COMPUTATIONAL TOXICOLOGY COMPUTATIONAL TOXICOLOGY Risk Assessment for Pharmaceutical and Environmental Chemicals Edited by SEAN EKINS WILEY-INTERSCIENCE A JOHN WILEY & SONS, INC., PUBLICATION Cover design/concept by Sean Ekins using images from Chapters 13, 16, and 19 Copyright © 2007 by John Wiley & Sons, Inc All rights reserved Published by John Wiley & Sons, Inc., Hoboken, New Jersey Published simultaneously in Canada No part of this publicatin may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose No warranty may be created or extended by sales representatives or written sales materials The advice and strategies contained herein may not be suitable for your situation You should consult with a professional where appropriate Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002 Wiley also publishes its books in a variety of electronic formats Some content that appears in print may not be available in electronic formats For more information about Wiley products, visit our web site at www.wiley.com Wiley Bicentennial logo: Richard J Pacifico Library of Congress Cataloging-in-Publication Data: Computational toxicology : risk assessment for pharmaceutical and environmental chemicals / edited by Sean Ekins p ; cm – (Wiley series on technologies for the pharmaceutical industry) Includes bibliographical references and index ISBN 978-0-470-04962-4 (cloth) Toxicology – Mathematical models Toxicology – Computer simulation QSAR (Biochemistry) I Ekins, Sean II Series [DNLM: Toxicology – methods Computer Simulation Drug Toxicity Environmental Pollutants – toxicity Risk Assessment QV 602 C738 2007] RA1199.4.M37C66 2007 615.9001′5118–dc22 2006100242 Printed in the United States of America 10 To Maggie It is very evident, that all other methods of improving medicine have been found ineffectual, by the stand it has been at these two or three thousand years; and that since of late mathematicians have set themselves to the study of it, men already begin to talk intelligibly and comprehensibly, even about abstruse matters, that it may be hop’d in a short time, if those who are designed for this profession are early, while their minds and bodies are patient of labour and toil, initiated in the knowledge of numbers and geometry, that mathematical learning will be the distinguishing mark of a physician from a quack: and that he who wants this necessary qualification, will be as ridiculous as one without Greek or Latin Richard Mead A mechanical account of poisons in several essays 2nd edition, London, 1708 CONTENTS SERIES INTRODUCTION PREFACE xiii ACKNOWLEDGMENTS CONTRIBUTORS PART I xi INTRODUCTION TO TOXICOLOGY METHODS An Introduction to Toxicology and Its Methodologies xv xvii Alan B Combs and Daniel Acosta Jr In vitro Toxicology: Bringing the In silico and In vivo Worlds Closer 21 Jinghai J Xu Physiologically Based Pharmacokinetic and Pharmacodynamic Modeling 33 Brad Reisfeld, Arthur N Mayeno, Michael A Lyons, and Raymond S H Yang Species Differences in Receptor-Mediated Gene Regulation 71 Edward L LeCluyse and J Craig Rowlands Toxicogenomics and Systems Toxicology 99 Michael D Waters, Jennifer M Fostel, Barbara A Wetmore, and B Alex Merrick vii viii CONTENTS PART II COMPUTATIONAL METHODS Toxicoinformatics: An Introduction 151 153 William J Welsh, Weida Tong, and Panos G Georgopoulos Computational Approaches for Assessment of Toxicity: A Historical Perspective and Current Status 183 Vijay K Gombar, Brian E Mattioni, Craig Zwickl, and J Thom Deahl Current QSAR Techniques for Toxicology 217 Yu Zong Chen, Chun Wei Yap, and Hu Li PART III APPLYING COMPUTERS TO TOXICOLOGY ASSESSMENT: PHARMACEUTICAL The Prediction of Physicochemical Properties 239 241 Igor V Tetko 10 Applications of QSAR to Enzymes Involved in Toxicology 277 Sean Ekins 11 QSAR Studies on Drug Transporters Involved in Toxicology 295 Gerhard F Ecker and Peter Chiba 12 Computational Modeling of Receptor-Mediated Toxicity 315 Markus A Lill and Angelo Vedani 13 Applications of QSAR Methods to Ion Channels 353 Alex M Aronov, Konstantin V Balakin, Alex Kiselyov, Shikha Varma-O’Brien, and Sean Ekins 14 Predictive Mutagenicity Computer Models 391 Laura L Custer, Constantine Kreatsoulas, and Stephen K Durham 15 Novel Applications of Kernel–Partial Least Squares to Modeling a Comprehensive Array of Properties for Drug Discovery 403 Sean Ekins, Mark J Embrechts, Curt M Breneman, Kam Jim, and Jean-Pierre Wery 16 Homology Models Applied to Toxicology 433 Stewart B Kirton, Phillip J Stansfeld, John S Mitcheson, and Michael J Sutcliffe 17 Crystal Structures of