Five years ago, Philip Karr's cancer had brought him to the brink of death. His white blood cells grew out of control, making him weak and vulnerable. The standard therapies failed and the most aggressive therapy of the time -- interferon -- had dramatically but only temporarily delayed his disease.

With his options running out, the retired Santa Barbara engineer went to Stanford University for one more roll of the dice: he signed up for an experimental treatment with monoclonal antibodies.

The Stanford team, lead by Dr. Ronald Levy, custom-made a batch of monoclonal antibodies to attack Karr's wildly growing B cells.

The results were dramatic. Within a short time, Karr's blood counts were back to normal. All signs of the disease disappeared.

Today, five years later, Karr is disease free, Levy told the country's leading monoclonal antibody researchers gathered at the National Institutes of Health last week.

But no one knows why the treatment worked. The next 13 patients treated exactly as Karr was treated have not done as well and their disease persisted.

Although Phil Karr's case remains unique, it convincingly demonstrates the power of monoclonal antibody technology to successfully attack America's most feared disease.

This technology exploits a fundamental power of the body's immune system: the ability to recognize molecules of a specific shape and attack the cells that carry those molecular targets.

Antibodies sort of look like lobsters. The main body supports two projections -- molecular claws -- that can "feel" their way along. Millions of different antibodies, each with a slightly different shaped claw, can be constructed by the body to recognize slightly different shaped molecules within the body. When a claw encounters the shape it recognizes, it clings fast, its tail waving free.

The waving tail signals other cells in the immune system -- great PacMan-like munchers such as the monocytes and macrophages -- to come and destroy the cell to which the antibodies are attached.

The antibody's claw recognizes antigens, proteins found on the surface of foreign cells, usually invaders such as viruses, bacteria, pollen or a transplanted organ. It can also recognize cancer.

In 1975, two English researchers discovered a way to make buckets of molecular claws that recognized only a single shape. They are called monoclonal antibodies because each batch is cloned, or duplicated, from a single white blood cell that produces only one kind of antibody. That discovery touched nearly every field of medicine and revolutionized many.

For example, a growing list of diagnostic tests rely on monoclonals to measure the stage of a cancer, quickly identify infections and determine proper blood concentrations of therapeutic drugs. One new diagnostic test relies on monoclonals to measure the heart damage after a heart attack.

Consumers, too, use monoclonal-based tests at home to test for pregnancy, detect hidden blood in the stool and determine the time of ovulation.

But more important, monoclonal antibodies are beginning to cross the threshold from diagnostic adjuncts to therapeutic agents.

Recently, the Food and Drug Administration approved the use of a monoclonal antibody treatment to control a rejection crisis for patients receiving transplanted kidneys.

A computerized search of the National Cancer Institute's Physician Data Query (PDQ) information system showed that there are more than 30 clinical studies to test monoclonals against cancers of the liver, lung, pancreas, gastrointestinal tract, colon, skin and blood.

But cancer remains the most important disease to be attacked with monoclonals.

Last week's conference, however, led to one conclusion: "There are no breakthroughs you can splash on page one," said Dr. Ian Trowbridge from the Salk Institute for Biological Studies in La Jolla, Calif. But there has been progress, steady and careful, with the usual amount of complications and setbacks for a field as complicated as human immunology and cancer.

Researchers have taken two basic approaches to killing cancer cells with monoclonals. The antibodies can be used alone, or they can be used to transport drugs, toxins or radioactive particles to the cancer cells.

Naked antibodies -- those not armed with drugs or radiation -- kill by coating the surface of the target cell in such a way that the rest of the body's immune system can recognize the marked cell and destroy it.

Naked antibodies were used to kill Phil Karr's lymphoma.

The ability of some cancer cells to mutate was the most difficult problem faced by the Stanford group, which concentrated on B cell lymphomas. Levy believes that the cancers were not eradicated by the treatment because the target for the monoclonal antibody on some of the cancer cells mutated into a different shape, which the antibody could not recognize.

By performing a genetic analysis, Levy's group discovered hot spots for mutations along certain regions of the patients' genes. Karr's genes, however, did not show mutation activity in a critical region, and that may explain why his cancer was not able to escape the treatment, and he was, apparently, cured.

Because the cancer cells mutate and produce several slightly different populations of cancer cells, "a single monoclonal antibody can only see a portion of the tumor population," Levy said.

To solve this problem, Levy and others suggest, it will be necessary to use two or three or more kinds of monoclonal antibodies in a single dose to treat cancer successfully.

Researchers at Memorial-Sloan Kettering Cancer Center in New York have used naked antibodies against malignant melanoma -- a skin cancer -- in more than 20 patients. Only four patients responded to the treatment, said Dr. Alan Houghton.

