Search

The focus of CNS drug pharmacokinetics programs has recently shifted from determining the total concentrations in brain and blood to considering also unbound fractions and concentrations. Unfortunately, assessing unbound brain exposure experimentally requires demanding in vivo and in vitro studies. We propose a physical model, based on lipid binding and pH partitioning, to predict in silico the unbound volume of distribution in the brain. The model takes into account the partition of a drug into lipids, interstitial fluid and intracellular compartments of the brain. The results are in good agreement with the experimental data, suggesting that the contributions of lipid binding and pH partitioning are important in determining drug exposure in brain. The predicted values are used, together with predictions for plasma protein binding, as corrective terms in a second model to derive the unbound brain to plasma concentration ratio starting from experimental values of total concentration ratio. The calculated values of brain free fraction and passive permeability are also used to qualitatively determine the brain to plasma equilibration time in a model that shows promising results but is limited to a very small set of compounds. The models we propose are a step forward in understanding and predicting pharmacologically relevant exposure in brain starting from compounds' chemical structure and neuropharmacokinetics, by using experimental total brain to plasma ratios, in silico calculated properties and simple physics-based approaches. The models can be used in central nervous system drug discovery programs for a fast and cheap assessment of unbound brain exposure. For existing compounds, the unbound ratios can be derived from experimental values of total brain to plasma ratios. For both existing and hypothetical compounds, the unbound volume of distribution due to lipid binding and pH partitioning can be calculated starting only from the chemical structure.

Drug metabolism, toxicity, and their interaction profiles are major issues in the drug-discovery and lead-optimization processes. The cytochromes P450 (CYPs) 2D6 and 2C9 are enzymes involved in the oxidative metabolism of a majority of marketed drugs. Therefore, the prediction of the binding affinity towards CYP2D6 and CYP2C9 would be beneficial for identifying cytochrome-mediated adverse effects triggered by drugs or chemicals (e.g., toxic reactions, drug-drug, and food-drug interactions). By identifying the binding mode by using pharmacophore prealignment, automated flexible docking, and by quantifying the binding affinity by multidimensional QSAR (mQSAR), we validated a model family of 56 compounds (46 training, 10 test) and 85 compounds (68 training, 17 test) for CYP2D6 and CYP2C9, respectively. The correlation with the experimental data (cross-validated r²=0.811 for CYP2D6 and 0.687 for CYP2C9) suggests that our approach is suited for predicting the binding affinity of compounds towards CYP2D6 and CYP2C9. The models were challenged by Y-scrambling and by testing an external dataset of binding compounds (15 compounds for CYP2D6 and 40 for CYP2C9). To assess the probability of false-positive predictions, datasets of nonbinders (64 compounds for CYP2D6 and 56 for CYP2C9) were tested by using the same protocol. The two validated mQSAR models were subsequently added to the VirtualToxLab (VTL, http://www.virtualtoxlab.org).

Broad modifications of various positions of the minimal natural epitope recognized by the myelin-associated glycoprotein (MAG), a blocker of regeneration of neurite injuries, produced sialosides with nanomolar affinities. However, important pharmacokinetic issues, for example, the metabolic stability of these sialosides, remain to be addressed. For this reason, the novel non-carbohydrate mimic 3 was designed and synthesized from (-)-quinic acid. For the design of 3, previously identified beneficial modifications of side chains of Neu5Ac were combined with the replacement of the ring oxygen by a methylene group and the substitution of the C(4)-OH by an acetamide. Although docking experiments to a homology model of MAG revealed that mimic 3 forms all but one of the essential hydrogen bonds identified for the earlier reported lead 2, its affinity was substantially reduced. Extensive molecular-dynamics simulation disclosed that the missing hydrogen bond of the former C(8)-OH leads to a change of the orientation of the side chain. As a consequence, an important hydrophobic contact is compromised leading to a loss of affinity.

A series of novel aryl-substituted triazolyl D-galactosamine derivatives was synthesized as ligands for the carbohydrate recognition domain of the major subunit H1 (H1-CRD) of the human asialoglycoprotein receptor (ASGP-R). The compounds were biologically evaluated with a newly developed competitive binding assay, surface plasmon resonance and by a competitive NMR binding experiment. With compound 1b, a new ligand with a twofold improved affinity to the best so far known D-GalNAc was identified. This small, drug-like ligand can be used as targeting device for drug delivery to hepatocytes.

