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

Review article: Recommendations for detection, assessment and management of suspected drug-induced liver injury during clinical trials in oncology patients

BACKGROUND

Drug-induced liver injury (DILI) is a major concern for oncology drugs in clinical practice and under development. Monitoring cancer patients for hepatotoxicity is challenging as these patients may have abnormal liver tests pre-treatment or on-study for many reasons including liver injury due to past oncology treatments, hepatic metastases, medical co-morbidities such as heart failure, and concomitant medications. At present, there are no regulatory guidelines or position papers that systematically address best practices pertaining to DILI detection, assessment and management in oncology patients.

AIMS

The goals of this review are (1) to examine and interpret the available evidence and (2) to make recommendations for detection, monitoring, adjudication, and management of suspected hepatocellular DILI during oncology clinical trials.

METHODS

This manuscript was developed by the IQ Consortium (International Consortium for Innovation and Quality in pharmaceutical development) DILI Initiative that consists of members from 17 pharmaceutical companies, in collaboration with academic and regulatory DILI experts. The manuscript is based on extensive literature review, expert interpretation of the literature, and several rounds of consensus discussions.

RESULTS

This review highlights recommendations for patient eligibility for clinical trials with or without primary/metastatic liver involvement, as well as changes in liver tests that should trigger increased monitoring and/or discontinuation of study drug. Guidance regarding causality assessment for suspected DILI events, rechallenge and dose-modification is provided.

CONCLUSIONS

This review brings together evidence-based recommendations and expert opinion to provide the first dedicated consensus for best practices in detection, assessment, and management of DILI in oncology clinical trials.

Improved Detection of Drug-Induced Liver Injury by Integrating Predicted In Vivo and In Vitro Data

Drug-induced liver injury (DILI) has been a significant challenge in drug discovery, often leading to clinical trial failures and necessitating drug withdrawals. Over the last decade, the existing suite of in vitro proxy-DILI assays has generally improved at identifying compounds with hepatotoxicity. However, there is considerable interest in enhancing the in silico prediction of DILI because it allows for evaluating large sets of compounds more quickly and cost-effectively, particularly in the early stages of projects. In this study, we aim to study ML models for DILI prediction that first predict nine proxy-DILI labels and then use them as features in addition to chemical structural features to predict DILI. The features include in vitro (e.g., mitochondrial toxicity, bile salt export pump inhibition) data, in vivo (e.g., preclinical rat hepatotoxicity studies) data, pharmacokinetic parameters of maximum concentration, structural fingerprints, and physicochemical parameters. We trained DILI-prediction models on 888 compounds from the DILI data set (composed of DILIst and DILIrank) and tested them on a held-out external test set of 223 compounds from the DILI data set. The best model, DILIPredictor, attained an AUC-PR of 0.79. This model enabled the detection of the top 25 toxic compounds (2.68 LR+, positive likelihood ratio) compared to models using only structural features (1.65 LR+ score). Using feature interpretation from DILIPredictor, we identified the chemical substructures causing DILI and differentiated cases of DILI caused by compounds in animals but not in humans. For example, DILIPredictor correctly recognized 2-butoxyethanol as nontoxic in humans despite its hepatotoxicity in mice models. Overall, the DILIPredictor model improves the detection of compounds causing DILI with an improved differentiation between animal and human sensitivity and the potential for mechanism evaluation. DILIPredictor required only chemical structures as input for prediction and is publicly available at https://broad.io/DILIPredictor for use via web interface and with all code available for download.

A disproportionality analysis of CDK4/6 inhibitors in the FDA Adverse Event Reporting System (FAERS)

OBJECTIVE

The FDA Adverse Event Reporting System (FAERS) was used to mine and evaluate adverse events (AEs) associated with cyclin-dependent kinase (CDK) 4/6 inhibitors, thereby providing a reference for clinical rational drug use.

METHODS

AE data related to CDK4/6 inhibitors from the first quarter of 2015 to the first quarter of 2023 were acquired from FAERS, while the signal mining was processed using the reporting odds ratio (ROR) method and Bayesian confidence propagation neural network (BCPNN) method.

RESULTS

The number of AE reports for CDK4/6 inhibitors was, respectively, 132,494 for palbociclib, 56,151 for ribociclib, and 7,014 for abemaciclib. The corresponding numbers of AE signals were 319, 517, and 59, with the number of involved System Organ Class (SOC) being 23, 23, and 15, mainly involving blood and lymphatic system disorders, respiratory, thoracic and mediastinal disorders, hepatobiliary disorders, skin and subcutaneous tissue disorders, etc.

CONCLUSION

CDK4/6 inhibitors could lead to pulmonary toxicity, myelosuppression, skin reactions, etc. Special attention should be paid to abemaciclib for interstitial lung disease (ILD), erythema multiforme, and thrombosis risk; ribociclib for cardiac toxicity, hepatotoxicity, and musculoskeletal toxicity; palbociclib for neurocognitive impairment and osteonecrosis of the jaw.

Drug interaction with UDP-Glucuronosyltransferase (UGT) enzymes is a predictor of drug-induced liver injury

BACKGROUND AND AIMS

DILI frequently contributes to the attrition of new drug candidates and is a common cause for the withdrawal of approved drugs from the market. Although some noncytochrome P450 (non-CYP) metabolism enzymes have been implicated in DILI development, their association with DILI outcomes has not been systematically evaluated.

APPROACH AND RESULTS

In this study, we analyzed a large data set comprising 317 drugs and their interactions in vitro with 42 non-CYP enzymes as substrates, inducers, and/or inhibitors retrieved from historical regulatory documents using multivariate logistic regression. We examined how these in vitro drug-enzyme interactions are correlated with the drugs' potential for DILI concern, as classified in the Liver Toxicity Knowledge Base database. Our study revealed that drugs that inhibit non-CYP enzymes are significantly associated with high DILI concern. Particularly, interaction with UDP-glucuronosyltransferases (UGT) enzymes is an important predictor of DILI outcomes. Further analysis indicated that only pure UGT inhibitors and dual substrate inhibitors, but not pure UGT substrates, are significantly associated with high DILI concern.

CONCLUSIONS

Drug interactions with UGT enzymes may independently predict DILI, and their combined use with the rule-of-two model further improves overall predictive performance. These findings could expand the currently available tools for assessing the potential for DILI in humans.

Improved Detection of Drug-Induced Liver Injury by Integrating Predicted in vivo and in vitro Data

Drug-induced liver injury (DILI) has been significant challenge in drug discovery, often leading to clinical trial failures and necessitating drug withdrawals. The existing suite of in vitro proxy-DILI assays is generally effective at identifying compounds with hepatotoxicity. However, there is considerable interest in enhancing in silico prediction of DILI because it allows for the evaluation of large sets of compounds more quickly and cost-effectively, particularly in the early stages of projects. In this study, we aim to study ML models for DILI prediction that first predicts nine proxy-DILI labels and then uses them as features in addition to chemical structural features to predict DILI. The features include in vitro (e.g., mitochondrial toxicity, bile salt export pump inhibition) data, in vivo (e.g., preclinical rat hepatotoxicity studies) data, pharmacokinetic parameters of maximum concentration, structural fingerprints, and physicochemical parameters. We trained DILI-prediction models on 888 compounds from the DILIst dataset and tested on a held-out external test set of 223 compounds from DILIst dataset. The best model, DILIPredictor, attained an AUC-ROC of 0.79. This model enabled the detection of top 25 toxic compounds compared to models using only structural features (2.68 LR+ score). Using feature interpretation from DILIPredictor, we were able to identify the chemical substructures causing DILI as well as differentiate cases DILI is caused by compounds in animals but not in humans. For example, DILIPredictor correctly recognized 2-butoxyethanol as non-toxic in humans despite its hepatotoxicity in mice models. Overall, the DILIPredictor model improves the detection of compounds causing DILI with an improved differentiation between animal and human sensitivity as well as the potential for mechanism evaluation. DILIPredictor is publicly available at https://broad.io/DILIPredictor for use via web interface and with all code available for download and local implementation via https://pypi.org/project/dilipred/.

Novel clinical phenotypes, drug categorization, and outcome prediction in drug-induced cholestasis: Analysis of a database of 432 patients developed by literature review and machine learning support

BACKGROUND

Serum transaminases, alkaline phosphatase and bilirubin are common parameters used for DILI diagnosis, classification, and prognosis. However, the relevance of clinical examination, histopathology and drug chemical properties have not been fully investigated. As cholestasis is a frequent and complex DILI manifestation, our goal was to investigate the relevance of clinical features and drug properties to stratify drug-induced cholestasis (DIC) patients, and to develop a prognosis model to identify patients at risk and high-concern drugs.

