Drug-induced liver injury classification model based on in vitro human transcriptomics and in vivo rat clinical chemistry data

Identification of early liver toxicity gene biomarkers using comparative supervised machine learning

Screening agrochemicals and pharmaceuticals for potential liver toxicity is required for regulatory approval and is an expensive and time-consuming process. The identification and utilization of early exposure gene signatures and robust predictive models in regulatory toxicity testing has the potential to reduce time and costs substantially. In this study, comparative supervised machine learning approaches were applied to the rat liver TG-GATEs dataset to develop feature selection and predictive testing. We identified ten gene biomarkers using three different feature selection methods that predicted liver necrosis with high specificity and selectivity in an independent validation dataset from the Microarray Quality Control (MAQC)-II study. Nine of the ten genes that were selected with the supervised methods are involved in metabolism and detoxification (Car3, Crat, Cyp39a1, Dcd, Lbp, Scly, Slc23a1, and Tkfc) and transcriptional regulation (Ablim3). Several of these genes are also implicated in liver carcinogenesis, including Crat, Car3 and Slc23a1. Our biomarker gene signature provides high statistical accuracy and a manageable number of genes to study as indicators to potentially accelerate toxicity testing based on their ability to induce liver necrosis and, eventually, liver cancer.

Predicting Drug-Induced Liver Injury Using Ensemble Learning Methods and Molecular Fingerprints

Drug-induced liver injury (DILI) is a major safety concern in the drug-development process, and various methods have been proposed to predict the hepatotoxicity of compounds during the early stages of drug trials. In this study, we developed an ensemble model using 3 machine learning algorithms and 12 molecular fingerprints from a dataset containing 1241 diverse compounds. The ensemble model achieved an average accuracy of 71.1 ± 2.6%, sensitivity (SE) of 79.9 ± 3.6%, specificity (SP) of 60.3 ± 4.8%, and area under the receiver-operating characteristic curve (AUC) of 0.764 ± 0.026 in 5-fold cross-validation and an accuracy of 84.3%, SE of 86.9%, SP of 75.4%, and AUC of 0.904 in an external validation dataset of 286 compounds collected from the Liver Toxicity Knowledge Base. Compared with previous methods, the ensemble model achieved relatively high accuracy and SE. We also identified several substructures related to DILI. In addition, we provide a web server offering access to our models (http://ccsipb.lnu.edu.cn/toxicity/HepatoPred-EL/).

Gene Expression Data Based Deep Learning Model for Accurate Prediction of Drug-Induced Liver Injury in Advance

Drug-induced liver injury (DILI), one of the most common adverse effects, leads to drug development failure or withdrawal from the market in most cases, showing an emerging challenge that is to accurately predict DILI in the early stage. Recently, the vast amount of gene expression data provides us valuable information for distinguishing DILI on a genomic scale. Moreover, the deep learning algorithm is a powerful strategy to automatically learn important features from raw and noisy data and shows great success in the field of medical diagnosis. In this study, a gene expression data based deep learning model was developed to predict DILI in advance by using gene expression data associated with DILI collected from ArrayExpress and then optimized by feature gene selection and parameters optimization. In addition, the previous machine learning algorithm support vector machine (SVM) was also used to construct another prediction model based on the same data sets, comparing the model performance with the optimal DL model. Finally, the evaluation test using 198 randomly selected samples showed that the optimal DL model achieved 97.1% accuracy, 97.4% sensitivity, 96.8% specificity, 0.942 matthews correlation coefficient, and 0.989 area under the ROC curve, while the performance of SVM model only reached 88.9% accuracy, 78.8% sensitivity, 99.0% specificity, 0.794 matthews correlation coefficient, and 0.901 area under the ROC curve. Furthermore, external data sets verification and animal experiments were conducted to assess the optimal DL model performance. Finally, the predicted results of the optimal DL model were almost consistent with experiment results. These results indicated that our gene expression data based deep learning model could systematically and accurately predict DILI in advance. It could be a useful tool to provide safety information for drug discovery and clinical rational drug use in early stage and become an important part of drug safety assessment.

Mango phenolics increase the serum apolipoprotein A1/B ratio in rats fed high cholesterol and sodium cholate diets

BACKGROUND

Serum lipoproteins are in dynamic equilibrium, partially controlled by the apolipoprotein A1 to apolipoprotein B ratio (APOA1/APOB). Freeze-dried mango pulp (FDM) is a rich source of phenolic compounds (MP) and dietary fiber (MF), although their effects on lipoprotein metabolism have not yet been studied.

