General overview

The AcuteTox project was structured into nine Workpackages (WPs). The nine WPs were interlinked as illustrated in the  Workplan Project Structure.pdf .

General overview of the ACuteTox Work packages

WP1: Generation of an in vivo database
Principal responsibilities of WP1 were to select project reference chemicals, as well as generate an in vivo acute toxicity database for the selected reference chemicals by searching and compiling information from documented sources. WP1 was also responsible for statistical review of the compiled animal data, conceived to provide explicit insight of the LD50 values and establish an objective basis for projected correlation of the in vivo data with results from corresponding in vitro assays. The review made statistical analyses of variability, inter-species correlation, relevance, and predictive capacity for toxicity classification. Results of these analyses provided a basis for performance assessment of in vitro methods and were relevant in development of strategies for evaluation of chemical toxicity.

WP2: Generation of an in vitro database
The aim of WP2 was to merge cytotoxicity data from two projects (NICETAM /ECVAM/ and MEIC/EDIT) and to increase the new database by extending the two studies by performing testing of additional compounds. The final outcome of WP2 is a database containing high quality in vitro toxicity data obtained with defined Standard Operating Procedures for the assay and culture methods for the testing of up to 97 different chemicals, which were used for identification of outliers in comparison with published in vivo data as well as when compared with in vitro data from the tissue specific test. The data were generated in mouse 3T3 Neutral Red Uptake (NRU) assay (3T3 NRU Anon 2006), the Normal Human Keratinocyte (NHK) NRU assay (NHK NRU Anon 2006), the lymphocyte HL60 ATP assay (Wakuri et al 1993), the liver derived Fa32 cells with an NRU and a fluorescent total protein endpoint assay (Dierickx 2000 and 2003) and the liver derived HepG2 fluorescent total protein assay (Dierickx 2005). In WP2 also solubility protocols were defined, which have been recommended for employment in all other WPs.

WP3: Database, statistical analysis, assay automation
WP3 had a central role in the project and it highly relied on input/data coming from other WPs, mainly WPs 1-2 and WPs 4-7. The main objectives of this WP were to develop of an internet-based database for storing and management of all project data; to adapt promising in vitro methods to robotic screening platforms; and finally, to perform statistical analyses of in vitro and in vivo data, with the final goal to identify outliers and to design of the preliminary algorithm/prediction model for prediction of acute oral toxicity. The main tasks in this WP included: 1. To provide AcutoxBase for efficient data management within the project. 2. To adapt commercially available high-throughput (HTS) robotic platforms for handling of different cell systems and different endpoints. 3. To estimate human LC50 values from acute sub-lethal and lethal blood concentration data obtained from the case report compiled in WP1. 4. To identify outliers from the in vivo/in vitro correlations of the data obtained from WP1 and WP2. 5. To perform statistical analysis of the data obtained from testing 57 reference chemicals in 75 in vitro endpoints by use of PLS multivariate analysis.

WP4: New cell systems and new endpoints
The role of WP4 in the project was to provide an innovative, alternative way to improve the predictivity of cell-based cytotoxicity assays by incorporating assays for immunotoxicity and haematotoxicity and new end-points for cytotoxicity. Cytotoxicity assays were approached by novel methodologies based on single-cell analysis, the so-called cytomic techniques (flow cytometry and high-content assays by bioimaging) that expanded the classical cytotoxicity endpoints by introducing novel early markers of cell stress and damage. In addition, WP4 was in charge of selecting reference chemicals with immunotoxic and haematotoxic potential and participated actively to define protocols (SOP) for all methods that should be considered at the selection of the best performing method.

WP5: Alerts and correctors in toxicity screening (I): Role of ADE
The overall objectives of WP5 were focused on the factors that are influencing the relationship between in vitro cytotoxicity data and in vivo doses. By assessing this relationship the estimation of acutely toxic doses on the basis of in vitro toxicity has been improved. The most crucial parts of the kinetic behaviour have been studied: absorption of compounds, distribution between blood and tissues and the passage of special barriers. In the context of acute toxicity, the blood brain barrier (BBB) is the most relevant special barrier and has received extra attention. Both in vitro and in silico models have been used to determine oral absorption and passage over the BBB. A crucial parameter is the use of the proper dose metric in in vitro experiments, i.e. the free concentration of the compound, which has also been measured. The data obtained from WP5 were used for biokinetic modeling in order to transfer the in vitro EC50 values to oral dose.

