Cancer researchers can sequence the genomes of tumor cells, scan them for aberrant gene activity, dissect their mutated proteins and study their growth in lab dishes, but researchers have not been able to follow the process by which cells form tumors. Now three independent research groups have done so in mice. Their findings support the idea that a small group of cells drives tumor growth, and that a cure for cancer may require the removal of these so-called tumor stem cells. It is not yet possible to confirm whether these conclusions from the brain, bowel and skin cancer studies apply to other types of tumors, but Luis? Parada believes that if they apply to other tumors, “it would profoundly change the current criteria for evaluating the efficacy of chemotherapy and developing clinical therapies.” Instead of just looking at whether a particular therapy shrinks tumors, researchers will be more concerned about whether the right cells are being killed. Parada and his colleagues wanted to test whether a genetic marker that specifically identifies healthy adult neural stem cells could also identify cancer stem cells in neuroblastoma. They found that all neuroblastoma samples had at least a few labeled cells – presumably stem cells. Unmarked cells can be killed by standard chemotherapy, but tumors can recover quickly. Further experiments showed that the unlabeled cells originated from labeled cell progenitors. When the researchers combined chemotherapy with genetic means of suppressing the marker cells, Parada said, the tumors shrank dramatically to the level of “residual remnants. In another study, stem cell biologists at the Hubrecht Institute in Utrecht, the Netherlands, focused on the gut. Using a drug-driven fluorescein marker expression system, they demonstrated in mice that a number of different types of tumor cells were actually derived from the same stem cells. Moreover, these stem cells are the driving force behind tumor development. For the study of skin cancer, Blanpain and his group labeled individual tumor cells rather than specifically stem cells. They found that the cells exhibited two distinct patterns of division of labor: they either divided to produce a few cells before slowly running out, or they produced many cells. This reaffirms that a distinct class of cell subpopulations is the driver of tumor growth. The next step in the research program, the researchers say, will be to figure out how the cells tracked in these experiments are linked to the putative cancer stem cells identified through years of transplantation experiments. The researchers have been working feverishly to find ways to kill these cells; now they have more tools to test whether such a strategy would work.