METHODS

DIC-related articles were searched by keywords and Boolean operators in seven databases. Relevant articles were uploaded onto Sysrev, a machine-learning based platform for article review and data extraction. Demographic, clinical, biochemical, and liver histopathological data were collected. Drug properties were obtained from databases or QSAR modelling. Statistical analyses and logistic regressions were performed.

RESULTS

Data from 432 DIC patients associated with 52 drugs were collected. Fibrosis strongly associated with fatality, whereas canalicular paucity and ALP associated with chronicity. Drugs causing cholestasis clustered in three major groups. The pure cholestatic pattern divided into two subphenotypes with differences in prognosis, canalicular paucity, fibrosis, ALP and bilirubin. A predictive model of DIC outcome based on non-invasive parameters and drug properties was developed. Results demonstrate that physicochemical (pKa-a) and pharmacokinetic (bioavailability, CYP2C9) attributes impinged on the DIC phenotype and allowed the identification of high-concern drugs.

CONCLUSIONS

We identified novel associations among DIC manifestations and disclosed novel DIC subphenotypes with specific clinical and chemical traits. The developed predictive DIC outcome model could facilitate DIC prognosis in clinical practice and drug categorization.

CDK4/6 inhibitors in drug-induced liver injury: a pharmacovigilance study of the FAERS database and analysis of the drug-gene interaction network

Objective

The aim of this study was to investigate the potential risk of drug-induced liver injury (DILI) caused by the CDK4/6 inhibitors (CDK4/6is abemaciclib, ribociclib, and palbociclib by comprehensively analyzing the FDA Adverse Event Reporting System (FAERS) database. Moreover, potential toxicological mechanisms of CDK4/6is-related liver injury were explored via drug-gene network analysis.

Methods

In this retrospective observational study, we collected reports of DILI associated with CDK4/6i use from the FAERS dated January 2014 to March 2023. We conducted disproportionality analyses using the reporting odds ratio (ROR) with a 95% confidence interval (CI). Pathway enrichment analysis and drug-gene network analyses were subsequently performed to determine the potential mechanisms underlying CDK4/6i-induced liver injury.

Results

We found positive signals for DILI with ribociclib (ROR = 2.60) and abemaciclib (ROR = 2.37). DILIs associated with liver-related investigations, signs, and symptoms were confirmed in all three reports of CDK4/6is. Moreover, ascites was identified as an unlisted hepatic adverse effect of palbociclib. We isolated 189 interactive target genes linking CDK4/6 inhibitors to hepatic injury. Several key genes, such as STAT3, HSP90AA1, and EP300, were revealed via protein-protein analysis, emphasizing their central roles within the network. KEGG pathway enrichment of these genes highlighted multiple pathways.

Conclusion

Our study revealed variations in hepatobiliary toxicity among the different CDK4/6 inhibitors, with ribociclib showing the highest risk of liver injury, followed by abemaciclib, while palbociclib appeared relatively safe. Our findings emphasize the need for cautious use of CDK4/6 inhibitors, and regular liver function monitoring is recommended for long-term CDK4/6 inhibitor use.

A comprehensive analysis of liver safety across zibotentan oncology trials: knowledge of the past offers new perspectives on the present

BACKGROUND

Endothelin receptor antagonists (ERAs) are associated with liver injury. We used data from previous oncology clinical trials to determine the liver safety profile of zibotentan, which is currently in clinical development (in combination with dapagliflozin) for chronic kidney disease and cirrhosis.

RESEARCH DESIGN AND METHODS

Six global, double-blinded, phase 2b and 3 clinical trials from the zibotentan oncology development program were pooled to analyze liver safety. Descriptive statistics, proportion of liver-related adverse events, liver biochemistry parameter elevation, and shifts from baseline were analyzed, with individual case assessment.

RESULTS

A total of 1532 patients received zibotentan for 285 days (mean), and 1486 patients received placebo for 320 days (mean). The frequency of any hepatic disorder preferred term was similar across zibotentan monotherapy (22/947 patients, 2.3%) and placebo monotherapy arms (30/881 patients, 3.4%). A total of 4 (0.4%) patients receiving zibotentan monotherapy experienced ALT elevations >5× ULN versus 8 (0.9%) receiving placebo. Of the seven patients receiving zibotentan who met criteria for potential Hy's Law, there were no cases of drug-induced liver injury.

CONCLUSIONS

We found no evidence of zibotentan-related liver biochemistry changes among cancer-treated patients, suggesting that hepatotoxicity of ERAs is molecule-dependent, and allowing exploration of zibotentan for new indications.

In silico modeling-based new alternative methods to predict drug and herb-induced liver injury: A review

New approach methods (NAMs) have been developed to predict a wide range of toxicities through innovative technologies. Liver injury is one of the most extensively studied endpoints due to its severity and frequency, occurring among populations that consume drugs or dietary supplements. In this review, we focus on recent developments of in silico modeling for liver injury prediction using deep learning and in vitro data based on adverse outcome pathways (AOPs). Despite these models being mainly developed using datasets generated from drug-like molecules, they were also applied to the prediction of hepatotoxicity caused by herbal products. As deep learning has achieved great success in many different fields, advanced machine learning algorithms have been actively applied to improve the accuracy of in silico models. Additionally, the development of liver AOPs, combined with big data in toxicology, has been valuable in developing in silico models with enhanced predictive performance and interpretability. Specifically, one approach involves developing structure-based models for predicting molecular initiating events of liver AOPs, while others use in vitro data with structure information as model inputs for making predictions. Even though liver injury remains a difficult endpoint to predict, advancements in machine learning algorithms and the expansion of in vitro databases with relevant biological knowledge have made a huge impact on improving in silico modeling for drug-induced liver injury prediction.

Yin/Yang associated differential responses to Psoralea corylifolia Linn. In rat models: an integrated metabolomics and transcriptomics study

ETHNOPHARMACOLOGICAL RELEVANCE

Psoralea corylifolia Linn. (BGZ) is a commonly used traditional Chinese medicine (TCM) for the treatment of kidney-yang deficiency syndrome (Yangsyn) with good curative effect and security. However, BGZ was also reported to induce liver injury in recent years. According to TCM theory, taking BGZ may induce a series of adverse reactions in patients with kidney-yin deficiency syndrome (Yinsyn), which suggests that BGZ-induced liver damage may be related to its unreasonable clinical use.

AIM OF THE STUDY

Liver injury caused by TCM is a rare but potentially serious adverse drug reaction, and the identification of predisposed individuals for drug-induced liver injury (DILI) remains challenging. The study aimed to investigate the differential responses to BGZ in Yangsyn and Yinsyn rat models and identify the corresponding characteristic biomarkers.

MATERIALS AND METHODS

The corresponding animal models of Yangsyn and Yinsyn were induced by hydrocortisone and thyroxine + reserpine respectively. Body weight, organ index, serum biochemistry, and Hematoxylin and Eosin (HE) staining were used to evaluate the liver toxicity effect of BGZ on rats with Yangsyn and Yinsyn. Transcriptomics and metabonomics were used to screen the representative biomarkers (including metabolites and differentially expressed genes (DEGs)) changed by BGZ in Yangsyn and Yinsyn rats, respectively.

RESULTS

The level changes of liver organ index, alanine aminotransferase (ALT), and aspartate aminotransferase (AST), suggested that BGZ has liver-protective and liver-damaging effects on Yangsyn and Yinsyn rats, respectively, and the results also were confirmed by the pathological changes of liver tissue. The results showed that 102 DEGs and 27 metabolites were significantly regulated related to BGZ's protective effect on Yangsyn, which is mainly associated with the glycerophospholipid metabolism, arachidonic acid metabolism, pantothenate, and coenzyme A (CoA) biosynthesis pathways. While 28 DEGs and 31 metabolites, related to the pathway of pantothenate and CoA biosynthesis, were significantly regulated for the BGZ-induced liver injury in Yinsyn. Furthermore, 4 DEGs (aldehyde dehydrogenase 1 family member B1 (Aldh1b1), solute carrier family 25 member 25 (Slc25a25), Pim-3 proto-oncogene, serine/threonine kinase (Pim3), out at first homolog (Oaf)) and 4 metabolites (phosphatidate, phosphatidylcholine, N-Acetylleucine, biliverdin) in the Yangsyn group and 1 DEG [galectin 5 (Lgals5)] and 1 metabolite (5-amino-1-(5-phospho-D-ribosyl)imidazole-4-carboxylate) in Yinsyn group were significantly correlated to the ALT and AST levels of BGZ treated and untreated groups (receiver operating characteristic (ROC) ≥ 0.9).

