Scientists seek to fight cancer with cancer.

Earlier this year, researchers injected two patients suffering from advanced melanoma with white blood cells modified to bear the genetic instructions for making a cancer-fighting protein. Now, they want to take their gene therapy experiment one step farther.

Study director Steven A. Rosenberg says his group plans to insert two cancer-fighting genes directly into tumor cells taken from additional volunteers with advanced melanoma -- a highly lethal skin cancer -- and then put the altered cells back into the patients. The proposed experiment, which he announced in Houston this week at a meeting of the American Society of Clinical Oncology, would mark the first use of genetically engineered cancer cells in humans.

Rosenberg, of the National Cancer Institute in Bethesda, Md., says he expects the altered tumor cells to prompt a stronger immune response against the cancer, in effect immunizing the patients against their own tumors. "We're trying to convince the cancer patient's body to generate a better and more effective antitumor response," he says.

The study is an outgrowth of a 1989 experiment in which Rosenberg and his co-workers treated melanoma patients with genetically modified cells called tumor-infiltrating lymphocytes (TIL). The researchers removed these cancer-fighting white blood cells from the patients' tumors, then supercharged the cells with interleukin-2 and added a "marker" gene as a label. When the supercharged TIL cells were reinjected into the patients, they homed in on and attacked tumors better than other lymphocytes from elsewhere in the body.

Last January, the team treated two melanoma patients with TIL cells engineered to carry a gene for tumor necrosis factor (TNF). Injections of this normal body protein have been shown to drastically reduce tumors in mice (SN:2/2/91, p. 69), but purified TNF is too toxic in humans to be injected in doses large enough to shrink their tumors. The researchers sought to circumvent this problem by using the altered TIL cells to release TNF only in tumors. Rosenberg says it's too early to present the results, but he reports that neither patient has shown serious side effects from the experimental treatment.

In the new experiment, he plans to eliminate the need for TIL cells by inserting genes for TNF or interleukin-2 directly into tumor cells removed from patients. The researchers will allow the engineered cells to multiply in a lab culture, then return them to the patients.

In mouse experiments, such altered tumor cells continued to multiply after reinjection, then died off as the mouse's immune system attacked them, Rosenberg says. But before the cells died, they enhanced the immune system's ability to recognize and kill all remaining tumor cells. Using this approach, "we have been able to immunize animals against their cancers," he reports.

The National Cancer Institute's bioethics review board has granted provisional approval to test the treatment in human patients. But Rosenberg must also win approval from three other National Institutes of Health committees and from the FDA -- a process that could take more than six months.

J. Gordon McVie, scientific director of Britain's Cancer Research Campaign, calls the proposals to engineer tumor cells instead of TIL cells "absolutely brilliant." He adds, "We haven't got the foggiest idea which one will work better."

Rosenberg is already planning yet another gene therapy approach to cancer. He believes he has isolated a gene for a protein present on the tumor cells of all melanoma patients. Although the discovery awaits confirmation, he hopes one day to insert the gene for this "melanoma tumor antigen" into Vaccinia viruses, commonly used as the basis for vaccines.

With such a vaccine, he suggests, "someday it may be possible . . . to immunize [people] against cancer" so that they never develop melanoma.