Background: Drug-induced liver injury (DILI) is a major source of acute liver failure and is one of the leading causes of drug development failures. As such, there remains an important unmet need for earlier identification and mitigation of DILI risk. The DILIsym™ model is a mechanistic representation of DILI in preclinical species and humans designed to address this need. The first generation model focused on hepatocyte drug interactions, with limited representation of inflammation. To facilitate the quantitative investigation of innate immune responses in DILI, we have expanded the representation to include liver macrophages, liver sinusoidal endothelial cells, and various mediators. The model represents the current understanding of mechanistic links between hepatocyte death, immune cell activation, mediator production, and mediator effects on hepatocyte death and regeneration. Selection of model parameters was informed by the literature. Mediator profiles and local accumulation of macrophages in the liver were aligned with reported data from acetaminophen (APAP) hepatotoxicity. HMGB1 is a classic alarmin or damage-associated molecular pattern (DAMP) molecule that can induce immune cell activation and is represented in the model. HMGB1 can be released from dying cells, including acetaminophen-exposed hepatocytes 1,2 and can also be produced by activated immune cells 3. Several reports demonstrate that in mice, neutralization of HMGB1 reduces APAP-mediated ALT elevation 4–6.
Results: In the model, HMGB1 neutralization reduces APAP-mediated ALT elevation (max 30% reduction) in the baseline simulated mouse. The simulated improvement fell within the range reported in the experimental literature (10-70% reduction). The model parameters were varied to create earlier or higher HMGB1 profiles. An earlier HMGB1 profile did not improve the effect of HMGB1 neutralization on APAP hepatotoxicity. In contrast, HMGB1 neutralization was more effective in alleviating APAP hepatotoxicity when HMGB1 levels were higher, largely due to an increase in HMGB1-mediated immune activation (55% reduction). Lastly, model parameters were varied such that HMGB1-mediated immune cell activation led to a greater proportion of immune (e.g., TNF-α) mediated hepatocyte death relative to reactive-metabolite mediated hepatocyte death. It was initially surprising to observe that while the degree of liver injury was similar, this change led to slower progressing liver injury. Closer examination revealed that less reactive-metabolite mediated cell death reduced the level of DAMP release, which slowed immune cell activation and the subsequent immune-mediated injury. HMGB1 neutralization was also more effective in alleviating APAP hepatotoxicity in this scenario (55% reduction).
Conclusion: This research reports on how different HMGB1 profiles are predicted to translate to HMGB1 neutralization response and illustrates the application of the DILIsym™ model to test hypotheses regarding the role of the innate immune response in DILI.
NEDMDG/ISSX Workshop: The Role of Drug Metabolism in Immune-Mediated Drug Toxicity: Molecular, Clinical and Mechanistic Aspects, April 17-19, 2013, Cambridge, MA
Lisl K.M. Shoda, Scott Q. Siler, David S. Pisetsky, Paul B. Watkins, Brett A. Howell