WP3: Gradual improvement of predictive capabilities of in vitro tests

 

WP3 Results pdf 120 kB, download the entire summary.

 

The objectives in WP3 have been to:

1. Provide a database platform (AcuBase) to facilitate storage of SOPs, transfer of data from all partners (in vivo human and animal data, in vitro experimental data) and statistical analyses of larger data sets
2. Identify outliers from the in vivo-in vitro comparisons of data obtained from WP1 and WP2
3. Support with competence in multivariate modelling as performed in the MEIC (Multicenter Evaluation of In Vitro Cytotoxicity) project (Ekwall et al., 1998, 2000) and to carry out modelling of new data generated in WP1, WP2, and WP4-7.
4. Adapt the in vitro methods, selected for inclusion in the testing strategy, to robotic platforms for high-throughput screening (HTS)

1. AcuBase

A pilot version of AcuBase was opened to all ACuteTox partners in March 2006.
The database consists of the following sections:
• Chemicals section
• Animal in vivo section
• Human poisoning cases
• Kinetics
• Reports

 

At the end of June 2007 AcuBase contained a full set of data regarding the selected 97 reference chemicals, including molecular structure, physicochemical properties and summary descriptions. Moreover, 2204 data from acute oral toxicity studies in vivo, 2902 data sets from in vivo human blood poisoning reports, 6512 files from in vitro experiments and 80 SOPs were introduced into the database by partners and/or administrators.

 

2. The robotic platform for HTS

In WP3, the 3T3/NR uptake and HepG2/MTT assays have successfully been adapted to two commercially available HTS robotic platforms.

 3-4. Multivariate modeling and identification of outliers

In vitro-in vivo modelling of LC50 values for humans and LD50 values for rats have been performed using different combinations of in vitro tests.

 

Furthermore, the in vitro-in vivo linear regression analysis has shown that the in vitro basal cytotoxicity assays correlated better with human peak lethal blood concentrations (LC50) than with LD50 rat values. From this analysis it was also concluded that the in vitro-in vivo correlation based on mol/l and mol/kg was superior to g/l and g/kg.