A very important limitation of most in vitro toxicity tests and in vitro toxicity data is the lack of measured exposure levels. Unfortunately, the direct analytical determination of exposure in high-throughput in vitro assays is generally difficult due to the combination of rich media and small well volumes. The rich medium leads to substantial binding of many chemicals, which in turn leads to low freely dissolved fractions of chemicals and may lead to unrealistically high effect concentrations (Groothuis et al., 2015). This asks for analytical techniques that can measure freely dissolved concentrations (Cfree) rather than the traditionally used nominal or total concentrations (Cnominal or Ctotal). Solid Phase Microextraction (SPME) is the most appropriate method for this purpose (Smith and Schäfer, 2016), and has been widely applied in simple systems using vials or 24 well plates (Heringa et al., 2003; Kramer et al., 2012). However, this technique is very difficult to apply correctly within the smaller volume of multiwell plates and for complex, dynamic assays with multiple cell types and co-cultures. While mass balance models have been very beneficial for studying and predicting the fate and exposure of test compounds within in vitro assays (Armitage et al., 2014), it remains imperative to develop and optimize methodology to actually measure this exposure and validate the prediction models and assess the applicability domain of such models. Ultimately, data from in vitro assays can only be used for quantitative in vitro to in vivo extrapolation (QIVIVE) if the freely dissolved and cellular concentrations (Ccell) are known.
The objective of this project is to bundle existing expertise to progress exposure assessment in in vitro bioassays used for high-throughput screening (HTS) in 384- and 1536-well plates and complex in vitro bioassays based on transwell and 3D cultures. Since direct assessment of Cfree by SPME is not feasible in 1536-well plates (typical volume of medium 6 μL) or only possible for chemicals with favourable physicochemical properties in 384-well plates (typical volume of medium 40 μL), we propose a pragmatic approach to characterise the fate of chemicals in the bioassay systems by a combination of measurement of Cfree and binding to system components in larger-volume systems (100 to 1000 μL) and modelling followed by the development of a routine experimental approach that can be applied for HTS on robotic systems. The common denominator of exposure assessment will be SPME methods based on different types of polymers for neutral and ionisable chemicals. These SPME methods will be applied to determine free concentrations (or free fractions) in the assay medium as well as fate processes like evaporation, binding to the plastic of the well plates, degradation, and binding to medium constituents and cells, in which proteins and lipids are the dominant binding phases.
The short-term expected outcomes are experimentally validated models to predict the freely dissolved and cellular effect concentrations ECfree and ECcell for existing in vitro toxicity data, including ToxCast and published data that were based on nominal effect concentrations only, which will then make them amenable for QIVIVE. The long-term benefit of the project will be routine analytical tools that will improve exposure assessment in HTS tests in the future and could revolutionise the application of HTS in risk assessment.
The project is composed of four work packages: WP 1 focuses on analytical measures for the free fraction and loss processes including sorption, volatilisation and degradation. In WP2 the developed methods will be applied to parameterise a time-resolved mass balance model for in vitro assays. Here we will measure the system parameters (dry weight, lipid and protein content), partitioning processes between the different organic phases, which are assumed to reach fast equilibrium, and the kinetics of those processes, for which this assumption does not hold, namely the sorption to the plate material and the uptake into cells (Stadnicka-Michalak et al., 2014). In WP3 more complex, dynamic in vitro assay systems will be evaluated. While WP 1 to 3 are targeted to understand and describe exposure in in vitro systems and to develop predictive models, in WP4 the methods and knowledge gained will be applied to develop a routine experimental approach for exposure assessment in HTS.