CHALLENGES TO BUILDING AND APPLYING AN AOP/ MOA ONTOLOGY
The challenges to building an AOP/MOA developmental toxicity ontology (DTO) include (1) the role of potency (and separating adaptive from adverse response), (2) the importance of maternal toxicity as a driver/confounder of in vivo responses, and (3) the importance of developmental stage susceptibility. Identifying a parsimonious testing strategy for identifying a specific AOP is a challenge to applying an AOP/MOA ontology. Additionally, most of the toxicology literature is descriptive and evaluates effects at the organ and organism level and generally does not contain information on mechanism of action, a least not at the granularity that is needed to support a relatively complete ontology.
Translation of an AOP/MOA ontology into a testing strategy containing assays covering the KEs (qualitatively and quantitatively) is necessary for an efficient assessment of the possible developmental toxicity potential of chemicals. Translation of the response magnitude in each KE-representing assay, in terms of adaptation versus adversity, is also required. In other words, thresholds of adversity need to be defined, either for individual assays or for combinations thereof. Moreover, the outcome of a developmental toxicity IATA should be accompanied by an uncertainty analysis, for which tools and approaches need to be defined and put in place.
AOPs describe physiological/toxicological routes as the elements of the ontology. Thus, AOPs can be seen as the bricks needed for building the ontology house, leading to an IATA. Components of IATA outside the domain of the ontology are chemical-related approaches such as structure/activity, physicochemical properties, in silico modelling, grouping, read-across, ADME, etc. This approach is fit for purpose if applied case-by-case in an iterative process of initially limited testing and deciding on the next testing step based on interim findings. This process ends if sufficient information has surfaced to underpin hazard and risk assessment for the tested compound.
The acceptability of a DTO-driven IATA for mechanistically based developmental toxicity hazard and risk assessment is heavily dependent on whether the DTO is comprehensive. Comprehensiveness is not necessarily determined by the level of detail of the description of the biology involved, but rather by the extent to which the DTO leads to an IATA that is sufficient to detect developmental toxicants with sufficient sensitivity and specificity, as agreed by risk assessors and risk managers. This would be facilitated by complete and open sharing of all in vivo toxicology data to ensure the comprehensiveness of the DTO.