Heightened exposure

Researchers promote exposomics to clarify cancer's contributors


  • Bryn Nelson PHD

After a lengthy probe in Toms River, New Jersey, researchers found statistical associations that tied a cluster of childhood cancers to chemical contamination. They could not, however, provide definitive closure for a town beset by industrial pollution. An even longer investigation in North Haven, Connecticut, found that workers at Pratt & Whitney, a jet engine manufacturer, were no more likely than the general population to develop malignant brain tumors. However, researchers could not allay all lingering suspicions.

In communities, factories, and other sites, epidemiologists and fellow researchers have sought explanations for seemingly high rates of leukemia, brain cancer, and other diseases that appear to be occurring in suspected clusters. Given the formidable difficulty of these studies (and the relative lack of them), some scientists say establishing any associations at all should be considered a success.

In the quest to help resolve some of the ambiguity, a growing number of researchers are promoting an ambitious concept in epidemiology. Scientists are pointing to exposomics, the study of the human exposome, as a way to provide a more objective analysis of potential disease contributors.

Rethinking the Instigators of Cancer

The term refers to an individual's lifetime exposure to radiation, pollution, drugs, diet, and other nongenetic input from the time of conception onward. How these external cues alter our internal chemical milieu, from endocrine disruptors and proteins to immune modulators and metals, can be measured by gauging changes in biomarkers found in blood or serum.

Stephen M. Rappaport, PhD, professor of environmental health at the University of California at Berkeley, says the goal is to account for the combined influence of everything except a person's genes on health and disease. “All of these different pieces of the exposome can be investigated separately, and ultimately integrated into an analysis to really get at what the potential nongenetic causes of disease might be,” he says.

The total contribution is likely to be significant. Although some well-known genomic abnormalities have been strongly linked to cancer development, they account for only a small percentage of the total variation in cancer incidence. “So if it's not the genes, or at least they're not acting alone in causing this disease, then what else is it?” Dr. Rappaport says.

Margaret R. Spitz, MD, MPH, professor of epidemiology at Baylor College of Medicine in Houston, Texas, agrees that researchers are reevaluating what they thought they knew about cancer genetics. “We're now learning that we need to integrate exposure information together with looking at the genetic architecture of the tumor and looking at the somatic changes,” she says. Exposures might have unique genetic profiles, she says, meaning that researchers could observe a sort of signature of a specific exposure in the tumor. Deciphering the individual signatures may depend on understanding how dynamic changes to DNA, epigenetic alterations, turn genes on and off in response to external cues. Traditional efforts to assess which external cues help to drive oncogenesis, however, have been hampered by a long list of potential limitations.

A Difficult Target

One challenge is what epidemiologist Kenneth Rothman, DrPH, described as the “Texas sharpshooter” problem. Gary Marsh, PhD, director of the Center for Occupational Biostatistics and Epidemiology at the University of Pittsburgh Graduate School of Public Health in Pennsylvania, explains it this way: a cowboy shoots the side of a barn, then draws a circle around his shots and says, “Look, I hit the bullseye every time.” In the same way, Dr. Marsh says, investigators analyzing suspicious cases often put circles around them and artificially magnify the problem, rather than first drawing a circle around a well-defined population and then seeing how many cases fall within the boundaries.

Detectable effects can be constrained by the size and diversity of communities, and by the difficulty in defining actual exposures. Intense public fear and anger over suspected cancer clusters also can lead to reporting bias. Many cancers have low incidence rates in the general population, and among solid tumors, some can take 20 years or more to develop.

“For rare diseases, you need a large population followed over a long period of time to generate a sufficient number of cases for a formal study that can detect a true excess in cases over an appropriate background rate,” Dr. Marsh says. Full-scale residential studies have been time-consuming, expensive, and rare.

Signs of Progress

Dr. Marsh has maneuvered around some of the limitations by focusing his efforts on occupational exposures. Workers often have much more defined exposures, based on their shifts and job responsibilities, and companies generally keep decades' worth of employee records. “The reasoning behind these studies is that, if you're going to see something related to this exposure, your chances of seeing it are greater in a group of workers who are exposed to it every day,” he says.

After a 10-year investigation of glioblastomas and other brain cancers at Pratt & Whitney, Dr. Marsh and his colleagues recently reported that overall they found no statistically significant increase in cancer rates among workers.[1] The massive study, funded by the manufacturer, analyzed the company records of approximately 223,000 workers at 8 sites over 50 years.

However, therein lies another conundrum: researchers often must rely on the very companies accused of exposing workers to toxins to provide records and pay for at least part of the investigation. Dr. Marsh says he and his collaborators insisted on and received full access to available records and the academic freedom to publish all findings. Nevertheless, he concedes that it can be hard at times to fully shake the perception that the results are somehow biased when a company or trade organization is involved.

Accessing an archive of past exposures housed within an individual's body may sidestep questions about bias while offering a more holistic view of cancer contributors. Cataloging a lifetime of these potential contributors will be no small undertaking. Then again, Dr. Rappaport says, researchers have already developed multiple methods for measuring metals, chemicals, and other molecules in the blood, and more are currently in development.

As a proof of concept that a blood-based analysis of specific molecules can reveal previously hidden links to cancer, he points to recent reports by researchers in Saskatchewan, Canada. Their latest article suggests that lower-than-normal blood serum levels of a recently identified metabolite known as gastrointestinal tract acid-446 (GTA-446) are strongly associated with an increased risk of developing colorectal cancer.[2] Linking molecules such as GTA-446 to specific cancers, either as protectors or aggressors, will require extensive research, but Dr. Rappaport says, “we will do it because it's the only way we're ever going to find out what really causes these chronic diseases.”

Although still in its infancy, exposomics has attracted the notice of several agencies. In 2012, the European Commission awarded a combined $22.4 million to 2 exposome studies, including one in which sensor-equipped smart phones will measure participants' exposures. And in the United States, the National Institute of Environmental Health Sciences awarded a $#x00024;4 million grant to establish HERCULES (Health and Exposome Research Center: Understanding Lifetime Exposures) at Emory University in Atlanta, Georgia. Researchers such as Drs. Spitz and Marsh are enthusiastic about the potential, if cautious about when exposomics might become a force. “It may take decades before this gets into the mainstream practice of epidemiology,” Dr. Marsh says, “but it's got to start somewhere.”