Vascular Niche Nurtures Brain Tumor Stem Cells

"The finding that brain CSCs exist in protective vascular (blood vessel) niches helps explain the origin of brain tumors and suggests a new strategy for eliminating them," said Richard Gilbertson, M.D., Ph.D., co-director of the Neurobiology and Brain Tumor Program at St. Jude. Gilbertson is senior author of a report on this work that appears in the January issue of Cancer Cell.

"Our data indicate that brain CSCs are nurtured by these vascular niches and that disrupting them blocks tumor growth by removing CSCs from tumors," he said. "These niches might also protect CSCs from chemotherapy drugs and irradiation therapy. So our findings could explain why aggressive tumors rapidly produce new blood vessels and why brain tumors reappear following treatment."

The St. Jude team first determined that CSCs are located in vascular niches by identifying cells carrying a protein called Nestin that marks stem cells (Nestin+ cells) in four types of brain cancer removed from patients: medulloblastoma, ependymoma, oligodendroglioma and glioblastoma. They found that tumors with the densest system of tiny blood vessels contained the greatest number of Nestin+ cells, and that Nestin+ cells are located next to blood vessels in brain tumors.

The investigators then examined thin sections of brain tumors and found that more than one-third of the Nestin+ cells next to blood vessels in the vascular niches had a mutation known to be linked to cancer, which suggested they were CSCs, Gilbertson said. About 30 percent of these cells were multiplying abnormally and rapidly, as expected for cancer cells.

The team showed in mouse models that CSCs from brain tumors have a more natural tendency to associate closely with blood vessels than do non-CSC tumor cells. The researchers also demonstrated in test tube experiments that CSCs bind closely to cells isolated from human blood vessels. Further, the investigators found that human blood vessel cells release molecules that trigger brain CSCs to keep their identity as stem cells and continue to multiply rapidly.

"This is strong evidence that the cells making up the vascular niche send signals to CSCs in the brain, causing them to grow and multiply," Gilbertson said.

Gilbertson's team also studied the interaction of blood vessel cells with CSCs using mouse models of brain cancer. Mixing brain CSCs with human blood vessel cells dramatically increased the formation and growth of tumors. Brain CSCs inserted into the brain without human blood vessel cells produced tumors slowly, reaching a maximum size after seven weeks. In contrast, tumors formed by mixtures of brain CSCs and blood vessel cells grew much more rapidly, reaching a maximum growth after only four weeks. The blood vessel cells did not increase tumor growth by forming new vessels, but by associating with CSCs and stimulating these directly to produce tumors.

Finally, the investigators showed that increasing the numbers of blood vessels in mouse models of brain tumors markedly increased the numbers of CSCs in tumors. The scientists also showed that drugs that deplete blood vessels from tumors inhibit tumor growth by reducing the number of CSCs. For example, the team depleted blood vessels in tumors with Avastin® (bevacizumab), an anti-angiogenic drug that blocks a protein called VEGF. Anti-angiogenic drugs block the formation of new blood vessels.

"This strongly suggests that disrupting the blood vessels in brain tumors might block tumor growth by disrupting brain CSC niches," Gilbertson said. "This is important since the mechanism by which anti-angiogenic drugs, like Avastin, block tumor growth is largely unknown. Our data suggest a previously unrecognized way that anti-angiogenic agents inhibit tumor growth."

The St. Jude investigators have now translated these findings into a clinical trial to determine the effectiveness of Avastin and another drug, Traceva® (erlotinib), in eliminating tumors and preventing their recurrence in children with brain cancers.

The other authors of this study include Christopher Calabrese, Helen Poppleton, Mehmet Kocak, Twala L. Hogg, Christine Fuller, Blair Hamner, Eun Y. Oh, M. Waleed Gaber, David Finklestein, Meredith Allen, Adrian Frank, Ildar T. Bayazitov, Stanislav S. Zakharenko, Amar Gajjar and Andrew Davidoff.

This work was supported in part by the Sontag Foundation, The National Cancer Institute, the V Foundation for Cancer Research and ALSAC.