CONCLUSIONS

Yinsyn and Yangsyn are the predisposed syndromes for BGZ to exert liver damage and liver protection respectively, which are mainly related to the regulation of amino acid metabolism, lipid metabolism, energy metabolism, and metabolism of cofactors and vitamins. The results further suggest that attention should be paid to the selection of predisposed populations when using drugs related to the regulation of energy metabolism, and the Yinsyn/Yangsyn animal models based on the theory of TCM syndromes may be a feasible method for identifying the susceptible population to receive TCM.

Drug-induced liver injury: a comprehensive review

Drug-induced liver injury (DILI) remains a challenge in clinical practice and is still a diagnosis of exclusion. Although it has a low incidence amongst the general population, DILI accounts for most cases of acute liver failure with a fatality rate of up to 50%. While multiple mechanisms of DILI have been postulated, there is no clear causal relationship between drugs, risk factors and mechanisms of DILI. Current best practice relies on a combination of high clinical suspicion, thorough clinical history of risk factors and timeline, and extensive hepatological investigations as supported by the international Roussel Uclaf Causality Assessment Method criteria, the latter considered a key diagnostic algorithm for DILI. This review focuses on DILI classification, risk factors, clinical evaluation, future biomarkers and management, with the aim of facilitating physicians to correctly identify DILI early in presentation.

Clinical Pattern of Tolvaptan-Associated Liver Injury in Trial Participants With Autosomal Dominant Polycystic Kidney Disease (ADPKD): An Analysis of Pivotal Clinical Trials

RATIONALE & OBJECTIVE

Tolvaptan is associated with risk of drug-induced liver injury when used to treat autosomal dominant polycystic kidney disease (ADPKD). After this risk was described based on the clinical trials TEMPO 3:4 and TEMPO 4:4, additional data from the REPRISE trial and a long-term extension of TEMPO 4:4, REPRISE, and other tolvaptan trials in ADPKD have become available. To further characterize the hepatic safety profile of tolvaptan, an analysis of the expanded dataset was conducted.

STUDY DESIGN

Analysis of safety data from prospective clinical trials of tolvaptan.

SETTING & PARTICIPANTS

Multicenter clinical trials including more than 2,900 tolvaptan-treated participants, more than 2,300 with at least 18 months of drug exposure.

INTERVENTION

Tolvaptan administered twice daily in split-dose regimens.

OUTCOMES

Frequency of liver enzyme level increases detected by regular laboratory monitoring.

RESULTS

In the placebo-controlled REPRISE trial, more tolvaptan- than placebo-treated participants (38 of 681 [5.6%] vs 8 of 685 [1.2%]) experienced alanine aminotransferase level increases to >3× the upper limit of normal (ULN), similar to TEMPO 3:4 (40 of 957 [4.4%] vs 5 of 484 [1.0%]). No participant in REPRISE or the long-term extension experienced concurrent alanine aminotransferase level increases to >3× ULN and total bilirubin increases to >2× ULN ("Hy's Law" laboratory criteria). Based on the expanded dataset, liver enzyme increases most often occurred within 18 months after tolvaptan initiation and were less frequent thereafter. Increased levels returned to normal or near normal after treatment interruption or discontinuation. Thirty-eight patients were rechallenged with tolvaptan after the initial drug-induced liver injury episode, with return of liver enzyme level increases in 30; 1 additional participant showed a clinical "adaptation" after the initial episode, with resolution of the enzyme level increases despite continuation of tolvaptan.

LIMITATIONS

Retrospective analysis.

CONCLUSIONS

The absence of Hy's Law cases in REPRISE and the long-term extension trial support monthly liver enzyme monitoring during the first 18 months of tolvaptan exposure and every 3 months thereafter to detect and manage enzyme level increases, as is recommended on the drug label.

FUNDING

Otsuka Pharmaceutical Development & Commercialization, Inc.

TRIAL REGISTRATION

Trials included in the dataset were registered at ClinicalTrials.gov with study numbers NCT00428948 (TEMPO 3:4), NCT01214421 (TEMPO 4:4), NCT02160145 (REPRISE), and NCT02251275 (long-term extension).

Estimating drug-induced liver injury risk by in vitro molecular initiation response and pharmacokinetic parameters for during early drug development

Drug-induced liver injury (DILI) is a major factor influencing new drug withdrawal; therefore, an appropriate toxicity assessment at the preclinical stage is required. Previous in silico models have been established using compound information listed in large data sources, thereby limiting the DILI risk prediction for new drugs. Herein, we first constructed a model to predict DILI risk based on a molecular initiating event (MIE) predicted by quantitative structure-activity relationships, admetSAR parameters (e.g. cytochrome P450 reactivity, plasma protein binding, and water-solubility), and clinical information (maximum daily dose [MDD] and reactive metabolite [RM]) for 186 compounds. The accuracy of the models using MIE, MDD, RM, and admetSAR alone were 43.2%, 47.3%, 77.0%, and 68.9%, while the "predicted MIE + admetSAR + MDD + RM" model's accuracy was 75.7%. The contribution of MIE to the overall prediction accuracy was little effect or rather worsening it. However, it was considered that MIE was a valuable parameter and that it contributed to detect high DILI risk compounds in the early development stage. We next examined the effect of stepwise changes in MDD on altering the DILI risk and estimating the maximum safety dose (MSD) for clinical use based on structural information, admetSAR, and MIE parameters because it is important to estimate the dose that could prevent the DILI onset in clinical conditions. Low-MSD compounds might increase the DILI risk, as these compounds were classified as "most-DILI concern" at low doses. In conclusion, MIE parameters were especially useful to check the DILI concern compounds and to prevent the underestimation of DILI risk in the early stage of drug development.

Identification of intrinsic hepatotoxic compounds in Polygonum multiflorum Thunb. using machine-learning methods

ETHNOPHARMACOLOGICAL RELEVANCE

Polygonum multiflorum Thunb. (PM) is a herb, extracts of which have been used as Chinese medicine for years. Although it is believed to be beneficial to the liver, heart, and kidneys, it causes idiosyncratic drug-induced liver injury (DILI).

AIM OF THE STUDY

We propose that the intrinsic DILI caused by natural products in PM (NPPM) is an important complementary mechanism to PM-related herb-induced liver injury, and aim to identify the ingredients with high DILI potential by machine learning methods.

MATERIALS AND METHODS

One hundred and ninety-seven NPPM were collected from the literature to identify the intrinsic hepatotoxic compounds. Additionally, a DILI-labeled dataset consisting of 2384 compounds was collected and randomly split into training and test sets. A diparametric optimization method was developed to tune the parameters of extended-connectivity fingerprints (ECFPs), Rdkit, and atom-pair fingerprints as well as those of machine-learning (ML) algorithms. Subsequently, K means were employed to cluster the NPPM that were predicted to have a high DILI risk. An in vitro cell-viability assay was performed using HepaRG cells to validate the prediction results.

RESULTS

ECFPs with the top 35% of features ranked by the F-value with support vector machine (SVM) yielded the best performance. The optimized SVM model achieved an accuracy of 0.761 and recall value of 0.834 on the test dataset. The silico screening for NPPM resulted in 47 ingredients with high DILI potential, which were clustered into six groups based on the elbow method. A representative subgroup that contained 21 ingredients, of which two dianthrones exhibited the lowest IC₅₀ value (0.7-0.9 μM) and anthraquinones showed moderate toxicity (15-25 μM), was constructed.

CONCLUSION

Using ML methods and in vitro screening, two classes of compounds, dianthrones and anthraquinones, were predicted and validated to have a high risk of DILI. The diparametric optimization method used in this study could provide a useful and powerful tool to screen toxicants for large datasets and is available at https://github.com/dreadlesss/Hepatotoxicity_predictor.

Combined Risk Assessment of Food-derived Coumarin with in Silico Approaches

Hepatotoxicity associated with food-derived coumarin occurs occasionally in humans. We have, herein, assessed the data of existing clinical and nonclinical studies as well as those of in silico models for humans in order to shed more light on this association. The average intakes of food-derived coumarin are estimated to be 1-3 mg/day, while a ten-times higher level is expected in the worst-case scenarios. These levels are close to or above the tolerable daily intake suggested by a chronic study in dogs. The human internal exposure levels were estimated by a physiologically-based pharmacokinetic model with the use of virtual doses of coumarin in the amounts expected to derive from foods. Our results suggest that: (i) coumarin can be cleared rapidly via 7-hydroxylation in humans, and (ii) the plasma levels of coumarin and of its metabolite, o-hydroxyphenylacetic acid associated with hepatotoxicity, are considerably lower than those yielding hepatotoxicity in rats. Pharmacokinetic data suggest a low or negligible concern regarding a coumarin-induced hepatotoxicity in humans exposed to an average intake from foods. Detoxification of coumarin through the 7-hydroxylation, however, might vary among individuals due to genetic polymorphisms in CYP2A6 enzyme. In addition, the CYP1A2- and CYP2E1-mediated activation of coumarin can fluctuate as a result of induction caused by environmental factors. Furthermore, the daily consumption of food-contained coumarin was implicated in the potential risk of hepatotoxicity by the drug-induced liver injury score model developed by the US Food and Drug Administration. These results support the idea of the existence of human subpopulations that are highly sensitive to coumarin; therefore, a more precise risk assessment is needed. The present study also highlights the usefulness of in silico approaches of pharmacokinetics with the liver injury score model as battery components of a risk assessment.

