OASIS Suite for Human Health Endpoints
METABOLISM ACTIVATED TOXIC EFFECTS
The role of metabolism in prioritizing chemicals according to their toxicity and adverse health outcomes appears to be extremely important given the fact that innocuous parents could get transformed into toxic metabolites.
TIMES (TIssue MEtabolism Simulator) is a heuristic algorithm to generate plausible metabolic maps from a comprehensive library of biotransformations and abiotic reactions. It allows prioritization of chemicals according to toxicity of their metabolites. The list of transformations is prioritized based on estimated system-specific probabilities of occurrence of these transformations. The probabilistic prioritization of individual transformations allows better control of the propagation of metabolic maps. Thus, propagation of metabolic maps can end up at low probable metabolites, i.e., weak branches could be pruned. The metabolic maps can also be confined at highly soluble or excretable metabolites or by other more trivial constraints such as maximum tree depth, number of rival pathways or avoiding cyclic processes, such as redox cycling. Quantitative evaluation of the performance of biotransformations allows also prioritisation of metabolites according to their amount, reactivity, solubility, toxicity, etc., which solves one of the major troublesome in modelling metabolism and namely the huge number of predicted metabolites.
The ability of TIMES to predict on same platform the metabolism of chemicals and toxicity resulting from their metabolic activation is an important advantage of the method. Presently, TIMES platform is used to predict the following metabolism activated endpoints:
· Skin sensitization - combining skin metabolism simulator and reactivity model for protein binding
· AMES Mutagenicity - combining S9 liver metabolism simulator and reactivity model for DNA binding. Besides the model specifically for TA100, a model of general (across strains) mutagenicity are available.
· Chromosomal aberration - combining S9 liver metabolism simulator and reactivity model for DNA and protein binding.
· Receptor mediated endpoints – combining metabolic activation of chemicals in S9 liver and models for binding affinity with ER, AR, AhR.
Skin sensitization
The skin sensitization model was derived from a data set compiled from more than 800 chemicals tested in the LLNA, GPMT as well as from the BfR list. The basic set of the training chemicals are LLNA data, hence, the model simulate predominantly LLNA. In case, no LLNA data are available GPMT are used. Finally, when no LLNA and GPMT are available BfR activity data are used (based on expert) classification. The chemicals were scaled into three categories: chemicals with strong (EC3<10%), weak (10%<EC3<50%) or non-sensitizing effect (EC3>50%). Thus, within the used scale as “Strong sensitizers” are classified chemicals that are either extreme, strong or moderate sensitizers.
TIMES-SS (TIssue MEtabolism Simulator for Skin Sensitisation) is an expert system describing structure-toxicity and structure-metabolism relationships through a number of transformations simulating skin metabolism and interaction of generated reactive metabolites with skin proteins (Dimitrov et al, 2005a). The skin metabolism simulator mimics the molecular transformation which could be observed in skin. Metabolic pathways are generated based on a set of hierarchically ordered principle transformations including spontaneous reactions and enzyme-catalyzed biotransformation reactions (phase I and II). The covalent reactions with proteins are described by a list of alerting groups. The associated mechanisms are in accordance with the existing knowledge on electrophilic interaction mechanisms of various structural functionalities. For some alerting groups, such as aldehydes and conjugated double bonds systems with electron withdrawing groups, a specific common reactivity pattern (COREPA) approache were utilized to determine stereo-electronic characteristics that might enhance or inhibit activity. The COREPA model is derived in the form of a decision tree. Its logic boxes consist of decision rules based on the reactivity patterns described by a combination of global descriptors of molecular steric and electronic structure and local reactivity parameters associated with specific alerting groups. During metabolism simulation, chemicals having the specified alerts (aldehydes and conjugated double bonds systems with electron withdrawing groups) are screened by the COREPA models to determine the level of activation of the alerting groups by the rest of the molecules and subsequently their ability to interact with protein nucleophiles.
The system is mechanistically transparent providing to user enough metabolic and reactivity information to support his expert evaluation.
TIMES-SS utilizes also a three-stage applicability domain (AD) (Dimitrov et al, 2005b). The first stage comprises a global requirements domain with simple cut-offs for physicochemical parameters such as log Kow, molecular weight and water solubility. TIMES-SS specifically includes upper and lower thresholds for log Kow and MW based on the training set information.
The next stage of determining the AD consists of defining the structural domain which is used to gauge the extent to which a chemical is structurally similar to those in the training set of chemicals for which the model performs correctly or incorrectly (within a user defined accuracy threshold). The structural domain could be characterized by the atom centered fragments of these chemicals.
