Multiple interaction regions in the orthosteric ligand binding domain of the α7 neuronal nicotinic acetylcholine receptor.

Neuronal nicotinic receptors (nAChRs) belong to the Cys-loop family of ligand-gated ion channels and are formed from five subunits either as homologous or heterologous, oligomeric receptors, and are of interest as targets for treatment of a variety of central and peripheral nervous system disorders. Using a model of the homopentameric α7 nAChR extracellular region derived from the homologous acetylcholine binding protein (AChBP) from Aplysia California, binding modes of structurally diverse, high affinity α7 ligands were examined by docking to the orthosteric ligand binding domain. While all α7 ligands show similar interactions between the essential positively charged cationic center of the ligand and αTRP147 of the receptor (i.e., hydrogen bond to the tryptophan backbone carbonyl and cation-π interaction), docked poses of various ligands show the potential to interact with three additional regions within the binding domain, identified as regions 1, 2, and 3. Region 1 is located in the vicinity of Loop-E, involves ligand-protein interactions via a network of water-mediated hydrogen bonds, and is analogous to the region where pyridinyl groups are located in many of the AChBP-nicotinic ligand cocrystal structures. Ligands interacting with region 2 probe an area that spans from Loop-E to Loops-D and -F and may contribute to α7-selectivity over other nAChR subtypes. Several high affinity α7 ligands show strong interactions in this region. Region 3 is located near Loop-F of the protein and is analogous to an area involved in binding of an active metabolite derived from DMXBA, in an AChBP cocrystal structure. It appears that π-π interactions contribute to binding affinities of α7 nAChR ligands in this latter region, and further, this region may also contribute to α7-selectivity over other nAChR subtypes. Analysis of the resulting poses suggests that compounds with high α7 binding affinity do not require interactions across all regions simultaneously, but that interactions in multiple regions may enhance ligand binding and increase selectivity. Our results provide insight for further development of selective α7 nAChR ligands and may prove useful for the design of novel scaffolds for specific nicotinic therapeutic agents.

Quantitative structure-activity relationships of methotrexate and methotrexate analogues transported by the rat multispecific resistance-associated protein 2 (rMrp2).

Quantitative structure-activity relationship (QSAR) computational methods were performed to characterize structural requirements and molecular features for rMrp2-mediated methotrexate (MTX) transport. The compounds used in this analysis included MTX and 24 MTX analogues, with activity assessed by measuring inhibition of (3)H-DNP-SG ((3)H-S-(2,4)-dinitrophenyl glutathione) uptake in rat canalicular membrane vesicles. 2D-QSAR modeling using simulated annealing partial least squares (SA-PLS) method identified octanol/water partition coefficient, hydrophobicity, and negative charge as three important factors for MTX and MTX analogue affinity to rMrp2. Further analysis using 3D-QSAR method identified a pharmacophore model consisting of two hydrophobes, two aromatic rings, and a negative ionizable group as the critical molecular features that predict binding affinity of these compounds to rMRP2. The addition of a benzoyl ornithine group at a 9.3A distance and 136.5 degrees vector from the negative ionizable structure of MTX resulted in a 40-fold more potent inhibition of DNP-SG transport, suggesting that this chemical modification, while not essential for activity, contributes to the transport of MTX analogue by rMrp2. These observations provide important insights to the rationale development of analogues of MTX for the treatment of neoplastic and immunological diseases that may be devoid of hepatotoxicity or lack drug resistance.

Development and validation of k-nearest-neighbor QSPR models of metabolic stability of drug candidates.

