Mechanisms of Chemical Carcinogenicity and Mutagenicity: A Review with Implications for Predictive Toxicology

Automated detection of structural alerts (chemical fragments) in (eco)toxicology

This mini-review describes the evolution of different algorithms dedicated to the automated discovery of chemical fragments associated to (eco)toxicological endpoints. These structural alerts correspond to one of the most interesting approach of in silico toxicology due to their direct link with specific toxicological mechanisms. A number of expert systems are already available but, since the first work in this field which considered a binomial distribution of chemical fragments between two datasets, new data miners were developed and applied with success in chemoinformatics. The frequency of a chemical fragment in a dataset is often at the core of the process for the definition of its toxicological relevance. However, recent progresses in data mining provide new insights into the automated discovery of new rules. Particularly, this review highlights the notion of Emerging Patterns that can capture contrasts between classes of data.

Predicting chemical toxicity effects based on chemical-chemical interactions

Toxicity is a major contributor to high attrition rates of new chemical entities in drug discoveries. In this study, an order-classifier was built to predict a series of toxic effects based on data concerning chemical-chemical interactions under the assumption that interactive compounds are more likely to share similar toxicity profiles. According to their interaction confidence scores, the order from the most likely toxicity to the least was obtained for each compound. Ten test groups, each of them containing one training dataset and one test dataset, were constructed from a benchmark dataset consisting of 17,233 compounds. By a Jackknife test on each of these test groups, the 1^st order prediction accuracies of the training dataset and the test dataset were all approximately 79.50%, substantially higher than the rate of 25.43% achieved by random guesses. Encouraged by the promising results, we expect that our method will become a useful tool in screening out drugs with high toxicity.

Improvement of carcinogenicity prediction performances based on sensitivity analysis in variable selection of SVM models

A new sensitivity analysis (SA) method for variable selection in support vector machine (SVM) was proposed to improve the performance level of the QSAR model to predict carcinogenicity based on the correlation coefficient (CC) method used in our preceding study. The performances of both methods were also compared with that of the F-score (FS) method proposed by Chang and Lin. The 911 non-congeneric chemicals were classified into 20 mutually overlapping groups according to contained substructures, and a specific SVM model created on chemicals belonging to each group was optimized by searching the best set of SVM parameters while successively omitting descriptors of lower absolute values of sensitivity, CC or FS until the maximum predictive performance was obtained. The SA method improves the overall accuracy from 80% of CC and FS to 84%, which is considerably higher than those of existing models for predicting the carcinogenicity of non-congeneric chemicals. It selects the optimum sets of effective descriptors fewer than the CC and FS methods, and is not time-consuming and can be applied to a large set of initial descriptors. It is concluded that SA is superior as a variable selection method in SVM models.

Screening reactive metabolites bioactivated by multiple enzyme pathways using a multiplexed microfluidic system

A multiplexed, microfluidic platform to detect reactive metabolites is described, and its performance is illustrated for compounds metabolized by oxidative and bioconjugation enzymes in multi-enzyme pathways to mimic natural human drug metabolism. The device features four 8-electrode screen printed carbon arrays coated with thin films of DNA, a ruthenium-polyvinylpyridine (RuPVP) catalyst, and multiple enzyme sources including human liver microsomes (HLM), cytochrome P450 (cyt P450) 1B1 supersomes, microsomal epoxide hydrolase (EH), human S9 liver fractions (Hs9) and N-acetyltransferase (NAT). Arrays are arranged in parallel to facilitate multiple compound screening, enabling up to 32 enzyme reactions and measurements in 20-30 min. In the first step of the assay, metabolic reactions are achieved under constant flow of oxygenated reactant solutions by electrode driven natural catalytic cycles of cyt P450s and cofactor-supported bioconjugation enzymes. Reactive metabolites formed in the enzyme reactions can react with DNA. Relative DNA damage is measured in the second assay step using square wave voltammetry (SWV) with RuPVP as catalyst. Studies were done on chemicals known to require metabolic activation to induce genotoxicity, and results reproduced known features of metabolite DNA-reactivity for the test compounds. Metabolism of benzo[a]pyrene (B[a]P) by cyt P450s and epoxide hydrolase showed an enhanced relative DNA damage rate for DNA compared to cyt P450s alone. DNA damage rates for arylamines by pathways featuring both oxidative and conjugative enzymes at pH 7.4 gave better correlation with rodent genotoxicity metric TD(50). Results illustrate the broad utility of the reactive metabolite screening device.

