• Home
  • News
  • Calendar
  • About DF/HCC
  • Membership
  • Visitor Center

Understanding environmental and genetic factors in GI malignancies

CDK8 was identified as a colon cancer oncogene by integrating three different genomic approaches. Specifically, CDK8 was the only gene that was essential for proliferation of colon cancer cells, regulated beta catenin activity (known to be important in colon cancer) and amplified in colon cancers. Subsequent analysis showed that up to half of human colon cancers have abnormally high copies of the CDK8 gene.

Whether you are currently studying gastrointestinal cancers or not, Charles Fuchs, MD, MPH (DFCI) may well try to recruit you as a member of the Gastrointestinal (GI) Malignancies Program at DF/HCC. He reaches across Harvard institutions and all scientific disciplines to find investigators whose extraordinary research in other cancers may offer new insight into GI malignancies, which encompass cancers of the digestive tract as well as biliary tree and neuroendocrine tumors.

The close collaboration among the diverse group of investigators in the program has recently led to a SPORE grant — one of only five in the country — and sparked additional funding for the program’s ambitious agenda: to sort out the entire spectrum of genetic mutations driving and sustaining colorectal (CRC) and pancreatic cancers, to understand the underlying biology and environmental factors that increase risk, and to design novel strategies to improve treatment and prevention.

The impact of vitamin D

One of the best examples of interdisciplinary collaboration, says Fuchs, involves ongoing research using two Harvard cohorts: the Nurses’ Health Study (NHS) and the Health Professionals Follow-Up Study (HPFS). This research has already identified genetic and environmental risk factors — such as vitamin D deficiency — in colon and pancreatic cancers as well as factors influencing patient survival. In one SPORE project, co-PIs Fuchs, epidemiologist Edward Giovannucci, MD, ScD (HSPH), and public health researcher Karen Emmons, PhD (DFCI), are leading a team of investigators to define the optimal doses of vitamin D supplementation required to study chemoprevention of colorectal cancer (CRC) in American blacks — and eventually to reduce racial disparities in risk and mortality.

This work evolved from previous analysis of the HPFS which found that blacks and African Americans have about half the blood level of vitamin D (15 ng/mL) compared to whites (30 ng/mL). Most of the vitamin D in the body comes from sun exposure, not diet, explains Giovannucci, and dark skin pigmentation slows and reduces vitamin D production. Differences in vitamin D levels may account for the higher incidence of CRC and other GI cancers found among blacks, says Giovannucci, especially those living in the northeastern United States where solar intensity is low. “Increasing evidence shows an association between lower vitamin D level and higher risk of these cancers,” he says. Yet it remains unclear which dose of vitamin D supplementation is sufficient to elevate blood levels of blacks to 30 ng/mL, presumed to be protective against CRC and other cancers of the digestive tract.

“Piggybacking” on the Open Doors to Health project — a CRC prevention trial, led by Emmons and colleagues, among blacks in low-income housing sites — Giovannucci is enrolling 320 black participants from that cohort for the vitamin D study. He and colleagues are testing three doses of vitamin D supplementation: 1,000 I.U., 2,000 I.U., and 4,000 I.U. The study will also analyze interim endpoints, such as inflammation, and study the genes related to vitamin D metabolism to understand which, if any, correlate with higher or lower levels of vitamin D.

“This study will give us the preliminary evidence needed to conduct a randomized intervention trial,” explains Giovannucci. “It has been poorly appreciated how much vitamin D is required to achieve optimal levels. The recommended daily intake may turn out to be 10 times too low for this population.”

Untangling genetic pathways of colorectal cancer

Other research in the GI program is focusing on signaling pathways. The beta-catenin pathway, for example, is altered or dysregulated in about 90% of colorectal cancers, says Ron Firestein, MD, PhD, an instructor of pathology at BWH and a postdoc in the laboratory of William Hahn, MD, PhD (DFCI). Despite its profound influence in CRC, however, the beta-catenin pathway has eluded scientists searching for molecules that modulate its signaling. A recent study, led by Hahn, took a rational genomic approach to finding a beta-catenin protein that might be a potential therapeutic target, and focused on enzymes, known for their drug-amenable properties.

