Development and partial validation of a mouse model for predicting drug hypersensitivity reactions

Abstract

Animal models that can be used to predict the allergenic potential of drug candidates have not been adequately optimized, validated, or characterized. While initial validation data from an inter-laboratory study of the mouse lymph node proliferation assay (LNPA) appeared promising, no additional investigations in this model have been reported. The objectives of this study were to use positive and negative control drugs to further optimize and validate the LNPA utilizing a non-radioactive endpoint and determine the sensitivity, specificity, and predictivity of the model. Drugs associated with hypersensitivity reactions in the literature were chosen to test in the model in addition to drugs with few or no reports of hypersensitivity. Mice received a subcutaneous injection of drug or vehicle into the scruff of the neck once daily for a period of 3 days. On Day 6, draining lymph nodes were harvested, single cell suspensions prepared, and total cell numbers determined for each animal by flow cytometry. A stimulation index was calculated by dividing the mean total cell number for the drug-treated group by the mean total cell number for the vehicle-treated animals. Based on statistical analysis of the data, animals with a total cell number ≥2.5× the mean of the vehicle group were classified as ‘responders’. Based on data generated to date with 12 positive control and six negative control drugs, the model had a sensitivity of 75%, a specificity of 74%, and a relatively good predictive value (measured by the Receiver Operating Characteristic AUC of 0.80). The data here suggest that this model may be a useful tool for identifying drug candidates with the potential to produce allergic responses in the clinic. Future studies will investigate the mechanism(s) for the lymph node responses in order to develop additional endpoints that may increase the sensitivity and specificity of the model.