Exposure Assessment Methods (Past Initiative)

The National Cancer Institute (NCI) and National Institute of Environmental Health Sciences (NIEHS) are cofunding 11 research projects in response to a Request for Applications (RFA) issued to invite proposals to develop exposure assessment methods applicable to epidemiologic research on cancer. This initiative is to stimulate interdisciplinary research to develop, test, validate, and compare technology, methods, and biomarkers that could provide accurate estimates of retrospective and current environmental exposures (for example, chemical and physical substances) relevant to cancer development.

The grants, which were awarded in September 2001, are for:

Biomarkers of Carcinogen Exposure and Oxidative Injury
David Christiani, M.D., M.P.H., M.S.
Harvard School of Public Health
Boston, Mass.

Cancer is the second leading cause of death in the United States after heart disease. The great majority of cancers are believed to be preventable because they are due to known environmental factors that influence their incidence. Air pollution, which is predominantly composed of particulates including polycyclic aromatic hydrocarbons (PAH) and metals, is known to contain significant levels of reactive oxygen species (ROS) that can lead to deleterious effects within a normal cell. Continued and repeated exposure to ROS (as seen in occupational exposures) leads to oxidative injury within a cell that can directly affect DNA, cell signaling and growth, and the induction of mitosis, and can initiate the multistep cascade of carcinogenesis.

Previous epidemiologic studies have demonstrated increased cancer incidence among workers exposed to particulates that contain PAHs and metals. David Christiani, M.D., M..P.H., M.S., and colleagues are using a repeated measures, short-term prospective approach to study exposure to such particulates and their relationship to oxidative injury biomarkers as intermediate factors in development of cancer in a cohort of occupationally exposed individuals.

The research addresses important gaps in our knowledge of epidemiologic exposure assessment and provides the opportunity to characterize both biologic markers of exposure (1-hydroxypyrene, mononuclear PAH-DNA adducts, urinary metals and urinary 7,8-dihydro-2'- deoxyguanosine;8-OH-dG) and biomarkers of early environmental carcinogen damage (PAH-DNA adducts, urinary 8-OH-dG).

Specifically, the research will characterize and quantify exposures and exposed populations through evaluation of several biomarkers that are collected simultaneously in surrogate tissue (blood DNA adducts, urinary metals, urinary 8-OH-dG and urinary 1- hydroxypyrene). A better understanding of these risks will lead to improved strategies for cancer prevention.

Historic Traffic Exposure Maps for Cancer Studies
Paul English, Ph.D., M.P.H.
Environmental Health Investigations Branch
California Department of Health Services, Oakland, Calif.

Exposure to traffic exhaust is common in urban areas, and its components have been found to be associated with lung cancer and leukemia in epidemiological studies. Determining whether these associations are causal in nature has been especially limited by the lack of historical individual-level exposure data.

Recently, there have been calls for the development of exposure maps showing modeled concentrations of ambient air pollution using geographic information systems (GIS) to investigate long-term health effects. A study subject's residence can be a good predictor of exposure if suitably modeled levels of air pollution are achieved at a proper spatial scale. Modeling of air pollution typically has been done either at a regional (air basin) or localized level. The regional approach results in coarse exposure maps that lack the spatial resolution needed for epidemiological studies and can result in exposure misclassification. The localized models achieve a proper spatial scale but require complex data inputs and are too data intensive for mapping exposure over wide areas.

Paul English, Ph.D., M.P.H., and colleagues previously developed GIS traffic exposure maps at an appropriate scale for epidemiological studies that account for properties of wind and dispersal behavior of specific pollutants. In this research project, they are further developing this model by accounting for the surface texture of the landscape using GIS land use layers.

They also are evaluating and validating this model and several other models of external exhaust exposure in one California County with nitrogen dioxide (NO₂) field measurements using passive diffusion tubes. The predicted exposure level from each model to actual NO₂ concentrations at each monitored location will be compared. The researchers will assess the amount of bias due to exposure misclassification for each model by geocoding the addresses of lung cancer cases from the California Cancer Registry and a random control series, and comparing the predicted and observed NO₂ measurements. They will evaluate and validate this model and several other models of traffic exhaust exposure.

