Last month, the world's biggest drug company, Pfizer Inc., signed Medarex to a decade-long deal that could bring up to $500 million and a stream of future royalties. Pfizer, which has yet to win approval for any antibody drugs, will try to develop up to 50 using Medarex technology.

Already, about 20 experimental drugs using Medarex antibodies are in various stages of testing — about one-third by the 17-year-old Princeton-based company and the rest by its collaborators and licensing partners. Those 50 partners include Abbott Laboratories, Eli Lilly & Co., Novartis Pharmaceuticals and Schering-Plough Corp., along with biotech companies Celltech Group, Genmab A/S, Human Genome Sciences Inc., Immunex and MedImmune.

"There are so many opportunities with this technology that they go beyond what we can do on our own," said Donald Drakeman, Medarex's co-founder and chief executive officer. "The point is to take advantage of all these possibilities for the benefit of medicine and for the benefit of our company."

Medarex, which has laboratories in Bloomsbury, N.J., and in Milpitas and Sunnyvale, Calif., was founded in 1987 specifically to do work on monoclonal antibodies, which are made by repeatedly cloning, or copying, cells that make one specific antibody.

The company went public in 1991 and has raised about $1 billion to finance research, but the early going was tough. In 1997, Drakeman took a big gamble, spending about two-thirds of the company's capital to acquire biotech company GenPharm, which had figured out how to make the immune systems of mice produce human antibodies.

The "transgenic" mice, bred from standard brown ones used in research labs worldwide, are altered by inactivating two key genes and adding two human genes, said Nils Lonberg, GenPharm's top scientist and now scientific director of Medarex. That's done through "molecular biology tricks" on the cells of mouse embryos.

To get specific human antibodies for use in drug development, Lonberg's team injects the mice with cells or material from a tumor, infectious agent or other target.

"The mice respond with a human antibody response instead of a mouse antibody response," he said.

Medarex researchers then put the antibody-making cells from the mice into special nutrients inside a vessel much like a beer-brewing vat so they multiply — producing the monoclonal antibodies. The company's collaborators then use those antibodies for their own research, or sometimes have the mice shipped overnight to them once they have begun producing antibodies.

The company's main competitor, Abgenix Inc. of Fremont, Calif., also has created mice that can produce human antibodies, and also is developing antibody drugs on its own and with collaborators.

One breakthrough could come from his company's leading drug candidate, MDX-010, for melanoma in patients who have failed prior treatments and had the deadly skin cancer spread to internal organs.

The final stage of human testing began in September and the drug could be approved in 2006, according to biotech analyst Brian Rye of Janney Montgomery Scott LLC. Medarex still owns full marketing rights to MDX-010, meaning it would receive all the revenues if the drug is approved.

Medarex also owns full rights to three other cancer drugs in human testing.

Rye said Medarex is well structured, has strong patent protection for its technology and enough money to finance operations for at least three or four years. The company, with about 430 employees, had revenues of about $11 million and a net loss of $129 million last year.