The leukemia stem cells isolated proved to be surprisingly different from normal blood stem cells -- a finding that may be good news for developing a drug that selectively targets them.

Cancer stem cells are self-renewing cells that are likely responsible for maintaining or spreading a cancer, and may be the most relevant targets for cancer therapy. The discovery provides answers to the longstanding questions of whether cancer stem cells must be similar to normal stem cells, and what type of cell first becomes abnormal in leukemia, the most common form of cancer in childhood. The journal Nature has posted the study's findings online in advance of print publication.

It had been speculated that leukemia begins in a totally undifferentiated stem cell that can become any type of specialized blood cell and has the ability to renew itself almost without limit.

Instead, the scientists showed that they could create leukemia stem cells, which also are self-renewing, from partially committed, non-self-renewing progenitor cells. The latter are short-lived cells that can turn into several types of blood cells, but are more committed than stem cells, which can become any kind of blood cell and also are virtually immortal.

"Our data supports the idea that leukemia stem cells do not have to originate from normal blood stem cells. Furthermore, we have shown that fully developed leukemia stem cells do not necessarily have the same genetic program as normal stem cells," said Scott Armstrong, MD, PhD, of Dana-Farber and Children's Hospital and senior author of the paper. "This is an important finding, because it indicates that in the future we should be able to specifically target leukemia stem cells without killing normal stem cells."

Leukemias are cancers of the blood-forming tissues of the bone marrow in which white blood cells proliferate abnormally, with life-threatening effects. About 35,000 diagnoses of all types of leukemia will be made in 2006, according to the American Cancer Society, with about 22,280 deaths. Some forms of leukemia have a high rate of cure. In other forms, chemotherapy may initially put the patient's disease into remission, but after months or years the cancer reappears and may be fatal.

Many scientists believe that relapses are caused by the survival of a handful of leukemia stem cells mixed in with the population of cancer cells. These cells have gained self-renewal capabilities, and, if not killed by chemotherapy, can lie dormant in the bone marrow and eventually trigger new growth of the leukemia. Current thinking is that cure rates of leukemia and other cancers could be improved if the cancer stem cells could be identified and selectively targeted with designer drugs.

To test this hypothesis, the researchers sought to transform a normal, partially committed progenitor blood cell from a mouse into a leukemia stem cell, and then determine whether that stem cell was more like a normal blood stem cell or instead resembled the progenitor. As a first step, they inserted an abnormal gene, MLL-AF9, which causes a type of acute myelogenous leukemia (AML) in humans, into partially committed mouse blood cells known as granulocyte macrophage progenitors, or GMPs. These genetically altered cells were injected into mice, which subsequently developed AML.

Through several steps of purification, the researchers winnowed down the leukemia cells from the mice to a small population that contained a large percentage of leukemia stem cells -- as evidenced by the fact that they could induce cancer in normal mice using successively smaller amounts of cells, since only the stem cells cause the disease when injected. "Such a pure population of leukemia stem cells had not been isolated before," said Andrei Krivstov, PhD, of Children's Hospital Boston, the paper's lead author. "We are the first to transplant as few as four cells and induce leukemia in the mice."

The investigators next compared gene activity in the leukemia stem cells with that in the original partially committed progenitor cells, and in normal uncommitted blood stem cells. Using microarray technology, they compared the cells' gene expression patterns -- that is, which genes were turned on and which were turned off.

In terms of gene activity, "the leukemia stem cell looks most like the committed progenitor," said Armstrong, who is also an assistant professor of pediatrics at Harvard Medical School. "But there's a program of a few hundred genes that are turned on in the progenitor, which appears to give it the ability to self-renew. It's almost as if the abnormal gene we inserted knows what to do to turn on the program that makes it a self-renewing cancer stem cell."

The scientists referred to the gene activity pattern they discovered as a "signature" of self-renewal. Their next efforts will be to determine which genes among the several hundred that were particularly active or inactive are the most responsible for the cancer cell's behavior. These genes might eventually become targets for new types of drugs.

Moreover, said Armstrong, knowing the gene signature of an individual patient's leukemia might be useful in predicting how difficult it will be treat it and for evaluating the success of treatment. So far, researchers have not identified and isolated a pure population of leukemia stem cells in humans with the disease. The gene expression signature might be used to identify leukemia stem cells in the human disease, and the presence of a large number of leukemia stem cells could indicate a poor prognosis.

dfci.harvard