Computational ADME (absorption, distribution, metabolism, and excretion) models may be used early in the drug discovery process in order to flag drug candidates with potentially problematic ADME profiles. We report the development, validation, and application of quantitative structure-property relationship (QSPR) models of metabolic turnover rate for compounds in human S9 homogenate. Biological data were obtained from uniform bioassays of 631 diverse chemicals proprietary to GlaxoSmithKline (GSK). The models were built with topological molecular descriptors such as molecular connectivity indices or atom pairs using the k-nearest neighbor variable selection optimization method developed at the University of North Carolina (Zheng, W.; Tropsha, A. A novel variable selection QSAR approach based on the k-nearest neighbor principle. J. Chem. Inf. Comput. Sci., 2000, 40, 185-194.). For the purpose of validation, the whole data set was divided into training and test sets. The training set QSPR models were characterized by high internal accuracy with leave-one-out cross-validated R(2) (q(2)) values ranging between 0.5 and 0.6. The test set compounds were correctly classified as stable or unstable in S9 assay with an accuracy above 85%. These models were additionally validated by in silico metabolic stability screening of 107 new chemicals under development in several drug discovery programs at GSK. One representative model generated with MolConnZ descriptors predicted 40 compounds to be metabolically stable (turnover rate less than 25%), and 33 of them were indeed found to be stable experimentally. This success (83% concordance) in correctly picking chemicals that are metabolically stable in the human S9 homogenate spells a rapid, computational screen for generating components of the ADME profile in a drug discovery process.

Targacept active conformation search: a new method for predicting the conformation of a ligand bound to its protein target.

Targacept active conformation search (TACS) is a novel variation of well-established three-dimensional quantitative structure--activity relationship methodologies that seeks to determine probable conformation(s) of ligands bound to their protein targets. A combination of affinity or activity data and energetically accessible conformational ensembles, each conformer described by three-dimensional (3-D) sensitive descriptors, forms the basis of the TACS data model. Recursive pruning is used to reduce the size of both the conformational ensemble and the descriptor space until the TACS data model contains just enough information to determine probable conformation(s) of ligands bound to their protein targets. The TACS algorithm is comprised of five components: (1) conformational ensemble generation, (2) 3-D sensitive descriptor calculation, (3) ensemble descriptor preprocessing, (4) model generation, and (5) prediction of bound conformation(s). Significantly, this method precludes the need for subjective or objective molecular alignment. We report the application of this technique to five benchmark protein-ligand couples where the conformation of a bound ligand has been previously established using X-ray crystallography: 9-cis-retinoic (1) and 9-trans-retinoic acid (2), both agonists for the retinoic acid receptor gamma, compounds KH1060 (3) and MC1288 (4), which bind to the vitamin D3 receptor, and R04 (5), an inhibitor bound to human rhinovirus 14 thermolysin. The binding conformations predicted by TACS were compared to the crystallographic structures extracted from their respective binding sites using root-mean-squared deviation (rmsd) criteria. Three of the conformations found using TACS were within crystallographic error. 9-cis-Retinoic acid, 9-trans-retinoic acid, and MC1288, when superimposed on their crystallographic structures, gave rmsd values of 0.22, 0.17, and 0.34 A, respectively. The rmsd values for KH1060 (1.54 A) and R04 (1.01 A) were larger but still reasonable.

Antitumor agents. 213. Modeling of epipodophyllotoxin derivatives using variable selection k nearest neighbor QSAR method.

We have applied a variable selection k nearest neighbor quantitative structure-activity relationship (kNN QSAR) method to develop predictive QSAR models for 157 epipodophyllotoxins synthesized previously in our ongoing effort to develop potential anticancer agents. QSAR models were generated using multiple topological descriptors of chemical structures, including molecular connectivity indices (MCI) and molecular operating environment descriptors. The 157 compounds were separated into several training and test sets. The robustness of QSAR models was characterized by the values of the internal leave one out cross-validated R2 (q2) for the training set and external predictive R2 for the test set. The significance of the training set models was confirmed by statistically higher values of q2 for the original data set as compared to q2 values for the same data set with randomly shuffled activities. kNN QSAR models were compared with those obtained with the comparative molecular field analysis method; the kNN QSAR approach afforded models with higher values of both q2 and predictive R2. One of the best models obtained from kNN analysis using MCI as descriptors provided q2 and predictive R2 values of 0.60 and 0.62, respectively. QSAR models developed in these studies shall aid in future design of novel potent epipodophyllotoxin derivatives.

Quantitative structure-activity relationship analysis of functionalized amino acid anticonvulsant agents using k nearest neighbor and simulated annealing PLS methods.

