Computational Toxicology最新文献

Aromatase is a crucial enzyme in the aromatization process, which catalyzes the conversion of androgenic steroids to estrogens. Aromatase dysregulation, as well as elevated estrogen levels, have been linked to a variety of malignancies, including breast cancer. Herein, we present the results of the optimization of Xanthones employing density functional theory (DFT) using the B3LYP/6-311G+(d, p) basis set to determine their frontier molecular orbitals, Mulliken charges, and chemical reactivity descriptors. According to the DFT results, Erythrommone has the smallest HOMO-LUMO gap (3.85 Kcal/mol), as well as the greatest electrophilicity index (5.19) and basicity (4.47). Xanthones and their derivatives were docked into the active site cavity of CYP450 to examine their structure-based inhibitory effect. The docking simulation studies predicted that Erythrommone has the lowest binding energy (-7.43 Kcal/mol), which is consistent with the DFT calculations and may function as a powerful CYP450 inhibitor equivalent to its known inhibitor, Exemestane, which has a binding affinity of −8.13 Kcal/mol. The high binding affinity of Xanthones was linked to the existence of hydrogen bonds as well as various hydrophobic interactions between the ligand and the receptor's essential amino acid residues. The findings demonstrated that Xanthones are more powerful inhibitors of the Aromatase enzyme than the recognized inhibitor Exemestane.

Established cytotoxicity assays are commonly used for assessing the hepatotoxic risk of a compound. The addition of gene expression measurements from high-dimensional RNAseq experiments offers the potential for improved classification. However, it is generally not clear how best to summarize the high-dimensional gene measurements into meaningful variables. We propose several intuitive methods for dimension reduction of gene expression measurements toward interpretable variables and explore their relevance in predicting hepatotoxicity, using a dataset with 60 compounds.

Different advanced statistical learning algorithms are evaluated as classification methods and their performances are compared on the dataset. The best predictions are achieved by tree-based methods such as random forest and xgboost, and tuning the parameters of the algorithm helps to improve the classification accuracy. It is shown that the simultaneous use of data from cytotoxicity assays and from gene expression variables summarized in different ways has a synergistic effect and leads to a better prediction of hepatotoxicity than both sets of variables individually. Further, when gene expression data are summarized, different strategies for the generation of interpretable variables contribute to the overall improved prediction quality. When considering cytotoxicity assays alone, the best classification method yields a mean accuracy of 0.757, while the same classification method and an optimal choice of variables yields a mean accuracy of 0.811. The overall best value for the mean accuracy is 0.821.

This work estimates benchmarks for new approach method (NAM) performance in predicting organ-level effects in repeat dose studies of adult animals based on variability in replicate animal studies. Treatment-related effect values from the Toxicity Reference database (v2.1) for weight, gross, or histopathological changes in the adrenal gland, liver, kidney, spleen, stomach, and thyroid were used. Rates of chemical concordance among organ-level findings in replicate studies, defined by repeated chemical only, chemical and species, or chemical and study type, were calculated. Concordance was 39–88%, depending on organ, and was highest within species. Variance in treatment-related effect values, including lowest effect level (LEL) values and benchmark dose (BMD) values when available, was calculated by organ. Multilinear regression modeling, using study descriptors of organ-level effect values as covariates, was used to estimate total variance, mean square error (MSE), and root residual mean square error (RMSE). MSE values, interpreted as estimates of unexplained variance, suggest study descriptors accounted for 52–69% of total variance in organ-level LELs. RMSE ranged from 0.41 to 0.68 log₁₀-mg/kg/day. Differences between organ-level effects from chronic (CHR) and subchronic (SUB) dosing regimens were also quantified. Odds ratios indicated CHR organ effects were unlikely if the SUB study was negative. Mean differences of CHR - SUB organ-level LELs ranged from − 0.38 to − 0.19 log₁₀ mg/kg/day; the magnitudes of these mean differences were less than RMSE for replicate studies. Finally, in vitro to in vivo extrapolation (IVIVE) was employed to compare bioactive concentrations from in vitro NAMs for kidney and liver to LELs. The observed mean difference between LELs and mean IVIVE dose predictions approached 0.5 log₁₀-mg/kg/day, but differences by chemical ranged widely. Overall, variability in repeat dose organ-level effects suggests expectations for quantitative accuracy of NAM prediction of LELs should be at least ± 1 log₁₀-mg/kg/day, with qualitative accuracy not exceeding 70%.

Human CYP2D6 contributes extensively to the biotransformation of important therapeutic drugs. CYP2D6 substrate and inhibitor specificity may be affected by genetic polymorphism. This study aimed to characterize interactions of three typical ligands, 3-cyano-7-ethoxycoumarin, fluoxetine and terbinafine with six CYP2D6 variants using molecular docking simulations. The compounds were docked individually to the CYP2D6 models based on published crystal structure (PDB code: 3TBG). All ligands bound within the active site pocket near the heme. Binding involved residues found in critical secondary structures that formed the active site boundary: B-C loop, F helix, F-G loop, β-1 strands and I helix. Twenty-five amino acids were involved in the binding, and all were located in the known substrate recognition sites. Hydrophobic bonds involving phenylalanine (Phe120, Phe384) dominated CEC binding whereas electrostatic bonds between the protonated nitrogen with acidic residues (Glu216, Glu222, Asp301) dominated in binding of fluoxetine and terbinafine. Collectively, the subtle structural changes in the active site and substrate access channels induced by the mutations in the variants contributed to differential ligand docking poses. This study has provided insights into important molecular properties for CYP2D6 catalysis and inhibition, and formed basis for further exploration of structural determinants for potency and specificity of CYP2D6 ligands.

