Predicting idiosyncratic drug-induced liver injury: some recent advances

Role of Gut Microecology in the Pathogenesis of Drug-Induced Liver Injury and Emerging Therapeutic Strategies

Drug-induced liver injury (DILI) is a common clinical pharmacogenic disease. In the United States and Europe, DILI is the most common cause of acute liver failure. Drugs can cause hepatic damage either directly through inherent hepatotoxic properties or indirectly by inducing oxidative stress, immune responses, and inflammatory processes. These pathways can culminate in hepatocyte necrosis. The role of the gut microecology in human health and diseases is well recognized. Recent studies have revealed that the imbalance in the gut microecology is closely related to the occurrence and development of DILI. The gut microecology plays an important role in liver injury caused by different drugs. Recent research has revealed significant changes in the composition, relative abundance, and distribution of gut microbiota in both patients and animal models with DILI. Imbalance in the gut microecology causes intestinal barrier destruction and microorganism translocation; the alteration in microbial metabolites may initiate or aggravate DILI, and regulation and control of intestinal microbiota can effectively mitigate drug-induced liver injury. In this paper, we provide an overview on the present knowledge of the mechanisms by which DILI occurs, the common drugs that cause DILI, the gut microbiota and gut barrier composition, and the effects of the gut microbiota and gut barrier on DILI, emphasizing the contribution of the gut microecology to DILI.

Computational models for predicting liver toxicity in the deep learning era

Drug-induced liver injury (DILI) is a severe adverse reaction caused by drugs and may result in acute liver failure and even death. Many efforts have centered on mitigating risks associated with potential DILI in humans. Among these, quantitative structure-activity relationship (QSAR) was proven to be a valuable tool for early-stage hepatotoxicity screening. Its advantages include no requirement for physical substances and rapid delivery of results. Deep learning (DL) made rapid advancements recently and has been used for developing QSAR models. This review discusses the use of DL in predicting DILI, focusing on the development of QSAR models employing extensive chemical structure datasets alongside their corresponding DILI outcomes. We undertake a comprehensive evaluation of various DL methods, comparing with those of traditional machine learning (ML) approaches, and explore the strengths and limitations of DL techniques regarding their interpretability, scalability, and generalization. Overall, our review underscores the potential of DL methodologies to enhance DILI prediction and provides insights into future avenues for developing predictive models to mitigate DILI risk in humans.

VenomPred: A Machine Learning Based Platform for Molecular Toxicity Predictions

The use of in silico toxicity prediction methods plays an important role in the selection of lead compounds and in ADMET studies since in vitro and in vivo methods are often limited by ethics, time, budget and other resources. In this context, we present our new web tool VenomPred, a user-friendly platform for evaluating the potential mutagenic, hepatotoxic, carcinogenic and estrogenic effects of small molecules. VenomPred platform employs several in-house Machine Learning (ML) models developed with datasets derived from VEGA QSAR, a software that includes a comprehensive collection of different toxicity models and has been used as a reference for building and evaluating our ML models. The results showed that our models achieved equal or better performance than those obtained with the reference models included in VEGA QSAR. In order to improve the predictive performance of our platform, we adopted a consensus approach combining the results of different ML models, which was able to predict chemical toxicity better than the single models. This improved method was thus implemented in the VenomPred platform, a freely accessible webserver that takes the SMILES (Simplified Molecular-Input Line-Entry System) strings of the compounds as input and sends the prediction results providing a probability score about their potential toxicity.

