Two recently published studies may help explain why glioblastoma, one of the most aggressive types of human cancer, is so difficult to treat. According to the studies, some glioblastoma cells can differentiate into endothelium cells, which line the interior of blood vessels, thereby forming vascular networks that help to feed the tumors.

Researchers have previously noted that glioblastoma tumors have extensive, abnormally structured blood vessels that consist of large, rapidly dividing cells. It was thought that these blood vessels were derived from normal brain vascular tissue, but no one had taken a close look, as Viviane Tabar, of Memorial Sloan-Kettering Cancer Center in New York, told The Scientist.

Tabar and colleagues first analyzed endothelial cells in glioblastoma samples for chromosomal abnormalities (Nature doi:10.1038/nature09624; published online 21 November 2010). They found that some of the endothelial cells within these samples had the same mutations as those found within the tumor cells. The team then used antibodies that stick to specific proteins to separate the glioblastoma cells. Further analyses showed that stem-like glioblastoma cells give rise to endothelium cells that constitute blood vessels. Tabar and her team injected these human cancer cells into the brains of immunodeficient mice. They found that the resulting tumor implants consisted of blood vessels of human origin, meaning that they were derived from the stem-like cancer cells.

In Italy, Ruggero De Maria of the Istituto Superiore di Sanità in Rome, Roberto Pallini of the Catholic University of Rome and colleagues were working on the same issue (Nature doi:10.1038/nature09557; published online 21 November 2010). They found that between 20% and 90% of endothelial cells in glioblastoma tumors were derived from the tumors. The team isolated the tumor cells that could become endothelial cells and injected these stem-like cells into immunodeficient mice. They found that the majority of the blood vessels in the resulting tumors were of human, instead of mouse, origin.

These studies might help to explain why a drug used to inhibit angiogenesis, or blood vessel growth, has limited effectiveness against glioblastoma. Further in vitro experiments carried out by Tabar's group showed that exposure to this drug inhibited maturation of tumor endothelial cells but was unable to inhibit the formation of blood vessels. On the other hand, exposure to an inhibitor of Notch signaling, which controls multiple cell differentiation processes, stopped the differentiation of the stem-like cancer cells into endothelial progenitors. Further research could possibly lead to the development of drugs that block this differentiation process.