Toxicity modeling and prediction with pattern recognition

Modeling MEK4 Kinase Inhibitors through Perturbed Electrostatic Potential Charges

MEK4, mitogen-activated protein kinase kinase 4, is overexpressed and induces metastasis in advanced prostate cancer lesions. However, the value of MEK4 as an oncology target has not been pharmacologically validated because selective chemical probes targeting MEK4 have not been developed. With advances in both computer and biological high-throughput screening, selective chemical entities can be discovered. Structure-based quantitative structure-activity relationship (QSAR) modeling often fails to generate accurate models due to poor alignment of training sets containing highly diverse compounds. Here we describe a highly predictive, nonalignment based robust QSAR model based on a data set of strikingly diverse MEK4 inhibitors. We computed the electrostatic potential (ESP) charges using a density functional theory (DFT) formalism of the donor and acceptor atoms of the ligands and hinge residues. Novel descriptors were then generated from the perturbation of the charge densities of the donor and acceptor atoms and were used to model a diverse set of 84 compounds, from which we built a robust predictive model.

Classification models for safe drug molecules

Frequent failure of drug candidates during development stages remains the major deterrent for an early introduction of new drug molecules. The drug toxicity is the major cause of expensive late-stage development failures. An early identification/optimization of the most favorable molecule will naturally save considerable cost, time, human efforts and minimize animal sacrifice. (Quantitative) Structure Activity Relationships [(Q)SARs] represent statistically derived predictive models correlating biological activity (including desirable therapeutic effect and undesirable side effects) of chemicals (drugs/toxicants/environmental pollutants) with molecular descriptors and/or properties. (Q)SAR models which categorize the available data into two or more groups/classes are known as classification models. Numerous techniques of diverse nature are being presently employed for development of classification models. Though there is an increasing use of classification models for prediction of either biological activity or toxicity, the future trend will naturally be towards the development of classification models capable of simultaneous prediction of biological activity, toxicity, and pharmacokinetic parameters so as to accelerate development of bioavailable safe drug molecules.

Levels of perfluorinated chemicals in municipal drinking water from Catalonia, Spain: public health implications

In this study, the concentrations of 13 perfluorinated compounds (PFCs) (PFBuS, PFHxS, PFOS, THPFOS, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTDA, and PFOSA) were analyzed in municipal drinking water samples collected at 40 different locations from 5 different zones of Catalonia, Spain. Detection limits ranged between 0.02 (PFHxS) and 0.85 ng/L (PFOA). The most frequent compounds were PFOS and PFHxS, which were detected in 35 and 31 samples, with maximum concentrations of 58.1 and 5.30 ng/L, respectively. PFBuS, PFHxA, and PFOA were also frequently detected (29, 27, and 26 samples, respectively), with maximum levels of 69.4, 8.55, and 57.4 ng/L. In contrast, PFDoDA and PFTDA could not be detected in any sample. The most contaminated water samples were found in the Barcelona Province, whereas none of the analyzed PFCs could be detected in two samples (Banyoles and Lleida), and only one PFC could be detected in four of the samples. Assuming a human water consumption of 2 L/day, the maximum daily intake of PFOS and PFOA from municipal drinking water would be, for a subject of 70 kg of body weight, 1.7 and 1.6 ng/kg/day. This is clearly lower than the respective Tolerable Daily Intake set by the European Food Safety Authority. In all samples, PFOS and PFOA also showed lower levels than the short-term provisional health advisory limit for drinking water (200 ng PFOS/L and 400 ng PFOA/L) set by the US Environmental Protection Agency. Although PFOS and PFOA concentrations found in drinking water in Catalonia are not expected to pose human health risks, safety limits for exposure to the remaining PFCs are clearly necessary, as health-based drinking water concentration protective for lifetime exposure is set to 40 ng/L for PFOA.

Promises and pitfalls of quantitative structure-activity relationship approaches for predicting metabolism and toxicity

The description of quantitative structure-activity relationship (QSAR) models has been a topic for scientific research for more than 40 years and a topic within the regulatory framework for more than 20 years. At present, efforts on QSAR development are increasing because of their promise for supporting reduction, refinement, and/or replacement of animal toxicity experiments. However, their acceptance in risk assessment seems to require a more standardized and scientific underpinning of QSAR technology to avoid possible pitfalls. For this reason, guidelines for QSAR model development recently proposed by the Organization for Economic Cooperation and Development (OECD) [Organization for Economic Cooperation and Development (OECD) (2007) Guidance document on the validation of (quantitative) structure-activity relationships [(Q)SAR] models. OECD Environment Health and Safety Publications: Series on Testing and Assessment No. 69, Paris] are expected to help increase the acceptability of QSAR models for regulatory purposes. The guidelines recommend that QSAR models should be associated with (i) a defined end point, (ii) an unambiguous algorithm, (iii) a defined domain of applicability, (iv) appropriate measures of goodness-of-fit, robustness, and predictivity, and (v) a mechanistic interpretation, if possible [Organization for Economic Cooperation and Development (OECD) (2007) Guidance document on the validation of (quantitative) structure-activity relationships [(Q)SAR] models. The present perspective provides an overview of these guidelines for QSAR model development and their rationale, as well as the promises and pitfalls of using QSAR approaches and these guidelines for predicting metabolism and toxicity of new and existing chemicals.

