You are here:
Post-hoc Development of a Quantitative Adverse Outcome Pathway Prediction Model for Lung Cancer
Citation:
Hill, T. AND R. Conolly. Post-hoc Development of a Quantitative Adverse Outcome Pathway Prediction Model for Lung Cancer. NC Chapter SOT, Research Triangle Park, North Carolina, October 15, 2018.
Impact/Purpose:
The US EPA has expressed great interest in the adverse outcome pathway (AOP) concept as a chemical-agnostic tool to prospectively evaluate chemical safety in an efficient, rapid and effective manner that protects the public health. A central premise of the AOP concept is that pathway progression from the molecular initiating event (MIE) implies a definable “response-response” (R-R) relationship exists between each key event (KE) preceding the adverse outcome. Computational description of these R-R relationships in a quantitative AOP (qAOP) enables dose-response consideration of probabilities and uncertainty, cross taxa scaling, and flagging of special at-risk populations or sentinel species. These mathematical descriptions are integrated into the overarching qAOP construct and can be coupled to environmental fate and exposure data needed to inform level of risk. The qAOP also provides a platform to utilize New Approach Methodologies (NAMs; e.g. early genomic or in vitro data streams) for rapid, less resource-intensive hazard prediction. However, most qAOP development to date has relied upon directed, prospective experimentation for model construction. A main goal of this work is construction of the AOP and qAOP entirely from findings in previously published, peer reviewed literature, using publicly-available freeware modeling programs. This case study will encourage the increased use of qAOP in the regulatory community and supports the US EPA’s efforts on use of quantitative effect data for adversity predictions and incorporation of NAMs into the cancer risk assessment process.
Description:
Traditional methods for carcinogenicity testing are resource-intensive and time consuming. The continuous backlog of unexamined compounds has generated interest in the adverse outcome pathway (AOP) concept as a chemical-agnostic tool to prospectively evaluate chemical safety in an efficient, rapid and effective manner that protects the public health. A central premise of the AOP concept is that pathway progression from the molecular initiating event (MIE) implies a definable “response-response” (R-R) relationship exists between each key event (KE) preceding the adverse outcome. Computational description of these R-R relationships in a quantitative AOP (qAOP) enables dose-response consideration of probabilities and uncertainty, cross taxa scaling, and flagging of special at-risk populations or sentinel species. These mathematical descriptions are integrated into the overarching qAOP construct and can be coupled to environmental fate and exposure data needed to inform level of risk. The qAOP also provides a platform to utilize New Approach Methodologies (NAMs; e.g. early genomic or in vitro data streams) for rapid, less resource-intensive hazard prediction. However, most qAOP development to date has relied upon directed, prospective experimentation for model construction. This poster describes both a novel AOP and qAOP for a type of murine lung cancer that is considered specific to the mouse, i.e., to not also occur in humans. A main goal of this work is construction of the AOP and qAOP entirely from findings in previously published, peer reviewed literature, , using publicly-available freeware modeling programs . The AOP (Fig. 1) is independent of route of exposure and grounded in heavily published, overlapping mechanistic events for naphthalene, styrene, ethyl benzene, isoniazid and fluensulfone in mice. The collated data describing KER and progression to the adverse outcome were programmed into the modeling software (GNU Octave; www.gnu.org) as the mathematical descriptors for the R-R relationships in the qAOP. The AOP/qAOP (Fig. 2) is structured in accordance with the 2-stage clonal growth model for cancer. It begins with an MIE of CYP2F2-specific formation of reactive metabolites and KE for protein/nucleic acid adducts, inducing rapid proliferation of CC10 deficient Club cells by direct DNA damage (genotoxic) or cell death/repair (cytotoxic) mechanisms. Either route takes advantage of errors in cell division and effectively increases the incidence of an initiated cellular genotype. Continued stimulation of the pathway then activates progression to a mature cancer cell, which culminates in the adverse outcome of club cell positive adenocarcinoma tumors in the airway. Data indicates there are about 3.36 x106 club cells in the mouse lung, making 80% of the total lung tissue a target for this process. Quantification of mechanistic relationships and exposure (PBPK, dose-response and time-course) was supported using archival data in published literature that was collated in accordance with US EPA guidelines for BMD analysis. The modeling supports a distinction between a purely genotoxic (A) or a cytotoxic (B) driven mutation process and suggests that the “unique” susceptibility to lung tumor development through CYP2F in the mouse has a quantitative, rather than a qualitative, basis. Our hypothesis is that the observed mouse-human difference is due to the localized magnitude of disturbance in an R-R relationship, rather than absolute genomic (genetic or phenotypic expression) differences between humans and mice. Additional study findings include evaluation of data depth and breadth supporting the cancer qAOP, suitability for characterization of R-R relationships, and identification of data gaps or additional research required. This approach supports the US EPA’s efforts on use of quantitative effect data for adversity predictions