Canadian scientists have found cancer stem cells that seem to fuel the uncontrolled growth of malignant brain tumors.

If they could be targeted with a drug, these troublemaking cells might hold the key to successful treatment of an aggressive and deadly cancer.

The finding appears in the Nov. 18 issue of Nature.

"It's very preliminary, but very interesting," said Dr. Robert Aiken, director of medical neurological oncology at Continuum Cancer Centers in New York City. "It basically demonstrates a way of discriminating within a tumor those cells that perpetuate the tumor from those which just basically take up space."

The replicating cells, Aiken added, are the ones that doctors would want to destroy.

Cells with stem cell properties had already been found in leukemia and in breast cancer. They represent only a fraction of the total number of cells making up the tumor, though.

"These cancer stem cells are probably the most primitive or most embryonic of the cells in a tumor, and replicate and go on to cause all of the mischief and damage that tumors cause," Aiken said.

The cells "have nothing to do with stem cells that we work on to transplant or to cure Parkinson's disease (news - web sites)," added Dr. Viviane Tabar, an assistant professor of neurosurgery and a stem cell scientist at Memorial Sloan-Kettering Cancer Center in New York City.

In this study, researchers from the University of Toronto used information gleaned from previous laboratory experiments to distinguish replicating cells from non-replicating cells in human tumor samples. The replicating cells have a specific marker on their surface that identifies them. These cells, known as CD133 cells, were then transplanted into the brains of mice. The cells caused deadly brain tumors -- medulloblastomas and glioblastomas -- to grow in 16 of 19 rodents.

The research helps provides a new framework with which to think about cancer, Tabar said.

"We have been thinking of cancer -- and particularly brain tumors -- as a large conglomerate of cells that you try to treat," Tabar explained. "We're starting to think that the tumor is not really organized like this but is a hierarchy, whereby there are cells that matter a lot more than others even though they all look the same."

The next step, she added, would be to determine if there is a pathway that is associated with the cell that could be targeted with a medication.

"We know that these surface markers exist on the surface of the self-perpetuating cancer stem cells, and it certainly does provide a means of discriminating those stem cells from the rest of the tumor," Aiken said. "If it's possible to generate a molecule of some sort to identify these clone cells, it may be possible to selectively destroy the tumor. But it remains entirely an empirical effort to determine if that can be done safely."

"It is always exciting to achieve a better understanding of how a tumor is formed and how it evolves," Tabar added. "I'm sure, from the patient's perspective, it would be a lot more exciting if we can associate it with treatment."