A series of historic traffic exhaust exposure maps will be developed using the best-evaluated traffic model and retrospective data on traffic counts, land use, meteorology, point sources, and ambient air monitoring data. The resulting exposure maps then may be applied to existing cohorts of study subjects to assign previous exposures and dramatically reduce the time and cost of expensive prospective studies of traffic exhaust exposure and cancer risk.

Quantifying the Impact of Diet on Carcinogen Exposure
James Felton, Ph.D.
University of California Lawrence Livermore National Laboratory
Livermore, Calif.

Epidemiologic and animal studies have shown that diet plays a role in the etiology of human cancer. Cooked muscle meats contain potent mutagens and rodent carcinogens belonging to the heterocyclic amine (HA) class of compounds. Based on analyses of cooked meats, humans are exposed to HAs at levels that may exceed a hundred parts-per-billion. Although it is possible to reliably measure the amount of HAs present in cooked meats, quantification of the biologically effective dose of these compounds is currently limited by imperfect understanding of the effects of individual physiology and the interactions of these carcinogens with other foods in the diet.

James Felton, Ph.D., and colleagues are developing an exposure assessment method that will predict the effect of digestion parameters, intestinal transport and diet upon the low-dose exposure to commonly ingested food carcinogens. The aims are to:

• develop an in vitro digestion procedure to estimate the bioavailability of the four most commonly occurring HAs;
• determine the absorption of the carcinogens from the digestates using the human colon carcinoma cell line Caco2, a commonly used model for intestinal absorption;
• determine the effects of single and multiple food interactions on carcinogen bioavailability in a controlled environment; and
• determine the biologically effective dose of these compounds in humans and validate the in vitro method.

Cooked meats will be digested in the laboratory, and HAs that are made available from the meat matrix will be measured using the HPLC method for measuring Has in food. The laboratory digestates will then be coupled to Caco-2 intestinal cells to measure transport across the intestinal cell wall. Using this method, the effect of other foods on HA bioavailability will be determined by digesting the meat concurrently with vegetables, fruits, liquids, carbohydrates, and fats.

The results of the laboratory study will be compared to two human studies to validate the method. There presently is no accurate way to determine bioavailability of HAs. The in vitro method proposed in this study could provide valuable insight into the dietary interactions that modulate the bioavailability of these compounds.

These experiments also will provide a basis for more accurate exposure measurements for epidemiological investigations of diet and cancer. Better estimates of the bioavailable dose of these food carcinogens and an understanding of how other foods affect bioavailability will make it possible to devise strategies that could potentially reduce exposure levels. Because the results will be based upon interactions of common foods, they can be easily conveyed to the public. Once the impact of dietary factors is understood, then it will be possible to emphasize prevention strategies that will eliminate or modify risk for cancer.

Technology for DNA Damage Analysis and Biomarkers
X.C. Le, Ph.D.
University of Alberta
Edmonton, Canada

Characterization of environmental determinants of cancer is an important goal for development of effective public health strategies for cancer prevention and control. Informative epidemiological investigations of relationships between the environment and cancer rely on accurate, quantifiable exposure measurements. In traditional population surveys that attempt to link a substance with cancer occurrence, exposure data often is obtained indirectly by questionnaires or personal interviews. The uncertainty associated with errors in exposure assessment has obscured or limited identification of cancer etiologies. Rather than the current reliance on external dose estimation and extrapolation from high-dose exposure with rodent models, sufficiently sensitive biomarkers based on DNA damage measurements could permit more realistic assessment of environmental risk.

X.C. Le, Ph.D., and colleagues are working to develop innovative, cost-effective, and reliable technology for quantifying exposures, especially low-level exposures. They aim to develop highly sensitive technology for measuring minute amounts of DNA damage, and improve biomarkers for epidemiological investigations of environment and cancer relationships. The technology combines the specificity of antibodies and capillary electrophoresis separation with the high sensitivity of laser-induced fluorescence detection.