Evaluation of the toxicity forecasting capability of EPA's ToxCast Phase I data: can ToxCast in vitro assays predict carcinogenicity?

Long-term rodent bioassays have played a central role in protecting human health from carcinogens; for ethical and practical reasons their use is decreasing whereas genotoxicity testing has taken a pivotal role. However, this strategy--as presently implemented--is not sensitive enough to detect all genotoxic carcinogens, and cannot detect nongenotoxic carcinogens. Among the alternative approaches under study there is the ToxCast/Tox21 project. Following a previous study from our laboratory, here we present a new, more extensive analysis of ToxCast Phase I results, indicating that at the present state-of-art this approach is not able to predict the carcinogenicity of chemicals. Possible reasons for this mediocre performance are discussed, and opinions on ways to tune up the project in the next phases are presented.

Comparison of in silico models for prediction of mutagenicity

Using a dataset with more than 6000 compounds, the performance of eight quantitative structure activity relationships (QSAR) models was evaluated: ACD/Tox Suite, Absorption, Distribution, Metabolism, Elimination, and Toxicity of chemical substances (ADMET) predictor, Derek, Toxicity Estimation Software Tool (T.E.S.T.), TOxicity Prediction by Komputer Assisted Technology (TOPKAT), Toxtree, CEASAR, and SARpy (SAR in python). In general, the results showed a high level of performance. To have a realistic estimate of the predictive ability, the results for chemicals inside and outside the training set for each model were considered. The effect of applicability domain tools (when available) on the prediction accuracy was also evaluated. The predictive tools included QSAR models, knowledge-based systems, and a combination of both methods. Models based on statistical QSAR methods gave better results.

Studies on the interaction mechanism of pyrene derivatives with human tumor-related DNA

Pyrene derivatives can be carcinogenic, teratogenic and mutagenic, thus having the potential to cause malignant diseases. In this work, the interactions of two selected pyrene derivatives (1-OHP and 1-PBO) and human tumor-related DNA (p53 DNA and C-myc DNA) are investigated by spectroscopic and non-native polyacrylamide gel electrophoresis (PAGE) methods. Using fluorescence spectrometry and circular dichroism (CD), DNA interactions of pyrene derivatives are confirmed to occur mainly via the groove binding mode supported by the intercalation into the base pairs of DNA. There is an obvious binding order of pyrene derivatives to the targeted DNA, 1-OHP > 1-PBO. The binding constants of 1-OHP are 1.16 × 10⁶ L·mol⁻¹ and 4.04 × 10⁵ L·mol⁻¹ for p53 DNA and C-myc DNA, respectively, while that of 1-PBO are only 2.04 × 10³ L·mol⁻¹ and 1.39 × 10³ L·mol⁻¹ for p53 DNA and C-myc DNA, respectively. Besides, the binding of pyrene derivatives to p53 DNA is stronger than that for C-myc DNA. CD and PAGE results indicate that the binding of pyrene derivatives can affect the helical structures of DNA and further induce the formation of double-chain antiparallel G-quadruplex DNA of hybrid G-rich sequences.

Density functional study on the cytochrome-mediated S-oxidation: identification of crucial reactive intermediate on the metabolic path of thiazolidinediones

