Research Finding Could Lead To Vaccines Against A Range Of Cancers

This simple gene therapy also protected the mice against subsequent injection of tumor cells that had not been altered, meaning their immune systems remembered the initial challenge - a necessary step for any vaccine therapy.

"When we introduced the gene into cells that didn't have it and then injected the tumor cells into mice, they clearly stimulated a variety of different elements of the immune system, including natural killer cells and killer T cells," said David H. Raulet, the Choh Hao Li Professor of Immunology in the College of Letters & Science at UC Berkeley. "The response in mice was really very dramatic and unexpected - the tumors were all uniformly rejected."

He is optimistic that the gene - actually two families of similar genes - will make an effective vaccine therapy, joining the growing arsenal of cancer immunotherapies in clinical trials today.

Raulet, along with UC Berkeley postdoctoral fellow Andreas Diefenbach, graduate student Amanda M. Jamieson and postdoctoral fellow Eric R. Jensen, report the findings in the Sept. 13 issue of Nature. All are part of the campus's Health Sciences Initiative.

The findings hark back to a theory popular a generation ago - that the immune system constantly surveys the body for cancer cells and kills them on sight. Only when the growth of cancer cells outstrips the immune system's ability to eliminate them do tumors form. The idea fell by the wayside when scientists found that most cancers were no more common in mice lacking certain components of the immune system than in normal mice. More recent studies, however, suggest that the mice used in the first reports were not entirely immunodeficient, and that cancer is indeed more common in immunodeficient mice.

One way the immune system could detect tumors is if cells in the process of converting into cancer cells put up flags, or protein markers, that alert the immune system and draw its fire. The protein products of the genes that Raulet and his team have identified could be one of these common flags.

"The surveillance theory may be on the rebound," Raulet said. "The immune system may be cleaning up many tumors, but we haven't, until now, had the tools to see this."

Raulet and his laboratory colleagues found the proteins while pursuing their main interest, natural killer cells. These generalized attack cells of the immune system are among the first and fastest line of defense against invading tumors and viruses and keep them at bay until the more specialized T and B cells have had a chance to ramp up production.

While searching for proteins that activate the NKG2D receptor on natural killer cells, the researchers came across two families of proteins - called H60 and Rael - that bind the receptor and stimulate natural killer cells to kill as well as secrete gamma-interferon, a cytokine that activates other cells of the immune system.

Interestingly, the NKG2D receptor is also found on two other immune cells - macrophages, which engulf and eat invaders, and CD8 "killer" T cells. Last year, Raulet and his team reported in Nature Immunology that tumor cells producing the H60 and Rae1 proteins trigger macrophages to switch to attack mode.

What intrigued Raulet the most, however, was that the proteins were found on most cancer cells tested, but not on normal cells. Apparently, the protein flags are often at such low levels on tumor cells that they fail to trigger an effective response by natural killer cells or macrophages. Perhaps, he thought, boosting the amount tumors produce would tip the scale.

To test this, Diefenbach put the genes for H60 and Rae1-Beta in separate retroviruses able to shuttle the genes into cells, and used them to transform three different kinds of tumor cell lines that do not normally display the proteins: a melanoma (skin cancer), a T cell lymphoma and a thymoma (cancer of the thymus). He then injected them into laboratory mice at levels that would typically kill them.

All mice injected with transformed cells survived, while none of those injected with untransformed tumor cells survived. They were surprised, however, to find that the survivor mice developed long-lasting immunity to the tumor cells. When these mice were injected 8-12 weeks later with the same tumor cells but lacking H60 and Rae1-Beta, the mice were able to successfully fight off the cells. This protective immunity was mediated by CD8 "killer" T cells.

"We didn't expect protective immunity like this," Diefenbach said. "That and the immune specificity are hallmarks of a T cell response."

Even more surprising was that the immunity was effective against a melanoma cell line that is normally unable to induce even a weak immune response.

Raulet suggests that one reason the immune system is unable to eliminate some tumors entirely is that it unwittingly selects for cells that express low levels of the protein marker the immune system homes in on. That is, tumor cells expressing high levels of the markers are quickly cleaned out, leaving only cells that slip beneath the radar of natural killer cells, macrophages and T cells. These cells then grow unchecked.

"A good therapeutic approach could be to engineer the tumor to produce very high levels of the ligand, to get a strong immune response against it," Raulet said. "This may combine well with other immunotherapies being used to get the immune system jazzed up."

He cautioned that clinical trials must determine if the approach will work in humans.

The team is continuing its studies to see if the technique works as a cancer therapy. They have achieved encouraging preliminary results by injecting both transformed and untransformed tumor cells at the same time, and eventually will test the effect of injecting the transformed cells into mice that already have established tumors.

The research was supported by the National Institutes of Health.