Bhat VS. 2014. Benchmark dose-response modeling approaches to integrate toxicogenomics into quantitative risk assessment. Dissertation, University of California, Riverside, June 2014.
The 21st century vision for toxicity testing and risk assessment is to become more efficient at evaluating a broader range of chemicals with minimal animal testing. The objective of this research was to determine if short-term global gene expression profiling can be used to inform potency and chemical mode of action (MOA) determinations and thus collectively serve as a strategy to prioritize future or data-poor chemicals for further evaluation. Dose-response modeling was used to estimate low-dose potency for activation or repression of key genes and pathways, and these transcriptional benchmark dose (BMDT) estimates were compared to benchmark dose estimates for more traditional toxicological (i.e., apical) responses (BMDA) at the cellular or tissue level which were observed after the same exposure duration. These BMDT and BMDA were also compared to BMDA for the ultimate adverse outcome that might traditionally serve as the basis of quantitative risk assessment. The case studies included (1) liver gene expression at 30 days for conazole antifungal agents, some of which are hepatocarcinogens in mice, (2) liver gene expression at 2 and 7 days for phenobarbital, a prototype nongenotoxic rodent hepatocarcinogen that activates the constitutive androstane receptor (CAR), and (3) urinary bladder gene expression at 7 days and 20 weeks for diuron, a substituted urea pesticide associated with urinary bladder cytotoxicity and tumorigenesis in rats. BMDT and BMDA estimates are particularly useful in informing threshold dose-response and risk assessment approaches for nongenotoxic carcinogens. Our studies showed that BMDT estimates for key genes and pathways were highly concordant and thus phenotypically anchored to BMDA for target tissue responses observed after the same exposure duration. Our short-term BMDT were also consistent with potency estimates for the tumorigenic outcome or precursor key events such as hyperplasia. In conclusion, this research demonstrated the utility of short-term global gene expression profiling for informing potency with respect to MOA determinations and quantitative risk assessment of long-term exposures. Further, these case studies encompassed multiple rodent species, target tissues, MOA, and exposure durations ranging from 2 days to 20 weeks, thus demonstrating the potential utility of this approach as a chemical testing and prioritization strategy.