Nearly 100 years ago, a New York surgeon named William B. Coley claimed that he could cure cancer by injecting people with a concoction of killed, disease-causing bacteria.

Coley's toxins, as the unorthodox potion was called, never caught on, and eventually the notion was largely forgotten.

Today, medical researchers in several centers have found that Coley was on to something, and this fall they plan to test on people an anticancer drug inspired by his findings. They have found that he also opened the door to other possibilities.

These include a treatment for victims of potentially fatal shock caused by certain bacteria and a drug that really might do what countless quack remedies have claimed: cause the body to shed fat.

"It truly is remarkable what this is leading to," says Bruce Beutler, a biochemist at Rockefeller University in New York. "People were working in totally different areas, and now it looks like there's one explanation for all the different phenomena."

The trail began in the 1890s, when Coley noticed that some cancer patients seemed to recover spontaneously after suffering a major bacterial infection.

Reasoning that the microbes somehow triggered the death of cancer cells, he developed a mixture of certain heat-killed bacteria and gave it to cancer patients. Some, against all odds, recovered, their tumors shrinking away. But others died anyway.

Coley's colleagues were largely unconvinced, and the toxins were tried only sporadically afterward, usually with mixed results. Eventually the American Cancer Society branded them an "unproven remedy," lumping them with various quack nostrums.

In the 1970s, scientists at New York's Sloan-Kettering Institute for Cancer Research, led by Lloyd Old, exhumed Coley's findings and tried similar experiments on mice. They discovered that Coley's toxins somehow induced one type of white blood cell, the macrophage, to manufacture a previously unknown protein that killed many types of tumor cells without killing normal cells. They named it tumor necrosis factor (TNF).

Alerted by that knowledge, researchers in many centers speculated that if enough TNF could be given to cancer patients, it might cure their cancer. More recently, at least three genetic engineering companies cloned the human gene for TNF, planning to make enough of the protein in engineered bacteria to treat patients.

Genentech Inc., in South San Francisco, plans human tests this fall. Cetus Corp., in Emeryville, Calif., plans to follow next year.

Meanwhile, researchers elsewhere were working on a widespread medical problem known as cachexia, from Greek words meaning "bad condition." Patients with cachexia (pronounced kak-hexia) lose weight inexorably, often wasting away to death. It is common in the end stages of many fatal infectious diseases.

Anthony Cerami at Rockefeller University, working with Beutler, has found that cachexia is brought on by a particular substance that is part of the cell wall of a large class of microbes known as Gram-negative bacteria, named for the way they take up a stain under the microscope.

This class includes bacteria that cause meningitis, typhoid, cholera, plague, undulant fever and several other diseases. The class also includes Escherichia coli, or E. coli, which inhabits the human gut. When these normally useful bacteria escape the intestinal tract and get into the bloodstream, they can cause potentially fatal septic shock.

Cerami has found that when the bacterial substance enters the body in large enough amounts, it causes macrophages, the same blood cell that makes TNF, to make a protein that blocks cells' ability to store fat and even causes them to dump fat they have accumulated.

Cerami, concluding that the substance was responsible for the weight loss and wasting away of infected patients, named the substance cachectin.

Cerami, collaborating with Gordon Ringold and his colleagues at Stanford University, recently found that cachectin is the same substance as tumor necrosis factor.

The finding has raised questions about the wisdom of giving large doses of TNF to cancer patients, but the genetic engineering companies say they plan to go ahead. They are trying modified forms of the TNF molecule that may not cause severe wasting. And experiments with mice show that, although they lost weight during treatment, they regained it later.

Some researchers believe that TNF may be a cause of the wasting that often afflicts dying cancer patients. It may be that their bodies are producing large amounts of TNF in an effort to combat the tumors but that the cancer cells are resistant.

The obesity-treating potential of TNF, or cachectin, has not been lost on researchers. Several scientists note that it may be possible to find a safe dose that will be high enough to make the body shed fat but not cause dangerous wasting.

Although some biotechnology firms are reportedly interested in this application, they must deal with the finding that very low doses cause severe shock in monkeys. A few millionths of a gram can kill mice.

"I wouldn't advocate giving it to obese people right away," Beutler says. "Cachectin appears to produce a wide range of metabolic effects, and it will take some work to learn how to use it."

Beutler and Cerami said they believe that cachectin, or TNF, is the major cause of septic shock. Like the shock that results from a sudden, massive loss of blood, a severe bacterial infection can cause blood pressure to drop dangerously low, which means vital organs such as the heart, brain and kidneys no longer receive enough blood to function and death can result.

The Rockefeller researchers believe it may be possible, however, to treat impending shock by giving patients a substance that would stop cachectin from entering cells. A leading candidate would be an antibody to cachectin.

Because cachectin is made in the patient's body, the immune system makes no antibodies to attack it. But, Cerami speculates, if human cachectin were injected into an animal, its immune system would recognize it as a foreign protein and make antibodies that bind to it.

Animal antibodies, which can be harvested from the animal's blood, could be injected into shock victims in the hope that they would bind with the patient's cachectin, preventing it from acting on cells.

Still unexplained are the precise details of TNF's effect on tumor cells and its ability to cause shock. Cerami and Ringold said they suspect that its antitumor activity is not the chief reason the body makes TNF, or cachectin.

They speculate that it, like the macrophage cells that make it, is part of the immune system for fighting infection. The microbes that cause malaria and African sleeping sickness, though not bacteria, also cause macrophages to make cachectin.

If an infection is mild, they speculate, the body makes just enough to mobilize energy stored in fat to fuel the fight against the microbe. Larger quantities induce septic shock.

Middling doses, such as those brought on by Coley's toxins, may throw the body into a mild shock but also somehow disrupt the special metabolism of cancer cells.