Mechanistic agent-based damage and repair models as hypotheses for patterns of necrosis caused by drug induced liver injury

Increasing model reuse and facilitating repurposing is expected to expand simulation use for better understanding biological phenomena. We demonstrate doing so in the context of liver diseases caused by toxic exposure to xenobiotics. A clinical goal is improved mechanistic explanations of how damage is generated, which can lead to new strategies to block and/or reverse injury. A goal for this work is to provide concrete, plausible explanations for acetaminophen induced liver injury (AILI) in mice. We instantiate mechanistic hypotheses that map to cellular damage and repair pathways and begin identifying plausible simulated causal cascades capable of generating the characteristic AILI spatial and temporal patterns. We use discrete event simulation of agent-based, multiscale, biomimetic models and Monte Carlo sampling. We use an Iterative Refinement protocol for implementing and validating/falsifying mechanistic hypotheses on a previously validated In Silico Liver. We simulated an observed necrosis pattern. Further approach improvement will yield new methods that combine iterations of in-silico and wet-lab experiments. Abbreviations: APAP: acetaminophen; AILI: APAP induced liver injury; CV: central vein; D: damage objects; DILI: drug induced livery injury; GSH: glutathione and analog counterpart objects; ISL: In Silico Liver; IR: Iterative Refinement; NAPQI: reactive metabolite of APAP; N: analog counterpart objects of NAPQI; PV: portal vein; R: Repair objects; SM: similarity measure; SCyc: simulation cycle; TA: targeted attribute