Environmental Factor, April 2011, National Institute of Environmental Health Sciences
Thomas explores advances in risk assessment using genomics
NTP Molecular Toxicologist Scott Auerbach, Ph.D., hosted Thomas' presentation as part of the BSB seminar series and fielded the questions that followed. (Photo courtesy of Steve McCaw)
In the course of his talk, Thomas maintained, "The degree of correlation between the transcriptional BMD values and those for the traditional apical endpoints suggests that they may be used as potential surrogates for both cancer and non-cancer points of departure." (Photo courtesy of Steve McCaw)
Among the many attendees who asked questions in the course of the extended question and answer session was SRA contractor Ruchir Shah, Ph.D., a specialist in genomics and bioinformatics. (Photo courtesy of Steve McCaw)
Thomas left the roomful of scientists with an appreciation for the potential of bringing about a change in the established methodology of risk assessment and stimulated an animated discussion of questions that remain to be answered. (Photo courtesy of Steve McCaw)
Russell Thomas, Ph.D., director of the Center for Genomic Biology and Bioinformatics and a senior investigator at The Hamner Institutes for Health Sciences, presented a talk at NIEHS March 21 on the application of non-traditional approaches to risk assessment, as part of the NTP Biomolecular Screening Branch (BSB) seminar series.
In his talk, titled "Application of Transcriptional Benchmark Dose Values in Quantitative Cancer and Noncancer Risk Assessment," Thomas(http://www.thehamner.org/scientists/entry/russell-s-thomas) explored the current status of risk assessment and the potential for incorporating genomic technology into an emerging paradigm of predictive toxicology. He began his survey with a review of several of the issues being addressed by the Tox21(https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2721892) consortium made up of the NTP, NIH Chemical Genomics Center (NCGC), U.S. Environmental Protection Agency (EPA), and its most recent partner, the U.S. Food and Drug Administration (FDA).
Current status of risk assessment
One of the current challenges in risk assessment is that "chemicals without published reference values are not considered quantitatively in the overall hazard index calculation when evaluating contaminated sites where multiple chemicals exist," explained Thomas. He pointed to some limitations of the existing risk assessment paradigm, such as the cost, time, and number of animals used in the traditional toxicity tests that support a risk assessment. As a result, a vast number of chemicals are largely ignored when risk assessment is performed at Superfund sites, and potentially harmful ones can be interpreted as posing no risk to human health, he continued.
Incorporation of genomic technology in risk assessment
In explaining why genomics may offer a useful approach for filling the data void, Thomas said that the use of genomic technology in toxicology began more than a decade ago and since then has transitioned from being a specialized laboratory technique to an off-the-shelf commodity. Current genomic technology now includes the capability to survey the entire transcriptome.
As Thomas explained, "The technology has become more reliable and reproducible." Hence, he sees the primary advantages of applying genomic technology to risk assessment as "being able to sensitively and comprehensively examine the molecular changes resulting from chemical exposure." An important application of this technology lies in providing both quantitative and qualitative information on the dose at which cellular processes are transcriptionally affected in what Thomas referred to as "a transcriptional point of departure for chemical risk assessment" (see text box).
A lively discussion
In addressing such questions as whether we know which transcriptional markers are correlated with specific pathological endpoints and how tissue-specific non-cancer endpoints can be assessed, Thomas responded that even though different chemicals may act by different mechanisms to elicit these pathological endpoints, common processes or pathways that are transcriptionally altered, which are what he called "bioindicators of effect," can still be identified, even though they may be independent of the mechanism. He went on to describe possible ways this technology may be applied to chemicals with unknown effects or target organs.
(Mamta Behl, Ph.D., is a research fellow in the NTP Toxicology Branch)
Transitioning from traditional end points to newer metrics of biological perturbation
Thomas explained that a risk assessment for non-cancer and cancer endpoints is usually conducted using dose response modeling approaches where a dose is identified that causes a defined increase in an adverse effect.
"The dose response modeling approaches are referred to as benchmark dose analysis, and a benchmark dose (BMD) can be thought of as a transition point between a no observed adverse effect level (NOAEL) and a lowest observed adverse effect level (LOAEL) within a dose-response curve," said Thomas. Referring to his recent study(https://www.ncbi.nlm.nih.gov/pubmed/21097997) on the topic, he explained how BMD values are calculated for cancer and non-cancer endpoints and provided examples of the way BMD methods have been applied to previous NTP studies for both cancer and non-cancer endpoints.
Thomas described a method his group has developed that allows for transcriptomic data to be used for calculating genomic-based BMD values for individual genes and sets of genes with shared biological roles (referred to as gene ontologies). He showed how the most sensitive BMD values derived from liver and lung transcriptomic measurements following a subchronic exposure were slightly more sensitive than BMD values for traditional cancer and non-cancer apical endpoints.
Thomas explained that this comparison demonstrates, counter to common belief, that transcriptome changes at 90 days are not orders of magnitude more sensitive than apical toxicity endpoints. He went on to show that the genomic BMDs for gene ontologies related to stress signaling and inflammatory processes correlate strongly with BMDs for cancer and non-cancer apical endpoints, respectively. Finally, he outlined an approach by which genomic-based BMDs may be incorporated into chemical risk assessment.
Citation: Thomas RS, Clewell HJ 3rd, Allen BC, Wesselkamper SC, Wang NC, Lambert JC, Hess-Wilson JK, Zhao QJ, Andersen ME.(https://www.ncbi.nlm.nih.gov/pubmed/21097997) 2011. Application of transcriptional benchmark dose values in quantitative cancer and noncancer risk assessment Toxicol Sci 120(1):194-205.