Early in this century, patients with cancer would often seek medical attention only in the final stages of their disease, after their tumors had become massive. Surgeons would attempt to remove these tumors to alleviate their patients’ pain. But since sterile operative techniques were in their infancy and the discovery of antibiotics almost half a century away, such surgery often caused massive and frequently fatal infections.
In a few instances, however, tumor remnants of infected patients would disappear, leaving them healthy and doctors puzzled. Today, after decades of research into how the immune system works, scientists have learned that the seemingly miraculous cures resulted from the infections themselves, which set in motion complex immune reactions.

Understanding the relationship between cancer and the immune system-whose purpose is to find and destroy any abnormal cells in an organism among a mass of normal ones-has resulted partly from so-called “experiments of nature.” A few diseases involving immune deficiencies are associated with higher rates of cancer, for example. A person infected with HIV, for instance, is 100 times more likely to develop a malignancy than an uninfected person. Also, organ-transplant recipients treated for a long time with drugs that suppress the immune system, thereby preventing rejection of the new organs, are many times more likely to develop cancer than the rest of the population. In fact, an elegant demonstration of the immune system’s role in protecting against cancer is the fact that tumors occasionally regress when doctors remove immunosuppressing drugs.

Armed with such discoveries as well as modern biotechnology techniques, researchers have begun using several elements of the immune system to systematically destroy tumor cells. (Of course, immunotherapy is only one of a number of innovative ideas in cancer treatment that hold significant promise. The public investment into cancer research since the Nixon administration declared war on cancer in 1971 is finally starting to produce a variety of innovative techniques, such as methods that exploit the genetics of how a cell becomes malignant.) The work, being conducted on both animal models and people with cancers difficult to eliminate by traditional means, has a long way to go but suggests a more effective way of treating many cancers in the future.

The Problem with Recurrent Cancer

The standard methods for eliminating malignant tumors have long been surgery, radiation, and chemotherapy. The underlying crude assumption is that doctors can generally destroy a cancer without losing the patient to the therapy’s toxic effects. Of course, as researchers have developed supportive care such as powerful antibiotics and growth factors-natural body proteins that can be manufactured using genetic-engineering techniques and that cause the white-blood-cell count (which plummets after treatments) to recover more quickly-doctors have been able to administer more intensive chemotherapy. These treatments have played a role in the recent good news that the number of U.S. cancer deaths has finally dropped for several consecutive years. Still, 500,000 U.S. residents annually die of malignancies-a number hard on the heels of the country’s number-one killer, cardiovascular disease.

For everyone who has witnessed cancer firsthand-whether as patient, family member, friend, or medical professional-one of the most worrisome and disheartening aspects is that while today’s therapies can generally eradicate all measurable evidence of disease initially, any remaining cells may proliferate and cause a relapse of cancer. And because the first set of remaining cells has resisted chemotherapy, their offspring have a selective advantage to do the same, leaving the person with a recurrence of cancer that is often widespread and much less easily treated with chemotherapy or other techniques.

A critical need, therefore, is to find and eliminate the few remaining cancer cells left viable after conventional therapy. The immune system’s elegant and complex methods of attacking invaders suggest a variety of ways to assail the remaining bits of malignancies. (Another goal is to discard today’s techniques altogether in favor of an entirely different method. But eliminating treatments that now work better than anything else is not judicious until viable alternatives develop.)

At its simplest level, the immune system discriminates between “self” and “nonself” and destroys the latter. In most responses, the system relies partly on antibodies-proteins with two specialized ends. One end of an antibody binds to the infecting cell or virus while the other end attracts immune cells that engulf and digest the invader. The attack also relies on different immune cells, which, upon contact with foreign cells, secrete compounds that make the offenders’ membranes porous so that their vital contents leak out and the cells eventually die.

With this general understanding of the immune system and knowledge of the early reports of tumor regressions after infection, in the 1960s and ’70s clinicians injected some solid tumors with killed or weakened microbes, such as BCG (bacille Calmette-Gurin, an altered form of tuberculosis) and Cornybacteria (the cause of diphtheria), in the hope that such outside organisms would stimulate an immune response. And indeed they did, sometimes destroying the tumor. Although the results were inconsistent, the work demonstrated that “immunotherapy” might someday have a place in the cancer-fighting armamentarium.

