"The clinical result was far below our ultimate goal, no question," said Robert Vonderheide, senior author on the study and an associate professor at Penn's Abramson Family Cancer Research Institute. "Now we have insights that will allow us to really go forward in an accelerated way."
Usually diagnosed only after it has spread and begun causing symptoms, pancreatic cancer kills about 37,000 patients a year. Fewer than 6 percent are alive five years after diagnosis.
Treatments that harness the body's immune system are a hot area of cancer research now. Last year, Provenge, a vaccine for prostate-cancer patients, became the first FDA-approved cancer vaccine.
The Penn pancreatic-cancer study, Vonderheide said, shows that immune treatments "might be more complex and carry more possibilities than we previously thought."
In an unusual approach that Vonderheide believes will speed drug development, Penn tested the new monoclonal antibody - Pfizer's CP-870,893 - in 21 people while doing a parallel study with a similar drug in mice. The mouse study helped researchers understand what was going on in the tumors at a cellular level.
"More and more at Penn, this is a high priority," he said, referring to linked animal and human studies. "It's only been possible because the type of animal model has improved so much in the last five years."
In the study, people with pancreatic cancer were given gemcitabine, a standard chemotherapy drug, and the experimental drug, which binds and stimulates a cell surface receptor called CD40. It regulates T cell activation. The researchers thought the T cells would attack the tumor.
Although all of the patients eventually relapsed and died, five had tumors that shrank more than 30 percent, Vonderheide said. The surprise was that scientists didn't find T cells in tumor samples. Instead, they saw lots of the more basic cells called macrophages that usually get co-opted into helping the tumors.
They got the same results with treated mice. They then discovered that the macrophages were attacking the stroma, a layer of scar tissue that surrounds pancreatic tumors. Though the stroma does not itself include cancer cells, it can account for the bulk of the tumor and can promote tumor growth in multiple ways. It also can serve as a barrier that prevents chemotherapy drugs from reaching cancer cells.
The result of the treatment was that both the stroma and cancerous areas shrank.
Now researchers will have to figure out why the treatment wore off and what kinds of drug combinations might improve its effectiveness.
The results hold hope for other solid tumors as well.
Penn is not currently testing the Pfizer drug - which had few side effects - in pancreatic-cancer patients, but it is enrolling patients in a metastatic-melanoma trial.
Jonathan Brody, a pancreatic-cancer expert who is co-director of the Jefferson Pancreas, Biliary Tract and Related Cancer Center, said the small number of patients, lack of a current control group, and some differences between human and mouse cancer cells were cause for caution, but "as a first study it's a very promising approach."
Given the size of the stroma in pancreatic cancer, he said it made sense to pay more attention to it. The new study, he said, underscored "the importance of going after the environment that surrounds the pancreatic tumor cells. . . . They're really opening up a new paradigm in which to go after the disease."
Christina Tabarrini, local coordinator for the Pancreatic Cancer Action Network, said a couple of extra months can mean a lot to a pancreatic cancer patient. "It could mean someone being able to walk their daughter down the aisle," she said. "Unfortunately, you don't have many options with this disease."
Tabarrini, a Schwenksville woman whose father died of pancreatic cancer in 2007, has been lobbying for more government funding for research. She thinks immune treatments deserve more study. "At this point, you've got to look at everything," she said. "If it's out there, let's research it."
Contact staff writer Stacey Burling at 215-854-4944 or email@example.com.