A logistic regression model based on inpatient health records to predict drug-induced liver injury caused by ramipril—An angiotensin-converting enzyme inhibitor

Drug-induced liver injury (DILI) is a rare side effect of angiotensin-converting enzyme inhibitors (ACEIs). Ramipril is a widely used ACE compound because of its effectiveness in the treatment of hypertension and heart failure, as well as its low risk of adverse effects. However, the clinical features of ramipril, and the risk of DILI, have not been adequately studied. A retrospective cohort study was performed based on data from 3909 inpatients to compare the risk of DILI conferred by ramipril and other ACEIs. A logistic regression model was then constructed and validated against data from 1686 patients using ramipril, of which 117 patients were diagnosed with DILI. The use of ramipril increased the risk of DILI by 2.68 times (odds ratio = 2.68; 95% confident interval (CI):1.96–3.71) compared with the group using other ACEIs. The clinical features of DILI in the ramipril group were similar to those from the ACEI group (P>0.05), except that the ALT level was higher (P<0.05). A logistic regression model including Body mass index (BMI), comorbidity, liver disease, daily dose, alanine aminotransferase (ALT), and alkaline phosphatase (ALP) was built and successfully validated for DILI risk prediction, with the area under the receiver operating characteristic curve of the model of 0.82 (95% CI: 0.752–0.888). We recommend that clinicians should be aware of the levels of ALT and ALP as well as BMI, comorbidities, and liver disease before prescribing ramipril to avoid the risk of DILI in patients.

Incidence and risk factors of drug-induced liver injury

The epidemiology and aetiology of drug-induced liver injury (DILI) vary across different countries and populations. Overall, DILI is rare in the general population but has become more prevalent in hospitalized patients, especially among patients with unexplained liver conditions. In addition, drugs implicated in DILI differ between Western and Eastern countries. Antibiotics are the leading drugs implicated in DILI in the West, whereas traditional Chinese medicine is the primary cause implicated in DILI in the East. The incidence of herbal and dietary supplements-induced hepatotoxicity is increasing globally. Several genetic and nongenetic risk factors associated with DILI have been described in the literature; however, there are no confirmed risk factors for all-cause DILI. Some factors may contribute to the risk of DILI in a drug-specific manner.

Pharmacokinetic and toxicodynamic concepts in idiosyncratic, drug-induced liver injury

INTRODUCTION

Idiosyncratic drug-induced liver injury (IDILI) causes morbidity and mortality in patients and leads to curtailed use of efficacious pharmaceuticals. Unlike intrinsically toxic reactions, which depend on dose, IDILI occurs in a minority of patients at therapeutic doses. Much remains unknown about causal links among drug exposure, a mode of action, and liver injury. Consequently, numerous hypotheses about IDILI pathogenesis have arisen.

AREAS COVERED

Pharmacokinetic and toxicodynamic characteristics underlying current hypotheses of IDILI etiology are discussed and illustrated graphically.

EXPERT OPINION

Hypotheses to explain IDILI etiology all involve alterations in pharmacokinetics, which lead to plasma drug concentrations that rise above a threshold for toxicity, or in toxicodynamics, which result in a lowering of the toxicity threshold. Altered pharmacokinetics arise, for example, from changes in drug metabolism or from transporter polymorphisms. A lowered toxicity threshold can arise from drug-induced mitochondrial injury, accumulation of toxic endogenous factors or harmful immune responses. Newly developed, interactive freeware (DemoTox-PK; https://bit.ly/DemoTox-PK) allows the user to visualize how such alterations might lead to a toxic reaction. The illustrations presented provide a framework for conceptualizing idiosyncratic reactions and could serve as a stimulus for future discussion, education, and research into modes of action of IDILI.

Liver Injury with Nintedanib: A Pharmacovigilance-Pharmacokinetic Appraisal

Drug-induced liver injury (DILI) with nintedanib has emerged as an adverse event of special interest in premarketing clinical trials. We characterized DILI with nintedanib in the real world and explored the underlying pharmacological basis. First, we assessed serious hepatic events reported to the Food and Drug Administration’s Adverse Event Reporting System by combining the disproportionality approach [reporting odds ratio (ROR) with 95% confidence interval (CI)] with individual case assessment. Demographic and clinical features were inspected (seriousness, onset, discontinuation, dechallenge/rechallenge, concomitant drugs) to implement an ad hoc causality assessment scoring system. Second, we appraised physiochemical and pharmacokinetic parameters possibly predictive of DILI occurrence. Significant disproportionality was found for nintedanib as compared to pirfenidone (N = 91; ROR = 4.77; 95% CI = 3.15–7.39). Asian population, low body weight (59 kg), and rapid DILI onset (13.5 days) emerged as clinical features. Hospitalization and discontinuation were found in a significant proportion of cases (32% and 36%, respectively). In 24% of the cases, at least two potentially hepatotoxic drugs (statins, proton pump inhibitors, antibiotics) were recorded. Causality was at least possible in 92.3% of the cases. High lipophilicity and predicted in silico inhibition of liver transporters emerged as potential pharmacokinetic features supporting the biological plausibility. Although causality cannot be demonstrated, clinicians should consider early monitoring and medication review on a case-by-case basis.

Study on the Characteristics of Small-Molecule Kinase Inhibitors-Related Drug-Induced Liver Injury

Background and Aim: More than half of the small-molecule kinase inhibitors (KIs) induced liver injury clinically. Meanwhile, studies have shown a close relationship between mitochondrial damage and drug-induced liver injury (DILI). We aimed to study KIs and the binding between drugs and mitochondrial proteins to find factors related to DILI occurrence.

Methods: A total of 1,223 oral FDA-approved drugs were collected and analyzed, including 44 KIs. Fisher’s exact test was used to analyze DILI potential and risk of different factors. A total of 187 human mitochondrial proteins were further collected, and high-throughput molecular docking was performed between human mitochondrial proteins and drugs in the data set. The molecular dynamics simulation was used to optimize and evaluate the dynamic binding behavior of the selected mitochondrial protein/KI complexes.

Results: The possibility of KIs to produce DILI is much higher than that of other types (OR = 46.89, p = 9.28E-13). A few DILI risk factors were identified, including molecular weight (MW) between 400 and 600, the defined daily dose (DDD) ≥ 100 mg/day, the octanol–water partition coefficient (LogP) ≥ 3, and the degree of liver metabolism (LM) more than 50%. Drugs that met this combination of rules were found to have a higher DILI risk than controls (OR = 8.28, p = 4.82E-05) and were more likely to cause severe DILI (OR = 8.26, p = 5.06E-04). The docking results showed that KIs had a significant higher affinity with human mitochondrial proteins (p = 4.19E-11) than other drug types. Furthermore, the five proteins with the lowest docking score were selected for molecular dynamics simulation, and the smallest fluctuation of the backbone RMSD curve was found in the protein 5FS8/KI complexes, which indicated the best stability of the protein 5FS8 bound to KIs.

Conclusions: KIs were found to have the highest odds ratio of causing DILI. MW was significantly related to the production of DILI, and the average docking scores of KI drugs were found to be significantly different from other classes. Further analysis identified the top binding mitochondrial proteins for KIs, and specific binding sites were analyzed. The optimization of molecular docking results by molecular dynamics simulation may contribute to further studying the mechanism of DILI.

Clinical characteristics of antiepileptic-induced liver injury in patients from the DILIN prospective study

BACKGROUND & AIMS

Antiepileptic drugs (AEDs) are a common cause of drug-induced liver injury (DILI). Over the last few decades, several newer AEDs were approved for marketing in the United States, and they are increasingly prescribed for indications other than seizures. Contemporaneous data related to trends and characteristics of AED-related liver injury are sparse.

METHODS

We report the trends, characteristics, and outcomes of patients with AED-related DILI enrolled into the DILIN Prospective Study between 2004 and 2020.