We report the development of rigorously validated quantitative structure-activity relationship (QSAR) models for 48 chemically diverse functionalized amino acids with anticonvulsant activity. Two variable selection approaches, simulated annealing partial least squares (SA-PLS) and k nearest neighbor (kNN), were employed. Both methods utilize multiple descriptors such as molecular connectivity indices or atom pair descriptors, which are derived from two-dimensional molecular topology. QSAR models with high internal accuracy were generated, with leave-one-out cross-validated R(2) (q(2)) values ranging between 0.6 and 0.8. The q(2) values for the actual dataset were significantly higher than those obtained for the same dataset with randomly shuffled activity values, indicating that models were statistically significant. The original dataset was further divided into several training and test sets, with highly predictive models providing q(2) values greater than 0.5 for the training sets and R(2) values greater than 0.6 for the test sets. These models were capable of predicting with reasonable accuracy the activity of 13 novel compounds not included in the original dataset. The successful development of highly predictive QSAR models affords further design and discovery of novel anticonvulsant agents.

Modeling of p38 mitogen-activated protein kinase inhibitors using the Catalyst HypoGen and k-nearest neighbor QSAR methods.

We have employed in parallel the Catalyst HypoGen pharmacophore modeling approach and the variable selection k-nearest neighbor quantitative structure-activity relationship (kNN QSAR) method to model a diverse data set of p38 mitogen-activated protein (MAP) kinase inhibitors. The HypoGen pharmacophore model, developed from a novel automated training set selection protocol, identified chemical functional features that were characteristic of the active compounds and differentiated the active from the inactive inhibitors. The kNN QSAR modeling employed topological descriptors and afforded predictive QSAR models with consistently high values of both leave-one-out cross-validated R2 for the training set and predictive R2 for the test set. The results of both modeling approaches were sensitive to the selection of the training and test sets used for model development and validation. The resulting Catalyst pharmacophore and kNN QSAR models can be used concurrently for rapid virtual screening of chemical databases to identify novel p38 MAP kinase inhibitors.

Quantitative structure-pharmacokinetic parameters relationships (QSPKR) analysis of antimicrobial agents in humans using simulated annealing k-nearest-neighbor and partial least-square analysis methods.

We have developed quantitative structure-pharmacokinetic parameters relationship (QSPKR) models using k-nearest-neighbor (k-NN) and partial least-square (PLS) methods to predict the volume of distribution at steady state (Vss) and clearance (CL) of 44 antimicrobial agents in humans. The performance of QSPKR was determined by the values of the internal leave-one-out, crossvalidated coefficient of determination q(2) for the training set and external predictive r(2) for the test set. The best simulated annealing (SA)-kNN model was highly predictive for Vss and provided q(2) and r(2) values of 0.93 and 0.80, respectively. For all compounds, the model produced average fold error values for Vss of 1.00 and for 93% of the compounds provided predictions that were within a twofold error of actual values. The best SA-kNN model for prediction of CL yielded q(2) and r(2) values of 0.77 and 0.94, respectively, and had an average fold rror of 1.05. Use of PLS methods resulted in inferior QSPKR models. The SA-kNN QSPKR approach has utility in drug discovery and development in the identification of compounds that possess appropriate pharmacokinetic characteristics in humans, and will assist in the selection of a suitable starting dose for Phase I, first-time-in-man studies.

Discovery of (2S,3R)-N-[2-(pyridin-3-ylmethyl)-1-azabicyclo[2.2.2]oct-3-yl]benzo[b]furan-2-carboxamide (TC-5619), a selective α7 nicotinic acetylcholine receptor agonist, for the treatment of cognitive disorders.

(2S,3R)-N-[2-(Pyridin-3-ylmethyl)-1-azabicyclo[2.2.2]oct-3-yl]benzo[b]furan-2-carboxamide (7a, TC-5619), a novel selective agonist of the α7 neuronal nicotinic acetylcholine receptor, has been identified as a promising drug candidate for the treatment of cognitive impairment associated with neurological disorders. 7a demonstrated more than a thousand-fold separation between the affinities for the α7 and α4ß2 receptor subtypes and had no detectable effects on muscle or ganglionic nicotinic receptor subtypes, indicating a marked selectivity for the central nervous system over the peripheral nervous system. Results obtained from homology modeling and docking explain the observed selectivity. 7a had positive effects across cognitive, positive, and negative symptoms of schizophrenia in animal models and was additive or synergistic with the antipsychotic clozapine. Compound 7a, as an augmentation therapy to the standard treatment with antipsychotics, demonstrated encouraging results on measures of negative symptoms and cognitive dysfunction in schizophrenia and was well tolerated in a phase II clinical proof of concept trial in patients with schizophrenia.

