If, as many scientists believe, tumours come from cancer stem cells, then where do cancer stem cells come from? Led by Joerg Huelsken at the Ecole Polytechnique Federale de Lausanne, a group of researchers believe that they have answered that question, at least for some skin cancers1. In doing so, they may have identified a way to attack cancer stem cells without harming healthy ones.

Skin is maintained, in part, by bulge stem cells residing in hair follicles. The researchers chemically induced cutaneous tumours in mice, and looked for cells the carryin markers CD34, which is associated with bulge stem cells. Compared to skin tissue, tumours contained 9 times as many cells with this marker. Closer examination of cells isolated from tumours showed that they also had additional markers of bulge stem cells. If transferred from one mouse to another, cells displaying the marker CD34 were 100 times more able to initiate secondary tumours than were unsorted cells. These cells could even start tumours when transferred a third time. Cells lacking this marker never produced tumours. Lineage tracing studies also showed that bulge stem cells were contributing to tumours.

Next, the researchers looked at signaling within the cells. In particular, they wanted to see if a signaling pathway that controls how skin stem cells differentiate was necessary for the cancer stem cells they'd identified to form tumours. They found that it was. Activity of a protein called beta-catenin is periodic in normal stem cells but was consistently strong in cancer stem cells. The researchers engineered cells that could delete the beta-catenin gene on command, and found that when the gene was deleted, the tumours regressed completely. Moreover, the proportion of cells displaying CD34 shrank dramatically. Cells that displayed CD34 but lacked the beta-catenin gene were unable to initiate tumours.

Next, the researchers looked to confirm their results with squamous cell carcinomas from humans. The human carcinomas showed high beta-catenin activity, but did not express CD34, so the researchers could not use it as a marker to isolate cancer stem cells. However, they did check to see whether blocking beta-catenin signalling with shRNA would have an effect. It did. Human tumours transplanted onto mice shrank when beta catenin activity was blocked. Further, areas on tumours that did show growth also tested positive for beta-catenin activity, indicating that the gene had not been completely knocked down in these cells. Thus, skin cancer stem cells seem to require beta-catenin and normal skin stem cells do not. That difference could, one day, be used to develop drugs that kill cells that lead to tumours but not healthy skin.