Environmental Systems Biology Group
Modeling of toxicokinetic data aids in the design and interpretation of toxicity and carcinogenicity studies with respect to the relationship between external exposure, toxic effects, target organ dosimetry and species differences in physiology and metabolism. Through continued cooperation with several research labs, new model structures and methods have been developed to characterize concentration, metabolite formation and biochemical interactions in tissues across multiple species. These new model structures provide mechanistic insight on the origin of biochemical changes that result from particular exposures and that contribute to adverse effects. In addition, a number of compounds of interest to the NTP and other groups have been modeled using physiologically-based pharmacokinetic (PBPK) models in order to provide a context through which the NTP can evaluate whether species differences in toxic response are due to physiological and/or metabolic differences.
The analysis and interpretation of gene expression data linked to environmental exposures (toxicogenomic data) in the context of making public health decisions is a major area of research at the NIEHS and at numerous other research institutions. The mathematical analysis of toxicogenomic data derived from microarrays has received considerable study over the last few years. In general, most of these methods are qualitative and intended for identifying significant changes in gene expression, for identifying clusters in changes in gene expression, or for identifying network relationships between the genes. Only a few recently published methods are sufficiently quantitative in nature that they could be used for quantitative evaluation of health risks through direct linkage with other models like the PBPK models described above and the morbidity/mortality models described below.
A major focus of the modeling effort within the ESBG has been on the development of statistically sound, mechanism-based methods for the identification and quantification of gene-expression networks using microarray data. The group’s approach has focused on classical likelihood-based methods and on Bayesian networks to provide insight into the limitations of microarray data for quantitative modeling, and to provide study design guidance to overcome these limitations. In addition, the ESBG is addressing design and analysis issues related to ongoing studies being developed by the NTP to ensure utility of the study results for both hazard identification and dose-response analysis.
Current areas of research are intended to allow for incorporation of other types of molecular data, specifically gene sequence, proteomic and metabolomic data, into modeling to better identify and quantify cellular networks and improve the efficiency of experimental designs for use in health risk assessments.
Since most environmental health risk assessments are focused on the rates of morbidity and mortality in human populations following an environmental exposure, the group also has an active research program in developing mechanism-based models for toxicity endpoints. Such models are needed to provide a sound linkage between laboratory-based mechanistic research and quantitative estimates of human health risks. Because these are generally linked to counting events or failure of an entire organ system to function properly, they require a different mathematical treatment than the mathematical treatment applied to toxicokinetics and toxicogenomics data. Mechanistic models of morbidity and mortality lag far behind those for toxicokinetics, partially due to the difficulties in the mathematical treatment of these endpoints and partially due to the large number of possible mechanisms combined with gaps in understanding how these mechanisms can be numerically encoded. The goal of this project is the development of improved methods and novel models of toxicity that more closely reflect the current biological understanding of disease processes. As before, these approaches are generally tied to NTP studies and research activities and are used to address questions of experimental design as well as the analysis and interpretation of study results.