He said that the patients suffered some side effects, including nausea, vomiting, fever and diarrhea.

Dr. Albert LoBuglio of the University of Atlanta Comprehensive Cancer Center described the treatment of 20 gastrointestinal cancer patients with a naked monoclonal antibody.

"The clinical response data is rather disappointing," Houghton said. One patient had a remission for six months, four patients remained stable and the disease in all the others continued to progress.

The failure of naked antibodies to eradicate cancer has led researchers to arm them with something that can kill cancer cells -- either a drug, a toxin or a radioactive particle.

In this approach, the monoclonal antibody is like the rocket and targeting system of a guided missile, and the cargo -- drug or radioactive particle -- is like the warhead.

The antibody provides the specificity -- it guides the ensemble to the cancer cell target -- and the cargo destroys the cancer. That, at least, is the idea.

Human trials already are under way to determine whether the concept actually works.

Dr. Ralph Reisfeld of the Scripps Clinic and Research Foundation in La Jolla, Calif., described efforts to treat lung cancer with a monoclonal antibody carrying the drug methotrexate.

"In two patients," Reisfeld said of the 11 patients treated so far, "we saw a decrease in tumor size by CT scan."

But, he cautioned, the result cannot yet be considered significant. "It is much too early to say anything at this stage in the game."

The absence of toxic side effects to the antibody and its deadly cargo, however, was significant. The reduced toxicity -- presumably because the drug was getting to the tumor while sparing normal cells -- allowed the researchers to inject single doses of 25 milligrams of methotrexate. A normal dose during cancer treatment is 2.5 to 5.0 milligrams per day.

Drugs, even those attached to a monoclonal antibody, have a drawback. Hundreds of molecules of the drug must get inside the cancer cell to kill it.

Toxins, on the other hand, are much more efficient. Because toxins themselves are enzymes that selectively destroy key components of a cell, only one or two toxin molecules must get into a cell to be lethal.

Ricin, a potent plant toxin, is one such toxin now being tested.

"Ricin will kill all cells," said Dr. Richard Youle of the National Institute of Neurological and Communicative Disorders and Stroke. "The monoclonal antibody is used to provide specificity."

Youle has used ricin-carrying antibodies against leukemia in guinea pigs. "Twenty-eight percent have survived long-term," he said. "These may be tumor-free at this point."

"We may be able to improve the cell killing" by getting more toxin-carrying antibody to the cancer cells, Youle said. They are now working with an antibody that concentrates 10 times more toxin on the cancer cell than the current antibody.

This spring, the Xoma Corporation in Berkeley, Calif., reported the first clinical use of a toxin-loaded antibody. Twenty-two patients with malignant melanoma that had spread throughout the body were treated. The tumors regressed in five patients and remained stable in five others.

Some researchers are loading monoclonal antibodies with a bit of radioactive material to lethally irradiate the cancer cell.

These tiny nuclear warheads may solve some of the problems with tumors that lack the target molecule for which the monoclonal searches. If enough of the antibodies bind their radioactive load to the surface of cancer cells that do carry the antibody's target, the radiation will kill all nearby cells, both those carrying the antibody target and cancer cells that lack the antibody target.

Dr. Steven Rosen from Northwestern University presented some dramatic slides of patients with severe skin lesions caused by T-cell lymphomas that virtually disappeared, in some cases, after treatment with a radioactive monoclonal antibody.

Only five patients have been treated, however, and their remissions lasted only one to three months. The study, however, did show that it was possible to dose the tumor with radiation while sparing healthy but radiation-sensitive tissues such as the bone marrow, spleen and thyroid.

Dr. Stanley Order at Johns Hopkins University used an antibody preparation, though not a monoclonal preparation, linked to radioactive iodine to treat more than 100 liver cancer patients. Nearly half, 48 percent, of the patients had a complete or partial remission. Standard treatments have only a 15 percent remission rate.

While none of the therapies are producing cures, and many technical problems remain to be dealt with, the studies do show the potential for this fundamentally new approach to targeting and killing cancer cells.

"The monoclonal antibody field is only 10 years old," said Salk's Trowbridge. "Unlike other cancer therapies which have come and gone, this one is still around and still exciting."

At 72, Phil Karr, who was nearly dead five years ago, is still around, too. He still sails his 22-foot sloop out of Santa Barbara harbor. Recently he came back from a ocean cruise to Mexico and now is heading for a several-week tour of Alaska.

*"My body was at the right place at the right time," Karr said in a recent telephone interview. "It was like winning the lottery."