Computer-modeling-based QSARs for analyzing experimental data on biotransformation and toxicity

Over the past decades the description of quantitative structure-activity relationships (QSARs) has been undertaken in order to find predictive models and/or mechanistic explanations for chemical as well as biological activities. This includes QSAR studies in toxicology. In an approach beyond the classical QSAR approaches, attempts have been made to define parameters for the QSAR studies on the basis of quantum mechanical computer calculations. The conversion of relatively small xenobiotics within the active sites of biotransformation enzymes can be expected to follow the general rules of chemistry. This makes the description of QSARs on the basis of only one parameter, chosen on the basis of insight in the mechanism, feasible. In contrast, toxicological endpoints can very often be the result of more than one physico-chemical interaction of the compound with the model system of interest. Therefore the description of quantitative structure-toxicity relationships often does not follow a one-descriptor mechanistic approach but starts from the other end, describing QSARs by multi-parameter approaches. The present paper focuses on the possibilities and restrictions of using computer-based QSAR modeling for analyzing experimental toxicological data, with emphasis on examples from the field of biotransformation and toxicity.

Activation of respiratory burst in human granulocytes by polychlorinated biphenyls: A structure-activity study

The respiratory burst in human granulocytes activated by 33 different congeners of polychlorinated biphenyls (PCBs) was measured as luminol-amplified chemoluminescence. The selection of 20 (training set) compounds was based on multivariate chemical characterization, laying the groundwork for covering the whole chemical series of tetra- through hepta-chlorinated PCBs. In addition 6 congeners were used as a validation set, and 7 were mono- to tri-chlorinated congeners representing low-chlorinated compounds not covered by the training set. Only ortho-substituted biphenyls activate respiratory burst. There is a correlation between activated respiratory burst and the total surface area of congeners up to 230 x 10(-20) m(2). Congeners of larger size show a reduced activity. There is also a correlation between respiratory burst activity and the number of ortho-substituents. Furthermore, there is also a correlation with parameters that describe absolute hardness of the molecule and respiratory burst activity. Congeners with a 2,4, 6-substitution on one biphenyl ring are optimal activators. In conclusion, all three factors, size, rotation, and electronic properties, which are not independent of each other, are important for the activity of the PCBs.

Potential pathways and exposure to explain the human body burden of organochlorine compounds: a multivariate statistical analysis of human monitoring in Würzburg, Germany

By means of multivariate statistical analysis of human monitoring in the vicinity of Würzburg it is examined whether the organochlorine levels in blood data of 309 persons can in part be explained by pollution from outdoor and/or indoor air. Although diet is the main source for intake this study concentrated on the remaining sources. Using 12 personal and geographical variables, which describe some aspects or components of indoor or outdoor exposure, the quantities of the different pathways for the intake are calculated and shown. Besides the impact of the indoor air, which can be pointed out for intake at the workplace for some PCB-congeners, also the hypothesis of the existence of an intake from outdoor air for some organochlorine compounds cannot be rejected. Here, the origin is the spatial variation of the pollution, which can be directly traced back to micro-climatic processes, local relief conditions, and the functional set-up of the town.

Analogy Models for Prediction of Human Toxicity

Estimating the toxicity to humans of chemicals by testing on human subjects is not considered to be ethically acceptable, and toxicity testing on laboratory animals is also questionable. Therefore, there is a need for alternative methods that will give estimates of various aspects of human toxicity. Batteries of in vitro tests, together with physicochemical and toxicokinetic data, analysed by efficient data analytical methods, may enable analogy models to be constructed that can predict human toxicity. It may be possible to model non-specific toxicity relating to lipophilicity, or basal cytotoxicity, for a series of diverse compounds with large variation in chemical structure and physicochemical properties. However, local models for a series of similar compounds are generally expected to be more accurate, as well as being capable of modelling more-specific interactions. Analogy models for the prediction of human toxicity are discussed and exemplified with physicochemical and cytotoxicity data from the first ten chemicals in the multicenter evaluation of in vitro cytotoxicity (MEIC) project.

Multivariate classification of histochemically stained human skeletal muscle fibres by the SIMCA method

The SIMCA (soft independent modelling of class analogy) method of pattern recognition has been used to classify four muscle fibre types: I, IIA, IIB and IIC. The samples were histochemically stained human skeletal sections from biopsy material. Disjoint (separate) class modelling gave information about variables, i.e., the combinations of alkaline, acidic and Ca2+-containing preincubation procedures with appropriate discrimination power, and showed satisfactory separation of the classes (fibre types). Two serial stained muscle sections represent a minimum for a proper classification of the four fibre groups. A comparison of biopsy samples from two different persons showed significant variation in the data structure between similar fibre types, probably caused by intermuscle variations. It is suggested that the introduction of computer-assisted classification by the application of such multivariate analytical techniques both facilitates the classification of muscle fibres and improves the precision and reliability of fibre typing.