The mechanistic domain comprises two sub-domains: transformation performance and interpolation space. The first explores the domain of functional or reactive groups, i.e., the performance of the alerts with respect to the interactions with protein. The second stage holds only for chemicals for which an additional COREPA model is required. It estimates the position of the target chemicals in the population density plot built in the parametric space defined by the explanatory variables of the model by making use the training set chemicals.
An overall call for the domain status is reported as the total domain. If one of the domains fails for a given chemical – the overall outcome is reported as “outside of domain”.
The advantage of processing query chemicals through all stages of the applicability domain is the increased reliability of prediction for those chemicals that satisfy all conditions for inclusion in the AD. The cost of applying this rigorous approach is that the number of chemicals for which reliable predictions are eventually made is reduced but this increases confidence in reliability of the final prediction.
References:
AMES Mutagenicity
The reactivity model describing interactions of chemicals with DNA is based on an alerting group approach. Only those toxicophores extracted from the training set (almost 2800 chemicals) having clear interpretation for the molecular mechanism causing the ultimate effect were included in the model. The consideration of structural alerts as necessary attributes for eliciting a single molecular interaction mechanism allowed the training data set to be segmented according to independent interaction mechanisms that eventually cause a mutagenic effect. Subsequently, the COREPA approach was applied to each subset of chemicals associated with specific alerts to identify concomitant molecular parameters defining the degree of activation of the alert by the rest of the molecule.
The derived models are combined with metabolic simulator used for predicting metabolic activation of chemicals with the S9 mix. When a new chemical is submitted for predicting mutagenicity, first the COREPA model is used to screen the parent structure. Independent of the prediction for the parent chemical all generated metabolites are submitted for screening using the same COREPA model. The user is then informed about the mutagenicity of the parent structure and its metabolites. Mutagenicity could be due to the parent chemical only or as a result of its metabolic activation (i.e., the parent is inactive but it is transformed to a mutagenic metabolite), or both parent structures and metabolites could be mutagenic.
References:
Serafimova R., Todorov M., Pavlov T., Kotov S., Jacob E., Aptula N., Mekenyan O.. Identification of the structural requirements for mutagencitiy, by incorporating molecular flexibility and metabolic activation of chemicals II. General Ames mutagenicity model., Chem. Res. Toxicol., 2007, 20 (4), pp 662–676.
Chromosomal aberration
The assessment of genotoxic potential is a critical point for approval and registration of new chemicals and drugs. Since no single test is capable of detecting all relevant genotoxic end-points, a battery of in vitro and in vivo tests for genotoxicity is recommended. One of the end-points of interest is the ability of the chemicals to cause damages on the chromosomes.
Chromosomal aberrations (CA) are abnormalities in the structure or number of chromosomes and are often responsible for genetic disorders. Therefore, highlighting of structural fragments associated with this biological effect appears of a crucial importance. Modeling the potential of chemicals to induce chromosomal damage is impeded by the diversity of mechanisms which condition this biological effect. One of the underlying mechanisms responsible for this effect is the direct binding of a chemical to DNA. The latter is also responsible for bacterial mutagenicity. Disturbance of DNA synthesis due to inhibition of topoisomerases and interaction of chemicals with nuclear proteins associated with DNA (e.g., histone proteins) were identified as additional mechanisms leading to chromosomal aberrations (CA). A comparative analysis of in vitro genotoxic data for a large number of chemicals revealed that over 80% of chemicals that elicit bacterial mutagenicity (as indicated by the Ames test) also induce CA; alternatively, only 60% of chemicals that induce CA have been found to be active in the Ames test. In agreement with this relationship the Ames model for bacterial mutagenicity, which has already been derived (see the above section) was combined with a newly derived model accounting for interaction of chemicals with specific proteins (histone proteins and topoisomerase) under the assumption that direct interaction with DNA and/or proteins lead to CA. The models for DNA and protein reactivity are based on the classical concept of reactive alerts. Some of the alerts interact directly with DNA or nuclear proteins, whereas others need to be fired by the rest of the molecules. Two- or three-dimensional quantitative structure-activity relationship models are used to assess the degree of activation of the alerts from the rest of the molecules. The use of each of the alerts has been justified by a mechanistic interpretation of the interaction. In combination with a rat liver S9 metabolism simulator, the model explained the CA induced by metabolically activated chemicals that do not elicit activity in the parent form. The model can be applied in two ways: with and without metabolic activation of chemicals.
References:
Mekenyan O., Serafimova R., Todorov M., Stoeva S., Aptula A., Jacob E., Finking R. Identification of the structural requirements for chromosomal aberration, by incorporating molecular flexibility and metabolic activation of chemicals. (2007) Chem. Res. Toxicol. 20 (12), 1927–1941.