RESULTS

Among 1,711 participants with definite, highly likely, or probable DILI, 66 (3.9%) had AED-related DILI (lamotrigine [n = 18], phenytoin [n = 16], carbamazepine [n = 11], valproate [n = 10], gabapentin [n = 4], and others [n = 7]). The frequency of AED-related liver injury significantly decreased during the study period (from 8.5% of cases during 2004-2007 to 2.6% during 2015-2020, p = 0.01). AEDs other than phenytoin were commonly prescribed for non-seizure indications. Compared to non-AEDs, patients with AED-related liver injury were younger (mean age 38.5 vs. 50.1 years-old, p <0.001) and more likely African American (27% vs. 12%, p = 0.008). DRESS was common with liver injury caused by lamotrigine, phenytoin, and carbamazepine, but not valproate or gabapentin. Liver injury severity was moderate to severe in the majority: 5 died, and 3 underwent orthotopic liver transplantation (OLT). No patient with lamotrigine-related DILI, including 13 with hepatocellular jaundice, died or needed OLT, while 3 out of 16 patients (19%) with phenytoin-related DILI either died or required OLT.

CONCLUSION

The frequency of AED-related liver injury significantly decreased over the last 2 decades in our experience. AED-related liver injury has several distinctive features, including a preponderance in African American patients and those with immunoallergic skin reactions, with outcomes depending on the type of AED involved.

LAY SUMMARY

Medications used to treat epilepsy may sometimes cause severe liver injury. However, several new medications have been approved over the last 2 decades and they may not be as toxic to the liver as older antiepileptic medications (AEDs). This study shows that overall liver injury due to AEDs is decreasing, likely due to decreasing use of older AEDs. Liver injury due to AEDs appears to be more common in African Americans and is commonly associated with allergic skin reactions.

In silico approaches in organ toxicity hazard assessment: current status and future needs in predicting liver toxicity

Hepatotoxicity is one of the most frequently observed adverse effects resulting from exposure to a xenobiotic. For example, in pharmaceutical research and development it is one of the major reasons for drug withdrawals, clinical failures, and discontinuation of drug candidates. The development of faster and cheaper methods to assess hepatotoxicity that are both more sustainable and more informative is critically needed. The biological mechanisms and processes underpinning hepatotoxicity are summarized and experimental approaches to support the prediction of hepatotoxicity are described, including toxicokinetic considerations. The paper describes the increasingly important role of in silico approaches and highlights challenges to the adoption of these methods including the lack of a commonly agreed upon protocol for performing such an assessment and the need for in silico solutions that take dose into consideration. A proposed framework for the integration of in silico and experimental information is provided along with a case study describing how computational methods have been used to successfully respond to a regulatory question concerning non-genotoxic impurities in chemically synthesized pharmaceuticals.

Liver Injury with Ulipristal Acetate: Exploring the Underlying Pharmacological Basis

Introduction

The European Medicines Agency has suspended the use of ulipristal acetate (UPA) in the treatment of uterine fibroids and is reassessing its association with a risk of liver injury.

Objectives

Our objectives were to characterize the post-marketing reporting of drug-induced liver injury (DILI) with UPA and investigate the underlying pharmacological basis.

Methods

We queried the worldwide FDA Adverse Event Reporting System and performed a disproportionality analysis, selecting only hepatic designated medical events (DMEs) where UPA was reported as suspect. The reporting odds ratios (RORs) were calculated, and we considered a lower limit of the 95% confidence interval (LL95% CI) > 1 as significant. Physiochemical/pharmacokinetic features were extracted to assess the risk of hepatotoxicity by applying predictive DILI risk models. Mifepristone and leuprolide were selected as comparators.

Results

A significantly higher proportion of liver disorders was reported for UPA than for mifepristone (2.9 vs. 0.8%; p < 0.00001) and leuprolide (2.9 vs. 1.6%; p = 0.015). As regards hepatic DMEs, statistically significant RORs were found for autoimmune hepatitis (N = 5; LL95% CI 16.8), DILI (n = 5; LL95% CI 5.9), and acute hepatic failure (N = 5; LL95% CI 9.3). No signals of DILI emerged for mifepristone and leuprolide acetate. UPA and mifepristone showed high lipophilicity and hepatic metabolism (predicted intermediate DILI risk). Leuprolide exhibited contrasting features, resulting in no DILI concern. Inhibition of different liver transporters and the presence of a reactive metabolite were also recognised for UPA.

Conclusion

Different drug properties previously linked to the occurrence of DILI may partially explain the reporting pattern observed with UPA. Our “bedside-to-bench” approach may support regulators in the risk–benefit assessment of UPA.

Electronic supplementary material

The online version of this article (10.1007/s40264-020-00975-8) contains supplementary material, which is available to authorized users.

Predicting Drug-Induced Liver Injury Using Machine Learning on a Diverse Set of Predictors

A major challenge in drug development is safety and toxicity concerns due to drug side effects. One such side effect, drug-induced liver injury (DILI), is considered a primary factor in regulatory clearance. The Critical Assessment of Massive Data Analysis (CAMDA) 2020 CMap Drug Safety Challenge goal was to develop prediction models based on gene perturbation of six preselected cell-lines (CMap L1000), extended structural information (MOLD2), toxicity data (TOX21), and FDA reporting of adverse events (FAERS). Four types of DILI classes were targeted, including two clinically relevant scores and two control classifications, designed by the CAMDA organizers. The L1000 gene expression data had variable drug coverage across cell lines with only 247 out of 617 drugs in the study measured in all six cell types. We addressed this coverage issue by using Kru-Bor ranked merging to generate a singular drug expression signature across all six cell lines. These merged signatures were then narrowed down to the top and bottom 100, 250, 500, or 1,000 genes most perturbed by drug treatment. These signatures were subject to feature selection using Fisher's exact test to identify genes predictive of DILI status. Models based solely on expression signatures had varying results for clinical DILI subtypes with an accuracy ranging from 0.49 to 0.67 and Matthews Correlation Coefficient (MCC) values ranging from -0.03 to 0.1. Models built using FAERS, MOLD2, and TOX21 also had similar results in predicting clinical DILI scores with accuracy ranging from 0.56 to 0.67 with MCC scores ranging from 0.12 to 0.36. To incorporate these various data types with expression-based models, we utilized soft, hard, and weighted ensemble voting methods using the top three performing models for each DILI classification. These voting models achieved a balanced accuracy up to 0.54 and 0.60 for the clinically relevant DILI subtypes. Overall, from our experiment, traditional machine learning approaches may not be optimal as a classification method for the current data.

In-Vitro Approaches to Predict and Study T-Cell Mediated Hypersensitivity to Drugs

Mitigating the risk of drug hypersensitivity reactions is an important facet of a given pharmaceutical, with poor performance in this area of safety often leading to warnings, restrictions and withdrawals. In the last 50 years, efforts to diagnose, manage, and circumvent these obscure, iatrogenic diseases have resulted in the development of assays at all stages of a drugs lifespan. Indeed, this begins with intelligent lead compound selection/design to minimize the existence of deleterious chemical reactivity through exclusion of ominous structural moieties. Preclinical studies then investigate how compounds interact with biological systems, with emphasis placed on modeling immunological/toxicological liabilities. During clinical use, competent and accurate diagnoses are sought to effectively manage patients with such ailments, and pharmacovigilance datasets can be used for stratification of patient populations in order to optimise safety profiles. Herein, an overview of some of the in-vitro approaches to predict intrinsic immunogenicity of drugs and diagnose culprit drugs in allergic patients after exposure is detailed, with current perspectives and opportunities provided.

Association of CYP1A1 and CYP1B1 inhibition in in vitro assays with drug-induced liver injury

Drug-induced liver injury (DILI) is one of the major causes for the discontinuation of drug development and withdrawal of drugs from the market. Since it is known that reactive metabolite formation and being substrates or inhibitors of cytochrome P450s (P450s) are associated with DILI, we systematically investigated the association between human P450 inhibition and DILI. The inhibitory activity of 266 DILI-positive drugs (DILI drugs) and 92 DILI-negative drugs (no-DILI drugs), which were selected from Liver Toxicity Knowledge Base (US Food and Drug Administration), against 8 human P450 forms was assessed using recombinant enzymes and luminescent substrates, and the threshold values showing the highest balanced accuracy for DILI discrimination were determined for each P450 enzyme using receiver operating characteristic analyses. The results showed that among the P450s tested, CYP1A1 and CYP1B1 were inhibited by DILI drugs more than no-DILI drugs with a statistical significance. We found that 91% of drugs that showed inhibitory activity greater than the threshold values against CYP1A1 or CYP1B1 were DILI drugs. The results of internal 5-fold cross-validation confirmed the usefulness of CYP1A1 and CYP1B1 inhibition data for the threshold-based discrimination of DILI drugs. Although the contribution of these P450s to drug metabolism in the liver is considered minimal, our present findings suggest that the assessment of CYP1A1 and CYP1B1 inhibition is useful for screening DILI risk of drug candidates at the early stage of drug development.