WP6 and WP7.3: Alerts and correctors in toxicity screening (II and V): Role of metabolism and hepatotoxicity
The toxicity of a compound may be caused not only to the parent molecule but also by its metabolites. Metabolism, indeed, can result in a bioactivation phenomenon rather than in a detoxification process leading to metabolism-dependent toxicity. The occurrence of this phenomenon cannot be assessed by cell lines used for basal cytotoxicity screening, as they lack biotransforming enzymes. Despite many chemicals can cause toxicity directly (active toxins), it is not infrequent that the metabolites formed in situ by hepatic or other metabolic-competent cells be ultimately responsible for the observed toxicity (latent toxins). This explains the differences in the intrinsic in vitro toxicity observed for certain compounds when assessed with metabolically competent vs. non-competent cell lines. The main objective of WP6 was to set up of an assay to generate an “alert” about metabolism-dependent toxicity of a given compound, by testing its effects in a metabolic competent model (primary hepatocytes) and in non-metabolising cells (cell lines), as part of a general acute cytotoxicity testing. Another important aim was to develop a new software not only to determine the IC50s, but to compare up to 3 dose-response curves in testing of metabolism-dependent toxicity and hepatotoxicity and to manage and integrate all the data. Strategies based on engineered cells including expression vectors for transient and controllable expression of biotransformation enzymes for CYP 1A2, 2A6, 2C9, 2E1 and 3A4 were also evaluated as a way to overcome their intrinsic limitations by generation of metabolically competent cell lines. Furthermore, in vitro models were evaluated regarding their metabolic capacity and the possible use of these models for generating data on liver clearance for PBBK-TD modelling (in collaboration with WP5). In addition, metabolite formation of selected compounds in vitro were analyzed and compare the results obtained to in vivo literature data as well as to predictions made with the METEOR software from Lhasa Ltd and where possible, improve the predictive properties of this software. The ultimate goal of WP7.3 was to provide quantifiable information that could be integrated in a wider assessment of in vitro cytotoxicity that could anticipate in vivo toxicity of chemicals. Metabolic competent cells (rat hepatocytes), non-competent hepatic cells (HepG2) and non hepatic cells (3T3 mouse fibroblasts) were used to investigate the effects of a selected list of reference compounds. In WP7.3, also cells systems suitable for hepatic transport assays of anions bile acids and/or xenobiotics, including double transfected cells and ATP dependent transport systems were generated. Furthermore, pilot experiments, using the newly developed fluorescent bile acid derivatives and cell systems, were performed to determine its robustness and suitability medium-throughput testing.

WP7.1: Alerts and correctors in toxicity screening (III): Neurotoxicity
Acute toxicity may be a result of impaired neuronal function, either in the peripheral or the central nervous system. The overall objective for WP7.1 was to carry out testing of the ACuteTox reference chemicals in an optimised neurotoxicity test battery, according to well defined protocols (Standard Operating Procedures), and to deliver high quality in vitro data to AcutoxBase. The selected assays were identified as the best performing assays out of a larger set of assays in which 16 general and 10 neurotoxic reference compounds were tested. The criteria for selection were the ability of the neurotoxicity assays to (i) identify “neurotoxic alerts”, i.e. indicating alteration of neuronal function at lower concentrations than the general cytotoxicity indicated in the 3T3/NRU assay (see WP2), and (ii) “correct” under-estimated toxicity as determined by the 3T3/NRU assay. Eight different cell models for the nervous system were used for the studies on approximately 70 endpoints; in pure enzymes, native or differentiated human neuroblastoma SH-SY5Y cells, primary cultures of mouse or rat cortical or cerebellar granule neurons, mouse brain slices and mature re-aggregated rat brain cells. Several of the endpoints were analyzed in more than one cell model.

WP7.2: Alerts and correctors in toxicity screening (IV): Nephrotoxicity
The kidney is especially susceptible to toxicity because of its role in excreting compounds, which involves a high blood supply, concentrating, metabolizing and transporting compounds. The focus in WP7.2 was on developing in vitro assays which reflect the role of the kidneys in vivo based on functional parameters including transport and barrier function involving a transepithelial cell layer and transport. For the measurement of nephrotoxicity, transepithelial resistance (TER) was chosen as the functional assay and the LLC-PK1 proximal tubular cell line as the test system. The functional assay reflects the in vivo transporting capabilities of the renal proximal tubules. The functional assay was compared to a viability assay namely the resazurin (alamar blue) assay under exactly the same experimental conditions and testing was carried out in the 96-well plate format for both assays. The REMS automated device was selected for measurement of TER

WP8 and WP9: Optimisation of the testing strategy and Prevalidation
Currently, acute oral toxicity is assessed in rats in accordance with the OECD Test Guidelines 420 (OECD, 2001 a), 423 (OECD, 2001 b), 425 (OECD, 2001 c). One of the main goals of the ACuteTox project was to develop and optimise an in vitro testing strategy for predicting human acute oral toxicity and further pre-validate it. The first step consisted in the identification of the methods as promising building blocks for the testing strategy on the basis of an in depth statistical analysis of the large dataset generated with the training set of 57 compounds used during the first phase of the project. During the prevalidation phase, the selected test methods were challenged with a new set of 32 compounds. The work performed during this challenging exercise was focused mainly on the assessment of the predictive capacity of the proposed tiered testing strategies and the identification of the combination that gives the best prediction. Specific objectives of WP9 were: To guarantee that the protocols of the selected methods were sufficiently standardised and optimised (module 1 ECVAM‟s modular approach). To assess the reproducibility (module 2 of ECVAM‟s modular approach) of each building block composing the final tiered testing strategy using a defined set of chemicals. Preliminary assessment of the relevance of the final tiered testing strategy (module 5 ECVAM‟s modular approach).