Public Release: 

Scientists using gene chips identify a unique form of leukemia

Dana-Farber Cancer Institute

BOSTON -- Dana-Farber Cancer Institute scientists have discovered that a rare but lethal blood cancer that strikes infants in their first year is a genetically distinct type of leukemia that might someday be treatable with designer drugs specifically targeted to it.

The discovery will help researchers single out the abnormal genes that make the leukemia so difficult to treat, as well as identifying new targets within the cancer cells for drugs that would be more selective than those used today. Nature Genetics will publish the findings on its Web site on Dec. 3 and in its January 2002 print edition.

"This finding is very exciting to us because it forces us to think about this as a separate disease and to think about other therapies" that could be designed to attack its specific weak points, says Scott Armstrong, MD, PhD, lead author of the paper.

Currently, physicians diagnose and treat the rare cancer as a particularly aggressive form of the more common Acute Lymphoblastic Leukemia, or ALL. The cancer may respond to chemotherapy at first, but tends to recur fatally. That's why the prognosis is so much worse than for most types of childhood leukemia, which today can usually be treated effectively.

The Dana-Farber scientists and collaborators propose the name Mixed Lineage Leukemia, or MLL, for the newly identified disease, which affects fewer than 100 hundred babies annually in the United States but is typically fatal in about 60 percent of the cases.

Using powerful "gene chips," the researchers say they've found that the genetic abnormalities in the cells of this infant cancer are dramatically different from ALL, and also from AML (Acute Myelogenous Leukemia), the other common type of childhood leukemia. The pattern of gene activity in the infant cancer is so distinct that the disease warrants its own name and might be tamed with drugs designed to hit its unique set of vulnerable points.

Armstrong, who is an instructor in pediatric oncology at Dana-Farber and Harvard Medical School, worked with Dana-Farber's Stanley Korsmeyer, MD, and Todd Golub, MD, on the project.

Golub, who also is affiliated with the Whitehead Institute/MIT Center for Genome Research in Cambridge, is a pioneer in using data from the Human Genome Project to obtain genetic "profiles" of different cancers - that is, determining which genes in the cancer cell are active and which are not - as a way of classifying and diagnosing them. Traditionally, scientists have categorized cancers on the basis of what body organs they originated in and how their cells appear under the microscope. The genetic profiles, or "signatures," not only could provide a new system of classification but also spells out the particular genetic abnormalities in each cancer.

The gene activity pattern in the infant cancer cells was captured by so-called gene chips that can determine the "off" or "on" status of thousands of genes in the cell - about one-third of all the cell's genes. The chips, glass or silicon wafers with 12,600 DNA segments attached to them, are used to probe genetic material (RNA) from cancer cells to determine which of the genes are being expressed.

In their study, the scientists found that about 1,000 genes were silent or underactive in the MLL cells compared to cells from patients with conventional ALL, while about 200 genes were overactive compared to ALL. "Thus, MLL shows a dramatically different gene expression profile from conventional ALL," they wrote.

That was not the only important difference. Leukemias involve blood cells that haven't reached full maturity, and the researchers found that the cells responsible for MLL are stuck at an earlier stage than those comprising ALL or AML.

"The message here is that when a leukemia behaves differently, doesn't respond to chemotherapy in the same way others do, and has this different genetic profile, then it should be recognized as a novel type of leukemia," said Armstrong.

The gene snapshots also singled out a single gene that may prove critical in causing the leukemia's uncontrolled growth. It "jumped off the page" in the data analysis, said Golub. The gene is called flt-3 and it makes an enzyme known as a tyrosine kinase that fuels cell growth. In MLL, the flt-3 gene is stuck in the "on" position and may contribute to the defective cell reproducing without limit.

The gene may represent a vulnerable point in the cancer cell that specifically targeted drugs could exploit. "So if you can block flt-3 you might be able to treat MLL," says Armstrong.

"This is a very, very exciting observation," said Stephen Sallan, MD, chief of the medical staff at Dana-Farber and an author on the paper.

About 3,000 to 4,000 children are stricken with leukemia annually in the United States, and 80 percent of them have ALL, the great majority of whom can be cured with chemotherapy. The remainder have AML. The great majority of children with ALL can be cured with chemotherapy. Dana-Farber has the highest cure rate in the world for ALL, well over 80 percent. But a minority of ALL patients have a much worse outlook. Among these high-risk cases, which amount to a few hundred babies a year, a small number develop the blood cancer as newborns or within the first year of life, and have such a stubborn form of disease that only about 40 percent of the infants survive.

What stands out about most of these infants - at least 70 percent - is that their blood cells are the scene of a chromosomal accident. In the majority of cases, portions of their 4th and 11th chromosomes have broken off and swapped places. As a result a gene known as MLL is altered, leading to the runaway growth that typifies cancer.

A similar finding in another rare cancer, Chronic Myelogenous Leukemia, or CML, led to the development of a drug to block the overactive enzyme. That drug, Gleevec, has been markedly successful and with few side effects in treating CML.


Other authors of the paper include Jane E. Staunton of the Whitehead/MIT genome center, Lewis B. Silverman, MD, of Pediatric Oncology and Harvard Medical School; Eric S. Lander of the Whitehead/MIT genome center, and collaborators in the Netherlands and Toronto.

The research was funded by the National Institutes of Health, the American Society of Hematology, the Belfer Cancer Genomics Center, and Bristol-Myers Squibb, Millennium Pharmaceuticals and Affymetrix.

Dana-Farber Cancer Institute ( is a principal teaching affiliate of the Harvard Medical School and is among the leading cancer research and care centers in the United States. It is a founding member of the Dana-Farber/Harvard Cancer Center (DF/HCC), a designated comprehensive cancer center by the National Cancer Institute.

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.