Katarzyna R. Przybylak

In silico models for drug-induced liver injury – current status

Introduction: Drug-induced liver injury (DILI) is one of the most important reasons for drug attrition at both pre-approval and post-approval stages. Therefore, it is crucial to develop methods that will detect potential hepatotoxicity among drug candidates as early and quickly as possible. However, the complexity of hepatotoxicity endpoint makes it very difficult to predict. In addition, there is still a lack of sensitive and specific biomarkers for DILI that consequently leads to a scarcity of reliable hepatotoxic data, which are the key to any modelling approach. Areas covered: This review explores the current status of existing in silico models predicting hepatotoxicity. Over the past decade, attempts have been made to compile hepatotoxicity data and develop in silico models, which can be used as a first-line screening of drug candidates for further testing. Expert opinion: Most of the predictive methods discussed in this review are based on the structural properties of chemicals and do not take into account genetic and environmental factors; therefore, their predictions are still uncertain. To improve the predictability of in silico models for DILI, it is essential to better understand its mechanisms as well as to develop sensitive toxicogenomics biomarkers, which show relatively good differentiation between hepatotoxins and non-hepatotoxins.

Hepatotoxicity: A scheme for generating chemical categories for read-across, structural alerts and insights into mechanism(s) of action

Abstract The ability of a compound to cause adverse effects to the liver is one of the most common reasons for drug development failures and the withdrawal of drugs from the market. Such adverse effects can vary tremendously in severity, leading to an array of possible drug-induced liver injuries (DILIs). As a result, it is not surprising that drug development has evolved into a complex and multifaceted process including methods aiming to identify potential liver toxicities. Unfortunately, hepatotoxicity remains one of the most complex and poorly understood areas of human toxicity; thus it is a significant challenge to identify potential hepatotoxins. The performance of existing methods to identify hepatotoxicity requires improvement. The current study details a scheme for generating chemical categories and the development of structural alerts able to identify potential hepatotoxins. The study utilized a diverse 951-compound dataset and used structural similarity methods to produce a number of structurally restricted categories. From these categories, 16 structural alerts associated with observed human hepatotoxicity were developed. Furthermore, the mechanism(s) by which these compounds cause hepatotoxicity were investigated and a mechanistic rationale was proposed, where possible, to yield mechanistically supported structural alerts. Alerts of this nature have the potential to be used in the screening of compounds to highlight potential hepatotoxicity, whilst the chemical categories themselves are important in applying read-across approaches. The scheme presented in this study also has the potential to act as a knowledge generator serving as an excellent starting platform from which to conduct additional toxicological studies.

Assessing toxicological data quality: basic principles, existing schemes and current limitations

Existing toxicological data may be used for a variety of purposes such as hazard and risk assessment or toxicity prediction. The potential use of such data is, in part, dependent upon their quality. Consideration of data quality is of key importance with respect to the application of chemicals legislation such as REACH. Whether data are being used to make regulatory decisions or build computational models, the quality of the output is reflected by the quality of the data employed. Therefore, the need to assess data quality is an important requirement for making a decision or prediction with an appropriate level of confidence. This study considers the biological and chemical factors that may impact upon toxicological data quality and discusses the assessment of data quality. Four general quality criteria are introduced and existing data quality assessment schemes are discussed. Two case study datasets of skin sensitization data are assessed for quality providing a comparison of existing assessment methods. This study also discusses the limitations and difficulties encountered during quality assessment, including the use of differing quality schemes and the global versus chemical-specific assessments of quality. Finally, a number of recommendations are made to aid future data quality assessments.

Strategies for the optimisation of in vivo experiments in accordance with the 3Rs philosophy.

There are a large number of chemicals in current use for which adequate toxicity data are not available. Whilst there are clear ethical and legal obligations to obtain data from sources other than in vivo experiments wherever possible, in certain cases in vivo assays may be deemed necessary. In such circumstances, it is essential to ensure that the maximum amount of high quality data is obtained from the minimum number of animals, using the most humane procedures, in accordance with the philosophy of reduction, refinement and replacement (3Rs). The aim of this report is to provide a strategy for anyone involved in animal experimentation, for either toxicological or pharmacological purposes, as to how in vivo experiments may be optimised. The impact of generic and endpoint specific sources of variability has been highlighted in a proof-of-principle analysis considering the variation in protocols for assays for four human health endpoints (skin sensitisation, reproductive/developmental toxicity, mutagenicity and carcinogenicity). Other factors such as operator training, experimental/statistical design, use of lower species and use of combined assays are also discussed. Recommendations for optimisation of in vivo assays, in terms of the 3Rs philosophy, applied to performing tests, harvesting data and appropriate reporting are summarised as a checklist of issues to be addressed prior to undertaking such assays.