Trade-off Predictivity and Explainability for Machine-Learning Powered Predictive Toxicology: An in-Depth Investigation with Tox21 Data Sets

Selecting a model in predictive toxicology often involves a trade-off between prediction performance and explainability: should we sacrifice the model performance to gain explainability or vice versa. Here we present a comprehensive study to assess algorithm and feature influences on model performance in chemical toxicity research. We conducted over 5000 models for a Tox21 bioassay data set of 65 assays and ∼7600 compounds. Seven molecular representations as features and 12 modeling approaches varying in complexity and explainability were employed to systematically investigate the impact of various factors on model performance and explainability. We demonstrated that end points dictated a model's performance, regardless of the chosen modeling approach including deep learning and chemical features. Overall, more complex models such as (LS-)SVM and Random Forest performed marginally better than simpler models such as linear regression and KNN in the presented Tox21 data analysis. Since a simpler model with acceptable performance often also is easy to interpret for the Tox21 data set, it clearly was the preferred choice due to its better explainability. Given that each data set had its own error structure both for dependent and independent variables, we strongly recommend that it is important to conduct a systematic study with a broad range of model complexity and feature explainability to identify model balancing its predictivity and explainability.

Hepatotoxicity of FDA-approved small molecule kinase inhibitors

Introduction: Given their importance in cellular processes and association with numerous diseases, protein kinases have emerged as promising targets for drugs. The FDA has approved greater than fifty small molecule kinase inhibitors (SMKIs) since 2001. Nevertheless, severe hepatotoxicity and related fatal cases have grown as a potential challenge in the advancement of these drugs, and the identification and diagnosis of drug-induced liver injury (DILI) are thorny problems for clinicians.Areas covered: This article summarizes the progression and analyzes the significant features in the study of SMKI hepatotoxicity, including clinical observations and investigations of the underlying mechanisms.Expert opinion: The understanding of SMKI-associated hepatotoxicity relies on the development of preclinical models and improvement of clinical assessment. With a full understanding of the role of inflammation in DILI and the mediating role of cytokines in inflammation, cytokines are promising candidates as sensitive and specific biomarkers for DILI. The emergence of three-dimensional spheroid models demonstrates potential use in providing clinically relevant data and predicting hepatotoxicity of SMKIs.

Application of a Rat Liver Drug Bioactivation Transcriptional Response Assay Early in Drug Development That Informs Chemically Reactive Metabolite Formation and Potential for Drug-induced Liver Injury

Drug-induced liver injury is a major reason for drug candidate attrition from development, denied commercialization, market withdrawal, and restricted prescribing of pharmaceuticals. The metabolic bioactivation of drugs to chemically reactive metabolites (CRMs) contribute to liver-associated adverse drug reactions in humans that often goes undetected in conventional animal toxicology studies. A challenge for pharmaceutical drug discovery has been reliably selecting drug candidates with a low liability of forming CRM and reduced drug-induced liver injury potential, at projected therapeutic doses, without falsely restricting the development of safe drugs. We have developed an in vivo rat liver transcriptional signature biomarker reflecting the cellular response to drug bioactivation. Measurement of transcriptional activation of integrated nuclear factor erythroid 2-related factor 2 (NRF2)/Kelch-like ECH-associated protein 1 (KEAP1) electrophilic stress, and nuclear factor erythroid 2-related factor 1 (NRF1) proteasomal endoplasmic reticulum (ER) stress responses, is described for discerning estimated clinical doses of drugs with potential for bioactivation-mediated hepatotoxicity. The approach was established using well benchmarked CRM forming test agents from our company. This was subsequently tested using curated lists of commercial drugs and internal compounds, anchored in the clinical experience with human hepatotoxicity, while agnostic to mechanism. Based on results with 116 compounds in short-term rat studies, with consideration of the maximum recommended daily clinical dose, this CRM mechanism-based approach yielded 32% sensitivity and 92% specificity for discriminating safe from hepatotoxic drugs. The approach adds new information for guiding early candidate selection and informs structure activity relationships (SAR) thus enabling lead optimization and mechanistic problem solving. Additional refinement of the model is ongoing. Case examples are provided describing the strengths and limitations of the approach.

Drug Induced Liver Injury (DILI). Mechanisms and Medicinal Chemistry Avoidance/Mitigation Strategies

Adverse drug reactions (ADRs) are a common cause of attrition in drug discovery and development and drug-induced liver injury (DILI) is a leading cause of preclinical and clinical drug terminations. This perspective outlines many of the known DILI mechanisms and assessment methods used to evaluate and mitigate DILI risk. Literature assessments and retrospective analyses using verified DILI-associated drugs from the Liver Tox Knowledge Base (LTKB) have been used to derive the predictive value of each end point, along with combination approaches of multiple methods. In vitro assays to assess inhibition of the bile salt export pump (BSEP), mitotoxicity, reactive metabolite (RM) formation, and hepatocyte cytolethality, along with physicochemical properties and clinical dose provide useful DILI predictivity. This Perspective also highlights some of the strategies used by medicinal chemists to reduce DILI risk during the optimization of drug candidates.

Machine-Learning Prediction of Oral Drug-Induced Liver Injury (DILI) via Multiple Features and Endpoints

Drug discovery is a costly process which usually takes more than 10 years and billions of dollars for one successful drug to enter the market. Despite all the safety tests, drugs may still cause adverse reactions and be restricted in use or even withdrawn from the market. Drug-induced liver injury (DILI) is one of the major adverse drug reactions, and computational models may be used to predict and reduce it. To assess the computational prediction performance of DILI, we curated DILI endpoints from three databases and prepared drug features including chemical descriptors, therapeutic classifications, gene expressions, and binding proteins. We trained machine-learning models to predict the various DILI endpoints using different drug features. Using the optimal feature sets, the top-performing models obtained areas under the receiver operating characteristic curve (AUC) around 0.8 for some DILI endpoints. We found that some features, including therapeutic classifications and proteins, have good prediction performance towards DILI. We also discovered that the severity of DILI endpoints as well as the selection of negative samples may significantly affect the prediction results. Overall, our study provided a comprehensive collection, curation, and prediction of DILI endpoints using various drug features, which may help the drug researchers to better understand and prevent DILI during the drug discovery process.

Mechanism-based integrated assay systems for the prediction of drug-induced liver injury

Drug-induced liver injury (DILI) can cause hepatic failure and result in drug withdrawal from the market. It has host-related and compound-dependent mechanisms. Preclinical prediction of DILI risk is very challenging and safety assessments based on animals inadequately forecast human DILI risk. In contrast, human-derived in vitro cell culture-based models could improve DILI risk prediction accuracy. Here, we developed and validated an innovative method to assess DILI risk associated with various compounds. Fifty-four marketed and withdrawn drugs classified as DILI risks of "most concern", "less concern", and "no concern" were tested using a combination of four assays addressing mitochondrial injury, intrahepatic lipid accumulation, inhibition of bile canalicular network formation, and bile acid accumulation. Using the inhibitory potencies of the drugs evaluated in these in vitro tests, an algorithm with the highest available DILI risk prediction power was built by artificial neural network (ANN) analysis. It had an overall forecasting accuracy of 73%. We excluded the intrahepatic lipid accumulation assay to avoid overfitting. The accuracy of the algorithm in terms of predicting DILI risks was 62% when it was constructed by ANN but only 49% when it was built by the point-added scoring method. The final algorithm based on three assays made no DILI risk prediction errors such as "most concern " instead of "no concern" and vice-versa. Our mechanistic approach may accurately predict DILI risks associated with numerous candidate drugs.

Understanding Idiosyncratic Toxicity: Lessons Learned from Drug-Induced Liver Injury

Idiosyncratic adverse drug reactions (IADRs) encompass a diverse group of toxicities that can vary by drug and patient. The complex and unpredictable nature of IADRs combined with the fact that they are rare makes them particularly difficult to predict, diagnose, and treat. Common clinical characteristics, the identification of human leukocyte antigen risk alleles, and drug-induced proliferation of lymphocytes isolated from patients support a role for the adaptive immune system in the pathogenesis of IADRs. Significant evidence also suggests a requirement for direct, drug-induced stress, neoantigen formation, and stimulation of an innate response, which can be influenced by properties intrinsic to both the drug and the patient. This Perspective will provide an overview of the clinical profile, mechanisms, and risk factors underlying IADRs as well as new approaches to study these reactions, focusing on idiosyncratic drug-induced liver injury.

