Drug-induced liver injury (DILI) is not only a major concern for all patients requiring drug therapy, but also for the pharmaceutical industry. Many new in vitro assays and pre-clinical animal models are being developed to help screen compounds for the potential to cause DILI. This study demonstrates that mechanistic, mathematical modeling offers a method for interpreting and extrapolating results. The DILIsym™ model (version 1A), a mathematical representation of DILI, was combined with in vitro data for the model hepatotoxicant methapyrilene (MP) to carry out an in vitro to in vivo extrapolation. In addition, simulations comparing DILI responses across species illustrated how modeling can aid in selecting the most appropriate pre-clinical species for safety testing results relevant to humans. The parameter inputs used to predict DILI for MP were restricted to in vitro inputs solely related to ADME (absorption, distribution, metabolism, elimination) processes. MP toxicity was correctly predicted to occur in rats, but was not apparent in the simulations for humans and mice (consistent with literature). When the hepatotoxicity of MP and acetaminophen (APAP) was compared across rats, mice, and humans at an equivalent dose, the species most susceptible to APAP was not susceptible to MP, and vice versa. Furthermore, consideration of variability in simulated population samples (SimPops™) provided confidence in the predictions and allowed examination of the biological parameters most predictive of outcome. Differences in model sensitivity to the parameters were related to species differences, but the severity of DILI for each drug/species combination was also an important factor.

Troglitazone (TGZ) caused delayed, life-threatening drug-induced liver injury (DILI) in some patients, but was not hepatotoxic in rats. This study investigated altered bile acid (BA) homeostasis as a mechanism of TGZ hepatotoxicity using a systems pharmacology model incorporating drug/metabolite disposition, BA physiology/pathophysiology, hepatocyte life cycle, and liver injury biomarkers. In the simulated human population, TGZ (200–600mg/day×6months) resulted in delayed increases in serum ALT>3× ULN in 0.3–5.1% of the population with concomitant bilirubin elevations>2× ULN in 0.3–3.6%. In contrast, pioglitazone (15–45mg/day×6months) did not elicit hepatotoxicity, consistent with clinical data. TGZ was not hepatotoxic in the simulated rat population. In summary, mechanistic modeling based only on BA effects accurately predicted the incidence, delayed presentation, and species differences in TGZ hepatotoxicity, and the relative liver safety of pioglitazone. Systems pharmacology models integrating physiology and experimental data can evaluate DILI mechanisms and may be useful to predict hepatotoxic potential of drug candidates.

Tolvaptan is a selective vasopressin V2 receptor antagonist, approved in several countries for the treatment of hyponatremia and autosomal dominant polycystic kidney disease (ADPKD). No liver injury has been observed with tolvaptan treatment in healthy subjects and in non-ADPKD indications, but ADPKD clinical trials showed evidence of drug-induced liver injury (DILI). Although all DILI events resolved, additional monitoring in tolvaptan-treated ADPKD patients is required. In vitro assays identified alterations in bile acid disposition and inhibition of mitochondrial respiration as potential mechanisms underlying tolvaptan hepatotoxicity. This report details the application of DILIsym software to determine whether these mechanisms could account for the liver safety profile of tolvaptan observed in ADPKD clinical trials. DILIsym simulations included physiologically based pharmacokinetic estimates of hepatic exposure for tolvaptan and2 metabolites, and their effects on hepatocyte bile acid transporters and mitochondrial respiration. The frequency of predicted alanine aminotransferase (ALT) elevations, following simulated 90/30  mg split daily dosing, was 7.9% compared with clinical observations of 4.4% in ADPKD trials. Toxicity was multifactorial as inhibition of bile acid transporters and mitochondrial respiration contributed to the simulated DILI. Furthermore, simulation analysis identified both pre-treatment risk factors and on-treatment biomarkers predictive of simulated DILI. The simulations demonstrated that in vivo hepatic exposure to tolvaptan and the DM-4103 metabolite, combined with these 2 mechanisms of toxicity, were sufficient to account for the initiation of tolvaptan-mediated DILI. Identification of putative risk-factors and potential novel biomarkers provided insight for the development of mechanism-based tolvaptan risk-mitigation strategies.

N-acetylcysteine (NAC) is the treatment of choice for acetaminophen poisoning; standard 72-h oral or 21-h intravenous protocols are most frequently used. There is controversy regarding which protocol is optimal and whether the full treatment course is always necessary. It would be challenging to address these questions in a clinical trial. We used DILIsym, a mechanistic simulation of drug-induced liver injury, to investigate optimal NAC treatment after a single acetaminophen overdose for an average patient and a sample population (n ϭ 957). For patients presenting within 24 h of ingestion, we found that the oral NAC protocol preserves more hepatocytes than the 21-h intravenous protocol. In various modeled scenarios, we found that the 21-h NAC infusion is often too short, whereas the full 72-h oral course is often unnecessary. We found that there is generally a good correlation between the time taken to reach peak serum alanine aminotransferase (ALT) and the time taken to clear N-acetyl-p-benzoquinone imine (NAPQI) from the liver. We also found that the most frequently used treatment nomograms underestimate the risk for patients presenting within 8 h of overdose ingestion. V max for acetaminophen bioactivation to NAPQI was the most important variable in the model in determining interpatient differences in susceptibility. In conclusion, DILIsym predicts that the oral NAC treatment protocol, or an intravenous protocol with identical dosing, is superior to the 21-h intravenous protocol and ALT is the optimal available biomarker for discontinuation of the therapy. The modeling also suggests that modification of the current treatment nomograms should be considered.

