Florian Simon (1), Matthieu Jacobs (1), Hélène Aerts, (2), Olivier Wattrelos (2), Sylvain Fouliard (1)
(1) Translational Pharmacometrics - Quantitative Pharmacology, Servier, Suresnes, France, (2) Biologie Servier, Non-clinical safety, Gidy, France
Objectives:
Hepatotoxicity is one of the most frequent cause of safety related withdrawals for the past fifty years. It is difficult to predict due to many underlying mechanisms and poor predictivity of animal models. The FDA has strongly encouraged the application of modeling in drug development to improve product safety (1).
A quantitative system toxicology model of drug-induced liver injury, DILIsym®, was developed (2). It integrates in vitro data on several mechanisms related to hepatotoxicity such as bile acids accumulation, mitochondrial dysfunction or oxidative stress production in addition to the pharmacokinetics (PK) of the compound in order to quantitively predict the viability of hepatocytes and the dynamics of the release of hepatotoxicity biomarkers (e.g. alanine aminotransferase (ALT) and bilirubin).
Drug S is currently in early clinical development and some hepatic concerns were observed in the form of asymptomatic, reversible and dose-dependent bilirubin elevations. In this work, DILIsym was used in order to evaluate its ability to reproduce the clinical observations and confirm the underlying mechanisms related to hepatotoxicity.
Methods:
In vitro assays :
Several in vitro assays were performed in order to measure the impact of Drug S on :
– Bile acids accumulation : BSEP inhibition assay on primary human hepatocytes.
– Mitochondrial dysfunction : Seahorse assay performed in hepG2 cells to measure changes of oxygen consumption rate.
– Oxidative stress production : ROS generation measured by high-content analysis in hepaRG cells.
– Bilirubin metabolism and transport : Inhibition assay of UGT1A1 in human liver microsomes and human recombinant system, and OATP1B1 in human transfected cells.
Clinical data :
Mean observed clinical PK profiles following 50mg, 100mg, 150mg and 200mg daily administration at steady-state from phase 1 were used in this work.
Plasma bilirubin elevations have been observed in around 10% of patients.
QST simulations :
DILIsym software was used to perform the QST simulations. The in vitro data were converted into input parameters for use in DILIsym.
Simulations were performed using a subset (N=10) of an healthy volunteer population including variability on bilirubin metabolizing-enzymes and transporters parameters (e.g. UGT1A1, OATP, MRP2 and MRP3 expression levels). Treatment administrations were simulated for 2 months for each dose level.
The dynamic and the elevation magnitude of hepatotoxicity biomarkers were then compared to the clinical observations.
Results:
In vitro results did not show toxicity signals regarding BSEP inhibition and mitochondrial respiration even at the highest tested concentrations of Drug S, respectively 100µM and 70µM. ROS production ratio reached 1.3 after 4h exposure at the highest tested concentration of 50µM. Regarding bilirubin metabolism and transport, Drug S inhibited UGT1A1 with an IC50 of 0.55µM-1.46µM depending on the in vitro setting and a concentration dependent inhibition of OATP1B1 with an IC50 of 2.3µM was also measured.
These mechanistic data were combined with the observed mean PK profiles for each dosing regimen.
Elevation of total bilirubin in plasma above 1.5-fold the upper limit of normal range (ULN) was found to be dose dependent. However, the few observed plasma ALT increases were not predicted by the model. Drug S inhibition of UGT1A1 seems to be the main driver of these changes in the model, leading to an inhibition of bilirubin metabolism without inducing hepatocytes necrosis that explain the absence of predicted ALT release in plasma.
Conclusions:
DILIsym model was able to describe the dose dependent bilirubin elevations observed in phase I, this demonstrates how combining in vitro experimental methods with QST can lead to improved predictions about the underlying mechanisms behind drug-induced toxicity.
References:
[1] Temple R. Guidance for Industry. Drug Safety. 2009;28.
[2] Watkins PB. The DILI‐sim Initiative: Insights into Hepatotoxicity Mechanisms and Biomarker Interpretation. Clin Transl Sci. mars 2019;12(2):122‑9
Reference: PAGE 30 (2022) Abstr 10033 [www.page-meeting.org/?abstract=10033]
Poster: Drug/Disease Modelling - Safety