Scientists at Harvard Medical School have determined for the first time the molecular structure of a protein that instructs the body to make blood vessels grow, opening the possibility of new ways to treat cancer.

By sending new capillaries into surrounding tissue, the protein is believed to play a key role for tumors; the capillaries supply the tumors with oxygen and nutrients needed for growth. It is possible that the new knowledge may be used to develop substances capable of blocking the protein's action and thus limit the tumors' blood supply.

Such a treatment has long been a goal of cancer researchers who speculate that it may be possible to stunt tumors while they are still very small. It is known that tumors cannot grow more than a few millimeters in diameter without requiring new blood vessels to feed them.

The Harvard findings may also prove useful in speeding up the healing of wounds and in treating diseases in which tissues suffer from inadequate blood supply. In many forms of heart disease, for example, narrowed arteries leave the heart muscle starved for oxygen. It is conceivable, the researchers say, that doses of the protein to the heart could trigger new blood vessels.

The newly identified protein is not the only one that plays a role in making new blood vessels grow. In recent years scientists in many laboratories have found that various natural substances in the body can do this. Until now, however, none has been reported to be as well understood in its chemical structure.

The Harvard group is the first to determine the complete sequence of units, called amino acids, that make up the protein and the corresponding sequence of the gene that directs its manufacture in cells.

It has also found that the gene for the protein is naturally present in all cells. Presumably, the gene is dormant in most cells, becoming active only when cells start proliferating, as in healing wounds and growing tumors.

The new findings, announced yesterday in Boston, culminate 10 years of research under the leadership of Bert L. Vallee, who heads Harvard's Center for Biochemical and Biophysical Sciences and Medicine.

Vallee has named the protein angiogenin, from the Greek terms for "vessel" and "to produce."

"It's very gratifying to be able to understand a protein as well as we do this one," Vallee said. "I think this has great potential."

Judah Folkman, one of Vallee's colleagues at Harvard who has been working separately to understand the same phenomenon, hailed the development as "very stunning work."

"In the 1970s, angiogenesis was just a phenomenon," Folkman said. "Now we have a specific molecule and the gene for it. This should make it possible to study angiogenesis in great detail."

Lance Liotta, a National Cancer Institute specialist in the mechanisms of tumor growth, praised Vallee's findings but cautioned against too much optimism because several other natural substances have also been found to trigger blood-vessel growth.

"I would like to believe he has the key angiogenesis factor," Liotta said, "It would be fantastic." But, Liotta said, angiogenesis is a multistep process in which it is not yet clear whether angiogenin is an essential participant.

Vallee said that he was not claiming angiogenin is the sole agent responsible for making blood vessels grow, but because it works in extremely low concentrations, it is a more potent inducer of vessel formation than any other.

One test of angiogenin's power involves injecting very small doses into the center of corneas in rabbits' eyes. The cornea, though composed of living cells, is transparent and normally without blood vessels. Within days of injection with angiogenin, tiny blood vessels begin growing through the cornea, moving from the edges toward the center.

Vallee suspects that the gene for angiogenin is active during an individual's early growth and development but then shuts down -- ending the production of angiogenin by the cell -- until reactivated by special circumstances.

One such circumstance would be a wound, in which new tissues grow to repair damage and must induce existing blood vessels to keep up with them. Angiogenin, Vallee suspects, diffuses out of the new cells to act on the nearest blood vessels. In women the gene may also be triggered in the lining of the uterus during every menstrual cycle to prepare for implantation of a fertilized egg.

In most normal cases, the gene is active only briefly. In tumors, on the other hand, it may remain switched on indefinitely.

Vallee said he was now looking for ways to turn angiogenin on and off.

One widely discussed possibility would be to make antibodies that bind to the protein, deforming its shape and thus canceling its power to make capillaries grow. Antibodies are normally produced by the immune system to attack invading microbes, but, in recent years, biologists have developed a way of making large quantities of antibodies, called monoclonal antibodies, tailored to bind to a given protein.

These antibodies would be a prime candidate for the role of blocking angiogenin produced by tumor cells.