Dr. Fong said that the target protein carcinoembryonic antigen is overexpressed in colorectal cancer, non-small cell lung cancer, and breast cancer. Dendritic cells are potent antigen-presenting cells that prime T-cell immunity to new antigens. However, they are rare, representing less than 1% of all white blood cells. To increase their number, Dr. Fong and colleagues injected 12 patients who had metastatic lung or colorectal cancer with a hematopoeitic growth factor Flt3 ligand (20 mcg/kg), which expanded the number of dendritic cells twenty-fold.

"This turns the patient into a dendritic cell factory," said Dr. Fong. The same ligand also expands monocytes and bone marrow precursors, he added.

In the next step, researchers used leukapheresis of two blood volumes to harvest the dendritic cells. The cells were then loaded with an epitope derived from carcinoembryonic antigen and injected into the patients. Dr. Fong asserted that the antigen-loaded dendritic cells could then prime CD4 and CD8 T-cell immunity in vivo with minimal toxicity.

"This treatment can lead to a clinical response correlated with in vivo expansion of tetramer+ CD8 T-cells," he said.

Clinically, tumors regressed in two patients, a mixed response occurred in one additional patient, and previously progressive disease stabilized in two more. There were no serious treatment-related side effects.

More important than the specific clinical response, said Dr. Fong, was the display of proof of concept. Previous work has concentrated on more immunogenic tumors such as those of non-Hodgkin's lymphoma, but Dr. Fong's trial showed the technique could work against solid tumors, as well. The advantage of expansion of dendritic cells within the patient's body also meant that less manipulation was needed.

"You're using the patient's own cells, not trying to manipulate other cells to turn them into dendritic cells," Dr. Fong said.

This process may have applications for other types of cancers. In prostate cancer, for instance, prostate specific antigen might be used as the target protein rather than carcinoembryonic antigen, yet the principle would remain the same.