The “Six Pack” is a set of animal toxicity studies that are widely used by industry and regulatory agencies to evaluate the toxicity of chemicals. It consists of three systemic toxicities (acute oral toxicity, acute inhalation toxicity and acute dermal toxicity) and three specific organ endpoints (eye damage/irritation, skin corrosion/irritation and skin sensitization). In the last two decades there has been a growing effort in the scientific community, as well as in regulatory agencies, to reduce and replace animal tests through implementation of alternative approaches. Computational methods in combination with in vitro measurements are pursued actively as the integrative approach for accurate and reliable assessment of chemical toxicity. Here, we generated structural alerts and developed a set of ten classification models for all six-pack endpoints using different molecular descriptors and machine learning techniques. The coverage of active chemicals by structural alerts was in the range from 24 % for acute inhalation toxicity to 52 % for acute oral toxicity. To establish confidence in model predictions, we used two different approaches to estimate the applicability domain (AD). The first approach was based on similarity distance between the query chemical and chemicals in the training set. In the second approach, the AD was estimated based on distance to model. The prediction accuracy of models evaluated using the validation sets was in the range from 0.67 for acute inhalation toxicity to 0.78 for acute dermal toxicity. The evaluation of models for chemicals within the similarity-based AD showed similar accuracy compared with the whole validation set. On the other hand, improvement of model performance was observed by using the distance to model approach to estimate AD, e.g. when distance to model was set to 0.3 the accuracy of predictions ranged from 0.75 for acute inhalation toxicity to 0.86 for acute oral toxicity. The combination of structural alerts and classification models provide a rapid means to screen a list of compounds for six-pack toxicity and to prioritize chemicals for in vitro toxicity evaluation.

Over-activation of extra-synaptic NMDARs by excessive glutamate is known to cause excitotoxicity. The molecular mechanism of how this excitotoxicity occurs was revealed recently. This paper presents the results of in silico studies aimed at finding potential small-molecule inhibitors that can block this mechanism, namely the extra-synaptic NMDAR/TRPM4 interaction. We screened for small molecules according to 2D (at least Tanimoto threshold was 90%) and/or 3D similarity, molecular weight, lipophilicity using control compounds (C8 and C19) targeting this interaction. We then pre-filtered these molecules according to their drug-likeness and toxicity profiles. After pre-filtering, we performed a docking study against the extra-synaptic NMDAR/TRPM4 interaction with the remaining 26 compounds. In addition, we determined that selected compounds exhibit low affinity for classical NMDAR ligand binding sites. Ultimately, we identified four novel compounds (C8-12, C8-15, C19-3, C19-4) that could block the extra-synaptic NMDAR/TRPM4 interaction without inhibiting the normal function of synaptic NMDARs.

The change in age-related mortality patterns is an important characteristic of the population that can be used as a metric of risk by comparing exposed and non-exposed populations.

In this paper, the mortality parameters were predicted for populations exposed to crystalline silica, a proven lung carcinogen.

Seven hazard functions were tested for a dose–response relationship between lung cancer and characteristics of exposure. Life tables were calculated, along with parameters of the Gompertz-Makeham model for the force of mortality.

It was demonstrated, in particular, that exposure to crystalline silica in the range from 0.03 to 0.3 mg/m³ for 40 years starting at age 20 causes a predicted drop in average life expectancy in the range of from 0.15 to 1.38 years.

It was demonstrated that the lost life expectancy linearly correlates with relative risk (R = 0.995, R² = 0.989, p< 0.00001). The probability of the life expectancy increasing while relative risk decreases was as low as 0.01.

It was found that exponential parameter α of the Gompertz-Makeham equation increases with crystalline silica exposure, while the two linear parameters A and R (which are negatively correlated between each other) increase or decrease with exposure depending on the duration and onset age. Modal age of death in the cohort decreases with cumulative exposure with R = -0.977, R² = 0.954, p < 0.0001.

Based on several different approaches, it was suggested that the threshold of cumulative crystalline silica exposure concentration causing statistically significant change in the cohort life tables can be found in the range from 1.81 to 2.50 mg/m³-years. The change of average age of death in exposed male population does not exceed 1% below cumulative exposure of 3.5 mg/m³-years, and does not exceed 5% at cumulative exposure less than 9.8 mg/m³-years. It shows that no significant acceleration of death rate with age is happening even at the high levels of exposure to crystalline silica.

The study demonstrated the value and advantages of the use of life expectancy and other lifetable characteristics as a tool for quantitative risk assessment.

