The usual path for medical research - with luck - leads from mice to humans. But in the June 2007 issue of Cancer Cell, a group from the National Cancer Institute that began its work with humans goes in the opposite direction, searching for molecular markers in mice that distinguish tumor blood vessels from regular ones.

The Cancer Cell study, which compared gene expression signatures from normal and tumor blood vessels, builds on a paper from 2000 that compared normal to tumor blood vessels in humans. (See BioWorld Today, Aug. 18, 2000.)

That paper, published in the Aug. 18, 2000, issue of Science, found nearly 80 genes that were more highly expressed in colorectal tumor tissues than adjacent noncancerous colon tissue from the same patient. But it suffered from a necessary limitation of studies on human tissues. "You're always limited to normal tissue that's right adjacent to the tumor," Brad St. Croix told BioWorld Today.

And that limitation has two potential problems. For one, tissue right adjacent to a tumor may not be typical in all respects. For another, it's impossible to tell whether the specific markers for tumor blood vessels are in fact expressed in other parts of the body.

St. Croix, who is first author of the Science study and was a postgraduate researcher at Johns Hopkins University at the time it was published, has since gone on to establish his own laboratory at the National Cancer Institute.

In the Cancer Cell paper, he and his team went to mice to compare gene expression patterns in a wider variety of tissues.

In their research, the group took advantage of the respectable regenerative capacity of the liver: Cut out two-thirds of a mouse's liver, and it will regrow it in less than a week. That made it possible to compare the blood vessels of normal mature tissues to both new blood vessels forming after an injury and those forming to supply liver tumors. The scientists also compared the gene expression of the blood vessels of seven other normal resting organs.

St. Croix, who is the senior author of the new paper, and his group found 13 markers that were expressed at least 10-fold higher in the endothelial cells, which line blood vessels, of tumors than in those of normal or regenerating livers. Seven of them were found to be cell surface receptors, making them the most promising in terms of potential therapeutic development, which St. Croix said he is "absolutely" interested in, though he also cautioned that the studies represent the earliest stage of any pharmaceutical endeavors.

As a final step, the scientists went back from mice to men, and tested the most differentially expressed cell surface marker, CD276, on a variety of human tumor samples. They found that CD276 was expressed on tumor blood vessels - but not just there; many tumor cells express the marker as well. Normal epithelial cells, in contrast, never stained positive for CD276.

St. Croix said the discovery that CD276 is expressed on tumor as well as tumor vessel cells makes it "an even more exciting target" because if clinical targeting pans out, any agents that selectively bind to CD276 would target tumor cells as well as blood vessels.

"It was unexpected to us," he said. "But we see it as a bonus."