RESULTS

Thirty male Wistar rats were fed with four different isocaloric diets (3.4 kcal g^-1 ) for 12 weeks: control diet, high cholesterol (8 g kg^-1 ) + sodium cholate (2 g kg^-1 ) diet either alone or supplemented with MF (60 g kg^-1 ), MP (1 g kg^-1 ) or FDM (50 g kg^-1 ). MP and FDM reduced food intake, whereas MF and MP tended to increase serum APOA1/APOB ratio, independently of their hepatic gene expression. This suggests that lipoprotein metabolism was favorably altered by mango bioactives, MP also mitigated the non-alcoholic steatohepatitis that resulted from the intake of this diet.

CONCLUSION

We propose that phenolics are the most bioactive components of mango pulp, acting as anti-atherogenic and hepatoprotective agents, with a mechanism of action tentatively based on changes to the main protein components of lipoproteins. © 2018 Society of Chemical Industry.

Transcriptomics in toxicology

Xenobiotics, of which many are toxic, may enter the human body through multiple routes. Excessive human exposure to xenobiotics may exceed the body's capacity to defend against the xenobiotic-induced toxicity and result in potentially fatal adverse health effects. Prevention of the adverse health effects, potentially associated with human exposure to the xenobiotics, may be achieved by detecting the toxic effects at an early, reversible and, therefore, preventable stage. Additionally, an understanding of the molecular mechanisms underlying the toxicity may be helpful in preventing and/or managing the ensuing adverse health effects. Human exposures to a large number of xenobiotics are associated with hepatotoxicity or pulmonary toxicity. Global gene expression changes taking place in biological systems, in response to exposure to xenobiotics, may represent the early and mechanistically relevant cellular events contributing to the onset and progression of xenobiotic-induced adverse health outcomes. Hepatotoxicity and pulmonary toxicity resulting from exposure to xenobiotics are discussed as specific examples to demonstrate the potential application of transcriptomics or global gene expression analysis in the prevention of adverse health effects associated with exposure to xenobiotics.

Prediction models for drug-induced hepatotoxicity by using weighted molecular fingerprints

Background

Drug-induced liver injury (DILI) is a critical issue in drug development because DILI causes failures in clinical trials and the withdrawal of approved drugs from the market. There have been many attempts to predict the risk of DILI based on in vivo and in silico identification of hepatotoxic compounds. In the current study, we propose the in silico prediction model predicting DILI using weighted molecular fingerprints.

Results

In this study, we used 881 bits of molecular fingerprint and used as features describing presence or absence of each substructure of compounds. Then, the Bayesian probability of each substructure was calculated and labeled (positive or negative for DILI), and a weighted fingerprint was determined from the ratio of DILI-positive to DILI-negative probability values. Using weighted fingerprint features, the prediction models were trained and evaluated with the Random Forest (RF) and Support Vector Machine (SVM) algorithms. The constructed models yielded accuracies of 73.8% and 72.6%, AUCs of 0.791 and 0.768 in cross-validation. In independent tests, models achieved accuracies of 60.1% and 61.1% for RF and SVM, respectively. The results validated that weighted features helped increase overall performance of prediction models. The constructed models were further applied to the prediction of natural compounds in herbs to identify DILI potential, and 13,996 unique herbal compounds were predicted as DILI-positive with the SVM model.

Conclusions

The prediction models with weighted features increased the performance compared to non-weighted models. Moreover, we predicted the DILI potential of herbs with the best performed model, and the prediction results suggest that many herbal compounds could have potential to be DILI. We can thus infer that taking natural products without detailed references about the relevant pathways may be dangerous. Considering the frequency of use of compounds in natural herbs and their increased application in drug development, DILI labeling would be very important.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-017-1638-4) contains supplementary material, which is available to authorized users.

