Allison explained that the flip side of tumor immunity is autoimmunity, since tumor cells, arise from normal cells and therefore differ little from other cells in the body. "The body is trying to balance recognition of self and recognition of altered self, and when that fails in either direction you get autoimmunity or cancer," he said.
He and his colleagues reasoned that blocking CTLA-4's actions on immune cells known as T lymphocytes might allow the killer cells to destroy the cancerous cells rather than fade away after the tumor repels their initial attack. To accomplish this, the investigators raised a monoclonal antibody against CTLA-4 and administered it to mice with various types of cancer.
"CTLA-4 blockade using this monoclonal antibody worked, but not particularly well in anything but the least aggressive types of tumors," said Allison. "We wanted something that would work in tumors that are hard to kill."
At about the same time, other researchers were testing the effectiveness of a vaccine made from melanoma cells that were both irradiated, to kill them, and genetically engineered to produce granulocyte/macrophage colony-stimulating factor (GM-CSF). This compound, which the immune system can produce naturally, triggers scavenger cells known as macrophages to travel to a site of infection and engulf invading microorganisms. Trials with this vaccine, however, produced about the same results as those from Allison's laboratory with CTLA-4 blockade -- some success, but mostly failure.
"So we decided to use both approaches, and the results were very positive, much better than with either approach alone," said Allison. Well-developed melanomas vanished in 80 percent of the mice with already established cancer. Moreover, the therapy protected the animals from developing melanoma again even after receiving additional doses of tumor cells.