Combined in silico and 3D in vitro approaches for the accurate prediction of human drug induced liver injury

Abstract: Utilising 3D in vitro approaches to de-risk drug induced liver injury (DILI) is an emerging trend. These approaches primarily use primary human hepatocytes (PHH) and kupffer cells to form microtissues (MTs). Various studies have now illustrated the advantages of these approaches over traditional 2D assays utilising a single biochemical endpoint ATP1. We have further developed this approach by applying multiparametric confocal high content imaging to determine mitochondrial membrane potential, reactive oxygen species formation and GSH content in addition to ATP and have evaluated PHH in co-culture as well as HepaRG 3D human liver MTs2. CYP450 characterisation of MTs showed significantly higher activities in HepaRG MTs compared to PHH MTs (CYP3A4: 210pM/min/MT HepaRG, 80pM/mn/MT PHH). Following characterisation DILI reference compounds were screened through each model and confocal high content imaging was used in addition to ATP activity to determine DILI risk.

Prediction of human DILI was improved in the HepaRG MT when compared to the overall response in the PHH MTs, with hepatocyte donor differences observed. In addition the accuracy of the prediction of DILI was increased further using a multiparametric approach compared to measurement of ATP alone. Furthermore, incorporation of in silico human liver exposure data further enhanced the sensitivity of this approach.

Our research has demonstrated that 3D MTs enhances our ability to predict DILI risk over traditional 2D techniques. This prediction of DILI is also significantly improved by using multi-parametric approaches and in silico human exposure data. In vitro models such as the system detailed above are amendable to high throughput screening, are cost effective and as such could play an important role in the pharmaceutical strategies for mitigating DILI risk early in drug discovery.