Predicting drug-induced liver injury in human with Naïve Bayes classifier approach

Drug-induced liver injury (DILI) is one of the major safety concerns in drug development. Although various toxicological studies assessing DILI risk have been developed, these methods were not sufficient in predicting DILI in humans. Thus, developing new tools and approaches to better predict DILI risk in humans has become an important and urgent task. In this study, we aimed to develop a computational model for assessment of the DILI risk with using a larger scale human dataset and Naïve Bayes classifier. The established Naïve Bayes prediction model was evaluated by 5-fold cross validation and an external test set. For the training set, the overall prediction accuracy of the 5-fold cross validation was 94.0 %. The sensitivity, specificity, positive predictive value and negative predictive value were 97.1, 89.2, 93.5 and 95.1 %, respectively. The test set with the concordance of 72.6 %, sensitivity of 72.5 %, specificity of 72.7 %, positive predictive value of 80.4 %, negative predictive value of 63.2 %. Furthermore, some important molecular descriptors related to DILI risk and some toxic/non-toxic fragments were identified. Thus, we hope the prediction model established here would be employed for the assessment of human DILI risk, and the obtained molecular descriptors and substructures should be taken into consideration in the design of new candidate compounds to help medicinal chemists rationally select the chemicals with the best prospects to be effective and safe.

Key Challenges and Opportunities Associated with the Use of In Vitro Models to Detect Human DILI: Integrated Risk Assessment and Mitigation Plans

Drug-induced liver injury (DILI) is a major cause of late-stage clinical drug attrition, market withdrawal, black-box warnings, and acute liver failure. Consequently, it has been an area of focus for toxicologists and clinicians for several decades. In spite of considerable efforts, limited improvements in DILI prediction have been made and efforts to improve existing preclinical models or develop new test systems remain a high priority. While prediction of intrinsic DILI has improved, identifying compounds with a risk for idiosyncratic DILI (iDILI) remains extremely challenging because of the lack of a clear mechanistic understanding and the multifactorial pathogenesis of idiosyncratic drug reactions. Well-defined clinical diagnostic criteria and risk factors are also missing. This paper summarizes key data interpretation challenges, practical considerations, model limitations, and the need for an integrated risk assessment. As demonstrated through selected initiatives to address other types of toxicities, opportunities exist however for improvement, especially through better concerted efforts at harmonization of current, emerging and novel in vitro systems or through the establishment of strategies for implementation of preclinical DILI models across the pharmaceutical industry. Perspectives on the incorporation of newer technologies and the value of precompetitive consortia to identify useful practices are also discussed.

Exploiting microRNA and mRNA profiles generated in vitro from carcinogen-exposed primary mouse hepatocytes for predicting in vivo genotoxicity and carcinogenicity

The well-defined battery of in vitro systems applied within chemical cancer risk assessment is often characterised by a high false-positive rate, thus repeatedly failing to correctly predict the in vivo genotoxic and carcinogenic properties of test compounds. Toxicogenomics, i.e. mRNA-profiling, has been proven successful in improving the prediction of genotoxicity in vivo and the understanding of underlying mechanisms. Recently, microRNAs have been discovered as post-transcriptional regulators of mRNAs. It is thus hypothesised that using microRNA response-patterns may further improve current prediction methods. This study aimed at predicting genotoxicity and non-genotoxic carcinogenicity in vivo, by comparing microRNA- and mRNA-based profiles, using a frequently applied in vitro liver model and exposing this to a range of well-chosen prototypical carcinogens. Primary mouse hepatocytes (PMH) were treated for 24 and 48h with 21 chemical compounds [genotoxins (GTX) vs. non-genotoxins (NGTX) and non-genotoxic carcinogens (NGTX-C) versus non-carcinogens (NC)]. MicroRNA and mRNA expression changes were analysed by means of Exiqon and Affymetrix microarray-platforms, respectively. Classification was performed by using Prediction Analysis for Microarrays (PAM). Compounds were randomly assigned to training and validation sets (repeated 10 times). Before prediction analysis, pre-selection of microRNAs and mRNAs was performed by using a leave-one-out t-test. No microRNAs could be identified that accurately predicted genotoxicity or non-genotoxic carcinogenicity in vivo. However, mRNAs could be detected which appeared reliable in predicting genotoxicity in vivo after 24h (7 genes) and 48h (2 genes) of exposure (accuracy: 90% and 93%, sensitivity: 65% and 75%, specificity: 100% and 100%). Tributylinoxide and para-Cresidine were misclassified. Also, mRNAs were identified capable of classifying NGTX-C after 24h (5 genes) as well as after 48h (3 genes) of treatment (accuracy: 78% and 88%, sensitivity: 83% and 83%, specificity: 75% and 93%). Wy-14,643, phenobarbital and ampicillin trihydrate were misclassified. We conclude that genotoxicity and non-genotoxic carcinogenicity probably cannot be accurately predicted based on microRNA profiles. Overall, transcript-based prediction analyses appeared to clearly outperform microRNA-based analyses.