A research team led by scientists at the Whitehead Institute for Biomedical Research has identified a promising new ingredient for cancer vaccines using a new system for assessing anti-tumor activity.
The new ingredient, described in a recent issue of the Proceedings of the National Academy of Sciences, is a natural blood growth factor called GM-CSF, or granulocyte-macrophage colony stimulating factor. In animal studies, a vaccine consisting of GM-CSF modified tumor cells stimulated potent, long-lasting, and specific anti-tumor immunity.
Unlike conventional vaccines meant to prevent disease, cancer vaccines are a form of therapy. They are designed to stimulate the body's own natural defenses to seek out and destroy tumor cells.
Dr. Richard Mulligan, a member of the Whitehead Institute and professor in the Department of Biology at MIT, explains that scientists have been seeking ways to augment anti-tumor immunity for almost a century. Most cancer vaccines have consisted of the patient's own tumor cells, either alone or mixed with a substance known to stimulate the immune system in a non-specific way. "A few cancer patients have appeared to benefit from such vaccines, but the responses generally have been partial or short-lived," Dr. Mulligan said.
"Over the past several years, we and others have developed a new approach to cancer vaccines using gene transfer techniques," Dr. Mulligan explained. "We insert genes into tumor cells to make them produce substances known to stimulate the growth and development of immune cells. The hope is that when the immune system encounters these substances in close proximity to tumor antigens -protein markers on the surface or inside of tumor cells-it will be stimulated to kill all similarly marked tumor cells throughout the body."
First Comparative Study
Federal agencies have approved clinical studies of several different genetically modified cancer vaccines based on data from experiments that examined the activity of a single gene product. The new Whitehead study was the first to compare the relative benefits of many different potential immunoregulatory molecules.
Dr. Glenn Dranoff, a postdoctoral fellow at the Whitehead Institute, Dr. Mulligan, and their collaborators compared the activity of 10 different cytokines, other growth factors, and cell adhesion molecules against a single type of tumor, a rodent melanoma. Some, such as interleukin-2 (IL- 2) and tumor necrosis factor-alpha (TNF-a), had been tested in earlier trials, while others had not previously been considered for cancer vaccines. The researchers also compared vaccines made from live tumor cells with vaccines made from inactivated tumor cells-cells treated with low-level radiation. Remarkably, irradiated tumor cells bearing a foreign cytokine gene were found to produce substantial levels of the desired cytokines even though they could not grow or divide.
"Our study had two very important findings," Dr. Dranoff said. "First, we showed that the inactivated tumor cells performed as well as or better than the live tumor cells in the animal studies. This is important because the manipulation and reimplantation of live tumor cells in cancer patients might pose a small risk of generating a more aggressive form of malignancy. The use of irradiated tumor cells will virtually eliminate this risk.
Specific Protection with GM-CSF
The second major finding was that GM-CSF was the most powerful immunostimulant of the 10 molecules tested. In the mouse model, irradiated tumor cells expressing GM-CSF conferred protection against challenge with unmodified melanoma cells and also led to the rejection of small, pre-established tumors. The protection was highly specific; melanoma cells expressing GM-CSF protected against challenge with melanoma cells but not against challenge with lung cancer cells. (The results with GM-CSF were unexpected because the molecule had been thought to be primarily involved in hematopoiesis, the process by which red and white blood cells are produced.) Irradiated cells expressing the growth factors IL-4 and IL-6 showed a lower level of activity. The Whitehead researchers did not find any activity with IL-2 or TNF-a, two cytokines previously approved for clinical trials.
"We won't know how these results transfer to the clinical setting until we begin tests of irradiated GM-CSF modified tumor cells in patients," Dr. Mulligan says. "Several different collaborations are being established to evaluate the safety and effectiveness of GM-CSF vaccines."
A New Viral Vector
The transfer of this new vaccine strategy to clinical practice will be greatly enhanced by one special feature of the model system described in the PNAS paper: the modified virus used by the scientists to insert cytokine genes into the tumor cells is more powerful than any virus vector previously used for this purpose.
The new vector, developed in the Mulligan laboratory, has a unique structure that greatly increases the efficiency with which it infects target cells. In addition, cells modified with this vector produce higher levels of cytokine proteins than cells modified with other virus vectors.
This increased efficiency is advantageous for several reasons. First, the time required to produce a vaccine is greatly reduced because there is no need to select genetically modified cells from the target population and grow them in culture to obtain sufficient cells for vaccination; also, the population of modified tumor cells is more diverse, thereby increasing the likelihood that any existing tumor cells in the body will be recognized and destroyed.
In addition, reducing the need for selection means that it is relatively easy to introduce multiple gene products into tumor cells-multiple cytokines may make a more effective cancer vaccine than one cytokine alone.
"While we believe that GM-CSF expressing tumor vaccines may have wide application in the treatment of different forms of cancer, it is possible that some types of tumors will respond better to a different cytokine," Dr. Mulligan said. "The new screening system, using the improved virus vector and irradiated tumor cells, will allow us to identify the best cytokine, or cytokine combination, for each type of cancer."
Authors of the PNAS paper in addition to Drs. Dranoff and Mulligan are Katja Brose, Valerie Jackson, and Hirofumi Hamada of the Whitehead Institute; Audrey Lazenby of the Department of Pathology at Johns Hopkins University School of Medicine; and Elizabeth Jaffee, Paul Golumbek, Hyam Levitsky, and Drew Pardoll of the Department of Oncology at Johns Hopkins.
A version of this article appeared in MIT Tech Talk on April 28, 1993.