S-Oxidation is an important cytochrome P450 (CYP450)-catalyzed reaction, and the structural and energetic details of this process can only be studied by using quantum chemical methods. Thiazolidinedione (TZD) ring metabolism involving initial S-oxidation leads to the generation of reactive metabolites (RMs) and subsequent toxicity forcing the withdrawal of the glitazone class of drugs, thus, the study of the biochemical pathway of TZD ring metabolism is a subject of interest. The S-oxidation of the TZD ring and the formation of the isocyanate intermediate (ISC) was implicated as a possible pathway; however, there are several questions still unanswered in this biochemical pathway. The current study focuses on the CYP450-mediated S-oxidation, fate of the sulfoxide product (TZDSO), ring cleavage to ISC, and formation of nucleophilic adducts. The process of S-oxidation was explored by using Cpd I (iron(IV)-oxo porphyrin, to mimic CYP450) at TZVP/6-311+G(d) basis set. The barriers were calculated after incorporating dispersion and solvent corrections. The metabolic conversion from TZDSO to ISC (studied at B3LYP/6-311++G(2df,3pd)//B3LYP/6-31+G(d)) required a novel protonated intermediate, TZDSOH(+). The effect of higher basis sets (6-311+G(d,p), aug-cc-pvqz) on this conversion was studied. TZDSOH(+) was observed to be more reactive and thermodynamically accessible than ISC, indicating that TZDSOH(+) is the actual reactive intermediate leading to toxicity of the TZD class of compounds.

Activity, distribution and function of indole-3-acetic acid biosynthetic pathways in bacteria

The capacity to produce the phytohormone indole-3-acetic acid (IAA) is widespread among bacteria that inhabit diverse environments such as soils, fresh and marine waters, and plant and animal hosts. Three major pathways for bacterial IAA synthesis have been characterized that remove the amino and carboxyl groups from the α-carbon of tryptophan via the intermediates indolepyruvate, indoleacetamide, or indoleacetonitrile; the oxidized end product IAA is typically secreted. The enzymes in these pathways often catabolize a broad range of substrates including aromatic amino acids and in some cases the branched chain amino acids. Moreover, expression of some of the genes encoding key IAA biosynthetic enzymes is induced by all three aromatic amino acids. The broad distribution and substrate specificity of the enzymes suggests a role for these pathways beyond plant-microbe interactions in which bacterial IAA has been best studied.

Strategies for the generation, validation and application of in silico ADMET models in lead generation and optimization

INTRODUCTION

The most desirable chemical starting point in drug discovery is a hit or lead with a good overall profile, and where there may be issues; a clear SAR strategy should be identifiable to minimize the issue. Filtering based on drug-likeness concepts are a first step, but more accurate theoretical methods are needed to i) estimate the biological profile of molecule in question and ii) based on the underlying structure-activity relationships used by the model, estimate whether it is likely that the molecule in question can be altered to remove these liabilities.

AREAS COVERED

In this paper, the authors discuss the generation of ADMET models and their practical use in decision making. They discuss the issues surrounding data collation, experimental errors, the model assessment and validation steps, as well as the different types of descriptors and statistical models that can be used. This is followed by a discussion on how the model accuracy will dictate when and where it can be used in the drug discovery process. The authors also discuss how models can be developed to more effectively enable multiple parameter optimization.

EXPERT OPINION

Models can be applied in lead generation and lead optimization steps to i) rank order a collection of hits, ii) prioritize the experimental assays needed for different hit series, iii) assess the likelihood of resolving a problem that might be present in a particular series in lead optimization and iv) screen a virtual library based on a hit or lead series to assess the impact of diverse structural changes on the predicted properties.

Established and emerging trends in computational drug discovery in the structural genomics era

Bioinformatics and chemoinformatics approaches contribute to hit discovery, hit-to-lead optimization, safety profiling, and target identification and enhance our overall understanding of the health and disease states. A vast repertoire of computational methods has been reported and increasingly combined in order to address more and more challenging targets or complex molecular mechanisms in the context of large-scale integration of structure and bioactivity data produced by private and public drug research. This review explores some key computational methods directly linked to drug discovery and chemical biology with a special emphasis on compound collection preparation, virtual screening, protein docking, and systems pharmacology. A list of generally freely available software packages and online resources is provided, and examples of successful applications are briefly commented upon.