Strategies for Early Prediction and Timely Recognition of Drug-Induced Liver Injury: The Case of Cyclin-Dependent Kinase 4/6 Inhibitors

The idiosyncratic nature of drug-induced liver injury (DILI) represents a current challenge for drug developers, regulators and clinicians. The myriad of agents (including medications, herbals, and dietary supplements) with recognized DILI potential not only strengthens the importance of the post-marketing phase, when urgent withdrawal sometimes occurs for rare unanticipated liver toxicity, but also shows the imperfect predictivity of pre-clinical models and the lack of validated biomarkers beyond traditional, non-specific liver function tests. After briefly reviewing proposed key mechanisms of DILI, we will focus on drug-related risk factors (physiochemical and pharmacokinetic properties) recently proposed as predictors of DILI and use cyclin-dependent kinase 4/6 inhibitors, relatively novel oral anticancer medications approved for breast cancer, as a case study to discuss the feasibility of early detection of DILI signals during drug development: published data from pivotal clinical trials, unpublished post-marketing reports of liver adverse events, and pharmacokinetic properties will be used to provide a comparative evaluation of their liver safety and gain insight into drug-related risk factors likely to explain the observed differences.

Risk profiling using metabolomic characteristics for susceptible individuals of drug-induced liver injury caused by Polygonum multiflorum

Idiosyncratic drug-induced liver injury (IDILI) is a rare but potentially severe adverse drug reaction. To date, identifying individuals at risk for IDILI remains challenging. This is a prospective study, where a nested case-control (1:5) design was adopted. For six patients who had abnormalities in liver function test after Polygonum multiflorum Thunb. (PM) ingestion (susceptible group), 30 patients with normal liver function were matched (tolerant group). Based on liquid chromatography-mass spectrometry, metabolomics analysis was done on serum samples prior to PM ingestion, to screen the differential metabolites and characterize metabolomic profiles of patient serum in the two groups. Multivariate analysis showed that there were remarkable separations between susceptible and tolerant groups. A total of 25 major differential metabolites were screened out, involving glycerophospholipid metabolism, sphingolipid metabolism, fatty acid metabolism, histidine metabolism and aromatic amino acid metabolism. Wherein, the area under the curve of the receiver operating characteristic curves of metabolites PE 22:6, crotonoyl-CoA, 2E-tetradecenoyl-CoA, phenyllactic acid, indole-5,6-quinone, phosphoribosyl-ATP were all greater than 0.9. The overall serum metabolic profile comprising of 25 metabolites could clearly distinguish susceptible and tolerant groups. This proof-of-concept study used metabolomics to characterize the metabolic profile of IDILI risk individuals before drug ingestion for the first time. The metabolome characteristics in patient serum before PM ingestion may predict the risk of liver injury after PM ingestion.

Predicting Drug-Induced Liver Injury with Bayesian Machine Learning

Drug induced liver injury (DILI) can require significant risk management in drug development and on occasion can cause morbidity or mortality, leading to drug attrition. Optimizing candidates preclinically can minimize hepatotoxicity risk, but it is difficult to predict due to multiple etiologies encompassing DILI, often with multifactorial and overlapping mechanisms. In addition to epidemiological risk factors, physicochemical properties, dose, disposition, lipophilicity, and hepatic metabolic function are also relevant for DILI risk. Better human-relevant, predictive models are required to improve hepatotoxicity risk assessment in drug discovery. Our hypothesis is that integrating mechanistically relevant hepatic safety assays with Bayesian machine learning will improve hepatic safety risk prediction. We present a quantitative and mechanistic risk assessment for candidate nomination using data from in vitro assays (hepatic spheroids, BSEP, mitochondrial toxicity, and bioactivation), together with physicochemical (cLogP) and exposure (Cmax_total) variables from a chemically diverse compound set (33 no/low-, 40 medium-, and 23 high-severity DILI compounds). The Bayesian model predicts the continuous underlying DILI severity and uses a data-driven prior distribution over the parameters to prevent overfitting. The model quantifies the probability that a compound falls into either no/low-, medium-, or high-severity categories, with a balanced accuracy of 63% on held-out samples, and a continuous prediction of DILI severity along with uncertainty in the prediction. For a binary yes/no DILI prediction, the model has a balanced accuracy of 86%, a sensitivity of 87%, a specificity of 85%, a positive predictive value of 92%, and a negative predictive value of 78%. Combining physiologically relevant assays, improved alignment with FDA recommendations, and optimal statistical integration of assay data leads to improved DILI risk prediction.

Rotenone Increases Isoniazid Toxicity but Does Not Cause Significant Liver Injury: Implications for the Hypothesis that Inhibition of the Mitochondrial Electron Transport Chain Is a Common Mechanism of Idiosyncratic Drug-Induced Liver Injury

Idiosyncratic drug reactions (IDRs) significantly increase the risk of failure in drug development. The major IDR leading to drug candidate failure is idiosyncratic drug-induced liver injury (IDILI). Although most evidence suggests that IDRs are mediated by the immune system, there are other hypotheses, such as mitochondrial dysfunction. Many pharmaceutical companies routinely screen for mitochondrial toxicity in an attempt to "derisk" drug candidates. However, the basic hypothesis has never been rigorously tested. A major assay used for this screening involves measurement of inhibition of the mitochondrial electron transport chain. One study found that the combination of rotenone and isoniazid, which inhibit mitochondrial complex I and II, respectively, were synergistic in causing hepatocyte toxicity in vitro and suggested the combination of another drug that inhibited complex I would increase the risk of isoniazid-induced liver injury in patients. We tested this hypothesis in vivo where wild-type and PD-1-/- mice administered anti-CTLA-4, our impaired immune tolerance mouse model, were given 0.02% (w/v) rotenone in water or 0.1%, 0.05%, and 0.01% (w/w) rotenone alone or in combination with isoniazid in food. The cotreatment led to lethality in 100% of the animals receiving 0.1% rotenone and 0.2% isoniazid and 83% of the animals cotreated with 0.05% rotenone and 0.2% isoniazid in food. Nevertheless, there was no significant increase in GLDH or histological evidence of liver injury. No signs of toxicity were observed in any of the mice given rotenone or isoniazid alone. Even though inhibition of the mitochondrial electron transport chain did not lead to significant liver toxicity, it could provide danger signals that promote immune-mediated liver injury. However, rotenone did not significantly increase the liver injury induced by isoniazid in our impaired immune tolerance model. Overall, we conclude that inhibition of the mitochondrial electron transport chain is not a significant mechanism of IDILI.

Application of hepatocyte-like cells to enhance hepatic safety risk assessment in drug discovery

Hepatic stress and injury from drugs continues to be a major concern within the pharmaceutical industry, leading to preclinical and clinical attrition precautionary warnings and post-market withdrawal of drugs. There is a requirement for more predictive and mechanistically accurate models to aid risk assessment. Primary human hepatocytes, subject to isolation stress, cryopreservation, donor-to-donor variation and a relatively short period of functional capability in two-dimensional cultures, are not suitable for high-throughput screening procedures. There are two areas within the drug discovery pipeline that the generation of a stable, metabolically functional hepatocyte-like cell with unlimited supply would have major impact. First, in routine, cell health risk-assessment assays where hepatic cell lines are typically deployed. Second, at later stages of the drug discovery pipeline approaching candidate nomination where bespoke/investigational studies refining and understanding the risk to patients use patient-derived induced pluripotent stem cell (iPSC) hepatocytes retaining characteristics from the patient, e.g. HLA susceptibility alleles, iPSC hepatocytes with defined disease phenotypes or genetic characteristics that have the potential to make the hepatocyte more sensitive to a particular stress mechanism. Functionality of patient-centric hepatocyte-like cells is likely to be enhanced when coupled with emerging culture systems, such as three-dimensional spheroids or microphysiological systems. Ultimately, the aspiration to confidently use human-relevant in vitro models to predict human-specific hepatic toxicity depends on the integration of promising emerging technologies.This article is part of the theme issue 'Designer human tissue: coming to a lab near you'.