Toward Specific Immune Therapy

Then along came genetic engineering and the recognition of its value in identifying and manufacturing proteins called cytokines. Some of these can act as anticancer agents because they regulate the activation of immune cells. They essentially put the immune system at a higher state of alertness so that it can better detect tumor cells. Researchers had discovered some cytokines in the mid 1970s, before the era of genetic engineering, but that technology made possible the production of a large enough quantity of the proteins for study and treatment.

Over the years, investigators have identified several cytokines that lead to the destruction of tumor cells. Indeed, the U.S. Food and Drug Administration has approved the intravenous use of the cytokine interleukin-2 for this purpose against kidney cancer. Researchers have also found this cytokine can help against one form of leukemia and malignant melanoma, and are studying its efficacy against a broad range of other cancers. And scientists have discovered that the cytokine alpha interferon can activate immune cells against a form of leukemia. Although for reasons not yet understood investigators have found that cytokine injections generally induce responses in only 20 to 30 percent of patients, researchers worldwide are continuing to explore which of a large variety of tumors, prognostic factors, doses, and dose schedules work best with particular cytokines.

Meanwhile, following some early work injecting cytokines to fight cancer, some scientists recognized that the amount of immune-cell activation that can take place in the body through cytokine injection is limited by the dose of cytokines that can be administered that way. The treatment’s side effects, such as alternating fever and chills, low blood pressure, and difficulty in breathing, can become severe enough that patients need to stop the therapy. The investigators thought a way to raise the amount of activated immune cells would be to identify some of those cells, remove a few from the body, and mix them with a high dose of cytokines in a test tube. The idea was that the resulting large mass of activated immune cells, when infused back into the body, would travel to tumor sites and destroy cancer cells.

In 1982 Elizabeth Grimm, then a cancer expert at the National Cancer Institute (NCI), and two colleagues were the first to identify a kind of immune cell-which they labeled a lymphokine-activated killer (LAK) cell-that led to the death of tumor cells. Using the cytokine interleukin-2, the team next activated quantities of LAK cells in test tubes. Working in the laboratory of Steven Rosenberg, chief of NCI’s surgery branch, the group ultimately found that LAK infusions destroyed cancer cells, especially those associated with malignant melanoma and kidney cancer. But other researchers concluded that the effects were no stronger than those seen from simply injecting interleukin-2. Apparently generating LAK cells outside the body was no more effective that creating them within the body through infusions of interleukin-2.

Then in 1986 Rosenberg isolated an immune cell from inside a tumor itself. Calling this cell a tumor-infiltrating lymphocyte (TIL), he postulated that it recognizes proteins unique to a tumor, moves into that entity, and attempts to generate an immune response. He extracted TIL cells from a surgically removed tumor and, in a test tube using interleukin-2, stimulated their growth up to about 100 times their original number. After this process, which took three weeks, Rosenberg injected the complete mass of TIL cells into animal models and later patients. The therapy was more potent than LAK treatments, with the tumors in 11 of 20 patients shrinking or disappearing altogether.

In follow-up research to determine why the TIL technique didn’t always work as expected, however, in 1990 Richard Barth of the same laboratory determined that TIL cells do not kill tumor cells directly. Instead, after traveling to the tumor site they secrete more cytokines, such as tumor necrosis factor (TNF) and gamma interferon. These presumably recruit still other elements of the immune system in an attempt to destroy the tumor-an operation analogous to a reconnaissance team finding an enemy and then firing flares to bring reinforcements. Armed with this knowledge, in a subsequent experiment the lab tried to genetically engineer TIL cells to raise their secretion of TNF and gamma interferon. The idea was that the manufactured material, after being injected into the body, would home to areas surrounding malignant melanoma tumors-a particularly virulent form of cancer. But the researchers couldn’t consistently engineer the TIL cells as they wanted to.

Interest in the method, as well as in injecting amplified but unaltered TIL, has subsequently waned because of such practical difficulties. Moreover, growing TIL cells is costly because of extensive involvement of lab technicians over three to six weeks, and contamination of cultures with bacteria or fungi is always possible. Repeated surgeries are also required to provide tumor cells to stimulate continual TIL growth. Still, the notion underlying Barth’s research-genetically manipulating tumor cells-has provided the intellectual impetus for the next stage of work: cancer vaccines.