We report a neural network modeling approach combined with genetic algorithm for prediction of experimental binding affinity to human Estrogen Receptor alpha and beta (ER-alpha and ER-beta) of a diverse set of chemicals. The counterpropagation artificial neural network is used as a modeling method. Structural features of ligands having the strongest influence to the binding affinities were investigated. The molecular descriptors have been selected in the variable selection procedure based on the genetic algorithm (GA). The 3D descriptors of molecular structures were calculated for the minimal energy conformation of isolated ligands. All the optimized models were tested by an internal and an external set of compounds. The models served for classification and prediction of binding affinities. The optimized models were 100% correct in the classification part, where the active molecules were separated from the inactive ones. The best predictive model of active molecules was assessed with the internal test set yielding the error in prediction RMS = 0.12, while the predictions for the external test set contain some outliers, which are ascribed to the incompatibility of individual compounds concerning the structural domain of our model. The influence of the receptor on the conformation of the ligands in the ligand-protein complex is described and discussed in the accompanying paper.

Molecular interactions of odorants with their olfactory receptors (ORs) are of central importance for the ability of the mammalian olfactory system to detect and discriminate a vast variety of odors with a limited set of receptors. How a particular OR binds and distinguishes different odorant molecules remains largely unknown on a structural basis. Here we investigated this question for the mouse eugenol receptor (mOR-EG). By screening a large odorant library, we discovered a wide range of chemical structures activating the receptor in heterologous mammalian cells. Potent agonists comprise (i) benzene, (ii) cyclohexane, or (iii) polycyclic structures substituted with alcohol, aldehyde, keto, ether, or esterified carboxylic groups. To detect those amino acids within the receptor that are in contact with a particular bound odorant molecule, we investigated how distinct mOR-EG point mutants were activated by the different odorant agonists found for the wild-type receptor. We identified 11 amino acids as a part of the receptor's ligand binding pocket. Molecular modeling predicted 10 of these residues in transmembrane helices TM3-TM6 and one in the extracellular loop between TM2 and TM3. These amino acids participate in odorant binding with variable importance depending on the type of odorant, revealing functional "fingerprints" of ligand-receptor interactions.

The VirtualToxLab is an in silico tool for predicting the toxic (endocrine-disrupting) potential of drugs, chemicals and natural products. It is based on a fully automated protocol and calculates the binding affinity of any molecule of interest towards a series of 12 proteins, known or suspected to trigger adverse effects and estimates the resulting toxic potential. In contrast to other approaches in the field, the technology allows to rationalize a prediction at the molecular level by interactively analyzing the binding mode of the tested compound with any target protein in 3D. The technology is accessible over the Internet (via a secure SSH protocol) and available for any science-oriented organization. The toxic potential - a complex value derived from the individual binding affinities, their standard deviation and the quality of the underlying model (number and ratio of training and test compounds, activity range covered) - of existing and hypothetical compounds is estimated by simulating and quantifying their interactions towards a series of macromolecular targets at the molecular level using automated flexible docking combined with multidimensional QSAR (mQSAR). Currently, those targets comprise 12 proteins: the androgen, aryl hydrocarbon, estrogen alpha/beta, glucocorticoid, mineralocorticoid, thyroid alpha/beta liver X and the peroxisome proliferator-activated receptor gamma as well as the enzymes cytochrome P450 3A4 (CYP 3A4) and 2A13 (CYP 2A13). Up to date, the technology has been used to predict the toxic potential for more than 2,000 drugs, chemicals and natural compounds. All results are posted in the Internet - in this account, a few will be discussed in detail with reference to the molecular mechanisms triggering the adverse effect.

The LXR model has been added in the VirtualToxLab, a fully automated technology that allows for the identification of the endocrine-disrupting potential of drugs, chemicals and natural products. This protocol has then been applied to screen a series of 161 natural compounds probing their binding to the LXR. The results of the simulation were compared with experimental data (where available) and suggest that the LXR model can be applied to predict the binding affinity of existing or hypothetical compounds for screening purposes. The binding of 52 ligands towards the liver X receptors (LXRs) was identified trough docking to the three-dimensional protein structure and quantified by multidimensional QSAR (mQSAR), an approach referred to as 'mixed-model QSAR'. The model was validated by the prediction of 17 external compounds (oxysterols) present neither in the training nor in the test set. The robustness of the model was verified by consensus scoring using a conceptually different methodology, and chance correlation was ruled out by a series of scramble tests.

We present a computational study on the human mineralocorticoid receptor (hMR) that is based on multi-dimensional quantitative structure–activity relationships (mQSAR). Therein, we identified the binding mode of 48 steroid and non-steroid homologues by flexible docking to the crystal structure (software Yeti) and quantified it using 6D-QSAR (software Quasar). The receptor surrogate, evolved using a genetic algorithm, converged at a cross-validated r2 of 0.810, and yielded a predictive r2 of 0.661. The model was challenged by a series of scramble tests and by consensus scoring (software Raptor: r2 = 0.844, predictive r2 = 0.620). The model was then employed to predict the binding affinity of 26 anabolic steroids, demonstrating to which extent they might disrupt the endocrine system via binding to the hMR. The model for the hMR was added to the VirtualToxLab™, a technology developed by the Biographics Laboratory 3R, allows the identification of the endocrine-disrupting potential of drugs, chemicals and natural products in silico.

