Zeng and his colleagues conducted a series of experiments under carefully controlled laboratory conditions to determine whether, as they suspected, HGF stimulates head and neck tumor cells to release pro-angiogenic proteins. To their surprise, they found that was not the case. Tumor cells stimulated by HGF strongly promoted the formation of a capillary-like network compared with secreted factors induced by HGF alone. "At this point, we didn't have any idea of what was going on," said Wang. "But we started to think that it must involve the direct interaction between the tumor and endothelial cells."
Wang said that's where luck entered the picture. He and his colleagues decided to take a closer look at a vast body of data that they had generated a few years earlier showing thousands of genes that HGF activates in head and neck tumor cells. The gene that was among the most expressed is called jagged1, which is known to bind to a specific protein on the surface of endothelial cells. "I thought, 'Oh, this makes sense,'" said Wang. "The jagged1 protein is not secreted but is displayed on the surface of the tumor cells. I speculated that HGF induced jagged1 levels to increase, leading to a direct surface to surface interaction between the tumor and endothelial cells."
Wang's hunch also made good intuitive sense for another reason. The jagged1 protein bound in a hand-in-glove manner to a protein on the endothelial cells called notch. Other laboratories have shown that notch plays a key role during human development in forming blood vessels. Oddly, Wang noted, the possible role of notch in tumor angiogenesis has