By David N. Leff

Q: Why does a solid tumor resemble an unhealed wound?

A: "They both display angiogenesis - formation of new blood vessels - as required for replacement tissue to grow properly."

So said molecular geneticist Brad St. Croix, at the Johns Hopkins Oncology Center in Baltimore. "Our studies show," he observed, "that the molecules turned on in these new tumor-induced arterioles, venules and capillaries are also turned on in vessels that take part in healing wound injuries."

St. Croix, a postgraduate research fellow at the Oncology Center, is first author of a paper in today's Science, dated Aug. 18, 2000, titled: "Genes expressed in human tumor endothelium." Kenneth Kinzler and Bert Vogelstein, founders and co-directors of the center, are the paper's senior authors.

The research they report presents for the first time, St. Croix told BioWorld Today, "fundamental differences between the endothelial cells lining healthy and malignant blood vessels." He continued, "I think of it as the most comprehensive analysis of tumor endothelium to date, and I believe it has promise - at least in the long term - for making antitumor therapeutics and diagnostics. This is a necessary first step," he added, "in order to capitalize clinically on the whole concept of tumor anti-angiogenesis."

That concept drove the media and Wall Street into a feeding frenzy in 1997 and 1998, over a novel drug, endostatin, which aborted tumor angiogenesis in mice by crimping the growth of the endothelial cells lining their blood vessels. "That frenzy is still going full force in the scientific community," St. Croix observed. "It's definitely a very hot topic. In terms of the preclinical trials that are now ongoing with endostatin - in mice, not people - I think the jury is still out."

When St. Croix and his co-authors set out to seek fundamental features distinguishing tumorous from healthy endothelium, he recalled, "it wasn't clear that we would find any differences at all. The reasons we thought there might be is that blood vessels in tumors look morphologically distinct from normal vessels in the body. They're very erratic and tortuous in their patterns because they go every which way. The blood that flows through them is often static; it's just not moving along. It's hard even to classify them as arterioles or venules because they just don't have the appearance of normal vessels. These morphological differences caused us to think that new vessels are probably distinct at the molecular level from normal ones."

Piratical Tumors Kidnap Innocent Cells

St. Croix made the point that "the angiogenic blood vessels that infiltrate and grow into the tumors are not themselves malignant. They're actually normal host cells that the tumor pirates, in order to expand. A solid tumor is thought not to be able to grow beyond the size of a pea - 2 to 3 millimeters - without inducing a steady supply of new blood.

"The network of new blood vessels feeds the tumor," St. Croix related, "allows wastes to escape from it, and also provides a route for cancerous cells to escape and metastasize to different organs. But the endothelium lining those vessels," he went on, "is a very normal cell type. Its DNA is stable, and it's diploid - in sharp contrast to the tumor cells, which are genetically unstable, and changing their genetic makeup all the time.

"The gene expression study we report in Science allowed us to characterize all the genes in these two cell types, and we found that they were differentially expressed. In fact, there was a significant number of them that we could find by using a screening technique called SAGE - serial analysis of gene expression." Kinzler, Vogelstein and a third co-author, molecular geneticist Victor Velculescu, invented SAGE, which the University has licensed to Genzyme Molecular Oncology - a subsidiary of Genzyme Corp., in Framingham, Mass. (See BioWorld Today, Sept. 11, 1998, p. 1)

SAGE counts cellular messenger RNAs as a measure of gene expression. The co-authors started by analyzing endothelial cells isolated from both normal and tumor tissue of a patient with colorectal cancer. They focused initially on this tumor type because of its high incidence of mortality and frequent resistance to treatment.

The team used a panel of antibodies to separate out cells from the tissues, then enlisted their SAGE computerized technology to screen 200,000 pieces of genetic material from those cells. This assay identified 46 genes that were overexpressed up to 10-fold in tumors as compared to normal endothelium, and 33 genes underexpressed.

Diagnosis, Therapy: Far Off But Real

Now that SAGE has pulled these genes out of the closet, St. Croix envisages three modes of applying them clinically - diagnosis, chemotherapy and imaging.

"One can imagine a very simplistic but not unreasonable way of thinking about this," he suggested. "If something is turned on and expressed by a gene on the surface of these tumor endothelial cells, you can envision injecting intravenously some drug that's linked up to an antibody that recognizes this particular molecule, homes in on it, and kills the cells that express it. Presumably those would be the tumor blood vessels, and without them to feed it, the tumor would collapse."

As for diagnostic potential, he went on, "If the new vessels are secreting factors into the bloodstream, you might be able to measure in the blood whether or not angiogenesis is ongoing. If it is when it shouldn't be, that would lead you to suspect a tumor there.

"Another form of diagnostics," St. Croix proposed, "is imaging. Say you put a tracer linked to an antibody - just like the therapy - but in this case the tracer would hook up to something on the surface of the endothelial cells in tumors that would allow you to scan and visualize those tumors under the microscope."

He observed, "This clinical potential might be useful in a variety of tumor types - not just colon. We've looked at expression in breast, brain, lung and pancreas. This is a very early study," St. Croix concluded, "but we want to get it out there so everyone else can help us to figure out which ones are going to be useful."