QST and the Transformation in Drug Safety Assessment
Paul B. Watkins, M.D. FAASLD
Director, Institute for Drug Safety Sciences , University of North Carolina in
Chapel Hill
Wednesday December 16, 2020, 12:00 to 1:00 pm EDT
Register for free at https://www.rosaandco.com/webinars
Abstract:
Establishing the safety of new drug candidates is a major hurdle to drug
development as standard preclinical toxicology does not reliably predict human
adverse drug events. Liver toxicity is a potentially fatal adverse event that
has been particularly challenging to predict from preclinical studies.
Moreover, abnormalities in serum liver chemistries are commonly observed in
clinical trials raising suspicion of liver safety liability that can currently
only be removed with very large clinical trials. This talk will focus on the
progress of a public-private partnership (the DILI-sim Initiative) that for the
last decade has been developing a Quantitative Systems Toxicology (QST) model
(DILIsym(r)) to improve mechanistic understanding and therefore prediction of
liver safety liabilities of new drug candidates.
The DILIsym model uses PBPK and other available data to determine the
concentration of parent drug and major metabolites inside the hepatocyte during
various dosing regimens. Also fed into the model are the exposure dependent
effects of parent drug and major metabolites on oxidative stress, bile acid
homeostasis, and mitochondrial function as measured in in vitro or cellular
systems. Parameters in the model have been varied to reflect genetic and
non-genetic variability to create a virtual healthy human population as well as
disease-specific populations. With the data inputs, DILIsym will predict the
incidence and severity of liver injury that will be observed in a simulated
patient population as a function of dosing regimen. Results of DILIsym modeling
are increasingly used in decision making within Pharma and have also been
helpful in interactions with regulators.
DILIsym provides an example of how increased application of QST modeling should
transform the safety assessment of new drug candidates as well as risk
management in clinical trials and post-approval.