Fighting Tumors with Their Own Kind

When most people think of vaccines, their reference is to doctors giving these agents to prevent illness in the first place. But vaccines can be more than that, since they are simply modified infectious agents, or portions of agents, that stimulate an immune response when administered into a body. Thus a vaccine can be given after a person is sick, with the intention of stimulating an immune response strong enough to enable the patient to heal and then stay well. This approach underlies the idea of cancer vaccines.

The method used to construct most cancer vaccines has entailed introducing genes for cytokines into tumor cells and injecting these into the body. (To ensure safety, they have been treated with high-dose radiation so they cannot proliferate but can still carry out functions such as releasing the cytokines.) Researchers believe that physically bringing together these agents, plus immune cells that are always circulating in the body, approximates what happens during the natural generation of an immune response against cancer. The additional tumor cells should theoretically secrete extra-large doses of cytokines that will in turn activate heightened immune responses. And theoretically, the treatment should work for a long time, since immune responses produce other cells known as memory T cells, which provide long-term immunity.

Experiments introducing cytokine genes into animals’ tumor cells began in the late 1980s, with researchers having now tested more than 10 of some 50 isolated cytokines. Working with mice, scientists have found that 4 cytokine genes-those that produce interleukin-2, interleukin-4, gamma interferon, and GM-CSF-suppress the growth of existing malignancies of various kinds, including melanoma, colon, kidney, breast, and leukemia.

The encouraging results have led to more than 80 human clinical trials. Researchers have treated over 100 people with interleukin-2 vaccines, more than 80 with GM-CSF, and 30-plus with gamma interferon. The investigators are just starting to publish their findings. In the first trial to be described that has had a significant number of patients, for instance, investigators at Somatix of Alameda, Calif., inserted the GM-CSF gene into tumor cells of more than 40 extremely sick patients for whom other treatments for malignant melanoma had failed. The researchers then injected the modified cells into the patients. While because of the stage of their illness many of the patients died during the treatment period, 3 of 13 people have had some tumors partly or completely shrink after therapy. Since ordinarily none of these tumors would have been expected to shrink, that number is encouraging enough to suggest that further trials are warranted.

Still, the general technique is not ideal to use across the board, since it requires genetically engineering each patient’s tumor cells. That’s necessary because an individual will reject another individual’s cells. To avoid this problem, some researchers have come up with the idea of using something other than whole cells: tumor antigens-proteins found on the surface of tumor cells. Antigens, which are not necessarily unique to individuals and therefore not something that bodies reject, lead to an immune response against abnormal entities such as tumors. If investigators can figure out the antigens produced by particular kinds of tumors, the thinking goes, large amounts of such compounds, or better yet, their genes, could be placed, along with cytokine genes, into “vectors.” These are viruses manipulated so that they cannot replicate and therefore cannot cause disease but can still insert themselves into the DNA of host cells. The engineered vectors could be used to infect cells in a person with a particular malignancy. Presumably the resulting antigens would induce an immune response against the cancer.

In 1991, Thierry Boon, director of the Ludwig Institute for Cancer Research branch in Brussels, published a technique for identifying antigens found on the surface of tumor cells. Transferring into mice a gene he discovered using this method-for an antigen associated with a mouse tumor-Boon has generated immune responses that have protected against subsequent injection of tumor cells. Steven Rosenberg’s group at the National Cancer Institute has since started clinical trials to evaluate the effect of transferring genes for that antigen, as well as two others subsequently discovered, into melanoma patients. Results are not yet available.

Inhibiting Inhibitors

Other researchers have taken an altogether different tack in developing cancer vaccines. Recognizing that some cytokines associated with tumors do not activate an immune response but actually inhibit it, a group led by Habib Fakhrai, director of brain-tumor gene therapy at University of California, Los Angeles, School of Medicine, has researched techniques for preventing tumor cells from secreting inhibitory cytokines. This past year Fakhrai published an approach for treating an aggressive brain tumor that secretes an inhibitory cytokine called transforming growth factor-beta (TGF-). This cancer usually causes patients’ deaths within a year of diagnosis.