Doctors usually measure how well a cancer treatment works by how much tumor it destroys. The more cancer cells killed the better. The eradication of every last mutated cell is ideal, because nearly every tumor cell can regenerate the disease. Or so cancer biologists believed.
A radically different idea is transforming cancer research. In this view, a relatively few dangerous cells fuel cancer growth. The implications are profound: find and kill this minority of crucial cells, and the rest of the tumor may languish or even self-destruct. This popular notion is called the cancer stem cell hypothesis.
A new study of human breast cancer tissue challenges this hypothesis in solid tumors. The presence of purported breast cancer stem cells in the primary tumors increased the risk of distant metastases, the researchers found, but the metastases were packed with more differentiated non-stem cancer cells. An experimental molecular therapy that shut down the “stem cells” in culture did not affect the more differentiated cells.
“Patients are killed by the metastases composed of cancer non-stem cells, even though it could have been the cancer stem cells that metastasized,” said senior author Kornelia Polyak, HMS associate professor of medicine at Dana–Farber Cancer Institute. Both groups of cancer cells are bad, she and her co-authors conclude in the March Cancer Cell, with distinct molecular pathways requiring differently targeted drugs.
The new study is not the only thing that should give cancer researchers pause about the cancer stem cell hypothesis, Polyak said. Enthusiasm has outpaced supporting data and may be coloring scientists’ interpretations of results, she said. One reviewer criticized her paper for not replicating results from injecting human tumor cells into mice.
“How can I answer that?” she said. “To me, the gold standard is what’s happening in patients. We have to use models, but they have to be validated and correlate with what’s happening in cancer patients. It’s important we figure out what’s going on in patients.”
Roots of a Concept
The cancer stem cell concept dates back to 1875, but convincing molecular evidence first arose in leukemia 10 years ago. Researchers wondered if blood cancers are organized in a hierarchy like the components of the healthy blood system.
John Dick of the University of Toronto made the first rough cut separating leukemia stem cells from other leukemia cells. He and his colleagues sorted out human cancer cells using fluorescent-tagged antibodies for stem cell markers and high-speed flow cytometry. He injected the different groups into immunodeficient mice and then transplanted those cancer cells into secondary recipients. Cancer stem cells reconstituted tumors, including the more differentiated cells; the others did not. That assay remains the accepted rigorous standard of proof for identifying cancer stem cells.
Leukemia stem cell research builds on the foundation of 30 years of careful scientific work to characterize the development and continuous replenishing of healthy white blood cells, red blood cells, platelets, and more from normal hematopoietic stem cells in the bone marrow, said Gary Gilliland, director of the cancer program at the Harvard Stem Cell Institute.
“I don’t think there is any question that cancer stem cells exist,” he said. The question is how broadly applicable the paradigm is to all solid tumors.
“It’s an extremely important concept,” said Gilliland, also a Howard Hughes investigator and an HMS professor of medicine at Brigham and Women’s Hospital. “If you take leukemia as the paradigmatic example, we can induce complete remission in well over 80 percent of adults with acute myelogenous leukemia, as measured by sophisticated molecular techniques that show no evidence that disease persists. Yet 80 percent ultimately will relapse if not treated definitively with a bone marrow transplant, which is risky but effective if the person is young enough and has a good donor. It seems so obvious in retrospect that those agents we use don’t target the leukemia stem cell ultimately responsible for the relapse.”
Even the latest targeted molecular anticancer drugs, such as Gleevec, which can better banish certain cancers and prolong survival with far fewer side effects, do not usually cure the disease. Now researchers think that these drugs are targeting the tumor bulk and not the tumor stem cells.
From Blood to Solid Tumors
In the past five years, more evidence has accumulated for leukemia stem cells. In chronic myelogenous leukemia, the cancer actually originates with a chromosome relocation in a hematopoietic stem cell. But in other leukemias, cancer stem cells result from other cells that accumulate enough mutations to activate part of the stem cell program in those cells, said Scott Armstrong, HMS assistant professor of pediatrics at Children’s Hospital Boston and DFCI. Last summer, he and his colleagues, including Gilliland, showed that one oncogene mutation was enough to convert a committed progenitor cell into a cancer stem cell. Armstrong is searching for the origin of cancer stem cells and wants to learn about any molecular differences that will allow anticancer drugs to target cancer stem cells without harming healthy stem cells.
Research in cancer stem cells of solid tumors is lagging in part because scientists do not understand development in other organ systems as well as they do in blood. Yet researchers have reported cancer stem cells in some brain, breast, and colon cancers, based on the mouse assay. And one of the goals of the Harvard Stem Cell Institute is to prospectively purify true cancer stem cells from the full spectrum of solid tumors, Gilliland said.
In 2003, Michael Clarke’s group at the University of Michigan published the first evidence of stem cells in solid tumors, using human breast cancer tissues. In their new study, first author Michail Shipitsin and colleagues in the Polyak lab probed the molecular profile of the putative breast cancer “stem cells,” called CD44+ cells, and a less tumorigenic group, known as CD24+ cells.
Polyak’s team found that a stronger CD44+ cell gene expression signature in tumors predicted a higher risk of distant metastases in patients. “It is a more aggressive tumor cell, more angiogenic, and more invasive,” Polyak said. Surprisingly, they found, metastases were highly populated by the more differentiated CD24+ cells.
In cultured CD44+ cells, an experimental molecular therapy targeting the TGF-beta pathway was effective, but the same drug did not work on the more differentiated CD24+ cells making up much of the metastases.
The drug is poised for phase 1 clinical safety testing. One immediate implication is that the first round of pending clinical trials, usually conducted on people with metastasized cancer, may miss the most effective timing for the drug, Polyak said.
More broadly, this study “says things may be more complicated than a single kind of cancer stem cell,” said Whitehead Institute cancer biologist Robert Weinberg. The cancer stem cell hypothesis remains plausible, Weinberg said. “What’s important is the notion that these cells need to become the new targets for killing, not the cells forming the bulk of the tumor.”
Polyak believes cancer cells in solid tumors can be sneakier and that they can change phenotypes as conditions dictate. “Cancer cells continuously have different genetic changes and evolve,” she said. “They can reactivate progenitor pathways or can differentiate, depending on what’s better for them at the time. From a clinical point of view, all these arguments become semantic. Whether or not you call them stem cells is not important. What’s important is how to target cancer cells and prevent recurrence and progression.”