After investigators engineered a colon cancer cell line with a high level of beta-catenin activity, they conducted two loss-of-function RNAi screens. The first screen, which targeted ~1,000 genes - those that produce all the human kinases and phosphatases as well as other proteins linked to cancer — turned up 34 enzymes essential for beta-catenin activity. The second screen, which cast a wider net to capture all the genes implicated in cellular proliferation, found ~160. “When we compiled the two screens,” says Firestein, “we discovered nine genes that were responsible for both beta-catenin activity and colon cancer cell proliferation.” Hahn and Firestein then called on Adam Bass, MD, a postdoc in the laboratory of Matthew Meyerson, MD, PhD (DFCI), to find out whether any of these nine genes were also altered in samples from patients with colorectal cancer. A SNP analysis of regions of DNA amplification revealed that a single gene, cyclin-dependent kinase 8 (CDK8), harbors copy number gains in a significant subset of the colon cancers.

Eager to understand how CDK8 regulates beta-catenin, Firestein used RNAi once again and found that suppressing the CDK8 gene reduces beta-catenin binding to transcription factors, such as c-myc . “This suggests that CDK8 directly impacts beta-catenin activity,” explains Firestein. However, the far more interesting finding, he says, is that the kinase activity of CDK8 is required to impact beta-catenin and mediate changes in cell growth. “When you inhibit CDK8’s catalytic activity, it can no longer function as an oncogene,” he notes. “We now have a starting point to test whether any of the cyclin-dependent kinase inhibitors now in clinical trials for other cancers might also target CDK8 in colon cancer.” In the meantime, he and colleagues are attempting to solve the many other pieces of the CDK8 puzzle and to find all the target genes that CDK8 activates. 

Emulating the biology of pancreatic cancer

Of all the GI malignancies, pancreatic ductal carcinoma (PDA) is among the most lethal. In an effort to understand the molecular and cellular mechanisms underlying PDA, Ron DePinho, MD, director of the Center for Applied Cancer Science at DFCI, and Nabeel El-Bardeesy, PhD, a basic scientist at MGH, have engineered mouse models with myriad combinations of the signature gene mutations found in human PDA: the Kras oncogene and the tumor suppressors Ink4a/Arf, p53, Lkb1, and Smad4.

These models have helped investigators elucidate how specific mutations affect progression of the disease through one of the three known precursor lesions: PanIN (pancreatic intraepithelial neoplasia), MCN (mucinous cystic neoplasia), or IPMN (intraductal papillary mucinous neoplasms). Bardeesy and DePinho discovered that activating mutations in Kras initiate PanIN, which if followed by mutations in p53 or Ink4a/Arf eventually progress to PAD; if Kras activating mutations are followed instead by a Lkb1 mutation, the disease advances to IPMN, which may also lead to adenocarcinoma; but if Kras mutations are followed by Smad4 mutations, a combination of IPMN and MCN results.

Inspired in part by the work of Jeffrey Engelman, MD, PhD (MGH) in lung cancer, Bardeesy and colleagues are now testing novel combinations of inhibitors in their mouse models. “Since PI3K is required for Kras to transform cells, Jeff’s work in the lung suggests that the combination of a PI3K and a Mek inhibitor might be effective in pancreatic tumors.” (See story on Engelman’s work). “We want to see whether a different constellation of mutations will respond to this combination treatment” — the very first in vivo study of two targeted therapies in pancreatic cancer.

“The GI program fosters increasingly effective collaborations, like this one, across institutions and disciplines,” remarks Fuchs. “Bringing investigators together toward a common cause is helping us understand the biology of GI cancers and develop more effective treatments for patients.”

—Lonnie Christiansen