Compensatory Transition of Bile Acid Metabolism from Fecal Disposition of Secondary Bile Acids to Urinary Excretion of Primary Bile Acids Underlies Rifampicin-Induced Cholestasis in Beagle Dogs

Drug induced cholestasis (DIC) is complexly associated with dysbiosis of the host-gut microbial cometabolism of bile acids (BAs). Murine animals are not suitable for transitional studies because the murine BA metabolism is quite different from human metabolism. In this work, the rifampicin (RFP) induced cholestasis was established in beagle dogs that have a humanlike BA profile to disclose how RFP affects the host-gut microbial cometabolism of BAs. The daily excretion of BA metabolites in urine and feces was extensively analyzed during cholestasis by quantitative BA profiling along the primary-secondary-tertiary axis. Oral midazolam clearance was also acquired to monitor the RFP-induced enterohepatic CYP3A activities because CYP3A is exclusively responsible for the tertiary oxidation of hydrophobic secondary BAs. RFP treatments caused a compensatory transition of the BA metabolism from the fecal disposition of secondary BAs to the urinary excretion of primary BAs in dogs, resulting in an infantile BA metabolism pattern recently disclosed in newborns. However, the tertiary BAs consistently constituted limitedly in the daily BA excretion, indicating that the detoxification role of the CYP3A catalyzed tertiary BA metabolism was not as strong as expected in this model. Multiple host-gut microbial factors might have contributed to the transition of the BA metabolism, such as inhibition of BA transporters, induction of liver-kidney interplaying detoxification mechanisms, and elimination of gut bacteria responsible for secondary BA production. Transitional studies involving more cholestatic drugs in preclinical animals with a humanlike BA profile and DIC patients may pave the way for understanding the complex mechanism of DIC in the era of metagenomics.

Prediction and mechanistic analysis of drug-induced liver injury (DILI) based on chemical structure

BACKGROUND

Drug-induced liver injury (DILI) is a major safety concern characterized by a complex and diverse pathogenesis. In order to identify DILI early in drug development, a better understanding of the injury and models with better predictivity are urgently needed. One approach in this regard are in silico models which aim at predicting the risk of DILI based on the compound structure. However, these models do not yet show sufficient predictive performance or interpretability to be useful for decision making by themselves, the former partially stemming from the underlying problem of labeling the in vivo DILI risk of compounds in a meaningful way for generating machine learning models.

RESULTS

As part of the Critical Assessment of Massive Data Analysis (CAMDA) "CMap Drug Safety Challenge" 2019 ( http://camda2019.bioinf.jku.at ), chemical structure-based models were generated using the binarized DILIrank annotations. Support Vector Machine (SVM) and Random Forest (RF) classifiers showed comparable performance to previously published models with a mean balanced accuracy over models generated using 5-fold LOCO-CV inside a 10-fold training scheme of 0.759 ± 0.027 when predicting an external test set. In the models which used predicted protein targets as compound descriptors, we identified the most information-rich proteins which agreed with the mechanisms of action and toxicity of nonsteroidal anti-inflammatory drugs (NSAIDs), one of the most important drug classes causing DILI, stress response via TP53 and biotransformation. In addition, we identified multiple proteins involved in xenobiotic metabolism which could be novel DILI-related off-targets, such as CLK1 and DYRK2. Moreover, we derived potential structural alerts for DILI with high precision, including furan and hydrazine derivatives; however, all derived alerts were present in approved drugs and were over specific indicating the need to consider quantitative variables such as dose.

CONCLUSION

Using chemical structure-based descriptors such as structural fingerprints and predicted protein targets, DILI prediction models were built with a predictive performance comparable to previous literature. In addition, we derived insights on proteins and pathways statistically (and potentially causally) linked to DILI from these models and inferred new structural alerts related to this adverse endpoint.

Species Differences of Bile Acid Redox Metabolism: Tertiary Oxidation of Deoxycholate is Conserved in Preclinical Animals

