Two other scientists in Folkman's lab, Oliver Kisker and Steven Pirie-Shepherd, later confirmed O'Reilly's finding when they saw the same protein being produced by human pancreatic cancer cells.

By implanting human tumors into mice and comparing tumor growth, scientists can try to identify which types of human tumors could possibly be used in future to inhibit cancers, according to the Science article.

Folkman's lab has previously reported successful experiments using the anti-angiogenic proteins endostatin and angiostatin on tumors in mice. In those experiments, the substances were shown to cause tumors in mice to regress to microscopic size.

These results have been duplicated by scientists at the National Cancer Institute, according to its recent press release.

Researchers at Bristol Myers Squibb and other biotechnology companies, however, say they are unable to reproduce the results, stirring some controversy over the validity of Folkman's methodology.

Mallinckrodt Professor of Chemistry George M. Whitesides '60 suggested that disparity between labs' results may stem from the different ways they are trying to relate the vascular system's growth to tumor growth.

"Folkman has his approaches to this problem; other research groups and companies have other approaches," he said. "But the point is that the concept of inhibition of angiogenesis is a fundamentally very important one, both because it seems to me very likely that it will work, at least in some