Dr. Thomas Devers

Predicting Colon Cancer:

UConn Health Researchers Redefine ‘Early Detection’

By Chris DeFrancesco

Illustration of colon with areas highlighted and given warning signs


Find cancer early enough and you can treat it. Predict it before it develops and you can prevent it altogether.

Thanks to volumes of epidemiological data they have amassed, UConn Health researchers believe they are closing in on ways to identify who’s most at risk for colorectal cancer by analyzing cells from lesions that, if they are to become cancerous, are years away from doing so.

Once this can be figured out, doctors and patients would have a larger window of time to take steps to stop the cancer before it starts.

In the laboratory of Daniel W. Rosenberg in UConn Health’s Center for Molecular Medicine, concurrent studies — epidemiological, genomic, and molecular — are ongoing and are expected to yield a series of published papers this year and next. One, which describes how the Rosenberg lab uncovered evidence of the origins of colorectal cancer that historically have been poorly understood, was published in the June edition of Molecular Cancer Research, a major scientific journal of the American Association for Cancer Research.

We believe identifying early molecular changes may uncover new targets that could be used for preventing early neoplasia from progressing.

Cancer evolves from what is known as neoplasia — or new, abnormal growth of tissue. Not all neoplasias become malignant tumors, but they are considered an early warning sign of possible cancer.

“Understanding early neoplasia has become such a major focus at the National Cancer Institute — how do we characterize these early changes so we can prevent cancer,” says Rosenberg, HealthNet Inc. Endowed Chair in Cancer Biology and professor of medicine, “and I believe we’re at the forefront of answering this question.”

MD/Ph.D. candidate Allen Mo, first author on the Molecular Cancer Research paper, says colon cancer is thought to be an epithelial disease, meaning that it starts with a mutation in the tissue that lines the surface of the colon (the epithelium), and grows and then invades the underlying support tissue.

“People historically believed these are separate compartments, that there is no interaction between the epithelium and the support tissue until the cells become cancerous and break through the membrane,” Mo says. “We’ve been able to demonstrate that, even at the very early stage, prior to the polyp stage, the supportive tissue is actually influencing how these epithelial initiate cells are evolving.”

That’s important because it provides another potential early intervention point — if scientists can figure out a reliable way to make alterations to the signaling pathways between the two tissue types, perhaps they could influence how the mutated cells progress.

The Devers Data

The work includes collaborators from both within and outside the institution, but central to all of it is Dr. Thomas Devers, a UConn Health gastroenterologist whose volume of consecutive colonoscopies over the past five-and-a-half years has yielded data from 5,000 patients. The resulting demographic database has been powering the engine driving the research that has helped substantiate epidemiological findings, such as how smoking completely cancels the protective properties of aspirin, how consumption of diets rich in Omega-3 fatty acids appear to reduce risk of early neoplasia in the colon, and how walnuts may have protective properties, as described in a study recently published in Cancer Prevention Research.

“Dr. Devers routinely screens patients at a resolution that only a handful of clinicians are doing,” Rosenberg says.

Devers uses a high-definition endoscope with contrast dye-spray that enables him to detect tiny — less than 5 mm — lesions that are scattered throughout the colon.

“Many of the subjects have already returned for follow-up [surveillance] colonoscopy, so we actually have genomic data from three and five years ago that we can use to predict the possibility they may develop advanced adenomas or even cancer,” Rosenberg says. “We’re actually at the point now where we can follow the impact of these early changes over time.”

The work involves the intensive application of bioinformatics, the collection and analysis of complex biochemical and biological information.

What the Earliest Changes Can Reveal

Among the tiny lesions of particular interest are those known as aberrant crypt foci (ACF), which represent the earliest detectable precancerous change in the human colon. ACF tend to be detected less frequently during conventional colonoscopy, occur throughout the colon, and are the source of tissue that the Rosenberg lab uses for many of its analyses.

“We’re one of the only places in the country that actually look for these very early lesions,” says Mo, whose work has been instrumental to a number of ongoing studies in the Rosenberg lab.

“Most people study colon cancer development in the context of polyps as the earliest lesion,” Rosenberg says. “But we’re going one step earlier, with our focus on ACF. ACF present a unique opportunity to study the risk factors that may predispose the development of colon neoplasia, and may help to guide us toward potential interventions that may actually eliminate neoplasia prior to the appearance of polyps.”

Devers says the ACF he’s been finding in one particular area of the colon can be very telling when it comes to predicting future cancer risk.

“Part of our hypothesis is, we’re going to find these tiny lesions on the right side of the colon that have a lot meaning, that are only present in a small percentage of the population compared to the people who have tiny lesions in the rectosigmoid (the lower part of the colon), which everybody has,” Devers says. “And by finding these tiny lesions in the right side of the colon, you may want to screen those people more frequently.”

Biopsies and data from about 300 patient research volunteers have been the basis of several studies, including a collaboration with scientists at the Van Andel Institute in Grand Rapids, Mich., and The City of Hope in Duarte, Calif.

“We’ve done a complete genome-wide analysis of the epigenetic changes present within these tiny lesions, something that has never been done before,” Rosenberg says. “We’re uncovering all these interesting changes that occur to a person’s epigenetic profile years before they may develop a more advanced neoplasia. Much of this transformative epigenetic work was been performed by Matthew Hanley, a fifth-year graduate fellow in my lab. The question is, why we are interested in this ‘predictive’ profile? We believe identifying early molecular changes may uncover new targets that could be used for preventing early neoplasia from progressing.”

Often Hidden, Likely Telling

Another finding: some people seem to have a higher likelihood of forming what are known as sessile serrated adenomas (SSA) in the upper part of their colon. SSA, which also tend to be harder to detect, carry a strong likelihood of progressing, and may contribute to 20 to 30 percent of colorectal cancers.

“We believe that missed right colon cancers, or interval colon cancers, are related to these serrated adenomas,” Devers says.

SSA are larger than ACF but are also very difficult to catch during colonoscopy because of their flat shape and their tendency to be camouflaged in mucus along the colon wall. It takes a high-definition scope and experience — Devers has both — to find them.

“We’re able to actually identify thepeople who form this lesion, then go back to the epidemiological database and develop risk profiles as to which people are more likely to form that type of lesion,” Rosenberg says.

On the molecular level, Rosenberg’s lab routinely uses laser capture microdissection, which enables scientists to select and retrieve small groups of cells from a single biopsy. From there, they can apply genomic technologies to particular cells and screen for cancer-related mutations and genome-wide alterations. This technique was used in the research behind the Molecular Cancer Research paper.

“Early neoplasia has become a very hot area, and because of Dr. Devers, here we probably have accumulated the largest repository of human early neoplastic lesions anywhere in the world,” Rosenberg says. “With this amazing resource, we can now begin to define many of the key changes that are happening at this very early stage. It’s never been done before.”