Timing of carcinogenicity studies and predictability of genotoxicity for tumorigenicity in anti-HIV drug development

The timing of carcinogenicity studies in parallel with the clinical development of anti-human immunodeficiency virus (HIV) drugs has been flexible for most cases in the past. This includes postponement of the initiation of the studies and submission of final audited reports to the US Food and Drug Administration (FDA) for a new drug application (NDA) approval. We address this regulatory practice for anti-HIV drugs for which, in the past, there had been no effective treatment. We also examine the correlation of genotoxicity data with carcinogenicity data for the varied subclasses of anti-HIV drugs. We suggest that this regulatory policy regarding the timing of carcinogenicity testing does not compromise the safety standards of FDA's drug evaluation and the approval process. The policy does facilitate availability of these agents to meet the medical needs of the target population. Our analysis on the profile of carcinogenicity findings of anti-HIV drugs shows trends of class effects. Additionally, both carcinogenicity and genotoxicity data show significant correlations, which provide useful insights into issues involving these 2 important areas of toxicological investigations.

Paradigm shift in toxicity testing and modeling

The limitations of traditional toxicity testing characterized by high-cost animal models with low-throughput readouts, inconsistent responses, ethical issues, and extrapolability to humans call for alternative strategies for chemical risk assessment. A new strategy using in vitro human cell-based assays has been designed to identify key toxicity pathways and molecular mechanisms leading to the prediction of an in vivo response. The emergence of quantitative high-throughput screening (qHTS) technology has proved to be an efficient way to decompose complex toxicological end points to specific pathways of targeted organs. In addition, qHTS has made a significant impact on computational toxicology in two aspects. First, the ease of mechanism of action identification brought about by in vitro assays has enhanced the simplicity and effectiveness of machine learning, and second, the high-throughput nature and high reproducibility of qHTS have greatly improved the data quality and increased the quantity of training datasets available for predictive model construction. In this review, the benefits of qHTS routinely used in the US Tox21 program will be highlighted. Quantitative structure-activity relationships models built on traditional in vivo data and new qHTS data will be compared and analyzed. In conjunction with the transition from the pilot phase to the production phase of the Tox21 program, more qHTS data will be made available that will enrich the data pool for predictive toxicology. It is perceivable that new in silico toxicity models based on high-quality qHTS data will achieve unprecedented reliability and robustness, thus becoming a valuable tool for risk assessment and drug discovery.

Alternatives to the carcinogenicity bioassay for toxicity prediction: are we there yet?

INTRODUCTION

For decades, traditional toxicology has been the ultimate source of information on the carcinogenic potential of chemicals; however, with increasing demand on regulation of chemicals and decreasing resources for testing, opportunities to accept 'alternative' approaches have dramatically expanded. The need for tools able to identify carcinogens in shorter times and at a lower cost in terms of animal lives and money is still an open issue, and the present strategies and regulations for carcinogenicity prescreening do not adequately protect human health.

AREAS COVERED

This paper briefly summarizes the theories on the early steps of carcinogenesis and presents alternative detection methods for carcinogens based on genetic toxicology, structure-activity relationships and cell transformation assays.

EXPERT OPINION

There is evidence that the combination of Salmonella and structural alerts for the DNA-reactive carcinogens, and in vitro cell transformation assays for nongenotoxic carcinogens, permits the identification of a very large proportion of carcinogens. If implemented, this alternative strategy could improve considerably the protection of human health.

Assessment and validation of US EPA's OncoLogic® expert system and analysis of its modulating factors for structural alerts

OncoLogic® is a software program able to screen chemical compounds for toxicological effects. The software predicts the potential carcinogenicity of chemicals by applying rules of structure activity relationship (SAR) analysis. To validate the predictivity of OncoLogic® (Version 7.0), 123 compounds tested with the long-term carcinogenicity bioassay on rodents were extracted from the ISSCAN database and were analyzed. The concordance between the OncoLogic® SAR analysis and the bioassay results was high. To better understand the strength of the SAR science in OncoLogic®, we investigated the influence of a select group of modulating factors on the predictions by the structural alerts.