Analgesic effects mediated by neuronal nicotinic acetylcholine receptor agonists: correlation with desensitization of α4ß2* receptors.

Nicotinic α4ß2* agonists are known to be effective in a variety of preclinical pain models, but the underlying mechanisms of analgesic action are not well-understood. In the present study, we characterized activation and desensitization properties for a set of seventeen novel α4ß2*-selective agonists that display druggable physical and pharmacokinetic attributes, and correlated the in vitro pharmacology results to efficacies observed in a mouse formalin model of analgesia. ABT-894 and Sazetidine-A, two compounds known to be effective in the formalin assay, were included for comparison. The set of compounds displayed a range of activities at human (α4ß2)(2)ß2 (HS-α4ß2), (α4ß2)(2)α5 (α4ß2α5) and (α4ß2)(2)α4 (LS-α4ß2) receptors. We report the novel finding that desensitization of α4ß2* receptors may drive part of the antinociceptive outcome. Our molecular modeling approaches revealed that when receptor desensitization rather than activation activitiesat α4ß2* receptors are considered, there is a better correlation between analgesia scores and combined in vitro properties. Our results suggest that although all three α4ß2 subtypes assessed are involved, it is desensitization of α4ß2α5 receptors that plays a more prominent role in the antinociceptive action of nicotinic compounds. For modulation of Phase I responses, correlations are significantly improved from an r(2) value of 0.53 to 0.67 and 0.66 when HS- and LS-α4ß2 DC(50) values are considered, respectively. More profoundly, considering the DC(50) at α4ß2α5 takes the r(2) from 0.53 to 0.70. For Phase II analgesia scores, adding HS- or LS-α4ß2 desensitization potencies did not improve the correlations significantly. Considering the α4ß2α5 DC(50) value significantly increased the r(2) from 0.70 to 0.79 for Phase II, and strongly suggested a more prominent role for α4ß2α5 nAChRs in the modulation of pain in the formalin assay. The present studies demonstrate that compounds which are more potent at desensitization of α4ß2* receptors display better analgesia scores in the formalin test. Consideration of desensitization propertiesat α4ß2* receptors, especially at α4ß2α5, in multiple linear regression analyses significantly improves correlations with efficacies of analgesia. Thus, α4ß2* nicotinic acetylcholine receptor desensitization may contribute to efficacy in the mediation of pain, and represent a mechanism for analgesic effects mediated by nicotinic agonists.

Discovery of 3-(5-chloro-2-furoyl)-3,7-diazabicyclo[3.3.0]octane (TC-6683, AZD1446), a novel highly selective α4ß2 nicotinic acetylcholine receptor agonist for the treatment of cognitive disorders.

Diversification of essential nicotinic cholinergic pharmacophoric elements, i.e., cationic center and hydrogen bond acceptor, resulted in the discovery of novel potent α4ß2 nAChR selective agonists comprising a series of N-acyldiazabicycles. Core characteristics of the series are an exocyclic carbonyl moiety as a hydrogen bond acceptor and endocyclic secondary amino group. These features are positioned at optimal distance and with optimal relative spatial orientation to provide near optimal interactions with the receptor. A novel potent and highly selective α4ß2 nAChR agonist 3-(5-chloro-2-furoyl)-3,7-diazabicyclo[3.3.0]octane (56, TC-6683, AZD1446) with favorable pharmaceutical properties and in vivo efficacy in animal models has been identified as a potential treatment for cognitive deficits associated with psychiatric or neurological conditions and is currently being progressed to phase 2 clinical trials as a treatment for Alzheimer's disease.

Combinatorial QSAR modeling of P-glycoprotein substrates.

