Machine Learning-Based Prediction of Drug-Induced Hepatotoxicity: An OvA-QSTR Approach

Drug-induced hepatotoxicity, also known as drug-induced liver injury (DILI), is among the possible adverse effects of pharmacotherapy. This clinical condition is accepted as one of the factors leading to patient mortality and morbidity. The LiverTox database was built by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) to predict potential liver damage from medications and take appropriate precautions. The database has classified medicines into seven risk categories (A, B, C, D, E, E*, and X) to avoid medicine-induced liver toxicity. The hepatic damage risk decreases from group A to group E. This study did not include the E* and X classes because they contained unverified and unknown data groups. Our study aims to predict potential liver damage of new drug molecules without using experimental animals. We predict which of the LiverTox risk category drugs with unknown liver toxicity potential will fall into using our one-vs-all quantitative structure-toxicity relationship (OvA-QSTR) model. Our dataset, consisting of 678 organic drug molecules from different pharmacological classes, was collected from LiverTox. The OvA-QSTR models implemented by Bayesian Network (BayesNet) performed well based on the selected descriptors, with the precision-recall curve (PRC) areas ranging from 0.718 to 0.869. Our OvA-QSTR models provide a reliable premarketing risk evaluation of pharmaceutical-induced liver damage potential and offer predictions for different risk levels in DILI.

Modeling Based on Ensemble Learning Methods for Detection of Diagnostic Biomarkers from LncRNA Data in Rats Treated with Cis-Platinum-Induced Hepatotoxicity

BACKGROUND

The first aim of this study is to perform bioinformatic analysis of lncRNAs obtained from liver tissue samples from rats treated with cisplatin hepatotoxicity and without pathology. Another aim is to identify possible biomarkers for the diagnosis/early diagnosis of hepatotoxicity by modeling the data obtained from bioinformatics analysis with ensemble learning methods.

METHODS

In the study, 20 female Sprague-Dawley rats were divided into a control group and a hepatotoxicity group. Liver samples were taken from rats, and transcriptomic and histopathological analyses were performed. The dataset achieved from the transcriptomic analysis was modeled with ensemble learning methods (stacking, bagging, and boosting). Modeling results were evaluated with accuracy (Acc), balanced accuracy (B-Acc), sensitivity (Se), specificity (Sp), positive predictive value (Ppv), negative predictive value (Npv), and F1 score performance metrics. As a result of the modeling, lncRNAs that could be biomarkers were evaluated with variable importance values.

RESULTS

According to histopathological and immunohistochemical analyses, a significant increase was observed in the sinusoidal dilatation and Hsp60 immunoreactivity values in the hepatotoxicity group compared to the control group (p < 0.0001). According to the results of the bioinformatics analysis, 589 lncRNAs showed different expressions in the groups. The stacking model had the best classification performance among the applied ensemble learning models. The Acc, B-Acc, Se, Sp, Ppv, Npv, and F1-score values obtained from this model were 90%, 90%, 80%, 100%, 100%, 83.3%, and 88.9%, respectively. lncRNAs with id rna-XR_005492522.1, rna-XR_005492536.1, and rna-XR_005505831.1 with the highest three values according to the variable importance obtained as a result of stacking modeling can be used as predictive biomarker candidates for hepatotoxicity.

CONCLUSIONS

Among the ensemble algorithms, the stacking technique yielded higher performance results as compared to the bagging and boosting methods on the transcriptomic data. More comprehensive studies can support the possible biomarkers determined due to the research and the decisive results for the diagnosis of drug-induced hepatotoxicity.

A review of challenges and prospects of 3D cell-based culture models used for studying drug induced liver injury during early phases of drug development

Drug-induced liver injury (DILI) is the leading cause of compound attrition during drug development. Over the years, a battery of in-vitro cell culture toxicity tests is being conducted to evaluate the toxicity of compounds prior to testing in laboratory animals. Two-dimensional (2D) in-vitro cell culture models are commonly used and have provided a great deal of knowledge; however, these models often fall short in mimicking natural structures of tissues in-vivo. Testing in humans is the most logical method, but unfortunately there are ethical limitations associated with human tests. To overcome these limitations better human-relevant, predictive models are required. The past decade has witnessed significant efforts towards the development of three-dimensional (3D) in-vitro cell culture models better mimicking in-vivo physiology. 3D cell culture has advantages in being representative of the interactions of cells in-vivo and when validated can act as an interphase between 2D cell culture models and in-vivo animal models. The current review seeks to provide an overview of the challenges that make biomarkers used for detection of DILI not to be sensitive enough during drug development and explore how 3D cell culture models can be used to address the gap with the current models.

