Scientists have long sought ways to sic the body's own immune system on cancer cells. But such immunotherapies, while elegant in theory, have proved elusive in practice. Part of the difficulty has come from trying to discover both if and how the immune system targets tumors for destruction, much as it attacks infectious pathogens. In today's issue of Nature, a trio of researchers describes the first conclusive evidence that the immune system does indeed prevent tumor formation, and their work also offers clues about the mechanism involved.

Robert D. Schreiber and student Vijay Shankaran of the Washington University School of Medicine in St. Louisalong with Lloyd J. Old, director and CEO of the Ludwig Institute for Cancer Research at the Memorial Sloan-Kettering Cancer Centerperformed a series of experiments on mice engineered to lack certain key immune system components. In some mice, the RAG2 gene was knocked out so that these animals did not produce immune cells called lymphocytes. In others, the gene for Stat1a protein required by immune system gamma interferon (IFNg) receptorswas missing. And in some animals, both genes were absent.

When they injected the mice with a chemical carcinogen MCA, they found that 58 percent of the RAG2-deficient subjects developed tumors, as did 72 percent of the animals lacking both RAG2 and Stat1. In contrast, tumors arose in only 19 percent of normal mice. Because there was little statistical difference in the incidence of tumors between the two knockout types, the researchers concluded that lymphocytes and IFNg receptors work in tandem to kill tumors. "IFNg makes tumor cells expose themselves to the immune system," Schreiber says. "After seeing the abnormal proteins in the tumor, the lymphocytes eliminate the tumor cells." The team also noted that after 15 months, two of 11 normal mice not exposed to MCA spontaneously developed noncancerous tumors; all 12 unexposed RAG2-deficient mice developed tumors, half of them malignant; and all 11 animals lacking both genes developed cancerous tumors in the same period.

By transplanting tumor cells between animals, the scientists further discovered that immune defenses against cancer cells are choosy. Thus, they propose that the immune system does not catch cells on the verge of malignancy but rather edits out only some types of tumor cells and alters others after the fact. "Immunoediting could explain why the immune system can select for tumors that are more capable of escaping the immune system as they continue to develop," Schreiber says. "If this process is always occurring, it can have multiple outcomes, one of which is protection. But if you're unlucky, transformed tumor cells might alter themselves so the immune system can pick out only a few. The others continue growing."

In one final experiment, the researchers added a protein called TAP1 to highly aggressive tumors harvested from healthy mice and transplanted them into other healthy mice. They found that tumors tagged in this way were rejected, whereas untagged tumors grew rapidly. RAG2-deficient mice did not, however, reject transplanted tagged tumors. "We showed that if a tumor is forced to reveal itself to the immune system, it often is rejected," Schreiber says. "We think that a tagged tumor could be used to train the immune system to reject others like it. This is very exciting because it indicates that immunotherapy has a significant potential use even for the treatment of tumors that are altered by the immunoediting process."