Growing Evidence Supports Stem Cell Hypothesis of Cancer

BETHESDA, Maryland—During the past 18 months, researchers have developed substantial evidence supporting the notion that stem cells play a critical role in the development of at least some cancers, their progression, and the prognosis of patients, including breast, brain, lung, and prostate cancer, multiple myeloma, and melanoma.

"The idea of stem cells in cancer is a very old one, but it is only recently that scientists have had experimental models that actually validate this," Max Wicha, MD, professor of internal medicine and director of the University of Michigan Comprehensive Cancer Center, Ann Arbor, said at a meeting of the National Cancer Advisory Board (NCAB). "It represents a paradigm shift in how we need to approach cancer because it has very wide clinical implications."

The stem cell hypothesis challenges the classic stochastic model, which holds that cancer results from a random mutation in a cell that reproduces and eventually forms a malignant neoplasm. In contrast, the stem cell model suggests that in many instances, stem cells or their immediate progeny are the cells transformed during carcinogenesis, and that only these cells are capable of self-replication within a tumor. All other cells in a cancer have lost their ability to self-renew and are in various stages of differentiation.

Moreover, if these differentiated cells should escape the primary tumor and travel to other parts of the body, they would not grow metastases of clinical consequence. This means that once a cancer develops, its growth is driven by a small number of cells—perhaps as few as 100—which have the two distinguishing properties of stem cells, namely the ability to make exact copies of themselves and to differentiate.

In the Michigan team's scenario, a stem cell that reaches a distant site might settle in a microenvironment that fails to support its immediate proliferation. This could explain the dormancy of tumors and their late emergence when the environment becomes right to put the cell back into cycle.

"If the stem cell model is correct, then we have to reexamine, in a very critical way, the preclinical models for therapeutic development," Dr. Wicha said. "We have to look at the endpoints for clinical trials, which may not be adequate because the tumor stem cells may be resistant to these therapies. And we think effective therapies will need to target the tumor stem cell population while sparing normal stem cells. Our laboratory and others are working on potential strategies to target this cancer stem cell population." Dr. Wicha is one of the founders of OncoMed Pharmaceuticals, a California-based biotech company that is developing technology to target cancer stem cells.

Five years ago, Canadian scientists suggested that stem cells were key players in leukemia. Dr. Wicha, working with Michael F. Clarke, MD, now at the Stanford University Comprehensive Cancer Center, began investigating the role of stem cells in breast cancer.

Dr. Wicha and his colleagues have isolated stem cells in the breast and characterized how they change as normal breast tissue becomes cancerous. They have concluded that the cells involved are mammary stem cells whose ability to differentiate is limited to the cell types that occur in the breast. These cells appear to undergo mutation as the result of a genetic instability or exposure to some damaging environmental cause, which, in turn, destabilizes the process that regulates the cells' self-renewal ability.

"More importantly for therapeutics, there is now emerging evidence that both normal stem cells and these transformed counterparts are highly resistant to our therapies, which has lots of implications," Dr. Wicha said. "Even the metastasis of cancer is probably related to the homing of stem cells, in that both normal and transformed stem cells use very similar receptors as they spread to distant sites."

The Michigan researchers employed cell markers to eventually identify two variations of a cell designated B38+CD44, an extracellular matrix receptor, that differed by having or not having the marker CD24, an adhesion molecule. Using NOD/SCID mice, the researchers made a surprising discovery. Mice injected with as many as 20,000 CD24-positive cells from human breast tumors failed to develop cancers. However, mice injected with CD24-negative cells all developed cancerous tumors within 12 weeks, even with as few as 200 of the cells injected. "If you take unsorted cells, you have to put in about 50,000 cells to get tumors," Dr. Wicha said.

Flow cytometry studies of cells positive and negative for CD24 have ruled out the possibility that the tumors that developed in the animals resulted from the inadvertent selection and injection of a highly metastatic clone, he said.

Investigators have identified several signaling pathways, eg, Notch, Hedgehog, Bmi-1, and Wnt, believed to be involved in stem cell self-renewal and tumorigenesis, which have yielded insights into the normal and abnormal regulation of the cells. Researchers have begun sorting out how the pathways may go awry and produce an expanded pool of stem cells that, in turn, provides additional targets for further transforming events. In studies of the Hedgehog pathway, Dr. Wicha and his colleagues have created human stroma in mice using human mammary stem cells obtained from tissue removed from women during reduction mammoplasties. When they perturbed the pathway, the mice developed human ductal carcinoma in situ, which suggests that a skewed Hedgehog signaling pathway plays a role in early breast cancer and may provide a target for halting or reversing a cell's momentum to malignancy.

The Michigan team also has identified Bmi-1, an important transcription factor, as one pathway that appears to regulate the unique renewal and differentiation properties of stem cells in normal and cancerous stem cells. Now they are trying to determine how its deregulation might contribute to the growth and spread of cancer.

"These stem cells have a very different genetic profile in terms of gene expression than do the vast majority of cells that form the bulk of the tumor," Dr. Wicha said. "What is clear is that patients who had the so-called stem cell profile had a very poor survival in comparison to those who did not have the stem cell phenotype."

Other researchers have made the same finding in other cancers, "suggesting a commonality of stem cells in a variety of cancers that carry this prognostic indication," he added. "The molecular profiling studies, which have been quite exciting, are actually telling us what the stem cell of the tumor is and the differentiated progeny that is produced."

Implications for Oncology

The stem cell model, if confirmed, carries enormous implications for oncology. In Dr. Wicha's view, identifying and eliminating mutated stem cells, or forcing them to differentiate, may one day become an important prevention strategy. Moreover, the genetic profiles may reflect the genetic make-up of the stem cell from which a tumor arose.

From their work, he and his colleagues have proposed that basal breast cancers, which have a poor prognosis, result from a stem cell that is estrogen-receptor (ER) negative. ER-positive breast tumors may have either a bad or good prognosis. ER-positive luminal B breast tumors, which have a very poor prognosis, appear driven by a stem cell that itself is ER negative. The stem cell driving luminal A tumors and their good prognoses apparently is ER positive.

Thus, effective therapy and, perhaps, prevention would need to focus on the stem cell underlying a cancer. Killing a tumor's differentiated cells but not its driving stem cells may explain why patients have tumor shrinkage that has no impact on their survival. "We need to be targeting the stem cell population," Dr. Wicha said. "And we believe that in certain tumors, like testicular cancer, that are curable by chemotherapy, that is exactly what happens."