Identification of New Toxicity Mechanisms in Drug-Induced Liver Injury through Systems Pharmacology

Among adverse drug reactions, drug-induced liver injury presents particular challenges because of its complexity, and the underlying mechanisms are still not completely characterized. Our knowledge of the topic is limited and based on the assumption that a drug acts on one molecular target. We have leveraged drug polypharmacology, i.e., the ability of a drug to bind multiple targets and thus perturb several biological processes, to develop a systems pharmacology platform that integrates all drug–target interactions. Our analysis sheds light on the molecular mechanisms of drugs involved in drug-induced liver injury and provides new hypotheses to study this phenomenon.

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.

MRI-based preclinical discovery of DILI: A lesson from paracetamol-induced hepatotoxicity

Worldwide, drug-induced liver injury (DILI) is a major cause of hepatic failure. It is also the leading cause of withdrawal, cautionary labeling, and restricted usage of licensed drugs; therefore, European Medicines Agency (EMA) and United States Food and Drug Administration (FDA) warn that the existing methods of assessing DILI are insufficient and that some of the translational biomarkers of hepatotoxicity must be relooked. Magnetic resonance imaging (MRI) seems to be a proper tool in elucidating the effects of DILI in both preclinical and clinical studies, providing excellent visualization of the morphology of the liver parenchyma. Therefore, herein, we propose preclinical MRI assessment of liver injury in experimental paracetamol-treated rats. Quantitative MRI clearly provides evidence of adverse effects in the liver tissue caused by a single overdose of paracetamol (1 g kg^-1 and 1.5 g kg^-1 b.w.). The results of the MRI were confirmed by the histopathological examination (H&E) of the rat liver specimen, however the adverse effects were not disclosed due to standard aminotransferase assays (ALT/AST) in rat blood serum. The results of our analysis demonstrate the successful application of MRI in the examination of paracetamol-induced hepatotoxicity in rats; it has a potential to serve as the early diagnostic tool for the prediction of DILI in preclinical evaluation.

Developmental toxicity and biological responses of zebrafish (Danio rerio) exposed to anti-inflammatory drug ketoprofen

Ketoprofen a nonsteroidal anti-inflammatory drug (NSAID) is widely used in over-the-counter to treat pain, swelling and inflammation. Due to extensive application these drugs has been detected in surface waters which may create a risk to aquatic organisms. The aim of the present study is to assess the ecotoxicity of ketoprofen at different concentrations (1, 10 and 100 μg/ml) on embryos and adult zebrafish (1, 10 and 100 μg L^-1) under laboratory conditions. In embryos, concentration dependent developmental changes such as edema, spinal curvature, slow heartbeat, delayed hatching, and mortality rate were observed. In adult zebrafish, biochemical enzymes such as AST, ALT and LDH activities were significantly (P < 0.05) increased whereas a decrease in Na⁺/K⁺-ATPase activity was noticed in all the tested concentrations of the drug ketoprofen. Similarly, exposure of ketoprofen caused a significant decrease in antioxidant levels in liver tissue (SOD, CAT, GSH, GPx, and GST). However, lipid peroxidation (LPO) level in liver tissue was found to be increased. The histopathological studies further evidenced the impact of ketoprofen in the liver tissue of zebrafish. The present result concludes that ketoprofen could have an impact on the development and biological endpoints of the zebra fish at above concentrations. The malformation in the development of the embryo and changes in the biological end points may provide integrated evaluation of the toxic effect of ketoprofen on zebrafish in a new perspective.

MiR-122 and other microRNAs as potential circulating biomarkers of drug-induced liver injury

INTRODUCTION

Drug-induced liver injury (DILI) is a severe adverse drug reaction which is of major concern to patients, clinicians and the pharmaceutical industry. Accurate and rapid detection of DILI is important for patient stratification and treatment in the clinic and benefits preclinical drug design and risk assessment. MicroRNAs (miRNAs) offer a potential new and improved class of circulating biomarkers of DILI over the current gold standard biomarkers. Areas covered: This review highlights the shortcomings of the currently used panel of biomarkers and how miRNAs, primarily miR-122, show an improved level of specificity and sensitivity in the prediction of DILI. Furthermore, the use of miRNAs as potential markers of progression of DILI and specific zonated damage within the liver is discussed. Expert commentary: MiRNAs offer more sensitive and specific markers over the current biomarkers for DILI. Combinations of different miRNAs may be able to relay the location of DILI and the progression of disease. More studies using different hepatotoxins apart from acetaminophen will ultimately strengthen the case for the clinical introduction of miRNAs as biomarkers of DILI.

