Medical Examiner

Killer T-Cells

The promise of using genetic engineering to treat cancer.

Once a week, it seems, there is news of a breakthrough in cancer treatment. But the announcement 10 days ago that scientists have succeeded in using gene therapy to treat malignant melanoma merits special interest. It involves not an advance in a standard treatment but a new method for fighting cancer—the engineering of the body’s own tumor-fighting cells to specifically target malignant ones. This may be the start of a new era in cancer treatment: Genetically modifying a patient’s own immune defense cells to fight tumors could be more effective and less invasive than the chemotherapy, radiation, and surgery we now depend on.

The laboratory of Dr. Steven A. Rosenberg of the National Cancer Institute, from which the new report comes, focused on cases of malignant melanoma (skin cancer) that had failed conventional treatment and spread widely. The lab used its experimental method to treat 17 patients. Of these, 15 showed little or no improvement, but two (about 12 percent) were alive and doing well a year and a half later.

For plenty of kinds of cancer—many lymphomas and leukemias, for example—long-term survival and even complete remission are common after treatment. So what’s the big deal about a 12 percent success rate? The answer is that malignant melanoma, which usually originates in the pigment-producing cells of the skin, is an especially dangerous form of cancer. It’s often aggressive, invasive, hard to control, and unresponsive to chemotherapy. And once malignant melanoma has metastasized, with cells from the primary tumor seeding themselves elsewhere in the body to grow secondary tumors, the prognosis becomes truly dismal. A patient with several distant metastases has only a 1 percent or 2 percent likelihood on average of living for more than a year, even with the most vigorous treatment. Since the patients in Rosenberg’s study had already failed conventional treatment, their prospects were probably still poorer. So, his lab’s achievement of a 12 percent survival rate at about 18 months seems remarkable, indeed.

Random events and mutations cause rare cells in each of us to become malignant. It’s the job of our powerful immunological surveillance system to detect and eliminate these aberrant cells before they can multiply. Cancer occurs when this surveillance system fails. Rosenberg’s innovation is to modify the body’s own cells to attack the aberrant cells once they’ve multiplied.

To recognize and destroy abnormal body cells, the body makes use of a special kind of white cell—the killer T-cell. Rosenberg and his colleagues cleverly began harnessing these cells in a previous 2005 study of 35 patients with malignant melanoma. For this research, the team collected samples of the patients’ immune-system cells and picked out the killer T-cells whose recognition site matched abnormal molecules (called antigens) on the surface of the melanoma cells. The researchers stimulated these T-cells to multiply like crazy in the lab. They were then injected into the patient after a brief blast of standard chemotherapy drugs made room for them to multiply by depleting the patient’s white blood cells. Because the killer T-cells came from the patient, his or her body would not reject them, and the killer cells would not attack the patient’s normal tissues.

In the 2005 study, half the treated patients showed clear improvement with reductions in tumor size, and three patients (9 percent) went into complete remission. That was a start. But Rosenberg’s initial method was not easily applied outside of an experimental setting. It is difficult and laborious (and sometimes impossible) to find the appropriate killer T-cells and stimulate them to multiply in the laboratory. And it is never certain that the best attack cells and target antigens have been chosen.

Which leads to the research just announced: In Rosenberg’s latest study, a patient’s own white cells (collected from his or her blood) were genetically engineered to produce the killer T-cells that targeted the melanoma cells. When the engineered cells were injected back into the patient, they performed like normal killer T-cells: They multiplied in the patients and (when things worked well) mounted a targeted lethal attack on the tumor cells, with virtually no side effects.

It’s a stunning advance because these were made-to-order killer T-cells, directed precisely at a target—the tumor. If ongoing studies teach us how to improve the success rate of this method, we may soon be able to use it to treat cancer without depending on the good luck of finding the right kind of killer T-cell in a patient’s immune system. Instead, we could examine a patient’s tumor, test to see which antigens are found on the tumor cells, and then construct the killer T-cells that are needed to go after it. Rosenberg’s method of genetic engineering could be directed not just at melanoma, but at virtually any kind of cancer. And, at least in principle, it could generate killer T-cells to attack cells infected with the viruses that cause AIDS and other chronic, lingering infections, like some forms of hepatitis.

Are there tumors that can fend off the killer T-cells? Can we clone the right genes to manufacture the T-cells needed to attack every patient’s tumor? Can we make the process more efficient so that a larger percentage of patients will be successfully treated? We don’t know the answers to these questions yet. And while this method seems safe so far, it has been tested on only 17 people.

These uncertainties notwithstanding, there is, of course, great desire to develop improved cancer treatments. Cancer chemotherapy, introduced nearly 50 years ago, has been largely based on the use of cell poisons. The chemicals are chosen because they’re more toxic to tumor cells than to normal tissue, but none has precisely targeted malignant cells. And there’s a familiar downside: Normal cells are also damaged by these chemicals, and the threat this poses can prevent treatment that’s aggressive enough to kill the cancer.

The last few years have brought several advances. Drugs like tamoxifen and Erbitux block access to certain tumor cell surface receptors and thus prevent the cell from being stimulated to grow and multiply. And drugs like Gleevec block enzyme systems that certain tumor cells need to function but normal tissues don’t need. Still other drugs (thalidomide and Avastin, for example, and others coming along) block the development of the supply of new blood vessels that cancers depend on. All of these new treatments are useful and most are less toxic than the old-line drugs. But most seem limited in the kinds of cancer they can treat and often need to be given in concert with the older drugs.

Rosenberg’s method offers a way to target tumor cells that spares normal cells. Any number of difficulties could yet emerge. But so far, the potential seems great—ultimately, this could change the way we think about treating cancer.