It was recently disclosed that CYP3A is responsible for the tertiary stereoselective oxidations of deoxycholic acid (DCA), which becomes a continuum mechanism of the host-gut microbial cometabolism of bile acids (BAs) in humans. This work aims to investigate the species differences of BA redox metabolism and clarify whether the tertiary metabolism of DCA is a conserved pathway in preclinical animals. With quantitative determination of the total unconjugated BAs in urine and fecal samples of humans, dogs, rats, and mice, it was confirmed that the tertiary oxidized metabolites of DCA were found in all tested animals, whereas DCA and its oxidized metabolites disappeared in germ-free mice. The in vitro metabolism data of DCA and the other unconjugated BAs in liver microsomes of humans, monkeys, dogs, rats, and mice showed consistencies with the BA-profiling data, confirming that the tertiary oxidation of DCA is a conserved pathway. In liver microsomes of all tested animals, however, the oxidation activities toward DCA were far below the murine-specific 6β-oxidation activities toward chenodeoxycholic acid (CDCA), ursodeoxycholic acid, and lithocholic acid (LCA), and 7-oxidation activities toward murideoxycholic acid and hyodeoxycholic acid came from the 6-hydroxylation of LCA. These findings provided further explanations for why murine animals have significantly enhanced downstream metabolism of CDCA compared with humans. In conclusion, the species differences of BA redox metabolism disclosed in this work will be useful for the interspecies extrapolation of BA biology and toxicology in translational researches. SIGNIFICANCE STATEMENT: It is important to understand the species differences of bile acid metabolism when deciphering biological and hepatotoxicology findings from preclinical studies. However, the species differences of tertiary bile acids are poorly understood compared with primary and secondary bile acids. This work confirms that the tertiary oxidation of deoxycholic acid is conserved among preclinical animals and provides deeper understanding of how and why the downstream metabolism of chenodeoxycholic acid dominates that of cholic acid in murine animals compared with humans.

An In Silico Model for Predicting Drug-Induced Hepatotoxicity

As one of the leading causes of drug failure in clinical trials, drug-induced liver injury (DILI) seriously impeded the development of new drugs. Assessing the DILI risk of drug candidates in advance has been considered as an effective strategy to decrease the rate of attrition in drug discovery. Recently, there have been continuous attempts in the prediction of DILI. However, it indeed remains a huge challenge to predict DILI successfully. There is an urgent need to develop a quantitative structure-activity relationship (QSAR) model for predicting DILI with satisfactory performance. In this work, we reported a high-quality QSAR model for predicting the DILI risk of xenobiotics by incorporating the use of eight effective classifiers and molecular descriptors provided by Marvin. In model development, a large-scale and diverse dataset consisting of 1254 compounds for DILI was built through a comprehensive literature retrieval. The optimal model was attained by an ensemble method, averaging the probabilities from eight classifiers, with accuracy (ACC) of 0.783, sensitivity (SE) of 0.818, specificity (SP) of 0.748, and area under the receiver operating characteristic curve (AUC) of 0.859. For further validation, three external test sets and a large negative dataset were utilized. Consequently, both the internal and external validation indicated that our model outperformed prior studies significantly. Data provided by the current study will also be a valuable source for modeling/data mining in the future.

Biomarkers of drug-induced liver injury

Drug-induced liver injury (DILI) is a major clinical and regulatory challenge. As a result, interest in DILI biomarkers is growing. So far, considerable progress has been made in identification of biomarkers for diagnosis (acetaminophen-cysteine protein adducts), prediction (genetic biomarkers), and prognosis (microRNA-122, high mobility group box 1 protein, keratin-18, glutamate dehydrogenase, mitochondrial DNA). Many of those biomarkers also provide mechanistic insight. The purpose of this chapter is to review major advances in DILI biomarker research over the last decade, and to highlight some of the challenges involved in implementation. Although much work has been done, more liver-specific biomarkers, more DILI-specific biomarkers, and better prognostic biomarkers for survival are all still needed. Furthermore, more work is needed to define reference intervals and medical decision limits.

Drug induced liver injury with analysis of alternative causes as confounding variables

AIMS

Drug-induced liver injury (DILI) is rare compared to the worldwide frequent acute or chronic liver diseases. Therefore, patients included in series of suspected DILI are at high risk of not having DILI, whereby alternative causes may confound the DILI diagnosis. The aim of this review is to evaluate published case series of DILI for alternative causes.

