| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© Oxford University Press 2007.
NEWS |
RESOURCEFUL RODENTS
Personal Mouse Colonies Give Hope for Pancreatic Cancer Patients
Despite much talk about individualized therapy, nearly all cancer patients are still treated based on clinical trials of patients with general tumors, not the particulars of their specific tumor.
"There are many drugs for which we don't have a clue on how to select patients," said Carlos Arteaga, M.D., of Vanderbilt University in Nashville, at the American Association for Cancer Research International Conference on Molecular Diagnostics last September.
The search for predictive tumor biomarkers has been hampered by the shortage of available tissue banks and delayed by the need to validate such markers in prospective clinical trials. But researchers at the Johns Hopkins Kimmel Cancer Center in Baltimore are trying one creative new approach to speed things up. There, a group led by Manuel Hidalgo, M.D., Ph.D., is creating large mouse colonies with tumor tissue obtained from individual patients with pancreatic cancer. By regrowing these tumors in mice—up to 150 mice per patient—enough tumor tissue becomes available for both extensive biomarker studies and for tests to see whether the biomarkers predict drug response. And, in an immediate effort to help patients, the Hopkins group has launched a clinical trial using the mice to help choose the best drugs for treating the pancreatic cancer patients after their tumors return. In theory, each patient's tumor will respond best to the drug that is most active against that patient's personal panel of mouse tumors.
"It's an exciting approach for individualizing therapy," says Daniel Von Hoff, M.D., head of translational research at the Translational Genomics Research Institute in Phoenix, Ariz. "Plus, [it] allows us a rather more rapid system to screen for new agents—or evaluate the agents we have, to see if they're worth taking into patients with pancreas cancer."
Mining for Markers
The mouse strategy originated more than a decade ago in the work of Johns Hopkins molecular biologist Scott Kern, M.D., who first implanted pancreatic cancer fragments in mice as a way of generating enough tissue for genetic studies. The research caught the attention of Hidalgo's group. "We thought ... why don't we take it from there and propagate [the tumors] more, and we can test drugs?" recalled Antonio Jimeno, M.D., Ph.D., who is coordinating the project. They began their work in 2005.
|
After surgery to remove the patient's tumor, a piece is given to technicians to implant under the skin in immune-deficient (nude) mice. About 80% of these mice develop full tumors, and fragments of these xenografts are then passed to two more generations of mice to build the colony. The whole process, including drug testing, takes 6–8 months. Because pancreatic cancer patients typically relapse in 9–12 months, the results should be available to guide an individual patient's next treatment. "We certainly hope that we will be able to be on time for most patients," Jimeno said. The immediate goal is to validate the mouse drug testing model by testing whether the patients respond to the same agents as the mice do.
The Hidalgo group has demonstrated that the mouse tumors retain their main genetic features despite passing across three or four generations of mice, putting to rest concerns that intergenerational passage would change the tumors and invalidate the model. But the system isn't perfect. For example, the mouse tumors do not metastasize—"a potential bias of the system," Jimeno noted.
And the overall effort is labor intensive and expensive. Although the Hopkins group is exploring ideas for testing tumors from other institutions, such individualized testing is not likely to spread beyond specialized academic centers. "There are enormous logistic limitations [for it] to be applied broadly," Jimeno acknowledged.
They hope that these individual biomarker studies will collectively yield a panel of predictive markers that will benefit all pancreatic cancer patients. Standard biomarker discovery is hard, because primary pancreatic tumor tissue can be scarce (and thus cannot be used for drug testing), cell lines harbor many genetic abnormalities, and traditional mouse xenografts poorly predict cancer response to drugs (see J Natl Cancer Inst 2006; 98:1176–8). The Hopkins model offers abundant live, genetically intact tissue for testing whether biomarkers can predict a therapy's effectiveness. So this approach, in theory, offers a more direct path to predictive pancreatic cancer biomarkers.
Personal Problems
As a tool for marker discovery and drug development, the Hopkins model has excellent potential, said James Abbruzzese, M.D., chairman of gastrointestinal medical oncology at the University of Texas M. D. Anderson Cancer Center in Houston. He's somewhat skeptical, however, about using mouse results to guide individual treatment. For one thing, tumors are not uniform, and "we don't know for sure that the metastases that develop in a given patient are going to be phenotypically the same as the tumor that came out and grew in the mouse," Abbruzzese said. So a drug that works against the primary tumor in the mouse may not work against the human metastases.
|
Another issue is money. Should doctors, on the basis of these tests, recommend off-label use of expensive cancer drugs not covered by insurance? (Only two drugs, gemcitabine and erlotinib, are approved now for pancreatic cancer.) Some patients may choose to pay, but others may be unable to. "This becomes a real societal issue, on top of a scientific one," Abbruzzese said.
Finally, the use of patient tumor samples to guide individual therapy remains unproven, despite more than three decades of clinical trials. "All of this has a very strong intuitive appeal," Abbruzzese said. "I'd love to, as an oncologist, take a tumor, study it, find out what that particular patient's tumor is sensitive to, and then treat the patient more specifically. My concern is that it's not going to be quite that easy."
Similar tests for guiding therapy, called chemoresistance assays and chemosensitivity assays, usually involve testing drugs in vitro against the patient's isolated tumor cells obtained from surgery or biopsy. (A few involved mouse tumor models.) Two careful reviews, published in the Sept. 1, 2004, issue of the Journal of Clinical Oncology, concluded that randomized trials are still needed to demonstrate that these assays work. Although several studies did show a higher response rate for assay-guided therapy than physician-guided therapy, few resulted in a survival advantage. "There does not appear to be a single assay that is ready for routine integration into the clinical setting," concluded one of the reviews, by a American Society of Clinical Oncology (ASCO) working group.
The ASCO group did add that new clinical trials of predictive assays should be a "priority" but few such trials are done anymore because of lack of funding. "We could never find any additional support to keep these trials going," said Von Hoff, who performed one of the few randomized trials of assay-guided therapy (See J Natl Cancer Inst 1990;82:110–6). So Von Hoff is heartened by the new mouse effort. "The Hopkins team should be applauded for taking this on," he said.
It's far too early to know whether the Hopkins personalized approach will work. The biomarker studies have already begun yielding candidate predictive markers, however, and many more will probably be found. Correlating them with drug response will then follow. As for the predictive therapy element, tumors from more than half of the 40 patients who have consented to the clinical trial have already been xenografted, but the final testing results will not be in for a few years. Patients have been eager to participate, Jimeno said. Because the 5-year survival rate for pancreatic cancer is less than 4%, the Hopkins mouse platform could make an impact. "It's a uniformly fatal disease, and one that robs the patients who are diagnosed of hope," said Johns Hopkins pathologist Ralph Hruban, M.D. "This approach provides a window of hope."
Related Articles in JNCI
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 1990 82: 110-116.
J Natl Cancer Inst 2006 98: 1176-1178.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||