The tetrasaccharide 1, a substructure of ganglioside GQ1b alpha, shows a remarkable affinity for the myelin-associated glycoprotein (MAG) and was therefore selected as starting point for a lead optimization program. In our search for structurally simplified and pharmacokinetically improved mimics of 1, modifications of the core disaccharide, the alpha(2-->3)- and the alpha(2-->6)-linked sialic acid were synthesized. Biphenylmethyl and (S)-lactate were identified as suitable replacements for the alpha(2-->6)-linked sialic acid. Combined with a core modification and the earlier found aryl amide substituent in the 9-position of the alpha(2-->3)-linked sialic acid, high affinity MAG antagonists were identified. All mimics were tested in a competitive target-based binding assay, providing relative inhibitory potencies (rIP). Compared to the reference tetrasaccharide 1, the rIPs of the most potent antagonists 59 and 60 are enhanced nearly 400-fold. Their K(D)s determined in surface plasmon resonance experiments are in the low micromolar range. These results are in semiquantitative agreement with molecular modeling studies. This new class of glycomimetics will allow to validate the role of MAG in the axon regeneration process.

The glucocorticoid receptor (GR) is a member of the nuclear receptor superfamily that affects immune response, development, and metabolism in target tissues. Glucocorticoids are widely used to treat diverse pathophysiological conditions, but their clinical applicability is limited by side effects. A prediction of the binding affinity toward the GR would be beneficial for identifying glucocorticoid-mediated adverse effects triggered by drugs or chemicals. By identifying the binding mode to the GR using flexible docking (software Yeti) and quantifying the binding affinity through multidimensional QSAR (software Quasar), we validated a model family based on 110 compounds, representing four different chemical classes. The correlation with the experimental data (cross-validated r(2)=0.702; predictive r(2)=0.719) suggests that our approach is suited for predicting the binding affinity of related compounds toward the GR. After challenging the model by a series of scramble tests, a consensus approach (software Raptor), and a prediction set, it was incorporated into our VirtualToxLab and used to simulate and quantify the interaction of 24 psychotropic drugs with the GR.

In the last decade, we have developed and validated an in silico concept based on multidimensional QSAR (mQSAR) for the prediction of the toxic potential of drugs and environmental chemicals. Presently, the VirtualToxLab includes eleven so-called virtual test kits for estrogen (?/?), androgen, thyroid (?/?), glucocorticoid, aryl hydrocarbon, mineralocorticoid and peroxisome proliferator-activated receptor ? as well as for the enzymes cytochrome P450 3A4 and 2A13. The surrogates have been tested against a total of 824 compounds and are able to predict the binding affinity close to the experimental uncertainty with only six of the 194 test compounds giving calculated results more than a factor of 10 off the experimental binding affinity and the maximal individual deviation not exceeding a factor of 15. These results suggest that our approach is suited for the in silico identification of endocrine-disrupting effects triggered by drugs and environmental chemicals. Most recently, the technology has been made available through the Internet for academic laboratories, hospitals and environmental organizations.

A QSAR study is reported, in which the relationship between chemical structure of a set of compounds and the binding affinity to human estrogen receptor alpha and beta (ER-alpha and ER-beta) is modelled. Counterpropagation neural networks are used to predict experimental binding affinity of a range of substances. Several compounds as estrogenic chemicals, phytoestrogens, and natural and synthetic estrogens are treated with a structure-based approach that involves the protein structure. The conformations obtained with a docking methodology are used to calculate molecular descriptors. The models are built up with the neural network training procedure, which encodes the information present in molecular descriptors and related binding affinities of the pre-selected training set of compounds. In order to reach the best possible models, a selection of the descriptors using genetic algorithm was conducted. The selection was directed by the error in the prediction of binding affinities of compounds from the test set. The final models obtained for estrogen receptor alpha and beta were tested with an external validation set and were compared with the models obtained from a receptor-independent approach reported in the accompanying paper.

It is fundamental to the development of any quantitative structure–activity relationship (QSAR) that the structures of the compounds be fully and accurately represented. It is most usual for the compounds in the chosen data set to be modeled on computers for which a range of molecular modeling and computational chemistry software is available. In order for this highly labor-intensive facet of a QSAR study to be successful, it clearly must result in a set of accurate chemical structures. The subsequent calculation of a range of molecular descriptors on which the QSAR will depend ideally will only be valid if the chemical structures are as close as possible to their actual characteristics in the aqueous phase and at ambient temperatures and pressures. Although it is not possible unambiguously to determine this by experiment, a combination of data from X-ray crystallography, NMR, and other spectroscopic techniques can give good clues to the structural features of chemicals, with which to validate computer molecular models.