In addition, specific DNA-recognition proteins will be used to bind with various DNA lesions to detect a range of DNA lesions. Automated and high throughput assay formats will be explored that make use of the high-sensitivity technology. The sensitive DNA damage measurements will be used as potential biomarkers for exposure assessment. The technology also will be useful to monitor DNA damage induced by anticancer agents and to study DNA repair, which is an essential protection mechanism against cancer initiation.

Exposure to Beef Growth Promoter in Breast Cancer
Young C. Lin, D.V.M., Ph.D.
Ohio State University
College of Veterinary Medicine
Laboratory of Reproductive and Molecular Endocrinology
Columbus, Ohio

The laboratory-based study is to develop an assay system to detect and measure low levels of Zeranol, a nonsteroidal agent with estrogenic activity, used as a growth promoter in U.S. beef and veal industries. This project is an initial effort to develop an accurate method to quantitate Zeranol in human specimens (e.g., breast tissue, serum, and urine) and in beef products.

Although there has been no evidence or reports of breast cancer linked with meat additives such as Zeranol, laboratory experiments previously conducted by Young C. Lin, D.V.M., Ph.D., and colleagues have noted estrogenic activity of the agent in breast cell and tissue cultures. They hypothesize that the proposed laboratory assay could potentially quantitate how much estrogenic activity, currently unknown, is correlated with measured levels of Zeranol in beef products available in supermarkets.

The researchers further anticipate that the assay will be able to contribute significantly toward answering the question about suspected environmental causes of breast cancer. For instance, since breast cancer is a hormone-related malignancy and appears to be associated with estrogen, knowing the amount of exposure to estrogenic substances such as Zeranol will allow future investigators to determine whether there is any association betweeen the agent and risk for breast cancer.

The aims are to:

• investigate relationships among beef consumption, levels of Zeranol in breast, serum, and urine, and breast cancer in patients of The Ohio State University Hospital;
• measure Zeranol levels in commercially available beef products purchased in supermarkets from 5 different regions across the United States;
• utilize established in vitro model systems to confirm the presence of biological activity (mitogenic activity; altered expression of mechanism-based molecular biomarkers) in specimens from human subjects and commercial beef products in which Zeranol levels were detected; and
• after the assay system has been validated in human specimens, possible relationships will be investigated among beef consumption, levels of Zeranol in breast, serum, and urine, and breast cancer in patients of The Ohio State University Hospital.

Passive PAH and Micro-camera Personal Monitoring System
Paul Lioy, Ph.D.
University of Medicine and Dentistry of New Jersey-Rutgers University
Environmental and Occupational Health Sciences Institute
Piscataway, N.J.

Paul Lioy, Ph.D, and colleagues are developing a new, sensitive diffusion (or passive) sampler and coupling its use to a miniaturized personal portable video camera system (MPPVCS) for measurement and assessment of personal exposure to carcinogenic organics in personal air. The research project will:

• determine the sampling rate under different sampling conditions (different air concentration, sampling duration, temperature, humidity, and face velocity);
• evaluate the performance of the sampler in side-by-side comparisons with proven active sampling method;
• evaluate the ability of subjects to successfully wear and manage battery and tape changes; and
• evaluate the combined effectiveness of the sampler and MPPVCS to capture information about volunteer personal exposures to carcinogenic organics.

The new passive sampler consists of a denuder, which is made from 10 0.2-cm (or 2-cm) sections of a multicapillary gas chromatography (GC) column (900 of 40-micrometer internal diameter). Due to the high value of diffusive area (inner cylindrical surface) to diffusive path length (the radius of the each column), the sampling rate will be about 1000 times higher than the traditional badge or tube-type diffusive samplers. The denuder-type geometry will allow constant sampling rates in a wide range of sampling conditions.

Also, the stationary phase of the capillary column allows direct thermal desorption of analytes to an analytical instrument, which avoids dilution of the sample by solvent and increases the sensitivity by a factor of 100. The sampler can be used repeatedly without any regeneration process. The sampler is light (<1 g), small (11.5 mm in diameter and 2 mm or 2 cm in length), and easy for a person to carry.

During the exposure session, a lightweight, unobtrusive MPPVCS will be attached to each study participant in order to obtain real time information about proximity to sources of carcinogenic organics and the individual's activities at a given period of time. The MPPVCS weighs less than one pound and can be comfortably worn. Use of MPPVCSs can reduce the costs and burden associated with having a videographer track the individual, reduce the burden associated with having study participants complete questionnaires and time/activity diaries, and avoid information loss from failure to complete questionnaires or the study participants being unaware of activities and the environment.

In summary, the researchers are developing and validating a sensitive, simple, and cost-effective technique to directly measure personal exposure to carcinogenic organics. Joint use of the new diffusive sampler and MPPVCS will capture both the levels of exposure and the locations of exposure, and possible sources. This combined approach will improve the accuracy of community or residential assessments of personal exposure to carcinogens for prospective epidemiological studies, and provide improved characterization of exposure for risk assessments.

Retrospective Assessment of Mixed Chemical Exposures
Gurumurthy Ramachandran, Ph.D.
University of Minnesota
Division of Environmental and Occupational Health
Minneapolis, Minn.

Misclassification of exposure is a persistent source of bias in epidemiologic studies of occupational cancer. This occurs because:

• exposure reconstructions are typically based on sparse data with significant uncertainty;
• biologically relevant doses are not estimated (for example, for exposures to inhaled dusts, the cumulative lung burdens of the worker do not account for retention and clearance of inhaled dusts); and
• exposures to multiple chemicals are not taken into account.

Gurumurthy Ramachandran, Ph.D., and colleagues are developing an improved exposure and dose assessment method for epidemiologic research on occupational cancer that accounts for the uncertainties in exposure reconstruction due to sparse data, determinants of etiologically relevant dose, and exposures to multiple chemicals, using a Bayesian probabilistic framework.

The methodology will be developed and demonstrated using a large occupational exposure dataset (1950-2000), from Falconbridge Ltd., Sudbury, Ontario, which is one of the world's leading primary nickel production companies. Workers at the company historically have been exposed to several nickel species (oxidic, sulfitic, and soluble), diesel particulate matter, and silica -- all of which are proven or suspected human lung carcinogens.

A novel Bayesian methodology that synthesizes expert judgment and historical measurements is being developed for exposure reconstruction. The exposure reconstruction will incorporate retention and clearance models for estimating the cumulative lung dose of oxidic, sulfitic, and soluble nickel species, diesel particulate matter, and airborne silica. The exposure reconstruction then will be available for a planned epidemiologic case-control study of lung cancer among the Falconbridge employees. The methods also will be applicable to other industry-based epidemiologic studies where available data are sparse and exposures to chemical mixtures are the norm.

Carcinogen Metabolism Enzyme Expression in Buccal Cells
Simon Spivack, M.D.
New York State Department of Health
Wadsworth Center
Albany, N.Y.

Simon Spivack, M.D., M.P.H., and colleagues are evaluating the validity of testing for gene expression in buccal cells as a way to accurately reflect gene expression in human lung tissue when both are exposed to tobacco smoke, as occurs in smokers. Previous studies of coding sequence polymorphisms in these genes have presented very mixed results as to whether they confer risk for lung cancer. The parallel induction of genes encoding carcinogen-metabolizing enzymes simultaneously in both the lung and an easily accessible surrogate tissue (buccal cells), suggests the possibility of being able to expression phenotype smokers for bioactivating gene inducibility.

If this carcinogen metabolism pathway is relevant to lung cancer susceptibility, then a high phase I- and low phase II-enzyme inducibility phenotype might be hypothesized to confer risk for lung cancer. The power of studies on carcinogen-metabolizing gene induction in lung cancer to date has been limited by the number of study participants enrolled and short supply of suitable, quality human lung tissue for expression or activity assays.

This study takes advantage of newly developed real-time methods for quantifying gene expression in appropriate, smoke-exposed epithelium, and compares that expression to that in precisely-defined microdissected smoke-exposed human lung epithelia.