RECEPTOR MEDIATED ENDPOINTS
Estrogen binding affinity
There is a diversity of toxic pathways of disrupting endocrine system which makes assessing the risk of potential endocrine disrupting chemicals (EDC) a particularly challenging task. The currently implemented model for EDC`s is focused on the direct interactions of chemicals with ER — that is, ER binding mediated endocrine disruption. This interaction is considered as a primary event that triggers the ultimate biological responses. In this respect, the structural tolerance of ER allowing a large number of exogenous chemicals to mimic the action of natural hormones is of high concern.
The COREPA (COmmon REactivity PAttern) technique which is used for predicting ER binding affinity is a probabilistic scheme for estimating the COREPA of biologically similar chemicals, accounting for their molecular flexibility. Analysis of reactivity patterns of the training set chemicals (645 chemicals with observed data for binding to human ER evaluated by in vitro test) is based on the distance between nucleophilic sites resulted in identification of distinct interaction types: a steroid-like A–B type described by frontier orbital energies and distance between nucleophilic sites with specific charge requirements; an A–C type where local hydrophobic effects are combined with electronic interactions to modulate binding; and mixed A–B–C (AD) type. The models are organized as rules combining prefiltering requirements and COREPA reactivity patterns. In this respect, each model could be considered as a small expert system based on ‘‘knowledge’’ for structural association with different interaction mechanisms and potency classes; subsequently the models are combined in an integral expert system (battery). The model is combined with a simulator of mammalian liver metabolism. In this way the model could predict the ER binding potential of metabolically activated chemicals that do not elicit activity in the parent form. The model can be applied in two ways: with and without metabolic activation of chemicals.
References:
Serafimova, R., Todorov, M., Nedelcheva, D., Pavlov, T., Akahori, Y., Nakai, M., and Mekenyan, O. (2007) QSAR and mechanistic interpretation of estrogen receptor binding, SAR QSAR Environ. Res. 18, 1-33.
Androgen binding affinity
The androgen receptor (AR) is an important cellular regulatory protein and plays a critical role in numerous physiological processes, including the development and maintenance of male secondary sexual characteristics. For some environmental incidences chemicals evidences were found that they can interfere with the AR signaling pathways by mimicking or inhibiting the action of endogenous androgens and may lead to endocrine disruption.
The multiparameter formulation of COmmon REactivity PAttern (COREPA) approach was used to describe the structural requirements for eliciting androgen potency. Structurally diverse training data set containing 202 chemicals was used for modeling. In agreement with previous models, chemical affinities for the ratAR were related to distances between nucleophilic sites and structural features describing electronic interactions between the receptor and ligands. A single binding mechanism was identified similar to the steroidal A-B electronic mechanism identified for ER binding activity.
The COREPA model predicting AR binders with relative binding affinity (RBA) higher then 10% is related to specific distance screen between nucleophilic sites (oxygen atoms) and additional discriminating parameters - maximal donor delocalizability and logKow. Almost same model was obtained for the moderate AR range (0.1<RBA<10%): again specific distance screen between nuclephilic sites in combination with parameters partial positive charged surface areas and hydrophobicity (logKow). The interaction mechanisms for chemicals belonging to the lowest AR activity range (0.001<RBA<0.1%) were found to be poorly defined. A series of class specific models may be more appropriate including polychlorinated chemicals, alkylphenols, DDTs, etc.
The integral screening tool for predicting binding affinity to AR was built as a battery of model each associated with different activity bins.
Aril hydrocarbon receptor binding affinity
Aryl hydrocarbon receptor (AhR) is a protein involved in development of both liver and immune system. The AhR is recognized as a transcription factor and member of a class structurally related DNA binding proteins, which binds to variety of receptor agonists eliciting the prototypical AhR-mediated biochemical responses. The increased interest in modeling of the AhR binding affinity is based on the fact that exposure of mammalians to these agonists produces not only biological but also toxic effect including thymus atrophy and immune suppression. The modeling was performed by utilization of the Common Reactivity Pattern (COREPA) approach providing mechanistic transparency of the AhR binding. Moreover, the derived model was also used to identify chemicals which bind to the receptor in antagonistic fashion. The properties of AhR agonists and antagonists correlated with their gene expression (GE) effects. The highest increase in GE was elicited by strong agonists, whereas a lower increase in GE was caused by weak agonists. The antagonists (and non AhR binders) had no GE. The identified correlation between AhR binding affinity and GE is used to define GE of the chemicals using predicted binding data and vice versa.
References:
Petko I. Petkov, J. C. Rowlands, R. Budinsky, B. Zhao, M. Denison, O.Mekenyan, Mechanism based common reactivity pattern (COREPA) modeling of AhR binding affinity. SAR and QSAR in Environmental Research (submitted).