METHODS

Relevant studies were identified using a computerized search of the Medline database for publications from 1993 through 30 October 2017. We used the following terms: drug hepatotoxicity, drug induced liver injury, hepatotoxic drugs combined with diagnosis, causality assessment and alternative causes.

RESULTS

Alternative causes as variables confounding the DILI diagnosis emerged in 22 published DILI case series, ranging from 4 to 47%. Among 13 335 cases of suspected DILI, alternative causes were found to be more likely in 4555 patients (34.2%), suggesting that the suspected DILI was probably not DILI. Biliary diseases such as biliary obstruction, cholangitis, choledocholithiasis, primary biliary cholangitis and primary sclerosing cholangitis were among the most missed diagnoses. Alternative causes included hepatitis B, C and E, cytomegalovirus, Epstein-Barr virus, ischemic hepatitis, cardiac hepatopathy, autoimmune hepatitis, nonalcoholic fatty liver disease, nonalcoholic steatohepatitis, and alcoholic liver disease.

CONCLUSIONS

In more than one-third of published global DILI case series, alternative causes as published in these reports confounded the DILI diagnosis. In the future, published DILI case series should include only patients with secured DILI diagnosis, preferentially established by prospective use of scored items provided by robust diagnostic algorithms such as the updated Roussel Uclaf causality assessment method.

Drug Induced Liver Injury: Can Biomarkers Assist RUCAM in Causality Assessment?

Drug induced liver injury (DILI) is a potentially serious adverse reaction in a few susceptible individuals under therapy by various drugs. Health care professionals facing DILI are confronted with a wealth of drug-unrelated liver diseases with high incidence and prevalence rates, which can confound the DILI diagnosis. Searching for alternative causes is a key element of RUCAM (Roussel Uclaf Causality Assessment Method) to assess rigorously causality in suspected DILI cases. Diagnostic biomarkers as blood tests would be a great help to clinicians, regulators, and pharmaceutical industry would be more comfortable if, in addition to RUCAM, causality of DILI can be confirmed. High specificity and sensitivity are required for any diagnostic biomarker. Although some risk factors are available to evaluate liver safety of drugs in patients, no valid diagnostic or prognostic biomarker exists currently for idiosyncratic DILI when a liver injury occurred. Identifying a biomarker in idiosyncratic DILI requires detailed knowledge of cellular and biochemical disturbances leading to apoptosis or cell necrosis and causing leakage of specific products in blood. As idiosyncratic DILI is typically a human disease and hardly reproducible in animals, pathogenetic events and resulting possible biomarkers remain largely undisclosed. Potential new diagnostic biomarkers should be evaluated in patients with DILI and RUCAM-based established causality. In conclusion, causality assessment in cases of suspected idiosyncratic DILI is still best achieved using RUCAM since specific biomarkers as diagnostic blood tests that could enhance RUCAM results are not yet available.

Patient-specific hepatocyte-like cells derived from induced pluripotent stem cells model pazopanib-mediated hepatotoxicity

Idiosyncratic drug-induced hepatotoxicity is a major cause of liver damage and drug pipeline failure, and is difficult to study as patient-specific features are not readily incorporated in traditional hepatotoxicity testing approaches using population pooled cell sources. Here we demonstrate the use of patient-specific hepatocyte-like cells (HLCs) derived from induced pluripotent stem cells for modeling idiosyncratic hepatotoxicity to pazopanib (PZ), a tyrosine kinase inhibitor drug associated with significant hepatotoxicity of unknown mechanistic basis. In vitro cytotoxicity assays confirmed that HLCs from patients with clinically identified hepatotoxicity were more sensitive to PZ-induced toxicity than other individuals, while a prototype hepatotoxin acetaminophen was similarly toxic to all HLCs studied. Transcriptional analyses showed that PZ induces oxidative stress (OS) in HLCs in general, but in HLCs from susceptible individuals, PZ causes relative disruption of iron metabolism and higher burden of OS. Our study establishes the first patient-specific HLC-based platform for idiosyncratic hepatotoxicity testing, incorporating multiple potential causative factors and permitting the correlation of transcriptomic and cellular responses to clinical phenotypes. Establishment of patient-specific HLCs with clinical phenotypes representing population variations will be valuable for pharmaceutical drug testing.