The aims are to:

• develop quantitative real-time RT-PCR assays for the evaluation of gene expression of a panel of carcinogen metabolizing enzymes in buccal mucosa, including Ahr, ER, CYP1B1, CYP1A1, GSTM1, GSTM3, GSTP1, GSTT1, NQO1;
• compare the gene expression in buccal mucosa to simultaneously sampled plasma assayed for nicotine and cotinine by state-of-the-art LC-MS-MS techniques;
• quantitatively compare the expression of carcinogen metabolizing enzyme genes from in situ tobacco-exposed human buccal mucosal cells with that from microdissected human lung bronchial and alveolar epithelial cells also exposed in situ; and
• measure the variability of this buccal cell carcinogen-metabolizing gene expression across individuals, accounting for expression confounders such as tobacco smoke exposure and dietary constituents, and thereby identify phase I and phase II high or low expressers.

The long-term goal is to develop a panel of biomarkers in an easily accessible, smoke-exposed epithelium that identifies smokers at high-risk for lung cancer, and thereby allows smoking-cessation, chemoprevention, and early-disease detection strategies to be focused on those individuals who are most susceptible.

Validation of Fumonisin and Microcystin Biomarkers
Jia-Sheng Wang, M.D., Ph.D.,
Texas Tech University System
Lubbock, Texas

The long-term goal of this research project is to study relationships between exposure to fumonisins and microcystins and human liver and esophageal cancer risks in high-risk populations. Fumonisins and microcystins are recently identified environmental biotoxins that are produced by toxicogenic fungi and cyanobacteria, respectively. Fumonisins are carcinogens, and both fumonisins and microcystins are strong tumor promoters in animal models.

Human populations in certain areas of world are exposed to higher levels of these toxins in their daily life mainly through contaminated dietary components and/or drinking water. Both toxins have been etiologically linked to high incidence of human primary liver and esophageal cancer in several areas of South Africa and China. Although analytical methods for detecting these toxins in environmental samples have been reported and potential biomarkers, such as disruption of sphingolipids metabolism by fumonisins and inhibition of protein phosphatases by microcystins, have been found in animal models, validation of these biomarkers in humans, especially in high-risk populations, has not been done or reported.

To date, methods are lacking for simultaneous detection of these toxins in environmental samples and in human body fluids. This is critical, however, for assessment of human cancer risks because coexposure to these biotoxins has been widely reported in high-risk populations.

Jia-Sheng Wang, M.D., Ph.D., and colleagues are developing and validating analytical array methods to simultaneously measure these biotoxins in both environmental samples and human body fluids. The researchers are using molecular epidemiological tools to validate methods for biomarkers in body fluids of cases of primary liver cancer and esophageal cancer and controls. The working hypothesis underlying this research is that long-term exposure to fumonisins and microcystins may induce synergistic carcinogenic effects in high-risk individuals, and application of validated biomarkers can improve the quantitative estimation of human cancer risk from exposure to these biotoxins.

The aims are to:

• develop rapid and sensitive analytical array method(s) for measuring fumonisin and microcystin biomarkers based on a previously developed immunoaffinity-HPLC method for aflatoxins and/or HPLC-enzyme-linked immunosorbent assay for microcystins;
• validate method(s) for measuring biomarkers of microcystins and fumonisins in food, water, and body fluids such as blood and urine samples collected from two one-week longitudinal biomonitoring studies in study participants from high-risk areas for primary liver and esophageal cancer; and
• perform molecular epidemiologic studies in two high-risk populations of primary liver cancer and esophageal cancer in Qidong and Huaian, China, to explore cancer risks from exposures to fumonisins and microcystins.

Validation of Indirect Pesticide Exposure Estimates
Timothy Wilcosky, Ph.D.
Research Triangle Institute
Research Triangle Park, N.C.

Many years typically elapse between etiologic carcinogenic exposures and the subsequent occurrence of cancer, and therefore, epidemiologists often must use indirect estimates of historical exposures for study participants. Earlier research suggests that historical data on crops and associated pesticide applications can be used to estimate the probability of agricultural pesticide exposures among persons living near fields, but this retrospective work could not ascertain whether nearby residents actually were exposed.