Evaluation of acute and sublethal effects of chloroquine (C18H26CIN3) on certain enzymological and histopathological biomarker responses of a freshwater fish Cyprinus carpio

•

Chloroquine (CQ) toxicity on fresh water fingerlings Cyprinus carpio was studied.

•

Median lethal concentration (96 h) was noted.

•

Acute (96 h) and sub-lethal (35 days) treatments was performed.

•

Enzymological activity and histological alteration was analysed.

•

Drug CQ has a toxic effect on non-target organism.

In this study the toxicity of antimalarial drug chloroquine (CQ) on certain enzymological (GOT, GPT and LDH) and histopathological alterations (Gill, liver and kidney) of a freshwater fish Cyprinus carpio was studied after acute (96 h) and sublethal (35 days) exposure. The median lethal concentration (96 h) of CQ was 31.62 mg/ml. During acute treatment (CQ at 31.62 mg/ml) the treated fish groups showed a significant increase in GOT and GPT activities in blood plasma; whereas LDH activity was decreased when compare to control groups. To analyse the effects of drug at the lowest concentration, the fish were exposed to 3.16 mg/ml (1/10th of 96 h LC50 value) for 96 h. In sublethal treatment (3.16 mg/ml) GOT activity increased up to 14th day and decreased during the rest of the exposure period (21, 28 and 35th day). A biphasic response in GPT activity was observed. LDH activity was found to be increased throughout the study period (35 days) compare to control groups. The alterations in enzyme activities in blood plasma were found to be significant at p < 0.05 (DMRT). Many histopathological changes in vital organs such as gill, liver and kidney of fish were observed in CQ treated group (acute and sub-lethal) compare to normal group. The alterations in the enzymological and histopathological study in the present investigation indicate that the drug CQ has toxic effects on non-target organisms. We conclude that the alterations in enzymological parameters and histopathological changes can be used as biomarker to assess the health of the aquatic organism/environment. Further data on molecular studies are needed to define the mode of action and toxicity of these emerging pollutants.

Circulating levels of miR-122 increase post-mortem, particularly following lethal dosing with pentobarbital sodium: implications for pre-clinical liver injury studies

microRNA-122 (miR-122) is increasingly being measured in pre-clinical and clinical settings due to greater sensitivity and hepatic specificity compared to the gold standard liver injury biomarker alanine aminotransferase (ALT). In pre-clinical studies, various culling methods can be employed prior to collection of blood samples, including lethal injection with pentobarbital sodium (Pentoject). However, little is known about whether such an approach could alter the circulating levels of miR-122 and compromise the interpretation of data. We therefore exposed C57BL/6J mice to saline or the model hepatotoxin paracetamol and collected blood samples pre-cull (via tail bleed) and post-cull (via cardiac puncture following exposure to a rising concentration of CO₂ or intraperitoneal injection of Pentoject). Compared to pre-cull levels there was a significant increase in serum miR-122 level in mice culled with CO₂ and, to a much greater extent, in mice culled with Pentoject. As a result, whilst the serum level of miR-122 increased in Pentoject-culled animals exposed to paracetamol, the higher level in saline-treated mice rendered this difference statistically non-significant, in contrast to findings in animals culled with CO₂. ALT levels were unaffected by sacrifice method. Consistent with the in vivo findings, exposure of primary mouse hepatocytes to Pentoject provoked a rapid and concentration-dependent release of miR-122 into the culture media. Thus, for optimal design and interpretation of data from pre-clinical liver injury studies in which miR-122 is to be used as a biomarker, we recommend that blood samples are collected pre-cull whenever possible, and that lethal injection with Pentoject is avoided.

