing therapies based on these discoveries will be a long process. “But, really, to some extent, it’s like the gold rush,” Downing says. “From a scientific point of view, what this says is that there’s gold in those hills. Now we know how to find it, and let’s go find it.”

    Downing predicts that within the next few years, this kind of study will be conducted on every human tumor. “As a result, an incredible amount of information is going to come out that will be a leap in our understanding of what causes cancer,” he says. “People are racing to do this, and that’s good. The competition will accelerate research, and we will end up getting answers much more quickly, which is what we are really after, especially in a place like St. Jude. We really don’t care about getting the credit; we just want to figure out how to improve treatment for kids with cancer.”


    Reprinted from Promise magazine, summer 2007.

    If you would like to comment on this article, click here.