Researchers discover tumor molecule that causes spread of pancreatic cancer

Testing could help guide treatment
An illustration of a traffic cop
Fred Hutch researchers have found that a tumor molecule called RUNX3 acts like a traffic cop of sorts, determining whether pancreatic cancer cells will spread, or metastasize, to distant parts of the body or whether they will stay put. Illustration by Kim Carney / Fred Hutch News Service

A single molecule switches on metastasis, or spread, in pancreas cancers, reports new research led by scientists at Fred Hutchinson Cancer Research Center. The researchers predict that testing for this molecule, called RUNX3, could soon help oncologists choose the most appropriate treatments based on the metastatic potential of each patient’s disease.

“We’re defining a readout that may help doctors in their approach to treatment of patients who have pancreatic cancer,” said Dr. Martin “Marty” Whittle, a postdoctoral researcher in the lab of Dr. Sunil Hingorani and first author on the paper, which was published online today in the journal Cell. “The gene that we identified can be used to give some insight as to whether a patient’s tumor is more likely to grow locally or metastasize.”

RUNX3, the researchers found, controls the activation of numerous genes involved in metastasis in a mouse model, triggering cancer cells to migrate to other parts of the body and turning on genes that help those metastatic cells take root and thrive once they invade distant tissues.

“It’s extraordinary — it seems to control an entire metastatic program,” said Hingorani, the senior researcher on the study and a physician-scientist at Fred Hutch who specializes in pancreatic cancer. “RUNX3 serves to both expel the seed and prepare the soil.”

A ‘devastating’ reality for patients

Pancreatic cancer has the highest metastatic drive of any malignancy, Hingorani said. By the time they are diagnosed, most pancreas cancer patients already have either metastatic disease or tumors that have grown too much to be surgically removed. Even when patients’ tumors are contained, making them eligible for surgery or focused radiation, many of them end up dying of metastatic cancer anyway — because their apparently localized tumors had nevertheless already started spreading to other sites.

For this reason, oncologists often treat patients who have early-stage pancreas tumors with chemotherapy before surgery in the hopes of killing off any distant microscopic metastases and prolonging life. But in the time it takes for a few rounds of chemo, almost a third of surgically removable tumors grow too large to be eligible for surgery, a “devastating” outcome, said Hingorani, which slashes patients’ average survival time from two years for a removable tumor to less than 11 months for a non-operable tumor.

By predicting a tumor’s metastatic behavior, doctors could choose the type of treatment that gives their patients the best chance at the longest survival time, Hingorani said. 

Dr. Sunil Hingorani
Dr. Sunil Hingorani is confident that his team's findings in a mouse model of pancreas cancer may soon be validated in humans. Photo by Stefanie Felix for Fred Hutch

“Even as we hope to develop more rational and targeted therapies for pancreas cancer, we can and must more intelligently apply the treatment modalities we have now,” he said. “And understanding how and when to use chemotherapy and radiation in the ideal context has the potential to impact patient survival and quality of life in the near term.”

With the caveat that they still must be validated in humans, Hingorani believes that his team’s findings will do just this.

“I haven’t found a compelling explanation yet for the unusual metastatic drive of pancreas cancer, and certainly not one that would reconcile some of the paradoxes that exist both in the fundamental biology but also in treatment response in patients,” Hingorani said. “So for us, the thing that, on one hand, was worth the many years we spent studying it and why we finally felt ready to communicate it to the world, is that it helps us understand both the biology of the disease and ― most excitingly I think ― it might soon influence what we do in the clinic.”

Pinpointing a role in disease

The researchers relied on their mouse model of pancreatic ductal adenocarcinoma (or PDA, the most common type of pancreatic cancer), which they developed more than a decade ago to faithfully mimic the accumulation of pancreas cancer mutations seen in human disease.

In this study, they varied the number of copies of a gene called Dpc4, a tumor suppressor gene, in the mice. Previous research in pancreatic cancer has linked both the presence and absence of Dpc4 to metastasis, a confusing picture the researchers hoped to clarify by examining the effects of deleting one versus two versus zero copies of the gene, an approach first suggested by a paper that Hingorani’s team published in 2007.

Then, they zoomed in on the signaling pathways activated by each combination of genes in the mice and tracked how the disease developed.

Each of the groups of mice developed a cancer with different potential for metastasis and primary tumor growth, they found. Why? A little-known transcription factor called RUNX3 stood out.

RUNX3, they found, is the focal point of several signaling pathways involved in pancreatic cancer metastasis, and it sits right at the decision point, molecularly speaking, of whether a cancer will stay put and proliferate in place or start sending metastatic cells all over the body.

When RUNX3 was activated in tumors with either two or zero copies of Dpc4, it, in turn, activated genes involved in metastasis. When there was just one copy of Dpc4, curiously, RUNX3 levels were very low, and the cancer focused its energy on local growth. And the picture stayed the same when the researchers followed up with human PDA cells.

Dr. Martin "Marty" Whittle
"The gene that we identified can be used to give some insight as to whether a patient's tumor is more likely to grow locally or metastasize," said Dr. Martin "Marty" Whittle. Photo by Bo Jungmayer / Fred Hutch News Service

Informing treatment choices

Because high RUNX3 means high metastatic potential and relatively lower risk of death from primary tumor growth, treating these patients first with systemic chemotherapy to kill distant metastases before surgically removing the tumor would offer patients the best chance, the researchers predicted. But if RUNX3 is high and the patient has lost both functional copies of the Dpc4 gene ― an “especially lethal” combination, Hingorani said, which happens in half of patients with PDA ― both primary tumor growth and metastasis are risks, which might mean that going after the primary tumor first might make the most sense.

Conversely, if RUNX3 is low, metastasis is a relatively low risk. It would then make the most sense to focus on treating the primary tumor with surgery and perhaps radiation, the authors predicted.

This finding could be built in right away into ongoing clinical trials and, if borne out in human studies, it could alter clinical decision-making, said Dr. Anirban Maitra, who was not involved in this research. Maitra is co-director and scientific director of the Sheikh Ahmed Bin Zayed Al Nahyan Center for Pancreatic Cancer Research at The University of Texas MD Anderson Cancer Center.

“Because of what the field has seen in potential discriminators of metastases in the past with Dpc4 ― and yet the lingering imperfections of that dichotomy that Dpc4 alone provides ― [it] has suggested for a long time that there’s clearly something else going on,” Maitra said. “With all the caveats ― and again, this is a mouse study and it needs significant validation in both prospective and retrospective datasets ― but with all of those caveats in mind, we now have something that you can assess in tissue specimens. The same biopsies [in which] you’re looking for Dpc4 histochemistry, now you can look and see what are the RUNX3 levels, and you can stratify patients based on that.”

Hingorani’s group already has studies underway to begin to translate their results into the clinic and learn more about how RUNX3 is regulated, they said. They are hopeful that it won’t be long before their discovery plays a role in extending patients’ lives.

“We’ve had such success in this mouse model in mimicking human PDA that we’re very hopeful that will translate well into humans,” Whittle said.

Susan Keown, a staff writer at Fred Hutchinson Cancer Research Center, has written about health and research topics for a variety of research institutions, including the National Institutes of Health and the Centers for Disease Control and Prevention. Reach her at

Solid tumors, such as those of the pancreas, are the focus of Solid Tumor Translational Research, a network comprised of Fred Hutchinson Cancer Research Center, UW Medicine and Seattle Cancer Care Alliance. STTR is bridging laboratory sciences and patient care to provide the most precise treatment options for patients with solid tumor cancers.

Help Us Eliminate Cancer

Every dollar counts. Please support lifesaving research today.