However, a prospective study that collects environmental and biological samples from households at different distances from agricultural fields would reveal whether a clear exposure gradient occurs as a function of distance from the fields, and indicate whether there is a correspondence between measured exposures and a records-based method for predicting household member pesticide exposures. If a predictable exposure gradient occurs with increasing distance from the fields, then retrospective exposure estimates could be further refined to account for this gradient.

Timothy Wilcosky, Ph.D., and colleagues are investigating the herbicide atrazine as a marker of pesticide exposure in persons living near corn fields in Illinois. These areas have pockets of rural housing located in proximity to cornfields, and towns and communities that lie in juxtaposition to agricultural lands.

To identify specific areas and specific houses, the researchers are using a variety of data including the U.S. Census of Agriculture and the Illinois Agricultural Statistics Service, aerial photography and/or satellite imagery, GIS data sets, localized census data (block, block group, and census tract levels), and topographic maps. From each of the 50 households enrolled, the researchers will include one adult and one child younger than 4 years-of-age from whom they will obtain a urine sample. They also will obtain a dust sample and drinking water sample from each home.

A questionnaire will capture information about behaviors and housing characteristics that might influence the participants' pesticide dose. This research will help reveal whether records-based techniques will be useful for reconstructing pesticide exposures, and whether use of additional exposure-related variables increases the precision of the exposure estimates.

Validation of PAH Biomarkers for Quantifying Cancer Risk
Junfeng Zhang, Ph.D.
University of Medicine and Dentistry of New Jersey-Rutgers University
Piscataway, N.J.

Biological monitoring is particularly useful for assessing exposure to pollutants, such as polycyclic aromatic hydrocarbons (PAHs), that are present in multiple environmental media. Although a number of PAH biomarkers have been developed, epidemiological studies relating cancer mortality to average biomarker concentrations are lacking, and the current risk assessment frame is based on air concentrations of benzo[a]pyrene (BaP).

Junfeng Zhang, Ph.D., and colleagues' research will fill several important gaps that have hampered the application of PAH biomarkers in quantifying cancer risk. Two urinary markers, 1-hydroxypyrene and 9-hydroxy-benzo[a]pyrene, will be validated.

The researchers will:

• Establish a quantitative relationship between urinary concentrations and BaP exposure, which will be measured as 24-hour average personal air concentration and 24-hour dietary intake.
• Examine whether there is a significant day-to-day variation in urinary marker concentrations in study participants who have stable daily PAH exposures. This is important because for a biomarker to be useful for estimating cancer risk, the marker should be able to predict people's average (steady) exposure.
• Examine whether a first morning urine sample can efficiently represent daily exposure or a 24-hour composite urine sample is required. This is of practical importance because 24-hour urine samples are considerably harder to get in free-living populations.
• Examine relationships between urinary markers and other PAH metrics, because other metrics of PAH mixture may prove to be a better index for future risk assessment.

The study is being conducted with 100 nonsmoking adults who have a wide range of exposure to airborne BaP, including both occupationally and non-occupationally exposed individuals. Inclusion of non-occupationally exposed individuals in evaluating the biomarkers is important because high PAH exposures also can occur in the course of their daily lives, including non-smokers.

Two sets of repeated measurements are being taken for each study participant. Each set includes 24-hour personal breathing-zone concentrations of gas-phase and particle-phase PAHs; 24-hour dietary PAH intakes; and biomarker concentrations in first-morning urine and in 24-hour-composite urine. All measurements will be made within the same 24-hour period. Two repeated measurements are separated by at least three months to maximize the representativeness of daily exposure.

Air samples and food samples are being analyzed for at least 16 common PAHs, including pyrene and BaP, using the established HPLC/fluorescence methods. The spot nature of urinary concentrations will be corrected with creatinine. Statistical methods to be used include descriptive analyses, scatter plots, correlation analyses, paired comparisons, and linear mixed models. Adjustments will be made for age, sex, height, weight, and body mass index in the mixed models.