Towards a Fuzzy Expert System on Toxicological Data Quality Assessment

Quality assessment (QA) requires high levels of domain‐specific experience and knowledge. QA tasks for toxicological data are usually performed by human experts manually, although a number of quality evaluation schemes have been proposed in the literature. For instance, the most widely utilised Klimisch scheme 1 defines four data quality categories in order to tag data instances with respect to their qualities; ToxRTool 2 is an extension of the Klimisch approach aiming to increase the transparency and harmonisation of the approach. Note that the processes of QA in many other areas have been automatised by employing expert systems. Briefly, an expert system is a computer program that uses a knowledge base built upon human expertise, and an inference engine that mimics the reasoning processes of human experts to infer new statements from incoming data. In particular, expert systems have been extended to deal with the uncertainty of information by representing uncertain information (such as linguistic terms) as fuzzy sets under the framework of fuzzy set theory and performing inferences upon fuzzy sets according to fuzzy arithmetic. This paper presents an experimental fuzzy expert system for toxicological data QA which is developed on the basis of the Klimisch approach and the ToxRTool in an effort to illustrate the power of expert systems to toxicologists, and to examine if fuzzy expert systems are a viable solution for QA of toxicological data. Such direction still faces great difficulties due to the well‐known common challenge of toxicological data QA that “five toxicologists may have six opinions”. In the meantime, this challenge may offer an opportunity for expert systems because the construction and refinement of the knowledge base could be a converging process of different opinions which is of significant importance for regulatory policy making under the regulation of REACH, though a consensus may never be reached. Also, in order to facilitate the implementation of Weight of Evidence approaches and in silico modelling proposed by REACH, there is a higher appeal of numerical quality values than nominal (categorical) ones, where the proposed fuzzy expert system could help. Most importantly, the deriving processes of quality values generated in this way are fully transparent, and thus comprehensible, for final users, which is another vital point for policy making specified in REACH. Case studies have been conducted and this report not only shows the promise of the approach, but also demonstrates the difficulties of the approach and thus indicates areas for future development.

Mechanism-based QSAR modeling of skin sensitization

Many chemicals can induce skin sensitization, and there is a pressing need for non-animal methods to give a quantitative indication of potency. Using two large published data sets of skin sensitizers, we have allocated each sensitizing chemical to one of 10 mechanistic categories and then developed good QSAR models for the seven categories that have a sufficient number of chemicals to allow modeling. Both internal and external validation checks showed that each model had good predictivity.

How Does the Quality of Phospholipidosis Data Influence the Predictivity of Structural Alerts

The ability of drugs to induce phospholipidosis (PLD) is linked directly to their molecular substructures: hydrophobic, cyclic moieties with hydrophilic, peripheral amine groups. These structural properties can be captured and coded into SMILES arbitrary target specification (SMARTS) patterns. Such structural alerts, which are capable of identifying potential PLD inducers, should ideally be developed on a relatively large but reliable data set. We had previously developed a model based on SMARTS patterns consisting of 32 structural fragments using information from 450 chemicals. In the present study, additional PLD structural alerts have been developed based on a newer and larger data set combining two data sets published recently by the United States Food and Drug Administration (US FDA). To assess the predictive performance of the updated SMARTS model, two publicly available data sets were considered. These data sets were constructed using different criteria and hence represent different standards for overall quality. In the first data set high quality was assured as all negative chemicals were confirmed by the gold standard method for the detection of PLD-transmission electron microscopy (EM). The second data set was constructed from seven previously published data sets and then curated by removing compounds where conflicting results were found for PLD activity. Evaluation of the updated SMARTS model showed a strong, positive correlation between predictive performance of the alerts and the quality of the data set used for the assessment. The results of this study confirm the importance of using high quality data for modeling and evaluation, especially in the case of PLD, where species, tissue, and dose dependence of results are additional confounding factors.

In Silico Studies of the Relationship Between Chemical Structure and Drug Induced Phospholipidosis

Drug‐induced phospholipidosis (PLD) is a side effect of the administration of cationic amphiphilic drugs (CADs). It is desirable to identify and screen compounds with the potential to induce PLD as early as possible in drug development. Recently, a number of in silico methods have been developed to predict PLD. These models are low‐cost and high‐throughput strategies; however, they produce a high number of false positive predictions. The aim of this study was to assess the predictive performance of existing in silico approaches and to develop new strategies for the rapid identification of the potential PLD‐inducers. Studies on 450 chemicals confirmed the high false positive rate of prediction of models based only on log P and pKa values. Modification of the methods by incorporating structural information gave moderate improvements in the prediction performance. Therefore, a new strategy, based on molecular fragments captured by SMARTS strings was developed. These structural fragments were able to identify potential PLD‐inducers and achieved a high sensitivity of 85 %. The results showed that the phospholipidosis is linked directly to the molecular structure of chemical; therefore the SMARTS pattern methodology could be used as a first line of screening of PLD potential during the drug discovery process.

In Silico Models for Hepatotoxicity.

In this chapter we review the challenges of predicting human hepatotoxicity. Principally, this is our partial understanding of a very complex biochemical system and our ability to emulate that in a predictive capacity. We give an overview of the published modeling approaches in this area to date and discuss their design, strengths, and weaknesses. It is interesting to note the shift during the period of this review in the direction of evidenced-based approaches including structural alerts and pharmacophore models. Proposals on how best to utilize the data emerging from modeling studies are also discussed.

Assessment and Reproducibility of Quantitative Structure–Activity Relationship Models by the Nonexpert

Model reliability is generally assessed and reported as an intrinsic component of quantitative structure-activity relationship (QSAR) publications; it can be evaluated using defined quality criteria such as the Organisation for Economic Cooperation and Development (OECD) principles for the validation of QSARs. However, less emphasis is afforded to the assessment of model reproducibility, particularly by users who may wish to use model outcomes for decision making, but who are not QSAR experts. In this study we identified a range of QSARs in the area of absorption, distribution, metabolism, and elimination (ADME) prediction and assessed their adherence to the OECD principles, as well as investigating their reproducibility by scientists without expertise in QSAR. Here, 85 papers were reviewed, reporting over 80 models for 31 ADME-related endpoints. Of these, 12 models were identified that fulfilled at least 4 of the 5 OECD principles and 3 of these 12 could be readily reproduced. Published QSAR models should aim to meet a standard level of quality and be clearly communicated, ensuring their reproducibility, to progress the uptake of the models in both research and regulatory landscapes. A pragmatic workflow for implementing published QSAR models and recommendations to modellers, for publishing models with greater usability, are presented herein.