Researchers identify biomarkers of early-stage pancreatic cancer in mice and man

Jun 10 2008

A multicenter team of researchers has identified a panel of proteins linked to early development of pancreatic cancer in mice that applies also to early stages of the disease in humans - a breakthrough that brings scientists a significant step closer to developing a blood test to detect the disease early, when cure rates are highest.

The findings, by senior author Samir "Sam" Hanash, M.D., Ph.D., of Fred Hutchinson Cancer Research Center, will be published in the June 10 issue of PLoS Medicine, a freely available, open-access online journal. Investigators from Massachusetts General Hospital, University of Michigan and Belfer Institute for Innovative Cancer Science at Dana-Farber Cancer Institute collaborated on the research.

"Our team identified, for the first time, protein changes associated with early-stage pancreatic-tumor development in genetically engineered mice that were also found to be associated with the presence of the disease in humans at an early, pre-symptomatic stage," said Hanash, head of the Hutchinson Center's Molecular Diagnostics Program. Hanash is an international leader in the field of proteomics, which seeks to identify clinically relevant trace proteins, or biomarkers, that are leaked by tumors into the blood. Scientists believe that such proteins could be used in screening blood tests for early and more accurate detection of cancer and other diseases.

"Our findings represent a breakthrough in the application of advanced proteomic technologies and mouse models to cancer-biomarker discovery," said Hanash, also a member of the Hutchinson Center's Public Health Sciences Division.

The five-biomarker panel, if developed into a commercial screening test, may be particularly useful when combined with a currently available test that measures a pancreatic-cancer biomarker called CA19.9, which is elevated in 80 percent of newly diagnosed patients but is not linked to asymptomatic, early-stage disease. Together, the biomarker panel and CA19.9 may significantly improve the detection of early-stage disease prior to the onset of symptoms and may also help better distinguish between cancer and pancreatitis, a noncancerous, inflammatory condition.

Early detection of cancer is crucial for long-term survival. Most solid tumors can be cured 90 percent of the time if they're detected and treated early, whereas cure rates for late-stage cancer are about 10 percent. Early detection is particularly relevant to pancreatic cancer, which is the fourth-leading cause of cancer death in the United States, with a five-year survival rate of only 3 percent. Because the disease is asymptomatic in the early stages, most patients are not diagnosed until the cancer has spread beyond the pancreas, which contributes significantly to the poor long-term survival rate.

"There is a substantial challenge in studying the early molecular changes in pancreatic cancer because most patients are diagnosed with advanced-stage disease and so there is a lack of suitable specimens for biomarker discovery," said paper co-author Nabeel Bardeesy, Ph.D., of the Massachusetts General Hospital Cancer Center.

Finding telltale proteins that can signal the earliest stages of cancer development can be like looking for the proverbial needle in a haystack, as blood contains a complex mixture of thousands of proteins. In addition, any two proteins may exist in concentrations more than a million-fold different from one another. "The ones that are likely to be useful for diagnosing cancer are probably the ones that exist at the lower end of the range, which makes them very hard to find with standard methods," Hanash said.

To guide their quest, Hanash and colleagues employed a variety of technologies to identify, measure and analyze blood proteins in mice and man. Since every protein is different, each has the equivalent of a distinguishing molecular "bar code." The goal is to identify protein signatures that are only present in cancer, which may then serve as biomarkers to detect early disease.

A significant boon to biomarker research - and one of the great scientific advances in the past century, according to paper co-author Ronald DePinho, M.D. - has been the incorporation of mouse models into many disciplines of science, including cancer research. "Our ability to now engineer mice with the same mutations that drive specific cancers in humans has provided powerful and accurate model systems to study virtually all aspects of the disease and then translate these new insights into improved prevention, detection and treatment strategies for cancer," said DePinho, director of the Belfer Institute at Dana-Farber Cancer Institute.

For this study, the researchers first analyzed blood samples from genetically engineered mouse models of pancreatic ductal adenocarcinoma at both early and late stages of tumor development. Of nearly 1,500 proteins identified in these mice, five were associated consistently with a precancerous condition known as pancreatic intraepithelial neoplasia, or PanIN, which, if left untreated, eventually progresses to full-blown pancreatic cancer.

The researchers then sought to determine whether the same biomarkers turned up in blood samples obtained from 30 recently diagnosed pancreatic-cancer patients. They also looked for the biomarkers in 13 people with asymptomatic, early-stage pancreatic cancer who had donated blood for another, unrelated study within a year of their cancer diagnosis. For comparison purposes, the researchers analyzed blood from 20 healthy subjects and 15 people diagnosed with chronic pancreatitis.

Next steps in the research will include validating additional candidate biomarkers and further testing of the biomarker panel they have assembled to see how well it distinguishes between pancreatitis and pancreatic cancer. The researchers also want to continue testing the value of a biomarker-panel approach for early detection of pancreatic cancer among those at increased risk, such as people with a strong family history of the disease.

The National Cancer Institute Mouse Models of Human Cancer Program, NCI Early Detection Research Network, National Institutes of Health, Canary Foundation, Paul Allen Foundation, Waxman Foundation, Verville Family Foundation and Deutsche Forschungsgemeinschaft funded the research.