Giving Up on The Mice

By Dick Thompson/Washington

It was time to give the mice a rest and try something else. For 35 years scientists laboring in the National Cancer Institute's screening program have injected more than 400,000 chemicals into leukemic mice, hoping to find chemotherapies that would help solve the riddles of cancer. All that frustrating work has produced only 36 licensed drugs. Most of them, while dramatically effective against leukemia, have shown only modest value in other forms of cancer. "Maybe," says David Korn, chairman of the National Cancer Institute's advisory board, "we've been using the wrong system as the screening device."

Maybe so. In a radical departure from the traditional methods, researchers have swapped their mice for a procedure that they hope will detect a drug's potency not only against leukemia but also in scores of different types of cancer cells. The new effort, which is being employed at the Developmental Therapeutics Program in Frederick, Md., uses an arsenal of automated devices and computers to test potential cancer-fighting drugs on real human cancer cells, grown in laboratories, rather than on mice. This enables scientists to test more than 300 chemicals a week. Many of these drugs had failed in the past when tested on mice, but the researchers hope their more sophisticated approach will produce fresh leads. Says Michael Boyd, founder and director of the program: "It's a high-risk venture, but this is a gamble worth taking."

At one time, the assumption was that cancer cells shared common characteristics and that therefore a drug effective against leukemia would kill, say, cancerous lung cells as well. With mouse screening, that technique brought solid advances in leukemia chemotherapy but yielded mixed results in other forms of cancer.

The theory behind the new program is that cancers from different organs share certain "family" characteristics. For example, brain cells that turn cancerous might share qualities with other brain cancers but differ dramatically from colon cancers or leukemias. To look for common weaknesses among different types of cancer, the automation process tests chemical compounds directly against a range of 60 lines of living tumor cells grown in Petri dishes and representing seven leading cancer killers: colon, lung, melanoma, kidney, ovarian, brain and blood.

Half the compounds are manufactured by chemical and pharmaceutical companies; the rest are provided by botanists and ethnobiologists who collect folk medicines and exotic living materials like the bark of the Pacific yew tree, from which scientists extract Taxol, shown to be effective against ovarian cancer cells. The researchers are looking for "natural" cell killers harvested from such remote locations as the Brazilian rain forest and Australia's Great Barrier Reef. Even ground-up seashells, sponges and coral starfish are studied for chemicals that might show some ability to fight cancer.

Some critics are concerned that the new $4 million-per-year program will fail to spot drugs that are enhanced naturally by the body's metabolism or immune system, and that the old mouse screens were better in that respect. In any case, new agents discovered by the automated screening may require years of additional testing in the lab -- and then on animals -- before any newly discovered therapies can be tried on human cancer patients. "All it will take," says NCI adviser Korn, dean of the Stanford University School of Medicine, "is one smashing winner. Then everyone will say it was worth it."