Botulinum neurotoxin (BoNT) is a highly toxic protein and a Tier 1 Biodefense Select Agent and Toxin. BoNT is also a widely used therapeutic and cosmetic. Despite the toxicological and pharmacological interest, little is known about its biodistribution in the body. The objective herein was to develop a dose-dependent, species-specific physiologically-based toxicokinetic (PBTK) model of BoNT biodistribution in rodents following a single intravenous dose. The PBTK model was based on published physiologically-based pharmacokinetic (PBPK) models of therapeutic monoclonal antibody (mAb) biodistribution because the size and charge of BoNT is nearly identical to a typical IgG₄ mAb and size/charge are main factors governing protein biodistribution. Physiological compartments included the circulation, lymphatics and tissues grouped by capillary pore characteristics. Host species-specific parameters included weight, plasma volume, lymph volume/flow, and tissue interstitial fluid parameters. BoNT parameters included extravasation from blood to tissues, charge, binding to internal lamella or cholinergic neuron receptors. Parameter values were obtained from the literature or estimated using an Approximate Bayesian Computation-Sequential Monte Carlo algorithm, to fit the model to published mouse BoNT low-dose, time-course plasma concentration data. Fits captured the low-dose mouse data well and parameter estimates appeared biologically plausible. The fully-parameterized model was then used to simulate mouse high-dose IV data. Model results compared well with published data. Finally, the model was re-parameterized to reflect rat physiology. Model toxicokinetics agreed well with published rat BoNT intravenous data for two different sized rats with different intravenous doses (an a priori cross-species extrapolation). These results suggested the BoNT model predicted dose-dependent biodistribution in rodents, and for rats, without any BoNT-specific data from rats. To our knowledge, this represented a first-in-kind physiologically-based model for a large protein toxin. Results are discussed in general and in the context of human simulations to support BoNT risk assessment and therapeutic research objectives.

Physiologically based pharmacokinetic (PBPK) modeling is used to predict the pharmacokinetic behavior of xenobiotics in humans. During pregnancy, anatomical and physiological parameters are modified leading to toxicokinetics’ changes of substances in the body. Considering these physiological parameters change in the building processes of pregnancy PBPK (p-PBPK) model is essential to have accurate estimates of tissue/organ concentrations for the pregnant women but also for the fetus.

The review aims to summarize which specific processes are considered in the building of p-PBPK models and may be useful at the early stages of p-PBPK modeling.

To achieve this objective, a literature search focusing on anatomical, physiological, and biochemical parameters impacted by pregnancy was conducted. Most of the time, p-PBPK models do not include all the specific processes identified but only the most impacting ones on the global kinetics, depending mainly on the substance of interest. Allometric relations were identified to be classically included in the pregnancy models to describe the modifications induced by pregnancy to overcome the lack of data usually observed for the gestation. However, more and more data are gathered for pregnancy leading to the introduction of more data-based equations in PBPK modeling.

The most common strategy for p-PBPK development is based on the development of adult PBPK models that are then adapted to specific populations such as pregnant women. The adult PBPK model structure is modified to account for the pregnancy by adding specific compartments of fetal development and also specific compartments that are impacted during the pregnancy such as fat or mammary glands. Extrapolation of pregnant rat model is the other common strategy option used more specifically for environmental substances.

Overall, further data on maternal and fetal pharmacokinetics are needed to validate the xenobiotic exposure predictions during pregnancy, using for example in vitro, in vivo or ex vivo experiments.

This paper presents the development of toxicity of some saturated and phthalate carboxylic diesters (CDE) quantified by experimentally measured (Mes) and calculated (C) values using the Hydractinia echinata (invertebrate) Toxicity Screening Test System (HeTSTS) and the Köln Model (KM) algorithm. The validity of the investigation model is confirmed by the results for three other aquatic organisms: the ciliate protozoan Tetrahymena pyriformis, the freshwater fish Pimephales promelas and the freshwater crustacean Daphnia magna test systems have shown that the evolution of effectiveness is similar, although the absolute values are different. CDE undergoes rapid, irreversible, selective and abiotic –OH^– nucleophilic catalyzed monohydrolysis with the formation of the substrate amphiphilic carboxylate monoester (CME), saturated or phthalate and alcohol (AL) as a xenobiotic (SbX) binary mixture in stoichiometric proportion. The Mes represents the inverse of the logarithm of the diester concentration (molL^-1), which determines the 50% reduction in metamorphosis of H. echinata from larva to polyp and is influenced by the saturated carbon atom (Cs) of the molecular substructure involved in monohydrolysis. According to the KM algorithm, Cs is the Elementary Specific Interaction Parameter (ESIP) with a specific and constant toxicity value – identical in different substances – depending on the nature of the organism that allows the calculation of toxicity predictions in C. AL is the fingerprint of the mixture (FP) because it influences the diffusion of CMEs through the cell membrane to cellular receptors (CRs). Generally, the Mes and C, are the predicted ECOSAR and calculated C* values form the Class Regulated Increased Toxicity (CRIT) and Class Regulated Decreased Toxicities (CRDT) series. The use of H. echinata in toxicity determinations is an alternative for the study of the relevant ecological impact of chemical oxidative stress on aquatic organisms and, consequently, on human health.