Toxicology Targets 469 Frank E Blaney and Ben G Tehan 18 Expert Systems Philip N Judson 521 ix CONTENTS 19 Strategies for Using Computational Toxicology Methods in Pharmaceutical R&D 545 Lutz Müller, Alexander Breidenbach, Christoph Funk, Wolfgang Muster, and Axel Pähler 20 Application of Interpretable Models to ADME/TOX Problems 581 Tomoko Niwa and Katsumi Yoshida PART IV 21 APPLYING COMPUTERS TO TOXICOLOGY ASSESSMENT: ENVIRONMENTAL The Toxicity and Risk of Chemical Mixtures 599 601 John C Lipscomb, Jason C Lambert, and Moiz Mumtaz 22 Environmental and Ecological Toxicology: Computational Risk Assessment 625 Emilio Benfenati, Giovanna Azimonti, Domenica Auteri, and Marco Lodi 23 Application of QSARs in Aquatic Toxicology 651 James Devillers 24 Dermatotoxicology: Computational Risk Assessment 677 Jim E Riviere PART V 25 NEW TECHNOLOGIES FOR TOXICOLOGY: FUTURE AND REGULATORY PERSPECTIVES Novel Cell Culture Systems: Nano and Microtechnology for Toxicology 693 695 Mike L Shuler and Hui Xu 26 Future of Computational Toxicology: Broad Application into Human Disease and Therapeutics 725 Dale E Johnson, Amie D Rodgers, and Sucha Sudarsanam 27 Computational Tools for Regulatory Needs 751 Arianna Bassan and Andrew P Worth INDEX 777 ‘Omics Databases Treatment Identify Gene/ Protein Functional Groups, Pathways, Networks Clinical Chemistry Other Effects Indexes Absorption, Distribution, Metabolism, Excretion Literature Mining and Knowledge via Curated Resources Sequence Anchoring to Genomes Microarrays/ Proteomics/ Metabolomics Histopathology Model Organism Databases Loader CEBS Knowledgebase Phenotypic Anchoring Effects Databases Iterative Biological Modeling Integrated Systems-Level Understanding Figure 5.2 Framework for systems toxicology The figure indicates the paths from the initial observation (rat in upper left) to an integrated toxicogenomics knowledgebase and thence to systems toxicology The -omics data stream is shown by the clockwise path from rat to knowledgebase, and the traditional toxicology approach is shown in the counterclockwise path The knowledgebase will integrate both data streams, along with literature knowledge, and, by virtue of iterative modeling, lead to a systems toxicology understanding The framework involves “phenotypic anchoring” (to toxicological endpoints and study design information) and “sequence anchoring” (to genomes) of multidomain molecular expression datasets in the context of conventional indices of toxicology, and the iterative biological modeling of resulting data DATABASE OF ARRAY, PROTEOMICS AND TOXICOLOGY DATA ON CHEMICALS, DRUGS AND STRESSORS DATABASE ON GENES AND GENE GROUPS RELEVANT TO ENVIRONMENTAL DISEASE COMPENDIA OF FUNCTIONAL GENE GROUPS WITH ASSOCIATED PATHWAYS AND NETWORKS Retrieve Associated Data DATABASE OF SNPs AND MUTANTS RELEVANT TO ENVIRONMENTAL DISEASE Store and Convert DICTIONARIES AND METADATA Literature and Clinical Info Resources, Tox DBs, NTP Pathways Links Function Protein DBs External Links Query NLM/NCBI Genomic Resources Compound/Class/Structure Effects Gene/Prot Descriptions Gene/Protein Functional Grps Figure 5.5 Conceptual framework for the development of the Chemical Effects in Biological Systems (CEBS) knowledgebase CEBS knowledgebase is a cross-species reference toxicogenomics information system chemicals/stressors and their effects The upper section indicates data associated in CEBS; the center section, external links from CEBS; and the lower section, sample query types that CEBS will support The boxes in the upper section include primary data (blue), important genetic loci (red) and genetic markers such as SNPs (green) The tasks CEBS will carry out are shown with gray boxes In the central section the links that are to databases are shown in gray, and the links to unstructured data are in green Abbreviations: NTP, National Toxicology Program; NLM, National Library of Medicine; NCBI, National Center for Biotechnology Information Figure 11.3 P-glycoprotein pharmacophore model derived from propafenones used for screening of the world drug index Blue: hydrophobic/aromatic; green: hydrophobic; brown: H-bond acceptor; red: positive ionisable N O N O O O N N OH N OH O N O N O N S S N O Figure 11.4 Self-organizing map used for screening the SPECS compound library Figure 12.2 Estradiol binding to the ER as obtained by X-ray crystallography The image was created with PyMol (Delano, WL, The PyMol user’s manual San Cartos, CA: Delano Scientific, 2002.) (a) (b) Figure12.4 (a)Dihydrotestosterone (DHT) and DHT benzoate binding to the androgen