Redox-Dependent HMGB1 Isoforms as Pivotal Co-Ordinators of Drug-Induced Liver Injury: Mechanistic Biomarkers and Therapeutic Targets

SIGNIFICANCE

High-mobility group box 1 (HMGB1) is a critical protein in the coordination of the inflammatory response in drug-induced liver injury (DILI). HMGB1 is released from necrotic hepatocytes and activated immune cells. The extracellular function of HMGB1 is dependent upon redox modification of cysteine residues that control chemoattractant and cytokine-inducing properties. Existing biomarkers of DILI such as alanine aminotransferase (ALT) have limitations such as lack of sensitivity and tissue specificity that can adversely affect clinical intervention.

RECENT ADVANCES

HMGB1 isoforms have been shown to be more sensitive biomarkers than ALT for predicting DILI development and the requirement for liver transplant following acetaminophen (APAP) overdose. Hepatocyte-specific conditional knockout of HMGB1 has demonstrated the pivotal role of HMGB1 in DILI and liver disease. Tandem mass spectrometry (MS/MS) enables the characterization and quantification of different mechanism-dependent post-translationally modified isoforms of HMGB1.

CRITICAL ISSUES

HMGB1 shows great promise as a biomarker of DILI. However, current diagnostic assays are either too time-consuming to be clinically applicable (MS/MS) or are unable to distinguish between different redox and acetyl isoforms of HMGB1 (ELISA). Additionally, HMGB1 is not liver specific, so while it outperforms ALT (also not liver specific) as a biomarker for the prediction of DILI development, it should be used in a biomarker panel along with liver-specific markers such as miR-122.

FUTURE DIRECTIONS

A point-of-care test for HMGB1 and the development of redox and acetyl isoform-targeting antibodies will advance clinical utility. Work is ongoing to validate baseline levels of circulating HMGB1 in healthy volunteers.

Subtoxic Alterations in Hepatocyte-Derived Exosomes: An Early Step in Drug-Induced Liver Injury?

Drug-induced liver injury (DILI) is a significant clinical and economic problem in the United States, yet the mechanisms that underlie DILI remain poorly understood. Recent evidence suggests that signaling molecules released by stressed hepatocytes can trigger immune responses that may be common across DILI mechanisms. Extracellular vesicles released by hepatocytes, principally hepatocyte-derived exosomes (HDEs), may constitute one such signal. To examine HDE alterations as a function of drug-induced stress, this work utilized prototypical hepatotoxicant acetaminophen (APAP) in male Sprague-Dawley (SD) rats, SD rat hepatocytes, and primary human hepatocytes. HDE were isolated using ExoQuick precipitation reagent and analyzed by quantification of the liver-specific RNAs albumin and microRNA-122 (miR-122). In vivo, significant elevations in circulating exosomal albumin mRNA were observed at subtoxic APAP exposures. Significant increases in exosomal albumin mRNA were also observed in primary rat hepatocytes at subtoxic APAP concentrations. In primary human hepatocytes, APAP elicited increases in both exosomal albumin mRNA and exosomal miR-122 without overt cytotoxicity. However, the number of HDE produced in vitro in response to APAP did not increase with exosomal RNA quantity. We conclude that significant drug-induced alterations in the liver-specific RNA content of HDE occur at subtoxic APAP exposures in vivo and in vitro, and that these changes appear to reflect selective packaging rather than changes in exosome number. The current findings demonstrate that translationally relevant HDE alterations occur in the absence of overt hepatocellular toxicity, and support the hypothesis that HDE released by stressed hepatocytes may mediate early immune responses in DILI.

A two-photon fluorescent sensor revealing drug-induced liver injury via tracking γ-glutamyltranspeptidase (GGT) level in vivo

Currently drug-induced liver injury (DILI) has become a major and challenging public health issue in terms of medicine development and clinical therapy. The level of γ-glutamyl transpeptidase (GGT) has long been regarded as a preclinical/clinical biomarker for prediction of DILI. Herein, we report a two-photon fluorescent sensor for tracking GGT level changes resulted from DILI in vivo. The sensor was prepared by linking a glutamic acid to a dicyanomethylene-4H-pyran (DCM) derivative; and the presence of GGT cleaves γ-glutamyl amide group from the sensor and thereby restores the fluorescence emission (at 635 nm) of DCM moiety under femtosecond pulses at 800 nm. This two-photon sensor exhibits superior sensing performance such as red emission, high photostability and low detection limit (∼0.057 U/L). On a two-photon microscope, the sensor shows a bright red fluorescence in GGT-overexpressing A2780 cells; and it can fluorescently respond to the GGT generated in the liver of zebrafishes as a result of clinical drug (phenytoin) treatment. These findings demonstrate that a commonly-used clinical drug phenytoin can cause remarkable elevation in GGT level in liver, and this sensor may be useful as a marker to detect clinical drug-induced organ damages.

