Increasing confidence in a skin sensitization assay using a measurement science approach
Elijah Petersen, Richard Uhl, Blaza Toman, John T. Elliott, John Gordon
Background: The Electrophilic Allergen Screening Assay (EASA) has emerged as a promising method for risk and hazard assessment for skin sensitization classification. This assay functions by assessing the depletion of one of two probe molecules (4-Nitrobenzenethiol (NBT) and Pyridoxylamine (PDA)). The initial development of this method utilized a single cuvette format resulting in multiple measurement challenges such as low throughput and inability to include sufficient control measurements. Objectives: In this study, we describe the redesign of this method into a 96-well format with process control measurements incorporated into a plate design to quantify key sources of variability each time the assay is run. Methods: Sixty compounds were tested in the 96-well plate design. Additional control experiments were performed to evaluate photodegradation of probe molecules, bias from bubbles caused by pipetting, and statistical power of the assay system. Results: One key insight revealed by this process was the interference from test compounds, namely producing an absorbance or fluorescence signal similar to that of the probe molecules, which would not have been previously detected using the single cuvette assay. Not taking the interference into account has been shown to lead to potential false negative identification. The data from the 60 compounds tested has approximately 73% concordance with LLNA data, a result higher than the agreement for the direct peptide reactivity assay (DPRA), an OECD test guideline protein binding assay, which had a concordance of 61 %. 96 % of compounds with > 20 % probe depletion in an assay were sensitizers according to in vivo data. Discussion: Overall, the measurement science approach described here provides steps that can be taken to increase confidence in in vitro assays by fully characterizing sources of variability and potential biases in the assay which will facilitate interlaboratory testing and standardization.