receptor Local induced fit is necessary to accommodate the additional volume of the benzoate group The image was created with PyMol See color plates Figure 12.5 Stereo view of 3,5-dichloro-3′-isopropyl-thyronine bound to the thyroid hormone receptor α Details of the binding pocket; dashed lines indicate stabilizing ligand-protein hydrogen bonds Figure 12.6 Stereo view of the Raptor surrogate for the thyroid receptor β with the largest ligand of the training set depicted The front section has been clipped to display inner (wireframe) and outer shells (smooth surface) Areas colored in brown represent hydrophobic properties; areas in red correspond to H-bond acceptors, areas in blue to H-bond donors and green reflects H-bond flip-flops Figure 12.7 AhR surrogate with a bound aza-PAH as generated by Quasar For clarity, the front section has been clipped Areas colored in gray/brown represent hydrophobic properties; areas in green H-bond donor functions; areas in yellow indicate H-bond acceptors and purple domains correspond to H-bond flip-flops No salt bridges are observed in this model as the ligands lack any charged groups (a) (b) Figure 12.8 (a) SR12813 and (b) hyperforin binding to the PXR as obtained by X-ray crystallography When hyperforin binds, two loops move out to provide additional space and a hydrophobic interaction is observed with Ile414 The image was created with PyMol a b c d e Figure 13.1 hERG potassium channel pharmacophores (a) Ekins hERG pharmacophore [26]; blue sphere = positive ionizable, green spheres = hydrophobic (b) Cavali hERG pharmacophore [19] (c, d, e) Aronov hERG pharmacophores [55]; red feature = hydrogen bond acceptor All pharmacophores recreated in PyMol [148] using distance geometry from the intrafeature distance (Å) tables supplied by the corresponding communications F O O N N H H 2N O O Cl cisapride Figure 13.2 Cisapride mapped to the Aronov hERG pharmacophore described in Figure 13.1e [55] recreated in PyMol [148] using distance geometry from the intrafeature distance tables b c Figure 13.4 (b) Catalyst hypogen pharmacophore generated with nine molecules that bind the human sodium channel Na1.3 The molecule conformers were generated with the BEST algorithm with an energy level of 20kcal/mol The model observed versus predicted correlation was r = 0.91, and the three most potent ligands TTX (blue), riluzole (red), and tetracaine (green) mapped to two hydrogen bond acceptor features (green) and one positive ionizable feature (blue) (c) Paroxetine aligned to the pharmacophore (predicted IC50 0.29 μM) Figure 16.2 Schematic representation of the crystal structure of CYP2D6 [44] illustrating the position of amino acids identified as key by model building Hydrogen bonds identified by modeling between Asp301 and the main chain amides of Val119 and Phe120 are denoted by dashed lines (Image produced using Pymol [15].) a b c S0 S1 P S2 V625 S3 T623 S624 S4 C0 C1 Central G648 N Scav Y652 S5 Cavity F656 C2 V659 S6 d e f Figure 16.4 (a) A homology model of Pore domain of hERG (S5–S6) in the closed state [128], based on KcsA Residues identified by mutagenesis to be important in drug binding are highlighted The Cavalli pharmacophore (see text) is shown positioned within the central cavity For clarity, only two of the four subunits of hERG are shown (b) A homology model of the pore domain of hERG in the open state [128], based on KvAP K+ ion binding sites are shown as purple spheres interspersed by water molecules (these sites are denoted Sext, S1, S1, S3, S4, and Scav) (c) The crystal structure of Kv1.2 in the open state [82] The Pro-X-Pro motif is highlighted in black to illustrate its influence on the central cavity (d–f ) Surface representations of the pore domains to illustrate the state of the modeled or crystallized structure, as viewed from the intracellular mouth (Image produced using PyMol [15].) H-G’ H-F H-G H-D H-B’ β4 β1 β3 H-E H-I H-A H-K’ H-H H-L H-K β2 H-C H-J H-B H-J’ (a) R132 D100 E98 D91 D52 E55 R129 D87 D84 R440 R269/K146 R88 R140 R450 (b) D53 D147 D148 (c) (d) Figure 17.2 (a) Crystal structure of CYP2D6 with the helices in purple and the βstrands in yellow The helices are labeled H–A through to H–L The oxygenated heme can be seen just below H–I The helical nature of H–G′ is evident although it was not detected in the software, and is therefore displayed in the cyan box at the top (b) The largely negative electrostatic potential on the distal face of CYP2D6 Positively charged ligands are attracted to this This is in contrast to (c) the proximal face of CYP2D6 