A New Structure-Activity Relationship (SAR) Model for Predicting Drug-Induced Liver Injury, Based on Statistical and Expert-Based Structural Alerts

The prompt identification of chemical molecules with potential effects on liver may help in drug discovery and in raising the levels of protection for human health. Besides in vitro approaches, computational methods in toxicology are drawing attention. We built a structure-activity relationship (SAR) model for evaluating hepatotoxicity. After compiling a data set of 950 compounds using data from the literature, we randomly split it into training (80%) and test sets (20%). We also compiled an external validation set (101 compounds) for evaluating the performance of the model. To extract structural alerts (SAs) related to hepatotoxicity and non-hepatotoxicity we used SARpy, a statistical application that automatically identifies and extracts chemical fragments related to a specific activity. We also applied the chemical grouping approach for manually identifying other SAs. We calculated accuracy, specificity, sensitivity and Matthews correlation coefficient (MCC) on the training, test and external validation sets. Considering the complexity of the endpoint, the model performed well. In the training, test and external validation sets the accuracy was respectively 81, 63, and 68%, specificity 89, 33, and 33%, sensitivity 93, 88, and 80% and MCC 0.63, 0.27, and 0.13. Since it is preferable to overestimate hepatotoxicity rather than not to recognize unsafe compounds, the model's architecture followed a conservative approach. As it was built using human data, it might be applied without any need for extrapolation from other species. This model will be freely available in the VEGA platform.

A Model to predict severity of drug-induced liver injury in humans

Drug-induced liver injury (DILI) is a major public health concern, and improving its prediction remains an unmet challenge. Recently, we reported the Rule-of-2 (RO2) and found lipophilicity (logP ≥3) and daily dose ≥100 mg of oral medications to be associated with significant risk for DILI; however, the RO2 failed to estimate grades of DILI severity. In an effort to develop a quantitative metrics, we analyzed the association of daily dose, logP, and formation of reactive metabolites (RM) in a large set of Food and Drug Administration-approved oral medications and found factoring RM into the RO2 to highly improve DILI prediction. Based on these parameters and by considering n = 354 drugs, an algorithm to assign a DILI score was developed. In univariate and multivariate logistic regression analyses the algorithm (i.e., DILI score model) defined the relative contribution of daily dose, logP, and RM and permitted a quantitative assessment of risk of clinical DILI. Furthermore, a clear relationship between calculated DILI scores and DILI risk was obtained when applied to three independent studies. The DILI score model was also functional with drug pairs defined by similar chemical structure and mode of action but divergent toxicities. Specifically, for drug pairs where the RO2 failed, the DILI score correctly identified toxic drugs. Finally, the model was applied to n = 159 clinical cases collected from the National Institutes of Health's LiverTox database to demonstrate that the DILI score correlated with the severity of clinical outcome.

CONCLUSIONS

Based on daily dose, lipophilicity, and RM, a DILI score algorithm was developed that provides a scale of assessing the severity of DILI risk in humans associated with oral medications. (Hepatology 2016;64:931-940).

Mechanism Profiling of Hepatotoxicity Caused by Oxidative Stress Using Antioxidant Response Element Reporter Gene Assay Models and Big Data

BACKGROUND

Hepatotoxicity accounts for a substantial number of drugs being withdrawn from the market. Using traditional animal models to detect hepatotoxicity is expensive and time-consuming. Alternative in vitro methods, in particular cell-based high-throughput screening (HTS) studies, have provided the research community with a large amount of data from toxicity assays. Among the various assays used to screen potential toxicants is the antioxidant response element beta lactamase reporter gene assay (ARE-bla), which identifies chemicals that have the potential to induce oxidative stress and was used to test > 10,000 compounds from the Tox21 program.