Integration of QSAR and SAR methods for the mechanistic interpretation of predictive models for carcinogenicity

The knowledge-based Toxtree expert system (SAR approach) was integrated with the statistically based counter propagation artificial neural network (CP ANN) model (QSAR approach) to contribute to a better mechanistic understanding of a carcinogenicity model for non-congeneric chemicals using Dragon descriptors and carcinogenic potency for rats as a response. The transparency of the CP ANN algorithm was demonstrated using intrinsic mapping technique specifically Kohonen maps. Chemical structures were represented by Dragon descriptors that express the structural and electronic features of molecules such as their shape and electronic surrounding related to reactivity of molecules. It was illustrated how the descriptors are correlated with particular structural alerts (SAs) for carcinogenicity with recognized mechanistic link to carcinogenic activity. Moreover, the Kohonen mapping technique enables one to examine the separation of carcinogens and non-carcinogens (for rats) within a family of chemicals with a particular SA for carcinogenicity. The mechanistic interpretation of models is important for the evaluation of safety of chemicals.

Flexible use of QSAR models in predictive toxicology: a case study on aromatic amines

Over the last years, predictive toxicology approaches based on Structure-Activity Relationships have emerged as fundamental tools in the regulatory assessments of chemicals, especially in those programs where regulatory constraints and assessment schemes limit the amount of data available from experimental test methods. Both the qualitative (e.g., Structural Alerts) and the quantitative (Quantitative Structure-Activity Relationships, QSAR) approach can play important roles. However, the two approaches are not familiar to the same extent to the regulators that most often use only the qualitative approach, so that the potentiality of the more sophisticated QSAR approach is neglected. In fact, QSAR is a very flexible tool that allows the user to modulate its response according to different goals and requirements. Here, we present a non-naïve approach to the use of a QSAR relative to a dichotomous biological activity (such as mutagen/nonmutagen), and we show how the user can maximize alternatively the reliability of the prediction of negative compounds (i.e., safe chemicals) or that of positive chemicals (i.e., chemicals that pose high hazard). Because of the environmental and industrial importance of the class of aromatic amines, we apply the approach to a previously published QSAR on the Salmonella typhimurium mutagenicity of these chemicals.

Dendritic antioxidants with pyrazole as the core: ability to scavenge radicals and to protect DNA

Chalcones with or without a para-hydroxyl group were condensed with phenylhydrazine-related compounds to form 1,3,5-triphenyl-1H-pyrazole (TPP), 4-(1,5-diphenyl-1H-pyrazol-3-yl)phenol (APP), 4-(1,3-diphenyl-1H-pyrazol-5-yl)phenol (BPP), and 4-(3,5-diphenyl-1H-pyrazol-1-yl)phenol (CPP), in which the phenyl group formed a dendritic structure with pyrazole as the core. Thus, the aim of this work was to explore the antioxidant capacities of TPP, APP, BPP, and CPP in trapping 2,2'-azinobis(3-ethylbenzothiazoline-6-sulfonate) cationic radical (ABTS(+•)) and 2,2'-diphenyl-1-picrylhydrazyl radical (DPPH) and in inhibiting Cu(2+)/glutathione (GSH)-, (•)OH-, and 2,2'-azobis(2-amidinopropane hydrochloride) (AAPH)-induced oxidation of DNA. TPP can react with ABTS(+•) and DPPH, indicating that the N atom in pyrazole possesses radical-scavenging ability. Moreover, APP, BPP, and CPP can trap 1.71, 1.81, and 1.58 radicals, respectively, in protecting DNA against AAPH-induced oxidation. Thus, the combination of pyrazole with a phenyl group exerted antioxidant ability although only one phenolic hydroxyl group was involved. However, these compounds showed weak protective effect against Cu(2+)/GSH-induced oxidation of DNA and even a pro-oxidant effect on (•)OH-induced oxidation of DNA.

Comparison and possible use of in silico tools for carcinogenicity within REACH legislation

Seven in silico models have been used to assess the prediction accuracy of chemical compound carcinogenicity. More than 1500 compounds with experimental values have been used to evaluate the models. Here we review the application of these models for toxicity prediction and their advantages and disadvantages, discussing the different approaches underlying the models and their main critical points. Some models have fewer false negatives while others are better at avoiding false positives. Since carcinogenicity is typically evaluated using a series of studies, identification of a strategy using one, or preferably a battery of in silico models, could reduce the number of animal studies needed.