Accumulation-depuration data collection in support of toxicokinetic modelling

www.nature.com/scientificdata Accumulation-depuration Data Descriptor data collection in support of toxicokinetic modelling OPEN Aude Ratier & Sandrine Charles ✉ Regulatory bodies require bioaccumulation evaluation of chemicals within organisms to better assess toxic risks. Toxicokinetic (TK) data are particularly useful in relating the chemical exposure to the accumulation and depuration processes happening within organisms. TK models are used to predict internal concentrations when experimental data are lacking or difficult to access, such as within target tissues. The bioaccumulative property of chemicals is quantified by metrics calculated from TK model parameters after fitting to data collected via bioaccumulation tests. In bioaccumulation tests, internal concentrations of chemicals are measured within organisms at regular time points during accumulation and depuration phases. The time course is captured by TK model parameters thus providing bioaccumulation metrics. But raw TK data remain difficult to access, most often provided within papers as plots. To increase availability of TK data, we developed an innovative database from data extracted in the scientific literature to support TK modelling. Freely available, our database can dynamically evolve thanks to any researcher interested in sharing data to be findable, accessible, interoperable and reusable. Background & Summary The Environmental Risk Assessment (ERA) workflow for chemical substances of interest, as described in some European regulations (e.g., for plant protection products in marketing authorisation applications (EU regulation No 283/2013)), requires a bioaccumulation test, for example on fish according to the OECD Test guideline 3051. Such a test consists in an accumulation phase followed by a depuration one and the time course of the internal concentration within fish is measured at regular time points during both phases. The resulting data allow us to model the time-course of the exposure within organisms (denoted toxicokinetic, TK), summarized in the end via bioaccumulation metrics (BCF/BSAF/BMF for water, sediment and food exposure, respectively). From a regulatory point of view, these bioaccumulation metrics are key decision criteria to determine the bioaccumulative property of chemical substances2, and to further assess potential risks that are associated with them according to the exposure sources. Kinetic bioaccumulation metrics are always defined as ratios between uptake and elimination rates, these latter being estimated from TK models1,3. More precisely, TK models relate the exposure concentration to a given chemical substance to the internal concentration within organisms, considering various processes such as absorption, depuration, metabolism and excretion (ADME)4. These last decades, different types of TK models have been proposed, all being compartment models5,6, where organs are considered as biological compartments connected through a fluid, usually blood. Organisms are thus described as whole or divided into organs, with input and output fluxes whose dynamics are described by TK models. Data collected from standard bioaccumulation tests are used to fit the TK models providing uptake and elimination rate estimates. They are also a way to provide a quantitative mechanistic framework to understand and simulate the time-course of the concentration of a chemical substance in various target organs accounting for body fluids in the case of physiologically-based toxicokinetic (PBTK) models. More generally, PBTK models allow to perform extrapolations that are inherent to risk assessment (e.g., extrapolations from one species to another, between exposure routes, from one exposure scenario to other ones, …), and to calculate bioaccumulation metrics5 thus helping decision-makers in the regulatory context. Université de Lyon, Université Lyon 1, CNRS UMR5558, Laboratoire de Biométrie et Biologie Evolutive, 69100, Villeurbanne, France. ✉e-mail: Scientific Data | (2022) 9:130 | https://doi.org/10.1038/s41597-022-01248-y 1 www.nature.com/scientificdata/ www.nature.com/scientificdata Fig. 1 Conceptual framework of the collection of TK data and their storage in the database of MOSAICbioacc. Some years ago, the United States Environmental Protection Agency (US EPA) created a database for ecotoxicology data providing such bioaccumulation metrics according to species-compound combinations7. However, raw data are rarely provided in this database as this would require it to be collected directly from the corresponding scientific papers. Moreover, when available, raw data are mainly provided as plots. Nevertheless, to make updates or develop new TK modelling frameworks, it is crucial for researchers to be able to benefit from a collection of raw TK data in order to check the robustness of their new approaches. While developing the MOSAICbioacc web application8 (https://mosaic.univ-lyon1.fr/bioacc), we dealt with such a lack of raw data to fully test our innovative method. Indeed, MOSAICbioacc provides estimates of TK model parameters and bioaccumulation metrics with their uncertainties for a large range of species-compound combinations (e.g., aquatic or terrestrial organisms exposed to metals, hydrocarbons, active substances, etc.), encompassing different exposure routes and elimination processes. In particular, it was difficult to collect raw TK data with biotransformation processes being involved9, preventing us from fully testing the robustness of MOSAICbioacc for the widest possible diversity of input data types. This motivated us in creating a new publicly available database as presented in this paper. This new database gathers together more than 200 datasets of published bioaccumulation tests, concerning more than 50 genus and more than 120 chemical substances. Some datasets concern several exposure routes (water, soil or sediment, and/or food) and several possible elimination processes (excretion, growth dilution and/or biotransformation). All the collected data were standardized in the same units and uploaded in MOSAICbioacc, ensuring the use of the same methodology in the acces of bioaccumulation metrics for all the datasets. This database should allow the current lack of raw TK data in ecotoxicology to be overcome. Indeed, our purpose with this first version of the database is to motivate other researchers to share their data based on the Findability, Accessibility, Interoperability and Reuse (FAIR) principles which have become today almost a duty10. Our database can be considered as a proof of concept of the added-value of sharing raw TK data. It allows anyone to reuse data, for example to test new modelling frameworks. This database should also facilitate the design of new bioaccumulation experiments, as well as the comparison of results between several studies, benefiting from the unified calculation method for bioaccumulation metrics as provided by MOSAICbioacc. Finally, we (...truncated)