1 Introduction
Significant progress has been made in adopting New Approach Methodologies (NAMs) for chemical safety assessment. NAMs have been particularly successful for local toxicity endpoints like skin corrosion, eye damage, and skin sensitization (Sewell et al., 2024). To allow safety assessment of chemicals NAMs will also be needed for more complex endpoints. To this end, Next-Generation Risk Assessment (NGRA) approaches are increasingly being developed (Thomas et al., 2019; Berggren et al., 2017). These approaches are exposure-led, and hypothesis driven, using a tiered, iterative approach to make safety decisions, designed to prevent harm (Dent et al., 2018). The initial tier of such approaches is constructed to be protective of human health, often integrating high throughput assays (e.g., high throughput transcriptomics (HTTr) (Farmahin et al., 2017; Harrill et al., 2019) with more targeted tools (e.g., functional or binding assays for specific receptors) allowing broad biological coverage (Middleton et al., 2022; Zobl et al., 2024). Using multiple concentrations, points of departure (PoDs) can be calculated to identify concentrations at which a compound starts to cause biological perturbations (bioactivity) in a test system. These approaches have been evaluated in several case studies by calculating bioactivity:exposure ratios (BERs) from PoDs in combination with predicted systemic adult exposure estimates using physiologically based kinetic (PBK) models. Results from these evaluations demonstrate the protectiveness of these NAM based approaches mostly for systemic safety assessments (Baltazar et al., 2020; Dent et al., 2021; Middleton et al., 2022; Zobl et al., 2024; Cable et al., 2024). If needed an early tier can be followed up with more physiologically relevant cell systems for hazard testing or exposure predictions to refine outcomes (Thomas et al., 2019; Berggren et al., 2017). These new approaches have the potential to fundamentally transform chemical regulatory framework(s) by allowing more human-relevant decision-making to support sound human health safety decisions in diverse industrial sectors (cosmetics, industrial chemicals, pharmaceuticals, occupational health, etc.) (Magurany et al., 2023; Schmeisser et al., 2023).
To perform a comprehensive chemical safety assessment, it is crucial to ensure human exposures will not cause developmental and reproductive toxicity (DART). Due to the complexity and the distinct stages within the reproductive cycle, this was historically addressed using several OECD in vivo test guidelines (Knight et al., 2023) which assess changes in male and female reproductive function, gamete development and maturation, conception and embryo implantation, embryonic and fetal development, birth and weaning, the onset of puberty, attainment of full sexual function, and potential effects on subsequent generations (summarized in (EMA, 2023)). These DART-specific testing guidelines are employed to assess defined apical endpoints related to developmental or reproductive toxicity, such as pregnancy duration, fetal malformations, and the weight and morphology of reproductive organs, etc., but also evaluate non-specific/systemic effects like the body weight of the parental generation and the offspring, as well as the weight and morphological changes of reproductive as well as non-reproductive organs. The integration of DART and systemic testing endpoints serves as an approach protective of critical effect levels for human adverse outcomes (Browne et al., 2024). The first indication that NGRA approaches could also be protective for DART came from a study performed under the international government-to-government initiative “Accelerating the Pace of Chemical Risk Assessment (APCRA)”. By comparing PoDs from high-throughput assays with traditional hazard information for over 400 chemicals, including results from DART testing guidelines, this study demonstrated that for 89% of the compounds, the PoDs from NAMs were more conservative than PoDs derived from animal studies. No enrichment was found for compounds with data from DART studies within the cohort of 48 compounds in which the in vivo PoD was lower (Paul Friedman et al., 2020).
Previously we proposed an NGRA framework for DART (Rajagopal et al., 2022). The biological coverage of the NAMs within the proposed framework was evaluated by comparing cellular processes, signalling pathways and genes involved in known key stages in human reproduction and embryo-fetal development from an automated literature extraction to the read-outs from our NAM toolbox (including basic expression levels of cell lines). We showed ∼80% coverage of these processes based on gene numbers (Rajagopal et al., 2022). Knowledge of the biological coverage of our proposed framework and the previous work from APCRA (Paul Friedman et al., 2020) suggests that an NGRA approach could provide protection for DART, however conclusive evidence is still lacking. Therefore, in this study, we evaluated the protectiveness of our DART NGRA framework by testing 37 benchmark compounds. High and low-risk exposure scenarios for the 37 compounds were identified using DART-relevant data from authoritative sources as benchmarks. Within tier 0 of the framework in silico predictions covering general alerts for DART as well as for specific receptor activity were performed and results were compared to historical data to evaluate the predictive power of these tools. In tier 1 data from our DART NAM toolbox was generated and PoDs were calculated to estimate chemical bioactivity. Bioactivity was then compared to the estimated human exposure to calculate a BER for each exposure scenario (for an overview see Figure 1 and for a more detailed description of the workflow for the evaluation see material and methods).
To cover the different life stages of the reproductive cycle, it is essential to consider the exposure of non-pregnant adults, pregnant women, and fetal populations. This approach needs to take into account the anatomical and physiological changes in the pregnant woman and the gestational changes within the embryo, which may alter the absorption, distribution, metabolism, and excretion (ADME) of a compound, thereby impacting systemic exposure (Kapraun et al., 2019; Hudson et al., 2023). To broadly investigate the impact of population-based changes in pregnancy on internal exposures, clinical data from pregnant and fetal exposure as well as non-pregnant adult exposure was extracted from literature for the 37 compounds where available to inform on the exposure and exposure distribution between the three populations. PBK modelling was used to predict the internal exposure of the benchmark chemicals in female adults where no in vivo data could be found. BERs were calculated for all three subpopulations for each exposure scenario.
For the final evaluation of the NGRA approach BERs were used to group exposure scenarios into uncertain (BER <1) or low risk (BER >1). Conceptually a BER of 1 indicates that bioactivity would not be observed at human-relevant exposures. However, an experimentally derived BER threshold that would be considered protective for DART has not yet been proposed or agreed. Therefore, the purpose of this study was to assess whether a BER of 1 would be a protective of DART in humans and useful for decision making.
Hoặc