Test battery with the human cell line activation test, direct peptide reactivity assay and DEREK based on a 139 chemical data set for predicting skin sensitizing potential and potency of chemicals

Abstract

To develop a testing strategy incorporating the human cell line activation test (h-CLAT), direct peptide reactivity assay (DPRA) and DEREK, we created an expanded data set of 139 chemicals (102 sensitizers and 37 non-sensitizers) by combining the existing data set of 101 chemicals through the collaborative projects of Japan Cosmetic Industry Association. Of the additional 38 chemicals, 15 chemicals with relatively low water solubility (log Kow > 3.5) were selected to clarify the limitation of testing strategies regarding the lipophilic chemicals. Predictivities of the h-CLAT, DPRA and DEREK, and the combinations thereof were evaluated by comparison to results of the local lymph node assay. When evaluating 139 chemicals using combinations of three methods based on integrated testing strategy (ITS) concept (ITS-based test battery) and a sequential testing strategy (STS) weighing the predictive performance of the h-CLAT and DPRA, overall similar predictivities were found as before on the 101 chemical data set. An analysis of false negative chemicals suggested a major limitation of our strategies was the testing of low water-soluble chemicals. When excluded the negative results for chemicals with log Kow > 3.5, the sensitivity and accuracy of ITS improved to 97% (91 of 94 chemicals) and 89% (114 of 128). Likewise, the sensitivity and accuracy of STS to 98% (92 of 94) and 85% (111 of 129). Moreover, the ITS and STS also showed good correlation with local lymph node assay on three potency classifications, yielding accuracies of 74% (ITS) and 73% (STS). Thus, the inclusion of log Kow in analysis could give both strategies a higher predictive performance.

Tri-culture system for pro-hapten sensitizer identification and potency classification

“…The chemicals identified as false positives by the c-tree for the tri-culture system warrant notice; in particular, all three concentrations of vanillin were categorized as sensitizers. Vanillin has been shown to be weakly sensitizing in humans and guinea pigs and was identified as a “false” positive in U-SENS™, a THP-1 study using ROS to predict sensitization, and an in silico combined test strategy using h-CLAT, DPRA, and DEREK 37–39 . It stands to reason that vanillin could indeed be a weakly sensitizing agent and not a non-sensitizer, in which case, there was no error in classification.…”

An economic approach to non-animal toxicity testing for skin sensitisation

“…"2 out of 3" ITS approach: (Bauch et al, 2012;Natsch et al, 2013;Urbisch et al, 2015a); Kao ITS and Kao STS: (Nukada et al, 2013;Takenouchi et al, 2015); RIVM STS:(van der Veen et al, 2014a;van der Veen et al, 2014b): Stacking meta model: (Gomes, 2012); IDS: (Matheson, 2015;Strickland et al, 2016); BN ITS: Jaworska et al, 2013;Jaworska et al, 2015); ANN ITS: (Hirota et al, 2013;Tsujita-Inoue et al, 2014;Hirota et al, 2015); EC-JRC: (Dimitrov et al, 2005;Asturiol et al, 2016); Global and local regression models: ; IATA: (Patlewicz et al, 2014, Patlewicz et al, 2015 was checked using the leave-one-out validation (Strickland et al, 2016). For the ANN-ITS, the ability of the model to predict the final decision outcome on the skin sensitisation potential was validated using the 10-fold cross validation approach (Hirota et al, 2013).…”

“…For example, Kao DA (Nukada et al, 2013;Takenouchi et al, 2015) covers the first and third key event, whereas the EC-JRC DA covers the molecular initiating event (MIE) only (Dimitrov et al, 2005;Asturiol et al, 2016), which is considered to determine the final conclusion on the skin sensitisation potential (Asturiol et al, 2016). Given that only animal tests can provide information on the fourth key event (i.e.…”

“…Hence, ignoring testing costs means that existing DAs do not provide an endogenous stopping rule to testing. Rather, the stopping rule for deterministic approaches such as the "2 out of 3" ITS (Bauch et al 2012;Urbisch et al 2015b), the RIVM STS (van der Veen et al, 2014a;van der Veen et al, 2014b) and the Kao DA suggestions (Nukada et al, 2013;Takenouchi et al, 2015) is often based on exogenous decision rules such as AOP coverage. In probabilistic DAs such as the BN-ITS Jaworska et al, 2013;Jaworska et al, 2015), the stopping rule is exogenously determined setting information targets i.e.…”

“…Hence, they are presented in different ways, for example in the form of qualitative flowcharts (Grindon et al, 2008e;Mekenyan et al, 2010;ECHA, 2014a;Patlewicz et al, 2015), probabilistic approaches (machine learning) applying Artificial Neural Networks (Hirota et al, 2013;Tsujita-Inoue et al, 2014;Hirota et al, 2015) or Bayesian Networks (Jaworska and Hoffmann, 2010;Jaworska et al, 2013;Jaworska et al, 2015), and as deterministic approaches based on a "majority vote" decision rule for batteries of testing methods (Bauch et al, 2012;van der Veen et al, 2014a;van der Veen et al, 2014b;Urbisch et al, 2015a) or scorebased batteries of testing methods (Ellison et al, 2010;Nukada et al, 2013;Takenouchi et al, 2015). In addition, a regression analysis model ) and a quantitative model using the toxico-kinetics and toxico-dynamics modelling (MacKay et al, 2013) are used.…”