Quantitative structure-activity (property) relationship (QSAR/QSPR) models are typically generated with a single modeling technique using one type of molecular descriptors. Recently, we have begun to explore a combinatorial QSAR approach which employs various combinations of optimization methods and descriptor types and includes rigorous and consistent model validation (Kovatcheva, A.; Golbraikh, A.; Oloff, S.; Xiao, Y.; Zheng, W.; Wolschann, P.; Buchbauer, G.; Tropsha, A. Combinatorial QSAR of Ambergris Fragrance Compounds. J. Chem. Inf. Comput. Sci. 2004, 44, 582-95). Herein, we have applied this approach to a data set of 195 diverse substrates and nonsubstrates of P-glycoprotein (P-gp) that plays a crucial role in drug resistance. Modeling methods included k-nearest neighbors classification, decision tree, binary QSAR, and support vector machines (SVM). Descriptor sets included molecular connectivity indices, atom pair (AP) descriptors, VolSurf descriptors, and molecular operation environment descriptors. Each descriptor type was used with every QSAR modeling technique; so, in total, 16 combinations of techniques and descriptor types have been considered. Although all combinations resulted in models with a high correct classification rate for the training set (CCR(train)), not all of them had high classification accuracy for the test set (CCR(test)). Thus, predictive models have been generated only for some combinations of the methods and descriptor types, and the best models were obtained using SVM classification with either AP or VolSurf descriptors; they were characterized by CCR(train) = 0.94 and 0.88 and CCR(test) = 0.81 and 0.81, respectively. The combinatorial QSAR approach identified models with higher predictive accuracy than those reported previously for the same data set. We suggest that, in the absence of any universally applicable "one-for-all" QSAR methodology, the combinatorial QSAR approach should become the standard practice in QSPR/QSAR modeling.

Supervised self-organizing maps in drug discovery. 2. Improvements in descriptor selection and model validation.

The modeling of nonlinear descriptor-target relationships is a topic of considerable interest in drug discovery. We, herein, continue reporting the use of the self-organizing map-a nonlinear, topology-preserving pattern recognition technique that exhibits considerable promise in modeling and decoding these relationships. Since simulated annealing is an efficient tool for solving optimization problems, we combined the supervised self-organizing map with simulated annealing to build high-quality, highly predictive quantitative structure-activity/property relationship models. This technique was applied to six data sets representing a variety of biological endpoints. Since a high statistical correlation in the training set does not indicate a highly predictive model, the quality of all the models was confirmed by withholding a portion of each data set for external validation. Finally, we introduce new cross-validation and dynamic partitioning techniques to address model overfitting and assessment.

Supervised self-organizing maps in drug discovery. 1. Robust behavior with overdetermined data sets.

The utility of the supervised Kohonen self-organizing map was assessed and compared to several statistical methods used in QSAR analysis. The self-organizing map (SOM) describes a family of nonlinear, topology preserving mapping methods with attributes of both vector quantization and clustering that provides visualization options unavailable with other nonlinear methods. In contrast to most chemometric methods, the supervised SOM (sSOM) is shown to be relatively insensitive to noise and feature redundancy. Additionally, sSOMs can make use of descriptors having only nominal linear correlation with the target property. Results herein are contrasted to partial least squares, stepwise multiple linear regression, the genetic functional algorithm, and genetic partial least squares, collectively referred to throughout as the "standard methods". The k-nearest neighbor (kNN) classification method was also performed to provide a direct comparison with a different classification method. The widely studied dihydrofolate reductase (DHFR) inhibition data set of Hansch and Silipo is used to evaluate the ability of sSOMs to classify unknowns as a function of increasing class resolution. The contribution of the sSOM neighborhood kernel to its predictive ability is assessed in two experiments: (1) training with the k-means clustering limit, where the neighborhood radius is zero throughout the training regimen, and (2) training the sSOM until the neighborhood radius is reduced to zero. Results demonstrate that sSOMs provide more accurate predictions than standard linear QSAR methods.

Rational selection of training and test sets for the development of validated QSAR models.