Bile salt export pump (BSEP) inhibition has been proposed to be an important mechanism for drug-induced liver injury (DILI). Modeling can prioritize knowledge gaps concerning bile acid (BA) homeostasis and thus help guide experimentation. A submodel of BA homeostasis in rats and humans was constructed within DILIsym, a mechanistic model of DILI. In vivo experiments in rats with glibenclamide were conducted, and data from these experiments were used to validate the model. The behavior of DILIsym was analyzed in the presence of a simulated theoretical BSEP inhibitor. BSEP inhibition in humans is predicted to increase liver concentrations of conjugated chenodeoxycholic acid (CDCA) and sulfate-conjugated lithocholic acid (LCA) while the concentration of other liver BAs remains constant or decreases. On the basis of a sensitivity analysis, the most important unknowns are the level of BSEP expression, the amount of intestinal synthesis of LCA, and the magnitude of farnesoid-X nuclear receptor (FXR)-mediated regulation.

Inhibition of the bile salt export pump (BSEP) has been linked to incidence of drug-induced liver injury (DILI), presumably by the accumulation of toxic bile acids in the liver. We have previously constructed and validated a model of bile acid disposition within DILIsym®, a mechanistic model of DILI. In this paper, we use DILIsym® to simulate the DILI response of the hepatotoxic BSEP inhibitors bosentan and CP-724,714 and the non-hepatotoxic BSEP inhibitor telmisartan in humans in order to explore whether we can predict that hepatotoxic BSEP inhibitors can cause bile acid accumulation to reach toxic levels. We also simulate bosentan in rats in order to illuminate potential reasons behind the lack of toxicity in rats compared to the toxicity observed in humans. DILIsym® predicts that bosentan, but not telmisartan, will cause mild hepatocellular ATP decline and serum ALT elevation in a simulated population of humans. The difference in hepatotoxic potential between bosentan and telmisartan is consistent with clinical observations. However, DILIsym® underpredicts the incidence of bosentan toxicity. DILIsym® also predicts that bosentan will not cause toxicity in a simulated population of rats, and that the difference between the response to bosentan in rats and in humans is primarily due to the less toxic bile acid pool in rats. Our simulations also suggest a potential synergistic role for bile acid accumulation and mitochondrial electron transport chain (ETC) inhibition in producing the observed toxicity in CP-724,714, and suggest that CP-724,714 metabolites may also play a role in the observed toxicity. Our work also compares the impact of competitive and noncompetitive BSEP inhibition for CP-724,714 and demonstrates that noncompetitive inhibition leads to much greater bile acid accumulation and potential toxicity. Our research demonstrates the potential for mechanistic modeling to contribute to the understanding of how bile acid transport inhibitors cause DILI.

Tolcapone and entacapone are catechol‐O‐methyltransferase (COMT) inhibitors developed as adjunct therapies for treating Parkinson's disease. While both drugs have been shown to cause mitochondrial dysfunction and inhibition of the bile salt export protein (BSEP), liver injury has only been associated with the use of tolcapone. Here we used a multiscale, mechanistic model (DILIsym®) to simulate the response to tolcapone and entacapone. In a simulated population (SimPops™) receiving recommended doses of tolcapone (200 mg t.i.d.), increases in serum alanine transaminase (ALT) >3× the upper limit of normal (ULN) were observed in 2.2% of the population. In contrast, no simulated patients receiving recommended doses of entacapone (200 mg 8× day) experienced serum ALT >3× ULN. Further, DILIsym® analyses revealed patient‐specific risk factors that may contribute to tolcapone‐mediated hepatotoxicity. In summary, the simulations demonstrated that differences in mitochondrial uncoupling potency and hepatic exposure primarily account for the difference in hepatotoxic potential for tolcapone and entacapone.

The drug development industry faces multiple challenges in the realization of safe effective drugs. Computational modeling approaches can be used to support these efforts. One approach, mechanistic modeling, is new to the realm of drug safety. It holds the promise of not only predicting toxicity for novel compounds, but also illuminating the mechanistic underpinnings of toxicity. To increase the scientific community's familiarity with mechanistic modeling in drug safety, this article seeks to provide perspective on the type of data used, how they are used and where they are lacking. Examples are derived from the development of DILIsym ® software, a mechanistic model of drug-induced liver injury (DILI). DILIsym ® simulates the mechanistic interactions and events from compound administration through the progression of liver injury and regeneration. Modeling mitochondrial toxicity illustrates the type and use of in vitro data to represent biological interactions, as well as insights on key differences between in vitro and in vivo conditions. Modeling bile acid toxicity illustrates a case in which the over-arching mechanism is well accepted, but many mechanistic details are lacking. Modeling was used to identify measurements predicted to strongly impact toxicity. Finally, modeling innate immune responses illustrates the importance of time-series data, particularly in the presence of positive and negative feedback loops, as well as the need for data from different animal species for better translation. These concepts are germane to most mechanistic models, although the details will vary. The use of mechanistic models is expected to improve the rational design of new drugs.