A simple model to solve a complex drug toxicity problem

Linear drug toxicity models like therapeutic index (TI), physicochemical rules (rule of five, 3/75), ligand efficiency indices (LEI), ideal pharmacokinetic (PK) and pharmacodynamic (PD) profiles are widely used in drug discovery and development. In spite of this, predicting drug toxicity at various stages remains challenging and the overall productivity (<20%) and ultimate benefit to the patients remain low. A simple drug toxicity model, "Drug Toxicity Index" (DTI), is developed here using 711 oral drugs. DTI redefines drug toxicity as scaled biphasic and exponential functions of PD, PK and physicochemical parameters. PD, PK and physicochemical toxicity contributions were estimated from the on and off target IC₅₀, maximum unbound plasma drug concentration (free C _max), and log D values, respectively. These contributions are then scaled by molar dose and oral bioavailability and the logarithm of the sum of scaled contributions is DTI. Drugs with DTI above the WHO ATC drug category specific average values consistently have toxic profiles, while drugs with DTI below this average are relatively safe. DTI performs better than standard rules for lead optimization, LEI and exposure based TIs in identifying safe and toxic drugs. DTI classifies 392 drugs reported in the US-FDA's Liver Toxicity Knowledge Base (LTKB) with an AUC for ROC curves of 0.91-0.64 for different WHO ATC categories. DTI has been used to predict network meta-analysis results on relative toxicity within/across eight different therapeutic areas. It is useful in understanding PD, PK and physicochemical toxicity contributions and identifying potentially toxic drugs and the toxicity of recently approved drugs. Decision trees are proposed for applying the DTI concept in preclinical drug discovery and clinical trial settings. DTI can potentially reduce failure in drug discovery and might be useful in therapeutic drug monitoring and in xenobiotic and environmental toxicity studies.

The Development of a Database for Herbal and Dietary Supplement Induced Liver Toxicity

The growing use of herbal dietary supplements (HDS) in the United States provides compelling evidence for risk of herbal-induced liver injury (HILI). Information on HDS products was retrieved from MedlinePlus of the U.S. National Library of Medicine and the herbal monograph of the European Medicines Agency. The hepatotoxic potential of HDS was ascertained by considering published case reports. Other relevant data were collected from governmental documents, public databases, web sources, and the literature. We collected information for 296 unique HDS products. Evidence of hepatotoxicity was reported for 67, that is 1 in 5, of these HDS products. The database revealed an apparent gender preponderance with women representing 61% of HILI cases. Culprit hepatotoxic HDS were mostly used for weight control, followed by pain and inflammation, mental stress, and mood disorders. Commonly discussed mechanistic events associated with HILI are reactive metabolites and oxidative stress, mitochondrial injury, as well as inhibition of transporters. HDS⁻drug interactions, causing both synergistic and antagonizing effects of drugs, were also reported for certain HDS. The database contains information for nearly 300 commonly used HDS products to provide a single-entry point for better comprehension of their impact on public health.

Drug-Induced Liver Injury Associated with the Use of Everolimus in a Liver Transplant Patient

Drug-induced liver injury (DILI) has not been previously reported as a complication of treatment with everolimus. A 56-year-old Caucasian male liver transplant recipient developed DILI after receiving everolimus. Elevations in transaminase levels occurred within a week of starting everolimus and an upward trend in the transaminase levels continued with supporting histopathologic changes confirmed by liver biopsy. Within one week of drug discontinuation, his liver enzymes normalized to baseline. This report includes a brief review of the pharmacokinetic properties of everolimus, a review of the relevant literature, and an analysis using the RUCAM and Naranjo algorithms.

Impact of Physicochemical Properties on Dose and Hepatotoxicity of Oral Drugs

A database containing maximum daily doses of 1841 marketed oral drugs was used to examine the influence of physicochemical properties on dose and hepatotoxicity (drug induced liver injury, DILI). Drugs in the highest ∼20% dose range had significantly reduced mean lipophilicity and molecular weight, increased fractional surface area, increased % of acids, and decreased % of bases versus drugs in the lower ∼60% dose range. Drugs in the ∼20-40% dose range had intermediate mean properties, similar to the mean values for the full drug set. Drugs that are both large and highly lipophilic almost invariably do not have doses in the upper ∼20% range. The results show that oral druglike physicochemical properties are different according to these dose ranges, and this is consistent with maintenance of acceptable safety profiles as efficacious exposure increases. Verified DILI annotations from a compilation of >1000 approved drugs (Chen, M.; et al. Drug Discov. Today, 2016, 21, 648 ) were used. The drugs classified as "No DILI" ( n = 163) had significantly lower dose and lipophilicity, and higher Fsp3 (fraction of carbon atoms that are sp³ hybridized) versus the "Most DILI" ( n = 163) drugs. The percentages of acids were reduced and bases increased in the "No DILI" versus the "Most DILI" groups. Drugs classified as "Less DILI" or "Ambiguous DILI" had intermediate mean values of dose, lipophilicity, Fsp3, and % acids and bases. The impact of lipophilicity and Fsp3 on DILI increases in the upper 20% versus the lower 80% dose range, and a simple decision tree model predicted "No DILI" versus "Most DILI" outcomes with 82% accuracy. The model correctly classified 19 of 22 drugs (86%) that failed in development due to human hepatotoxicity. Because many oral drugs lacking DILI annotations are predicted to be "Most DILI", the model is best used preclinically in conjunction with experimental DILI mitigation.

Evaluation of the Relevance of DILI Predictive Hypotheses in Early Drug Development: Review of In Vitro Methodologies vs BDDCS Classification

Drug-induced liver injury (DILI) is a major safety concern; it occurs frequently; it is idiosyncratic; it cannot be adequately predicted; and a multitude of underlying mechanisms has been postulated. A number of experimental approaches to predict human DILI have been proposed utilizing in vitro screening such as inhibition of mitochondrial function, hepatobiliary transporter inhibition, reactive metabolite formation with and without covalent binding, and cellular health, but they have achieved only minimal success. Several studies have shown total administered dose alone or in combination with drug lipophilicity to be correlated with a higher risk of DILI. However, it would be best to have a predictive DILI methodology early in drug development, long before the clinical dose is known. Here we discuss the extent to which Biopharmaceutics Drug Disposition Classification System (BDDCS) defining characteristics, independent of knowing actual drug pharmacokinetics/pharmacodynamics and dose, can be used to evaluate prior published predictive proposals. Our results show that BDDCS Class 2 drugs exhibit the highest DILI severity, and that all of the short-lived published methodologies evaluated here, except when daily dose is known, do not yield markedly better predictions than BDDCS. The assertion that extensively metabolized compounds are at higher risk of developing DILI is confirmed, but can be enhanced by differentiating BDDCS Class 2 from Class 1 drugs.

CONCLUSION

Our published analyses suggest that comparison of proposed DILI prediction methodologies with BDDCS classification is a useful tool to evaluate the potential reliability of newly proposed algorithms, although BDDCS classification itself is not sufficiently predictive. Almost all of the predictive DILI metrics do no better than just avoiding BDDCS Class 2 drugs, although some early data with microliver platforms enabling long-enduring metabolic competency show promising results.

N-acetylcysteine and prednisolone treatment improved serum biochemistries in suspected flupirtine cases of severe idiosyncratic liver injury

BACKGROUND & AIMS

The analgesic flupirtine has been linked to cases of severe idiosyncratic drug-induced liver injury (sFILI). We therefore examined whether N-acetylcysteine (NAC) and glucocorticoid therapy is effective in the management of sFILI.

METHODS

In a retrospective cohort study efficacy of NAC-infusion and oral prednisolone treatments on liver-function-tests (LFTs) and clinical outcome of 21 sFILI cases was evaluated by comparing it to an external cohort of 30 sFILI cases not receiving the antidote. LFTs were also assayed in human hepatocyte cultures treated with flupirtine for 96 hours. Additionally, modulation of glutathione stores in cultures of human hepatocytes treated with 80 drugs was investigated.

RESULTS

Upon admission, patients presented Alanine aminotransferase (ALT), aspartate aminotransferase (AST) and bilirubin elevations of up to 90-, 35- and 30-fold of upper limits of normal (ULN) respectively. The average INR was 2.2, and 50% of patients had a high Model for end-stage liver disease (MELD) score of ≥25 and included 5 cases of 28-32. The combined NAC/prednisolone treatment was well tolerated and led to significant ALT, AST and INR improvements within 2 weeks. However, the hyperbilirubinaemia resolved slowly. Two patients with late start of NAC/prednisolone treatment developed hepatic encephalopathy and required liver transplantation; the remaining patients recovered without long-term sequela. Based on serum biochemistries sFILI cases resolved more rapidly (P < .01) when compared to untreated sFILI patients and included a case with fatal outcome. Additionally, in vitro studies revealed glutathione depletion as a common culprit for drugs with liver liabilities.

CONCLUSIONS

Based on these initial findings a prospective randomized clinical trial (RCT) is needed to confirm safety and efficacy of NAC/prednisolone treatment in sFILI.