OBJECTIVE

The ARE-bla computational model and HTS data from a big data source (PubChem) were used to profile environmental and pharmaceutical compounds with hepatotoxicity data.

METHODS

Quantitative structure-activity relationship (QSAR) models were developed based on ARE-bla data. The models predicted the potential oxidative stress response for known liver toxicants when no ARE-bla data were available. Liver toxicants were used as probe compounds to search PubChem Bioassay and generate a response profile, which contained thousands of bioassays (> 10 million data points). By ranking the in vitro-in vivo correlations (IVIVCs), the most relevant bioassay(s) related to hepatotoxicity were identified.

RESULTS

The liver toxicants profile contained the ARE-bla and relevant PubChem assays. Potential toxicophores for well-known toxicants were created by identifying chemical features that existed only in compounds with high IVIVCs.

CONCLUSION

Profiling chemical IVIVCs created an opportunity to fully explore the source-to-outcome continuum of modern experimental toxicology using cheminformatics approaches and big data sources.

CITATION

Kim MT, Huang R, Sedykh A, Wang W, Xia M, Zhu H. 2016. Mechanism profiling of hepatotoxicity caused by oxidative stress using antioxidant response element reporter gene assay models and big data. Environ Health Perspect 124:634-641; http://dx.doi.org/10.1289/ehp.1509763.

QSAR Models at the US FDA/NCTR

Quantitative structure-activity relationship (QSAR) has been used in the scientific research community for many decades and applied to drug discovery and development in the industry. QSAR technologies are advancing fast and attracting possible applications in regulatory science. To facilitate the development of reliable QSAR models, the FDA had invested a lot of efforts in constructing chemical databases with a variety of efficacy and safety endpoint data, as well as in the development of computational algorithms. In this chapter, we briefly describe some of the often used databases developed at the FDA such as EDKB (Endocrine Disruptor Knowledge Base), EADB (Estrogenic Activity Database), LTKB (Liver Toxicity Knowledge Base), and CERES (Chemical Evaluation and Risk Estimation System) and the technologies adopted by the agency such as Mold(2) program for calculation of a large and diverse set of molecular descriptors and decision forest algorithm for QSAR model development. We also summarize some QSAR models that have been developed for safety evaluation of the FDA-regulated products.

Drug-induced liver injury: Interactions between drug properties and host factors

Idiosyncratic drug-induced liver injury (DILI) is a common cause for drug withdrawal from the market and although infrequent, DILI can result in serious clinical outcomes including acute liver failure and the need for liver transplantation. Eliminating the iatrogenic "harm" caused by a therapeutic intent is a priority in patient care. However, identifying culprit drugs and individuals at risk for DILI remains challenging. Apart from genetic factors predisposing individuals at risk, the role of the drugs' physicochemical and toxicological properties and their interactions with host and environmental factors need to be considered. The influence of these factors on mechanisms involved in DILI is multi-layered. In this review, we summarize current knowledge on 1) drug properties associated with hepatotoxicity, 2) host factors considered to modify an individuals' risk for DILI and clinical phenotypes, and 3) drug-host interactions. We aim at clarifying knowledge gaps needed to be filled in as to improve risk stratification in patient care. We therefore broadly discuss relevant areas of future research. Emerging insight will stimulate new investigational approaches to facilitate the discovery of clinical DILI risk modifiers in the context of disease complexity and associated interactions with drug properties, and hence will be able to move towards safety personalized medicine.

Nonsteroidal anti-inflammatory drug-induced liver injury

Nonsteroidal anti-inflammatory drugs (NSAIDs) are among the most widely used drugs in clinical practice. It is generally accepted that drug-induced liver injury (DILI) is a relatively rare adverse reaction to NSAIDs, however, DILI related to NSAIDs is of outstanding importance as the wide use of these drugs. NSAIDs are the second leading cause of DILI after antimicrobial drugs. This review presents an overview of current knowledge of NSAID-induced liver injury (N-DILI) with emphasis on the causative drugs.