Circulating Kidney Injury Molecule 1 Predicts Prognosis and Poor Outcome in Patients With Acetaminophen-Induced Liver Injury

Acute kidney injury in the context of acetaminophen (APAP; paracetamol)-induced liver injury is an important predictor of the requirement for urgent liver transplantation (LT) to avoid death. However, the prognostic biomarker used to report kidney dysfunction (serum creatinine concentration) has suboptimal sensitivity and specificity. Kidney injury molecule 1 (KIM-1) can be quantified in plasma as a sensitive and specific biomarker of kidney injury in both clinical and preclinical studies. Therefore, plasma KIM-1 has potential as a sensitive prognostic biomarker of patient outcome post-APAP overdose. In a cohort of APAP overdose patients (N = 74) with and without established liver injury, we quantified plasma KIM-1 by immunoassay on the first day of admission to a LT unit and assessed its diagnostic performance to predict outcome compared with serum creatinine concentration. Day 1 plasma KIM-1 was significantly elevated in patients that died or required LT, compared to spontaneous survivors (1,182 ± 251 vs. 214 ± 45 pg/mL; P < 0.005). Receiver operator characteristic analysis demonstrated the superiority of KIM-1 (area under the curve [AUC]: 0.87; 95% confidence interval [CI]: 0.78-0.95; 0.56 sensitivity at 0.95 specificity), compared with serum creatinine (AUC, 0.76; 95% CI: 0.64-0.87; 0.08 sensitivity at 0.95 specificity) and other current prognostic indicators, when measured on the first day of enrollment into the study. Furthermore, KIM-1 was found to be a statistically significant independent predictor of outcome at the 5% level (P < 0.0386) in a multivariable logistic regression model that considered all measured factors (pseudo-R^2 = 0.895).

CONCLUSION

Early measurement of plasma KIM-1 represents a more sensitive predictor of patient outcome than serum creatinine concentration post-APAP overdose. With further development, plasma KIM-1 could significantly improve prognostic stratification.

Pathways involved in acetaminophen hepatotoxicity with specific targets for inhibition/downregulation

Insights on molecular/immunological mechanisms involve in APAP hepatotoxicity and pave way for researchers/clinicians/pharma bodies to identify novel biomarkers, effective bioactive candidates and fruitful therapy against APAP hepatotoxicity.

Biomarkers distinguish apoptotic and necrotic cell death during hepatic ischemia/reperfusion injury in mice

Hepatic ischemia/reperfusion (IRP) injury is a significant clinical problem during tumor-resection surgery (Pringle maneuver) and liver transplantation. However, the relative contribution of necrotic and apoptotic cell death to the overall liver injury is still controversial. To address this important issue with a standard murine model of hepatic IRP injury, plasma biomarkers of necrotic cell death such as micro-RNA 122, full-length cytokeratin 18 (FK18), and high-mobility group box 1 (HMGB1) protein and plasma biomarkers of apoptosis such as plasma caspase-3 activity and caspase-cleaved fragment of cytokeratin 18 (CK18) coupled with markers of inflammation (hyperacetylated HMGB1) were compared by histological features in hematoxylin and eosin-stained and terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling (TUNEL)-stained liver sections. After 45 minutes of hepatic ischemia and 1 to 24 hours of reperfusion, all necrosis markers increased dramatically in plasma by 40- to >10,000-fold over the baseline with a time course similar to that of alanine aminotransferase. These data correlated well with histological characteristics of necrosis. Within the area of necrosis, most cells were TUNEL positive; initially (≤3 hours of reperfusion), the staining was restricted to nuclei, but it later spread to the cytosol, and this is characteristic of karyorrhexis during necrotic cell death. In contrast, the lack of morphological evidence of apoptotic cell death and relevant caspase-3 activity in the postischemic liver correlated well with the absence of caspase-3 activity and CK18 (except for a minor increase at 3 hours of reperfusion) in plasma. A quantitative comparison of FK18 (necrosis) and CK18 (apoptosis) release indicated dominant cell death by necrosis during IRP and only a temporary and very minor degree of apoptosis. These data suggest that the focus of future research should be the elucidation of necrotic signaling mechanisms to identify relevant targets, which may be used to attenuate hepatic IRP injury.