Quantitative Structure-Activity Relationship (QSAR) models are used increasingly to screen chemical databases and/or virtual chemical libraries for potentially bioactive molecules. These developments emphasize the importance of rigorous model validation to ensure that the models have acceptable predictive power. Using k nearest neighbors (kNN) variable selection QSAR method for the analysis of several datasets, we have demonstrated recently that the widely accepted leave-one-out (LOO) cross-validated R2 (q2) is an inadequate characteristic to assess the predictive ability of the models [Golbraikh, A., Tropsha, A. Beware of q2! J. Mol. Graphics Mod. 20, 269-276, (2002)]. Herein, we provide additional evidence that there exists no correlation between the values of q2 for the training set and accuracy of prediction (R2) for the test set and argue that this observation is a general property of any QSAR model developed with LOO cross-validation. We suggest that external validation using rationally selected training and test sets provides a means to establish a reliable QSAR model. We propose several approaches to the division of experimental datasets into training and test sets and apply them in QSAR studies of 48 functionalized amino acid anticonvulsants and a series of 157 epipodophyllotoxin derivatives with antitumor activity. We formulate a set of general criteria for the evaluation of predictive power of QSAR models.

Combinatorial QSAR of ambergris fragrance compounds.

A combinatorial quantitative structure-activity relationships (Combi-QSAR) approach has been developed and applied to a data set of 98 ambergris fragrance compounds with complex stereochemistry. The Combi-QSAR approach explores all possible combinations of different independent descriptor collections and various individual correlation methods to obtain statistically significant models with high internal (for the training set) and external (for the test set) accuracy. Seven different descriptor collections were generated with commercially available MOE, CoMFA, CoMMA, Dragon, VolSurf, and MolconnZ programs; we also included chirality topological descriptors recently developed in our laboratory (Golbraikh, A.; Bonchev, D.; Tropsha, A. J. Chem. Inf. Comput. Sci. 2001, 41, 147-158). CoMMA descriptors were used in combination with MOE descriptors. MolconnZ descriptors were used in combination with chirality descriptors. Each descriptor collection was combined individually with four correlation methods, including k-nearest neighbors (kNN) classification, Support Vector Machines (SVM), decision trees, and binary QSAR, giving rise to 28 different types of QSAR models. Multiple diverse and representative training and test sets were generated by the divisions of the original data set in two. Each model with high values of leave-one-out cross-validated correct classification rate for the training set was subjected to extensive internal and external validation to avoid overfitting and achieve reliable predictive power. Two validation techniques were employed, i.e., the randomization of the target property (in this case, odor intensity) also known as the Y-randomization test and the assessment of external prediction accuracy using test sets. We demonstrate that not every combination of the data modeling technique and the descriptor collection yields a validated and predictive QSAR model. kNN classification in combination with CoMFA descriptors was found to be the best QSAR approach overall since predictive models with correct classification rates for both training and test sets of 0.7 and higher were obtained for all divisions of the ambergris data set into the training and test sets. Many predictive QSAR models were also found using a combination of kNN classification method with other collections of descriptors. The combinatorial QSAR affords automation, computational efficiency, and higher probability of identifying significant QSAR models for experimental data sets than the traditional approaches that rely on a single QSAR method.

Genetic algorithms and self-organizing maps: a powerful combination for modeling complex QSAR and QSPR problems.

Modeling non-linear descriptor-target activity/property relationships with many dependent descriptors has been a long-standing challenge in the design of biologically active molecules. In an effort to address this problem, we couple the supervised self-organizing map with the genetic algorithm. Although self-organizing maps are non-linear and topology-preserving techniques that hold great potential for modeling and decoding relationships, the large number of descriptors in typical quantitative structure-activity relationship or quantitative structure-property relationship analysis may lead to spurious correlation(s) and/or difficulty in the interpretation of resulting models. To reduce the number of descriptors to a manageable size, we chose the genetic algorithm for descriptor selection because of its flexibility and efficiency in solving complex problems. Feasibility studies were conducted using six different datasets, of moderate-to-large size and moderate-to-great diversity; each with a different biological endpoint. Since favorable training set statistics do not necessarily indicate a highly predictive model, the quality of all models was confirmed by withholding a portion of each dataset for external validation. We also address the variability introduced onto modeling through dataset partitioning and through the stochastic nature of the combined genetic algorithm supervised self-organizing map method using the z-score and other tests. Experiments show that the combined method provides comparable accuracy to the supervised self-organizing map alone, but using significantly fewer descriptors in the models generated. We observed consistently better results than partial least squares models. We conclude that the combination of genetic algorithms with the supervised self-organizing map shows great potential as a quantitative structure-activity/property relationship modeling tool.