where a large number of basic residues are found This is a perfect compliment to the reductase partner in (d), which has clusters of acidic residues which form a tight saltbridged interface with the cytochrome (a) (d) (b) (e) (c) (f) Figure 17.15 Ab initio calculations carried out on debrisoquine in a model reaction system derived on the left from a docking in our homology model and on the right, from the crystal structure of CYP2D6 (a) The spin density is centered on the iron (b) The largest LUMO eigenvector is on the C4 position of debrisoquine, and (c) The largest positive electrostatic potential is again on C4 These features indicate a classic nucleophilic reaction With the orthogonal approach likely from the crystal structure (d) debrisoquine is in a radical state, (e) the LUMO is centered on the heme, and ( f ) the charge is distributed around the complex Figure 17.16 PPARγ/RXRα heterodimer crystal structure with retinoic acid bound to RXR (left) and rosiglitazone in PPARγ (right) Helix 12 can be seen in the front of the PPARγ structure crossing horizontally and blocking the active site pocket S O O OH O NH2 Cl O Br N Cl withdrawn O HO hepatotoxic O OH hepatotoxic at high doses NH2 O N H Bromfenac H N Acetaminophen withdrawn hepatotoxic at high doses O Controls N N O Nefazodone withdrawn N Tacrine N N N Troglitazone Diclofenac Cl N N H O HO NH Cl O O O N O O Clozapine O agranulocytosis O O N N N N N O O Caffeine Acetylsalicilic acid Cl Indomethacine blood disorders H N S N N N N O O H N O N O S NH O S N Olanzapine Rosiglitazone Pioglitazone apparently safe O apparently safe apparently safe Figure 19.11 Examples of pharmaceuticals that are associated with evidence for human hepatotoxicity of varying levels of severity (caffeine and acetylsalicylic acid serve as negatives examples) Functional groups that are believed to be associated with metabolism pathways toward a reactive intermediate are highlighted in red Note that despite structural alerts and proven bioactivation liabilities, some drugs given at comparably low dose are considered as safe (olanzapine, rosiglitazone, and pioglitazone) Their structural analogues given at relatively high doses (clozapine, troglitazone) have been associated with severe clinical side effects Figure 20.3 Modeled structure of a hERG channel Amino acids are colored according to their physicochemical properties Amino acids reported to participate in the binding of drugs [22] are labeled For simplicity, only two of the four subunits that form the channel are shown a b Figure 20.4 X-ray structures of the binding sites of the Abl kinase/imatinib and cytochrome P450 3A4 /erythromycin complexes (PDB: Protein Data Bank) The amino acids located within Å of ligands are depicted The hydrogen bonding interactions are shown by dashed lines, and the amino acids participating in hydrogen-bonding interactions are labeled in bold The images were produced with Pymol (DeLano Scientific LLC) Figure 25.1 Schematics of typical nano and micro fabrication techniques: Photolithography, soft lithography, hot embossing, and direct writing Figure 25.3 pathways A typical μCCA device filled with red dye for visualization of fluidics Wiley Series on Technologies for the Pharmaceutical Industry Sean Ekins, Series Editor Editorial Advisory Board Dr Renee Arnold (ACT LLC, USA) Dr David D Christ (SNC Partners LLC, USA) Dr Michael J Curtis (Rayne Institute, St Thomas’ Hospital, UK) Dr James H Harwood (Pfizer, USA) Dr Dale Johnson (Emiliem, USA) Dr Mark Murcko (Vertex, USA) Dr Peter W Swaan (University of Maryland, USA) Dr David Wild (Indiana University, USA) Prof William Welsh (Robert Wood Johnson Medical School University of Medicine & Dentistry of New Jersey, USA) Prof Tsuguchika Kaminuma (Tokyo Medical and Dental University, Japan) Dr Maggie A.Z Hupcey (PA Consulting, USA) Dr Ana Szarfman (FDA, USA) Computational Toxicology: Risk Assessment for Pharmaceutical and Environmental Chemicals Edited by Sean Ekins ... Cataloging-in-Publication Data: Computational toxicology : risk assessment for pharmaceutical and environmental chemicals / edited by Sean Ekins p ; cm – (Wiley series on technologies for the pharmaceutical. . .COMPUTATIONAL TOXICOLOGY COMPUTATIONAL TOXICOLOGY Risk Assessment for Pharmaceutical and Environmental Chemicals Edited by SEAN EKINS WILEY-INTERSCIENCE A JOHN WILEY... Introduction to Toxicology Methods Computational Methods Applying Computers to Toxicology Assessment: Pharmaceutical Applying Computers to Toxicology Assessment: Environmental New Technologies for Toxicology: