DRUG DISCOVERY STRATEGIES AND METHODS

DRUG DISCOVERY STRATEGIES METHODS EDITED BY ALEXANDROSMAKRIYANNIS DIANEB~EGEL Center for Drug Discovery University o...

Author: Alexandros Makriyannis | Diane Biegel

43 downloads 1102 Views 8MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

DRUG DISCOVERY

STRATEGIES METHODS EDITED BY

ALEXANDROSMAKRIYANNIS

DIANEB~EGEL

Center for Drug Discovery University of Connecticut Storrs, Connecticut, U.S.A.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Although great care has been taken to provide accurate and current information, neither the author(s) nor the publisher, nor anyone else associated with this publication, shall be liable for any loss, damage, or liability directly or indirectly caused or alleged to be caused by this book. The material contained herein is not intended to provide speciﬁc advice or recommendations for any speciﬁc situation. Trademark notice: Product or corporate names may be trademarks or registered trademarks and are used only for identiﬁcation and explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress. ISBN: 0-8247-0691-9 This book is printed on acid-free paper. Headquarters Marcel Dekker, Inc. 270 Madison Avenue, New York, NY 10016 tel: 212-696-9000; fax: 212-685-4540 Eastern Hemisphere Distribution Marcel Dekker AG Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland tel: 41-61-260-6300; fax: 41-61-260-6333 World Wide Web http://www.dekker.com The publisher oﬀers discounts on this book when ordered in bulk quantities. For more information, write to Special Sales/Professional Marketing at the headquarters address above. Copyright n n 2004 by Marcel Dekker, Inc. All Rights Reserved. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microﬁlming, and recording, or by any information storage and retrieval system, without permission in writing from the publisher. Current printing (last digit): 10 9 8 7 6 5 4 3 2 1 PRINTED IN THE UNITED STATES OF AMERICA

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Preface

Drug research encompasses diverse branches of science united by a common goal, namely, developing novel therapeutic agents and understanding their molecular mechanisms of action. This process is a lengthy, exacting, and expensive undertaking that involves integration of data from diﬀerent ﬁelds and culminates in the ﬁnal product—a new drug in the marketplace. In the past decade, progress in drug research has ﬂourished because of major contributions from a variety of disciplines. The material presented in this volume focuses on a number of research topics that have provided critical information in the ﬁeld of drug discovery. Several chapters present techniques that extend our understanding of the three-dimensional structure of macromolecules, principally proteins, but also nucleic acid polymers and organized lipid and carbohydrate assemblies. As greater structural data on the these molecules become available, information can be obtained on their interactions with small endogenous ligand drug molecules as well as on the interactions between two or more of these biopolymers. Such knowledge enhances our overall understanding of the biochemical systems of interest and their relevance for therapeutic discovery. In addition to the basic knowledge gained by such research, the data provide a solid basis for the development of novel drugs with greater potencies, higher speciﬁcities of action, and reduced side eﬀects. Another area of research covered in this book is the in vivo anatomical localization of potential therapeutics using PET and SPECT analysis (Chapter 5). These techniques allow researchers to pinpoint the localization of high-aﬃnity ligands in the living organism with high accuracy, thus giving researchers a window on the functions of the brain and other organs and on the sites of action of potential therapeutic agents. Such studies will provide a blueprint for the design of pharmacological

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

agents that will target speciﬁc regions of aﬀected organs and deliver therapeutic actions rapidly and with high speciﬁcity. High throughput methods have increased our capacity for appropriate candidate compounds selection and also for developing libraries of novel compounds from which such candidates can be selected. Chapter 7 discusses the use of solid-phase synthesis for the high throughput production of peptides and other small molecules. In addition, as discussed in Chapter 6 on peptidomimetics, the swift production of novel leads holds considerable promise for future discovery of novel therapeutic agents. The investigation of therapeutic targets for cannabinoid sites of action has already generated considerable interest within the ﬁeld of drug discovery, and Chapter 4, which details the results of such studies, highlights the importance of target-based studies. The enhanced appreciation of the role of stereochemistry in drug action has focused eﬀorts on understanding the conformation of drugs as they bind to their target receptor. Studies of the diverse eﬀects of cannabinoids and the development of compounds that employ the information gleaned from the ligand/receptor data should provide substantial insight into their molecular mechanisms of action. Future research will promote the development of drugs that are capable of higher speciﬁcity. longer half-lives, and lessened toxicity. In studies of potential antiviral therapies, the understanding of viral target molecules is essential for the production of eﬀective medications that interact speciﬁcally in the viral life cycle and gene products, which will result in lowering drug toxicity to the host and enhancing the antiviral activity of the pharmacotherapy. As the nature of viral infectivity, cell growth, death, and receptor biology are elucidated, the methods and paradigms for development of highly speciﬁc medications will provide superior treatments for a number of diseases that pose a terrible burden worldwide (Chapters 10 and 11). From the ﬁelds of proteomics and genomics that hold signiﬁcant promise for unique medications, several areas of biology have also found applications in the drug discovery arena. The study of regulatory molecules and oncogenes has opened new avenues in drug therapy, as discussed in Chapter 8 on G-protein-coupled receptors and Chapter 2 on SRC homology domains. Research on protein misfolding (Chapter 9), which has been implicated in neurodegenerative diseases, has highlighted the need to enhance our understanding of structural alterations in normal proteins products. Chapter 1 details the development of such research, and asserts that only as we understand the basic physical mechanisms of such alterations can new therapeutic regimens be proposed and tested.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

The topics included in this volume are not intended to be allinclusive. Our approach has been eclectic, in an eﬀort to bring the reader the most exciting aspects of drug discovery, along with the methods that show the most promise in enhancing the discovery process. The chapters presented in this book have been contributed by specialists in their areas of research and will provide a contemporary picture of the overall ﬁeld of drug discovery to scientists from diverse disciplines. Alexandros Makriyannis Diane Biegel

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Contents

Preface Contributors

1. Protein Crystallography in Structure-Based Drug Design Xiayang Qiu and Sherin S. Abdel-Meguid 2. Src Homology-2 Domains and Structure-Based, SmallMolecule Library Approaches to Drug Discovery Chester A. Metcalf III and Tomi Sawyer 3. Three-Dimensional Structure of the Inhibited Catalytic Domain of Human Stromelysin-1 by Heteronuclear NMR Spectroscopy Paul R. Gooley 4. Cannabinergics: Old and New Possibilities Andreas Goutopoulos and Alexandros Makriyannis 5. Development of PET and SPECT Radioligands for Cannabinoid Receptors S. John Gatley, Andrew N. Giﬀord, Yu-Shin Ding, Ruoxi Lan, Qian Liu, Nora D. Volkow, and Alexandros Makriyannis

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

6. Structural and Pharmacological Aspects of Peptidomimetics Peter W. Schiller, Grazyna Weltrowska, Ralf Schmidt, Thi M.-D. Nguyen, Irena Berezowska, Carole Lemieux, Ngoc Nga Chung, Katharine A. Carpenter, and Brian C. Wilkes 7. Linkers and Resins for Solid-Phase Synthesis: 1997-1999 Pan Li, Elaine K. Kolaczkowski, and Steven A. Kates 8. Allosteric Modulation of G-Protein-Coupled Receptors: Implications for Drug Action Angeliki P. Kourounakis, Pieter van der Klein, and Ad P. I. IJzerman 9. Protein Misfolding and Neurodegenerative Disease: Therapeutic Opportunities Harry LeVine III 10. Uncoating and Adsorption Inhibitors of Rhinovirus Replication Guy D. Diana and Adi Treasurywala 11. Proﬁles of Prototype Antiviral Agents Interfering with the Initial Stages of HIV Infection E. De Clercq

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Contributors

Sherin S. Abdel-Meguid Suntory Pharmaceutical Research Laboratories, Cambridge, Massachussets, U.S.A. Irena Berezowska Quebec, Canada

Clinical Research Institute of Montreal, Montreal,

Katharine A. Carpenter treal, Quebec, Canada Ngoc Nga Chung Quebec, Canada

Clinical Research Institute of Montreal, Mon-

Clinical Research Institute of Montreal, Montreal,

Eric De Clercq Rega Institute for Medical Research, Katholieke Universiteit Leuven, Leuven, Belgium Guy D. Diana ViroPharma, Inc. Exton, Pennsylvania, U.S.A. Yu-Shin Ding U.S.A.

Brookhaven National Laboratory, Upton, New York,

S. John Gatley U.S.A.

Brookhaven National Laboratory, Upton, New York,

Andrew N. Giﬀord York, U.S.A. Paul R. Gooley

Brookhaven National Laboratory, Upton, New

University of Melbourne, Parkville, Victoria, Australia

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Andreas Goutopoulos Serono Reproductive Biology Institute, Rockland, Massachusetts, U.S.A. Ad P. IJzerman

Leiden University, Leiden, The Netherlands

Steven A. Kates

Surface Logix, Inc., Brighton, Massachusetts, U.S.A.

Elaine K. Kolaczkowski chussetts, U.S.A.

Vertex Pharmaceuticals, Cambridge, Massa-

Angeliki P. Kouranakis Greece

University of Thessaloniki, Thessaloniki,

Ruoxi Lan

University of Connecticut, Storrs, Connecticut, U.S.A.

Carole Lemieux Quebec, Canada

Clinical Research Institute of Montreal, Montreal,

Harry LeVine III

University of Kentucky, Lexington, Kentucky, U.S.A.

Pan Li Vertex Pharmaceuticals, Cambridge, Massachusettes, U.S.A. Qian Liu

University of Connecticut, Storrs, Connecticut, U.S.A.

Alexandros Makriyannis U.S.A. Chester A. Metcalf III sachusetts, U.S.A. Thi M.-D. Nguyen Quebec, Canada

University of Connecticut, Storrs, Connecticut,

ARIAD Pharmaceuticals, Inc., Cambridge, Mas-

Clinical Research Institute of Montreal, Montreal,

Xiayang Qiu SmithKline Beecham Pharmaceuticals, King of Prussia, Pennsylvania, U.S.A. Tomi Sawyer setts, U.S.A.

ARIAD Pharmaceuticals, Inc., Cambridge, Massachu-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Clinical Research Institute of Montreal, Montreal,

Peter W. Schiller Quebec, Canada Ralf Schmidt Canada

Clinical Research Institute of Montreal, Montreal, Quebec,

Pﬁzer Central Research, Groton, Connecticut, U.S.A.

Adi Treasurywala Pieter van der Klein Nora D. Volkow

NIDA, Bethesda, Maryland, U.S.A.

Grazyna Weltrowska Quebec, Canada Brian C. Wilkes Quebec, Canada

Leiden University, Leiden, The Netherlands

Clinical Research Institute of Montreal, Montreal,

Clinical Research Institute of Montreal, Montreal,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

1 Protein Crystallography in Structure-Based Drug Design Xiayang Qiu SmithKline Beecham Pharmaceuticals, King of Prussia, Pennsylvania, U.S.A.

Sherin S. Abdel-Meguid Suntory Pharmaceutical Research Laboratories, Cambridge, Massachusetts, U.S.A.

I.

INTRODUCTION

Proteins are responsible for a wide variety of important biological functions in living organisms and are commonly used as targets of therapeutic agents. A unique primary and tertiary structure is a hallmark property of a protein. Although several related and even unrelated proteins may share the same overall tertiary structure or fold, each will diﬀer from the others in the details. Knowledge of the detailed atomic three-dimensional structure of the protein and/or its ligand complexes should facilitate the design of novel, high aﬃnity ligands that interact with that protein. The process of elucidating the atomic structure of proteins and their complexes, and the design of novel, therapeutically relevant ligands based on these structure elucidations, is known as structure-based drug design. Proteins are complex molecules, typically containing several thousand atoms. Although Pauling and Corey proposed the a helix and the h sheet as the main secondary structural elements of proteins in 1951, and the crystal structure of myoglobin was reported by John Kendrew in 1958,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

crystal structure determination in the early days were hampered by numerous technical limitations and usually required many years of hard work. By the mid-1980s, substantial improvements in data acquisition software and hardware had considerably accelerated the speed with which a crystal structure could be determined. Trying to capitalize on the potential of structure-based drug design, several pharmaceutical companies built their own protein crystallography laboratories, and a number of structure-based drug design eﬀorts emerged in industrial and academic laboratories [1]. In the past 10 years, we have experienced a sudden burst in the number of protein three-dimensional structures determined. By the end of the twentieth century, merely 40 years since the ﬁrst protein structure was solved, there were over 11,000 structures deposited in the Protein Data Bank (PDB). Although each entry is not a unique protein, the number of novel structures deposited in the PDB has increased sharply during the last decade. These proteins include not only soluble proteins, but also a number of membrane proteins. Furthermore, structures of protein–protein and protein–nucleic acid complexes, viruses, and the ribosome are also available. This marvelous scientiﬁc achievement was mostly credited to the method of single-crystal x-ray diﬀraction (protein crystallography), although a notable number of structures were determined by means of NMR spectroscopy. Many factors in addition to the incredible advances in computer hardware and software contributed to the improved eﬃciency and precision in protein crystallography: the advent of molecular biology, which allows for cloning, mutation, and overexpression of many targets that are diﬃcult to isolate from natural sources; advances in protein puriﬁcation that facilitate the production of large amounts of highly puriﬁed proteins; improvement in protein characterization and crystallization strategies; enhancement of data acquisition techniques and equipments; access to powerful synchrotron radiation sources; and introduction of the selenomethionine multiple-wavelength anomalous diﬀraction (MAD) procedure for phase determination. Currently, almost all large pharmaceutical and numerous biotechnology companies have established in-house macromolecular crystallography units, and the crystallographic community is solving thousands of new structures every year. With structural information becoming more readily available, structure-based drug design has become an integral part of the modern drug discovery process and has begun to contribute to a signiﬁcant portion of the current drug discovery portfolio.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Identifying and bringing a successful small-molecule drug to the market requires considerable eﬀort, which typically costs millions of dollars and may span as much as10 years. With this time scale in mind, one must realize that structure-based drug design is still in its infancy, having started in earnest in the mid-1980s. While the concept of such a rational approach has been around for some time, for much of the work in the ﬁeld it is still too early to demonstrate market successes. Moreover, although structural knowledge may be used for lead generation and lead optimization, or even for addressing some developability issues, it does little to address other important issues in drug development ranging from the appropriateness of targets or disease models to government regulatory issues or changing market forces. In fact, drug discovery is a risky business in that only a very small number of compounds are able to ﬁnd their way to the market. Therefore, the successful structure-based design and the launch of inhibitors of HIV protease [2] and inﬂuenza virus neuraminidase [3] as drugs are particularly encouraging events for the ﬁeld [4–9]. In this chapter, we will introduce the technique of protein crystallography and its use in structure-based drug design, point out the technical challenges ahead of us, and report many practical lessons learned during the past decade of structure-based drug design.

II. THE DRUG DISCOVERY PROCESS The many steps of the complex and multidisciplinary drug discovery process can be grouped into four major phases: target identiﬁcation and validation, lead identiﬁcation, lead optimization, and biological testing (Fig. 1). Choosing an appropriate target is usually the ﬁrst step in the drug discovery process. Target selection requires an understanding of human diseases and the biological processes that lead to a particular disease. Although historically drugs (e.g., h-lactams) were discovered without knowledge of their molecular target, knowing the target greatly enhances one’s ability to discover novel drugs in a timely fashion. Recent advances in sequencing the human genome, as well as the genomes of many human pathogens, have provided a large pool of potential novel molecular targets. Most future drug discovery eﬀorts will start with a relatively unknown gene selected from a sequence database based on one or more attractive features that could provide a hint of its function, such as tissue distribution, genome

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 1

Simplified drug discovery process.

localization, and/or sequence homology or structural analogy to a known protein. Cloning, expression, puriﬁcation, and characterization of the protein target and other tool reagents such as antibodies or receptors will usually follow, to be used in target validation with a set of appropriate genetic and biological assays. The second step is to identify a suitable lead molecule to interact with the molecular target. This is usually achieved through high throughput screening of available chemical compound libraries and natural products, typically containing hundreds of thousands of compounds. Although the size of the library per se is not critical, a library that contains a large number of molecules is essential to assure molecular diversity. Novel lead molecules can also be designed by analysis of the threedimensional structure of the target molecule in a process known as de novo design. A desirable lead should usually have at least low micromolar binding potency against the target and should be amenable to further synthetic manipulations. The third step is to optimize the lead molecule through iterative chemical synthesis and biological testing, aiming to obtain molecules with the required potency (typically nanomolar), selectivity, bioavailability, and DMPK (drug metabolism and pharmacokinetics) properties. This step usually requires considerable time and resources; usually the synthesis of hundreds of compounds is needed to deduce a robust SAR (structure– activity relationship). Such resources can be considerably reduced and the

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

time signiﬁcantly shortened if optimization employs knowledge from the three-dimensional structure of complexes of leads with the target. The last step of the drug discovery process involves the testing of lead compounds to address issues such as eﬃcacy, bioavailability, and safety. Testing may include in vitro assays but ultimately would require a suitable disease model and studies in animals. Many compounds may need to be designed and synthesized to identify the one compound with all the desired properties. Such a compound can be advanced to preclinical studies and eventually to the clinic.

III. THE STRUCTURE-BASED DRUG DESIGN CYCLE Timely optimization of lead compounds requires knowledge of the threedimensional structure of target–ligand complexes. Protein crystallography has been the predominant technique used to elucidate the three-dimensional structure of proteins in structure-based drug design. Crystallographic studies usually consume tens of milligrams of pure protein and take several months to yield the ﬁrst crystal structure. Therefore, one should start crystallographic eﬀorts as soon as suitable material is available, preferably even before initiation of high throughput screening. Once a lead has been identiﬁed through high throughput screening or de novo design, structure determinations of target–ligand complexes should be pursued. The use of information derived from the structure determination of the target bound to the initial lead molecule should allow for the design and synthesis of new ligands with improved properties, as well as the initiation of further rounds of structure-based design. Through iterations of structure determination, design, synthesis, and biological testing (Fig. 2) a drug candidate should emerge. In addition to lead optimization and lead identiﬁcation, three-dimensional structures of the target–ligand complexes can contribute to the traditional drug discovery process in other ways. For example, structural information combined with genomic sequences may aid in target identiﬁcation by helping to classify genes with unknown functions. Structures can be used as templates for de novo design or in silico lead identiﬁcation through screening of virtual libraries. Structural information can provide a basis for the design of directed combinatorial libraries [10,11]. Moreover, structural studies of leads with serum albumin and various cytochrome P450s should allow for a better understanding of some of the developability issues that may arise during drug development.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 2

Structure-based design cycle.

IV. PROTEIN CRYSTALLOGRAPHY For most noncrystallographers, protein crystallography tends be a black box full of jargon. Here, we give a brief overview of the technology in an attempt to demystify some of the terms used.

A. Crystallization Obtaining large single crystals that diﬀract to high resolution remains the primary bottleneck of protein crystallography. The most widely used

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

crystallization method is the hanging-drop method of vapor diﬀusion (Fig. 3), in which a drop (1 or 2 AL) of protein is mixed with an equal volume of a precipitant on a glass coverslip and is sealed over a well containing the same precipitant added to the protein. Many factors are known to be important in protein crystallization: protein purity (preferably >95% pure) and concentration (typically 10 mg/mL), the nature of precipitant [e.g., poly(ethylene glycol) or various salts] and its concentration, the nature, concentration, and pH of the buﬀer, the presence or absence of additives (e.g., metal ions, reducing agents, protease inhibitors, metal chelators, detergents) and eﬀectors (e.g., ligands, cofactors, substrates, inhibitors), the rate of equilibrium between the protein and the precipitant, the crystallization temperature, and so on. Since there are no general rules to correlate all these factors to the eventual success in obtaining crystals, protein crystallization remains a trial-and-error process and a signiﬁcant bottleneck in protein crystallography: failure rate is typically 50% even with thousands of crystallization trials. Many methods and techniques have been employed to enhance one’s ability to obtain protein crystals. Molecular biology and biochemical methods have been utilized to generate domains of large proteins that may be less ﬂexible and thus more amenable to crystallization. Biophysical tools such as dynamic light scattering [12] and ultracentrifugation [13] have been used to study protein aggregation in solution. Molecular biology has been employed to generate mutants that do not aggregate or are more soluble. Crystallization trials using incomplete factorial designs [14] allow the screening of a much wider range of conditions with a modest number of experiments, and

Figure 3 Protein crystallization: diagrammatic representation of the hangingdrop method of vapor diffusion.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

thus less protein. Miniaturization and automation made possible by the use of advanced crystallization robots may also have a great impact on the future success of protein crystallization.

B. X-Ray Diffraction Data Acquisition The next step is to measure x-ray diﬀraction data from a single crystal (Fig. 4). Data are usually measured by means of an area detector such as a phosphorus image plate or a charge-coupled device (CCD). Through several steps of computational analysis, the position and amplitude or intensity of the each diﬀraction spot can be obtained. Because diﬀraction intensities are proportional to the volume of the crystal and generally decrease at higher resolution, protein crystals must be reasonably large to give strong enough diﬀraction signals at high resolution. While a cube of 0.1 to 0.5 mm in each dimension is still preferred by most crystallographers, the availability of powerful synchrotron radiation sources has made the analysis of much smaller crystals feasible. Crystals also must be stable enough in the x-ray beam to allow the measurement of a complete diﬀraction data set from a single crystal. In this regard, ﬂash-freezing of protein crystals under proper conditions at cryogenic temperatures [15] has virtually eliminated radiation decay problems.

Figure 4 Diagrammatic representation of single-crystal x-ray diffraction and data collection.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

C. Phasing The ultimate goal of an x-ray diﬀraction experiment is to produce an electron density map that is then used to build an atomic model of the molecule being studied (Fig. 5). The use of single-crystal x-ray diﬀraction techniques to determine the three-dimensional structure of molecules requires the measurement of amplitudes and the calculation of phases for each diﬀraction spot. Although amplitudes can be directly measured from diﬀracting crystals, as noted earlier, phases are indirectly determined. The inability to directly measure phases is known as the ‘‘phase problem’’ [16]. In practice, there are several ways to get around the phase problem. If the protein of interest is small (f100 amino acids) and highresolution data (1.2 A˚ or better) are available, phases can be obtained computationally by using the so-called direct method. This is basically the same technique used to determine crystal structures of small organic molecules. If the protein being studied is known to have a fold similar to that of a protein with a known three-dimensional structure, one uses the molecular replacement (MR) method, in which the known structure serves as a model to generate approximate phases that are then reﬁned against the experimental data obtained from crystals of the protein under study. Until recently, multiple isomorphous replacement (MIR) was the most widely used method for ab initio phase determination. This technique requires the introduction into the protein under study of atoms of high atomic number (heavy atoms) such as mercury, platinum, and uranium, without disrupting the protein’s three-dimensional structure or the packing in the crystal. This is achieved by soaking crystals in a solution

Figure 5 Steps in the use of protein crystallography for structure determination.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

containing the desired heavy atom. The binding of one or more heavy atoms to the protein alters the diﬀraction of the crystals from that of the underivatized (native) crystals. If the introduction of heavy atoms is truly isomorphous, the diﬀerences between the diﬀraction of the derivative and of the native will represent only contributions from the heavy atom(s). Thus, the problem of structure determination is reduced to locating the position of one or a few heavy atoms. Once their positions have been accurately determined, the heavy atoms are used to calculate phases for all diﬀraction intensities. In theory, one needs only two isomorphous derivatives, but in practice more are needed owing to errors that are introduced in data measurement as well as the lack of isomorphism. Multiple-wavelength anomalous dispersion (MAD) phasing, cited earlier, has gained popularity in the last 10 years, and this more recent technique for ab initio phase determination is now the predominant method in de novo structure determination. In the MAD technique, cells that overexpress the protein can be grown in media containing selenomethionine (Se-Met) instead of methionine, producing proteins that have Se-Met at all the methionine positions. Because of the unique absorption quality of Se, diﬀraction data can be measured by using a Se-Met-substituted crystal at three or four diﬀerent wavelengths around the Se absorption edge. These data can be analyzed by using computational methods to generate phase information, allowing an electron density map to be calculated [17]. Such an experiment calls for modern synchrotron facilities.

D. Model Building and Refinement Once an electron density map has become available, atoms may be ﬁtted into the map by means of computer graphics to give an initial structural model of the protein. The quality of the electron density map and structural model may be improved through iterative structural reﬁnement but will ultimately be limited by the resolution of the diﬀraction data. At low resolution, electron density maps have very few detailed features (Fig. 6), and tracing the protein chain can be rather diﬃcult without some knowledge of the protein structure. At better than 3.0 A˚ resolution, amino acid side chains can be recognized with the help of protein sequence information, while at better than 2.5 A˚ resolution solvent molecules can be observed and added to the structural model with some conﬁdence. As the resolution improves to better than 2.0 A˚ resolution, ﬁtting of individual atoms may be possible, and most of the

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 6 Electron density of an a-helix at different resolutions.

amino acid side chains can be readily assigned even in the absence of sequence information.

E. Understanding Structural Coordinates Once a crystal structure has been determined, the information is communicated in the form of an atomic coordinates ﬁle. In addition to a list of the atomic positions, the coordinates ﬁle contains other information that deserves an explanation and requires attention by the user. Some of the terms included in an atomic coordinates ﬁle are explained brieﬂy. It is hoped that the information will provide the reader with insights to evaluate the quality of the structure, distinguish between its well-deﬁned and ﬂexible regions, and make sensible decisions in structural analysis. The unit cell is the basic microscopic building block of the crystal. A crystal can be viewed as a three-dimensional stack of identical unit cells, each deﬁned by three cell edges (a, b, c, in angstroms), and three angles (a, h, g in degrees) between each pair of edges. Each unit cell may contain one or more protein molecules related by crystal symmetry. The unique portion of the unit cell (i.e., the portion that is not related to other portions by crystal symmetry) is called the asymmetric unit. There are only 230 diﬀerent combinations of symmetry elements in crystals; each of these is called a space group. However, since biological molecules are enantiomorphic,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

which means that a protein crystal cannot contain mirror planes, the number of space groups of relevance to protein crystallography is reduced to 65. It is possible to have more than one copy of the same protein in an asymmetric unit. However, these will be related by ‘‘noncrystallographic’’ symmetry. Therefore, all atoms of an asymmetric unit, along with the unit cell dimensions and the space group, must be given in the coordinates ﬁle for subsequent analysis and for regenerating the structure in any portion of the unit cell or the crystal, which may be important for studying intermolecular ‘‘crystal packing’’ interactions. The R-factor is probably the single most important number that provides a sense of the overall quality of the structure. It is deﬁned as [A||Fobs| k*|Fcalc||] / A|Fobs|, where Fobs is the observed structure factor (the square root of the measured diﬀraction intensity or amplitude), Fcalc is the structure factor calculated from the model, and k is a scaling factor. The factor R is a measurement of the agreement between the structural model and the observed diﬀraction data; the lower the number, the better. For a reﬁned crystal structure, the R factor is often approximately 10 times its resolution (e.g., 20% for a 2.0 A˚ resolution structure). Along with the traditional R factor, most of the recent structures also report an Rfree value, which is obtained from the part of the diﬀraction data (5–10%) set aside and not used during structural reﬁnement. Generally Rfree is 5–10% higher than R; larger discrepancies between the two may indicate that there is a problem in the structure model or diﬀraction data, or that the structure is overreﬁned against the data. Reducing R to below 20% used to be the goal for structural reﬁnement; but obtaining a sensible Rfree is now considered to be more important. Therefore, before analyzing a crystal structure on computer graphics, one should check the R factor and Rfree values to get a sense of the overall quality of the structure. It is important to note that these values can be reported as percentages (20%) or as fractions (0.20). The atomic temperature factor, or B factor, measures the dynamic disorder caused by the temperature-dependent vibration of the atom, as well as the static disorder resulting from subtle structural diﬀerences in diﬀerent unit cells throughout the crystal. For a B factor of 15 A˚2, displacement of an atom from its equilibrium position is approximately 0.44 A˚, and it is as much as 0.87 A˚ for a B factor of 60 A˚2. It is very important to inspect the B factors during any structural analysis: a B factor of less than 30 A˚2 for a particular atom usually indicates conﬁdence in its atomic position, but a B factor of higher than 60 A˚2 likely indicates that the atom is disordered.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

For a particular crystal, the number of diﬀraction data increases as the resolution increases, which means that more experimental data will be available for structural reﬁnement. There are four parameters to be reﬁned for each atom: x, y, z (atomic position), and B (temperature factor). If the crystal has normal solvent content (i.e., about 50%), the number of experimental data and reﬁnement parameters will be about the same at 2.8 A˚ resolution. This suggests that B factors for individual atoms should be reﬁned only when data have a resolution better than 2.8 A˚. Reﬁnement of atomic B factors at lower resolution will have no physical meaning, although a lower but meaningless R factor will result. Identiﬁcation and reﬁnement of solvent molecules (e.g., waters) become reliable only when the structure has at least a 2.5 A˚ resolution. Even then, before a water molecule is used in mechanistic or computational analysis, it is always wise to check its B factor for the existence of at least one hydrogen bond to hold the water to the protein. At times, spurious water molecules are added (such additions will result in a meaningless lower R factor). Unless the structure has been determined at a reasonably high resolution, electron density and reﬁnement often do not discriminate between the oxygen and nitrogen atoms of asparagines and glutamines, or the alternative conformations of histidine side chains. In a detailed structural analysis, it may be necessary to check alternative conformations of Asn, Gln, or His side chains and decide which one makes more sense chemically.

V. IN SILICO LEAD GENERATION Armed with the crystal structure of the protein–ligand complex and upto-date computer modeling software, one can design additional ligands. Numerous molecular modeling software programs are available for that purpose. However, it is important to note that current computational algorithms have their limitations and utilize many approximations. Therefore, while computer modeling software has been proven useful [4,18], further testing and structural validations are required to identify the best possible compound.

A. In Silico Screening of Virtual Compound Libraries Starting with the crystal structure of the target, it is possible to screen for leads in three-dimensional compound databases such as the Cambridge

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Structural Database [19] or the Chemical Abstracts Service Registry [20], or convert private databases to 3-D structures by programs such as CONCORD [21]. Several programs are available for such screening. For example, DOCK [22] works by using a set of overlapping spheres to create a complementary image of the ligand binding site and essentially matching the shape of a putative ligand with that of the image to generate a ‘‘goodness of ﬁt’’ score that is then used to rank the hits identiﬁed. Instead of comparing shapes, the program LUDI [23] uses parameters that describe hydrogen-bonding potential and hydrophobic complementarity to match the ligand and its binding site. These programs can rapidly search through three-dimensional databases of small molecules and rank each candidate. Typically, the 100 to 200 top-scoring compounds are examined graphically to identify the best 10 to 50 candidates for experimental testing. In the case of DOCK, 2 to 20% of these in silico hits may show micromolar binding aﬃnity [4]. Subsequently, crystallography can be used to optimize any leads identiﬁed.

B. Building Leads from Molecular Fragments Again starting with the crystal structure of the target, another strategy is to dock small chemical fragments into the ligand binding site, then grow the fragment to better complement the binding site. Programs such as GRID [24], AUTODOCK [25], and MCCS [26] can be used for the docking step. GRID uses small functional groups to probe the binding site and evaluate interaction energies by using an empirical Lennard-Jones energy function, as well as electrostatic and hydrogenbonding terms. AUTODOCK uses simulated annealing for ligand conformational search to dock small ligands of ﬂexible conformations onto a rigid binding site and a standard force ﬁeld for rapid grid-type energy evaluation. MCSS (multicopy simultaneous search) places thousands of copies of functional groups in the binding site and optimizes them simultaneously to generate energetically favorable positions and orientations in a ﬂexible binding site. Once selected, suitable binding fragments can be built into a single compound by manual modeling or by using linking programs such as CAVEAT [27], which attempts to identify a suitable cyclic linker from a database. Alternatively, programs like GroupBuild [28] can search compound libraries for potential leads that have the functional fragments identiﬁed by the programs just described.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

VI. STRUCTURE-BASED LEAD OPTIMIZATION Once a chemical lead has been identiﬁed, the structure-based lead optimization process goes through several iterations of structure determination, design, chemical synthesis, and biological testing. The goal is to optimize the lead in terms of electrostatic interactions, van der Waals contacts, and the ﬁt in the ligand binding pocket. The design process may be simple and intuitive if one starts with a relatively high aﬃnity lead. In this case only minor modiﬁcations to the existing lead are introduced at each of the iterations of the drug design cycle. Many of these modiﬁcations may be either proposed from personal knowledge or derived by computer modeling. However, it is important to note that computational methods are still not reliable in predicting binding modes and aﬃnities of ligands, mainly because of inaccuracies in force ﬁelds, limitations in dealing with ligand and target ﬂexibility, and the lack of a reliable scoring functions, as well as the diﬃculties in treating solvent molecules. Therefore, even for seemingly minor modiﬁcations of the leads, it is still necessary to conﬁrm the binding mode experimentally; there are countless examples in which the mode of binding signiﬁcantly changes upon introduction of minor modiﬁcations to the original lead.

VII.

EXPERIENCE WITH STRUCTURE-BASED DRUG DESIGN

Any summary of experience gained during the last 15 years in the area of structure-based drug design is in some way a work in progress, and clearly there is much that we still need to learn.

A. Design Should Be Based on Liganded Structures Many proteins undergo considerable conformational change upon binding to their ligands. Initiating ligand design based on an unliganded structure may be misleading if that structure is of a protein that will change its conformation upon ligand binding. To be on the safe side, one should always start ligand design based on a liganded structure of the target protein. An example of a protein that undergoes large conformational change upon ligation is EPSP (5-enol-pyruvyl-3-phosphate) synthase. The

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

unliganded structure [29] shows a large cavity at the active site (Fig. 7), much of which disappears upon ligation. Sometimes, diﬀerent ligands lead to diﬀerent conformational changes of the protein target, making the designing even more challenging.

B. Design of Small Molecules to Interfere with Protein–Protein Interaction Requires the Structure of the Complex Most protein–protein interfaces are large hydrophobic surfaces. For example, the interface area between growth hormone and its receptor [30] is about 2100 A˚2. To rationally design a small molecule to interfere with such large surfaces is a considerable challenge that requires atomic details of the receptor surface, which may diﬀer for unliganded and liganded forms. Generally, success in this arena is rare. Occasionally, protein–protein interactions consists of only a small number of contacts, such as the RGD interaction with its receptors [31]. In such a case, the

Figure 7 Stereoview of the structure of EPSP synthase in its open conformation.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

design task becomes essentially a small-molecule–protein interaction problem and is much more likely to be successful.

C. Allow for Flexibility in the Design of Enzyme Inhibitors to Assure Optimal Fit in an Often Rigid Active Site Cavity It is often very diﬃcult to design a highly constrained ligand that complements and ﬁts snugly in an enzyme active site. Although rigidity of the ligand is important to reduce entropy and to ensure greater aﬃnity, it is often wise to initially introduce some ﬂexibility to ensure proper ﬁt in an often rigid active site. Much of this ﬂexibility could be reduced as much as possible in later iterations of the drug design cycle.

D. Synthetic Accessibility Is Essential It is important to design ligands that can be synthesized in a timely fashion from readily available or easily obtained starting material. Given that many potential drugs fail for reasons that have nothing to do with their binding aﬃnity, it is important that one go through a design cycle as fast as possible to obtain feedback on the suitability of the designed ligands as drugs.

E. Every Water Molecule Is Special Incorporation of the position of water molecules that are ﬁrmly bound to the protein can impart aﬃnity and novelty to the designed ligand. A prime example is the design of a class of HIV protease cyclic urea inhibitors by DuPont scientists that incorporated a water molecule bound to both ﬂaps of the enzyme into their ligand [32]. The crystal structure of the HIV protease–cyclic urea complex [32] shows the urea carbonyl oxygen substituting for the position of the water molecule.

F. Fill Available Space and Maximize Interactions A major goal of ligand design should be to ﬁll as much of the space in the binding site as possible without rendering the designed ligand too large. Ligands greater than 500 Da have lower probability of being orally bioavailable. It is also important to maximize both polar and nonpolar interactions with the protein.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

G. Beware of Crystal Contacts In the crystal, it is possible for a ligand to make important contacts with residues from a neighboring molecule producing an artiﬁcial mode of binding that is not possible in solutions. Thus it is important to analyze all crystal contacts in the vicinity of the ligand prior to proceeding with the design of new ligands.

H. Use of Surrogate Enzymes Can Lead to Important Insights, But Optimization Requires the Target Protein When the target enzyme is diﬃcult to obtain, related enzymes could be used to provide insights in the design of novel ligands. For example, papain was used to design a class of potent cathepsin K inhibitors [33] spanning both sides of the papain active site. However, ﬁne-tuning these inhibitors to produce more potent ones required the use of the crystal structure of cathepsin K itself [34].

I. Iterative Design Is Essential It is a rarity that the ﬁrst ligand to be designed is the ﬁnal one. Thus, it is common to go through the structure-based design cycle (Fig. 2) several times with each class of inhibitors being designed. This iteration should continue until the ideal molecule that will be advanced to development has been identiﬁed.

J. Solubility of Ligands Matters One of the bottlenecks associated with structure-based design is poor aqueous solubility of many ligands. If the ligands are insoluble in water, it is often diﬃcult to form complexes under conditions of crystallization. Unlike the crystallization of small organic molecules, proteins must be crystallized from aqueous solutions or using solvents that are highly miscible with water. Therefore, it is sensible to introduce polar or charged groups to improve inhibitor solubility, making the target molecule more amenable to structural studies.

K. No Substitute for Experience Structure-based drug design is no diﬀerent from most other areas, in that experience counts greatly.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

L. Dedicated Molecular Biology and Protein Purification Groups Are Essential Protein crystallography often requires special constructs or mutants to facilitate crystallization; it also requires large quantities of highly puriﬁed protein. Thus to move forward in a timely fashion, it is important that an industrial structural biology group employ molecular biologists and individuals with expertise in protein puriﬁcation.

VIII. OUTLOOK Structure-based drug design is now an integral part of most if not all drug discovery programs. It is a given that structure-based design is part of each drug discovery eﬀort whenever the target is a soluble protein. However, a large segment of targets—namely, membrane proteins, particularly Gprotein-coupled receptors—are excluded. It is hoped that this situation will be remedied in the near future.

REFERENCES 1. Hol WGJ. Protein crystallography and computer graphics-toward rational drug design. Angew Chem 1986; 25:767–778. 2. Wlodawer A, Erickson JW. Structure-based inhibitors of HIV-1 protease. Annu Rev Biochem 1993; 62:543–585. 3. Kim CU, Chen X, Mendel DB. Neuraminidase inhibitors as anti-inﬂuenza virus agents. Antiviral Chem Chemother 1999; 10:141–154. 4. Kuntz ID. Structure-based strategies for drug design and discovery. Science 1992; 257:1078–1082. 5. Verlinde LMJ, Hol WGJ. Structure-based drug design: progress, results and challenges. Structure 1994; 2:577–587. 6. Blundell TL. Structure-based drug design. Nature 1996; 384(suppl):23–26. 7. Babine RE, Bender SL. Molecular recognition of protein–ligand complexes: application to drug design. Chem Rev 1997; 97:1359–1472. 8. Hol WGJ, Verlinde LMJ. Macromolecular crystallography and medicine. In: Rossmann MG, Arnold E, eds. International Tables of Crystallography. Vol. F. Dordrecht, The Netherlands: Kluwer Academic, 2002. 9. Amzel LM. Structure-based drug design. Curr Opin Biotech 1998; 9:366– 369.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

10. Hogan JC Jr. Directed combinatorial chemistry. Nature 1996; 384 suppl: 17–19. 11. Salemme FR, Spurlino J, Bone R. Serendipity meets precision: the integration of structure-based drug design and combinatorial chemistry for eﬃcient drug discovery. Structure 1997; 5:319–324. 12. Ferre-D’Amare AR, Burley SK. Use of dynamic light scattering to assess crystallizability of macromolecules and macromolecular assemblies. Structure 1994; 2:357–359. 13. Hensley P. Deﬁning the structure and stability of macromolecular assemblies in solution: the re-emergence of analytical ultracentrifugation as a practical tool. Structure 1996; 4:367–373. 14. Carter CW Jr, Carter CW. Protein crystallization using incomplete factorial experiments. J Biol Chem 1979; 254:12219–12223. 15. Rodgers DW. Cryocrystallography. Structure 1994; 2:1135–1140. 16. Drenth J. Principles of Protein X-Ray Crystallography. New York: SpringerVerlag, 1994. 17. Hendrickson WA, Horton JR, LeMaster DM. Selenomethionyl proteins produced for analysis by multiwavelength anomalous diﬀraction (MAD): a vehicle for direct determination of three-dimensional structure. EMBO J 1990; 9:1665–1672. 18. Bohm HJ. Current computational tools for de novo ligand design. Curr Opin Biotechnol 1996; 7:433–436. 19. Allen FH, et al. Watson DG. The development of versions 3 and 4 of the Cambridge Structural Database system. J Chem Inf Comput Sci 1991; 31:187–204. 20. Fisanick W, Cross KP, Forman JC, Rusinko A III. Experimental system for similarity and 3D searching of CAS Registry substances. 1. 3D substructure searching. J Chem Inf Comput Sci 1993; 33:548–559. 21. Pearlman RS. Three-dimensional structures: how do we generate them and what can we do with them? Chem Des Auto News 1993; 8–10:44–47. 22. Kuntz ID, Blaney JM, Oatley SJ, Langridge R, Ferrin TE. A geometric approach to macromolecule–ligand interactions. J Mol Biol 1982; 161:269– 288. 23. Bohm H-J. The computer program LUDI: a new method for the de novo design of enzyme inhibitors. J Comput-Aided Mol Des 1992; 6:61–78. 24. Goodford PJ. A computational procedure for determining energetically favorable binding sites on biologically important macromolecules. J Med Chem 1985; 28:849–857. 25. Goodsell DS, Olsen AJ. Automated docking of substrates to proteins by simulated annealing. Proteins 1990; 8:195–202. 26. Miranker A, Karplus M. Functionality maps of binding sites: a multiple copy simultaneous search method. Proteins 1991; 11:29–34. 27. Bartlett PA, Shea GT, Telfer SJ, Waterman S. CAVEAT: a program to

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

28. 29.

30.

31.

32.

33.

34.

facilitate the structure-derived design of biologically active molecules. In: Roberts SM, ed. Molecular Recognition in Chemical and Biological Problems. Cambridge: Royal Society of Chemistry, 1989:182–196. Rotstein SH, Murcko MA. GroupBuild: a fragment-based method for de novo drug design. J Med Chem 1993; 36:1700–1710. Stallings WC, Abdel-Meguid SS, Lim LW, Shieh H-S, Dayringer HE, Leimgruber NK, Stegeman RA, Anderson KS, Sikorski JA, Padgette SR, Kishore GM. Structure and topological symmetry of the glyphosphate 5-enol-pyruvylshikimate-3-phosphate synthase: a distinctive protein fold. Proc Natl Acad Sci USA 1991; 88:5046–5050. Cunninghum BC, Ultsch M, De Vos AM, Mulkerrin MG, Clauser KR, Wells JA. Dimerization of the extracellular domain of the human growth hormone receptor by a single hormone molecule. Science 1991; 254:821–825. Ku TW, Miller WH, Bondinell WE, Erhard KF, Keenan RM, Nichols AJ, Peishoﬀ CE, Samanen JM, Wong AS, Huﬀman WF. Potent non-peptide ﬁbrinogen receptor antagonists which present an alternative pharmacophore. J Med Chem 1995; 38:9–12. Lam PYS, Jadhav PK, Eyermann CJ, Hodge CN, Ru Y, Bacheler LT, Meek JL, Otto MJ, Rayner MM, Wong YN, Chang C-H, Weber PC, Jackson DA, Sharpe TR, Erickson-Viitanen S. Rational design of potent, bioavailable, nonpeptide cyclic ureas as HIV protease inhibitors. Science 1994; 263:380– 384. LaLonde JM, Zhao B, Smith WW, Janson CA, DesJarlais RL, Tomaszek TA, Carr TJ, Thompson SK, Oh HJ, Yamashita DS, Veber DF, AbdelMeguid SS. Use of papain as a model for the structure-based design of cathepsin K inhibitors: crystal structures of two papain–inhibitor complexes demonstrate binding to SV-subsites. J Med Chem 1998; 41:4567–4576. Thompson SK, Smith WW, Zhao B, Halbert SM, Tomaszek TA, Tew DG, Levy MA, Janson CA, DAlessio KJ, McQueney MS, Kurdyla J, Jones CS, DesJarlais RL, Abdel-Meguid SS, Veber DF. Structure-based design of cathepsin K inhibitors containing a benzyloxy-substituted benzoyl peptidomimetic. J Med Chem 1998; 41:3923–3927.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

2 Src Homology-2 Domains and Structure-Based, Small-Molecule Library Approaches to Drug Discovery Chester A. Metcalf III and Tomi Sawyer ARIAD Pharmaceuticals, Inc., Cambridge, Massachusetts, U.S.A.

The elucidation of cell-receptor-associated signal transduction pathways by means of the tools of biochemistry and molecular genetics has resulted in the identiﬁcation of a multitude of protein targets for therapeutic intervention (Table 1) [1]. The fact that many of these targets have x-ray crystallography and/or NMR spectroscopy to guide the syntheses of structurally biased single analogues and combinatorial libraries has ushered the pharmaceutical industry into a new era of drug discovery. Within cells there exists an enormously diverse data set of functional proteins and signaling pathways, involving both noncatalytic and catalytic processes, which are orchestrated through highly speciﬁc protein–protein interactions. In principle, such interactions can be disrupted or promoted, either directly or indirectly (via enzyme inhibition), through small-molecule intervention driven by structure-based methods. This chapter discusses the role of Src homology-2 (SH2) domains as mediators of protein–protein interactions in signal transduction, with a focus on the therapeutic implications of blocking the SH2 domain of the nonreceptor protein tyrosine kinase Src with

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 1 List of Possible Protein Targets for Therapeutic Intervention G-Protein-Coupled/Integrin Receptors Angiotensin (AT1, AT2) Bradykinin (B1, B2) Cholecystokinin (CCKA) Gastrin (CCKB) Endothelin (ETA, ETB) a–Melanotropin (MCR1) Adrenocorticotropin (MCR2) Substance P (NK1) Neurokinin-A (NK2) Neurokinin-B (NK3) y-opioid (Enkephalin) A-opioid (Endorphin) n-opioid (Dynorphin) Oxytocin Somatostatin (sst1–sst5) Vasopressin (V1A, V1B) Neuropeptide-Y (Y1-Y5) Calcitonin

Receptor/Nonreceptor Kinases/Phosphatases Receptor tyrosine kinases Epidermal growth factor Fibroblast growth factor Insulin Nerve growth factor Platelet-derived growth factor Nonreceptor tyrosine kinases Src and Src family (Lck, Hck) Abl, Syk, Zap-70 Receptor serine/threonine kinases Transforming growth factor Nonreceptor serine/threonine kinases cAMP-Dependent protein kinase Phosphoinositol-3-kinase (P13K) Cyclin-dependent kinases (CDKs)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Transcription Factors/Proteases NF-nB, STAT, NFAT, SMAD, CREB Proteases Aspartic proteases Pepsin Renin Cathepsins (D, E) HIV-1 protease Serinyl proteases Trypsin Thrombin Chymotrypsin-A Kallikrein Elastase Tissue plasminogen activator Factor Xa

Adenosine (A1-A3) Cathecholamine (a1, a2, h1-h3) Histamine (H1, H2) Muscarinic acetylcholine Seratonin (5HT1-5HT7) Melatonin (ML1A, ML1B) Dopamine (D1, D2, D4, D5) g-Amino butyric acid (GABAB) Leukotrienes (LTB4, LTC4, LTD4) Cell adhesion integrin receptors avh3 (Fibrinogen) aIIbh3 or gpIIaIIIb (Fibrinogen) a5h1 (Fibrinectin) a4h1 (VCAM-1)

Mitogen-activated protein kinase Protein kinase C (PKC) Janus family kinases (JAKs) InB family kinases (IKKs) Receptor/nonreceptor phosphatases Receptor tyrosine phosphatases CD45, LAR

Cysteinyl proteases Cathepsins (B, H, K, M, S, T) Proline endopeptidase Interleukin-converting enzyme Apopain (CPP-32) Picornavirus C3 protease Calpains

Metallo proteases Exopeptidase group Peptidyl dipeptidase-A (ACE) Aminopeptidase-M Nonreceptor serine/threonine phosphatases Carboxypeptidase-A PP-1 Calcineurin Endopeptidase group VH1 Endopeptidases (24, 11, 24, 15) Stromelysin Gelatinases (A, B) Collagenase Nonreceptor tyrosine phosphatases PTP1B, Syp

Source: Ref. 1.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

designed, nonpeptide small molecules. We highlight ARIAD’s approach to drug discovery by means of structure-based methods and parallel synthetic libraries to develop cell-active, in vivo eﬀective inhibitors of Src SH2-dependent signal transduction pathways, leading to novel drugs for the treatment of osteoporosis.

I. SIGNAL TRANSDUCTION AND PROTEIN–PROTEIN INTERACTIONS The network of protein–protein interactions that deﬁne signal transduction pathways in most cells originates at a cellular receptor and is triggered by the binding of speciﬁc external stimuli (e.g., growth factors, antigens, hormones). Such signal transduction pathways are then propagated within the cell to the nucleus resulting in speciﬁc gene activation and protein synthesis (Fig. 1) [2]. Listed in Table 2 are the modular domains [3] of various signal transduction proteins and the potential disease areas providing opportunity for therapeutic intervention through small-molecule inhibitory drugs [4]. For Src, these disease areas are osteoporosis and cancer. There are more than 50 known SH2-containing proteins, of which Src was the ﬁrst to be identiﬁed [5]. The SH2 domain of Src consists of approximately 100 amino acids and binds cognate

Figure 1 Representation of signal transduction pathways describing highly specific protein–protein interactions.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 2 Signal Transduction Proteins as Potential Therapeutic Targets Protein (Domains) Src (SH3-SH2-kinase) Hck (SH3-SH2-kinase) Syk (SH2-SH2-kinase) Zap70 (SH2-SH2-kinase) Syp (SH2-SH2-phosphatase) STATs (DNA binding-SH3-SH2) Grb2 (SH3-SH2-SH3) p85/PI3K [SH3-SH2-SH2 (p85 subunit)] Bcr/Abl (SH3-SH2-kinase)

Disease area Osteoporosis, cancer Immune disease, AIDS Allergy, asthma Autoimmune disease Anemia Inﬂammatory disease Cancer, chronic myelogenous leukemia Cancer Chronic myelogenous leukemia

Source: Ref. 4.

phosphotyrosine(pTyr)-containing proteins as well as synthetic peptides in a sequence-dependent manner [6]. In addition to the SH2 domain, Src possesses one SH3 domain (f 60 amino acids), which is characterized by its aﬃnity for proline-rich sequences, and a bilobed tyrosine kinase catalytic domain (f 300 amino acids) containing N-terminal (NT) and C-terminal (CT) domains (Fig. 2) that speciﬁcally phosphorylates tyrosine residues of cognate substrate proteins.

II. Src TYROSINE KINASE AND OSTEOPOROSIS Molecular insight into the protein conformation states of Src kinase has been revealed in a series of x-ray crystal structures of the Src SH3–SH2– kinase domain that depict Src in its inactive conformation [7]. This form maintains a ‘‘closed’’ structure, in which the tyrosine-phosphorylated (Tyr527) C-terminal tail is bound to the SH2 domain (Fig. 2). The x-ray data also reveal binding of the SH3 domain to the SH2–kinase linker [adopts a polyproline type II (PP II) helical conformation], providing additional intramolecular interactions to stabilize the inactive conformation. Collectively, these interactions cause structural changes within the catalytic domain of the protein to compromise access of substrates to the catalytic site and its associated activity. Signiﬁcantly, these x-ray structures provided the ﬁrst direct evidence that the SH2 domain plays a key role in the self-regulation of Src. The bone disease osteoporosis results when an imbalance occurs in the normal course of bone remodeling, a dynamic and highly regulated

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 2 Depiction of the active (‘‘open’’) and inactive (‘‘closed’’) conformations of Src kinase based on the analysis of x-ray structures of c-Src tyrosine kinase crystallized in its inactive state [7]. The stabilization of the inactive conformation is influenced by multiple events including intramolecular binding of the tyrosine-phosphorylated C-terminus tail to the SH2 domain as well as interactions between the SH3 domain and the SH2–kinase linker. CT, C-terminal; NT, N-terminal.

process that involves both bone degradation (resorption) and bone formation. Aberrantly high levels of bone resorption are associated with this disease, which eﬀects a net decrease in bone density and volume, resulting in fragile, brittle bones that are subject to premature breaks and fractures [8]. The most compelling evidence that Src is intimately involved in bone remodeling comes from genetically engineered Src knockout mice. In these Src (–/–) mice, the only major phenotype observed is excessive bone

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

formation; a condition termed osteopetrosis (the opposite of osteoporosis). This suggests that selective inhibition of Src, as a therapeutic treatment for osteoporosis, may shift the bone microenvironment from a state of perpetual bone degradation to one of normal bone turnover without deleteriously aﬀecting other Src-associated cellular processes in the body. The rationale for Src’s involvement in bone processes becomes apparent when the Src knockout eﬀects are examined at the cellular level of an osteoclast. Osteoclasts are multinucleated cells that are responsible for bone resorption. Two diﬀerent osteoclasts, a wild-type (normal) and a Src (–/–) osteoclast, are shown schematically in Figure 3. The wild-type cell shows the characteristics of a bone-resorbing osteoclast, including a ruﬄed border and ‘‘pit’’ formed by the bone-degrading actions of an active osteoclast. These features are absent in the Src (–/–) knockout osteoclast, albeit they are still viable and adhere to bone. In 2000 (Marzia et al. and Amling et al.) it was suggested that Src plays a negative regulatory role in osteoblasts (cells that are responsible for the

Figure 3 The effect of an Src (–/–) knockout in mice as shown by differences in function and appearance of wild-type and Src-minus osteoclasts. The Src-minus osteoclasts lack the ruffled borders of a normal, resorbing osteoclast, but are viable and can adhere to bone. (From Ref. 8.)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

formation of bone) as shown by enhanced bone formation and osteoblast diﬀerentiation rates in Src (–/–) mice [8]. Together, these data provide complementary, mechanistic evidence to validate Src as a viable therapeutic target for the treatment of bone diseases such as osteoporosis.

III. SH2 DOMAINS AND PHOSPHOPEPTIDE BINDING The question of whether ligand binding speciﬁcity exists among SH2 domains was addressed, quite elegantly, by Cantley et al. [6,9], who used synthesized combinatorial libraries of pTyr-containing peptides. A majority of the SH2 binding aﬃnity in pTyr-containing peptides can be attributed to a four-amino acid region represented by the sequence pTyrAaa-Bbb-Ccc. However, binding speciﬁcity exists in the three amino acids directly C-terminal to the pTyr (pY) group, referred to sequentially as pY+1 (Aaa), pY+2 (Bbb), and pY+3 (Ccc). The preferred pY+1, pY+2, and pY+3 amino acids for various SH2-containing proteins are listed in Table 3; the ﬁrst amino acid listed for each position represents the most preferred. For Src SH2 this sequence is pTyr-Glu-Glu-Ile. Such sequence speciﬁcity among SH2-containing proteins provides a rationale for the diﬀerentiation of their associated signal transduction pathways in cells. The successful design of small molecules to interact with a protein binding surface is markedly enhanced by an understanding of the target’s three-dimensional structure, preferably in the context of a bound ligand. In this regard, early x-ray structures of pTyr-containing peptides bound to Src SH2 [10,11] paved the way for the discovery of peptide, peptidomimetic, and nonpeptide ligands and the determination of their complexed structures with Src SH2 [12–14] or the highly homologous Lck SH2 [15,16]. In a landmark paper, Waksman, Kuriyan, and their colleagues reported [11] the ﬁrst x-ray structure of a high aﬃnity phosphopeptide (Glu-Pro-Gln-pTyr-Glu-Glu-Ile-Pro-Ile-Tyr-Leu) bound Src SH2 (KD = 3–6 nM), which uncovered key protein interactions with the pTyr-Glu-Glu-Ile sequence. In particular, this x-ray structure shows the bound phosphopeptide oriented perpendicular to a central h sheet and interacting with two major binding regions of the Src SH2 domain, namely, one for the ligand pTyr (pY pocket) and another for the ligand Ile (pY+3 pocket), to provide what has been described as a ‘‘twopronged’’ binding mode. The pY pocket is characterized mainly by

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 3 SH2 Specificity for Phosphotyrosine-Containing Peptides -Asp-Gly-[pTyr-Aaa-Bbb-Ccc]-Ser-Pro(pY)(pY+1)(pY+2)(pY+3) SH2 Domain

Aaa

Bbb

Ccc

Glu, Asp, Thr Glu, Thr, Gln

Glu, Asn, Tyr Glu, Asp

Ile, Met, Leu Ile, Val, Met

Glu, Thr, Met Gln, Thr, Glu Thr Gln, Tyr, Val

Asn, Glu, Asp Glu, Gln, Thr Thr Asn

Pro, Val, Leu Thr Ile, Leu, Met Tyr, Gln, Phe

Leu, Ile, Val Val, Ile, Leu Met, Ile, Val Ile, Glu, Tyr

Glu, Asp Ile, Leu — —

Leu, Ile, Val Pro, Val, Ile Met Ile, Leu, Met

a

Group 1A Src Lck Group 1Ba Abl Syk (N-SH2) Syk (C-SH2) Grb2 Group 3b PLCg (N-SH2) PLCg (C-SH2) p85 (N-SH2) SHC a

Group 1 contains Tyr or Phe at hD5. Group 3 contains Ile, Cys, or Leu at hD5. Sources: Refs. 6, 9. b

electrostatic interactions, while the pY+3 pocket involves interactions that are mostly hydrophobic. Figure 4 represents the speciﬁc Src SH2 binding interactions with pTyr-Glu-Glu-Ile sequences, as interpreted from x-ray structures [10,11]. The major intermolecular interactions in the pY pocket involve the phosphate oxygens of the ligand pTyr side chain with the conserved basic residues Arg158 and Arg178 of Src SH2. It is noted that Arg178 mutation results in essentially a total loss of binding aﬃnity [17]. Additional intermolecular hydrogen-bonding interactions are also observed with Ser180, Thr182, and the backbone NH of Glu181, whereas a hydrophobic contact occurs between the alkyl side chain of Lys206 and the phenyl ring of the ligand pTyr residue. The two adjacent glutamic acid residues (pY+1 and pY+2) form relatively weak interactions (electrostatic and hydrophobic) with the protein, albeit their extend side chain conformations (oriented away from each other) serve to align and rigidify the peptide backbone. This is an important feature from a drug discovery perspective and can be used in the design of rigid, nonpeptide templates

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 4 Representation of the binding interactions involving the phosphopeptide motif pTyr-Glu-Glu-Ile with Src SH2 as interpreted from complexed x-ray structures [10,11]. The binding regions of the protein, including the major pY and pY+3 pockets, are represented by their key binding residues. Also included are the observed structural waters and their interactions with the pY+1 Glu and pY+3 Ile phosphopeptide residues.

to advance Src SH2 inhibitors (see later: Sec. IV, Lead Discovery and Combinatorial Chemistry). The only direct ligand–protein hydrogen bond contact involves the backbone NH of the pY+1 Glu with the carbonyl oxygen of the His 204 residue. In addition to the hydrophobic interactions involving the Ile phosphopeptide residue and the pY+3 pocket, there exist potential hydrogen-bonding possibilities from Tyr205, Ile217, and a buried Tyr233 residue. Finally, two structural water molecules provide hydrogen-bonding networks between the pY+1 Glu (CO) and pY+3 Ile (NH) phosphopeptide residues, and the Lys206 (NH) and Ile217 (CO) Src SH2 protein residues, respectively. Such structural waters act as drug design elements to increase binding aﬃnity (through favorable entropic contributions) and can be exploited by small molecules that bind to or displace them (see later: Sec. VI, Structure-Based, Small-Molecule Libraries to Explore Src SH2 Binding). The importance of the pTyr group for SH2 binding is counterbalanced by the biological instability of the phosphate group to cellular

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 5 List of pTyr mimics containing nonhydrolyzable and reduced charge functionality, which were explored in the context of a pTyr-Glu-Glu-Ile peptide. (From Ref. 12.)

phosphatases as well as low cellular permeability posed by the highly charged phosphate group [18]. These issues have prompted the pursuit of pTyr mimics to discover cellulary active inhibitors. In a comparative binding study involving pTyr mimics, in the context of a pTyr-Glu-GluIle sequence, researchers explored the ability of a variety of functional groups to act as pTyr replacements (Fig. 5) [12]. The highest aﬃnity, nonhydrolyzable pTyr replacement was found to be the F2Pmp (diﬂuorophosphonomethyl phenylalanine) group [19]. Although some of the aforementioned pTyr replacements represent nonhydrolyzable moieties, the design of a stable pTyr mimic providing both high aﬃnity and adequate cell permeability has remained challenging [20].

IV. LEAD DISCOVERY AND COMBINATORIAL CHEMISTRY The integration of structural biology, drug design (molecular modeling and ‘‘druglike’’ assessment), and synthetic chemistry to discover novel small-molecule leads follows the general iterative process outlined in Figure 6. Available structural knowledge is used to design pharmaceutically driven compounds that will bind a desired protein target; these

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 6 The iterative drug discovery process integrating structural biology, drug design, synthetic chemistry, biological testing, and additional input from other related research areas.

compounds are then synthesized and tested in the appropriate assays. The biological data are analyzed in the context of available (x-ray or NMR) structural information to impact the design of the next series of analogues. This process is repeated until a lead compound or series of compounds possessing the desired biological activities are obtained. The database of available structural information during ARIAD’s initial investigation into compounds targeting Src SH2 was limited; cases involving ligand complexes utilized only peptide molecules [21]. Motivated by an interest to develop orally active Src inhibitors (i.e., nonpeptides) we adopted an exploratory approach to small-molecule lead discovery, using a combinatorial chemistry strategy. Combinatorial libraries were biased with a common phenyl phosphate group and systematically engineered with diversity elements (selection guided by modeling) to probe the protein surface for existing and new binding interactions (Fig. 7). Solid phase array synthesis encompassing a novel phosphate ester linker strategy [22] was

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 7 A novel, phosphate ester linking strategy [22] was used in the synthesis of phenyl phosphate–containing compound libraries, accomplished in 96deepwell reaction blocks [23]. Rigid, nonpeptide templates (A-group) and pY+3 substituents (B-group) satisfied the diversity sites of the molecules.

used to construct the compound libraries in a 96-deepwell plate format [23]. The diversity elements of the molecules included nonpeptide templates (A-group) and pY+3 substituents (B-group). The A-group diversity elements were typically rigid and provided access to both the pY and pY+3 pockets (in a manner similar to the aforementioned Glu-Glu sequence) as well as directionality for each attached substituent. Binding interactions targeting the pY+3 pocket were explored through hydrophobic B-group diversity elements. Finally, diversity building blocks were chosen to target ﬁnal products in the molecular weight range of 500 to 600. Compounds were screened in a high throughput, ﬂuorescent polarization (FP) binding assay [24], using estimated concentrations (relative to a 50 mM DMSO product stock solution assuming 100% synthetic conversion). To verify ﬁnal product formation, we performed qualitative analysis for all compounds using electrospray (+/–) mass spectroscopy. An HPLC peak area purity assessment was also conducted for selected compounds.

V. SOLID-PHASE PARALLEL SYNTHESIS AND NONPEPTIDE PHENYL PHOSPHATE LIBRARIES The combinatorial construction of compound libraries in 96-deepwell plates is eﬃciently accomplished by adhering to the two-dimensional grid

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

pattern of the plate. For example, if diﬀerent diversity elements are added across the rows of the plate (one diversity element, repeated 8 times, per row), with the 12 columns housing the second set of diversity elements, 96 discrete compounds (e.g., 8 A-group 12 B-group = 96 compounds; Fig. 7) will result. This format permits the rapid synthesis of relatively large structurally biased libraries by systematically combining sets of diversity groups. To streamline plate synthesis, we developed with Cyberlab, Inc. [25] a custom high throughput organic synthesizer designed to process the 96deepwell reaction blocks (Fig. 8). This instrument was constructed to tolerate a wide range of chemistries; therefore, all liquid contacts (syringes, needles, tubes, and valves) are made of glass, stainless steel, or Teﬂon. Coaxial tip needles with N2 inlets (connected to a bubbler) allow inert dispensing and withdrawal of liquid reagents from the closed vessels without excessive negative or positive pressure buildup. The instrument head, which can access all positions on the deck, is ﬁtted with single-needle and four-needle probes. The tandem use of both needle probes facilitates the transfer (via 5 mL syringe pumps, not shown) of all reaction intermediates from the reagent vials (left side) to the resin-containing 96deepwell reaction blocks (right side). The reaction block, which provides a fully enclosed reaction environment (Teﬂon, polypropylene, and silicone rubber seals) is a slightly modiﬁed version of a design ﬁrst disclosed by Sphynx Pharmaceuticals [23]. Figure 9 shows the reaction block and the reagent vials (100, 30, and 10 mL sizes) in their fully assembled and disassembled states. Holes at the bottom of the wells of the 96-deepwell polypropylene plate (sealed in fully assembled reaction block) allow the reaction solutions to be removed from the wells (via a separate vacuum plenum) and the functionalized resin (retained by Teﬂon frits) to be washed with solvents. The use of the phenyl phosphate group as both a solid support attachment site and a crucial binding element represents what has been referred to as a ‘‘pharmacophore-linking’’ strategy [26]. We explored a variety of phenyl phosphate tether functionalities to provide resins varying in substitution pattern and in chemical ﬂexibility (Scheme 1 and Table 4) [22]. All phenyl phosphate resins were synthesized in batch quantities of 20 g or more. Resin synthesis began with the addition of either p-methoxybenzyl alcohol or benzyl alcohol to commercially available bis(diisopropylamino)chlorophosphine, followed by addition of the diversity phenol [(R1)-OH, DIAT (diisopropylamino tetrazole)]. Displacement of the

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 8 High throughput organic synthesizer developed in collaboration with Cyberlab, Inc. [25] and designed to process the 96-deepwell reaction blocks. The instrument is capable of tolerating a wide range of chemistry (liquid contacts are glass, stainless steel, or Teflon) and accomplishes the transfer of reagents with coaxial tip (N2 inlets) single-needle and four-needle probes.

remaining diisopropylamino group from 1 with Wang resin and oxidation with 4-methylmorpholine N-oxide (NMO) provided the protected phenyl phosphate resins 2 and 3 in excellent yields, as shown in Table 4. Two types of functionality, namely, protected carboxy and amino groups, diﬀerentiated the starting phenols. Reaction schemes demonstrating compound synthesis using both phenol types are shown in Scheme 2 [22]. Mild Fmoc deprotection (1% DBU/DMA) of resin 2a and amide formation using standard coupling conditions [TBTU, DIEA, p-(CO2H)PhCH2NHFmoc)] resulted in attachment of the ﬁrst diversity element to provide resin product 4. A second deprotection followed by a double, one-pot

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 9 The 96-deepwell reaction block and the reagent vials (100, 30, and 10 mL sizes) used in the organic synthesizer in their fully assembled and disassembled states.

Scheme 1

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 4 Yields and Loading of Phenylphosphate Resins Resin 2a 2b 2c 2d 2e 3f 3g 3h 3i

R1

Loading (mmol/g)

Yield (%)

p-(FmocNHCH2CH2)Ph p-(FmocNHCH2)Ph m-(FmocNHCH2)Ph p-(Allyl-O2CCH2CH2)Ph p-(Allyl-O2CCH2)Ph m-(Allyl-O2CCH2)Ph p-[ p-(FmocNHCH2)PhO]Ph p-(Allyl-O2CCH=CH)Ph m-(Allyl-O2CCH=CH)Ph

0.659 0.637 0.627 0.730 0.672 0.807 0.552 0.535 0.796

93 89 88 92 84 98 81 66 98

Source: Ref. 22.

Scheme 2 Abbreviations: DBU, 1,8-diazabicyclo[5.4.0]undec-7-ene; DMA, N,N-dimethylacetamide; TBTU, O-benzotriazole-1-yl-N,N,NV,NV-tetramethyluronium tetrafluoroborate; DIEA, diisopropylethylamine; RA, reductive amination; TFA, trifluoroacetic acid; DCM, dichloromethane.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

reductive amination [Na(OAc)3BH, 2-ethylbutyraldehyde] provided the fully coupled compound with a branch point at the second diversity site. Mild conditions (30% TFA/CH2Cl2) to cleave the ﬁnal compound from the solid support as well as to remove the p-methoxybenzyl protecting group resulted in the isolation of compound 5 in 86% HPLC purity, following in vacuo concentration. For the synthesis of compound 7, the allyl ester of resin 2d was deprotected under palladium-mediated conditions, followed by amide coupling [TBTU, DIEA, m-(NH2CH2) PhCO2allyl] to generate the functionalized phenyl phosphate resin 6. A second deprotection and coupling [TBTU, DIEA, NH(Me)CH2Ph] provided the bisamide resin-bound compound, which was cleaved and isolated as described earlier to yield compound 7 in 66% HPLC purity. The compound types synthesized by using the foregoing combinatorial approach are represented in Figure 10. Variations in functional group connectivity (e.g., amides, oleﬁns, sulfonamides) reﬂect the wide range of chemistry that was pursued in the generation of these libraries. Some bifunctional A-group and monofunctional B-group diversity elements used in the coupling reactions are shown in Figure 11. Alkyl and aryl phosphate ester groups (R; see Fig. 10) were also explored to

Figure 10 Representations of some of the compound types synthesized in the nonpeptide phenyl phosphate libraries.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

investigate the binding consequences of reduced charge at the phosphate group. Initial Src SH2 screening produced hits, which were resynthesized and then retested in the binding assay. Some of the higher aﬃnity compounds are shown in Figure 12. Although more than 10,000 compounds were produced by this methodology, only marginal binding aﬃnities and no high-resolution x-ray or NMR structures were achieved; the poor aqueous solubility and undesirable physical properties of the molecules are likely to have hampered these eﬀorts. At this point a decision was made to pursue a much more structure-based approach. Compounds

Figure 11 List of some of the molecular diversity building blocks used in the construction of the nonpeptide phenyl phosphate libraries.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 12 Resynthesized library hits identified from the high throughput fluorescence-polarization assay along with their Src SH2 binding data.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

lending themselves to possible x-ray and/or NMR co structure determination were emphasized.

VI. STRUCTURE-BASED, SMALL-MOLECULE LIBRARIES TO EXPLORE Src SH2 BINDING Refocusing our drug-discovery strategy prompted us to revisit the initial lead compound, pTyr-Glu-Glu-Ile. It was clear that to generate high aﬃnity, small-molecule compounds for Src, we would likely need to maintain the key binding interactions of the pTyr-Glu-Glu-Ile motif, as well as to explore molecules capable of mimicking or interacting with the structural waters found in the Src SH2-phosphopeptide complexed x-ray structure. A template would be required that allowed access to both pockets (pY and pY+3), mimicking the ‘‘two-pronged’’ binding mode of pTyr-Glu-Glu-Ile. Noteworthy in this regard, a novel, de novo designed nonpeptide 8 was disclosed [14] with comparable binding to Ac-pTyrGlu-Glu-Ile-NH2 (phosphopeptide 9) (Fig. 13). Signiﬁcant interactions involving the benzamide functionality were revealed in an x-ray structure of 8 bound to Src SH2 [14]. In addition to interacting with several key sites of Src SH2 (e.g., the pY/pY+3 pockets and the CO of His204), this compound displaces both structural water molecules and makes a direct hydrogen bond contact with the backbone NH of Lys206 through its benzamide CO moiety. The eﬀect of this carboxamide group on Src SH2 binding is demonstrated by the related compounds 10 and 11 [14], in which the desamide compound 11 binds with over 15-fold lower aﬃnity than 10 (Fig. 13). ARIAD’s strategy was to utilize this high aﬃnity benzamide template to gain a better understanding of nonpeptide interactions with Src SH2, and then to advance a database of structure–activity relationships (SARs) to ultimately develop novel, proprietary Src SH2 inhibitors. Subsequent to the disclosure of compound 8, a second-generation, higher aﬃnity compound, containing a methylated benzamide template in the context of a pTyr group, was reported [27]. A literature procedure [27] was used to synthesize this compound (12, AP21733) [16], and a 2.5 A˚ x-ray crystal structure of Lck SH2 (S164C), a protein homologue of Src SH2, complexed with AP21733, was obtained (M. H. Hatada, unpublished results). The proposed S-conﬁguration of the benzylic methyl stereocenter of AP21733 was conﬁrmed through independent asymmetric synthesis [28]. The Lck SH2–nonpeptide structure reveals adherence to the historical pTyr-Glu-Glu-Ile interactions in the pY pocket and shows carboxamide

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 13 Series of de novo designed nonpeptides containing a benzamide template (exemplified by compound 12, AP21733) designed to interact favorably with Src SH2 and specifically to displace structural waters found in complexed Src SH2 structures [14,27]. The Src SH2 binding IC50 is shown for each compound, as well as a comparative IC50 for Ac-pTyr-Glu-Glu-Ile-NH2 (compound 9).

contacts with Lys182 (206 in Src) and Ile193 (217 in Src). The phenyl ring of the benzamide template also forms favorable stacking interactions with Tyr181 (205 in Src). Although the cyclohexylmethyl group interacts with the pY+3 pocket, the contacts are primarily surface type and do not extend as deeply into the pocket as the Ile of pTyr-Glu-Glu-Ile. Consequently, SAR exploration of the pY+3 pocket, which had not been rigorously studied with nonpeptide (peptidomimetic) small molecules [13,14], became the ﬁrst objective to be investigated. Parallel synthesis provides the means of rapidly preparing discrete analogues for both lead generation and lead optimization strategies, which makes it an attractive option for developing compound databases for therapeutic targets. Furthermore, the incorporation of structure-based methods into the design and evaluation of parallel synthetic libraries has proven to be a successful strategy for integrating the two drug discovery technologies [29]. For the synthesis of the benzamide-containing compounds, we devised a hitherto unreported solid phase, parallel synthetic

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

route focusing on pY+3 derivatives based on compound 13 and AP21733 [30]. Our synthetic philosophy adopts an integrated solid and solution phase strategy that diﬀers from the traditional unidirectional approach by recognizing the strengths and limitations of each synthetic method and then devising a route accordingly (Fig. 14) [31]. In addition, this strategy provides chemical ﬂexibility to incorporate, within the compound’s molecular design, the necessary functional group complexity dictated by our structure-based methods. The importance of the carboxamide group guided our decision to exploit this functionality both as a solid support attachment site and as a conserved binding element. A Rink amide linkage was chosen to provide facile coupling of the template, via its benzoic acid, and eventual generation of the critical benzamide binding moiety upon cleavage from the solid support. The protected salicylic acid template 14 (synthesized using a modiﬁcation of the solution phase literature procedure) [27] was coupled to Rink amide AM resin by means of standard protocols (EDC/HOBt) to provide the benzamide-linked resin 15

Figure 14 Parallel synthetic approaches demonstrating a traditional (unidirectional) strategy and a multifaceted, integrated strategy; the latter utilizes both solid and solution phase reactions.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

(Scheme 3) [30]. The pY+3 diversity alcohols (R1)-OH (Fig. 15) were attached to the template through a Mitsunobu coupling to provide ether derivatives of 16. Palladium-mediated Alloc deprotection followed by amide formation using the phosphate-ester-containing diversity acids (R2)-CO2H provided the fully coupled resin-bound products of 17. Cleavage from the resin with 95% TFA/H2O, which also aﬀorded benzyl phosphate deprotection, followed by reversed-phase (RP) semipreparative

Scheme 3 Abbreviations: EDC, 1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide)hydrochloride; HOBt, 1-hydroxybenzotriazole; DEAD, diethyl azodicarboxylate.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 15 The diversity alcohols (R1)-OH and carboxylic acids (R2)-CO2H used to synthesize compounds represented by 18 and 19. (From Ref. 30.)

Table 5 Src SH2 Binding (FP) for Analogs of Compound 18

Source: Ref. 30.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

HPLC puriﬁcation generated the ﬁnal compounds represented by 18 (mixture of diastereomers) and 19. The selection of the pY+3 diversity R1 groups was guided by a FLO docking program [32], utilizing 800 commercially available alcohols (preﬁltered by MW, H-bond donors, and reactive groups outside the OH). The R1 groups were computationally incorporated [33] into the benzamide template, docked into our Src SH2 binding site model [34], and then rank-ordered according to favorable ﬁt. The ﬁnal list of alcohols was

Figure 16 The predicted binding mode of compound 23 in the pY+3 pocket of the Src SH2 model. The branch point in the pY+3 bisallyl group allows favorable binding interactions to occur. (From Ref. 30.)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

selected according to predicted binding as well as the ability of the R1 group to impart beneﬁcial properties to the molecule as related to low molecular weight, increase solubility, and other factors. Table 5 contains the Src SH2 binding results for a selected set of pY+3-modiﬁed nonpeptide analog. Relative to compound 20, which was synthesized by means of our solid phase method to act as an internal standard, increases in binding aﬃnity appeared to track the degree of hydrophobicity at the R1 group as demonstrated by compounds 21 (methyl) and 22 (isopropyl). From a drug design perspective, the result of 22 is signiﬁcant because a four-carbon reduction took place, relative to the cyclohexylmethyl group (MW decrease by 54), without greatly compromising the binding aﬃnity (four-fold). An extension of the a-branch point of the isopropyl group to a bisallyl resulted in the highest aﬃnity analog, compound 23. Inspection of the docked structure of 23 in our Src SH2 model reveals how the branch point allows one allyl side chain to hug the surface of the protein, while the other is able to extend deeply into the pY+3 pocket (Fig. 16). A signiﬁcant decrease in binding aﬃnity occurs with the incorporation of a morpholine group, as exempliﬁed by compound 24. Presumably, this result reﬂects an incompatibility of the positively charged morpholine group (at pH 7.2 of the binding assay) in the hydrophobic pY+3 binding pocket of the Src SH2 domain; structurally, the pY+3 pocket according to our Src SH2 model accommodates this compound.

VII.

DISCOVERY OF AN IN VIVO EFFECTIVE Src SH2 INHIBITOR

The next logical step in the progression to a cellularly active Src SH2 inhibitor was to incorporate a high aﬃnity, biologically stable pTyr mimic into the benzamide template. Drug design eﬀorts at ARIAD led to a novel Src SH2 inhibitors containing 4-diphosphonomethylphenylalanine (Dmp), namely, compound 25 (AP21773; Fig. 17) [16]. The design concept for the Dmp group evolved from a 1.5 A˚ x-ray structure of Src SH2, crystallized from citrate buﬀer, that fortuitously contained a citrate molecule bound in the pTyr pocket. The x-ray structure reveals a number of additional hydrogen bonds that citrate makes compared with a pTyr group; this inspired the design of the Dmp moiety as a novel mimic of the citrate interactions. Armed with these designed hydrogen bond contact

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 17 Src SH2 binding IC50 (Fp) for compound 25 (AP21773), which contains a bone-targeted, 4-diphosphonomethylphenylalanine (Dmp) pTyr mimic. (From Ref. 16.)

groups, we expected the Dmp to bind with greater aﬃnity than pTyr, and the Src SH2 binding results for AP21773 (Dmp) and AP21733 (pTyr) conﬁrm this prediction (Figs. 13 and 17). X-ray and NMR structural studies involving AP21773 [16] verify these additional Dmp-related contacts in the pTyr pcket, as well as other key Src SH2 interactions observed earlier with this benzamide class as already discussed. The Dmp moiety not only increases Src SH2 binding aﬃnity, but also provides a mechanism for tissue selectivity by targeting bone [16,35]. This targeting feature provides a higher local concentration of compound on bone than

Figure 18 Solid phase synthetic scheme and molecular diversity groups for compound 27. (From Ref. 36.)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 6 Src SH2 Binding (FP), Rabbit Pit, and Rat TPTX Data for Analogs of Compounds 27 and 35 (AP22209)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

in solution, which in addition increases the amount of Src inhibitor delivered to the resorbing osteoclasts associated with bone. Compounds containing the Dmp group, including AP21773, also bind to hydroxyapatite (data not shown), a major component of bone [16]. Building on the SAR information obtained from the pY+3 study, we focused on improving binding aﬃnity and cellular potency by means of structure-based, parallel synthesis. A resin-bound, enantiomerically enriched benzamide template 26 (Fig. 18) [28] was synthetically elaborated in a manner similar to that described in Scheme 3 to provide the desired Dmp-containing products. A total of 22 structurally biased analogues of 27 were generated (not all combinations synthesized) having speciﬁc R1 and R2 groups as shown in Figure 18 [36]. Table 6 shows the SAR results for a selected series of the benzamide analogues. Similar to the earlier study, increasing hydrophobicity at the pY+3 position leads to increased binding aﬃnity, as demonstrated by compounds 28 to 30. Interestingly, the overall eﬀect on Src SH2 binding of the 3-pentyl group of compound 30 appears to be similar to that of the cyclohexylmethyl group of AP21773, although the latter group contains two more carbon atoms. All the derivatives show an approximately 5- to 10-fold reduction in Src SH2 binding aﬃnity with no substitution (R2 = H, compounds 31–34) at

Figure 19 Solid-phase synthetic scheme and molecular diversity groups for compound 36.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

the R2 position. Finally, in an eﬀort to select compounds for testing in an in vivo thyroparathyroidectomized (TPTX) animal model [37,38], Src SH2 inhibitors were evaluated in a cell-based resorption assay mediated by rabbit osteoclasts. A potent compound, 35 (AP22209; Table 6), was discovered and showed signiﬁcant bone resorption inhibition in test animals (55% inhibition at 25 mg/kg b.i.d.), thus providing in vivo validation for an Src SH2 inhibitor (C. A. Metcalf III, unpublished results). A recent series of proprietary, nonpeptide Src SH2 inhibitors synthesized by our solid phase, parallel synthetic method is outlined in Figure 19. This inhibitor series was based on a set of compounds disclosed earlier [35,39,40] and containing a novel, high-aﬃnity bicyclic benzamide template designed to interact favorably with the hydrophobic Tyr205 Src SH2 protein residue. A bone-targeting, 3,4-diphosphonophenylalanine (Dpp) mimic of pTyr was also incorporated [35,40]. The Dpp moiety can be correlated to both pTyr and citrate (Fig. 20). The biological data for the library analogs of 36 will be described elsewhere.

Figure 20 Representation of the design rationale for two novel, bone-targeting pTyr mimics, Dmp and Dpp, relative to an x-ray structure [16] of citrate complexed with Src SH2.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

VIII. CONCLUSION The increasing number of therapeutic targets available to drug discovery programs has challenged chemists to devise new and eﬃcient strategies for the advancement of lead compounds to clinical candidate status. One evolving approach, as described in this chapter, is the integration of synergistic technologies (e.g., structure-based drug design and combinatorial chemistry) into a focused program that emphasizes the strengths of each individual method. We have used this philosophy to direct our Src SH2 program toward achieving novel proprietary Src SH2 inhibitors such as AP22209, which exhibit promising antiresorptive activity both in an in vivo animal model and in cell-based osteoclast assays. The use of structure-based, small-molecule libraries allowed us to rationally design compounds relative to predicted binding interactions, while taking advantage of parallel synthesis to rapidly advance lead optimization. By adopting a synthetic strategy that utilizes both solid and solution phase chemistries, we were able to achieve the necessary chemical purity and diversity for SAR interpretation at all stages of the drug discovery process. This integrated drug design and combinatorial chemistry strategy is currently being adapted to other drug discovery programs at ARIAD.

ACKNOWLEDGMENTS The authors thank all our colleagues at ARIAD Pharmaceuticals, including Chi Vu, Virginia Jacobsen, Michael Yang, William Shakespeare, Regine Bohacek, Joseph Eyermann, Berkley Lynch, Shelia Violette, and Manfred Weigele, and especially Mayumi Uesugi, Vaibhav Varkhedkar, and Chad Haraldson, whose contributions were signiﬁcant to the success of this work. We also thank Chris Stearns for her help with the ﬁgures, David Dalgarno for his editorial suggestions, and Jay LaMarche for allowing us to buy all our expensive toys.

REFERENCES 1. Sawyer TK. Molecular, structural, and chemical biology in pharmaceutical research. In: Walker JM, Rapley R, eds. Molecular Biology and Biotechnology, 4th ed. London: Humana Press, 2000:237–260. 2. The reader is referred to two landmark papers discussing, respectively, the

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

3.

4. 5.

6.

7.

8.

9.

10.

11.

12.

disruption and promotion of protein–protein interactions in signal transduction and the resulting therapeutic implications: Brugge JS. New intracellular targets for therapeutic drug design. Science 1993; 260:918–919, and Spencer DM, Wandless TJ, Schreiber SL, Crabtree GR. Controlling signal transduction with synthetic ligands. Science 1993; 262: 1019–1024. These are considered to be functionally competent protein domains owing to their ability to maintain their structure and molecular recognition properties independent of the full-length protein (Cohen GB, Ren R, Baltimore D. Modular binding domains in signal transduction proteins. Cell 1995; 80: 237–248). Botﬁeld MC, Green J. SH2 and SH3 domains: choreographers of multiple signaling pathways. Ann Rep Med Chem 1995; 30:227–237. Koch CA, Anderson D, Moran MF, Ellis C, Pawson T. SH2 and SH3 domains: elements that control interactions of cytoplasmic signaling proteins. Science 1991; 252:668–674. Songyang Z, Shoelson SE, Chaudhuri M, Gish G, Pawson T, Haser WG, King F, Roberts T, Ratnofsky S, Lechleider RJ, Neel BG, Birge RB, Fajardo JE, Chou MM, Hanafusa H, Schaﬀhausen B, Cantley LC. SH2 domains recognize speciﬁc phosphopeptide sequences. Cell 1993; 72:767–778. Xu W, Harrison SC, Eck MJ. Three-dimensional structure of the tyrosine kinase c-Src. Nature 1997; 385:595–602. See also: Xu W, Doshi A, Lei M, Eck MJ, Harrison SC. Crystal structures of Src reveal features of its autoinhibitory mechanism. Mol Cell 1999; 3:629–638. For recent reviews describing Src’s role in the biological processes involving bone disease, see: Sawyer T, Boyce B, Dalgarno D, Iuliucci J. Src inhibitors: genomics to therapeutics. Expert Opin. Invest. Drugs 2001; 10:1327–1344, and Schwartzberg PL. Src in the osteoclast — a potential therapeutic target? Jikken Igaku 1998; 16(11):1386–1393. Songyang Z, Shoelson SE, McGlade J, Oliver P, Pawson T, Bustelo XR, Barbacid M, Sabe H, Hanafusa H, Yi T, Ren R, Baltimore D, Ratnofsky S, Feldman RA, Cantley LC. Speciﬁc motifs recognized by the SH2 domains of Csk, 3BP2, fps/fes, GRB-2, HCP, SHC, Syk, and Vav. Mol Cell Biol 1994; 14:2777–2785. Waksman G, Kominos D, Robertson SR, Pant N, Baltimore D, Birge RB, Cowburn D, Hanafusa H, Mayer BJ, Overduin M, Resh MD, Rios CB, Silverman L, Kuriyan J. Crystal structure of the phophotyrosine recognition domain SH2 of v-src complexed with tyrosine-phosphorylated peptides. Nature 1992; 358:646–653. Waksman G, Shoelson SE, Pant N, Cowburn D, Kuriyan J. Binding of a high aﬃnity phosphotyrosyl peptide to the Src SH2 domain: crystal structures of the complexed and peptide-free forms. Cell 1993; 72:779–790. Gilmer TM, Rodriguez M, Jordan S, Crosby R, Alligood KJ, Green M, Kimery M, Wagner C, Kinder D, Charifson P, Hassell AM, Willard D,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

13.

14.

15.

16.

17.

18.

19.

20.

Luther M, Rusnak D, Sternbach DD, Mehrotra M, Peel M, Shampine L, Davis R, Robbins J, Patel IR, Kassel D, Burkhart W, Moyer M, Bradshaw T, Berman J. Peptide inhibitors of Src SH3–SH2–phosphoprotein interactions. J Biol Chem 1994; 269:31711–31719. Also, for pTyr mimics see: Ye B, Akamatsu M, Shoelson SE, Wolf G, Giorgetti-Peraldi S, Yan X, Roller PP, Burke TR Jr. L-O-(2-Malonyl)tyrosine: a new phosphotyrosyl mimetic for the preparation of Src homology 2 domain inhibitory peptides. J Med Chem 1995; 38:4270–4275. Plummer MS, Holland DR, Shahripour A, Lunney EA, Fergus JH, Marks JS, McConnell P, Mueller WT, Sawyer TK. Design, synthesis, and cocrystal structure of a nonpeptide Src SH2 domain ligand. J Med Chem 1997; 40:3719–3725. Lunney EA, Para KS, Rubin JR, Humblet C, Fergus JH, Marks JS, Sawyer TK. Structure-based design of a novel series of nonpeptide ligands that bind to the pp60c-src SH2 domain. J Am Chem Soc 1997; 119:12471–12476. Eck MJ, Shoelson SE, Harrison SC. Recognition of a high-aﬃnity phosphotyrosyl peptide by the Src homology-2 domain of p56lck. Nature 1993; 362:87–91. Bohacek RS, Dalgarmo DC, Hatada M, Jacobsen VA, Lynch BA, Macek KJ, Merry T, Metcalf CA III, Narula SS, Sawyer TK, Shakespeare WC, Violette SM, Weigele M. X-ray structure of citrate bound to Src SH2 leads to a high aﬃnity, bone-targeted Src SH2 inhibitor. J Med Chem 2001; 44:660–663. Bibbins KB, Boeuf H, Varmus HE. Binding of the Src SH2 domain to phosphopeptides is determined by residues in both the SH2 domain and the phosphopeptides. Mol Cell Biol 1993; 13:7278–7287. For a discussion on the use of prodrugs to address bioreversibility and cell permeability for inhibitors of Src SH2, see: Rickles RJ, Henry PA, Guan W, Azimiora M, Shakespeare W, Violette S, Zoller MJ. A novel mechanismbased mammalian cell assay for the identiﬁcation of SH2-domain-speciﬁc protein–protein inhibitors. Chem Biol 1998; 5:529–538, and Stankovic CJ, Surendran N, Lunney EA, Plummer MS, Para KS, Shahripour A, Fergus JH, Marks JS, Herrera R, Hubbell SE, Humblet C, Saltiel AR, Stewart B, Sawyer TK. The role of 4-phosphonodiﬂuoromethyl- and 4-phosphonophenylalanine in the selectivity and cellular uptake of SH2 domain ligands. Bioorg Med Chem Lett 1997; 7:1909–1914. Burke TR Jr, Smyth M, Nomizu M, Otaka A, Roller PP. Preparation of ﬂuoro- and hydroxy-4-(phosphonomethyl)-D,L-phenylalanine suitably protected for a solid-phase synthesis of peptides containing hydrolytically stable analogues of O-phosphotyrosine. J Org Chem 1993; 58:1336–1340. For recent papers describing the use of pTyr mimics in the context of Src and Lck SH2 inhibitors, see: (a) Kawahata N, Yang MG, Luke GP, Shakespeare WC, Sundaramoorthi R, Wang Y, Johnson D, Merry T, Violette S, Guan W, Bartlett C, Smith J, Hatada M, Lu X, Dalgarno DC,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

21. 22.

23.

24.

25. 26. 27.

28. 29.

Eyermann CJ, Bohacek R, Sawyer TK. A novel phosphotyrosine mimetic 4V-carboxymethyloxy-3V-phosphonophenylalanine (CCP): exploitation in the design of nonpeptide inhibitors of pp60src SH2 domain. Bioorg Med Chem Lett 2001; 11:2319–2323. (b) Lesuisse D, Deprez P, Bernard D, Broto P, Delettre G, Jean-Baptiste V, Marquette JP, Sarubti E, Schoot B, Mandine E, Lange G. SAR by crystallography: a new approach combining screening and rational drug design. Application to the discovery of nanomolar Src SH2 binders. Chim Nouv 2001: 19:3240–3241. (c) Beaulieu PL, Cameron DR, Ferland J-M, Gauthier J, Ghiro E, Gillard J, Gorys V, Poirier M, Rancourt J, Wernic D, Llinas-Brunet M, Betageri R, Cardozo M, Hickey ER, Ingraham R, Jakes S, Kabcenell A, Kirrane T, Lukas S, Patel U, Proudfoot J, Sharma R, Tong L, Moss N. Ligands for the tyrosine kinase p56lck SH2 domain: discovery of potent dipeptide derivatives with monocharged, nonhydrolyzable phosphate replacements. J Med Chem 1999; 42:1757–1766. For examples of x-ray structures of nonpeptide compounds complexed with Src SH2, see Refs. 13 and 14, and for Lck SH2, see Ref. 16. Metcalf CA III, Vu CB, Sundaramoorthi R, Jacobsen VA, Laborde EA, Green J, Green Y, Macek KJ, Merry TJ, Pradeepan SG, Uesugi M, Varkhedkar VM, Holt DA. Novel phosphate ester-linked resins: the solidphase generation of phenyl phosphate-containing compounds for SH2 inhibition. Tetrahedron Lett 1998; 39:3435–3438. Meyers HV, Dilley GJ, Durgin TL, Powers TS, Winssinger NA, Zhu H, Pavia MR. Multiple simultaneous synthesis of phenolic libraries. Mol Diversity 1995; 1:13–20. Lynch BA, Loiacono KA, Tiong CL, Adams SE, MacNeil IA. A ﬂuorescence polarization based Src-SH2 binding assay. Anal Biochem 1997; 247:77–82. Cyberlab, Inc., Brookﬁeld, CT. Backes I, Ellman JA. Solid support linker strategies. Curr Opin Chem Biol 1997; 1:86. Lunney EA, Para KS, Plummer MS, Prasad JVNV, Saltiel AR, Sawyer TK, Shahripour A, Singh I, Stankovic CJ. Compounds, composition and methods for inhibiting the binding of proteins containing an SH2 domain to cognate phosphorylated proteins. PCT Int Appl WO97/12903, 1997. Luke GP, Holt DA. Synthesis of (S)-5-(1-aminoethyl)-2-(cyclohexylmethoxy) benzamide. Tetrahedron: Asymmetry 1999; 10:4393–4403. For examples of structure-based methods used in combinatorial chemistry, see: (a) Chang Y-T, Gray NS, Rosania GR, Sutherlin DP, Kwon S, Norman TC, Sarohia R, Leost M, Meijer L, Schultz PG. Synthesis and application of functionally diverse 2,6,9-trisubstituted purine libraries as CDK inhibitors. Chem Biol 1999; 6:361–375. (b) Antel J. Integration of combinatorial chemistry and structure-based drug design. Curr Opin Drug Discovery Dev

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

30.

31.

32. 33.

34.

35.

1999; 2:224–233. (c) Rohrer SP, Birzin ET, Mosley RT, Berk SC, Hutchins SM, Shen D-M, Xiong Y, Hayes EC, Parmar RM, Foor F, Mitra SW, Degrado SJ, Shu M, Klopp JM, Cai S-J, Blake A, Chan WWS, Pasternak A, Yang L, Patchett AA, Smith RG, Chapman KT, Schaeﬀer JM. Rapid identiﬁcation of subtype-selective agonists of the somatostatin receptor through combinatorial chemistry. Science 1998; 282:737–740. (d) Szardenings KA, Harris D, Lam S, Shi L, Tien D, Wang Y, Patel DV, Navre M, Campbell DA. Rational design and combinatorial evaluation of enzyme inhibitor scaﬀolds: identiﬁcation of novel inhibitors of matrix metalloproteinases. J Med Chem 1998; 41:2194–2200. (e) Kick EK, Roe DC, Skillman AG, Liu G, Ewing TJA, Sun Y, Kuntz ID, Ellman JA. Structurebased design and combinatorial chemistry yield low nanomolar inhibitors of cathepsin D. Chem Biol 1997; 4:297–307. (f) Combs AP, Kapoor TM, Feng S, Chen JK, Daude-Snow LF, Schreiber SL. Protein structure-based combinatorial chemistry: discover of non-peptide binding elements to Src SH3 domain. J Am Chem Soc 1996; 118:287–288. Metcal CA III, Eyermann CJ, Bohacek RS, Haraldson C, Varkhedkar VM, Lynch B, Bartlett C, Violette S, Sawyer TK. Structure-based design and solidphase parallel synthesis of phosphorylated nonpeptides to explore hydrophobic binding at the Src SH2 domain. J Comb Chem 2000; 2:305–313. Such a strategy follows the philosophy of ‘‘resin capture,’’ as introduced by Armstrong and Keating: Keating TA, Armstrong RW. Postcondensation modiﬁcations of Ugi four-component condensation products: 1-isocyanocyclohexane as a convertible isocyanide. Mechanism of conversion, synthesis of diverse structures, and demonstration of resin capture. J Am Chem Soc 1996; 118:2574–2583. See also: Brown SD, Armstrong RW. Synthesis of tetrasubstituted ethylenes on solid support via resin capture. J Am Chem Soc 1996; 118:6331–6332. FLO97, Graphics and Molecular Mechanics Software for Drug Design. Available from Colin McMartin: [email protected]. McMartin C, Bohacek RS. QXP: powerful, rapid computer algorithms for structure-based drug design. J Comput-Aided Mol Design 1997; 11: 333–344. The Src SH2 binding site model used in this study was developed based on a high resolution (1.0 A˚) crystal structure of Lck SH2 complexed with Ac-pTyr-Glu-Glu-Ile-NH2, obtained from the Brookhaven Protein Data Bank (reference code 1LKK). For a description of the X-ray structure, see: Tong L, Warren TC, King J, Betageri R, Rose J, Jakes S. Crystal structures of the human p56lck SH2 domain in complex with two short phosphotyrosyl peptides at 1.0 A˚ and 1.8 A˚ resolution. J Mol Biol 1996; 256(3):601–610. Violette SM, Guan W, Bartlett C, Smith JA, Bardelay C, Antoine E, Rickles RJ, Mandine E, van Schravendijk MR, Adams SE, Lynch BA, Shakespeare WC, Yang M, Jacobsen VA, Takeuchi CS, Macek KJ, Bohacek RS,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

36.

37. 38.

39.

40.

Dalgarno DC, Weigele M, Lesuisse D, Sawyer TK, Baron R. Bone-targeted Src SH2 inhibitors block Src cellular activity and osteoclast-mediated resorption. Bone 2001; 28:54–64. Weigele M, Bohacek R, Jacobsen VA, Macek K, Yang MG, Kawahata NH, Sundaramoorthi R, Wang Y, Takeuchi CS, Luke GP, Metcalf CA III, Shakespeare WC, Sawyer T. Synthesis of phosphono-containing amino acid derivatives and peptides as signal transduction inhibitors. PCT Int Appl W099/24442, 1999. Frost HM, Jee WSS. On the rat model of human osteopenias and osteoporosis. Bone Miner 1992; 18:227–236. Green JR, Muller K, Jaeggi KA. Preclinical pharmacology of CGP 42V446, a new, potent, heterocyclic biphosphonate compound. J Bone Miner Res 1994; 9:745–751. Violette SM, Shakespeare WC, Bartlett C, Guan W, Smith JA, Rickles RJ, Bohacek RS, Holt DA, Baron R, Sawyer TK. A Src SH2 selective binding compound inhibits osteoclast-mediated resorption. Chem Biol 2000; 7: 225–235. Shakespeare WC, Yang M, Bohacek R, Cerasoli F, Stebbins K, Sundaramoorthi R, Azimioara M, Vu C, Pradeepan S, Metcalf CA III, Haraldson C, Merry T, Dalgarno D, Narula S, Hatada M, Lu X, van Schravendijk M, Adams S, Violette S, Smith J, Guan W, Bartlett C, Herson J, Iuliucci J, Weigele M, Sawyer T. Structure-based design of an osteoclastselective, nonpeptide Src homology 2 inhibitor with in vivo antiresorptive activity. Proc Natl Acad Sci USA 2000; 97:9373–9378.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

3 Three-Dimensional Structure of the Inhibited Catalytic Domain of Human Stromelysin-1 by Heteronuclear NMR Spectroscopy Paul R. Gooley University of Melbourne, Parkville, Victoria, Australia

I.

INTRODUCTION

With the aid of isotopic enrichment it is now routine to determine the structure of moderate to large proteins (20 to 40 kDa) by multidimensional heteronuclear nuclear magnetic resonance (NMR) spectroscopy [1]. The advantages these heteronuclear experiments oﬀer are spectral simpliﬁcation and a reduced dependence on narrow proton linewidths. By spreading the 1H– 1H correlations of a 2-D NMR spectrum into a third and, perhaps, a fourth dimension, according to the chemical shift of the attached 13C or 15 N nucleus, considerable spectral simpliﬁcation is achieved. As the proton of interest is now correlated with its bound 13C or 15N, the information content for assignment is increased, and as the individual planes of the 3-D or 4-D spectra contain relatively fewer overlapping peaks, problems with assignment ambiguities are reduced. These experiments are more eﬃcient than their homonuclear counterparts because transfer of magnetization relies on the large one-bond heteronuclear couplings (11 to 140 Hz). The ﬁrst stage in solving the structure of a protein requires the acquisition of a large number of three-dimensional experiments for sequence-speciﬁc as-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

signment of backbone and side-chain atoms. Typically, a dozen or more 2D and 3-D experiments will need to be acquired, taking a total of 4 to 6 weeks of spectrometer time, thus requiring long-term sample stability or readily available large milligram quantities. The second stage in solving a structure remains largely dependent on the acquisition of NOE spectra and the assignment of many interresidue NOEs. In the special case of a protein complex, for example, an inhibited enzyme, where one component is isotopically labeled, the spectrum of the complex can be separated into subspectra of the two components to aid assignment of both protein and ligand resonances and most importantly, to determine protein-ligand contacts [2]. Using these methods, the spectrum of the ligand can be solved readily providing important information about the conformation of the ligand in the bound state. The application of these techniques relies on an abundant source of 13 C- and 15N-enriched proteins and, therefore, the application of heteronuclear NMR spectroscopy to solving the solution structure of proteins has relied on advances in molecular biology. Many proteins can be overexpressed and isotopically enriched with 13C and 15N by replacing the carbon source with 13C-glucose and the nitrogen source with a 15NH4+ salt. Eﬃcient enrichment is possible with media supplemented with minerals and vitamins, and using fermentation protocols [3]. Molecular biology has further contributed to the number of proteins that can be studied by NMR spectroscopy by overexpressing the catalytic or functional domains of large proteins, thus truncating the protein to a size (often less than 25 kDa) that is readily amenable to these techniques [4]. This chapter discusses the implementation of these methods to the catalytic domain of human stromelysin-1 (sfSTR), a matrix metalloendoproteinase (MMP), complexed to a N-carboxylalkyl inhibitor [5 –8] (Fig. 1). We will focus on the work where the structure of the protein complex was determined and compare this structure to other inhibited MMP catalytic domains.

II. THE MATRIX METALLOPROTEINASE FAMILY The matrix metalloendoproteinases (MMPs or matrixins) are a family of zinc and calcium dependent extracellular proteases that collectively degrade most of the protein constituents of the extracellular matrix [9]. There are at least 23 members of this family and are divided primarily on the basis of sequence homology and substrate speciﬁcity into the following grouping: collagenases (MMP-1, -8, -13, -18) gelatinases

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 1 N-carboxylalkyl inhibitor of sfSTR, N-[(R)-carboxyl-ethyl]-(S)-(2phenylethyl) glycyl-L-arginine-N-phenylamide [8]. The convention of Schechter and Berger [61] is used to describe the specificity subsites of the enzyme S1V, S2V, S3V which correspond to the side chains P1V, P2V, P3V of the inhibitor. (From Ref. 5.)

(MMP-2, -9), stromelysins (MMP-3, -7, -10, -11), membrane-associated (MMP-14, -15, -16, -17), and a group of ‘‘others’’ including macrophage elastase (MMP-12) and enamelysin (MMP-20). These enzymes participate in many normal biological processes such as embryonic development, bone remodeling, wound healing, angiogenesis, and apoptosis. The pharmaceutical industry has shown considerable attention to the regulation of these enzymes because pathological proteolysis by these enzymes accompanies many degradative diseases including arthritis, ulcerations (corneal, gastric, skin), and periodontal diseases. Degradation of the basement membrane by one or more of the MMPs is clearly essential in tumor progression. While the activity of the MMPs is controlled by endogenous inhibitors such as a-macroglobulin and tissue inhibitor of metalloproteases (TIMP-1 and -2), the disease state may be a consequence of an imbalance in the ratio of protease to protease inhibitor. In the disease state, a potent synthetic inhibitor may have therapeutic eﬀects by restoring the ratio of protease to protease inhibitor to normal physiological levels. The MMPs are synthesized as preproproteins and are secreted as latent proproteins. Most MMPs share a common domain structure of a propeptide (about 80 amino acids) that has a conserved cysteine ligated to the catalytic zinc thus maintaining latency [10], a catalytic domain (about 180 amino acids), and a C-terminal domain (about 210 amino acids)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

(Fig. 2). Not all MMPs have these domains, for example, matrilysin lacks the C-terminal domain and the MT-MMPs do not have the propeptide. Some MMPs have additional domains, for example, the gelatinases have three repeats of a ﬁbronectin-type II domain. The C-terminal hemopexin-like domain has been structurally characterized as a four-bladed h-propellor [11 – 13]. While the structural homology of this domain is clear, it has varied functions, for example, it is essential for cleavage of triple helical collagens by the collagenases [14]; however, it is required for activation of pro-MMP-2 by MT1-MMP [15]. The catalytic domain and the C-terminal domain are connected by a proline-rich linker that modeling experiments suggest may play a role in recognition and destabilization of collagen [16].

Figure 2 Domain structure of the MMPs: 92 kDa gelatinase-A (MMP-2), 72 kDa gelatinase-B (MMP-9), the collagenases (MMP-1, -8, and – 13), stromelysin-1 (MMP-3) and matrilysin (MMP-7). Matrilysin is the only known MMP that does not have a C-terminal hemopexin-like domain.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

The catalytic domain has been the focal point for drug discovery. This domain contains the motif HEXXHXXGXXH, which ligates a catalytic zinc, and a characteristic h-turn that contains a conserved methionine that is structurally located near the catalytic zinc. These structural properties have led to the classiﬁcation of the MMPs with several other families of metalloproteases (astacins, serralysins, and reprolysins) as ‘‘metzincins’’ [17]. The catalytic domain further contains a structural zinc and several calcium ions necessary for stability. The catalytic domain of several MMPs has been expressed in bacterial systems either as a soluble protein or in inclusion bodies [18, 19]. The proteins have been puriﬁed and/or refolded and shown to be fully active against small peptide substrates. Structures of the catalytic domains have been solved by both x-ray crystallography and NMR spectroscopy for collagenase-1, -2 and -3 [20 –25], stromelysin-1 [6,10,26], and matrilysin [27]. Furthermore, the pro-form of stromelsyin-1 [10], and the full-length proteins collagenase-1 [11] and progelatinase A [28] have been solved by x-ray crystallography. The structure of the catalytic domains appears identical in the truncated forms to that in the full-length protein, and therefore the smaller truncated form has been ideal for the structural analysis of the inhibited forms to aid inhibitor and drug design. In the following discussion we outline the strategy for determining the structure of the inhibited catalytic domain of stromelysin-1 by NMR methods.

III. ASSIGNMENT OF THE RESONANCES OF THE INHIBITED CATALYTIC DOMAIN OF STROMELYSIN-1 The assignment of the 1H, 13C, and 15N resonances depends on acquiring a large number of separate 3-D or 4-D triple resonance experiments. The experiments can be divided into intraresidue and interresidue and, when combined, lead to sequence-speciﬁc assignment through bonds (Fig. 3) [1]. Improvements and new pulse sequences continue; however, a common set of experiments to assign the backbone (frequently deﬁned as Ha, Ca, N, HN, C’, Ch) resonances of a protein are: 3-D HNCACB, CBCA(CO)NH, HCACO, HNCO, (HCA)CO(CA)NH, 4-D HCANNH and HCA(CO) NNH [29 –32]. Side-chain resonances are assigned using HCCH-TOCSY and HCCH-COSY [33]. Unambiguous stereospeciﬁc assignment of the methyl groups of Leu and Val are possible by preparing a 10% 13C-labeled sample and acquiring a 1H,13C-HSQC spectrum [34]. The incorporation

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 3 3-D and 4-D triple resonance experiments correlate interresidue or intraresidue nuclei. (A) The efficiency of the experiments depend on the large one-bond couplings; (B) atoms correlated in the 4-D HCANNH, an example of an intraresidue experiment; (C) the 4-D HCA(CO)NNH, an interresidue experiment; (D) the 3-D HNCO, an interresidue experiment; and (E) the 3-D HCACO, an intraresidue experiment.

of label is nonrandom such that Leu and Val residues are labeled as 13 Cy2H3-12CgH, 13Cy1H3-13CgH, and 13Cg2H3-12ChH, 13Cg1H3-13ChH, respectively. Consequently, the 13Cy2H3 of Leu and 13Cg2H3 of Val groups appear as singlets in the 1H,13C HSQC spectra and are thus readily stereoassigned. Measurement of 3JHNa in 3-D HNHA spectra [35] aids determination of f torsion angles and stereoassignment of h-methylene groups requires 3-D HNHB [36] and HACAHB [37] experiments. To determine the fold of the protein, a large number of interresidue NOEs must be assigned in 3-D 15N-NOESY [38], 3-D and 4-D 13C-NOESY experiments [39,40]. The assignment of the backbone resonances of the 13 15 C, N-enriched catalytic domain of stromelysin-1 were mostly accomplished with 4-D HCANNH and HCA(CO)NNH experiments (Fig.4) [5]. Side-chain atoms were assigned with 3-D HCCH-COSY and HCCHTOCSY experiments with the carrier located near 35 ppm for aliphatic side chains and at 124 ppm for aromatic side chains. Stereospeciﬁc assignment of the methyl groups were obtained with a 10% 13C-labeled

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 4 Sequential assignment of the backbone atoms for the segment Pro-109 to Val-113 of inhibited sfSTR by 4-D HCANNH and 4-D HCA(CO)NNH. Four planes are shown from each spectrum. The assigned backbone atoms are indicated in (A). In (B) the upper four planes in solid lines are from the 4-D HCANNH and the lower four planes in dashed lines are from the 4-D HCA(CO)NNH. The chemical shifts for the four correlated nuclei in each case are shown. The correlations continue for the segment Pro-109 to Pro-129. As Pro lacks a protonated N, this residue serves as a ‘‘stop’’ signal. The correlation of 19 residues with Pro at the N- and C-terminal ends is unique for this segment in the sequence of sfSTR, therefore these backbone atoms are specifically assigned without having to further assign side chain atoms. (From Ref. 5.)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

sample and a 2-D 1H, 13C HSQC spectrum, and couplings from 3-D HNHA and HNHB experiments aided deﬁning f and m1 angles. To determine which speciﬁc ring nitrogens of the six histidine residues ligate the catalytic and structural zincs a 2-D 1H, 15N HMQC spectrum was acquired, where the delay for generating antiphase 1H, 15N magnetization is set to 22 ms [41]. This experiment will favor the weak two- and threebond couplings between the histidine ring nitrogens and the Cy2H and Cq1H protons (Fig. 5) and clearly showed the Nq2 of His-151, -166, 201, -205 and -211, and the Ny1 of His-96 are the zinc ligands. Finally, the critical NOEs that describe the tertiary structure of the protein were assigned in 3-D 15N-NOESY, 3D and 4-D 13C-NOESY experiments.

IV.

ASSIGNMENT OF THE RESONANCES OF THE INHIBITOR AND NOEs BETWEEN THE PROTEIN AND THE INHIBITOR

To understand the interactions between a protein and a small ligand, we take advantage of the fact that the protein is enriched with 13C and 15N and the ligand is not. Pulse sequences can be designed to edit the spectrum of the protein-ligand complex into spectra (2-D 1 H, 1 H-COSY, TOCSY, NOESY) of the ligand or intermolecular NOEs between the labeled protein and unlabeled ligand in either 2-D or preferably 3-D NOESY spectra. These experiments are composed of X-half-ﬁlters [2,42,43] and either select or ﬁlter the 13C,15N-attached protons (Fig. 6). Consequently, a 2-D 13C doubly ﬁltered NOESY spectrum will show intramolecular NOEs for the ligand, whereas a 3-D 13C-ﬁltered, 13C-selected NOESY will show intermolecular NOEs between the protein and the ligand. The latter experiment is preferably acquired as a 3-D to minimize the ambiguities in assigning the protons of the protein that are involved in the interaction with the ligand. For the complex of the inhibited catalytic domain of stromelysin-1, 2-D doubly ﬁltered 1H,1H COSY and TOCSY experiments performed poorly. As these experiments depend on 1H, 1H couplings, the linewidths of the stromelysin-inhibitor complex must be too large for eﬃcient magnetization transfer. On the other hand, 2-D doubly ﬁltered NOESY experiments acquired in 2H2O and H2O showed correlations for all protons of the inhibitor (Fig. 7), and, as the inhibitor (Fig. 1) was quite simple, the resonances were readily assigned. 3-D 13C-ﬁltered, 13C-separated NOESY experiments were also successfully acquired and assignment of these NOEs were unambiguously obtained (Fig. 8).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 5 Part of a 2-D 1H, 15N HMQC spectrum of inhibited stromelysin-1 where the ring nitrogens of His resonate. The delay where antiphase magnetization evolves is set to 22 ms thus favoring the weak two- and three-bond couplings of Ny1 and Nq2 to the protons of Cy2H and Cq1H [41]. The deprotonated (h-type) nitrogen typically resonates near 249 ppm. These resonances for inhibited sfSTR are near 200 to 210 ppm, shifted upfield by ligation to the zinc ions. For the stable Ny1-H tautomer two strong couplings are observed from the deprotonated Nq2 nucleus to the Cy2H and Cq1H protons. For the stable Nq2-H tautomer only one strong coupling is observed from the deprotonated Ny1 nucleus to the Cq1H proton. For the imidazolium tautomer the resonances of the ring nitrogens are both near 176 ppm and equivalent couplings from these nitrogens to both Cy2H and Cq1H protons are observed. For inhibited sfSTR, His-151, -166, -201, -205 and -211 are in the Ny1-H tautomer, His-179 is in the Nq2-H tautomer and His-96 and -224 are in the imidazolium tautomer. Specific labeling of the Ny1 nucleus supports these assignments [5]. (From Ref. 5.)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 6 X-half filters used for filtering or selecting 13C and 15N-attached protons. Thick and thin closed rectangles are 180j and 90j pulses, respectively, open rectangles are spin lock pulses. (A) A simple X-half filter (2). The delay H is equal to (1/(2[1JXH]) where 1JXH is the one-bond coupling between proton and either 13C (120 to 140 Hz) or 15N (95 Hz). The second 90j pulse is the editing

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

V. STRUCTURE CALCULATIONS Peak intensity data from NOE experiments were accumulated and converted to interproton distances by calibrating against the expected short distances in secondary structure elements. These data were complimented with coupling constants determined from the 3-D HNHA and HNHB experiments. A total of 2589 peaks were assigned in all NOE experiments. After removal of nonconstraining and ambiguous NOEs, typically found in mobile regions, 1814 meaningful restraints remained: 325 intraresidue, 429 sequential, 324 short-range (i+2 to i+5), 665 long range ( > i+5), and 71 intermolecular. Using a gridsearch program [44] 379 dihedrals (140 f, 140 c, 99 m1) were generated from sequential and intraresidue NOEs and coupling constant data from HNHA and HNHB experiments. Structures were calculated using the variable target function algorithm DIANA [45], but it should be noted that in recent years this method has been replaced by torsion angle dynamics methods that are far more eﬃcient [46,47]. To determine the structure of the complex, a residue template of the inhibitor was built as a single residue covalently linked through an oxygen of the carboxylate moeity of the inhibitor (Fig. 1) to the zinc which was covalently bonded to the Nq2 of His-201. The residue template of His151 was created with the structural zinc covalently attached to its Nq2 atom. The structure calculation process is largely iterative with trial structures calculated and incompatible NOEs reassigned or removed and new NOEs assigned on the basis of agreement with the trial structure. In the ﬁnal calculations, and to reduce bias in structure selection, plots of rmsd and number of structures versus target function [48] of the ﬁnal 80

pulse. The phase cycling of this pulse with respect to the receiver determines whether X-nucleus attached protons are selected or filtered. If both signals are added to the receiver (x,x) X-nucleus attached protons are filtered; and if the receiver phase is alternated (x,-x) the X-nucleus attached protons are selected. (B) A doubly tuned half filter for filtering 13C attached protons [42]. In this experiment the filter consists of two delays (H 1,H 2) tuned to different 1JCH values resulting in superior suppression of artifacts. (C) A doubly tuned time-shared half filter for 13C/15N (43). In this experiment D = 1/(41JNH), D1 = 1/(41JCH) and D2 = [1/(41JNH 1/(41JCH)]. Phase cycling the receiver selects or filters both 13C and 15 N attached protons.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 7 2-D 13C doubly filtered NOESY of inhibited sfSTR using the X-half filters of Fig. 4A and B. The NOE correlations of the rings of the P1V and the P3V group are shown. The Hh and Hg protons of the P1V group were distinguished in a similar 2D 13C doubly filtered TOCSY. The specific assignment of the protons of the P3V group were determined by NOEs between the H2,6 and the NH of the P3V in 2D 13C,15N-filtered experiments using the time-shared doubly tuned half filter of Fig. 4C.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 8 Sections of a 3-D 13C-separated, 13C-filtered NOESY of inhibited sfSTR. Only NOEs between the 12C-attached protons of the inhibitor and the 13 C-attached protons of the protein are observed in this spectrum. These NOEs describe the S1V, S2V and S3V subsites of sfSTR. Not shown are several NOEs from Val-197 and His-201 to the ethylene group of P1V. (From Ref. 6.)

structures were used to select structures for energy minimization using the program FANTOM [49]. In the ﬁnal calculations, 30 structures were selected. Table 1 summarizes the DIANA and FANTOM statistics for these structures.

VI. STRUCTURE OF INHIBITED STROMELYSIN-1 A. The Protein Fold Superposition of residues 83 to 248 of the family of structures is shown in Figure 9 viewed along the long axis of the catalytic helix. Residues 249 to 255 are disordered and therefore are not shown. In Figure 10 ribbon drawings of two views of the molecule are shown, one from above the h-sheet and the other from below the S1’ subsite. The secondary structure of sfSTR consists of a ﬁve stranded h-sheet with four parallel strands and

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 1 Structural Statistics and Residual Violations of the 30 Conformers Used to Represent the Solution Structure of the Inhibited Catalytic Domain of Stromelysin-1 Parameter

DIANA

DIANA target function (A˚ ) FANTOM energy (kcal/mol) Lennard-Jones energy (kcal/mol) Distance constraint violations (A˚) sum maximum rmsd Exp. angle constraint violations (j) sum Maximum Rmsd Rmsd residues 83 – 250 (A˚) backbone (Ca,N,C’,O) all heavy atoms 2

10.01 F 0.76

FANTOM 191.0 F 52.8 605.4 F 48.4

35.2 F 1.2 0.48 F 0.06 0.06 F 0.01

61.6 F 1.4 0.39 F 0.03 0.08 F 0.01

93.4 F 11.1 7.5 F 0.9 0.90 F 0.08

112.1 F 19.9 12.6 F 5.0 1.2 F 0.3

0.48 F 0.06 0.94 F 0.06

0.55 F 0.06 0.97 F 0.05

Source: Ref. 7.

one antiparallel strand and the topology 1x, +2x, +2, 1, using the Richardson nomenclature [50]. The h-sheet lies on two helices (helix A and B); a third helix (helix C) is near the C-terminus. The molecule has two zincs: a catalytic zinc is located at the bottom of a cleft, and a structural zinc above the h-sheet. The overall fold of sfSTR may be described as follows. The N-terminus is located near the N-terminal end of helix C. The protein backbone forms a poorly deﬁned irregular strand for the ﬁrst 13 residues before entering strand I of the h-sheet, then descending through helix A. Helix A acts as a backbone to the protein, spanning its full length. The pronounced amphipaticity of this helix provides hydrophobic residues for internal packing to helix B and to the h-sheet, and the hydrophilic residues are exposed to the solvent. After helix A the protein backbone turns to form strand II of the h-sheet, which lies parallel to and outside strand I. This strand rises steeply, giving the h-sheet a distinctly twisted appearance. It is connected by a short loop to strand III, which is parallel to and inside of strand I. A long loop connects strands III and IV, crossing over strand V and placing strand IV along the ligand-binding cleft and antiparallel to strand V. Another small

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 9 Backbone (Ca, C’, N) trace from residues 83 to 250 of 30 conformers of inhibited sfSTR. Residues 251 to 255 are disordered and are not included. All the heavy atoms of the inhibitor are shown. The family of structures are viewed along the long axis of the catalytic helix B. The inhibitor (I) binds to the protein in a well-defined cleft and runs antiparallel to the outer strand of the h-sheet with the ring of P1V homophenylalanine (hP) buried in a bottomless S1V subsite and the P2V arginine (R) is exposed to the solvent.

loop connects strand IV to V, which runs parallel to strand III. The structural zinc is ligated by three His, one each from strands IV and V (His-166 and -179, respectively), and the third (His-151) from the long loop connecting strands III and IV. The fourth ligand of this zinc appears to be Asp-153. After strand V the backbone loops to form helix B. The two His residues of helix B, His-201 and -205, ligate the catalytic zinc. A short turn then enters an extended strand containing His-211, a third ligand of the catalytic zinc. From His-211 to Leu-218 several short range NOEs, in particular between the side chains of Ser-212 and Ala-217, ChH3 of Ala-217 to the NH protons of Leu-218 and Met-219, and the backbone atoms of Thr-215 to Ala-217, describes the presence of two tight turns. An invariant residue, Met-219, which is residue three in one of these turns, is positioned below the three His residues that ligate the catalytic zinc and shows NOEs to all three. Except for helix C, the remainder of the protein is irregular, but well-deﬁned. Helix C runs perpendicular to helix A; the segments C-terminal to these helices are near each other.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 10 Ribbon diagrams of a single conformer of inhibited sfSTR from residues 83 to 250. (A) The complex is viewed from above the h-sheet. The positions of the two zincs are indicated as large balls. The strands of the hsheet (I – V) and helices (A – C) are indicated. The heavy atoms of the inhibitor and residues of the protein that ligate zinc are shown. The inhibitor runs antiparallel to strand IV. The structural zinc lies above the h-sheet and is

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

B. Conformation of the Inhibitor The inhibitor binds to the protein in a well-deﬁned cleft (Figs. 9 and 10) and in an extended fashion, running antiparallel to strand IV of the sheet, as indicated by the strong CaH-CaH NOE between the P2V residue and Val163 (Fig. 8) and parallel to the nonregular loop region encompassing Pro221 to Tyr-223. One of the most striking features of the structure is that the S1V subsite appears to pass through the entire structure. Indeed the aromatic ring of the homophenylalanine group is clearly observed from below the S1V pocket [7]. The S1V subsite is lined with hydrophobic residues including Leu-164, Leu-197, His-201, Val-198, Leu-218, Tyr-220, Leu-222 and Tyr-223. The residues Leu-197, Val-198 and His-201 are from the catalytic helix, whereas Tyr-220 and -223 and Leu-218 and -222 are from the loop following this helix. Contacts between the protein and inhibitor are summarized in Figure 11. Despite the P1V group appearing in contact with a number of residues, the ring of this residue can clearly undergo ring ﬂips, as indicated by the degeneracy of the H3,5 and H2,6 resonances (Fig. 7) thus indicating that this ring is not especially restricted. Similarly not all residues of the S1V are restricted in motion. For example, both methyls of Leu-197 show intraresidue NOEs to the CaH proton of Leu-197 suggesting that motion around the torsion angles m1, m2 is present. We note that this residue shows strong NOEs to the homophenylalanine ring of the inhibitor (Fig. 8) indicating that it is in contact with the inhibitor. Analysis of spectra with other inhibitors with extensions to the homophenylalanine showed this residue became restricted in motion, and thus subtle changes to residue mobility is inhibitor dependent. The family of conformers were analyzed for hydrogen bonds, where acceptor-donor (N-H. . .O) distance was set to an upper limit of 2.4 A˚ and

ligated by His-166 from strand IV, His-179 from strand V, and His-151 and Asp-153 both from a 14 residue loop. The catalytic zinc is ligated by His-201 and – 205 from helix B and His-211. (B) The complex is viewed from below S1V subsite. The heavy atoms of the inhibitor and the residues that are in intermolecular contact (Leu-164, Leu-197, Val-198, His-201, Leu-218, Tyr-220, Leu-222, Tyr-223) with the P1V homophenylalanine are shown. To reduce crowding in the figure not all these residues are labelled. (*) marks Leu-218 and His-201. Val-198 is below Leu-197. Leu-164 is at the N-terminal end of the h-strand that appears above Leu-197 in this figure. The ribbon diagrams were produced by MOLSCRIPT [62].

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 11 Potential hydrogen bond partners to the backbone atoms of the inhibitor and the residues of the S1V subsite that are in intermolecular contact with the P1V homophenylalanine.

the angle to 35j. The analysis suggests that the NH of the P3V hydrogen bonds to the carbonyl of Asn-162; the carbonyl of P1V hydrogen bonds to the NH of the Leu-164 (which is slowly exchanging with deuterium); and the amine of P1V hydrogen bonds with the carbonyl of Ala-165. The structures described here do not show hydrogen bonds between the NH and the carbonyl of the P2V arginine to the protein, which is in contrast to reported crystal structures which show a hydrogen bond to the NH of Tyr223 [10]. Although Pro-221 and Tyr-223 are near atoms of the inhibitor, for example, the NH of Tyr-223 shows weak NOEs to the ring of P3V, their distances in the structure models are not in agreement with these residues participating in hydrogen bonds. The NH of Tyr-223 does not show slow exchange with 2H2O and analysis of 2-D saturation transfer diﬀerence 1 H,15N HSQC spectra suggested that the exchange rate of the NH of Tyr223 was one to two orders slower than a free amide proton further

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

supporting the lack of a strong hydrogen bond [7]. The modest protection of the amide from solvent exchange may simply be due to solvent accessibility. We also note that the conditions of data collection for the NMR solution structure were 40jC, which may destablize any weak hydrogen bonds observed in the crystal structure.

VII.

COMPARISON OF INHIBITED STROMELYSIN TO OTHER MMPS

The structures of the catalytic domain of a large number of MMPs have been solved by x-ray crystallography and NMR spectroscopy [7,10,20 – 27]. All cases show that this catalytic domain of the MMPs has a common fold to that described above, suggesting that the design of speciﬁc inhibitors will require detailed structural investigations that take advantage of diﬀerences of the speciﬁcity pockets. The most signiﬁcant diﬀerence to date has been the nature of the S1V subsite, which is clearly very deep in stromelysin and collagenase-3 [7,10,25], deep in collagenase-2 [21], to quite shallow for collagenase-1 [20,23,25] and matrilysin [27]. In collagenase-1 an arginine residue (equivalent to Leu-197 in stromelysin-1) hydrogen bonds to a structural water and delimits the S1V subsite. Consequently, many inhibitors with large bulky P1V groups show poor aﬃnity for collagenase-1. However, it has been observed that this protein can undergo a conformational change to accommodate such groups [25]. The catalytic domain of stromelysin-1 has been studied by NMR spectroscopy as a complex with a variety of inhibitors [7,10,51], with most binding with groups in the S1V to S3V subsites. Those with a thiadizole group ligating the catalytic zinc bind with groups in the nonprime (S1 to S3) subsites. These inhibitors show NOEs and thus contacts to residues in the S3 subsite including His-166, Tyr-155, and Tyr-168, which are located near the structural zinc. An advantage of NMR spectroscopy is the analysis of protein dynamics. Measurement and analysis of the relaxation parameters R1, R2, and the 15N NOE of 15N-labeled proteins leads to an order parameter (S2) that can describe the relative mobility of the backbone of the protein. Both collagenase-1 and stromelysin-1 have been studied either as inhibited complexes or the free protein [19, 52]. Stromleysin-1 was studied with inhibitors binding to prime or nonprime subsites. Presence or absence of inhibitors in the nonprime sites had minor eﬀects on the highly ordered structure of residues in these subsites, which are in contact with the

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

inhibitor. Inhibitors binding to the primed subsites induced considerable order in the regions 191 to 192 and 223 to 224. Most importantly, the amide proton of His-223 formed a hydrogen bond to a carbonyl group of the inhibitor. In addition to these changes residues remote from the inhibitor, but near the binding sites showed increased mobility. These results suggest that the rigidity of the S1 to S3 subsites are important for distinguishing between ligands, while the ﬂexible S1V to S3V subsites are more accommodating to a broad range of residues. The ﬂexibility of the S1V subsite is in agreement with our observations. Interestingly, similar studies on collagenase-1 with a hydroxymate inhibitor [19] bound to the S1V to S3V subsites showed the analogous region to 220-226 of stromelysin1 was disordered in both the presence or absence of inhibitor. All these results indicate that changes to mobility are complex and mostly unpre-

Figure 12 Catalytic mechanism of thermolysin and stromelysin-1. (A) The mechanism of thermolysin [54]. (B) The mechanism of stromleysin-1 [10]. Equivalent residues to Tyr-157 and His-231 are not observed for stromelysin-1. The proposed mechanism for collagenase-1 [53] is similar to stromelysin-1, but also involves Asn-180 (equivalent to Asn-162 in stromelysin-1). This residue cannot participate in stromelysin-1 due to an additional residue between Ala-165 and Asn-162. (Adapted from Ref. 10.)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

dictable: however, such analyses may prove useful in supporting and monitoring the presence of stabilizing interactions. The mechanism proposed for ﬁbroblast collagenase [53] and stromelysin-1 (10) is similar to that suggested for thermolysin [54] (Fig. 12). In thermolysin the zinc ion ligates the carbonyl of the substrate and with Tyr157 and His-231 stabilizes the tetrahedral intermediate. In collagenase-1 and stromelysin-1, however, the stabilization of the carbonyl of the substrate and the tetrahedal intermediate is by zinc alone. In thermolysin the NH of the scissile bond is stabilized by a peptide carbonyl of Ala-182 and the side chain carbonyl of Asn-112. For collagenase-1 similar interactions by the peptide carbonyl of Ala-182 and the carbonyl of the side chain of Asn-180 are suggested. For stromelysin-1, however, only the carbonyl of Ala-165 would be involved in the stabilization of the NH of the substrate; the equivalent Asn (Asn-162) is not involved as there is a residue insertion in the stromelysin-1 sequence compared with the collagenase-1 sequence. The proposed mechanisms of thermolysin, collagenase-1 and stromelysin-1 suggest that the Glu in the consensus sequence HEXXH would promote the nucleophilic attack of water on the scissile bond of the peptide substrate. The solution structure of stromelysin-1 described here lacks the rigidity expected for the side chain of this residue, Glu-202. In several members of the family of structures, however, this side chain does approach a position that is consistent with the mechanistic role of this residue.

VIII. CONCLUSION This chapter has discussed the use of heteronuclear NMR and isotope editing methods to determine the structure of protein complexes of therapeutically important drug targets. NMR methodology continues to develop with larger protein complexes being studied, and more accurate structures being determined. Developments include deuteration of proteins [1] to enhance relaxation properties, and experiment design, for example, Transverse Relaxation Optimized Spectroscopy (TROSY) [55], which takes advantage of favorable relaxation pathways thus allowing proteins of at least 60 kDa to be studied; inclusion of residual dipolar couplings as an orientation constraint in structure calculations [56,57] are increasing the accuracy of solution structures; and combining deuteration and TROSY experiments has allowed hydrogen bonds to be directly observed and also included in structure calculations [58]. An

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

additional and powerful application of NMR spectroscopy is the method of ‘‘NMR by SAR’’ developed by Fesik et al. [59], which has been applied to ﬁnding new drug leads for stromelysin-1 [60]. NMR spectroscopy has clearly become a powerful and essential tool in the design and development of novel drug leads.

REFERENCES 1. Clore GM, Gronenborn AM. Determining the structures of large proteins and protein complexes by NMR. Trends Biotechnol 1998; 16:22 – 34. 2. Otting G, Wu¨thrich K. Heteronuclear ﬁlters in two-dimensional [1H,1H] NMR spectroscopy: combined use with isotope labelling for studies of macromolecular conformation and intermolecular interactions. Q Rev Biophys 1990; 23:39 – 96. 3. Cai M, Huang Y, Sakaguchi K, Clore GM, Gronenborn AM, Craigie R. An eﬃcient and cost-eﬀective isotope labeling protocol for proteins expressed in Escherichia coli. J Biomol NMR 1998; 11: 97 – 102. 4. Campbell ID, Downing AK. NMR of modular proteins. N Struct Biol 1998; 5(suppl): 496 – 499. 5. Gooley PR, Johnson BA, Marcy AI, Cuca GC, Salowe SP, Hagmann WK, Esser CK, Springer JP. Secondary structure and zinc ligation of human recombinant short-form stromelysin by multidimensional heteronuclear NMR. Biochemistry 1993; 32:13098 – 13108. 6. Gooley PR, O’Connell JF, Marcy AI, Cuca GC, Salowe SP, Bush BL, Hermes JD, Hagmann WK, Esser CK, Springer JP, Johnson BA. The NMR structure of the inhibited catalytic domain of human stromelysin-1. Nat Struct Biol 1994; 1:111 – 118. 7. Gooley PR, O’Connell JF, Marcy AI, Cuca GC, Hagmann WK, Caldwell CG, Axel MG, Becker JW. Comparison of the structure of human recombinant short form stromelysin by multidimensional heteronuclear NMR and X-ray crystallography. J Biomol NMR 1996; 7:8 – 28. 8. Chapman KT, Kopka IE, Durette PI, Esser CK, Lanza TJ, IzquierdoMartin M, Niedzwiecki L, Chang B, Harrison RK, Kuo DW, Lin T-Y, Stein RL, Hagmann WK. Inhibition of matrix metalloproteinases by N-Carboxyalkyl peptides. J Med Chem 1993; 36:4293 – 4301. 9. Nagase H, Woessner JF Jr. Matrix metalloproteinases. J Biol Chem 1999; 274:2149 – 21494. 10. Becker JW, Marcy AI, Rokosz LL, Axel MG, Burbaum JJ, Fitzgerald PMD, Cameron PM, Esser CK, Hagmann WK, Hermes JD, Springer JP. Stromelysin-1: Three-dimensional structure of the inhibited catalytic domain and of the C-truncated proenzyme. Prot Sci 1995; 4:1966 – 1976.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

11. Li J, Brick P, O’Hare MC, Skarzynski T, Lloyd LF, Curry VA, Clark IM, Bigg HF, Hazleman BL, Cawston TE, Blow DM. Structure of fulllength porcine synovial collagenase reveals a C-terminal domain containing a calcium-linked, four-bladed h-propellor. Structure 1995; 3:541 – 549. 12. Gomis-Ru¨th FX, Gohlke U, Betz M, Kna¨uper V, Murphy G, Lo´pez-Otı´ n C, Bode W. The helping hand of collagenase-3 (MMP-13): 2.7 A˚ crystal structure of its C-terminal haemopexin-like domain. J Mol Biol 1996; 264:556 – 566. 13. Gohlke U, Gomis-Ru¨th FX, Crabbe T, Murphy G, Docherty AJP, Bode W. The C-terminal (haemopexin-like) domain structure of human gelatinase A (MMP-2): structural implications for its function. FEBS Lett 1996; 378: 126 – 130. 14. Bode W. A helping hand for collagenases: the hemopexin-like domain. Structure 1995; 3:527 – 530. 15. Strongin AY, Collier I, Bannikov G, Marmer BL, Grant GA, Goldberg GI. Mechanism of cell surface activation of 72-kDa type IV collagenase. J Biol Chem 1995; 270:5331 – 5338. 16. de Souza SJ, Pereira HM, Jacchieri S, Brentani RR. Collagen/collagenase interaction - does the enzyme mimic the conformation of its own substrate FASEB J 1996; 10:927 – 930. 17. Bode W, Gomis-Ru¨th FX, Sto¨cker W. Astacins, serralysins, snake venom and matrix metalloproteinases exhibit idential zinc-binding environments (HEXXHXXGXXH and Met-turn) and topologies and should be grouped into a common family, the ‘metzincins.’ FEBS Lett 1993; 331: 134 – 140. 18. Marcy AI, Eiberger LL, Harrison R, Chan HK, Hutchinson NI, Hagmann WK, Cameron PM, Boulton DA, Hermes JD. Human ﬁbroblast stromelysin catalytic domain: expression, puriﬁcation and characterization of a C-terminally truncated form. Biochemistry 1991; 30:6476 – 6483. 19. Moy FJ, Pisano MR, Chanda PK, Urbano C, Killar LM, Sung M-L, Powers R. Assignments, secondary structure and dynamics of the inhibitor-free fragment of human ﬁbroblast collagenase. J Biomol NMR 1997; 10:9 – 19. 20. Bartoki N, Winkler FK, Williams DH, D’Arcy A, Broadhurst MJ, Brown PA, Johnson WH, Murray EJ. Structure of the catalytic domain of human ﬁbroblast collagenase complexed with an inhibitor. Nat Struct Biol 1994; 1:106 – 110. 21. Stams T, Spurlino JC, Smith DL, Wahl RC, Ho TF, Qoronﬂeh MW, Banks TM. Structure of human neutrophil collagenase reveals large S1’ speciﬁcity pocket. Nat Struct Biol 1994; 1:119 – 123. 22. Bode W, Reinemer P, Huber R, Kleine T, Schnierer S, Tschesche H. The Xray crystal structure of the catalytic domain of human neutrophil

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

collagenase inhibited by a substrate analogue reveals the essentials for catalysis and speciﬁcity. EMBO J 1994; 13:1263 – 1269. Lovejoy B, Cleasby A, Hassell AM, Longley K, Luther MA, Weigl D, McGeehan G, McElroy AB, Drewry D, Lambert MH, Jordan SR. Structure of the catalytic domain of ﬁbroblast collagenase complexed with an inhibitor. Science 1994; 263:375 – 377. Moy FJ, Chanda PK, Chen JM, Cosmi S, Edris W, Skotnicki JS, Wilhelm J, Powers P. NMR solution structure of the catalytic fragment of human ﬁbroblast collagenase complexed with a sulfonamide derivative of a hydroxamic acid compound. Biochemistry 1999; 38:7085 – 7096. Lovejoy B, Welch AR, Carr S, Luong C, Broka C, Hendricks RT, Campbell JA, Walker KAM, Martin R, Van Watrt H, Browner MF. Crystal structures of MMP-1 and MMP-13 reveal the structural basis for selectivity of collagenase inhibitors. Nat Struct Biol 1999; 6:217 – 221. Van Doren SR, Kurochkin AV, Hu W, Ye Q-Z, Johnson LL, Hupe DJ, Zuiderweg ERP. Solution structure of the catalytic domain of human stromelysin complexed with a hydrophobic inhibitor. Prot Sci 1995; 4: 2487 – 2498. Browner MF, Smith WW, Castelhano AL. Matrilysin-inhibitor complexes: common themes among metalloproteases. Biochemistry 1995; 34:6602 – 6610. Morgunova E, Tuutila A, Bergmann U, Isupov M, Lindqvist Y, Schneider G, Tryggvason K. Structure of human pro-matrix metalloproteinase-2: activation mechanism revealed. Science 1999; 284: 1667 – 1670. Muhandiram DR, Kay LE. Gradient-enhanced triple resonance threedimensional NMR experiments with improved sensitivity. J Magn Reson 1994; 103:203 – 216. Lo¨hr F, Ru¨terjans H. A triple resonance experiment for the sequential assignment of backbone resonances in proteins. J Biomol NMR 1995; 6:189 – 197 Boucher W, Laue ED, Campbell SL, Domaille PJ. Improved 4D NMR experiments for the assignment of backbone nuclei in 13C/15N labelled proteins. J Biomol NMR 1992; 2:631 – 637. Powers R, Gronenborn AM, Clore GM, Bax A. Three-dimensional tripleresonance NMR of 13C/15N-enriched proteins using constant-time evolution. J Magn Reson 1991; 94:209 – 213. Kay LE, Xu G-Y, Singer AU, Muhandiram DR, Forman-Kay JD. A gradient-enhanced HCCH-TOCSY experiment for recording side-chain 1H and 13C correlations in H2O samples of proteins. J Magn Reson 1993; 101B:333 – 337. Senn H, Werner B, Messerle BA, Weber C, Traber R, Wu¨thrich K. Stereospeciﬁc assignment of the methyl 1H NMR lines of valine and leucine

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

in polypeptides by nonrandom 13C labelling. FEBS Lett 1989; 249: 113 – 118. Kuboniwa H, Grezesiek S, Delaglio F, Bax A. Measurement of HN-Ha J couplings in calcium-free calmodulin using new 2D and 3D water-ﬂip-back methods. J Biomol NMR 1994; 4:871 – 878. Archer SJ, Ikura M, Torchia DA, Bax A. An alternative 3D NMR technique for correlating backbone 15N with side chain Hh resonances in larger proteins. J Magn Reson 1991; 95:636 – 641. Grzesiek S, Kuboniwa H, Hinck AP, Bax A. Mutiple-quantum line narrowing for measurement of Ha-Hh J couplings in isotopically enriched proteins. J Am Chem Soc 1995; 117:5312 – 5315. Zhang O, Kay LE, Olivier JP, Forman-Kay JD. Backbone 1H and 15N resonance assignments of the N-terminal SH3 domain of drk in folded and unfolded states using enhanced sensitivity pulsed ﬁeld gradient NMR techniques. J Biol NMR 1994; 4:845 – 858. Ikura M, Kay LE, Tsudin R, Bax A. Three-dimensional NOESYHMQC spectroscopy of a 13C-labelled protein. J Magn Reson 1990; 86: 204 – 209. Vuister GW, Clore GM, Gronenborn AM, Powers R, Garrett DS, Tschudin R, Bax A. Increased resolution and improved spectral quality in fourdimensional 13C/13C-separated HMQC-NOESY-HMQC spectra using pulsed ﬁeld gradients. J Magn Reson 1993; 101B:210 – 213. Pelton JG, Torchia DA, Meadow ND, Roseman S. Tautomeric states of the active-site histidines of phosphorylated and unphosphorylated IIIGlc, a signal-transuding protein from Escherichia coli, using two-dimensional heteronuclearNMR techniques. Prot Sci 1993; 2:543 – 558. Petros AM, Kawai M, Luly JR, Fesik SW. Conformation of two nonimmunosuppressive FK506 analogs when bound to FKBP by isotopeﬁltered NMR. FEBS Lett 1992; 308:309 – 314. Burgering M, Boelens R, Kaptein R. Observation of intersubunit NOEs in a dimeric P22 Mnt repressor mutant by a time-shared [15N,13C] double halfﬁlter technique. J Biomol NMR 1993; 3:709 – 714. Gu¨ntert P, Braun W, Billeter W, Wu¨thrich K. Automated stereospeciﬁc 1H NMR assignments and their impact on the precision of protein structure determinations in solution. J Am Chem Soc 1989; 111:3997 – 4004. Gu¨ntert P, Braun W, Wu¨thrich K. Eﬃcient computation of three-dimensional protein structures in solution from nuclear magnetic resonance data using the program DIANA and the supporting programs CALIBA, HABAS and GLOMSA. J Mol Biol 1991; 217:517 – 530. Gu¨ntert P, Mumenthaler C, Wu¨thrich K. Torsion angle dynamics for NMR structure calculation with the new program DYANA. J Mol Biol 1997; 273:283 – 298.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

47. Bru¨nger AT, Adams PD, Clore GM, DeLano WL, Gros P, GrosseKunstleve RW, Jiang J-S, Kuszewski J, Nilges M, Pannu NS, Read RJ, Rice LM, Simonson T, Warren GL. Crystallography and NMR system: A new software suite for macromolecular structure determination. Acta Cryst 1998; d54:905 – 921. 48. Widmer H, Widmer A, Braun W. Extensive distance geometry calculations with diﬀerent noe calibrations - new criteria for structure selection applied to sandostatin and BPTI. J Biomol NMR 1993; 3:307 – 324. 49. Schaumann T, Braun W, Wu¨thrich K. The program FANTOM for energy reﬁnement of polypeptides and proteins using a Newton-Raphson minimizer in torsion angle space. Biopolymers 1990; 29: 679 – 694. 50. Richardson JS. h-Sheet topology and the relatedness of proteins. Nature 1977; 268:495 – 500. 51. Stockman BJ, Waldon DJ, Gates JA, Schaill TA, Kloosterman DA, Mizsak SA, Jacobsen EJ, Belonga KL, Mitchell MA, Mao B, Petke JD, Goodman L, Powers EA, Ledbetter SR, Kaytes PS, Vogeli G, Marshall VP, Petzold GL, Poorman RA. Solution structure of stromelysin complexed to thiadiazole inhibitors. Prot Sci 1998; 7:2281 – 2286. 52. Yuan P, Marshall VP, Petzold GL, Poorman RA, Stockman BJ. Dynamics of stromelysin/inhibitor interactions studied by 15N NMR relaxation measurements: Comparison of ligand binding to the S1-S3 and S1’-S3’ subsites. J Biomol NMR 1999; 15:55 – 64. 53. Spurlino JC, Smallwood AM, Carlton DD, Banks TM, Vavra KJ, Johnson JS, Cook ER, Falvo J, Wahl RC, Pulvino TA, Wendoloski JJ, Smith DL. 1.56 A˚ Structure of mature truncated human ﬁbroblast collagenase. Proteins 1994; 19:98 – 109. 54. Mathews BW. Structural basis of the action of thermolysin and related zinc peptidases. Acc Chem Res 1988; 21:333 – 340. 55. Pervushin K, Riek R, Wider G, Wu¨thrich K. Attenuated T2 relaxation by mutual cancellation of dipole-dipole coupling and chemical shift anisotropy indicates an avenue to NMR structures of very large biological macromolecules in solution. Proc Natl Acad Sci 1997; 94:12366 – 12371. 56. Tjandra N, Bax A. Direct measurement of distances and angles in biomolecules by NMR in a dilute liquid crystalline medium. Science 1997; 278:1111 – 1114. 57. Olejniczak ET, Meadows RP, Wang H, Cai M, Nettesheim DG, Fesik SW. Improved NMR structures of protein/ligand complexes using residual dipolar couplings. J Am Chem Soc 1999; 121:9249 – 9250. 58. Cornilescu G, Hu J-S, Bax A. Identiﬁcation of the hydrogen bonding network in a protein by scalar couplings J Am Chem Soc 1999; 121:2949 – 2950.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

59. Shuker SB, Hajduk PJ, Meadows RP, Fesik SW. Discovering high-aﬃnity ligands for proteins - SAR by NMR. Science 1996; 274:1531 – 1534. 60. Hajduk PJ, Sheppard G, Nettesheim DG, Olejniczak ET, Shuker SB, Meadows RP, Steinman DH, Carrera GM, Marcotte PA, Severin J, Walter K, Smith H, Gubbins E, Simmer R, Holzman TF, Morgan DW, Davidsen SK, Summers JB, Fesik SW. Discovery of potent nonpeptide inhibitors of stromelysin using SAR by NMR. J Am Chem Soc 1997; 119:5818 – 5827. 61. Schechter I, Berger A. On the size of the active site in proteases 1. Papain Biochem. Biophys Res Commun 1967; 27:157 – 162. 62. Kraulis PJ. MOLSCRIPT: A Program to Produce Both Detailed and Schematic Plots of Protein Structures. J Appl Crystallogr 1991; 24: 946 – 950.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

4 Cannabinergics: Old and New Therapeutic Possibilities Alexandros Makriyannis University of Connecticut, Storrs, Connecticut, U.S.A.

Andreas Goutopoulos Serono Reproductive Biology Institute, Rockland, Massachusetts, U.S.A.

I.

INTRODUCTION

Cannabis sativa, one of the oldest plants farmed by man, has been known for its medicinal properties for at least four millennia (Peters, 1999). The psychoactive–euphoric eﬀects of this plant, as well as its facile and wide climatic range of cultivation, have rendered it a very popular recreational drug. Today, cannabis, or marijuana, is still the focus of strong social, legal, and medical controversy over its therapeutic utility. Referenda in Arizona and California in 1997, and later, others in eight additional states, aimed at legalizing marijuana cigarettes for medical purposes. Two licensed, single-compound, cannabimimetic pharmaceuticals, Marinol (Dronabinol, delta-9-THC from Roxane Labs) and Cesamet (Nabilone, developed at Eli Lilly, now in use in the United Kingdom), are marketed for two purposes: to control the nausea and emesis produced by cancer chemotherapy and to serve as appetite stimulants in AIDS-related anorexia. In clinical trials with cancer chemotherapy patients, both these agents have proven to be superior to conventional antiemetics, such as perchlorperazine (Breivogel, 1998). Beyond this relatively limited medical use of cannabimimetics, the current, albeit long-delayed elucidation of their pharmacology is likely to lead to a wide expansion of the clinical potential and signiﬁcance of

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

these drugs. The oily, noncrystalline nature of the biologically active terpenoid ingredients of Cannabis sativa contributed to the lag in understanding of cannabinoid biology. The main active ingredient, (–)-delta-9tetrahydrocannabinol (delta-9-THC), was isolated and identiﬁed only in 1964 (Mechoulam, 1967), over a hundred years after the isolation of many important crystalline biologically active natural products, such as morphine and quinine. A second reason for the lack of progress in deﬁning the biology of cannabimimetics was the long-standing scientiﬁc misconception that the cannabinoid-induced pharmacological actions are mediated by perturbation of cellular membranes rather than through speciﬁc receptors. This hypothesis was a deterrent in the pursuit of possible speciﬁc cannabinoid binding sites. Owing to their high lipophilicity, cannabinoids were paralleled with general anesthetics in terms of their mechanism of action (Paton, 1975). Although cannabinoids were found to clearly perturb membranes (Makriyannis, 1987), such eﬀects were never proven to be directly responsible for their biological activity. The advent of synthetic cannabimimetics with a high degree of enantioselectivity (Johnson, 1986; Little, 1988) paved the road for the identiﬁcation of speciﬁc cannabinoid binding sites in rat brain (Devane, 1988). This discovery marked the onset of a revolution in the understanding of cannabinoid biology.

II. CANNABINOID RECEPTORS A. The CB1 Receptor Deﬁnitive proof of the existence of the cannabinoid receptor came with the isolation of the cDNA of a cannabinoid receptor from a rat cerebral cortex cDNA library and its expression in Chinese hamster ovary (CHO) cells (Matsuda, 1990). A year later, the corresponding human receptor, named CB1, was cloned and found to share a 97.3% homology with the rat receptor (Gerard, 1991). The CB1 472 amino acid sequence revealed (Matsuda, 1990; Gerard, 1991) that it is a member of the G-protein-coupled receptors (GPCRs). Receptors of this family are membrane embedded and consist of an extracellular N-terminus, seven transmembrane helices interconnected with intra- and extracellular loops, and an intracellular C-terminus. Bramblett et. al. (1995) constructed a model for CB1, using the known structure of bacteriorhodopsin as a starting point.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

The sites involved in interactions with G proteins of the Gi/o family are the third intracellular loop from the N-terminal side and the Cterminus (Howlett, 1998a). The C-terminus was found to bind with high aﬃnity to Gi and the synthetic C-terminus peptide was found to individually stimulate GTPgS binding to G protein and to inhibit adenylate cyclase (Howlett, 1998b). Similarly with other GPCRs, CB1 is allosterically regulated by sodium ions. It has been shown that sodium ions aﬀect both ligand binding and signal transduction by inducing a receptor conformational change (Houston, 1998). There is also evidence that an interhelical H-bonding interaction between helix II Asp and helix VII Asn is important for the stabilization of a receptor conformational state that has high aﬃnity for most cannabimimetic ligands (Tao, 1998), (Howlett, 1998a). Sodium ions presumably disrupt this H bond, and thus, result in a diﬀerent, low aﬃnity, receptor state. The CB1 receptor is coupled with Gi (Howlett, 1998a). CB1 activation leads to inhibition of adenylyl cyclase and, therefore, to reduction of cAMP levels. Many eukaryotic cells utilize cAMP as a second messenger that activates the cAMP-dependent protein kinase A (PKA), which in turn phosphorylates various proteins, regulating their function. One of the cAMP-dependent cannabinoid eﬀects is the enhancement of voltage-sensitive, outwardly rectifying potassium channels, which occurs as a result of decreased phosphorylation of the K+ channel protein by PKA (Deadwyler, 1995). Besides Gi, CB1 is coupled to Go (Howlett, 1999). Furthermore, apart from inhibition of adenylyl cyclase, CB1 utilizes several additional eﬀector systems (intracellular mediators) involving Gi/o proteins: the inhibition of N-type Ca2+ channels (Mackie, 1992); the activation of mitogen-activated protein kinase (MAP kinase) (Bouabula, 1995a); and the expression of immediate early genes like Krox-24 (Bouabula, 1995b). Other cannabinoid-induced cellular eﬀects include activation of inwardly-rectifying potassium channels (Pertwee, 1997) and possibly the activation of phospholipases A, C, or D (Felder, 1995). Diﬀerent G proteins or second messengers may be coupled to CB1 in diﬀerent brain regions and may mediate diﬀerent physiological eﬀects (Howlett, 1999). Utilization of diverse eﬀector systems by CB1 may explain how the response to cannabimimetics varies across diﬀerent cell types. Understanding which physiological responses are mediated by each of the foregoing intracellular signaling systems is of great signiﬁcance and may suggest new approaches for the design of selective cannabimimetic agents.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

B. The CB2 Receptor Homology cloning revealed the existence of a second cannabinoid receptor, CB2 (Munro, 1993). This receptor shows 44% homology to the total CB1 and 68% homology within the transmembrane regions. Not present in the brain in signiﬁcant levels, CB2 is found mainly in the periphery and particularly in tissues of the immune system, such as leukocytes, spleen, thymus, and tonsils (CB1 is found in some of these cells as well). Localization of CB2 in the immune system suggests an immunomodulatory role for this receptor. Thus, CB2 is likely to be the mediator of the long-known immunosuppressive properties of marijuana. Similarly to CB1, CB2 uses signal transduction pathways, such as inhibition of adenylyl cyclase and stimulation of MAP kinase. However, unlike CB1, CB2 does not aﬀect ion channels (Pertwee, 1997). Although the human genome does not contain genes with high homology to those of CB1 and CB2, other cannabinoid receptor types may exist nevertheless. An amino-terminal diﬀerentially spliced CB1 variant, CB1A, has been isolated from a human lung cDNA library and, akin to CB1, is expressed in the brain (Shire, 1995). The biological role and pharmacological implications of this variant are still unclear. The existence of a second peripheral CB2-like receptor is supported by the ﬁnding that palmitylethanolamide provided antinociception after intraplantar injection of formalin solution in mice paws (Calignano, 1998). This eﬀect was attenuated by a CB2 antagonist, SR144528, but not by SR14176A (a CB1 antagonist) nor by the opioid antagonist naloxone. Palmitylethanolamide has no signiﬁcant aﬃnity for either CB1 or CB2 (Khanolkar, 1996). However, in addition to the ﬁndings of Calignano et al., palmitylethanolamide is shown to have a down-regulating eﬀect on mast cell activity, presumably mediated through a CB2-like receptor present in these cells (Facci, 1995). Mouse vas deferens (MVD) seems to express CB1 and at least one CB2-like cannabinoid receptor type, as is demonstrated by the presence of CB1 and CB2-like mRNA as well as by data collected from experiments with cannabinoid receptor selective agonists and antagonists (Pertwee, 1999). Furthermore, evidence indicates that a CB1-like receptor exists in vascular endothelium, which upon activation produces signiﬁcant hypotension (Wagner, 1999). This receptor diﬀers from CB1 in its pharmacological response to some well-characterized cannabimimetics. None of these possible new CB variant receptors has been cloned yet; therefore, their existence is still putative. Thus, it is unclear whether

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

these observations are indicative of novel cannabinoid receptor types, results of alternative versions of known receptors coupled with diﬀerent eﬀector systems, or even results of diﬀerent aﬃnity states of a single cannabinoid receptor. Discovery and characterization of new cannabinoid receptors with diﬀerent distribution patterns and ligand aﬃnities is of major importance because it will provide new targets for the development of highly selective and clinically useful cannabinergic agents.

C. Cannabinoid Receptor Distribution The ubiquitous CB1 is found in the central nervous system (CNS), as well as in the periphery and in both neural and nonneural tissues; it is one of the most abundant GPCRs in the brain (Breivogel, 1998). As shown by autoradiographic studies in various mammalian brains (Herkenham, 1990; Gatley, 1998), CB1 density is highest in basal ganglia: substantia nigra pars reticulata, entopeduncular nucleus, and the external segment of globus pallidus. Moderately high CB1 density is found in putamen, cerebellum, and hippocampus, whereas moderate levels exist in cerebral cortex. The spinal cord shows a range of moderate densities, while thalamus and brain stem contain low to negligible levels. Autoradiography studies with [35S]GTPgS revealed that cannabinoid activity occurs with the same regional distribution as the receptors; however, the level of activity did not parallel receptor density (Breivogel, 1998). This pattern of CB1 distribution in the brain is similar to that of D1 receptors, which suggests that the cannabinoid system may be involved in the modulation of the dopaminergic activity (Gatley, 1998). In fact, CB1 mediates a negative feedback control over D2 in the striatum (Giuﬀrida, 1999). In the periphery, CB1 is found in the adrenal glands, bone marrow, heart, lung, prostate, testes, thymus, tonsils, spleen, lymphocytes, phagocytes, smooth muscle, vascular endothelium, peripheral neurons (e.g., in the gut), kidneys, uterus, and sperm as reviewed by Schuel et al. (1999). The CB2 receptor has a more limited distribution, being localized predominantly in the immune system. Among the human leukocytes, B lymphocytes express the highest levels of CB2, followed respectively by natural killer cells, monocytes, polymorphonuclear neutrophils, T8 lymphocytes, and T4 lymphocytes. It is also found in the lymph nodes, spleen, tonsils, and thymus (Cabral, 1999).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

D. Cannabinoids and Membranes Before the ﬁrst indication of the existence of cannabinoid receptors, the prevailing theory on the mechanism of cannabinoid activity was that cannabinoids exert their eﬀects by nonspeciﬁc interactions with cell membrane lipids (Makriyannis, 1990). Such interactions can increase the membrane ﬂuidity, perturb the lipid bilayer and concomitantly alter the function of membrane-associated proteins (Loh, 1980). A plethora of experimental evidence suggests that cannabinoids interact with membrane lipids and modify the properties of membranes. However, the relevance of these phenomena to biological activities is still only, at best, correlative. An important conundrum associated with the membrane theories of cannabinoid activity is uncertainty over whether cannabinoids can achieve in vivo membrane concentrations comparable to the relatively high concentrations used in in vitro biophysical studies (Makriyannis, 1995). It may be possible that local high concentrations are attainable under certain physiological circumstances, and, if so, some of the cannabinoid activities may indeed be mediated through membrane perturbation. Interactions of cannabimimetics with membranes may be of importance for auxiliary roles such as transport to their sites of action and proper orientation for optimum interaction with their receptors. The molecular features of cannabimimetics are shown to govern the manner by which these molecules cross cell membranes, including the brain–blood barrier (BBB) (Makriyannis, 1995). Most cannabimimetics are amphipathic, a property that aﬀects their orientation within the lipid bilayer. Strong experimental evidence has shown that the phenolic hydroxyl group of (–)-delta-9-THC anchors it at the polar interface of the membrane, whereas the tricyclic hydrophobic system remains imbedded in the bilayer and perpendicular to the fatty acid chains (Martel, 1993; Makriyannis, 1995). Other active dihydroxy THC derivatives adopt similar orientation, while many inactive analogs assume an orientation parallel to the lipids and a position deeper, closer to the center of the bilayer. The proper positioning and orientation of cannabinoids within the membrane may be crucial for reaching the receptor site, located within the transmembrane receptor helices, by lateral diﬀusion (Makriyannis, 1995).

III. THE ENDOGENOUS LIGANDS The discovery of the cannabinoid receptors and their G-protein-coupled nature strongly suggested the existence of endogenous cannabimimetic

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 1 The structure of anandamide.

ligand(s) able to exert physiological activity upon binding to these receptors. Initial eﬀorts to identify a possible protein (Nye, 1988) or a watersoluble endogenous cannabimimetic ligand were unsuccessful (Deadwyler, 1995). The hypothesis that such a putative endocannabinoid should be lipophilic, like the classical exogenous cannabinoids, led Mechoulam et al. to seek such a ligand in the hydrophobic fractions of porcine brain extracts (Devane, 1992). Repetitive fractionations and puriﬁcations led to the identiﬁcation of a substance that bound to CB1 in a saturable fashion. This compound was the ethanolamide of arachidonic acid (arachidonyl ethanolamide, AEA) (Fig. 1). The authors named this brain constituent anandamide from ananda, the Sanskrit word for bliss. Anandamide is found in human brain: 100 pmol/g in the hippocampus, 75 pmol/g in the thalamus, 60 pmol/g in the cerebellum, and 55 pmol/g in the striatum (Martin, 1999). The concentration of AEA increases postmortem, especially when the brain is kept at ambient temperature. Furthermore, AEA surges are observed when cerebellar granule cells are treated in hypoxic conditions (Hillard, 1997). Although such concentration increases may be artifacts of postmortem brain damage, they may also occur in living tissue under certain conditions, such as hypoxia. Outside the CNS, anandamide is found in the spleen and heart at approximately 10 pmol/g (Martin, 1999). It is also localized in rat testes and uterus in concentrations signiﬁcantly greater than those in the brain

Figure 2 The structures of two N-acylethanolamide (NAE) endocannabinoids.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 3

The structure of 2-arachidonyl glycerol.

(Schmid, 1997). Very low levels have been detected in serum, plasma, and cerebrospinal ﬂuid—a fact that indicates that anandamide is not hormonal in nature but rather is biosynthesized at or near its sites of action. In addition to anandamide, several other endogenous polyunsaturated fatty acid derivatives were also found to act as cannabimimetics. They are all now collectively referred to as endocannabinoids. Soon after the discovery of anandamide, two more fatty acid ethanolamides were isolated and found to bind to CB1 preparations with aﬃnities similar to that of anandamide (anandamide CB1 binding aﬃnity Ki = 39.2 nM, according to Hanus et al., 1993). These were the homo-g-linolenylethanolamide (CB1 Ki = 53.4 nM) and 7,10,13,16-docosatetraenylethanolamide (CB1 Ki = 34.4 nM) (Fig. 2). All three N-acylethanolamide endocannabinoids were found to be CB1 agonists in the MVD test (Pertwee, 1994). A diﬀerent type of endocannabinoid that is also an arachidonic acid derivative was ﬁrst isolated from canine gut and identiﬁed as 2-arachidonyl glycerol (2-AG) (Fig. 3) (Mechoulam, 1995a). Later 2-AG was also found in the brain (Stella, 1997) and spleen (Di Marzo, 1998). It was shown to be released in a calcium-dependent manner, reaching concentrations 170 times higher than that of anandamide in the brain (Stella, 1997). Like the other endocannabinoids, 2-AG was shown to produce the typical tetrad of cannabimimetic behavioral eﬀects and inhibit electrically evoked contractions of mouse MVD (Mechoulam, 1995a).

A. Anandamide Pharmacology Since the discovery of anandamide in 1992, a number of studies have examined its pharmacological properties. Although its roles are still elusive, a plethora of data supports the initial postulate that anandamide is the major endogenous cannabinoid ligand. As mentioned earlier, anandamide binds to CB1 from brain preparations and displaces various well-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

characterized cannabimimetic radioligands (Hillard, 1997). Furthermore, it binds to CB1 expressed in cells transfected with CB1 DNA (Vogel, 1993). Its CB1 aﬃnity is comparable to that of delta-9-THC. Anandamide does not have eﬀects on other than the cannabinoid receptors (Hillard, 1997) and has cannabimimetic properties, both in vitro and in vivo. Anandamide acts as a CB1 agonist, as demonstrated when it inhibited forskolinstimulated adenylyl cyclase activity in N18TG2 cells (IC50 = 540 nM) (Vogel, 1993), CB1 expressing CHO cells (IC50 = 160–322 nM) (Vogel, 1993), and cerebral membranes (IC50 = 1.9 AM) (Childers, 1994). It was found to have lower eﬃcacy (lower maximal eﬀect) than the high-aﬃnity cannabimimetics WIN55212-2 and CP-55,940; thus, anandamide was classiﬁed as a partial agonist (Vogel, 1993; Childers, 1994). Furthermore, it was found to have inhibitory eﬀects in N-type calcium currents through a pertussis-toxin-sensitive pathway in N18 neuroblastoma cells (Mackie, 1993). Anandamide, in vivo, was shown to produce the four characteristic eﬀects of cannabimimetics, namely, analgesia, hypothermia, hypoactivity, and catalepsy (Smith, 1994; Fride, 1993; Crawley, 1993). These four eﬀects are not unique to cannabimimetics; when they are produced together, however, they are highly predictive of cannabimimetic activity (Martin, 1991). Anandamide was found to be less potent than delta-9THC in producing these behavioral eﬀects in mice (Fride, 1993). It has quicker onset and shorter duration of action, the latter because of rapid catabolism. Cross-tolerance studies, in which pretreatment of mice with delta-9-THC produced tolerance to most of the pharmacological eﬀects of anandamide and vice versa, indicate that both drugs act on the same receptor (Jarbe, 1998). In addition, anandamide was found to parallel classical cannabinoid pharmacology in a series of nonbehavioral experimental systems. In isolated MVD, (Pertwee, 1992) and guinea pig ileum, it inhibited electrically evoked twitch responses (Pertwee, 1995). Moreover, anandamide was shown to decrease intraocular pressure in rabbits (Pate, 1995), to reduce sperm-fertilizing capacity in sea urchins by inhibition of the acrosome reaction (Schuel, 1994), and to produce hypotension in rats (Varga, 1995). All the foregoing pharmacological eﬀects of anandamide, in conjunction with the well-documented existence of speciﬁc systems for its biosynthesis, catabolism, and cellular reuptake to be discussed shortly, suggest that anandamide is indeed the endogenous cannabinoid ligand. The other two less studied N-acylethanolamide endocannabinoids and also 2-AG may serve similar functions. The diﬀerential roles of each of these four endocannabinoids are still unclear.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Despite all the research conducted on the pharmacological eﬀects of anandamide and other cannabimimetics, the exact role of the endogenous cannabinoid system remains elusive. The ubiquitous pharmacology of cannabimimetics suggests that endocannabinoids have diﬀering functions depending on the tissue or organ system. A recent report (Giuﬀrida, 1999) sheds light on the role of anandamide in the dorsal striatum. It was shown that anandamide, but not 2-AG, is released after D2 activation and subsequently suppresses motor activity, possibly by inhibiting postsynaptic GABAergic currents. Therefore, it was suggested that anandamide, at least in the striatum, plays the role of an autocoid (local neuromodulator) that has a negative feedforward regulatory eﬀect of D2-mediated locomotor behavior (Giuﬀrida, 1999). It may be that 2-AG has diﬀerent roles in the CNS, for it can reach 170 times higher concentrations than that of anandamide in the brain (Stella, 1997), even though it was undetectable in the striatum. In the hippocampus, 2-AG, but not anandamide, was released after glutamatergic activation (Stella, 1997). Sugiura (1999) found 2-AG to be a full CB1 agonist, whereas anandamide is a partial agonist, again pointing to alternative roles for 2-AG in comparison to anandamide.

B. Endocannabinoid Metabolism Biosynthesis of Anandamide Considerable advances have been made during the late 1990s toward understanding the physiological pathways that are involved in the synthesis and inactivation of endocannabinoids. The ﬁrst of these pathways to be observed, an enzymatic activity responsible for anandamide hydrolysis, led to lower apparent CB1 aﬃnities for anandamide analogs in studies involving structure–activity relationships (SARs) (Childers, 1994), (Abadji, 1994). Inclusion in the binding assay of phenylmethanesulfonyl ﬂuoride (PMSF), a general serine protease inhibitor, protected the anandamide analog from hydrolysis (Abadji, 1994; Khanolkar, 1996). Shortly after, an enzyme speciﬁc for this hydrolytic process was identiﬁed and characterized (Deutsch, 1993; Ueda, 1995). Initially, it was thought that this hydrolase, named anandamide amidase or fatty acid amidohydrolase (FAAH), was also responsible for the synthesis of anandamide by acting reversibly (Devane, 1994). However, the current belief is that anandamide amidase is unlikely to be physiologically responsible for anandamide synthesis because of the requirement for signiﬁcantly higher than normal

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

physiological concentrations of arachidonic acid and ethanolamine (up to 160 mM) for this enzyme to catalyze the reverse reaction (Piomelli, 1998). Therefore, it is currently believed that anandamide is formed from membrane phospholipids (Fig. 4) through a pathway that involves: (1) a trans-acylation of the amino group of phosphatidylethanolamine with arachidonate from the sn-1 position of phosphatidylcholine and (2) a Dtype phosphodiesterase activity on the resulting N-arachidonylphosphatidylethanolamide (NAPE). Synthesis of anandamide is presumably regulated at the levels of both enzymes, the N-acyltranferase and the phospholipase D, by stimuli that raise intracellular calcium or by receptors linked with cAMP and PKA. It has been shown that anandamide is formed when neurons are depolarized and, therefore, the intracellular calcium ion levels are elevated (Cadas, 1996). Biosynthesis of 2-AG Two possible pathways for the biosynthesis of 2-AG have been proposed: (1) a phospholipase C (PLC) hydrolysis of membrane phospholipids followed by a second hydrolysis of the resulting 1,2-diacylglycerol by diacylglycerol lipase or (2) a phospholipase A1 (PLA1) activity that generates a lysophospholipid, which in turn is hydrolyzed to 2-AG by lysophospholipase C (Fig. 5) (Piomelli, 1998). Alternative pathways may also exist from either triacylglycerols by a neutral lipase activity or lysophosphatidic acid by a dephosphorylase. The fact that PLC and diacylglycerol lipase inhibitors inhibit 2-AG formation in cortical neurons supports the contention that 2-AG is, at least predominantly, biosynthesized by the PLC pathway (Stella, 1997). However, a mixed pathway may also be plausible. As with the biosynthesis of anandamide, the biosynthesis of 2-AG is also triggered by increases of intracellular calcium ions that result from neuronal activity. High frequency stimulation of neurons produced a fourfold increase of 2-AG accumulation compared with controls, and this was prevented by sodium ion channel blocking or removal of calcium ions (Stella, 1997). The concentration of 2-AG in depolarized neurons reached 1 to 2 AM, signiﬁcantly higher than anandamide and suﬃcient to substantially activate CB1 (Stella, 1997). Based on the pathways just proposed for the biosynthesis of anandamide and 2-AG, the formation of these endogenous ligands must also be dependent on the composition of the precursor lipids. This dependence is of greater importance for anandamide rather than for 2-AG because

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 4

Anandamide biosynthesis.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 5 Possible 2-AG biosynthesis.

arachidonic acid is rarely esteriﬁed at the sn-1 position of phospholipids, whereas it is commonly found at the sn-2 position. Endocannabinoid Release Immediately after synthesis, endocannabinoids are released in the extracellular space, where they then act on the same or neighboring cells as autocrine or paracrine mediators (Di Marzo, 1999). Experimental evidence thus far indicates that anandamide and 2-AG, unlike other classical neurotransmitters, are not stored in vesicles. First, anandamide basal concentrations are extremely low (5–10 pmol/g), 100 to 10,000 times lower than those of classical neurotransmitters (Cadas, 1997). Second, stimulus-dependent anandamide release is linked with de novo NAPE and

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

subsequent anandamide biosynthesis (Cadas, 1996; Di Marzo, 1999). Therefore, it is currently believed that anandamide and 2-AG are produced and immediately released from the neurons upon demand (Di Marzo, 1999b; Piomelli, 1998). The poor water solubility of anandamide must preclude extensive free diﬀusion in the extracellular space. Since, however, anandamide is found in brain incubation media or perfusates of brain microdialysis experiments it obviously exits the cells (Giuﬀrida, 1999). Additionally, it is known that striatal astrocytes, which do not produce anandamide, do respond to it (Cadas, 1996). Therefore, it has been suggested that after cleavage from NAPE, anandamide is immediately released from the membrane with the assistance of a membrane transporter (such as a P-glycoprotein) (Ayotte, Picone, and Makriyannis; unpublished results) or a lipid binding protein (like a lipocalin) (Piomelli, 1998). Such a lipid binding protein may also facilitate the passage of anandamide through the aqueous extracellular space to its sites of action. Endocannabinoid Inactivation Anandamide is inactivated in two steps, ﬁrst by transport inside the cell and subsequently by intracellular enzymatic hydrolysis. The transport of anandamide inside the cell is a carrier-mediated activity, having been shown to be a saturable, time- and temperature-dependent process that involves some protein with high aﬃnity and speciﬁcity for anandamide (Beltramo, 1997). This transport process, unlike that of classical neurotransmitters, is Na+-independent and driven only by the concentration gradient of anandamide (Piomelli, 1998). Although the anandamide transporter protein has not been cloned yet, its well characterized activity is known to be inhibited by speciﬁc transporter inhibitors. Reuptake of 2-AG is probably mediated by the same facilitating mechanism (Di Marzo, 1999a,b; Piomelli, 1999). Once inside the cell, anandamide is hydrolyzed by a speciﬁc hydrolase, anandamide amidase (AEAase) or fatty acid amidohydrolase (FAAH) (Desarnaud, 1995; Deutsch, 1993). This enzyme is membrane associated and shows signiﬁcant speciﬁcity for anandamide (Desarnaud, 1995; Lang, 1999). There is some evidence that in cells with low anandamide amidase activity, such as platelets and neutrophils, anandamide is inactivated by an oxidative pathway involving 12(S)-lipoxygenase (Edgemond, 1998). Metabolism of anandamide by enzymes of the arachidonic acid cascade

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

(Fig. 6) may be of physiological importance and may lead to possible biologically active oxygenated anandamide analogs. These pathways have not been explored yet. Nevertheless, it was shown that the 11-, 12-, and 15lipoxygenases recognize anandamide and catalyze its hydroxylation in vitro (Hampson, 1995). Among the resulting oxygenated anandamides, only the product of 11-lipoxygenase showed aﬃnity to CB1 comparable to that of anandamide (Hampson, 1995). The physiological relevance of this ﬁnding, if any, is unknown at present. Less explored is the role and metabolic fate of 2-AG. It is possible that in many tissues, 2-AG is only an intermediate of a signaling pathway that generates 1,2-diacylglycerol and arachidonic acid, two well-known signaling molecules. In the brain however, 2-AG may have regulatory roles, since it escapes immediate metabolism and accumulates in response to stimuli-generated Ca2+ surges (Stella, 1997). These diﬀerences may arise

Figure 6 Anandamide metabolism: NAPE, N-arachidonylphosphatidyl-ethanolamides; PLD, phospholipase D; AEA, anandamide; AC, anandamide carrier protein; AT, anandamide transporter; AEAase, anandamide amidase; AA, arachidonic acid.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

from diﬀerences in the involved isoenzymes or their levels of expression from tissue to tissue. Anandamide amidase recognizes and hydrolyzes 2-AG (Goparaju, 1999; Di Marzo, 1999; Lang, 1999); however, there is evidence for the existence of another speciﬁc hydrolase [monoacylglycerol (MAG) lipase] that hydrolyzes 2-AG (D. Piomelli and A. Makriyannis, 2000, personal communication). In addition to this pathway, 2-AG diﬀuses rapidly into the cell membrane where it could be either hydrolyzed to arachidonic acid and glycerol or esteriﬁed back to phosphoglycerides (Di Marzo, 1999b).

IV. THE ENDOCANNABINOID SYSTEM It is apparent that a series of critical research breakthroughs during the last decade have unveiled a new signiﬁcant biological assemblage, the endocannabinoid system. This system, which is evolutionarily well conserved, consists of at least two receptor types, each with diﬀerent localization and functions; a family of endogenous ligands; and a speciﬁc molecular machinery for the synthesis, transport and inactivation of these ligands. Although information about this system is now emerging, many signiﬁcant questions still remain unanswered. The anandamide transporter and some of the endocannabinoid metabolic enzymes have yet to be cloned. The accomplishment of a highly quantitative and detailed mapping of the endocannabinoid system will produce more information about its physiological roles. The advent of speciﬁc cannabinoid receptor antagonists (Pertwee, 1995c), (Rinaldi-Carmona, 1994) has already facilitated pharmacological studies by enabling reversal of the endogenous cannabinoid tone, as well as veriﬁcation of the interaction of various agents with these receptors. The ﬁrst inhibitors of anandamide amidase (Boger, 2000; Deutch, 1997a; Deutch, 1997b; Pertwee, 1995b) and its transporter (Beltramo, 1997; Christie, 2001; Wilson, 2001) are becoming important tools in understanding the functions of the endocannabinoid system by producing an hypercannabinoid state. Thorough understanding of this system and its functions in physiological and disease conditions will likely lead to the development of new therapeutics. Invaluable tools for such studies are selective agents capable of interacting with the protein members of the cannabinoid system and, in turn, either activating or inhibiting them. Therefore, the study of the SAR of each of these targets and the identiﬁcation of diﬀerences in ligand recognition comprises a task of great signiﬁcance, one that can lead to the development of highly selective cannabinergic agents. The term ‘‘canna-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 7 Cannabinoid targets for drug design.

binergic’’ encompasses ligands that act on proteins of the endocannabinoid system, regardless of chemical classiﬁcation or type of resultant pharmacological activity. Therefore, this general term includes agents that act on the cannabinoid receptors, either as agonists or antagonists, as well as molecules that inhibit AEAase or the anandamide transporter (AT). The therapeutic potential that emanates from modulating these proteins renders them important yet unexploited targets for drug design and development (Fig. 7). All the aforementioned protein members of the cannabinoid system are large, membrane-bound proteins; therefore, it is particularly diﬃcult to obtain direct information about their tertiary structure. Thus, at the present time, structure-based drug design is not feasible. Detailed exploration of the SAR and subsequent ligand-based design are the most appropriate means for the development of molecular probes for these proteins.

V. MAJOR CLASSES OF CANNABINERGIC LIGANDS Based on chemical structure, cannabinergic ligands are classiﬁed into ﬁve major classes. Structures of representative members from each of the ﬁve chemical classes are shown in Figures 8 to 11.

A. Classical Cannabinoids Classical cannabinoids (CCs) are tricyclic terpenoid derivatives bearing a benzopyran moiety. This class includes the natural product (–)-deltanine-tetrahydrocannabinol (Fig. 8, 1) and the other pharmacologically active constituents of the plant Cannabis sativa.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 8

Structures of representative classical cannabinoids.

Figure 9

Nonclassical cannabinoids (NCCs).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 10 Representative cannabinergic aminoalkylindoles.

Many classical cannabinoid analogs have been synthesized and evaluated pharmacologically and biochemically (Razdan, 1986; Mechoulam, 1999). The CC structural features that seem to be important for cannabimimetic activity (Makriyannis, 1990) are as follows: 1.

The phenolic hydroxyl group, can be substituted by an amino group but not by a thiol group. In contrast to the traditional CC

Figure 11 Structures of representative endocannabinoid analogs.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

2.

3. 4.

5.

SAR, which considers the phenolic hydroxyl to be one of the necessary pharmacophoric groups, analogs lacking it or bearing it in its etheriﬁed form retain high receptor binding aﬃnity, (e.g., analog 2a) especially for CB2 (Huﬀman, 1996). The benzopyran ring is not essential for activity. The pyran oxygen can be substituted by nitrogen or can be eliminated in open-ring mono- or bisphenolic compounds. The recently developed CB2-selective ligand HU-308 (5) is an example of such a bicyclic cannabinoid (Hanus et al., 1999). Neither the double bond nor the 9-methyl group are necessary for activity. The alkyl chain is probably the most essential CC pharmacophoric group. Increased biological activity results from elongating the ﬁve-carbon delta-eight-THC chain to a sevencarbon chain substituted with 1V,1V- (e.g., 2) or 1V,2V-dimethyl or with 1V,1V-cyclic moieties (e.g., 3, AMG3). Oxygen atoms (ethers) and unsaturation (Papahatjis, 1998) within the chain, or terminal halogens, carboxamido, and cyano groups are well tolerated (Khanolkar, 2000). An additional pharmacophore introduced in the nonclassical cannabinoid series is the southern aliphatic hydroxyl (Makriyannis, 1990). A variation involves the highly potent classical/nonclassical cannabinoid hybrids (e.g., 4, AM919) (Drake, 1998).

B. Nonclassical Cannabinoids A second class of cannabimimetics was developed at Pﬁzer, in an eﬀort to simplify the structure of CCs while maintaining or improving activity (Johnson, 1986). This class includes bicyclic (e.g., 6) and tricyclic (e.g., 7) analogs lacking the pyran ring of CCs (Fig. 9). These compounds are collectively speciﬁed as ‘‘non-classical cannabinoids’’ (NCCs). The crystalline CP55,940 (6) and its tritiated analog show high aﬃnity, eﬃcacy, and stereoselectivity to both cannabinoid receptors and have been used extensively as pharmacological tools. The key compound that led to the discovery of CB1 was [3H]CP55,940 (Devane, 1988). The structural resemblance of NCCs and CCs, as well as their comparable SARs, indicate that they bind to CB1 in a similar fashion. The side chain and the phenolic hydroxyl of an NCC are crucial for activity. The hydroxypropyl chain of CP55,940 is not necessary for

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

activity. However, when present, its stereochemistry is important and shows a strong preference for the beta relative conﬁguration.

C. Aminoalkylindoles The third chemical class of cannabinergics is that of aminoalkylindoles (AAIs) (Fig. 10). They were developed at Sterling Winthrop as potential nonsteroidal anti-inﬂammatory agents (Bell, 1991). These ﬁrst analogs exhibited antinociceptive properties that were eventually attributed to interactions with the cannabinoid receptors. Compound 8 (WIN55212) is a potent CB1 and CB2 agonist with high stereoselectivity and a slight preference for CB2. AM630 (9), the ﬁrst CB2-selective antagonist derived from this class of compounds, was developed in our laboratory after longterm eﬀorts to obtain such an inhibitor (Pertwee, 1995a). We have recently reported the development of AM1241, a potent, highly CB2-selective agonist (Malan, 2001). This class of compounds diﬀers from the ﬁrst two by being considerably less lipophilic and more ‘‘druglike.’’ Labeling of CB1 with electrophilic AAIs almost abolished the receptor’s ability to bind to CP55,940, indicating that AAIs and NCCs (as well as CCs) share at least some points of interactions with CB1 (Yamada, 1996). Several models have attempted to deﬁne the pharmacophoric equivalency between the functional groups of AAIs, NCCs, and CCs (Xie, 1995), (Huﬀman, 1994). Although these three diﬀerent classes of cannabimimetics show similarities in their binding with CB1, they diﬀer considerably in the susceptibility of their binding aﬃnities to diﬀerent Na+-modulated allosteric receptor states (Houston, 1998). They also diﬀer in their aﬃnities to several CB1 mutants (Chin, 1998), as well as in the way they activate the receptor (Houston, 1998). These diﬀerences may be explained by the existence of more than one ligand binding motif, or by ligand binding to partially overlapping but distinct receptor binding subsites, or even by induction of diﬀerent receptor conformational changes upon binding of diﬀerent ligands (Howlett, 1998a). It has been proposed that structurally dissimilar ligands may evoke diﬀerent receptor–G-protein coupling (Houston, 1998). Therefore, analogs from diﬀerent cannabinoid ligand classes may evolve as selective pharmacological agents exhibiting only speciﬁc cannabimimetic eﬀects. Structural features of AAI important for cannabinergic activity are the 3-aroyl moiety and the 1-chain, which must contain nitrogen, most often in a heterocyclic ring (e.g., piperidino or morpholino). This chain can

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

be conformationally restricted as part of a six-membered ring fused to the indole nucleus (D’Ambra, 1992).

D. Endocannabinoids The class of the endogenous cannabinoids (endocannabinoids) was discovered in 1992 as molecules produced by mammalian cells with aﬃnity for the cannabinoid receptor (Devane, 1992). This class includes lipid molecules such as fatty acid ethanolamides, monoacylglycerols, and related synthetic analogs. The two prototypes in this class are the ethanolamide of arachidonic acid (anandamide) and 2-arachidonyl glycerol (2-AG). Its (R)1V-methylated analog, AM356 (10) (Fig. 11) shows higher aﬃnity and remarkable metabolic stability (Abadji, 1994). This analog, named Rmethanandamide, has been established as a standard CB1-selective agonist in the cannabinoid ﬁeld. The (R,R)-2,1V-dimethyl anandamide was reported recently to exhibit a threefold improved aﬃnity over R-methanandamide and signiﬁcant enantioselectivity (Goutopoulos, 2001). Other modiﬁcations that result in high CB1 aﬃnity include the substitution of the hydroxyl group with halogen, or the methyl group, and the substitution of the terminal n-pentyl chain with the dimethylheptyl chain, reminiscent of potent classical cannabinoid ligands (e.g., 12, O-1064) (Pertwee, 2000). This compound class also includes some fatty acid analogs designed for endocannabinoid targets other than the cannabinoid receptors. For instance, arachidonyltriﬂuoromethylketone (ATFMK) (13) and hexadecylsulfonyl ﬂuoride (14, AM374) are potent inhibitors of anandamide amidase. The ﬁrst inhibitor of the anandamide transporter to play an important role in the discovery of this transport process was AM404 (15) (Beltramo, 1997).

E. 1,5-Biarylpyrazoles The ﬁfth class, 1,5 biarylpyrazoles, was developed at Sanoﬁ in 1994 from a hit generated by high throughput screening for cannabinoid receptor ligands (Rinaldi-Carmona, 1994). Compounds of this class act as cannabinoid receptor antagonists. Figure 12 shows SR141716A (16), which was reported, simultaneously with AM630, as the ﬁrst CB1 antagonist and has since been used extensively as an important pharmacological tool. SR141716A shows selectivity for CB1 and often acts as an inverse agonist rather than a pure antagonist (Pertwee, 2000). Also developed at Sanoﬁ, SR144528 (17) acts as an antagonist/inverse agonist with selectivity for CB2. A useful radioimaging agent in PET and SPECT studies [123I]

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 12 1,5 Biarylpyrazole cannabinoid receptor antagonists.

AM281, a 123I-labeled 1,5-biarylpyrazole was synthesized in our laboratory (Gatley, 1998).

VI. THERAPEUTIC POTENTIAL OF CANNABINERGIC AGENTS Most known cannabimimetics today have very broad eﬀects on organ systems, several of which are still not completely delineated. The ubiquitous pharmacology of cannabimimetics is one of the reasons for the failure, thus far, of the clinical application of these drugs to reach its full potential. The sections that follow summarize the eﬀects of cannabinergics on the various physiological systems and the possible therapeutic uses that may arise from these biological activities.

A. Nervous System The primary system of cannabimimetic activity is the nervous system. The CB1 receptor is omnipresent in the brain, especially in areas that control functions aﬀected by cannabimimetics. One of the functions most pronouncedly inﬂuenced by cannabimimetics is motor behavior. Catalepsy, immobility, ataxia, and impairment of complex behavioral acts after acute administration of high doses of cannabimimetics are manifestations of such motor eﬀects (Pertwee, 1997). In lower doses cannabimimetics produce the opposite eﬀects. The very dense presence of CB1 in the cerebellum and the basal ganglia, areas responsible for motor activity, is

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

congruent with these observations. The GABA function in the basal ganglia is enhanced by CB1 agonists (Consroe, 1998). Cannabimimetics seem to exert an important modulatory action in basal ganglia output nuclei by inhibiting both inhibitory striatal input, which is tonically inactive, and excitatory subthalamic input, which is tonically active (Sanudo-Pena, 1999). The net cannabimimetic eﬀect on motor activity depends on the level of activity of each of these two functions. This may explain the biphasic eﬀect of cannabimimetics on motor behavior. An important recent discovery has advanced the current understanding of how cannabimimetics are implicated in the control of motor behavior (Giuﬀrida, 1999). Giuﬀrida et al. have reported that D2 activation in the striatum results in release of the endocannabinoid anandamide, which in turn seems to mediate a negative feedback control, counteracting dopamine-induced facilitation of motor activity (Giuﬀrida, 1999). Because of these eﬀects of cannabinergics on the basal ganglia and subsequently on motor activity, it has been suggested that cannabinergics may be useful agents in the treatment of motor disorders such as choreas, Tourette’s syndrome, dystonias, and Parkinson’s disease (Consroe, 1998). In general, by increasing hypokinetic features in the basal ganglia, CB1 agonists may alleviate the various hyperkinetic manifestations, such as choreic movements, that characterize basal ganglia disorders. Direct evidence suggesting the involvement of CB1 in Huntington’s chorea is the extensive loss of CB1 receptors in the substantia nigra and lateral globus pallidus (Glass, 1993). It is still unclear whether these observations are causative of Huntington’s disease or its results. However, this ﬁnding alone argues that a suitable CB1 ligand could potentially be useful as a diagnostic agent for this chorea. Furthermore, the presence of CB1 in the structures and pathways associated with the pathophysiology of Tourette’s syndrome, and especially the functional link between CB1 and D1, D2, also argues that the endocannabinoid system may have some involvement in this disorder as well (Consroe, 1998). In addition, it has been suggested that activation of CB1 receptors, also owing to their link with the dopaminergic system, may reduce dyskinesia produced by L-DOPA in patients with Parkinson’s disease (Brotsie, 1998). The CB1 receptors present in the hippocampus, amygdala, and cerebral cortex may be responsible for observations that cannabimimetics are eﬀective against some types of seizures (Consroe, 1998). The anticonvulsant and antispastic eﬀects of cannabinoids are well documented, however the mechanisms of these eﬀects are still unclear (Nahas, 1999).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Orally administered cannabimimetics can relieve some of the symptoms of multiple sclerosis and spinal cord injury such as muscle spasticity, pain, tremor, nystagmus, and nocturia (Pertwee, 2000). Recent studies (Baker, 2000; Baker, 2001) have shown that exogenously administered cannabimimetics control spasticity in a multiple sclerosis (MS) model. Possible implication of both CB1 and CB2 receptors has been suggested. Agents that elevate anandamide levels by inhibiting AEAase (AM374) or AT (AM404) also produced these antispastic eﬀects indirectly. Cannabinoid receptor antagonists blocked these antispastic eﬀects. Respectively, SR141716A and SR1445228, selective CB1 and CB2 antagonists/inverse agonists, produced enhanced spasticity when administered alone to the same animal model (Baker, 2000). Furthermore, it was evident that endocannabinoids are released during episodes of MS, during which they alleviate the spastic eﬀects of the disease (Pertwee, 2000). These ﬁndings conﬁrm, at least to some extent, the anecdotal reports that marijuana smoking alleviates the symptoms in MS patients and establishes cannabimimetics as exciting candidates for the development of agents that control spasticity and other abnormalities resulting from some neurodegenerative diseases. These agents may also control spasticity produced by spinal cord injury by acting on spinal as well as on supraspinal mechanisms (Consroe, 1998). It has been suggested that the eﬀect of cannabimimetics on the release of glutamate in the substantia nigra appears to be the most important supraspinal mechanism of cannabimimetic-induced control of spasticity (Consroe, 1998). The CB1-mediated inhibition of glutamate release in the hippocampus was also suggested to be the most likely mechanism of the neuroprotective eﬀects of WIN5521,2 observed in both the global and focal cerebral ischemia animal models (Nagayama, 1999). These eﬀects were stereoselective and were blocked by SR141716A. Therefore, cannabimimetics may ﬁnd potential therapeutic utility in the treatment of disorders resulting from cerebral ischemia, including stroke. Another neuroprotective activity of cannabimimetics was shown to be associated with the CB1-mediated inhibition of nitric oxide (NO) release from rat microglial cells (Waksman, 1999). This study suggests cannabimimetics as potentially useful agents in brain injury resulting from inﬂammatory neurodegenerative processes, especially those involving activation of microglial cells, such as AIDS-encephalitis. Another signiﬁcant cannabinoid activity that is mediated by the nervous system arises from the antinociceptive properties of these agents. Compelling evidence suggests that cannabimimetics are eﬀective in the

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

control of acute and chronic pain in a variety of antinociceptive tests in animals (Martin, 1998). Synthetic cannabimimetics have been classiﬁed as equal to morphine in potency and eﬃcacy (Walker, 1999). The mechanism of the cannabimimetic-induced analgesia is multifaceted and occurs at several levels: (1) directly on spinal cord mechanisms (Walker, 1999); (2) in supraspinal mechanisms, speciﬁcally in the thalamus and the periaqueductal gray (PAG) matters (Walker, 1999; Martin, 1998); and (3) in the periphery, possibly involving CB1-like and CB2-like receptors (Calignano, 1998). Other systems, such as n and A opiate receptors, as well as spinal noradrenergic mechanisms, seem to be involved in the cannabimimeticproduced analgesia (Walker, 1999). Evidence supports the suggestion that cannabimimetics are eﬀective in animal models of chronic pain, a type of pain that is poorly managed by opioids (Walker, 1999). It has also been suggested that CB1 agonists may be superior to morphine in suppressing pain caused by nerve damage (Pertwee, 2000). This type of pain is signaled by abnormal discharges of Ah and Adelta ﬁbers, which are much more populated by CB1 than A-opioid receptors. Another category of CNS-mediated cannabinoid eﬀects includes alterations in cognition and memory. Cannabimimetics have been shown to interfere with the mechanisms of long-term potentiation (LTP), a candidate mechanism for learning and memory. They also alter presynaptic release of GABA and glutamate from hippocampal neurons (Hampson, 1998). Hippocampus, a structure rich in CB1, plays a major role in memory processing, especially by enabling memory retrieval, whereas retrohippocampal areas with fewer CB1 receptors are responsible for memory storage. Hippocampal lesions in rodents impair short-term memory. Several behavioral studies have exhibited that cannabinoids disrupt information processing in the hippocampus, acting as ‘‘reversible’’ hippocampal lesions (Hampson, 1999). It is suggested that the role of CB1 in these regions is to regulate storage information by switching hippocampal memory circuits (Hampson, 1998). The role of the cannabinoid system in memory and cognition renders it a possible target for memory and cognition enhancing agents. This possibility is strongly supported by some recent advances in understanding the neurobiology of the endocannabinoid system (Wilson, 2001; Christie, 2001; Kreitzer, 2001; Ohno-Shosaku, 2001). Endocannabinoids were found to be the neurotransmitters responsible for the depolarization-induced suppression of inhibition (DCI) and excitation (DCE). Since DCI enhances memory in the hippocampus, drugs that inhibit the metabolism and especially the transport of endocannabinoids are very likely to have a beneﬁcial eﬀect on

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

memory by increasing the levels of endocannabinoids at the sites where DCI takes place (Christie, 2001). Direct cannabinoid receptor agonists ﬂood the endocannabinoid system, resulting in the well-known overall disruptive eﬀect in memory and cognition. Cannabinoids are long known for their psychoactive and euphoric ‘‘high’’ eﬀects and have been used for these properties for centuries. Their addictive potential and mechanisms appear to be qualitatively and quantitatively diﬀerent from those of other drugs of abuse. However, recent studies indicate that cannabimimetics, similar to other addictive drugs, activate the brain reward/reinforcement circuit (ventral tegmental area, nucleus pallidus, and ventral pallidum) and produce reward-related behaviors in laboratory animals (Gardner, 1998). Eﬀorts to separate these unwanted eﬀects from the desired ones have had only limited success thus far. This fact, along with the negative social perception of these drugs, has been a major hindrance to the development of cannabinergic therapeutics. However, the increasing understanding of the endocannabinoid system presents us with possibilities for the design of selective agents. Indirect activation of this system by increasing endocannabinoid levels only at the sites where they are physiologically produced through inhibition of endocannabinoid catabolism or transport may lead to increased selectivity and fewer undesired eﬀects than activation of the cannabinoid receptors with direct agonists (Pertwee, 2000). Endocannabinoids such as anandamide were shown to have a much lower physical dependence potential (Aceto, 1998). Other well-known central cannabimimetic eﬀects that nevertheless are not well understood are hypothermia, appetite stimulation, and antiemetic eﬀects. Cannabimimetic-induced hypothermia is thought to occur by decreasing the thermoregulatory set point through interactions with the relevant hypothalamic centers (Pertwee, 1995b). Cannabimimetics also stimulate hunger in humans and animals, particularly for solid, sweet tasting foods (Pertwee, 1995b). For this property, delta-9-THC (marinol) is clinically used today for the management of AIDS-wasting syndrome (Nahas, 1999). The advent of potent and CB1-selective ligands lacking the CB2-mediated immunosuppressive properties may present signiﬁcant advantages over the currently used delta-9-THC in the treatment of AIDS patients who are already severely immunocompromised. It is also conceivable that cannabinoid receptor antagonists may be proven eﬀective as appetite suppressants, as suggested by the results of a study showing that SR141716A, a selective CB1 antagonist/inverse agonist, suppressed rodent appetite for sucrose and ethanol (Arnone, 1998).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

A second current clinical indication of cannabimimetics is their antiemetic and antinausea eﬀects, especially in cancer chemotherapy patients. These eﬀects are mediated above the level of vomiting reﬂex and possibly through descending inhibitory connections to the lower brain stem centers (Levitt, 1986).

B. Immune System The discovery of the peripheral CB2 receptor, which localizes in cells of the immune system, is very likely linked to the well-known immunosuppression of marijuana smokers. Miskin (1985) found that delta 9-THC decreases host resistance to herpes virus type 2 in mice and guinea pigs by decreasing both cellular and humoral immunity. In vivo and in vitro studies indicate that macrophages are the major targets of cannabinoids. delta 9-THC inhibits, in a dosedependent manner, the extrinsic antiviral activity of macrophages (Cabral, 1991). It was also shown that cannabinoids cause morphological changes in macrophages (Cabral, 1991) and aﬀect their phagocytic and spreading ability (Spector, 1991). The involvement of CB2 (and possibly of CB1) receptor(s) in the immunosuppressive eﬀects of cannabinoids is not proven yet. The localization of CB2 in cells of the immune system and especially in macrophages and lymphocytes suggests that this receptor serves some immunoregulatory role(s). The ﬁrst strong piece of evidence that implicates CB2 in such a function came from Kaminski et al. (1994), who demonstrated that cannabinoid-induced suppression of humoral immunity was partially mediated through inhibition of adenylyl cyclase by a G-protein-coupled mechanism that is pertussis toxin sensitive. Involvement of a membrane perturbation mechanism in cannabinoid-induced immunosuppression is also possible, especially in areas exposed to high drug concentrations, such as lung alveolar macrophages of marijuana smokers (Cabral, 1999). The involvement of the cannabinoid system in the regulation of the immune system may suggest that cannabinergics could potentially serve as immunomodulatory agents. Although CB2 selective agents already exist, their clinical potential in some immunomodulatory role will not be realized until the CB2 physiological functions are better understood. Cannabidiol, a cannabis terpenoid ingredient lacking the pyran ring as well as signiﬁcant binding aﬃnity for CB1 and CB2, was shown to be an active antiinﬂammatory agent in the murine model of arthritis (Pertwee, 2000). The molecular basis of this observation is still unknown.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

C. Cardiovascular System Cannabinoids reduce platelet aggregation and also produce tachycardia and orthostatic hypotension due to peripheral vasodilation. A distinct, CB1-like, receptor is found in the endothelium of rat mesenteric arteries (Jarai, 1999). This receptor mediates a remarkable vasodilating eﬀect after activation by any of several CCs, anandamide, or some CB1-inactive CClike analog. This eﬀect is NO independent and is inhibited by the CB1 antagonists, SR141716A and AM251 (Batkai, 2001), and also by cannabidiol. It is possible that exploitation of this new cannabinoid target may lead to new types of hypotensive agents.

D. Reproductive System Cannabinoids produce increased ring and chain chromosomal translocations and morphological abnormalities in mouse sperm, as well as reduction of sperm concentration in humans (Zimmerman, 1999). Strong evidence indicates the presence of functional CB1, or CB1-like receptors, in human sperm (Schuel, 1999). Furthermore, the endogenous cannabimimetic anandamide is produced in the human uterus and testes (Schuel, 1999). These ﬁndings along with several observations on cannabinoidinduced eﬀects on reproductive functions suggest that the cannabinoid system may be directly involved in the regulation of sperm production, sperm motility, the acrosome reaction, and prevention of polyspermy (Schuel, 1999). The endocannabinoid system in the uterus appears to play a fundamental role in embryo implantation and early development. Anandamide inhibits these processes and, therefore, regulation of its levels seems to control the timing of these events (Paria, 1995). These ﬁndings are also in line with recent clinical observations that correlate the levels of AEAase expression with miscarriages in pregnant women (Maccarone, 2000). Further understanding of the endocannabinoid functions in the reproductive system will open perspectives for exploitation of cannabinergics for the treatment of some types of infertility or the development of contraceptives. Cannabimimetics are also shown to aﬀect reproductive and metabolic functions indirectly by hormonal modulation through the hypothalamic and pituitary regulatory centers. They are found to reduce serum levels of the luteinizing hormone, prolactin, growth hormone, and thyroid-stimulating hormone, and to increase corticotropin (Murphy, 1998).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

E. Eye Cannabinoids reduce intraocular pressure, probably by directly aﬀecting ocular ﬂuid outﬂow pathways. The mechanism of this eﬀect is unknown, and its link to cannabinoid receptors has yet to be established (Green, 1999). Marijuana smoking is allegedly helpful to glaucoma patients, and the potential use of cannabimimetics for the treatment of glaucoma has long been recognized. New formulation technologies, as well as the advent of less hydrophobic cannabimimetics, present us with opportunities to overcome the challenge of local drug delivery to the eye.

F. Respiratory System Cannabimimetics are known to produce bronchodilation, which is manifested by a marked increase in airway conductance and reduction in airway resistance (Vachon, 1973). Although the mechanism of this activity is not known, it probably does not directly involve adrenergic receptors. Possible involvement of CB1A (a CB1 variant found in the lung) in cannabinoid-induced bronchodilation is still unexplored (Shire, 1995). Recently, it was shown that anandamide is released in the lung upon Ca2+ stimulation and exerts a dual eﬀect on bronchial response. It strongly inhibits capsaicin-evoked bronchospasm and cough; however, it causes bronchoconstriction in vagotomized rodents (Calignano, 2000). These eﬀects are mediated by CB1 receptors present in axon terminals of airway nerves since they are blocked by SR141716A. This endocannabinoidmediated control of airway responsiveness may be exploited in the development of new antiasthmatic agents.

G. Gastrointestinal System Cannabimimetics reduce the intestinal motility by a CB1-mediated inhibitory activity on acetylcholine release from autonomic ﬁbers. An endocannabinoid, 2-AG, was isolated from dog intestine; however, its role there remains unknown (Mechoulam, 1995a).

VII. CONCLUSIONS With the discovery of anandamide and 2-arachidonyl glycerol as two new families of endocannabinoids, cannabinoid research has taken major

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

strides toward arriving at an understanding of the molecular mechanism of cannabinoid action. Currently, there are multiple characterized endocannabinoid proteins [at least two receptors, CB1 and CB2; an enzyme, arachidonylethanolamide amidohydrolase (AEAase); and a transport protein, anandamide transporter (AT)] as potential therapeutic targets for the development of useful medications in the treatment of a multitude of conditions such as drug addiction, pain, and motor disorders. A number of ligands (receptor-selective agonists/antagonists, inverse agonists, enzyme inhibitors, transport inhibitors) are also available which can serve as important research tools for exploring the endocannabinoid biochemical pathways and their role in the modulation of behavior, memory, cognition, and pain perception. This is signiﬁcant progress, considering that only about a decade ago the sites of action of cannabinoids had not yet been identiﬁed and their molecular mechanism of action was still under question. The future of endocannabinoid research is undoubtedly very exciting and full of promise.

REFERENCES 1. Abadji V, Lin SY, Taha G, Griﬀin G, Stevenson LA, Pertwee RG, Makriyannis A. (R)-Methanandamide: a chiral novel anandamide possessing higher potency and metabolic stability. J Med Chem 1994; 37:1889–1893. 2. Aceto MD, Scates SM, Razdan RK, Martin BR. Anandamide, an endogenous cannabinoid has a very low physical dependence potential. J Pharmacol Exp Ther 1998; 287:598–605. 3. Arnone M, Maruani J, Chaperon F. Selective inhibition of sucrose nd ethanol intake by SR141716a, an antagonist of central cannabinoid (CB1) receptors. Psychopharmacology 1998; 132:104–106. 4. Baker D, Pryce G, Croxford JL. Cannabinoids control spasticity and tremor in a multiple sclerosis model. Nature 2000; 404:84–87. 5. Baker D, Pryce G, Croxford JL, Brown P, Pertwee RG, Makriyannis A, Khanolkar A, Layward L, Fezza F, Bisogno T, Di Marzo V. Endocannabinoids control spasticity in a multiple sclerosis model. FASEB J 2001; 15: 300–302. 6. Batkai S, Jarai Z, Wagner J, Goparaju S, Varga K, Liu J, Wang L, Mirshahi F, Khanolkar A, Makriyannis A, Urbaschek R, Garcia N Jr, Sanyal A, Kunos G. Endocannabinoids acting at vascular CB1 receptors mediate the vasodilated state in advanced liver cirrhosis. Nat Med 2001; 7:827–833. 7. Bell MR, D’Ambra TE, Kumar V, Eissentat MA, Herrmann JL, Wetzel JR, Rosi D, Philion RE, Daum SJ, Hlasta DJ, Kulling RK, Ackerman JH, Haubrich DR, Luttinger DA, Baixman ER, Miller MS, Ward SJ. Antinociceptive (aminoalkyl)indoles. J Med Chem 1991; 34:1099–1110.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

8. Beltramo M, Stella N, Calignano A, Lin S, Makriyannis A, Piomelli D. Functional role of high-aﬃnity anandamide transport, as revealed by selective inhibition. Science 1997; 277:1094–1097. 9. Boger DL, Sato H, Lerner AE, Hedrick MP, Fesik RA, Miyauchi H, Wilkie GD, Austin BJ, Patricelli MP, Cravatt BF. Exceptionally potent inhibitors of fatty acid amide hydrolase: the enzyme responsible for degradation of endogenous oleamide and anandamide. Proc Natl Acad Sci USA 2000; 97: 5044–5049. 10. Bouabula MBB, Rinaldi-Carmona M, Shire D, LeFur G, Casellas P. Stimulation of cannabinoid receptor CB1 induces krox-24 expression in human astrocytoma cells. J Biol Chem 1995b; 270:13973–13980. 11. Bouabula M, Poinot.-Chazel C, Bourrie B, Canat X, Rinaldi-Carmona M, LeFur G, Casellas P. Activation of mitogen-activated protein kinases by stimulation of the central cannabinoid receptor CB1. Biochem J 1995a; 312:637–641. 12. Bramblett RD, Pann AM, Ballesteros JA, Reggio PH. Construction of 3D model of the cannabinoid CB1 receptor: determination of helix ends and helix orientation. Life Sci 1995; 56:1971–1982. 13. Breivogel CS, Childers SR. The functional neuroanatomy of brain cannabinoid receptors. Neurobiol Dis 1998; 5:417–431. 14. Brotsie JM. Adjuncts to dopamine replacement, a pragmatic approach to reducing the problem of dyskinesia in Parkinson’s disease. Mot Disord 1998; 13:871–876. 15. Cabral GA. Marijuana and the immune system. In: Nahas GG et al., eds. Marijuana and Medicine. Totowa, NJ: Humana Press, 1999. 16. Cabral GA, Vasquez R. Marijuana decreases macrophage antiviral and antitumor activities. Adv Biosci 1991; 80:93–105. 17. Cadas H, di Tomaso E, Piomelli D. Occurrence and biosynthesis of endogenous cannabinoid precursor, N-arachidonyl phosphatidylethanolamine, in rat brain. J Neurosci 1997; 17:1226–1242. 18. Cadas H, Gaillet S, Beltramo M, Venance L, Piomelli D. Biosynthesis of an endogenous cannabinoid precursor in neurons and its control by calcium and cAMP. J Neurosci 1996; 16:3934–3942. 19. Calignano A, LaRana G, Giuﬀrida A, Piomelli D. Control of pain initiation by endogenous cannabinoids. Nature 1998; 394:277–281. 20. Calignano A, Katona I, Desarnaud F, Giuﬀrida A, LaRana G, Mackie K, Freund TF, Piomelli D. Bidirectional control of airway responsiveness by endogenous cannabinoids. Nature 2000; 408:96–101. 21. Childers SR, Sexton T, Roy MB. Eﬀects of anandamide on cannabinoid receptors in rat brain membranes. Biochem Pharmacol 1994; 47:711–715. 22. Chin CN, Lucas-Lenard J, Abadji V, Kendall DA. Ligand binding and modulation of cyclic AMP levels depend on the chemical nature of residue 192 of the human cannabinoid receptor 1. J Neurochem 1998; 70:366–373.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

23. Christie MJ, Vaughan CW. Cannabinoids act backwards. Nature 2001; 410: 527–530. 24. Consroe P. Brain cannabinoid systems as targets for the therapy of neurological disorders. Neurobiol Dis 1998; 5:534–551. 25. Crawley JN, Corwin RL, Robinson J, Felder CC, Derane WA, Axelrod J. Anandamide, an endogenous ligand of the cannabinoid receptor, induces hypomotility and Hypothermia in vivo in rodents. Pharmacol Biochem Behav 1993; 46:967–972. 26. D’Ambra K, Estep G, et al. Conformational restricted analogues of pravadoline: nanomolar potent, enanantioselective, (aminoalkyl)indole agonists of the cannabinoid receptor. J Med Chem 1992; 35:124–135. 27. Deadwyler S, Hampson RE, Childers SR. Functional igniﬁcance of cannabinoid receptors in brain. In: Pertwee RG, ed. Cannabinoid Receptors. New York: Academic Press, 1995: 206–227. 28. Desarnaud F, Cadas H, Piomelli D. Anandamide amidohydrolase activity in rat brain microsomes. J Biol Chem 1995; 270:6030–6035. 29. Deutsch D, Chin SA. Enzymatic synthesis and degradation of ananadmide, a cannabinoid receptor agonist. Biochem Pharmacol 1993; 46:791–796. 30. Deutsch DG, Lin S, Hill WAG, Morse KL, Salehani D, Arreaza G, Omeir RL, Makriyannis A. Fatty acid sulfonyl ﬂuorides inhibit anandamide metabolism and bind to the cannabinoid receptor. Biochem Biophys Res Commun 1997; 231:217–221. 31. Deutsch DG, Makriyannis A. Inhibitors of anandamide breakdown. NIDA Res Monog 1997; 173:65–84. 32. Devane W, Dysarz FA III, Johnson MR, Melvin LS, Howlett AC. Determination and characterization of a cannabinoid receptor in rat brain. Mol Pharmacol 1988; 34:605–613. 33. Devane WA, Hanus L, Breuer A, Pertwee RG, Stevenson LA, Griﬀin G, Gibson D, Mandelbaum A, Etinger A, Mechoulam R. Isolation and structure of a brain constituent that binds to the cannabinoid receptors. Science 1992; 258:1946–1949. 34. Devane WA, Axelrod J. Enzymatic synthesis of anandamide, an endogenous ligand for the cannabinoid receptor, by brain membranes. Proc Natl Acad Sci USA 1994; 91:6698–6701. 35. Di Marzo V. 2-Arachidonyl-glycerol as an endocannabinoid: limelight for a formerly neglected metabolite. Biochem (Moscow) 1998; 63:13–21. 36. Di Marzo V, Bisogno T, De Petrocellis L, Melck D, Orlando P, Wagner JA, Kunos G. Biosynthesis and inactivation of the endocannabinoid 2-arachidonoylglycerol in circulating and tumoral macrophages. Eur J Biochem 1999; 264:258–267. 37. Di Marzo V, Bisogno T, De Petrocellis L, Melck D, Martin BR. Cannabimimetic fatty acid derivatives: anandamide family and other ‘endocannabinoids.’ Curr Med Chem 1999b; 6:721–744.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

38. Drake DJ, Tius M, Jensen S, Busch-Petersen J, Kawakami JK, FernandezGarcia MC, Fan P, Makriyannis A. Classical/nonclassical hybrid cannabinoids: southern aliphatic chain-functionalized C-6h methyl, ethyl and propyl analogues. J Med Chem 1998; 41:3596–3608. 39. Edgemond WS, Hillard CJ, Falck JR, Kearn CS, Cambell WB. Human platelets and polymorphonuclear leukocytes synthesize oxygenated derivatives of arachidonylethanolamide (anandamide): their aﬃnities for cannabinoid receptors and pathways of inactivation. Mol Pharmacol 1998; 54: 180–188. 40. Facci L, Dal Toso R, Romanello S, Buriani A, Skaper SD, Leon A. Mast cells express a peripheral cannabinoid receptor with diﬀerential sensitivity to anandamide and palmitoylethanolamide. Proc Natl Acad Sci USA 1995; 92:3376–3380. 41. Felder CC, Joyce KE, Briley EM, Mansouri J, Mackie K, Blond O, Lai Y, Mitchell RL. Comparison of the pharmacology and signal transduction of the human cannabinoid CB1 and CB2 receptors. Mol Pharmacol 1995; 48: 443–449. 42. Fride E, Mechoulam R. Pharmacological activity of the cannabinoid receptor agonist, anandamide, a brain constituent. Eur J Pharmacol 1993; 231: 313–314. 43. Gardner EL, Vorel R. Cannabinoid transmission and reward-related events. Neurobiol Dis 1998; 5:501–533. 44. Gatley SJ, Lan R, Volkow ND, Papas N, King P, Wong CT, Giﬀord AN, Pyatt B, Dewey SL, Makriyannis A. Imaging of the brain marijuana receptor—Development of a radioligand that binds to cannabinoid CB1 receptors in vivo. J Neurochem 1998; 70:417–423. 45. Gerard, CM, Molerau, C, Vassart, G, Parmentier, M. Molecular cloning of a human cannabinoid receptor which is also expressed in testis. Biochem J 1991; 279:129–134. 46. Giuﬀrida A, Parsons LH, Kerr TM, Rodriguez de Fonseca F, Navarro M, Piomelli D. Dopamine activation of endogenous cannabinoid signaling in dorsal striatum. Nat Neurosci 1999; 2:358–363. 47. Glass M, Faull RLM, Dregnow M. Loss of cannabinoid receptors in the substantia nigra in Huntington’s disease. Neuroscience 1993; 56:523–527. 48. Goutopoulos A, Fan P, Khanolkar AD, Xie XQ, Lin SY, Makriyannis A. Stereochemical selectivity of methanandamides for the CB1 and CB2 cannabinoid receptors and their metabolic stability. Bioorg Med Chem 2001; 9:1673–1684. 49. Goparaju SK, Ueda N, Taniguchi K, Yamamoto S. Enzymes of porcine brain hydrolyzing 2-arachidonoylglycerol, an endogenous ligand of cannabinoid receptors. Biochem Pharmacol 2001; 57:417–423. 50. Green K. Marijuana and intraocular pressure. In: Nahas GE et al, eds. Marijuana and Medicine. Totowa, NJ: Humana Press, 1999: 581–589.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

51. Hampson, AJ, Hill WA, Zan Phillips M, Makriyannis A, Leung E, Eglen, RM, Bornheim LM. Anandamide hydroxylation by brain lipoxygenase: metabolite structures and potencies at the cannabinoid receptor. Biophys. Biochim. Acta 1995; 1259:173–179. 52. Hampson RE, Deadwyler SA. Role of cannabinoid receptors in memory storage. Neurobiol Dis 1998; 5:474–482. 53. Hampson RE, Deadwyler SA. Cannabinoids, hippocampal function and memory. Life Sci 1999; 65:715–723. 54. Hanus L, Gopher A, Almog S, Mechoulam R. Two new unsaturated fatty acid ethanolamides in brain that bind to the cannabinoid receptor. J Med Chem 1993; 36:3032–3034. 55. Hanus L, Breuer A, Tchilibon S, Shiloah S, Goldenberg D, Horowitz M, Pertwee RG, Ross RA, Mechoulam R, Fride, E. HU-308: a speciﬁc agonist for CB2, a peripheral cannabinoid receptor. Proc Natl Acad Sci USA, 1999; 25:14228–14233. 56. Herkenham M, Lynn AB, Little MD, Johnson MR, Melvin LS, DeCosta BR, Rice KC. Cannabinoid receptor localization in brain. Proc Natl Acad Sci 1990; 87:1932–1936. 57. Hillard CJ, Cambell WB. Biochemistry and pharmacology of arachidonylethanolamide, a putative endogenous cannabinoid. J Lip Res 1997; 38: 2383–2398. 58. Houston D, Howlett AC. Diﬀerential receptor–G-protein coupling evoked by dissimilar cannabinoid receptor agonists. Cell Signal 1998; 10:667–674. 59. Howlett AC. The CB1 cannabinoid receptor in the brain. Neurobiol Dis 1998a; 5:405–416. 60. Howlett, AC, Mukhopadhyay S, Shim J-Y, Welsh WJ. Signal transduction of eicosanoid CB1 receptor ligands. Life Sci 1999; 65:617–625. 61. Howlett AC, Song C, Berglund BA, Wilken GH, Pigg JJ. Characterization of CB1 cannabinoid receptors using receptor peptide fragments and sitedirected mutagenesis. Mol Pharmacol 1998b; 53:504–510. 62. Huﬀman JW, Dai D, Martin BR, Compton DR. Design, synthesis and pharmacology of cannabimimetic indoles. Bioorg Med Chem Lett 1994; 4: 563–566. 63. Huﬀman JW, Yu S, Showalter V, Abood ME, Wiley JL, Compton DR, Martin BR, Bramblett RD, Reggio PH. Synthesis and pharmacology of a very potent cannabinoid lacking a phenolic hydroxyl with high aﬀinity for the CB2 receptor. J Med Chem 1996; 39:3875–3877. 64. Jarai Z, Wagner J, Varga K, Lake KD, Compton, DR, Martin BR, Zimmer AM, Bonner TI, Buckley NE, Mezey E, Rajdan RK, Zimmer A, Kunos G. Cannabinoid-induced mesenteric vasodilation through an endothelial site distinct form CB1 or CB2 receptors. Proc Natl Acad Sci USA 1999; 96:14136–14141. 65. Jarbe TV, Lamb RJ, Makriyannis A, Lin S, Goutopoulos A. Delta 9-THC

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

66.

67.

68.

69. 70.

71.

72.

73.

74. 75.

76. 77.

78.

training dose as a determinant for (R)-methanandamide generalization in rats. Psychopharmacology 1998; 140:519–522. Johnson MR, Melvin LS, The discovery of nonclassical cannabinoid analgesics. In: Mechoulam C, ed. Cannabinoids as Therapeutic Agents. Boca Raton, FL: CRC Press, 1986: 121–145. Kaminski NE, Koh WS, Yang KH, Lee M, Kessler FK. Suppression of the humoral immune response by cannabinoids is partially mediated through inhibition of adenylate cyclase by a pertussis toxin-sensitive G-proteincoupled mechanism. Biochem Pharmacol 1994; 48:1899–1908. Khanolkar AD, A., V.; Lin SY, Hill WAG, Taha G, Abouzid K, Meng Z, Fan P, Makriyannis A. Head group analogs of arachidonylethanolamide, the endogenous cannabinoid ligand. J Med Chem 1996; 39:4515–4519. Khanolkar AD, Palmer SL, Makriyannis A. Molecular probes for the cannabinoid receptors. Chem Phys Lipids 2000; 108:37–52. Kreitzer AC, Regehr WG. Retrograde inhibition of presynaptic calcium inﬂux by endogenous cannabinoids at excitatory synapses onto Purkinje cells. Neuron 2001; 29:717–727. Lang W, Qin C, Lin C, Khanolkar AD, Goutopoulos A, Fan P, Abouzid K, Meng Z, Biegel D, Makriyannis A. Substrate speciﬁcity and stereoselectivity of rat brain microsomal anandamide amidohydrolase. J Med Chem 1999; 42:896–902. Levitt M. Cannabinoids as antiemetics in cancer chemotherapy. In: Mechoulam R, ed. Cannabinoids as Therapeutic Agents. Boca Raton, FL: CRC Press, 1986: 73–79. Little PJ, Compton DR, Johnson MR, Melvin LS, Martin BR. Pharmacology and stereoselectivity of structurally novel cannabinoids in mice. J Pharmacol Exp Ther 1988; 247:1046–1051. Loh HH, Law PY. The role of membrane lipids in receptor mechanisms. Annu Rev Pharmacol Toxicol 1980; 20:201–234. Maccarone M, Valenise H, Bari M, Lazzarin N, Romani C, Finazzi-Agro A. Relationship between decreased anandamide hydrolase concentrations in human lymphocytes and miscarriage. Lancet 2000; 235:1326–1329. Mackie K, Hille B. Cannabinoids inhibit N-type calcium channels in neuroblastoma–glioma cells. Proc Natl Acad Sci 1992; 89:3825–3829. Mackie K, Devane WA, Hille B. Anandamide, an endogenous cannabinoid, inhibits calcium currents as a partial agonist in N-18 neuroblastoma cells. Mol Pharmacol 1993; 44:498–503. Makriyannis A, Banijamali A, Van der Schyf C, Jarrell H. The role of cannabinoid stereochemistry and absolute conﬁguration and the orientation of D9-THC in the membrane bilayer. Structure–activity relationships of cannabinoids. In: Rapaka RS, Makriyannis A, eds. Interactions of Cannabinoids with Membranes. National Institute on Drug Abuse Research Monograph 79. Rockville, MD: US Department of Health and Human Services, 1987.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

79. Makriyannis A, Rapaka RS. The molecular basis of cannabinoid activity. Life Sci 1990; 24:2173–2184. 80. Makriyannis A. In: Pertwee RG, ed. Cannabinoid Receptors. The Role of Cell Membranes in Cannabinoid Activity. New York: Academic Press, 1995: 87–115. 81. Malan TP Jr, Ibrahim MM, Deng H, Liu Q, Mata HP, Vanderah T, Porreca F, Makriyannis A. CB2 cannabinoid receptor-mediated peripheral analgesia. Pain In press. 82. Martel P, Makriyannis A, Mavromoustakos T, Kelly K, Jeﬀrey KR. Topography of tetrahydrocannabinol in model membranes using neutron diffraction. Biochim Biophys Acta 1993; 1151:51–58. 83. Martin BR, Compton DR, Thomas BF, Prescott WR, Little PJ, Razdan RK, Johnson MR, Melvin LS, Mechoulam R, Ward SJ. Behavioral, biochemical and molecular modeling evaluations of cannabinoid analogs in mice. Pharmacol Biochem Behav 1991; 40:471–478. 84. Martin B, Lichtman AH. Cannabinoid transmission and pain perception. Neurobiol Dis 1998; 5:447–461. 85. Martin BR, Mechoulam R, Razdan RK. Discovery and characterization of endogenous cannabinoids. Life Sci 1999; 65:573–595. 86. Matsuda LA, Lolait SJ, Brownstein, MJ, Young AC, Bonner TI. Structure of a cannabinoid receptor and functional expression of the cloned cDNA. Nature 1990; 346:561–564. 87. Mechoulam R, Ben-Shabat S, Hanus L, Ligumsky M, Kaminski NE, Schatz AR, Gopher A, Almog S, Martin B, Compton DR, Pertwee RG, Griﬀin G, Bayewitch M, Barg J, Vogel Z. Identiﬁcation of an endogenous 2-monoglyceride, present in canine gut, that binds to cannabinoid receptors. Biochem Pharmacol 1995a; 50:83–90. 88. Mechoulam R, Devane WA, Glaser R. Cannabinoid geometry and biological activity. In: Nahas GG et al., eds. Marijuana and Medicine. Totowa, NJ: Humana Press, 1999: 65–90. 89. Mechoulam R, Feigenbaum E. The unpaved road to the endogenous brain cannabinoid ligands, the anandamides. In: Pertwee RG, ed. Cannabinoid Receptors. New York: Academic Press, 1995b: 233–258. 90. Mechoulam R, Gaoni Y. Isolation, structure and partial synthesis of an active constituent of hashish. J Am Chem Soc 1967; 86:1646–1647. 91. Miskin EM, Cabral GA. Delta-9-tetrahydrocannabinol decreases host resistance to herpes simplex virus type 2 vaginal infection in B6C3F1 mice. J Gen Virol 1985; 66:2539–2549. 92. Munro S, Thomas KL, Abu-Shaar M. Molecular characterization of a peripheral receptor for cannabinoids. Nature 1993; 365:61–65. 93. Murphy LL, Munoz RM, Adrian BA, Villnua MA. Function of cannabinoid receptors in the neuroendocrine regulation of hormone secretion. Neurobiol Dis 1998; 5:432–446. 94. Nagayama T, Sinor AD, Simon RP, Chen J, Graham SH, Jin K, Green-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

95. 96.

97.

98.

99.

100.

101. 102.

103.

104. 105.

106.

107.

108.

berg DA. Cannabinoids and neuroprotection in global and focal cerebral ischemia. J Neurosci 1999; 18:2987–2995. Nahas GG, Sutin KM, Harvey DJ, Agurell S, eds. In: Nahas GG, et al., eds. Marijuana and Medicine. Totowa, NJ: Humana Press, 1999. Nye JS, Voglmaier S, Martenson RE, Snyder SH. Myelin basic protein is an endogenous inhibitor of the high aﬀinity cannabinoid binding site in brain. J Neurochem 1988; 50:1170–1178. Ohno-Shosaku T, Maejima T, Kano M. Endogenous cannabinoids mediate retrograde signals from depolarized postsynaptic neurons to presynaptic terminals. Neuron 2001; 29:729–738. Papahatjis DP, Kourouli T, Abadji V, Goutopoulos A, Makriyannis A. Pharmacophoric requirements for cannabinoid side chains, multiple bond and C1V-substituted delta-8-tetrahydrocannabinols. J Med Chem 1998; 41: 1195–1200. Paria BC, Das SK, Dey SK. The preimplantation embryo is a target for cannabinoid ligand–receptor signaling. Proc Natl Acad Sci USA 1996; 92: 9430–9464. Pate E, Jarvinsen K, Utri P, Jarho P, Jarvinen, T. Ophthalmic arachidonylethanolamide decreases intraocular pressure in normotensive rabbits. Curr Eye Res 1995; 14:791–797. Paton WDM. Pharmacology of marijuana. Annu Rev Pharmacol 1975; 15: 191–220. Pertwee RGS, L. A.; Elrick DB, Mechoulam R, Corbett AD. Inhibitory eﬀects of certain enantiomeric cannabinoids in the mouse vas deferens and the myenteric plexus preparation of guinea-pig small intestine. Br J Pharmacol 1992; 105:980–984. Pertwee RG. Pharmacological, physiological and clinical implications of the discovery of cannabinoid receptors. In: Pertwee RG, ed. Cannabinoid Receptors. New York: Academic Press, 1995a; 69–83. Pertwee RG. Pharmacology of cannabinoid CB1 and CB2 receptors. Pharmacol Ther 1997; 74:129–180. Pertwee RG. Cannabinoid receptor ligands: clinical and neuropharmacological considerations, relevant to future drug discovery and development. Exp Opin Invest Drugs 2000; 9:1–19. Pertwee RG, Fernardo SR, Griﬀin G, Abadji V, Makriyannis A. Eﬀect of phenylmethylsulphonyl ﬂuoride on the potency of anandamide as an inhibitor of electrically evoked contractions in two isolated tissue preparations. Eur J Pharmacol 1995b; 272:73–78. Pertwee RG, Griﬀin G, Fernando SR, Li X, Hill A, Makriyannis A. AM630, a competitive cannabinoid receptor antagonist. Life Sci 1995c; 56: 1949–1955. Pertwee RG, Griﬀin G, Hanus L, Mechoulam R. Eﬀects of two endogenous fatty acid ethanolamides on mouse vasa deferentia. Eur J Pharmacol 1994; 259:115–120.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

109. Peters H, Nahas GG. A brief history of four millenia (B.C. 2000–A.D. 1974). In: Nahas GG et al., eds. Marijuana and Medicine. Totowa, NJ: Humana Press, 1999: 3–7. 110. Piomelli D, Beltramo M, Glasnapp S, Lin SY, Goutopoulos A, Xie XQ, Makriyannis A. Structural determinants for recognition and translocation by the anandamide transporter. Proc Natl Acad Sci USA 1999; 10:5802– 5807. 111. Piomelli D, Beltramo M, Giuﬀrida A, Stella N. Endogenous cannabinoid signaling. Neurobiol Dis 1998; 5:462–473. 112. Razdan RK. Structure–activity relatioships in cannabinoids. Pharmacol Rev 1986; 38:75–134. 113. Rinaldi-Carmona M, Bath F, Heaulme M, Shire D, Calandra B, Congy C, Martinez S, Maruani J, Neliat G, Caput D, Ferrara P, Soubrie P, Breliere JC, Lefur G. SR14176A, a potent and selective antagonist of the brain cannabinoid receptor. FEBS Lett 1994; 350:240–244. 114. Sanudo-Pena MC, Tsou K, Walker M. Motor actions of cannabinoids in the basal ganglia output nuclei. Life Sci 1999; 65:703–713. 115. Schmid P, Paria BC, Krebsbach R, Schmid H, Dey SK. Changes in anandamide levels in mouse uterus are associated with uterine receptivity for embryo implantation. Proc Natl Acad Sci USA 1997; 94:4188–4192. 116. Schuel H, Goldstein E, Mechoulam R, Zimmerman AM, Zimmerman S. Anandamide (arachidonylethanlamide), a brain cannabinoid receptor agonist, reduces sperm fertilizing capacity in sea urchins by inhibiting the acrosome reaction. Proc Natl Acad Sci USA, 1994; 91:9460–9464. 117. Schuel H, Chang MC, Burkman LJ, Picone R, Makriyannis A, Zimmerman AM, Zimmerman S. Cannabinoid receptors in sperm. Marijuana and Medicine. In: Nahas GG et al., eds. Totowa, NJ: Humana Press, 1999: 317–325. 118. Shire DC, Carillon C, Kaghad M, Calandra B, Rinaldi-Carmona M, Le Fur G, Caput D, Ferrara P. An amino-terminal variant of the central cannabinoid receptor resulting from alternative splicing. J Biol Chem 1995; 270: 3726–3731. 119. Smith PB, Compton DR, Welch SP, Razdan RK, Mechoulam R, Martin BR, The pharmacological activity of anandamide, a putative endogenous cannabinoid, in mice. J Pharmacol Exp Ther 1994; 270:219–227. 120. Spector S, Lancz G. Suppression of human macrophage function in vitro by D9-tetrahydrocannabinol. J Leukocyte Biol 1991; 50:423–426. 121. Stella N, Schweitzer P, Piomelli D. A second endogenous cannabinoid that modulates long-term potentiation. Nature 1997; 388:773–778. 122. Sugiura T, Kotaka T, Nakane S, Miyashita T, Kondo S, Suhara Y, Takayama H, Waku K, Seki C, Baba N, Ishima Y. Evidence that the cannabinoid CB1 receptor is a 2-arachidonylglycerol receptor. J Biol Chem 1999; 274: 2794–2801. 123. Tao Q, Abood ME. Mutation of a highly conserved aspartate residue in

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

124.

125.

126.

127.

128. 129.

130. 131. 132.

133.

134.

the second transmembrane domain of the cannabinoid receptors, CB1 and CB2, disrupts G-protein coupling. J Pharmacol Exp Ther 1998; 285:651– 658. Ueda N, Kurahashi Y, Yamamoto S, Tokunaga T. Partial puriﬁcation and characterization of the porcine brain enzyme hydrolyzing and synthesizing anandamide. J Biol Chem 1995; 270:23823–23827. Vachon L, Fitzerald MX, Solliday NH, Gould IA, Gaensler EA. Singledose eﬀect of marijuana smoke: bronchial dynamics and respiratory-center sensitivity in normal subjects. N Engl J Med 1973; 288:985–989. Varga KLK, Martin BR, Kunos G. Novel antagonist implicates the CB1 receptor in the hypotensive action of anandamide. Eur J Pharmacol 1995; 278:279–283. Vogel Z, Barg R, Levy D, Saya E Heldman, Mechoulam R. Anandamide, a brain endogenous compound, interacts speciﬁcally with cannabinoid receptors and inhibits adenylate cyclase. J Neurochem 1993; 61:352–355. Wagner JA, Varga K, Jarai Z, Kunos G. Mesenteric vasodilation mediated by endothelial anandamide receptors. Hypertension 1999; 33:429–34. Waksman Y, Olson SM, Carlisle SJ, Cabral, GA. The central cannabinoid receptor (CB1) mediates inhibition of nitric oxide production by rat microglial cells. J Pharmacol Exp Ther 1999; 288:1357–1366. Walker JM, Hohmann G, Martin WJ, Strangman NM. The neurobiology of cannabinoid analgesia. Life Sci 1999; 65:665–673. Wilson RI, Nicoll RA. Endogenous cannabinoids mediate retrograde signaling at hippocampal synapses. Nature 2001; 410:588–592. Xie X-Q, Eissentat M, Makriyannis A. Common cannabimimetic pharmacophoric requirements between aminoalkylindoles and classical cannabinoids. Life Sci 1995; 56:1963–1970. Yamada K, Rice KC, Flippen-Anderson JL, Eissenstat MA, Ward SJ, Johnson MR, Howlett AC. (Aminoalkyl)indole isothiocyanates as potential electrophylic aﬀinity ligands for the brain cannabinoid receptor. J Med Chem 1996; 39:1967–1974. Zimmerman MA, Zimmerman S, Raj AY, Eﬀects of cannabinoids on spermatogenesis in mice. In: Nahas GG et al, eds. Marijuana and Medicine. Totowa, NJ: Humana Press, 1999: 347–358.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

5 Development of PET and SPECT Radioligands for Cannabinoid Receptors S. John Gatley, Andrew N. Gifford, and Yu-Shin Ding Brookhaven National Laboratory, Upton, New York, U.S.A.

Nora D. Volkow NIDA, Bethesda, Maryland, U.S.A.

Ruoxi Lan, Qian Liu, and Alexandros Makriyannis University of Connecticut, Storrs, Connecticut, U.S.A.

I.

INTRODUCTION

Marijuana is the most commonly illegal drug of abuse in the United States, but relatively little known about how activation of cannabinoid receptors leads to the psychoactive eﬀects desired by its abusers, or whether receptor densities are altered in addiction or detoxiﬁcation, or in other disease states. Futhermore, the two known G-protein-coupled receptors for cannabinoids (CB1, which is found in brain and some peripheral tissues, and CB2, whose distribution is believed to be restricted to immunological tissues in the periphery) are potentially important targets for drug development. Various dosage forms of D9-tetrahydrocannabinol (THC), the major psychoactive constituent of marijuana, as well as novel compounds, are under investigations by several major drug companies. For example, recently developed subtype-selective antagonists of cannabinoid receptors such as SR141716A may have beneﬁcial drug

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

actions in treating memory impairment and other disorders. A publication by the Institute of Medicine/National Academy of Sciences entitled ‘‘Marijuana and Medicine: Assessing the Science Base’’ discusses some of the issues involved in this area. Positron emission tomography (PET) is an imaging method used to measure the regional distribution and kinetics of chemical compounds labeled with short-lived positron-emitting isotopes such as 11C (half-life = 20 min) and 18F (110 min). It thus enables direct measurement of components of the neurochemical systems in the living human brain [1,2]. The SPECT (single-photon emission-computed tomography) methodology can also be used to measure some of the same components of neurochemical systems as PET. As discussed later, SPECT is inferior to PET in terms of spatial resolution, sensitivity, and quantitation. On the other hand, SPECT methodology is less expensive than PET and is far more widely available because of its advantages in clinical nuclear medicine. These include the use of radionuclides of longer half-life, such as 123I (13 h) and 99mTc (6 h). Both PET and SPECT have been used in studies of several drugs to image functional consequences of acute or chronic drug treatment, using radiotracers that measure changes in factors such as blood ﬂow, glucose metabolic rate, or dopamine release. They have also been used to image changes in the apparent brain concentrations of neuroreceptors with which the drugs of abuse directly or indirectly interact. For a review of this area, see Gatley and Volkow [3]. The brain – dopamine system, which may be involved in the actions of all drugs of abuse, has been an important target of PET studies [4].

II. IMAGING THE EFFECTS OF CANNABINOIDS ON METABOLISM AND BLOOD FLOW Relatively few human imaging studies have evaluated the eﬀects of marijuana or THC on metabolism or blood ﬂow. Acute intravenous THC in both normal controls and habitual marijuana users led to increased an increased regional cerebral metabolic rate (CMR) in the cerebellum. This increase is positively correlated both with concentrations of THC in the plasma and with the intensity of the subjective sense of intoxication [5]. In a 1997 PET/[15O]water study with 32 abusers [6], THC dose-dependently increased cerebral blood ﬂow (CBF) in the frontal regions, insula

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

and cingulate gyrus, and subcortical structures, with somewhat greater eﬀects in the right hemisphere. Self-ratings of THC intoxication were correlated most markedly with the right frontal region. Most subjects exhibited increased rCBF in cerebellum. However, those whose cerebellar CBF decreased also had a signiﬁcant alteration in time sense [7]. The average increase in rCMR after THC administration was less in marijuana users than in controls, and users had lower cerebellar metabolism than the controls at baseline [8]. Thus the cerebellum shows the greatest metabolic increase in response to acute THC and responds to chronic marijuana exposure with a decrease in baseline CMR. Habitual users but not controls responded to THC administration with increased rCMR in prefrontal cortex, orbitofrontal cortex, and basal ganglia. In contrast to the robust eﬀects of THC on relative rCMR, changes in global CMR in response to THC were quite variable, with increases, decreases, and no changes seen in equal numbers of subjects. There was also variability in subjective eﬀects, which were pleasurable for most subjects but either minimal or unpleasant (anxiety or paranoia) for others. The involvement of the cerebellum in the psychoactive eﬀects of marijuana and in changes in rCMR is consistent with the view that THC interacts with the high concentration of CB1 receptors in this brain area. Decreases in the cerebellar rCMR in habitual marijuana users may reﬂect the eﬀects of chronic exposure to the drug. Functions known to be associated with the cerebellum, such as motor coordination, proprioception, and learning, are adversely aﬀected both during acute marijuana intoxication and in habitual users. The cannabinoid CB1 receptor is the binding site in the brain for D9-tetrahydrocannabinol, the active principle of marijuana. This Gprotein-coupled receptor is abundant in speciﬁc brain areas including the cerebellum, the hippocampus, and the outﬂow nuclei of the basal ganglia. An in vivo radioligand for the CB1 receptor would allow us to evaluate disease and drug-induced changes in cannabinoid receptor densities, and possibly to investigate relationships between receptor occupancy by agonist and antagonist ligands and their behavioral and toxic eﬀects. Such studies would contribute not only to our understanding of the neural basis of marijuana abuse, but also to medication development, since pharmacological manipulation of the CB1 receptor system might prove useful in conditions such as chronic pain and multiple sclerosis. In addition to compounds with the classical cannabinoid skeleton, recently developed high aﬃnity agonists and antagonist of the brain cannabinoid receptor may serve as lead compounds for PET

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 1 Structures of representatives of classes of compounds that blind to cannabinoid receptors.

and SPECT radiotracers. These have nonclassical cannabinoid, aminoalkylindole, anandamide, and pyrazole strructures (Fig. 1) [9 – 13]. Autoradiographic studies with tritiated CP55,940 and other high aﬃnity agonists (see, e.g., Ref. 14) demonstrated high concentration of cannabinoid receptors in the basal ganglia and especially in its outﬂow nuclei, the globus pallidus and the substantia nigra. High concentrations are also found in the hippocampus and the cerebellum. The cerebral cortex also contains appreciable concentrations of cannabinoid receptors, the highest being the cingulate gyrus. Some other regions including most of the brain stem and the thalamus, contain low or negligible concentrations. The pattern of distribution of cannabinoid receptors in many brain regions is similar to that of dopamine D1 receptors, which has led to the suggestion that a function of the cannabinoid system may be to induced modulate brain dopaminergic activity [15].

III. ATTEMPTS TO DEVELOP RADIOLIGANDS The ﬁrst attempt to develop a PET radioligand for imaging brain CB1 receptors involved modiﬁcation of D8-THC by labeling with ﬂuorine-18 in the hydrocarbon side chain [16,17], as shown in Figure 2. Unfortunately,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

this compound did not produce PET images that showed any particular regional pattern of brain localization when injected into a baboon. It showed poor uptake but was widely distributed in the brain. Clearance of radioactivity that did appear to enter the brain was rapid. Furthermore, uptake of radioactivity in the skull was apparent, which suggested in vivo decomposition of the radiotracer, leading to the production of labeled ﬂuoride ion, which then accumulated in bone. It is likely therefore that the PET images represented only nonspeciﬁc uptake of the tracer with a negligible component due to speciﬁc binding to cannabinoid receptors. Studies with [18F] D8-THC supported the view that a successful radiotracer must have adequate metabolic stability and a fairly high aﬃnity for the CB1 receptor to ensure that radioactivity is retained in brain tissues long enough for tomographic measurement. Furthermore, a good radiotracer should exhibit high uptake into the brain. This is likely to be a diﬃculty for cannabinoid receptor radioligands, since these molecules are extremely lipophilic. High log P values are generally associated with poor blood – brain barrier penetrability, presumably because they remain dissolved in lipid structures in the blood during transit through the brain capillary bed. The nonclassical cannabinoid CP55,940, the aminoalkylindole WIN55, 212-2, and THC are reported to possess log P values of about 6, 5, and 7, respectively. Even the lowest log P value in this series (5) has been associated with poor brain penetration in other classes of molecules (see, e.g., Refs. 18, 19). Furthermore, although [3H]WIN55,212-2 has an aﬃnity about 10-fold higher than THC, it does not exhibit preferential localization in CB1receptor-rich areas of the mouse brain when injected intravenously (Giﬀord et al., unpublished).

Figure 2 Incorporation of fluorine-18 into 5V-[18F]fluoro-D8-THC.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 3 Relationships between structures of SR141716A, AM251, and AM281.

The introduction of the diarylpyrazole CB1 receptor antagonist SR141716A [12] immediately suggested that exploitation of this class of molecules might lead to development of a successful CB1 receptor radioligand. Not only is the aﬃnity of SR141716A of the order of 100-fold higher than that of THC, but the structure suggested that the log P value would be considerably lower. In addition, SR141716A is highly selective for the brain cannabinoid receptor (CB1) relative to the CB2 receptor found in cells of the immune system. It is also an antagonist, which is a potential advantage because, in the binding of antagonists of G-proteincoupled receptors, there is no discrimination between receptors in diﬀerent aﬃnity states. This is unlike the situation with agonists, which bind predominantly to a high-aﬃnity state of the receptor. Finally, SR141716A contains three chlorine atoms, suggesting that replacement of one of these with a radioactive iodine atom might produce a compound with the desired properties. Our ‘‘mark I’’ pyrazole radioligand, code-named AM251, was synthesized in nonradioactive and radioactive forms [20,21]. Following intravenous injection in mice and rats, the radioiodinated compound concentrated preferentially in brain areas known to contain densities of CB1 receptors [22]; however, it failed to enter the brain in SPECT experiments conducted with baboons [23]. On the hypothesis that this failure was associated with too high a log P value, we synthesized a related, ‘‘mark II’’ radioligand with an additional structural modiﬁcation. This was replacement of the piperidine ring of SR141716A and AM251 with the more polar morpholino ring (Fig. 3). This compound AM281, was able to visualize CB1 receptors in baboon (Fig. 4) and rodent (Fig. 5) brains in vivo [23].

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 4 Sagittal sections of the brains of two baboons injected intravenously with [123I]AM251 (left) or [123I]AM281 (right). These experiments indicated that AM281 is to penetrate the baboon brain much more readily than AM251.

Figure 5 Ex vivo autoradiography of [123I]AM281 in rat brain gave distribution patterns that were essentially identical to in vitro autoradiographs obtained using tritiated high affinity agonists.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 6 Comparison of the behavior of radioidinated AM251 and AM281 in mice.

The degree of failure of AM251 to enter the baboon brain, as assessed by SPECT scanning, was surprising because in rats and mice AM251 does clearly enter the brain. Figure 6 presents a comparison of AM251 and AM281 in mice. In absolute terms, the graphed data should be interpreted cautiously because the experiments with the two radioligands were not conducted simultaneously, and other experimental details were not identical [22,23]. However, apparently greater brain uptake of AM281 at early times is consistent with its smaller log P value. Moreover, it is clear that there was signiﬁcant clearance of AM281 between 30 and 120 min, whereas there was no signiﬁcant diﬀerence between the 30, 60, and 120 min data points for AM251. The in vivo brain uptake data, which show a more prolonged retention of AM251, were thus consistent with the in vitro binding data, which indicated that AM251 has an approximately threefold higher aﬃnity for the CB1 receptor than AM281. The mouse data shown in Figure 6, however, did not predict the large diﬀerence seen in baboons (Fig. 4). It may be that tight binding of AM251 to a speciﬁc blood protein, rather than a greater distribution of AM251 into lipophilic blood components, is responsible for its low brain penetration in baboons.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Although our SPECT studies with [123I]AM281 provided a proof of principle that in vivo imaging of CB1 receptors in primates is possible, the behavior of this radioligand is far from ideal in that, relative to radiotracers used to study other neurochemical systems, its uptake in the brain is low and its clearance is rapid. These factors limit the count rate obtained in radionuclide imaging studies, and thus the quality of the images. Because PET is more sensitive than SPECT by an order of magnitude, a CB1 receptor radioligand labeled with ﬂuorine-18 with pharmacokinetic properties to those of similar AM281 would probably be quite acceptable for human use. On the other hand, a radioiodinated compound for use with SPECT would be more useful if it had at least the initial brain uptake of AM281, but a higher aﬃnity, to ensure a longer clearance time. Since our initial published work with AM281, several other candidate radioligands have been prepared [24 – 26]. However, to our knowledge no reports of human studies have appeared.

IV. ONGOING WORK In our own laboratories, we have continued to synthesize and evaluate new labeled cannabinoid receptor radioligands. One of these is [18F]AM284, where the labeled atom is part of a ﬂuoropentyl group on position 1 of the pyrazole ring (Fig. 7; see also Table 1). Although this (unpublished) study demonstrated in vivo binding of AM284 to CB1 receptors, as shown by the fact that co-injection with SR141716A reduced brain binding, the

Figure 7 Structure of an 18F-labeled pyrazole ligand; for brain uptake (see Table 1).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 1 Brain Uptake Data for the

18

F-labeled Pyrazole Ligand of Figure 7 Injected activity in whole brain (%)a

Conditions 15 min 60 min (vehicle) 60 min (+ SR14176A)

AM281

AM284

0.80 F 0.05 0.66 F 0.05* 0.33 F 0.05***

0.07 F 0.008 0.028 F 0.004** 0.020 F 0.001****

*

p < 0.003; cf. 15 min time point. p < 0.001; cf. 15 min time point. *** p < 0.001; cf. vehicle. **** p < 0.035; cf. vehicle. a Values are the mean FSD (n = 5). **

binding was less than one-tenth that of AM281 measured simultaneously in a dual-isotope experiment. It would therefore not be a practicable PET radioligand. Other ﬂuorine-18 and radioiodinated cannabinoid receptor ligands are being synthesized and studied in our laboratories. While we have not yet started human PET or SPECT studies, we have used AM281 to conduct fundamental studies of the CB1 receptor system in

Figure 8 Comparison of ability of WIN 55, 212-2 to sedate mice (triangles) and to block specific binding of [131I]AM281 (squares).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

rodents. For example, during experiments in mice designed to measure the occupancy of CB1 receptors associated with physiological eﬀects of exogenous cannabinoids, we found that doses of WIN 55,212-2 that induced profound sedation did not reduce binding of AM281 to cerebellum and hippocampus (Fig. 8). This observation indicates that the occupancy of the CB1 receptor necessary for physiological eﬀects of cannabinoids is very low [27]. Experiments in superfused hippocampal slices prepared from rats were then conducted to compare inhibition of acetylcholine release by the cannabinoid receptor agonist WIN55,212-2 with inhibition of AM281 binding by this agonist. The results (Fig. 9) show that half-maximal response is achieved at less than 1% occupancy, conﬁrming that the agonist occupancy necessary to produce a physiological response is very low in the CB1 system [27]. A consequence of a very large receptor reserve for CB1 receptors would be that PET or SPECT could not be used to image the occupancy of the CB1 receptor by biologically signiﬁcant doses of agonist drugs. This is

Figure 9 Graph of inhibition of acetylcholine release in superfused hippocampal slices versus CB1 receptor occupancy estimated from inhibition of [131I]AM281 binding.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

because even high doses of an agonist would not displace an appreciable fraction of the radiotracer binding. On the other hand, these experiments indicated that the binding of exogenous CB1 receptor antagonist radioligands should not be aﬀected by changes in the levels of endogenous ligands such as anandamide. This would thus remove a possible confounding factor in imaging experiments designed to detect changes in cannabinoid receptor densities. In the light of these experiments with WIN55,212-2, it is of interest to speculate on the degree of occupancy of brain CB1 receptors that is achieved by doses of THC that induce desired eﬀects in human during the smoking of a marijuana cigarette. THC is a partial agonist with an eﬃcacy of 20 to 25% [28,29], so that it would act at a higher receptor occupany than a full agonist like WIN55,212-2. Intravenous doses of 0.5 mg/kg THC are eﬀective in humans [30], and if 1% of the injected dose is distributed in the brain [31], this would correspond to a concentration of about 15 nmol/L. Applying the mass action equation and assuming that an in vitro Kd value for the CB1 receptor of 100 nM is appropriate, and a Bmax value of 100 nmol/L, an occupancy of 7% is estimated. However, it is likely that the fraction of THC available for binding to the receptor in vivo is quite small, since as an extremely lipophilic molecule, it will be distributed in brain membranes. This is expected to increase the eﬀective Kd value in vivo and so lower the estimate of occupancy, possibly by more than one order of magnitude. These considerations, therefore, suggest that only a very small proportion of the brain CB1 receptors need be activated to induce psychoactive eﬀects in humans, consistent with our results in mice, and that PET studies will not be able to measure the degree of occupancy achieved by marijuana smokers. Similar studies were done to evaluate the relationship between level of occupancy of the CB1 receptor by nonradioactive AM281 and the degree to which the antagonist AM281 was able to reverse the sedative eﬀect of the agonist WIN55,212-2 [32]. The AM281 eﬀectively restored the activity to normal levels (Fig. 10). In addition, AM281 alone was found to signiﬁcantly stimulate locomotor activity between 1 and 2 h after its administration (Fig. 11). Both the antagonism of the eﬀect of WIN55,212-2 and the eﬀect of AM281 alone increased progressively with doses up to 0.3 mg/kg AM281, but did not further increase at 1 mg/kg. A 50% occupancy of the CB1 receptor, as assessed by inhibition of [131I]AM281 binding, was achieved at a dose of 0.45 mg/kg. These data are consistent with prior in vitro indications that AM281 is a CB1 receptor antagonist or inverse agonist [33] and that AM281 inhibits an

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 10 Effect of increasing doses of AM281 on binding of [131I]AM281 in cerebellum and hippocampus, using brain stem as reference tissue.

endogenous cannabinoid tone. This baseline activity of the CB1 system might be maintained either by constitutive activity of the receptor [34], or by endogenous agonists such as anandamide [11]. These experiments [32] indicate that PET could be used to measure the degree of occupancy of CB1 receptors by antagonist or inverse agonist drugs in the human brain, if these drugs turn out to have useful therapeutic eﬀects, such as reducing memory loss in the elderly [35 –37].

V. CONCLUSIONS Our development of [123I]AM281, an antagonist radioligand for brain cannabinoid receptors, has allowed us to image this receptor for the ﬁrst time in vivo. Ex vivo autoradiographic experiments have been conducted in rodents, and SPECT studies have been conducted in baboons. Research

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 11 Stimulation of locomotor activity at CB1 receptor occupancy levels calculated from the effect of increasing doses of AM281 on binding of [131I]AM281; squares, antagonism of the sedative effect of WIN55,212-2 at 0 to 15 min; triangles, induction of hyperactivity by AM281 alone at 61 to 120 min.

continues to develop superior radioligands for SPECT research and also to develop CB1 receptor radioligands that can be labeled with positronemitting nuclides for PET. The results of the animal work to date will provide the foundation for using AM281 and other cannabinoid receptor radioligands in human imaging experiments. It is predicted from our animal data that psychoactive or medicinal doses of agonists such as THC will not alter CB1 receptor radioligand binding in the human brain. On the other hand, PET or SPECT is likely to be useful in determining the degree of CB1 receptor occupancy necessary for therapeutic eﬀects of antagonist drugs, as well as in evaluating CB1 receptor changes in addiction and in other diseases.

ACKNOWLEDGMENTS This research was carried out at the Brookhaven National Laboratory under contract DE-AC02-98CH10886 with the U.S. Department of

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Energy and supported by its Oﬃce of Health and Environmental Research. The research was also supported by awards from the National Institute on Drug Abuse to AM (DA 07515, DA 09158) and to ANG (DA 12412).

REFERENCES 1. Wolf AP, Fowler JS. Positron emission tomography. Biomedical research and clinical application. Neuroimaging Clin N Am 1995; 5: 87 – 101. 2. Fowler JS, Wolf AP. The heritage of radiotracers for positron emission tomography. Acta Radiol Suppl 1990; suppl 374: 13 – 16. 3. Gatley SJ, Volkow ND. Addiction and imaging of the living human brain. Drug Alcohol Depend 1998; 51: 97 – 108. 4. Volkow ND, Fowler JS, Gatley SJ, Logan J, Wang GJ, Ding YS, Dewey SL. PET evaluation of the dopamine system of the human brain. J Nucl Med 1996; 37: 1242 – 1256. 5. Volkow ND, Gillespie H, Mullani N, Tancredi L, Grant C, Ivanovic M, Hollister L. Cerebellar metabolic activation by delta-9-tetrahydrocannabinol in human brains. A study with positron emission tomography and 18F-2deoxy-2-ﬂuoro-D-glucose. Psychiat Res 1991; 40: 69 – 80. 6. Mathew RJ, Wilson WH, Coleman RE, Turkington TG, DeGrado TR. Marijuana intoxication and brain activation in marijuana smokers. Life Sci 1997; 60: 2075 – 2089. 7. Mathew RJ, Wilson WH, Turkington TG, Coleman RE. Cerebellar activity and disturbed time sense after THC. Brain Res 1998; 797: 183 – 189. 8. Volkow ND, Gillespie H, Mullani N, Tancredi L, Grant C, Valentine A, Hollister L. Brain glucose-metabolism in chronic marijuana users at baseline and during marijuana intoxication. Psychiatr Res Neuroimaging 1996; 67: 29 – 38. 9. Jansen EM, Haycock DA, Ward SJ, Seybold VS. Distribution of cannabinoid receptors in rat brain determined with aminoalkylindoles. Brain Res 1992; 575: 93 – 102. 10. Herkenham M, Lynn AB, Little MD, Johnson MR, Melvin LS, DeCosta BR, Rice KC. Cannabinoid receptor localization in brain. Proc Natl Acad Sci USA 190; 87: 1932 – 1936. 11. Felder CC, Briley EM, Axelrod J, Simpson JT, Mackie K, Devane WA. Anandamide, an endogenous cannabimimetic eicosanoid, binds to the cloned human cannabinoid receptor and stimulates receptor-mediated signal transduction. Proc Natl Acad Sci USA 1993; 90: 7656 – 7660. 12. Rinaldi-Carmona M, Barth F, Heaulme M, Shire D, Calandra B, Congy C, Martinez S, Maruani J, Neliat G, Caput D, Ferrara P, Soubrie P, Breliere

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

13.

14.

15. 16.

17.

18.

19.

20.

21.

22.

23.

24.

JC, Lefur G. SR141716A, a potent and selective antagonist of the brain cannabinoid receptor. FEBS Lett 1994; 350: 240 – 244. Devane WA, Breuer A, Sheskin T, Ja¨rbe TU, Eisen MS, Mechoulam, R. A novel probe for the cannabinoid receptor. J Med Chem 1992; 35: 2065 – 2069. Herkenham M, Lynn AB, Johnson MR, Melvin LS, de Costa BR, Rice KC. Characterization and localization of cannabinoid receptors in rat brain: a quantitative in vitro autoradiographic study. J Neurosci 1991; 11: 563 – 583. Biegon A, Kerman I. In: Biegon A, Volkow ND, eds. Sites of Drug Action in the Human Brain. Boca Raton, FL: CRC Press, 1995:65 – 74. Charalambous A, Marciniak G, Shiue CY, Dewey SL, Schlyer DJ, Wolf AP, Makriyannis A. PET studies in the primate brain and biodistribution in mice using ( – )-5V-18F-delta 8-THC. Pharmacol Biochem Behav 1991; 40: 503 – 507. Marciniak G, Charalambous A, Shiue CY, Dewey SL, Schlyer DJ, Wolf AP. Fluorine-18 labeled tetrahydrocannabinol: synthesis and PET studies in a baboon. J Labelled Compd Radiopharm 1991; 30: 413 – 415. John E, Bott A, Green M. Preparation and biodistribution of copper-67 complexes with tetradentate Schiﬀ-base ligands. J Pharm Sci 1994; 83: 587 – 590. Welch MJ, Chi DY, Mathias CJ, Kilbourn MR, Brodack JW, Katzenellenbogen JA. Biodistribution of N-alkyl and N-ﬂuoroalkyl derivatives of spiroperidol: radiopharmaceuticals for PET studies of dopamine receptors. Appl Radiat Isot 1986; 13: 523 – 526. Lan R, Gatley SJ, Makriyannis A. Preparation of iodine-123 labeled AM251, a potential SPECT radioligand for the brain cannabinoid CB1 receptor. J Labelled Comp Radiopharm 1996; 38: 875 – 881. Lan R, Liu Q, Fan P, Lin S, Fernando SR, McCallion D, Pertwee R, Makriyannis A. Structure – activity relationships of pyrazole derivatives as cannabinoid receptor antagonists. J Med Chem 1999; 42: 769 – 776. Gatley SJ, Giﬀord AN, Volkow ND, Lan R, Makriyannis A. Iodine-123 labeled AM251: a radioiodinated ligand which binds in vivo to the mouse brain CB1 cannabinoid receptor. Eur J Pharmacol 1996; 307: 301 – 308. Gatley SJ, Lan R, Volkow ND, Pappas N, King P, Wong CT, Giﬀord AN, Pyatt B, Dewey SL, Makriyannis A. Imaging the brain marijuana receptor: development of a radioligand that binds to cannabinoid CB1 receptors in vivo. J Neurochem 1998; 70: 417 – 423. Mathews WB, Ravert HR, Musachio JL, Frank RA, Rinaldi-Carmona M, Barth F, Dannals RF. Synthesis of [18F] SR144385: a selective radioligand for positron emission tomographic studies of brain cannabinoid receptors. J Labelled Compd Radiopharm 1994; 42: 589 – 596.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

Margonelli A, Angelini G, Attina A, Cartoni A, Finazzi Agro A, Maccarrone M. New Eﬀective chemical radiosynthesis of (2)-15,6,8,9, 11,12,14,15, -VH(N)1-Arachidonoyl-glycerol, (2-AG). A Putative endongenous cannabinoid receptor ligand. J Labelled Compd Radiopharm 1999; suppl 1: S787 – S789. Tamagnan GXJL, Gao Y, Amici LA, Baldwin RM, Innis RB. Potential SPECT radioligand for cannabinoid receptor. J Labelled Compd Radiopharm 1999; suppl 1: S191 – S192. Giﬀord AN, Bruneus M, Gatley SJ, Lan R, Makriyannis A, Volkow ND. Large receptor reserve for cannabinoid actions in the central nervous system. J Pharmacol Exp Ther 1999; 2: 478 – 483. Sim LJ, Hampson RE, Deadwyler SA, Childers SR. Eﬀects of chronic treatment with delta-9-tetrahydrocannabinol on cannabinoid-stimulated [35S]GTPgamma-S autoradiography in rat brain. J Neurosci 1996; 16: 8057 – 8066. Burkey TH, Quock RM, Consroe P, Roeske WR, Yamamura HI. delta-9Tetrahydrocannabinol is a partial agonist of cannabinoid receptors in mouse brain. Eur J Pharmacol 1997; 323: R3 – R4. Hollister LE, HK, G. The benzodiazepine receptor antagonist ﬂumazenil does not block clinical eﬀects of delta-9-tetrahydrocannabinol. Life Sci 1990; 47: 1655 – 1650. Schou J, Prockop LD, Dahlstrom G, Rohde C. Penetration of delta-9tetrahydrocannabinol and 11-OH-delta-9-tetrahydrocannabinol through the blood – brain barrier. Acta Pharmacol Toxicol (Copenh) 1977; 41: 33 – 38. Cosenza M, Giﬀord AN, Gatley SJ, Pyatt B, Liu Q, Makriyannis A, Volkow ND. Locomotor activity and occupancy of brain cannabinoid CB1 receptors by the antagonist/inverse agonist AM281. Synapse 2000; 38: 477 – 482. Giﬀord AN, Tang Y, Gatley SJ, Volkow ND, Lan R, Makriyannis A. Eﬀect of the cannabinoid receptor SPECT agent, AM 281, on hippocampal acetylcholine release from rat brain slices. Neurosci Lett 1997; 238: 84 – 86. Bouaboula M, Perrachon S, Milligan L, Canat X, Rinaldi-Carmona M, Portier M, Barth F, Calandra B, Pecceu F, Lupker J, Maﬀrand JP, Le Fur G, Casellas P. A selective inverse agonist for central cannabinoid receptor inhibits mitogen-activated protein kinase activation stimulated by insulin or insulin-like growth factor 1. Evidence for a new model of receptor/ligand interactions. J Biol Chem 1997; 272: 22330 – 22339. Collins DR, Pertwee RG, Davies SN. Prevention by the cannabinoid antagonist, SR141716A, of cannabinoid-mediated blockade of long-term potentiation in the rat hippocampal slice. Br J Pharmacol 1995; 115: 869 – 870.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

36. Katona I, Sperlagh B, Magloczky Z, Santha E, Kofalvi A, Czirjak S, Mackie K,Vizi ES, Freund TF. GABAergic interneurons are the targets of cannabinoid actions in the human hippocampus. Neuroscience 2000; 100: 797 – 804. 37. Giﬀord AN, Ashby CR. Electrically evoked acetylcholine release from hippocampal slices is inhibited by the cannabinoid receptor agonist, WIN 55212-2, and is potentiated by the cannabinoid antagonist, SR 141716A. J Pharmacol Exp Ther 1996; 277: 1431 – 1436.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

6 Structural and Pharmacological Aspects of Peptidomimetics Peter W. Schiller, Grazyna Weltrowska, Ralf Schmidt,* Thi M.-D. Nguyen, Irena Berezowska, Carole Lemieux, Ngoc Nga Chung, Katharine A. Carpenter,* and Brian C. Wilkes Clinical Research Institute of Montreal, Montreal, Quebec, Canada

I.

INTRODUCTION

Small linear neuropeptides are structurally ﬂexible molecules, capable of existing in a number of diﬀerent conformations of comparably low energy. This structural ﬂexibility precludes the determination of the bioactive conformation in solution and, furthermore, may be responsible for the lack of receptor selectivity of many of the naturally occurring peptide hormones and neurotransmitters, since conformational adaptation to diﬀerent receptor topographies takes place. In recent years, the introduction of conformational constraints into peptides either locally or at a particular amino acid residue (Na or Ca methylation, substitution of dehydro- or cyclic amino acids, etc.) or more globally (peptide cyclizations) emerged as a successful concept in the design of peptide analogues and peptidomimetics. In many cases, conformationally restricted peptide analogues showed high receptor selectivity and greatly improved conformational integrity. That diﬀerent receptor types for a given peptide hormone or neurotransmitter diﬀer from one another in their conforma-

*Present address: AstraZeneca Research Centre Montreal, St. Laurent, Quebec, Canada.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

tional requirements toward the peptide ligand was ﬁrst unambiguously demonstrated in the case of opioid receptors by comparing the receptor aﬃnity proﬁles of a cyclic enkephalin analogue and its corresponding open-chain analogue [1]. Furthermore, conformationally restricted peptide analogues often showed enhanced stability against enzymatic degradation because a scissile peptide bond in the native peptide may no longer be susceptible to enzymatic cleavage when incorporated in a conformationally constrained structure. Conformational restriction of peptides represents a ﬁrst step in the rational approach to develop peptidomimetics because conformationally constrained analogues may serve as relatively rigid templates for further structural modiﬁcation aimed at removing some of the less attractive peptide structural features (e.g., peptide backbone replacements). In this chapter we illustrate these principles with examples from the opioid peptide ﬁeld. The existence of three major classes (A, y, n) of opioid receptors is now well established [2,3]. Furthermore, the results of classical pharmacological testing and of opioid receptor binding studies indicate that various opioid receptor subtypes might also exist [4,5]. Recent work in the cloning of opioid receptors conﬁrmed the existence of a y [6,7], a n [8], and a A [9] receptor, but so far has not led to the identiﬁcation of receptor subtypes. Like all other G-protein-linked receptors, the three opioid receptors show seven putative transmembrane helices and considerable sequence similarity (60–70%) among themselves. Highest sequence identity is observed in the membrane-spanning segments and in the intracellular loops, whereas lower homology is seen in the extracellular regions. The ligand binding site is thought to be located in the cavity within the transmembrane domain of the receptors, whereas the extracellular loops may act as ﬁlters for the ligands, thus possibly playing a role in ligand selectivity. The precise threedimensional structures of the opioid receptors remain elusive because x-ray diﬀraction or NMR spectroscopic analysis cannot be used for their determination. However, theoretical analyses led to approximate models that are of interest for ligand docking studies and for the testing of hypotheses related to ligand design [10]. Since the discovery of the enkephalins in 1975 [11] a large number of endogenous opioid peptides have been detected in mammals, and at present three distinct families of opioid peptides are known (for a review, See Ref. 12). These are the enkephalins, the endorphins (a-, h-, and g-), and the dynorphins and neoendorphins. The recently discovered endomorphins [13] also may represent endogenous opioid peptides. Peptides with opioid activity have also been isolated from tryptic digests of milk casein

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

(h-casomorphins) [14] and from frog skin—dermorphins [15] and deltorphins [16]. Like the morphine-related opiates, opioid peptides produce a large spectrum of central and peripheral eﬀects, including analgesia, tolerance and physical dependence, respiratory depression, euphoria, dysphoria and hallucinations, sedation, feeding and other behavioral eﬀects, hypothermia/hyperthermia, miosis, eﬀects on tumor growth, control of release of several peptide hormones and catecholamines, eﬀects on transit in the gut, and various cardiovascular eﬀects. It has not yet been possible to establish clear-cut relationships linking speciﬁc opioid receptor types to distinct opioid eﬀects. This is mainly because until recently, potent, stable agonists and antagonists with high speciﬁcity for the various receptor types have not been available. The results of several studies suggest that the A receptor plays a primary role in mediating analgesia; however, there is some evidence that y and n interactions may result in analgesic eﬀects as well. Among the various isolated organ preparations used in the in vitro bioassays, the guinea pig ileum (GPI) contains primarily A receptors but also n receptors; in the mouse vas deferens (MVD) y receptors are predominant, even though A and n receptors are also present (for a review, see Ref. 17). It is now well recognized that the binding of agonists to opioid receptors leads to inhibition of adenylate cyclase via interaction with a guanine nucleotide regulatory protein [18]. The various endogenous opioid peptides resulting from processing of the three mammalian precursor molecules display only limited selectivity toward the diﬀerent receptor types [12]. The only naturally occurring opioid peptides with high receptor speciﬁcity discovered so far are the yselective deltorphins [16] and, possibly, the endomorphins [13]. Medicinal chemists and peptide chemists have made numerous eﬀorts to develop opioid agonists and antagonists with improved receptor selectivity. Substantial progress has been made in the development of selective nonpeptide opioid receptor ligands (for a review, see Ref. 19). The design of opioid peptide analogues with high receptor selectivity has also been very successful (for reviews, see Refs. 20 and 21). Structural modiﬁcation of the enkephalins, h-casomorphin (morphiceptin), dermorphin, and the deltorphins through various amino acid and end group substitutions produced linear analogues that turned out to be highly selective A agonists, y agonists, or y antagonists. The preparation of numerous linear analogues of dynorphin A led to several potent n agonists with high n-receptor selectivity (for a review, see Ref. 22), whereas potent and selective n opioid antagonists structurally derived from dynorphin A have not yet been reported. A most successful strategy in opioid peptide analogue design

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

has been the incorporation of conformational constraints through peptide cyclizations or substitution of conformationally restricted amino acids. In particular, conformational restriction of the enkephalins or the dermorphins and deltorphins produced potent and highly selective A agonists, y agonists and y antagonists. Agonists and antagonists showing high speciﬁcity for a particular opioid receptor class are valuable as pharmacological tools and may also have potential as therapeutic agents. Futhermore, it has been recognized that the development of opioid compounds with mixed agonist/antagonist properties may lead to improved analgesics with minimal side eﬀects. In this chapter we discuss the development of highly selective A agonists and y antagonists, the ﬁrst known compounds with mixed A agonist/y antagonist properties and a new class of dipeptide y agonists.

II. M OPIOID AGONISTS A. Cyclic Peptides with M-Agonist Properties The ﬁrst conformationally restricted opioid peptide reported in the literature was the cyclic enkephalin analogue H-Tyr-c[-D-A2bu-Gly-PheLeu-] (A2bu = a,g-diaminobutyric acid), which showed considerable preference for A receptors over y receptors [23]. Homologues of this compound containing a D-ornithine or D-lysine residue in place of DA2bu also were A-selective agonists [24]. Analogues of H-Tyr-c[-D-LysGly-Phe-Leu-] having one or two reversed amide bonds showed further improved A selectivity and excellent stability against enzymatic degradation [25]. One of the most selective cyclic opioid peptides with A agonist properties reported to date is the dermorphin-related tetrapeptide H-TyrD-Orn-Phe-Asp-NH2 [26]. A theoretical conformational analysis performed with this compound revealed that the 13-membered peptide ring structure was highly constrained and that the lowest energy conformer was characterized by a tilted stacking interaction between the Tyr1 and Phe3 aromatic rings [27]. Expansion of the peptide ring structure in this analogue, as achieved by replacement of Asp with Glu, resulted in the compound H-Tyr-D-Orn-Phe-Glu-NH2, which showed only slight preference for A receptors over y receptors [28]. The results of a molecular dynamics simulation carried out with this analogue showed that its 14membered peptide ring structure had moderate structural ﬂexibility, while the exocyclic Tyr1 residue and the Phe3 side chain enjoyed considerable orientational freedom [29]. j

j

j

j

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

To limit the structural ﬂexibility of these two aromatic residues, conformationally restricted analogues of Phe and/or Tyr were substituted [28,30] (Fig. 1). Replacement of Phe in the parent peptide with the cyclic phenylalanine analogue 2-aminoindan-2-carboxylic acid (Aic) resulted in a compound that showed only four times lower A aﬃnity but 65 times lower aﬃnity for y receptors and, consequently, markedly improved A selectivity (K yi /K Ai =49.6) (Table 1). The analog H-Tyr-D-Orn-Atc-Glu-NH2, containing the conformationally constrained phenylalanine analogue 2-aminotetralin-2-carboxylic acid (Atc) at the 3 position of the peptide sequence, also retained high A-agonist potency and showed further improved A selectivity. Interestingly, the diastereomeric D-Atc3 analogue also displayed good A receptor aﬃnity and high A selectivity. This observation is in contrast to the weak aﬃnity observed with the D-Phe3 analogue in comparison to the L-Phe3 parent peptide. Thus, stereospeciﬁcity was lost as a consequence of side chain conformational restriction, presumably because the D-Atc3 analogue binds to the receptor in a manner diﬀerent from that of the D-Phe3 analogue. Replacement of Tyr1 in the cyclic parent peptide with 6-hydroxy-2-amino-tetralin-2-carboxylic acid (Hat) produced the compound H-Hat-D-Orn-Phe-Glu-NH2, with only about three times reduced aﬃnity for A and y receptors. Its diastereomer, H-D-Hat-DOrn-Phe-Glu-NH2, was a full agonist at both the A and the y receptor but showed a substantial decrease in potency. Finally, the conformationally highly constrained analogue H-( D , L)-Hat-D -Orn-Aic-Glu -NH 2 also showed high A-receptor aﬃnity and marked A selectivity. This compound essentially contains only two freely rotatable bonds and represents the most rigid, rationally designed opioid peptidomimetic reported to date. j

j

j

j

j

j

Figure 1

j

j

Structural formulas of cyclic analogues of phenylalanine and tyrosine.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 1 Opioid Receptor Affinities of Cyclic Dermorphin Analoguesa Compound j

j

H-Tyr-D-Orn-Phe-Glu-NH2 H-Tyr-D-Orn-D-Phe-Glu-NH2 H-Tyr-D-Orn-Phe-Glu-NH2 H-Tyr-D-Orn-Atc-Glu-NH2 H-Tyr-D-Orn-D-Atc-Glu-NH2 H-Hat-D-Orn-Phe-Glu-NH2 H-D-Hat-D-Orn-Phe-Glu-NH2 H-(D,L)-Hat-D-Orn-Aic-Glu-NH2 j

j

j

j

j

j

j

j

j

j

j

j

j

a

j

K Ai (nM)

K yi (nM)

K yi /K Ai

0.981 1660 4.21 8.26 26.3 2.91 54.2 7.68

3.21 14,000 209 1,570 3,510 10.8 74.7 119

3.27 8.43 49.6 190 133 3.71 1.38 15.5

Displacement of [3H]DAMGO (A-selective) and [3H]DSLET (y-selective) from rat brain membrane binding sites.

j

j

B. Conformational Study of H-Hat-D-Orn-Aic-Glu -NH2 The results of a molecular mechanics study indicated that the lowest energy conformation of H-Hat-D-Orn-Aic-Glu-NH2 is still characterized by a tilted stacking interaction of the aromatic rings of the residues in positions 1 and 3 of the peptide sequence (B. C. Wilkes and P. W. Schiller, unpublished results) (Fig. 2). It had been suggested that this tilted stacking arrangement of the two aromatic rings might represent a structural requirement for high A-receptor aﬃnity of the tetrapeptide HTyr-D-Orn-Phe-Asp-NH2 and structurally related cyclic dermorphin analogues [31]. An alternative model of the A-receptor-bound conformation based on conformational analysis of morphiceptin analogues is characterized by a larger distance (f10 A˚) between the Tyr1 and Phe3 aromatic rings [32]. The A-selective morphiceptin analogue H-Tyr-ProPhe(NMe)-D-Pro-NH2 (PL017) in this proposed bioactive conformation is depicted in Figure 2. A conformer of H-Hat-D-Orn-Aic-Glu-NH2 with an energy 1.1 kcal/mol higher than that of the lowest energy structure showed good spatial overlap of its Tyr1 and Phe3 aromatic rings and N-terminal amino group with the corresponding moieties in morphiceptin in this proposed bioactive conformation, the root-mean-square deviation being 1.1 A˚ (Fig. 2). Thus, reasonable low energy conformers consistent with either one of the two proposed bioactive conformations can be assumed by H-Hat-D-Orn-Aic-Glu-NH2, and both models remain plausible candidate structures for the A-receptor pharmacophore. j

j

j

j

j

j

j

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

j

j

j

Figure 2 Spatial overlap of low energy conformers of H-Hat-D-Orn-Aic-GluNH2 (heavy lines) with proposed models of the A-receptor-bound conformation (light lines) based on conformational analysis of H-Tyr-D-Orn-Phe-Asp-NH2 [24] (left panel) and H-Tyr-Pro-Phe(NMe)-D-Pro-NH2 (PL017) [29] (right panel). j

j

III. D OPIOID ANTAGONISTS The ﬁrst y antagonists derived from opioid peptides were obtained through diallylation of the N-terminal amino group of enkephalin analogues. The best known compound of this type is N,N-diallyl-Tyr-Aib-Aib-Phe-LeuOH (ICI 174,864; Aib = a-aminoisobutyric acid), which is quite y selective (K Ai /K yi = 128) but not very potent (K yi = 199 nM; Ke = 69 nM in the MVD assay) [33,34]. The nonpeptide y antagonist naltrindole (NTI) [35] is highly potent but displays only modest y selectivity (K Ai /K yi = 21.2). A benzofuran analogue of naltrindole, NTB, showed improved y selectivity but somewhat lower y-antagonist potency [36]. However, both NTI and NTB also turned out to be antagonists against A and n agonists in the GPI assay, with potencies (Ke = 29–48 nM) about 100 to 300 times lower than those observed against y agonists in the MVD assay (Ke = 0.13 and 0.27 nM, respectively) [36]. The recently discovered TIP(P) peptides represent a novel class of potent and highly selective y-opioid antagonists [37]. The two prototype antagonists were the tetrapeptide H-Tyr-Tic-Phe-Phe-OH (TIPP;Tic = 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid) and the

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

tripeptide H-Tyr-Tic-Phe-OH (TIP). TIPP showed high antagonist potency against various y agonists in the MVD assay (Ke = 3–5 nM), high y-receptor aﬃnity (K yi = 1.22 nM) in the rat brain membrane receptor binding assay, and extraordinary y selectivity (K Ai /K yi = 1410). Importantly, TIPP displayed no A- or n- antagonist properties in the GPI assay at concentrations as high as 10 AM. In comparison with TIPP, the tripeptide TIP was a somewhat less potent and less selective y antagonist.

A. Structure–Activity Studies of TIP(P) Peptides Methylation of the N-terminal amino group of TIPP produced a compound, Tyr(NMe)-Tic-Phe-Phe-OH, with four fold enhanced y-antagonist potency and further improved y selectivity [38] (Table 2). Replacement of Tyr1 in TIPP with 2V,6V-dimethyltyrosine (Dmt) led to the analog H-DmtTic-Phe-Phe-OH (DIPP), with y-antagonist potency in the subnanomolar range (Ke = 0.196 nM in the MVD assay) and with still excellent y selectivity. This compound turned out to be four times more potent than naltrindole as antagonist in the MVD assay (Table 2) and it represents the most potent y-opioid antagonist reported to date. The results of stability studies indicated that TIPP is stable in aqueous solution for extended periods of time but undergoes slow diketopiperazine formation and concomitant cleavage of the Tic2—Phe3 peptide bond in DMSO or MeOH [39]. To prevent this spontaneous degradation, a TIPP analogue containing a reduced peptide bond between Tic2 and Phe3 was synthesized. The resulting pseudopeptide, H-Tyr-TicC[CH2–NH]Phe-Phe-OH (TIPP[C]), retained y-antagonist potency comparable to that of the parent peptide and showed extraordinary y selectivity in the receptor binding assays (K Ai /K yi = 10500) [40], being about 500 times more y selective than naltrindole and 17 times more y selective than [D-Ala2]deltorphin II (Table 2). Moreover, TIPP[C] was shown to be highly stable against chemical and enzymatic degradation. It also showed selectivity ratios exceeding 10,000 against all A- and n- receptor subtypes (A1, A2, n1, n2, n3) [41] and thus represents an excellent pharmacological tool. The corresponding pseudotripeptide, H-Tyr-TicC[CH2-NH]Phe-OH (TIP[C]), also retained high y-antagonist potency, and its y selectivity was 40 times greater than that of its parent, TIP. Methylation of the secondary amino group of the reduced peptide bond in TIPP[C] produced the compound H-Tyr-TicC[CH2-NCH3]PhePhe-OH, which retained the high y antagonist potency of the parent pseudopeptide and showed even higher y selectivity (K Ai /K yi = 15,900, Table 2). This compound is nearly 300 times more y selective than DPDPE

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 2 Antagonist Potencies and Opioid Receptor Affinities of TIPP Analogues Compound H-Tyr-Tic-Phe-Phe-OH (TIPP) H-Tyr-Tic-Phe-OH (TIP) Tyr(NMe)-Tic-Phe-Phe-OH H-Dmt-Tic-Phe-Phe-OH H-Tyr-TicC[CH2-NH]Phe-Phe-OH (TIPP[C]) H-Tyr-TicC[CH2-NH]Phe-OH (TIP[C]) H-Tyr-TicC[CH2-NCH3]Phe-Phe-OH H-Tyr(3V-I)-Tic-Phe-Phe-OH H-Tyr(3V-I)-Tic-Phe-OH H-Tyr(3V-I)-TicC[CH2-NH]Phe-Phe-OH H-Tyr-Tic-Leu-Phe-OH H-Tyr-Tic-Ile-Phe-OH H-Tyr-Tic-Cha-Phe-OH (TICP) H-Tyr-TicC[CH2-NH]Cha-Phe-OH (TICP[C]) H-Tyr(3V-I)-Tic-Cha-Phe-OH H-Dmt-Tic-OH Naltrindole DPDPE [D-Ala2]deltorphin II a b

Ke (nM)a 5.86 11.7 1.22 0.196 2.89

K Ai (nM)b

K yi (nM)b

K Ai /K yi

1,720 1,280 13,400 141 3,230

1.22 1,410 9.07 141 1.29 10,400 0.248 569 0.308 10,500

9.06 10,800 4.76 13,400 Agonist 5,230 141 12,100 19.2 2,660 7.32 904 12.7 6,460 0.438 3,600 0.219 1,050

1.94 5,570 0.842 15,900 24.8 211 60.0 202 2.08 1,280 2.84 318 4.37 1,480 0.611 5,890 0.259 4,050

12.7 6.55 0.636 Agonist Agonist

4,010 3.33 1,360 1.84 3.86 0.182 943 16.4 3,930 6.43

1,200 739 21.2 57.5 611

Determined against DPDPE in the MVD assay. Binding assay based on displacement of [3H]DAMGO (A-selective) and [3H]DSLET (y-selective) from rat brain membrane binding sites.

and shows even slightly higher y-receptor selectivity than the TIPP[C] parent peptide. For the purpose of opioid receptor binding studies, TIPP was also radioiodinated. Surprisingly, [125I]TIPP binding to y receptors in N4TG1 neuroblastoma cells was substantially reduced in the presence of Na+ and Gpp(NH)p [42]. These results indicated that substitution of an iodine atom at the 3V position of Tyr1 in TIPP had turned the y antagonist into a y agonist. The corresponding ‘‘cold’’ analogue, H-Tyr(3’-I)-Tic-Phe-PheOH, was then synthesized and shown to be a full agonist in the MVD assay (IC50 = 97 nM). This agonist eﬀect was antagonized by TIPP (Ke = 11 nM) [38]. Corresponding iodination of the Tyr residue in TIP and TIPP[C] did not result in agonism, but somewhat reduced antagonist

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

potency was observed (Table 2). It therefore appears that the astonishing conversion observed with the tetrapeptide TIPP may be due to an overall conformational eﬀect rather than to a direct, local eﬀect of the iodine substituent. Interestingly, substitution of a bromine or chlorine atom at the 3V position of Tyr1 in TIPP produced partial agonists with respective intrinsic eﬃcacies of 0.16 and 0.12, whereas the Tyr(3V-F)-analogue was again a pure antagonist (Ke = 13.0 nM) [38]. Thus, systematic substitution of halogen atoms beginning with iodine and in the order of the periodic table produced a progressive decrease in intrinsic activity and a concomitant increase in aﬃnity at the y receptor (K yi = 24.2, 3.62, 3.00 and 1.62 nM, respectively). Replacement of the Phe3 residue in TIPP with the aliphatic amino acid residues Leu or Ile resulted in analogues that retained high yantagonist potency and considerable y selectivity (Table 2). This result is in agreement with the weak y-antagonist activity that had been reported for the tripeptide H-Tyr-Tic-Ala-OH [43]. Obviously, an aromatic residue at the 3 position of the peptide sequence of TIP(P) peptides is not absolutely required for y antagonist activity. Most interestingly, saturation of the Phe3 aromatic ring in TIPP, as achieved through substitution of cyclohexylalanine (Cha), led to H-Tyr-Tic-Cha-Phe-OH [TICP], a compound showing substantially increased y-antagonist potency and higher y selectivity than the parent peptide [44]. The corresponding pseudopeptide, H-Tyr-TicC[CH 2 -NH]Cha-Phe-OH (TICP[C]), showed a further improvement in y-antagonist activity. Its y-antagonist potency is comparable to that of the analogue H-Dmt-Tic-Phe-Phe-OH but, in comparison with the latter peptide, it is seven times more y selective (K Ai /K yi = 4050) [44]. Both TIPP[C] and TICP[C] were prepared in tritiated form [45,46] and should turn out to be valuable new radioligands for y receptor labeling studies in vitro and in vivo. The analogue H-Tyr(3V-I)-Tic-Cha-Phe-OH was an antagonist in the MVD assay with a potency about 30 times lower than that of TICP. Thus, unlike in the case of TIPP, introduction of an iodine substituent at the 3V position of Tyr1 in TICP did not produce a y agonist. This result demonstrates once again how a relatively subtle structural modiﬁcation, such as the saturation of an aromatic ring, can have a determinant eﬀect on agonist versus antagonist behavior. In 1995 the dipeptide H-Dmt-Tic-OH was reported to be a y-opioid antagonist with unprecedented y-receptor aﬃnity (K yi = 0.022 nM) and y receptor selectivity (K Ai /K yi = 150,000) [47]. However, in a direct comparison under identical assay conditions, this compound showed about 30 times lower y-antagonist potency and 6 times lower y-receptor selectivity

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

than TICP[C] [48] (Table 2). Similar results were obtained in a more recent study [49], which conﬁrmed that H-Dmt-Tic-OH had much lower y receptor aﬃnity [IC50(y) = 1.6 nM] and much lower y selectivity [IC50(A)/ IC50(y) = 558] than had originally been reported. Furthermore, H-DmtTic-OH was found to be unstable in organic solvents owing to diketopiperazine formation (P. W. Schiller and T. M.-D. Nguyen, unpublished results).

B. Conformational Studies of TIP and TIPP A molecular mechanics study (grid search and energy minimization) of the tripeptide y-antagonist TIP resulted in several low energy conformers having energies within about 2 kcal/mol of that of the lowest energy structure [50]. The centrally located Tic residue imposes a number of conformational constraints on the N-terminal dipeptide segment; however, the results of molecular dynamics simulations indicate that this tripeptide still shows some structural ﬂexibility at the Phe3 residue. Attempts to demonstrate spatial overlap between the pharmacophoric moieties of low-energy conformers of TIP and the structurally rigid nonpeptide y antagonist naltrindole were made by superimposing either the Tyr1 and Phe3 aromatic rings and the N-terminal amino group or the Tyr1 and Tic2 aromatic rings and the N-terminal amino group of the peptide with the corresponding aromatic rings and nitrogen atom in the alkaloid structure. In each case the investigators found a conformer of TIP with an energy very close to that of the lowest energy structure (2.1 kcal/mol higher). However, the low-energy conformer showing spatial overlap of its Tic2 aromatic ring with the six-membered aromatic ring of the indole moiety in naltrindole (Fig. 3) appears to be a more plausible candidate structure of the y-receptor-bound conformation for two reasons: 1. 2.

The Tic2 aromatic ring has been shown to be of crucial importance for y antagonist activity [51]. The y-antagonist properties are maintained upon replacement of the Phe3 residue in the peptide with an aliphatic amino acid residue (see earlier).

This model of the receptor-bound conformation of TIP is characterized by a clustered conﬁguration of the three aromatic moieties with the Phe3 aromatic ring sandwiched between the Tyr1 and Tic2 aromatic rings. A molecular mechanics study of TIPP and TIPP[C] produced about 70 structures within 3 kcal/mol of the lowest energy conformation in each

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 3 Superimposition of a low energy conformer of TIP (heavy lines) with the minimized structure of naltrindole (light lines). The Tyr1 and Tic2 aromatic rings and the N-terminal amino group of the peptide are superimposed with the corresponding moieties in the alkaloid structure. The superimposed molecules are shown in two different orientations.

case [52]. The lowest energy conformers of both TIPP and TIPP[C] showed good overlap of their Tyr1 and Tic2 aromatic rings and N-terminal amino group with the corresponding pharmacophoric moieties of naltrindole. Thus, these results are in agreement with the model of the receptor-bound conformation of TIP proposed earlier. This model is characterized by all-trans peptide bonds and was deﬁnitely conﬁrmed by conformational analyses of two TIPP analogues (y antagonists) in which a cis peptide bond between the Tyr1 and Tic2 residues is sterically forbidden [53]. Both TIPP and TIPP[C] are very hydrophobic peptides, and the results of the theoretical conformational analyses clearly indicated that they enjoy considerable structural ﬂexibility, particularly in their Cterminal dipeptide segment. There is no doubt that their conformations are quite dependent on the environment. According to our theoretical analysis, a crystal structure of TIPP published in 1994 [54] is about 3 kcal/ mol higher in energy than the lowest energy structure and shows no similarity to any of the calculated low energy structures [52,53]. The crystal structure of TIPP appears to be stabilized by a large number of intermolecular hydrophobic contacts between layers of TIPP molecules in the crystal and by several hydrogen bonds to solvent (AcOH) molecules. There is no reason to believe that it resembles the y-receptor-bound conformation of TIPP. In an aqueous environment TIP(P) peptides may undergo a so-called hydrophobic collapse [55]. It is possible that subtle structural modiﬁcations, such as introduction of an iodine sub-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

stituent at the 3V position of Tyr1, saturation of the Phe3 aromatic ring, or reduction of the Tic2—Phe3 peptide bond, may produce diﬀerent patterns of aromatic ring clustering that could result in either y-agonist or yantagonist activity, as described earlier.

C. Effect of D-Opioid Antagonists on the Development of Morphine Tolerance and Dependence Blockade of y receptors with the nonpeptide y antagonist naltrindole concurrently with chronic morphine treatment has been reported to attenuate the development of tolerance and the severity of the precipitated withdrawal syndrome in mice [56]. In an eﬀort to corroborate these results, the eﬀects of TIPP, TIPP[C], and naltrindole on the development of morphine tolerance and dependence were examined. Each of the antagonists was continuously infused into the lateral ventricle of rats treated chronically with subcutaneous morphine [57]. After a 6-day period of drug administration, rats treated with TIPP[C] showed no morphine tolerance and a greatly reduced incidence of withdrawal symptoms following injection of naloxone. Naltrindole and TIPP also signiﬁcantly decreased the amount of time spent in withdrawal but did not attenuate the development of morphine tolerance. More recently, morphine was shown to retain its A-receptor-mediated analgesic activity in y-opioid receptor knockout mice without producing analgesic tolerance upon chronic administration [58]. These interesting ﬁndings clearly demonstrate that y-opioid receptors play a major role in the development of morphine tolerance and dependence and suggest the possibility of the combined use of a A type opioid analgesic and a y-opioid antagonist in the treatment of chronic pain. Even more interestingly, these results indicate that mixed A agonist/y antagonists can be expected to be analgesics with low propensity to produce tolerance and dependence and, therefore, might be of beneﬁt in the management of chronic pain.

IV. MIXED M AGONIST/D ANTAGONISTS A. Prototypes and Structure–Activity Relationships The ﬁrst known example of a mixed A agonist/y antagonist was a TIPP analogue in which the free C-terminal carboxylate function had been replaced by a carboxamide function [37]. This compound, H-Tyr-Tic-Phe-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Phe-NH2 (TIPP-NH2), was found to be a moderately potent A agonist in the GPI assay and a potent y antagonist in the MVD assay (Table 3). Replacement of Tyr1 in TIPP-NH2 with Dmt produced the compound H-Dmt-Tic-Phe-Phe-NH2 (DIPP-NH2), showing an increase in both Aagonist potency and y-antagonist activity by nearly two orders of magnitude [59,60]. In the receptor binding assays DIPP-NH2 displayed very high y-receptor aﬃnity and still some preference for y receptors over A receptors. In comparison with DIPP-NH2, the corresponding analogue with a reduced peptide bond between the Tic2 and Phe3 residues, H-Dmt-TicC [CH2-NH]Phe-Phe-NH2 (DIPP-NH2[C]), was about twice as potent as agonist in the GPI assay and about half as potent as antagonist in the MVD assay. Showing A- and y-receptor aﬃnities that were both in the subnanomolar range, DIPP-NH2[C] was essentially nonselective (K Ai /K yi = 2.11; Table 3). Therefore, DIPP-NH2[C] represents the ﬁrst known opioid compound with balanced A-agonist/y-antagonist properties [59,60]. In the rat tail ﬂick test, DIPP-NH2[C] given intracerebroventricularly (ICV) produced a potent analgesic eﬀect, being about three times more potent than morphine. It produced less acute tolerance than morphine, but still a certain level of chronic tolerance. Unlike morphine, DIPP-NH2[C] produced no physical dependence upon chronic administration at high doses. Thus, DIPP-NH2[C] fulﬁlled to a large extent the expectations based on the mixed A-agonist/y-antagonist concept [60]. Surprisingly, elimination of the C-terminal carboxylate function of the tripeptide y antagonist TIP resulted in H-Tyr-Tic-NH-(CH2)2-Ph (Ph = phenyl), a compound that was a moderately potent full y agonist (Table 3) [61]. Interestingly, lengthening of the phenylethyl substituent by insertion of an additional methylene group restored y antagonism, as indicated by the ﬁnding that the dipeptide derivative H-Tyr-Tic-NH(CH2)3-Ph was a moderately potent y antagonist in the MVD assay and a relatively weak partial A agonist in the GPI assay. This remarkable dependence of y-agonist versus y-antagonist behavior on the length of the phenylalkyl substituent may be due to conformational eﬀects resulting in diﬀerent clustering of the three aromatic moieties present in these molecules. The analogue H-Dmt-Tic-NH-(CH2)3-Ph showed very high aﬃnity for both A and y receptors and was a potent y antagonist in the MVD assay (Ke = 1.69 nM). That this compound displayed relatively modest agonist potency in the GPI assay suggests that it also may have partial A-agonist properties and that it may represent a mixed partial A agonist/y antagonist. In the case of high aﬃnity A-receptor ligands, partial agonism is not always directly apparent in the GPI assay because the

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 3 In Vitro Opioid Activities and Receptor Affinities of Mixed A Agonist/y Antagonists GPI Compound

IC50, nM

H-Tyr-Tic-Phe-Phe-NH2 H-Dmt-Tic-Phe-Phe-NH2 H-Dmt-TicC[CH2-NH]Phe-Phe-NH2 H-Tyr-Tic-NH-(CH2)2-Ph H-Tyr-Tic-NH-(CH2)3-Ph H-Dmt-Tic-NH-(CH2)3-Ph H-Tyr-c[-D-Orn-Phe-D-Pro-Gly-] H-Tyr-c[-D-Orn-2-Nal-D-Pro-Gly-] H-Dmt-c[-D-Orn-2-Nal-D-Pro-Gly-]

1700 18.2 7.71 3010 (42 %)c 102 2.14 384 7.88

Binding assaysa

MVD IC50, (nM)

Ke (nM)b

K Ai (nM)

18.0 0.209 0.537

78.8 1.19 0.943 69.1 160 0.386 0.881 5.89 0.460

82.0 41.9 1.69 4.89 233 2.13

Displacement of [3H]DAMGO (A-selective) and [3H]DSLET (y-selective) from rat brain membrane binding sites. Determined against DPDPE. c Maximal inhibition of the contractions at 10 AM. a b

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

K yi (nM) 3.00 0.118 0.447 5.22 3.01 0.0871 13.2 17.2 0.457

K Ai /K yi 26.3 10.1 2.11 13.2 53.2 4.43 0.0667 0.342 1.01

ileum has a very high A-receptor reserve. Substituted Tyr-Tic-dipeptide amides with mixed A-agonist/y-antagonist properties are of interest because their small molecular size and lipophilic character may facilitate their passage across the blood–brain barrier (BBB). Further eﬀorts aimed at strengthening the A-agonist component of this class of compounds may be required. Another prototype of a mixed A agonist/y antagonist is the cyclic hcasomorphin analogue H-Tyr-c[-D-Orn-2-Nal-D-Pro-Gly-] [62]. This compound turned out to be a fairly potent A agonist in the GPI assay and showed relatively modest y-antagonist potency in the MVD assay (Table 3). The 2-naphthylalanine (2-Nal) residue in this compound is a key structural determinant for its y-antagonist behavior, since the corresponding Phe3 analogue, H-Tyr-c[-D-Orn-Phe-D-Pro-Gly-] was found to be a full y agonist in the MVD assay [62]. As expected, an analogue containing Dmt in place of Tyr1, H-Dmt-c-[-D-Orn-2-Nal-D-Pro-Gly-], showed greatly increased A-agonist and y-antagonist potency [63]. This pentapeptide displayed almost equal aﬃnities for A and y receptors in the subnanomolar range and, thus, represents another example of a balanced A agonist/y antagonist. In comparison with DIPP-NH2[C], H-Dmt-c[-DOrn-2-Nal-D-Pro-Gly-] has the same A-agonist potency in the GPI assay and is about four times less potent as a y antagonist in the MVD assay. The various compounds described in this section represent the only known mixed A-agonist/y-antagonist substances reported to date. Analgesic testing of all these prototypes will reveal which type of compound has the greatest potential for the development of viable analgesics. Further analogues may have to be prepared and examined to determine the ratio between A-agonist and y-antagonist potency required for optimal attenuation of tolerance and dependence development. Additional structural modiﬁcations may be necessary to increase analgesic potency and bioavailability.

B. Conformational Study of H-Tyr-c[-D-Orn-2-Nal-DPro-Gly-] The conformation of the mixed A agonist/y antagonist H-Tyr-c[-D-Orn-2Nal-D-Pro-Gly-] in comparison to that of H-Tyr-c[-D-Orn-Phe-D-Pro-Gly-] was studied in DMSO-d6 by NMR spectroscopy and by molecular mechanics calculations [62,64]. Neither peptide showed nuclear Overhauser eﬀects between CaH protons or chemical exchange cross peaks in spectra obtained by total correlation and rotating frame Overhauser enhance-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

ment spectroscopy (TOCSY, ROESY). These results indicated that the average preferred solution conformation of both peptides was characterized by all-trans peptide bonds. The results of temperature-dependence studies of the amide proton chemical shifts in conjunction with those of the molecular mechanics studies indicated that the two analogues had backbone conformations that were both stabilized by Tyr1-COHNPhe3 (or 2-Nal3) and D-Orn2-COHNy-D-Orn2 hydrogen bonds. Furthermore, ROESY experiments revealed a close proximity between the aromatic moiety of the 3-position residue and the pyrrolidine ring of the 4 D-Pro residue in these two compounds. The comparison of all calculated low-energy conformations with the various proton NMR parameters led to proposals for the solution conformation of these two peptides (Fig. 4). Inspection of the structures reveals that the Phe3- and 2-Nal3analogues have similar backbone conformations and the same side chain orientation at the 3 position. These results suggest that the y-antagonist

Figure 4 Proposed solution conformations of H-Tyr-c-[-D-Orn-2-Nal-D-ProGly-] (left panel) and H-Tyr-c-[D-Orn-Phe-D-Pro-Gly-] (right panel).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

properties of the 2-Nal3 analogue may not be due to a diﬀerence in its overall conformation in comparison to the Phe3 analogue but rather may be the result of a direct interference of the 2-naphthyl moiety per se at the receptor binding site, preventing proper alignment of the peptide such as required for signal transduction.

V. D AGONISTS y-Opioid agonists are known to produce analgesic eﬀects and look promising because they induce less tolerance and physical dependence than morphine, no respiratory depression, and few or no adverse gastrointestinal eﬀects [65,66]. Selective peptide y agonists currently available include the enkephalin analogues H-Tyr-D-Thr-Gly-Phe-LeuThr-OH (DTLET), H-Tyr-c[D-Pen-Gly-Phe-D-Pen]OH (DPDPE), and H-Tyr-c[D-Cys-Phe-D-Pen-OH]OH (JOM-13), as well a the deltorphins H-Tyr-D-Met-Phe-His-Leu-Met-Asp-NH2 (dermenkephalin), H-Tyr-DAla-Phe-Asp-Val-Val-Gly-NH2 (deltorphin I), and H-Tyr-D-Ala-PheGlu-Val-Val-Gly-NH2 (deltorphin II) (for reviews, see Refs. 20 and 21). However, these peptides are of relatively large molecular size and for this reason their ability to cross the BBB is very limited. Nonpeptide y agonists that were developed in the early to mid-1990s include the racemic compound BW373U86 [67] and its chemically modiﬁed enantiomer SNC80 [68], as well as the compound TAN-67 [69]. However, BW373U86 produced signiﬁcant toxicity, manifested behaviorally as convulsions and barrel rolling, in mice [70], and TAN-67 showed no signiﬁcant antinociceptive activity in the mouse tail ﬂick test [69]. Evidently, there is still a need for the development of new potent y opioid agonists of low molecular weight and high lipophilicity. In an eﬀort to increase the moderate y-agonist potency and the yreceptor selectivity of the dipeptide H-Tyr-Tic-NH-(CH2)2-Ph [61], structural modiﬁcations of the C-terminal phenylethyl group were performed by introduction of an additional substituent either in ortho position of the phenyl ring or at the h carbon [44] (Table 4). The analogue H-Tyr-Tic-NH(CH2)2-Ph(o-Cl) was a 10-fold more potent y agonist than the parent peptide in the MVD assay and was ﬁve times more y-receptor selective. Introduction of a second phenyl group at the h carbon of the phenylethylamine moiety led to the compound H-Tyr-Tic-NH-CH2-CH(Ph)2, with 20fold increased y-agonist potency and 2-fold improved y selectivity. The corresponding N-methylated analogue, Tyr(NMe)-Tic-NH-CH2-CH(Ph)2

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 4 In Vitro Opioid Activities of Dipeptide y-Opioid Agonists Compound

IC50 (nM)a

K yi (nM)b

K Ai (nM)b

K Ai /K yi

H-Tyr-Tic-NH-(CH2)2-Ph H-Tyr-Tic-NH-(CH2)2-Ph(o-Cl) H-Tyr-Tic-NH-CH2-CH(Ph)2 Tyr(NMe)-Tic-NH-CH2-CH(Ph)2 H-Tyr-Tic-NH-CH2-CH(Ph)COOEt (I) H-Tyr-Tic-NH-CH2-CH(Ph)COOEt (II) H-Hmt-Tic-NH-CH2-CH(Ph)2 DPDPE

82.0 8.77 3.77 0.261 1.28 8.64 0.630 5.30

5.22 1.43 0.981 0.581 0.569 3.03 2.00 16.4

69.1 96.9 28.8 12.7 886 153 1670 943

13.2 67.8 29.4 21.9 1560 50.5 835 57.5

a b

Determined in the MVD assay. Binding assay based on displacement of [3H]DSLET (y-selective) and [3H]DAMGO (A-selective) from rat brain membrane binding sites.

displayed subnanomolar y-agonist potency and still marked y-receptor selectivity. One of the isomers (I) of H-Tyr-Tic-NH-CH2-CH(Ph)COOEt was also found to be a potent y agonist with very high preference for y receptors over A receptors. An analogue containing 2V-hydroxy,6V-methyltyrosine (Hmt) in place of Tyr1, H-Hmt-Tic-NH-CH2-CH(Ph)2, turned out to be particularly remarkable because it showed both subnanomolar y agonist potency (IC50 = 0.630 nM) and very high y-receptor selectivity (K Ai /K yi = 835). In a direct comparison under identical assay conditions, this compound was 8 times more potent than the well-known y agonist DPDPE and 15 times more y selective. None of these compounds had signiﬁcant binding aﬃnity for n-opioid receptors. From these results it can be concluded that the dipeptide derivatives described here represent a new class of potent and selective y-opioid agonists. It is expected that these compounds should be able to cross the BBB to some extent because of their low molecular weight and high lipophilicity. Therefore, they have potential as centrally acting analgesics that may produce fewer side eﬀects than the currently used A type opiates.

VI. CONCLUSIONS Application of the concept of conformational restriction to opioid peptides has produced fruitful results, insofar as peptide analogues and mimetics with interesting opioid activity proﬁles and high stability against enzymatic

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

degradation were obtained. The conformationally restricted analogues that were developed were amenable to meaningful conformational analysis permitting the elaboration of models of the bioactive conformation at the A or y receptor. The multiple conformational restriction of dermorphin-related tetrapeptide analogues that was performed represents a rational design of opioid peptidomimetics characterized by a high degree of structural rigidiﬁcation. This is indicated by the fact that the A-selective agonist H-Hat-DOrn-Aic-Glu-NH2 contains only two freely rotatable bonds, whereas there are 14 freely rotatable bonds in [Leu5]enkephalin. The discovery of the TIP(P) peptides and their further structural modiﬁcation led to y opioid antagonists with unprecedented potency and selectivity. The observation that very subtle structural modiﬁcations of these ﬂexible and hydrophobic peptides in some cases converted a y antagonist into a y agonist and vice versa is most intriguing and unique in the peptide ﬁeld. This behavior may be explained with changes in the patterns of aromatic ring clustering in these hydrophobically collapsed molecules as a consequence of the minor structural alterations (introduction of a halogen substituent, peptide bond reduction, saturation of an aromatic ring, etc.) that were performed. The TIP(P) peptides are of therapeutic interest because y antagonists have been shown to attenuate the development of morphine tolerance and dependence [56,57] and to have an immunosuppressive eﬀect [71]. The three prototype mixed A agonist/y antagonists described in this chapter have excellent potential as analgesics with low propensity to produce tolerance and dependence. The pseudotetrapeptide DIPPNH2[C] has already been shown to produce a potent analgesic eﬀect, less tolerance than morphine, and no physical dependence upon chronic administration. In preliminary experiments, the tetrapeptides DIPP-NH2 and DIPP-NH2[C] were shown to cross the BBB to some extent, but further structural modiﬁcations need to be performed in order to improve the BBB penetration of these compounds. The Tyr-Tic dipeptide derivatives can also be expected to penetrate into the central nervous system because they are relatively small, lipophilic molecules. In this context, it is of interest to point out that the structurally related dipeptide H-Dmt-DAla-NH-(CH2)3-Ph (SC-39566), a plain A-opioid agonist, produced antinociception in the rat by subcutaneous and oral administration [72]. As indicated by the results of the NMR and molecular mechanics studies, the conformation of the cyclic h-casomorphin analogue H-Tyr-c[-D-Orn-2Nal-D-Pro-Gly-] is stabilized by intramolecular hydrogen bonds. Therej

j

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

fore, this mixed A agonist/y antagonist has a reduced capacity for intermolecular hydrogen bonding with water molecules and, consequently, should have a reasonable chance to cross the BBB as well. The dipeptide y agonists may turn out to be interesting pharmacological tools, since some of them are more potent and more selective than the y agonists currently in use. Furthermore, these compounds represent a new class of y agonists and have potential for pain treatment because they may also be small enough and lipophilic enough to cross the BBB and to produce a centrally mediated analgesic eﬀect.

ACKNOWLEDGMENTS The work described in this chapter was supported by operating grants from the Medical Research Council of Canada (MT-5655) and the National Institute on Drug Abuse (DA-04443).

REFERENCES 1. Schiller PW, DiMaio J. Opiate receptor subclasses diﬀer in their conformational requirements. Nature (London) 1982; 297:74–76. 2. Martin WR, Eades CJ, Thompson GA, Huppler RA, Gilbert PE. The eﬀects of morphine- and nalorphine-like drugs in the nondependent and morphinedependent chronic spinal dog. J Pharmacol Exp Ther 1976; 197:517–523. 3. Lord JAH, Waterﬁeld AA, Hughes, J, Kosterlitz HW. Endogenous opioid peptides: multiple agonists and receptors. Nature (London) 1977; 267: 495–499. 4. Clark JA, Liu L, Price M, Hersh B, Edelson M, Pasternak GW. Kappa opiate receptor multiplicity: evidence for two U50,488-sensitive n1 subtypes and a novel n3 subtype. J Pharmacol Exp Ther 1989; 251:461–468. 5. Jiang Q, Takemori AE, Sultana M, Portoghese PS, Bowen BD, Mosberg HI, Porreca F. Diﬀerential antagonism of opioid delta antinociception by [D-Ala2,Leu5,Cys6]enkephalin and naltrindole 5V-isothiocyanate: evidence for delta receptor subtypes. J Pharmacol Exp Ther 1991; 257:1069–1075. 6. Evans CJ, Keith Jr., DE, Morrison H, Magendzo K, Edwards RH. Cloning of a delta opioid receptor by functional expression. Science 1992; 258: 1952–1955. 7. Kieﬀer BL, Befort K, Gaveriaux-Ruﬀ C, Hirth CG. The y-opioid receptor: isolation of a cDNA by expression cloning and pharmacological characterization. Proc Natl Acad Sci USA 1992; 89:12048–12052.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

8. Yasuda K, Raynor K, Kong H, Breder CD, Takeda J, Reisine T, Bell GI. Cloning and functional comparison of n and y opioid receptors from mouse brain. Proc Natl Acad Sic USA 1993; 90:6736–6740. 9. Chen Y, Mestek A, Liu J, Hurley JA, Yu L. Molecular cloning and functional expression of a A-opioid receptor from rat brain. Mol Pharmacol 1993; 44:8–12. 10. Pogozheva ID, Lomize AL, Mosberg HI. Opioid receptor three-dimensional structures from distance geometry calculations with hydrogen bonding constraints. Biophys J 1998; 75:612–634. 11. Hughes J, Smith TW, Kosterlitz HW, Fothergill LA, Morgan BA, Morris RH. Identiﬁcation of two related pentapeptides from the brain with potent opiate agonist activity. Nature (London) 1975; 258:577–579. 12. Ho¨llt V. Opioid peptide processing and receptor selectivity, Annu Rev Pharmacol Toxicol 1986; 26:59–77. 13. Zadina JE, Hackler L, Ge LJ, Kastin AJ. A potent and selective endogenous agonist for the mu opiate receptor. Nature (London) 1997; 386: 499–502. 14. Henschen A, Lottspeich F, Brantl F, Teschemacher H. Novel opioid peptides derived from casein (h-casomorphins). Hoppe-Seyler’s Z. Physiol Chem 1979; 360:1217–1224. 15. Montecucchi PC, de Castiglione R, Piani S, Gozzini L, Erspamer V. Amino acid composition and sequence of dermorphin, a novel opiate-like peptide from the skin of Phyllomedusa sauvagei. Int J Peptide Protein Res 1981; 17:275–283. 16. Erspamer V, Melchiorri P, Falconieri-Erspamer G, et al. Deltorphins: a family of naturally occurring peptides with high aﬃnity and selectivity for y opioid binding sites. Proc Natl Acad Sci USA 1989; 86:5188–5192. 17. Leslie FM. Methods used for the study of opioid receptors. Pharmacol Rev 1987; 39:197–249. 18. Blume A. Interaction of ligands with the opiate receptors of brain membranes: regulation by ions and nucleotides. Proc Natl Acad Sci USA 1978; 75:1713–1717. 19. Zimmerman DM, Leander JD. Selective opioid receptor agonists and antagonists: research tools and potential therapeutic agents. J Med Chem 1990; 33:895–902. 20. Hruby VJ, Gehring CA. Recent developments in the design of receptor speciﬁc opioid peptides. Med Res Rev 1989; 9:343–401. 21. Schiller PW. Development of receptor speciﬁc opioid peptide analogs. In: Ellis GP, West BG, eds. Progress in Medicinal Chemistry. Vol. 28. Amsterdam: Elsevier, 1991:301–340. 22. Hruby VJ, Agnes RS. Conformation–activity relationships of opioid peptides with selective activities at opioid receptors. Biopolymers (Peptide Sci) 1999; 51:391–410. 23. DiMaio J, Schiller PW. A cyclic enkephalin analog with high in vitro opiate activity. Proc Natl Acad Sci USA 1980; 77:7162–7166.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

24.

25.

26.

27. 28.

29.

30.

31. 32.

33.

34.

35.

36.

37.

DiMaio J, Nguyen TM-D, Lemieux C, Schiller PW. Synthesis and pharmacological characterization in vitro of cyclic enkephalin analogs: eﬀect of conformational constraints on opiate receptor selectivity. J Med Chem 1982; 25:1432–1438. Berman JM, Goodman M, Nguyen TM-D, Schiller PW. Cyclic and acyclic partial retro-inverso enkephalins: mu receptor selective enzyme resistant analogs. Biochem Biophys Res Commun 1983; 115:864–870. Schiller PW, Nguyen TM-D, Maziak LA, Lemieux C. A novel cyclic opioid peptide analog showing high preference for A-receptors. Biochem Biophys Res Commun 1985; 127:558–564. Wilkes BC, Schiller PW. Theoretical conformational analysis of a A-selective cyclic opioid peptide analog. Biopolymers 1987; 26:1431–1444. Schiller PW, Weltrowska G, Nguyen TM-D, Lemieux C, Chung NN, Marsden BJ, Wilkes BC. Conformational restriction of the phenylalanine residue in a cyclic opioid peptide analogue: eﬀects on receptor selectivity and stereospeciﬁcity. J Med Chem 1991; 34:3125–3132. Wilkes BC, Schiller PW. Molecular dynamics simulations of opioid peptide analogs containing multiple conformational restrictions. Int J Peptide Protein Res 1992; 40:249–254. Schiller PW, Weltrowska G, Nguyen TM-D, Lemieux C, Chung NN. Conformational restriction of the Phe3 residue in a cyclic dermorphin analog: eﬀects on receptor selectivity and stereospeciﬁcity. In: Smith JA, Rivier JE, eds. Peptides: Chemistry and Biology (Proceedings of the 12th American Peptide Symposium). Leiden: The Netherlands: Escom Science Publishers, 1992:97–99. Wilkes BC, Schiller PW. Conformation–activity relationships of cyclic dermorphin analogues. Biopolymers 1990; 29:89–95. Yamazaki T, Ro S, Goodman M, Chung NN, Schiller PW. A topochemical approach to explain morphiceptin bioactivity. J Med Chem 1993; 36: 708–719. Cotton R, Giles MG, Miller L, Shaw JS, Timms, D. ICI 174864: a highly selective antagonist for the opioid y-receptor. Eur J Pharmacol 1984; 97: 331–332. Corbett AD, Gillan MGC, Kosterlitz HW, McKnight AT, Paterson SJ, Robson LE. Selectivities of opioid peptide analogues as agonists and antagonists at the y-receptor. Br J Pharmacol 1984; 83:271–279. Portoghese PS, Sultana M, Nagase H, Takemori AE. Application of the message–address concept in the design of highly potent and selective nonpeptide y opioid receptor antagonists. J Med Chem 1988; 31:281–282. Portoghese PS, Nagase H, Maloney Huss KE, Lin CE, Takemori AE. Role of spacer and address components in peptidomimetic y opioid receptor antagonists related to naltrindole. J Med Chem 1991; 34:1715–1720. Schiller PW, Nguyen TM-D, Weltrowska G, Wilkes BC, Marsden BJ, Lemieux C, Chung NN. Diﬀerential stereochemical requirements of A vs y

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

38.

39.

40.

41. 42.

43.

44.

45.

46.

47.

48.

opioid receptors for ligand binding and signal transduction: development of a class of potent and highly y-selective peptide antagonists. Proc Natl Acad Sci USA 1992; 89:11871–11875. Schiller PW, Nguyen TM-D, Weltrowska G, Wilkes BC, Marsden BJ, Schmidt R, Lemieux C, Chung NN. TIPP opioid peptides: development of extraordinarily potent and selective y antagonists and observation of astonishing structure–intrinsic activity relationships. In: Hodges RS, Smith RS, eds. Peptides: Chemistry, Structure and Biology (Proceedings of the 13th American Peptide Symposium). Leiden, The Netherlands: Escom Science Publishers, 1994:483–486. Marsden BJ, Nguyen TM-D, Schiller PW. Spontaneous degradation via diketopiperazine formation of peptides containing a tetrahydroisoquinoline3-carboxylic acid residue in the 2-position of the peptide sequence. Int J Peptide Protein Res 1993; 41:313–316. Schiller PW, Weltrowska G, Nguyen TM-D, Wilkes BC, Chung NN, Lemieux C. TIPP[C]: a highly potent and stable pseudopeptide y opioid receptor antagonist with extraordinary y selectivity. J Med Chem 1993; 36:3182–3187. Visconti, L.M., Standifer, K.M., Schiller, P.W., Pasternak, G.W. TIPP[C]: a highly selective y ligand. Neurosci Lett 1994; 181:47–49. Lee PHK, Nguyen TM-D, Chung NN, Schiller PW, Chang KJ. Tyrosineiodination converts the delta opioid peptide antagonist TIPP to an agonist. Eur J Pharmacol 1995; 280:211–214. Temussi PA, Salvadori S, Amodeo P, Bianchi C, Guerrini R, Tomatis R, Lazarus LH, Picone D, Tancredi T. Selective opioid dipeptides. Biochem Biophys Res Commun 1994; 198:933–939. Schiller PW, Weltrowska G, Schmidt R, Berezowska I, Nguyen TM-D, Lemieux C, Chung NN, Carpenter KA, Wilkes BC. Subtleties of structure– agonist versus antagonist relationships of opioid peptides and peptidomimetics. J Receptor Signal Transduction Res 1999; 19:573–588. Nevin ST, To´th G, Weltrowska G, Schiller PW, Borsodi A. Synthesis and binding characteristics of tritiated TIPP[C], a highly speciﬁc and stable delta opioid antagonist. Life Sci (Pharmacol Lett) 1995; 56:PL225– 230. Szatma´ri I, To´th G, Kerte´sz I, Schiller PW, Borsodi A. Synthesis and binding characteristics of the tritiated TIPP analogue TICP[C], a highly speciﬁc and stable y opioid antagonist. Peptides 1999; 20:1079–1082. Salvadori S, Attila M, Balboni G, Bianchi C, Bryant SD, Crescenzi O, Guerrini R, Picone D, Tancredi T, Temussi PA, Lazarus LH. y Opioidmimetic antagonists: prototypes for designing a new generation of ultraselective opioid peptides. Mol Med 1995; 1:678–689. Schiller PW, Schmidt R, Weltrowska G, Berezowska I., Nguyen TM-D, Dupuis S, Chung NN, Lemieux C, Wilkes BC, Carpenter KA. Conforma-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

49.

50.

51.

52.

53.

54.

55. 56.

57.

58.

59.

tionally constrained opioid peptide analogs with novel activity proﬁles. Lett Peptide Sci 1998; 5:209–214. Page´ D, McClory A, Mischki T, Schmidt R, Butterworth J, St-Onge S, Labarre M, Payza K, Brown W. Novel Dmt-Tic dipeptide analogues as selective delta-opioid receptor antagonists. Bioorg Med Chem Lett 2000; 10:167–170. Wilkes BC, Schiller PW. Theoretical conformational analysis of the opioid y antagonist H-Tyr-Tic-Phe-OH and the A agonist H-Tyr-D-Tic-Phe-NH2. Biopolymers 1994; 34:1213–1219. Schiller PW, Nguyen TM-D, Berezowska I, Weltrowska G, Schmidt R, Marsden BJ, Wilkes BC, Lemieux C, Chung NN. The TIPP opioid peptide family: development of a new class of highly potent y-receptor antagonists with extraordinary y-selectivity. In: Yanaihara N, ed. Peptide Chemistry 1992 (Proceedings of the 2nd Japanese Symposium on Peptide Chemistry. Leiden, The Netherlands: Escom Science Publishers, 1993:337–340. Wilkes BC, Schiller, PW. Comparative analysis of various proposed models of the receptor-bound conformation of TIP(P)-related y opioid antagonists. Biopolymers (Peptide Sci) 1995; 37:391–400. Wilkes BC, Nguyen TM-D, Weltrowska G, Carpenter KA, Lemieux C, Chung NN, Schiller PW. The receptor-bound conformation of H-Tyr-Tic(Phe-Phe)-OH related y-opioid antagonists contains all trans peptide bonds. J Peptide Res 1998; 51:386–394. Flippen-Anderson JL, George C, Deschamps JR, Reddy PA, Lewin AH, Brine GA. X-ray structures of the y opioid antagonist TIPP and a protected derivative of the y opioid antagonist ICI 174,864. Lett Peptide Sci 1994; 1:107–115. Wiley RA, Rich DH. Peptidomimetics derived from natural products. Med Res Rev 1993; 13:327–384. Abdelhamid EE, Sultana M, Portoghese PS, Takemori AE. Selective blockage of delta opioid receptors prevents the development of morphine tolerance and dependence in mice. J Pharmacol Exp Ther 1991; 258: 299–303. Fundytus ME, Schiller PW, Shapiro M, Weltrowska G, Coderre TJ. The highly selective y-opioid antagonist H-Tyr-TicC[CH2-NH]Phe-Phe-OH (TIPP[C]) attenuates morphine tolerance and dependence. Eur J Pharmacol 1995; 286:105–108. Zhu Y, King MA, Schuller AGP, Nitsche JF, Reidl M, Elde RP, Unterwald E, Pasternak GW, Pintar JE. Retention of supraspinal delta-like analgesia and loss of morphine tolerance in y opioid receptor knockout mice. Neuron 1999; 24:243–252. Schiller PW, Weltrowska G, Schmidt R, Nguyen, TM-D, Berezowska I, Lemieux C, Chung NN, Carpenter KA, Wilkes BC. Four diﬀerent types of opioid peptides with mixed A agonist/y antagonist properties. Analgesia 1995; 1:703–706.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

60. Schiller PW, Fundytus ME, Merovitz L, Weltrowska G, Nguyen TM-D, Lemieux C, Chung NN, Coderre TJ. The opioid A agonist/y antagonist DIPPNH2[c] produces a potent analgesic eﬀect, no physical dependence and less tolerance than morphine in rats. J Med Chem 1999; 42:3520–3526. 61. Schiller PW, Weltrowska G, Nguyen TM-D, Lemieux C, Chung NN. Novel opioid peptide analogs with mixed A agonist/y antagonist properties. In: Maia HLS, ed. Peptides 1994 (Proceedings of the 23rd European Peptide Symposium) Leiden, The Netherlands: Escom Science Publishers, 1995: 632–633. 62. Schmidt R, Vogel D, Mrestani-Klaus C, Brandt W, Neubert K, Chung NN, Lemieux C, Schiller PW. Cyclic h-casomorphin analogues with mixed A agonist/y antagonist properties: synthesis, pharmacological characterization and conformational aspects. J Med Chem 1994; 37:1136–1144. 63. Schmidt R, Chung NN, Lemieux C, Schiller PW. Development of cyclic casomorphin analogs with potent y antagonist and balanced mixed A agonist/y antagonist properties. In: Kaumaya PTP, Hodges RS, eds. Peptides: Chemistry, Structure and Biology (Proceedings of the 14th American Peptide Symposium). Leiden, The Netherlands: Escom Science Publishers, 1996:645-646. 64. Mrestani-Klaus C, Brandt W, Schmidt R, Neubert K, Schiller PW. Proton NMR conformational analysis of cyclic h-casomorphin analogues of the type Tyr-cyclo[-NN-D-Orn-Xaa-Gly-]. Arch Pharm Pharm Med Chem 1996; 329:133–142. 65. Cowan A, Zhu XZ, Mosberg HI, Omnaas JR, Porreca F. Direct dependence studies in rats with agents selective for diﬀerent types of opioid receptor. J Pharmacol Exp Ther 1988; 246:950–955. 66. Galligan JJ, Mosberg HI, Hurst R, Hruby VJ, Burks TF. Cerebral delta opioid receptors mediate analgesia but not the intestinal motility eﬀects of intracerebroventricularly administered opioids. J Pharmacol Exp Ther 1984; 229:641–648. 67. Chang KJ, Rigdon GC, Howard JL, McNutt RW. A novel, potent and selective nonpeptidic delta opioid receptor agonist. J Pharmacol Exp Ther 1993; 267:852–857. 68. Calderon SN, Rothman RB, Porreca F, Flippen-Anderson JL, McNutt RW, Xu H, Smith LE, Bilsky EJ, Davis P, Rice KC. Probes for narcotic receptor mediated phenomena. 19. Synthesis of (+)-4-[(aR)-a-((2S,5R)-4allyl-2,5-dimethyl-1-piperazinyl)-3-methoxybenzyl]-N,N-diethylbenzamide (SNC80): a highly selective, nonpeptide delta opioid receptor agonist. J Med Chem 1994; 37:2125–2128. 69. Kamei J, Saitoh A, Ohsawa M, Suzuki T, Misawa M, Nagase H, Kasuya Y. Antinociceptive eﬀects of the selective non-peptidic delta-opioid receptor agonist TAN-67 in diabetic mice. Eur J Pharmacol 1995; 276:131–135. 70. Comer SD, McNutt, RW, Chang, K-J, DeCosta BR, Mosberg HI, Woods

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

JH. Discriminative stimulus eﬀects of BW373U86: a nonpeptide ligand with selectivity for delta opioid receptors. J Pharmacol Exp Ther 1993; 267: 888–895. 71. Arakawa K, Akami T, Okamoto M, Akioka K, Nakai I, Oka T, Nagase H. Immunosuppression by delta opioid receptor antagonist. Transplant Proc 1993; 25:738–740. 72. Hammond DL, Stapelfeld A, Drower EJ, Savage MA, Tam L, Mazur RH. Antinociception produced by oral, subcutaneous or intrathecal administration of SC-39566, an opioid dipeptide arylalkylamide, in the rodent. J Pharmacol Exp Ther 1994; 268:607–615.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

7 Linkers and Resins for Solid-Phase Synthesis Pan Li and Elaine K. Kolaczkowski Vertex Pharmaceuticals, Inc., Cambridge, Massachusetts, U.S.A.

Steven A. Kates Surface Logix, Inc., Brighton, Massachusetts, U.S.A.

I.

INTRODUCTION

Organic chemistry in the last half of the twentieth century has evolved to a level of extreme sophistication in which complex macromolecules thought only to exist in nature were prepared in a laboratory hood. The process typically involves performing a reaction in an organic solvent followed by isolating, purifying, and analyzing the compound. This tedious, timeconsuming procedure requires considerable expertise. Bruce Merriﬁeld was the ﬁrst to recognize an alternative approach for the preparation of organic compounds. He applied this method to synthetic peptides and was awarded the Nobel Prize in 1984 for this discovery [1]. The concept was to perform the chemistry proven in solution but add a covalent attachment step that links the target to an insoluble polymeric support. Key advantages to the solid-phase technique include simple ﬁltration, washing without manipulative losses, and ease of automation. Peptide synthesis was amenable to solid-phase techniques since the process was repetitive. The C-terminal amino acid is attached to polymeric surface and the peptide chain is assembled via a two-step process: coupling of the incoming amino acid that has the alpha-amino group protected

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

and removal of this protecting group. Following chain elongation, the peptide is liberated from the solid support with concomitant release of the side-chain-protecting groups. Subsequent to Merriﬁeld’s discovery, other repetitive processes such as DNA synthesis and protein sequencing were adapted to solid-phase methods. Most peptides occur in nature as C-terminal acids or amides and methods were developed to release a peptide from a solid support to provide these two chemical functionalities. During the 1980s, research in peptide synthesis methods focused on releasing peptides from a polymeric support using various conditions (high acid concentration, low acid concentration, basic, etc.). During the early 1990s, there was a realization that solid-phase techniques could be applied to the construction of small molecules for drug discovery. Since many drugs do not contain carboxylic acids or amides, there was a need to expand the resulting chemical functionality at the anchoring position of the molecule following cleavage from solid support. Thus there has been a recent resurgence in research for solid-phase methods. A critical aspect to solid-phase synthesis is the anchoring of the molecule to the polymeric support. A solid support or resin is required to possess a functional group that is the starting point for the construction of the molecule. In addition, resins should possess the following properties: (1) mechanically robust; (2) stable to temperature variation; (3) good swelling in a broad range of solvents; (4) acceptable bead size and loadings; (5) stable with acidic, basic, reducing, and oxidizing conditions; (6) compatible with radical, carbene, carbanion, and carbenium ion chemistry; (7) biocompatible and swelling in aqueous buﬀers; (8) little nonspeciﬁc binding to biomolecules; (9) mobile, well-solvated, and reagent-accessible sites. The two most commonly used supports are polystyrene (PS) functionalized with a chloromethyl 1 (original Merriﬁeld polymer) or amino group 2 at the terminus (Fig. 1) [2]. For peptide synthesis, the cesium salt of a protected amino acid is anchored to chloromethyl-PS via a nucleophilic displacement of chlorine. Following chain elongation, the peptide is released from the support by treatment with a strong acid such as HF to provide a peptide acid. Amino acids anchored to aminomethyl-PS form an

Figure 1

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 2

amide bond and are stable to HF conditions. p-Methylbenzhydrylamine (MBHA, 3) is an amine functionalized polystyrene that liberates Cterminal amides upon exposure to HF (Fig. 2) [3]. A second strategy is to attach a linker (also referred to as a handle or anchor) to the resin followed by assembly of the molecule. A linker is bifunctional spacer that serves to link the initial synthetic unit to the support in two discrete steps (Fig. 3). To attach a linker to a chloromethylPS resin, a phenol functionality such as handle 4 is used to form an ether bond (Fig. 4). To attach the same handle to an amino-functionalized support, acetoxy function 5 or a longer methylene spacer of the corresponding phenol is applied to form an amide bond. Both of these resins perform similarly and only diﬀer in their initial starting resin [4]. An alternative approach is to prepare a preformed handle in which the ﬁrst building block is prederivatized to the linker and this moiety is attached to the resin. For peptide synthesis, this practice is common for the preparation of C-terminal peptide acids in order to reduce the amount of racemization of the a-carbon at the anchoring position [5]. There are three features of a linker that will determine which support is applicable to a synthetic scheme: (1) the functionality of the molecule at the anchoring position required for attachment; (2) cleavage conditions; and (3) the resulting functionality at the anchoring position of the molecule after the cleavage. As a continuing review on resins and linkers, this discussion will focus on the work that has been developed from 1997 to 1999 (Refs. 6–9 and references cited therein) and is described according to

Figure 3

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 4

the resulting functionality at the anchoring position after the cleavage. In addition, references and structures will be provided herein for the key linkers/resins that are commercially available or routinely used in solidphase peptide and organic synthesis.

II. RESINS AND LINKERS FOR CARBOXYLIC ACID GENERATION Most peptides contain a carboxylic acid or carboxamide at the Cterminus of the polymer chain. Since Merriﬁeld introduced solid-phase peptide synthesis in 1963, peptide chemists developed linkers that will yield these two functionalities upon release from the solid support using a variety of chemical conditions. The hydroxyl-containing resins (Fig. 5) based upon alkoxy-substituted benzyl alcohols were developed to supplement chloromethyl- and hydroxymethyl-PS. Carboxylic acids are esteriﬁed to the resin typically using N,NV-diisopropylcarbodiimide (DIPCDI). The acid strength required for release of the molecule from the solid support is related to the electron donor substituents on the benzene ring which stabilize the transient resin-bound carbocation. With greater resonance stabilization conveyed by additional alkoxy groups and aryl rings, milder acidic conditions are required for cleavage. PAM ( phydroxymethylphenylacetic acid) resin 6 [10] with no electron-donating groups requires HF treatment while Wang 7 [11], HMPA 8 (4-hydroxymethylphenylacetic acid, also referred to as PAB [ p-alkoxylbenzyl] or PAC [peptide acid]) [12] and DHPP 9 (4-(1V,1V-dimethyl-1V-hydroxypropyl)phenoxyacetyl) [13] contain resonance stabilizing groups and cleavage is aﬀected with triﬂuoroacetic acid (TFA). The hydroxymethylbenzyl

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 5

linker 10 (HMB) [14] contains a carbonyl group para to the ester anchor and is activated to nucleophilic attack such as hydroxide ion and is stable toward acid. Alkoxybenzyl derivatives with greater electron donor strength (Fig. 6) such as SASRIN (super-acid-sensitive resin) 11 [15], Rink acid 12 [16], and HAL (hyper-acid sensitive) 13 [17] resin allow carboxylic acids to be cleaved using a lower acid concentration (typically

Figure 6

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 7

1–10% TFA) in DCM (dichloromethane). Linkers for carboxylic acids have also been designed to eﬀect cleavage via photolysis (3-nitro-4hydroxymethylbenzoic acid, ONb 14) [18] and ﬂouridolysis (N-3 or 4) ((4-hydroxymethyl)-phenoxy-t-butylphenylsilyl)phenyl pentanedioic acid monoamide (PBs) 15 [19] and quinonemethide-based handle 16 [20] (Fig. 7). Fluorenone derived linker 17 prepared in two steps was coupled to aminomethyl-PS via DIPCDI [21]. Due to the presence of an electronwithdrawing carboxamide group, the release of carboxylic acids from this support requires strong acids, such as triﬂuoromethanesulfonic acid (TFMSA) (Scheme 1). Insertion of an oxygen adjacent to the biphenyl rings to the ﬂuorenone scaﬀold provides xanthene 18 handle [22]. The oxygen is strategically located to decrease the acid concentration required

Scheme 1

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 2

for cleavage and carboxylic acids are released using TFA (Scheme 2). Resin bound diazo linker 19 was synthesized starting from Wang resin and was further oxidized to a benzyl aldehyde (Scheme 3) [23]. Carboxylic acids are anchored to the support in a rapid, colorimetric reaction and are released upon TFA treatment. Photolabile linkers play an important role in solid-phase organic synthesis (SPOS) due to their stability under both acidic and basic conditions. The ONb photolabile linker was modiﬁed to improve cleavage rates and yields; Fmoc-Tos-OH was released in 87% yield after 23 h (Scheme 4) [24]. Speciﬁcally, the primary alcohol was changed to a secondary benzylic alcohol and the attachment to the resin was through an alkyl chain as opposed to an amide function. Linker 20 was used for the production of carboxylic acids or carbohydrates. A second example

Scheme 3

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 4

incorporated a dithiane function to serve as a safety catch against premature photoreaction [25–26]. A carboxylic acid functionality was coupled to linker 21 via DIPCDI, the dithiane protecting group was removed by an S-alkylating reagent such as methyl triﬂate, and release of the molecule was accomplished with UV irradiation in THF-methanol (Scheme 5). Based on 2-pivaloylglycerol, photolabile linker 22 was prepared in six steps from the dimer of 1,3-dihydroxyacetone (Scheme 6) [27]. The handle was attached to TentaGel S NH2 amino resin, the protecting groups from the hydroxyl functions were removed, and a series of peptides were assembled. Cleavage rates were reported to be faster than other photolabile linkers. Silyl-based linker 23, cleaved by either basic (TBAF) or neutral (CsF) ﬂuoridolysis to release carboxylic acids, alcohols, or amines, was prepared by treatment of a Grignard reagent to an aldehyde resin [28]. To demonstrate the utility of this handle, p-bromobenzoic acid was

Scheme 5

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 6

attached to the support and cleavage was accomplished in TBAF in DMF at 65jC or CsF in DMF at 90jC in 78% and 77% yield, respectively (Scheme 7). Redox-sensitive resin 24 designed for solid-phase peptide synthesis (SPPS) [29] was prepared from commercially available 2,5-dimethylbenzoquinone in seven steps [30] and loaded to a support via a Wittig reaction. Release of the peptide occurs using two sequential mild conditions, reduction with NaBH4 followed by TBAF-catalyzed cyclic ether formation (Scheme 8) which provide orthogonality to acid sensitive reactions. Allylic hydroxycrotyl-oligoethylene glyco-n-alkanoyl (HYCRON) linker 25 was applied to the synthesis of protected peptides and glycopeptides [31]. HYCRON is stable to both acidic and basic conditions and is compatible with Boc- and Fmoc-based chemistry. The preparation of this novel linker is only two steps from commercially available materials. HYCRON linker can be cleaved under neutral conditions using Pd catalyst (Scheme 9).

Scheme 7

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 8

III. RESINS AND LINKERS FOR GENERATION OF AMIDE FUNCTION Functionalized supports with amino groups such as benzhydrylamine (BHA) 26 [32] and 4-methylbenzhydrylamine (MBHA) 3 [3] provided Cterminal amides upon HF cleavage (Fig. 2). Polyalkoxyaminobenzyl and alkoxydiphenylamino resins such as PAL (5-(4-aminomethyl-3,5-dime-

Scheme 9

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 8

thoxyphenoxy)valeric acid) 27 [33], Rink amide support (RAM) 28 [34] and 4-(4V-methoxyvenzhydryl)phenoxyacetic acid (Dod) linker 29 [35] contain more electron-donating groups and were designed on the same principles as discussed above for the hydroxymethyl supports (Fig. 8). These three linkers are the most widely used in SPOS and require TFA for cleavage. Xanthone-based handles XAL (xanthenyl amide linder) 30 [36] and Seiber 31 [37] resin were designed to release amides using low concentrations of TFA (Fig. 9). Handles which contain an aminomethyl and o-nitrobenzyl function (Nb [nitrobenzyl] 32 [38], NBHA [nitrobenzylamine] 33 [39], and a-methyl-6-nitroveratrylamine) 34 [40] are cleaved by photolysis and are based upon the same principles discussed for hydroxyl resins (Fig. 10).

Figure 9

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 10

In an extension to the xanthenyl theme, the benzyl hydrogen was replaced with a substituted p-methoxyphenyl ring to give linker 35 (Scheme 10) [41]. Peptide amides were cleaved rapidly and in high purity with TFADCM (1:9) for 15 min or as a protected fragment with TFA-DCM (1:99) for 3–10 min. Silyl-derived linker 36 was prepared in three steps from a silyl ether of serine and incorporated for Fmoc/tBu-based assembly of protected glycopeptide blocks (Scheme 11) [42]. The a-carboxylic acid function of serine was protected as an allyl ester. Deprotection by a Pd(0) catalyst in the presence of dimedone liberated the carboxylic acid in order for subsequent

Scheme 10

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 11

coupling with amines, alcohols, and carbohydrates. The ﬁnal glycopeptide product was released from the support by ﬂuoridolysis (CsF). As an extension to the p-carboxybenzenesulfonamide ‘‘safety-catch’’ linker [43,44], alkanesulfonamide handle 37 was developed [45]. This linker tethers carboxylic acids to the solid support to give an acylated sulfonamide which is stable to both basic and acidic conditions (Scheme 12). Products were released by treatment with iodoacetonitrile followed by the addition of a nucleophile. Aryl hydrazide linker 38 stable to both acid and base was utilized in SPPS [46]. Treatment of the resin with a copper(II) catalyst in the presence of a base and nucleophile gave the corresponding acid, amide, or ester (Scheme 13).

Scheme 12

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 13

IV. RESINS AND LINKERS FOR N-SUBSTITUTED AMIDE GENERATION Linker 39 with an aldehyde attachment point permits amine anchoring via reductive amination (Fig. 11) [47]. In peptide synthesis, the handle attaches the amino as opposed to carboxylic acid function of the Cterminal residue to the support followed by chain elongation (attachment to the peptide occurs via a backbone nitrogen). The same strategy for developing handles functionalized with an aldehyde is similar to the concepts described above. Backbone amide linker (BAL) 40 was prepared from the Fmoc-based tris(alkoxy)benzylamide handle PAL [48]. In peptide synthesis, BAL allows the preparation of sequences having a variety of C-terminal functionalities such as alcohols, N-alkyl amides, and head-to-tail cyclic peptides that are devoid of a trifunctional amino acid. Due to the electron-donating groups contained in the handle, release of the peptide is accomplished with a high concentration of TFA. Based upon the BAL concept, the acid sensitive methoxybenzal-

Figure 11

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 14

dehyde polystyrene resin (AMEBA) 41 was reported for the solid-phase synthesis of sulfonamides, amides, ureas, and carbamates [49]. Reductive amination of aldehydes and ketones with sodium cyanoborohydride to Rink amide linker generated N-alkyl amines. Acylation followed by cleavage with TFA provided a method to generate a series of difunctional amines and N-substituted amide derivatives [50]. Subsequently, backbone linker 42 for Boc-based peptide was developed from a 4alkoxybenzyl derivative in which products were released upon HF treatment (Scheme 14) [51]. Contrary to an alkoxy benzene scaﬀold, secondary amides were generated via novel aldehyde linker 43 based upon an indole scaﬀold (Scheme 15) [52]. The indole resin was prepared from indole-3-carboxyaldehyde in two steps and reacted with amines under reductive conditions to generate resin-bound secondary amines. Treatment of the resin with

Scheme 15

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 16

various reagents followed by TFA yielded amides as well as sulfonamides, carbamates, and ureas in high yields. MAMP (Merriﬁeld, Alpha-MethoxyPhenyl) resin 44 is an alternative to aldehyde linkers to construct N-substituted amides [53]. Nucleophilic displacement of the benzylic chloride with an amine followed by acylation yielded a secondary amide which was released upon a low (f10%) concentration of TFA (Scheme 16).

V. RESINS AND LINKERS FOR HYDROXYL AND GENERATION OF AMINO FUNCTION Dihydropyran (DHP) linker 45 is a common handle that couples an alcohol to a solid support with subsequent release upon mild TFA conditions (Fig. 12) [54]. An alternative approach is to prepare an active carbonate linker. N,NV-Disuccinimidyl carbonate (DSC), a valuable reagent for converting hydroxymethyl-based supports to their corresponding carbonates, was reacted with 4-hydroxymethylpolystyrene 46 and 4-nitrobenzamido (Nbb) 47 resins to anchor alcohols and phenols (Scheme 17) [55]. The ﬁnal products were released from the solid support by HF and photolysis, respectively.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 12

p-Benzyloxybenzylamine (BOBA) 48 is a new class of an amine support and was prepared from Merriﬁeld resin in two steps [56]. BOBA resin was treated with an aldehyde in the presence of an acid to give an imine that subsequently reacted with Yb(OTf )3-catalyzed silyl enolates (Scheme 18). Cleavage with trimethylsilyl triﬂate (TMSOTf) or 2,3dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) gave either phenols or amines, respectively. 9-Phenylﬂuoren-9-yl polystyrene (Phﬂ) based resin 49 was applied in the solid-phase synthesis of hydroxyl and amino functions [57,58]. This resin has higher acid stability compared to the structurally similar trityl resin. Final release of the product is accomplished with TFA in high purity (Scheme 19). Trialkylsilane resin (PS-DES) 50 was incorporated for solid-phase glycosylation by anchoring a glycosyl donor via their corresponding thiophenyl ether or h-glucopyranosyl ﬂuorides (Scheme 20) [59]. Disaccharides

Scheme 17

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 18

were prepared by reaction with a glycosyl acceptor followed by cleavage with acetic acid (AcOH). Phenols were constructed form novel serine-derived handle 51, which was stable to acids (TFA) and bases (pyridine) (Scheme 21) [60]. The ﬁnal products were released from the support by ﬂuoride ion. A variety of cleavage conditions have been reported for the release of amines from a solid support. Triazene linker 52 prepared from Merriﬁeld resin in three steps was used for the solid-phase synthesis of aliphatic amines (Scheme 22) [61]. The triazenes were stable to basic conditions and the amino products were released in high yields upon treatment with mild acids. Alternatively, base labile linker 53 synthesized from a-bromo-p-toluic acid in two steps was used to anchor amino functions (Scheme 23) [62]. Cleavage was accomplished by oxidation of the thioether to the sulfone with m-chloroperbenzoic acid followed by helimination with a 10% solution of NH4OH in 2,2,2-triﬂuoroethanol. A linker based on 1-(4,4V-dimethyl-2,6-dioxocyclohexylidene)ethyl (Dde) primary amine protecting group was developed for attaching amino functions (Scheme 24) [65]. Linker 54 was stable to both acidic and basic conditions and the ﬁnal products were cleaved from the resin by treatment with hydrazine or transamination with n-propylamine.

Scheme 19

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 20

REM linker 55 (regenerated after cleavage, functionalized by Michael addition) is a traceless handle for anchoring secondary amines (Fig. 13) [64]. Tertiary amines were prepared on this linker via basedinduced Hofmann elimination of the subsequent quaternary ammonium salt. An analogous vinyl sulfone linker was prepared from Merriﬁeld resin to perform the identical synthetic strategy (Scheme 25) [65]. Similar to REM, vinyl sulfone handle 56 was regenerated following cleavage, but was more stable to acids and nucleophiles such as Grignard reagents than the former. An extension to the vinyl sulfone theme was demonstrated by inserting a carbamate function at the anchoring position for the assembly of 2- and 2,4-substituted pyrrolidines (Scheme 26) [66]. The acid stable, base labile (ﬁnal cleavage accomplished with NaOMe) support 57 was used for N-acyliminium ion reactions.

Scheme 21

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 22

Scheme 23

Scheme 24

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 13

Scheme 25

Scheme 26

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 27

N-Protected amines were assembled on solid-phase via sulfonamidebased handle 58 (Scheme 27) [67]. Tertiary sulfonamides were generated upon reaction with allylic, benzylic and primary alcohols under Mitsunobu conditions. Secondary amines were released from the support using mild nucleophilic conditions such as treatment with thiophenol and potassium carbonate. A versatile approach for the solid-phase synthesis of aminopyridazines used the anchoring of 3,6-dichloropyridazine to resin-bound thiophenol 59 (Scheme 28) [68]. Treatment with nucleophilic amines released the aminopyridazine products from the solid support without further oxidation. Traceless linker 60 based on a benzotriazole scaﬀold was reacted with amines and aldehydes to produce Mannich-type amine products [69]. Final product release was achieved by treatment with Grignard reagents (Scheme 29).

Scheme 28

Scheme 29

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 14

VI. RESINS AND LINKERS FOR HYDROXAMIC GENERATION OF ACID FUNCTIONS Hydroxamic acids are an important class of compounds targeted as potential therapeutic agents. N-Fmoc-aminooxy-2-chlorotrityl polystyrene resin 61 allowed the synthesis and subsequent cleavage under mild conditions of both peptidyl and small molecule hydroxamic acids (Fig. 14) [70]. An alternative hydroxylamine linkage 62 was prepared from trityl chloride resin and N-hydroxyphthalimide followed by treatment with hydrazine at room temperature (Scheme 30) [71]. A series of hydroxamic acids were prepared by the addition of substituted succinic anhydrides to the resin followed by coupling with a variety of amines, and cleavage with HCOOH-THF(1:3).

Scheme 30

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

VII. RESINS AND LINKERS FOR GENERATION OF SULFONAMIDE, UREA, AND GUANIDINE FUNCTIONS Aminosulfonyl ureas were constructed from a sulfonylcarbamate linkage (Scheme 31) [72]. Reaction of chlorosulfonyl isocyanate (CSI) with Wang resin provided a chlorosulfonylcarbamate 63 which was then converted to substituted amino sulfonylcarbamate compounds by reaction with excess amines. The ﬁnal aminosulfonyl urea products were cleaved from the resin by treatment with amines in HF at reﬂux temperature for overnight. Urea libraries were assembled via thiophenoxy carbonyl linker 64 readily available in two steps from Merriﬁeld resin (Scheme 32) [73]. Treatment of this linker with primary or secondary amines, followed by basic cleavage with amines generated the ureas. An alternative approach for the synthesis of ureas was to treat p-nitrobenzophenone oxime resin with phosgene to give p-nitrophenyl(polystyrene)ketoxime (Phoxime resin) 65 [74]. The addition of primary amines to the phosgenated oxime linker gave a resin-bound carbamate. Ureas were genated by reaction with a second set of amines at temperatures greater than 80jC (Scheme 33). Traditional SPPS anchors the peptide to the support via the acarboxylic acid of the C-terminal residue. Novel sulfonyl linker 66 was prepared to side-chain anchor the guanidine function of arginine (Scheme 34) [75]. To demonstrate the utility of the linker, tripeptide H-Phe-Arg-Ala-OMe was assembled in which amino acids were extended to the anchoring residue in both the C- and N-terminal directions. HF cleavage released the peptide from the support. Small molecules containing guanidines were constructed from carbonylimidazole handle 67 generated from Wang resin (Scheme 35) [76]. Treatment of the carbonylimidazole linker with thiourea basic conditions aﬀorded

Scheme 31

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 32

Scheme 33

Scheme 34

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 35

a resin-linked thiourea product. Sulfur displacement of the thiourea resin with primary and secondary amines followed by TFA cleavage provided guanidine-containing products. Both mono- and disubstituted guanidines were prepared in good yields and purities using acyl isothiocyanate resin 68 prepared from carboxypolystyrene in two simple steps (Scheme 36) [77]. Reaction of a variety of amines with this resin produced the corresponding acyl thioureas under mild conditions. The guanidine moiety formation was achieved by exposing the acyl thiourea resin to a primary or secondary amine. Cleavage of the acyl guanidine was eﬀected by treatment with TFA-CHCl3-MeOH (1:1:1).

VIII. RESINS AND LINKERS FOR GENERATION OF ALDEHYDE FUNCTIONS The Leznoﬀ acetal linker 69 was used to anchor an aldehyde to the solid support and following a series of reactions, the aldehyde was released by acidic cleavage [78]. An application of this resin was demonstrated for a biaryl aldehyde library synthesis which incorporated a Suzuki–Miyaura reaction (Scheme 37) [79]. Cleavage was eﬀected by a solution of 3 M HCl

Scheme 36

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 37

at 80jC to give excellent yields for most of the products. An alternative strategy implemented serine or threonine as a linker to anchor an aldehyde to the solid support [80]. Unlike the acetal formation described above, this linker reacts with an aldehyde to form an oxazoline with release from the support by aqueous acid (such as HOAc) at 60jC. A second strategy employed tartaric acid-based linker 70 prepared from an amino PEGA resin in which C-terminal a-oxo-aldehydes were generated by an oxidative cleavage (Scheme 38) [81]. Following linear assembly of the peptide by Fmoc chemistry, TFA treatment removed the side-chain protecting groups and converted the anchoring acetonide to a 1,2-diol which was oxidized to the aldehyde with NaIO4.

IX. RESINS AND LINKERS FOR GENERATION OF OTHER FUNCTIONS Cleavage of all the linkers described above provide a functional group (carboxylic acid, amide, amine, etc) at the anchoring position. Silyl-based handles 71,72, and 73 as well as germanium-based handle 74 insert a C-H bond at the anchoring position and are referred to as traceless (Fig. 15) [82–

Scheme 38

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 15

85]. A further extension to this concept was (dimethylsilyl)propionic acid linker 75 used for the solid-phase synthesis of aryl-containing organic compounds [86]. The linker was cleaved smoothly with TFA and has been used for the synthesis of compounds which involved alkylation, acylation, and Mitsunobu reactions. Silicon linker 76 was used for direct loading of aromatic compounds to supports for the assembly of pyridine-based tricyclics (Scheme 39) [87]. Following the initial coupling of an aromatic bromide to the resin by halogen/metal exchange in the presence of tert-butyl lithium, a

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 39

series of reactions including TFA deprotection of Boc, alkylation under strong basic condition, SnCl2 reduction, and ring cyclization were performed. The ﬁnal tricyclic products were released from the polymer via basic ﬂuoride (Bu4N+-F) in THF at room temperature. A similar trialkylsilane linker was synthesized from Merriﬁeld resin in two steps [88]. Piperazine linker 77 was treated with propargyltriphenylphosphine bromide to provide a resin-bound Wittig reagent (Scheme 40) [89]. Base treatment followed by aldehyde addition produced a resin-bound 2-aminobutadiene which was implemented in [4+2] cycloadditions. Alternatively, treatment with 3% TFA in CH2Cl2 released a,h-unsaturated methylketones in high yields.

Scheme 40

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 41

Acetal handle 78 synthesized from Merriﬁeld resin and 4-hydroxybenzaldehyde was applied to the solid-phase synthesis of carbohydrates and 1-oxacephams (Scheme 41) [90]. For the latter, a 1,3-diol was initially anchored to the support to form a cyclic acetal. A ring opening reaction with DIBAL generated a resin-bound alcohol which was converted to the corresponding triﬂate for N-alkylation with 4-vinyloxyazetidin-2-one. A Lewis acid catalyzed ring closure released 1-oxacephams from the support. Aryl hydrazide-based linker 79 was developed as a traceless handle that released products under mild oxidative conditions (Scheme 42) [91]. Polymeric bound p–iodophenylhydrazide was subjected to a variety of Pd0-catalyzed coupling reactions (Heck, Suzuki, Sonogashira, and Stille). Oxidation with Cu(OAc)2 in MeOH and pyridine released the ﬁnal products in 50–96% yield. A traceless linker for solid-phase homo- and hetero-Diels-Alder reactions was based upon resin bound quinodimethane precursors

Scheme 42

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Scheme 43

(Scheme 43) [92]. Reaction of dienophiles such as 4-nitrobenzaldehyde with linker 80 at high temperature gave Diels-Alder products. Dihydropyrans were released from the support by Bronsted or Lewis acid-nucleophile combinations in moderate to good yield with stereoselectivity for the anti isomer.

X. CONCLUSION The past decade witnessed a renaissance in drug discovery due to the emergence of solid-phase synthesis. Initially, the solid supports and linkers used for the repetitive process of biomolecule assembly applied to the construction of small molecule libraries and scaﬀolds were required to contain a carboxylic acid or amide in order to anchor to the polymeric support. Thus, the linkers from solid-phase peptide synthesis such as Rink, Wang, and PAL were commonly employed in the synthetic strategy. As new bond-forming reactions have been adapted for solid phase as well as the construction of novel lead compounds, synthetic pathways are requiring additional handles that release compounds into solution upon various cleavage conditions and provide additional functionality at the anchoring position. In a retrosynthetic analysis of a library, one should plan the anchoring strategy as it relates to the functionality of the molecule as well as to insure that the cleavage conditions are compatible with the synthetic scheme. Although there are now a plethora of linkers that have been described in the literature, novel handles still provide medicinal chemists the tools to expand molecular diversity with the ultimate reward of discovering a new drug candidate.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Appendix A. HANDLES AND DERIVATIZED SOLID SUPPORTS

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

REFERENCES 1. 2.

3. 4.

5.

6. 7. 8. 9. 10.

11.

12. 13.

14.

15.

Merriﬁeld RB. Solid phase peptide synthesis I. The synthesis of a tetrapeptide. J Am Chem Soc 1963; 85:2149–2154. Adams JH, Cook RM, Hudson D, Jammalamadaka V, Lyttle MH, Songster MF. A reinvestigation of the preparation, properties, and applications of aminomethyl and 4-methylbenzhydrylamine polystyrene resins. J Org Chem 1998; 63:3706–3716. Matsueda GR, Stewart JM. A p-methylbenzhydrylamine resin for improved solid-phase synthesis of peptides amides. Peptides 1981; 2:45–50. Albericio F, Barany G. Improved approach for anchoring N a-9-ﬂuorenylmethyloxycarbonylamino acids as p-alkoxybenzyl ester anchoring linkages. Int J Peptide Protein Res 1985; 26:92–97. Bernatowicz MS, Kearney T, Neves RS, Koster H. An eﬃcient method for racemization-free attachment of 9-ﬂuorenylmethyloxycarbonyl-amino acids to peptide synthesis supports. Tetrahedron Lett 1989; 30:4341–4344. Blackburn C, Albericio F, Kates SA. Recent Res Devel Org Chem 1997; 1:477–497. Blackburn C, Albericio F, Kates SA. Drugs Future 1997; 22:1007–1025. Resins and handles for solid-phase synthesis. In: Kates SA, ed. Biopolymers: Peptide Science. 1998; 47:309–411. Kates SA, Albericio F, eds. Solid-Phase Synthesis: A Practical Guide. New York: Marcel Dekker, 2000. Mitchell AR, Kent SBH, Engelhard M, Merriﬁeld R. A new synthetic route to tert-butyloxycarbonylaminoacyl-4-oxymethyl)phenylacetamidomethylresin, an improved support for solid-phase peptide synthesis. J Org Chem 1978; 43:2845–2852. Wang SS. p-Alkoxybenzyl alcohol resin and p-alkoxybenzyloxycarbonylhydrazide resin for solid-phase synthesis of protected peptide fragments. J Am Chem Soc 1973; 95:1328–1333. Sheppard RC, Williams BJ. Acid-labile resin linkage agents for use in solid phase peptide synthesis. Int J Peptide Protein Res 1982; 20:451–454. Akaji K, Kiso Y, Carpino LA. Fmoc-based solid-phase peptide synthesis using a new t-alcohol type 4-(1V,1V-dimethyl-1V-hydroxypropyl)phenoxyacetyl handle (DHPP)-resin (Fmoc=9-ﬂuorenyloxycarbonyl). J Chem Soc Chem Comm 1990; 584–586. Atherton E, Logan CJ, Sheppard RC, Peptide synthesis, part 2: Procedures for solid phase synthesis using Na-ﬂuorenylmethoxycarbonylamino acids on polyamide supports: Synthesis of substance P and of acyl carrier protein 64–74 decapeptide. J Chem Soc Perkin Trans 1 1981; 538–546. Mergler M, Tanner R, Gosteli J, Grogg P. Peptide synthesis by a combination of solid-phase and solution methods. I: A new very acid-labile

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

16. 17.

18.

19.

20.

21. 22.

23. 24.

25.

26.

27.

28.

29.

30.

anchor group for the solid-phase synthesis of fully protected peptide fragments. Tetrahedron Lett 1988; 29:4005–4008. Rink H. Solid-phase synthesis of protected peptide fragments using a trialkoxy diphenyl methylester resin. Tetrahedron Lett 1987; 28:3787–3790. Albericio F, Barany G. Hypersensitive acid-labile (HAL) tris(alkoxy)benzyl ester anchoring for solid-phase synthesis of protected peptide segments. Tetrahedron Lett 1991; 32:1015–1018. Rich DH, Gurwara SK. Preparation of a new o-nitrobenzyl resin for solidphase synthesis of tert-butyloxycarbonyl-protected peptide acids. J Am Chem Soc 1975; 97:1575–1579. Mullen DG, Barany G. A new ﬂuoridolyzable linkage for orthogonal solidphase peptide synthesis: Design, preparation, and application of the N-(3 or 4)((4-hydroxymethyl)-phenoxy-tbutylphenylsilyl)phenyl pentanedioic acid monoamine (Pbs) handle. J Org Chem 1988; 53: 5240–5248. Ramage R, Barron CA, Bielecki S, Holden R, Thomas DW. Solid phase peptide synthesis: Fluoride ion release of peptide from the resin. Tetrahedron Lett 1987; 28:4105–4108. Henkel B, Bayer E. 9-Hydroxy-9-(4-carboxyphenyl)ﬂuorene—a new linker for solid phase synthesis. Tetrahedron Lett 1998; 39:9401–9402. Henkel B, Zeng W, Bayer E. 9-Hydroxy-9-(4-carboxyphenyl)xanthene- a new linker for the synthesis of peptide amides. Tetrahedron Lett 1997; 38:3511–3512. Bhalay G, Dunstan AR. Facile solid phase synthesis of an activated diazo linker. Tetrahedron Lett 1998; 39:7803–7806. Rodebaugh R, Fraser-Reid B, Geysen HM. A new o-nitrobenzyl photocleavable linker for solid phase synthesis. Tetrahedron Lett 1997; 38: 7653– 7656. Routledge A, Abell C, Balasubramanian S. The use of a dithiane-protected benzoin photolabile safety catch linker for solid-phase synthesis. Tetrahedron Lett 1997; 38:1227–1230. Lee HB, Balasubramanian S. Studies on a dithiane-protected benzoin photolabile safety catch linker for solid-phase synthesis. J Org Chem 1999; 64:3454–3460. Peukert S, Giese B. The pivaloyglycol anchor group: A new platform for a photolabile linker in solid-phase synthesis. J Org Chem 1998; 63: 9045–9051. Routledge A, Stock HT, Flitsch SL, Turner NJ. New ﬂuoride-labile linkers for solid-phase organic synthesis. Tetrahedron Lett 1997; 38: 8287–8290. Zheng A, Shan D, Shi X, Wang B. A model resin linker for solid-phase peptide synthesis which can be cleaved using two sequential mild reactions. J Org Chem 1999; 64:7459–7466. Zheng A, Shan D, Wang B. A redox-sensitive resin linker for the solid

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

31.

32. 33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

phase synthesis of C-terminal modiﬁed peptides. J Org Chem 1999; 64: 156–161. Seitz O, Kunz H. HYCRON, an allylic anchor for high-eﬃciency solid phase synthesis of protected peptides and glycopeptides. J Org Chem 1997; 62:813–826. Pietta PG, Marshall GR. Amide protection and amide supports in a solidphase peptide synthesis. J Chem Soc D 1970; 650–651. Albericio F, Kneib-Cordonier N, Biancalana S, Gera L, Masada RI, Hudson D, Barany G. Preparation and application of the 5-(4-(9ﬂuorenylmethoxycarbonyl)aminomethyl-3, 5-dimethoxyphenoxy)-valeric acid (PAL) handle for the solid-phase synthesis of C-terminal peptide amides under mild conditions. J Org Chem 1990; 55:3730–3743. Rink H. Solid-phase synthesis of protected peptide fragments using a trialkoxy diphenyl methylester resin. Tetrahedron Lett 1987; 28:3787– 3790. Breipohl G, Knolle J, Stuber W. Synthesis and application of acid labile anchor groups for the synthesis of peptide amides by Fmoc-solid-phase peptide synthesis. Int J Peptide Protein Res 1989; 34:262–267. Han Y, Bontems SL, Hegyes P, Munson MC, Minor CA, Kates SA, Albericio F, Barany G. Preparation and applications of xanthenylamide (XAL) handles for solid-phase synthesis of C-terminal peptide amides under particularly mild conditions. J Org Chem 1996; 61:6326. Seiber P. A new acid-labile anchor groups for the solid phase synthesis of C-terminal peptide amides by the Fmoc method. Tetrahedron Lett 1987; 28:2107–2110. Hammer RP, Albericio F, Gera L, Barany G. Practical approach to solid-phase synthesis of C-terminal peptide amides under mild conditions based on photolysable anchoring linkage. Int J Peptide Protein Res 1990; 36:31–45. Ajayaghosh A, Pillai VNR. Solid-phase synthesis and C-terminal amidation of peptides using a photolabile o-nitrobenzhydrylaminopolystyrene support. Tetrahedron Lett 1995; 36:777–780. Holmes CP, Jones DG. Reagents for combinatorial organic synthesis: Development of a new o-nitrobenzyl photolabile linker for solid-phase synthesis. J Org Chem 1995; 60:2318–2319. Meisenbach M, Echner H, Voelter W. New methoxy-substituted 9-phenylxanthen-9-ylamine linkers for the solid phase phase synthesis of protected peptide amides. Chem Commun (Cambridge) 1997; 849–850. Nakamura K, Hanai N, Kanno M, Kobayashi A, Ohnishi Y, Ito Y, Nakahara Y. Design and synthesis of silyl ether-based linker for solid-phase synthesis of glycopeptides. Tetrahedron Lett 1999; 40:515–518. Kenner GW, McDermott JR, Sheppard RC. The safety catch principle in solid-phase peptide synthesis. J Chem Soc Chem Comm 1971; 636–637.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

44. Backes BJ, Ellman JA. Carbon-carbon bond-forming methods on solid support. Utilization of Kenner’s ‘‘safety-catch’’ linker. J Am Chem Soc 1994; 116:11171–11172. 45. Backes BJ, Ellman JA. An alkanesulfonamide ‘‘safety-catch’’ linker for solid-phase synthesis. J Org Chem 1999; 64:2322–2330. 46. Millington CR, Quarrell R, Lowe G. Aryl hydrazides as linkers for solid phase synthesis which are cleavable under mild oxidative conditions. Tetrahedron Lett 1998; 39:7201–7204. 47. Boojmara CG, Burow K, Ellman JA. An expedient and high-yielding method for the solid-phase synthesis of diverse 1,4-benzodiazepine-2,5diones. J Org Chem 1995; 60:5742–5743. 48. Jensen KJ, Alsina J, Songster MF, Vagner J, Albericio F, Barany G. Backbone amide linker (BAL) strategy for solid-phase synthesis of Cterminal-modiﬁed and cyclic peptides. J Am Chem Soc 1998; 120:5441– 5452. 49. Fivush AM, Willson TM. AMEBA: An acid sensitive aldehyde resin for solid phase synthesis. Tetrahedron Lett 1997; 38:7151–7154. 50. Brown EG, Nuss JM. Alkylation of Rink’s amide linker on polystyrene resin: a reductive amination approach to modiﬁed amine-linkers for the solid phase synthesis of N-substituted amide derivatives. Tetrahedron Lett 1997; 38:8457–8460. 51. Bourne GT, Meutermans WDF, Alewood PF, McGeary RP, Scanlon M, Watson AA, Smythe ML. A backbone linker for Boc-based peptide synthesis and on-resin cyclization: Synthesis of stylostatin 1. J Org Chem 1999; 64:3095–3101. 52. Estep KG, Neipp CE, Stramiello LM, Stephens A, Mavis D, Allen MP. Indole resin: A versatile new support for the solid-phase synthesis of organic molecules. J Org Chem 1998; 63:5300–5301. 53. Brown DS, Revill JM, Shute RE. Merriﬁeld Alpha-Methoxyphenyl (MAMP) Resin; A new versatile solid support for the synthesis of secondary amides. Tetrahedron Lett 1998; 39:8533–8536. 54. Thompson LA, Ellman JA. Straightforward and general method for coupling alcohols to solid supports. Tetrahedron Lett 1994; 35:9333– 9336. 55. Alsina J, Chiva C, Ortiz M, Rabanal F. Active carbonate resins for solidphase synthesis through the anchoring of a hydroxyl function. Synthesis of cyclic and alcohol peptides. Tetrahedron Lett 1997; 38:883–886 and references cited therein. 56. Kobayashi S, Aoki Y. p-Benzyloxybenzylamine (BOBA) resin. A new polymer-supported amine used in solid-phase organic synthesis. Tetrahedron Lett 1998; 39:7345–7348. 57. Bleicher KH, Wareing JR. Phﬂ Polystyrene: A new resin for solid phase organic synthesis. Tetrahedron Lett 1998; 39:4587–4590.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

58. Bleicher KH, Wareing JR. Phﬂ Acetic Acid: A new linker for solid phase organic synthesis. Tetrahedron Lett 1998; 39:4591–4594. 59. Doi T, Sugiki M, Yamada H, Takahashi T, Porco JA, Jr. Glycosidation of solid-supported glycosyl donors tethered by a trialkylsilane linker. Tetrahedron Lett 1999; 40:2141–2144. 60. Chou Y-L, Morrissey MM, Mohan R. Novel serine-based linker for the solid-phase synthesis of organic compounds. Tetrahedron Lett 1998; 39:757–760. 61. Brase S, Kobberling J, Enders D, Lazny R, Wang M, Brandtner S. Nitrogen-based linker. 3. Triazenes as robust and simple linkers for amines in solid-phase organic synthesis. Tetrahedron Lett 1999; 40:2105– 2108. 62. Garcia-Echeverria C. A base labile handle for solid phase chemistry. Tetrahedron Lett 1997; 38:8933–8934. 63. Chhabra SR, Khan AN, Bycroft BW. Versatile Dde-based primary amine linkers for solid phase synthesis. Tetrahedron Lett 1998; 39:3585–3588. 64. Gayo KM, Suto MJ. Traceless linker: Oxidative activation and displacement of a sulfur-based linker. Tetrahedron Lett 1997; 38:211–214. 65. Kroll FEK, Morphy R, Rees D, Gani D. Resin-immobilized benzyl and aryl vinyl sulfones: New versatile traceless linkers for solid-phase organic synthesis. Tetrahedron Lett 1997; 38:8573–8576. 66. Veerman JJN, Rutjes FPJT, Van Maarseveen JH, Hiemstra H. A novel acid stable/base labile carbamate linker for N-acyliminium ion reactions on solid support. Tetrahedron Lett 1999; 40:6079–6082. 67. Kay C, Murray PJ, Sandow L, Holmes AB. A novel, chemically robust, amine-releasing linker. Tetrahedron Lett 1997; 38:6941–6944. 68. Parrot I, Wermuth C-G, Hibert M. Resin-bound thiophenols as SNARlabile linkers: Application to the solid phase synthesis of aminopyridazines. Tetrahedron Lett 1999; 40:7975–7978. 69. Schiemann K, Showalter HDH. Development of polymer-supported benzotriazole as a novel traceless linker for solid-phase organic synthesis. J Org Chem 1999; 64:4972–4975. 70. Mellor SL, McGuire C, Chan WC. N-Fmoc-aminooxy-2-chlorotrityl polystyrene resin: A facile solid-phase methodology for the synthesis of hydroxamic acids. Tetrahedron Lett 1997; 38:3311–3314. 71. Bauer U, Ho W-B, Koskinen AMP. A novel linkage for the solid-phase synthesis of hydroxamic acids. Tetrahedron Lett 1997; 38:7233–7236. 72. Fitzpatrick LJ, Rivero RA. Solid phase synthesis of substituted aminosulfonylureas using a sulfonylcarbamate linker. Tetrahedron Lett 1997; 38:7479–7482. 73. Dressman BA, Singh U, Kaldor SW. Solid phase synthesis of urea libraries using a diversiﬁable thiophenoxy carbonyl linker. Tetrahedron Lett 1998; 39:3631–3634.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

74. Scialdone MA, Shuey SW, Soper P, Hamuro Y, Burns DM. Phosgenated pnitrophenyl(polystyrene)ketoxime or phoxime resin. A new resin for the solid-phase synthesis of ureas via thermolytic cleavage of oxime-carbamates. J Org Chem 1998; 63:4802–4807. 75. Zhong HM, Greco MN, Maryanoﬀ BE. Solid-phase synthesis of argininecontaining peptides by guanidine attachment to a sulfonyl linker. J Org Chem 1997; 62:9326–9330. 76. Josey JA, Tarlton CA, Payne CE. Novel linker for the solid-phase synthesis of guanidines. Tetrahedron Lett 1998; 39:5899–5902. 77. Wilson LJ, Klopfenstein SR, Li M. A traceless linker approach to the solid phase synthesis of substituted guanidines utilizing a novel acyl isothiocyanate resin. Tetrahedron Lett 1999; 40:3999–4002. 78. Leznoﬀ CC, Wong JY. The use of polymer supports in organic synthesis. III. Selective chemical reactions on one aldehyde group of symmetrical aldehydes. Can J Chem 1973, 51:3756–3764. 79. Chamoin S, Houldsworth S, Kruse CG, Bakker WI, Snieckus V. The Suzuki-Miyaura cross coupling reactions on solid support. Link to solution phase directed ortho metalation. The Leznoﬀ acetal linker approach to biaryl and heterobiaryl aldehydes. Tetrahedron Lett 1998; 39: 4179–4182. 80. Ede NJ, Bray AM. A simple linker for the attachment of aldehydes to the solid phase. Application to solid phase synthesis by the Multipin method. Tetrahedron Lett 1997; 38:7119–7122. 81. Fruchart J-S, Gras-Masse H, Melnyk O. A new linker for the synthesis of C-terminal peptide a-oxo-aldehydes. Tetrahedron Lett 1999; 40:6225–6228. 82. Chenera B, Finkelstein JA, Veber DF. Protodetachable arylsilane polymer linkages for use in solid phase organic synthesis. J Am Chem Soc 1995; 117:11999–12000. 83. Plunkett MJ, Ellman JA. A silicon-based linker for traceless solid-phase synthesis. J Org Chem 1995; 60:6006–6007. 84. Ellman JA, Plunkett MJ. Germanium and silicon linking strategies for traceless solid-phase synthesis. J Org Chem 1997; 62:2885–2893. 85. Boehm TL, Showalter HDH. Development of a novel silyl ether linker for solid-phase organic synthesis. J Org Chem 1996; 61:6498–6499. 86. Newlander KA, Chenera B, Veber DF, Yim NCF Moore ML. Simple silyl linker for the solid phase organic synthesis of aryl-containing molecules. J Org Chem 1997; 62:6726–6732. 87. Woolard FX, Paetsch J, Ellman JA. A silicon linker for direct loading of aromatic compounds to supports. Traceless synthesis of pyridine-based tricylics. J Org Chem 1997; 62:6102–6103. 88. Hu Y, Porco JA, Jr, Labadie JW, Gooding OW, Trost BM. Novel polymersupported trialkylsilanes and their use in solid–phase organic synthesis. J Org Chem 1998; 63:4518–4521.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

89. Hird NW, Irie K, Nagai K. Solid phase synthesis of 2-aminobutadienes using a piperazine linker. Tetrahedron Lett 1997; 38:7111–7114. 90. Furman B, Thurmer R, Kahuza Z, Voelter W, Chmielewski M. A new acetal resin valuable for the solid-phase synthesis of 1-oxacephams via a cyclization/cleavage step. Tetrahedron Lett 1999; 40:5909–5912. 91. Stieber F, Grether U, Waldmann H. An oxidation-labile traceless linker for solid-phase synthesis. Angew Chem, Int Ed 1999; 38:1073–1077. 92. Craig D, Robson MJ, Shaw SJ. Traceless linkers for solid-phase synthesis. Homo- and hetero Diels-Alder reactions of ortho-quinodimethanes. Synlett 1998: 12:1381–1383.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

8 Allosteric Modulation of G-Protein-Coupled Receptors: Implications for Drug Action Angeliki P. Kourounakis University of Thessaloniki, Thessaloniki, Greece

Pieter van der Klein and Ad P. IJzerman Leiden University, Leiden, The Netherlands

I.

INTRODUCTION

Representing one of the largest superfamilies of proteins in the human body, G-protein-coupled receptors (GPCRs) play a crucial role in the regulation of a variety of physiological processes, particularly within the central nervous system and the cardiovascular and endocrine systems. It is estimated that this superfamily comprises about 500 and possibly over 1000 receptor (sub)types having similar structural and/or sequence motifs, while operating via common transduction mechanisms to mediate the transmission of extracellular signals into biochemical or electrophysiological responses in a cell. A speciﬁc endogenous molecule, such as a neurotransmitter or hormone, acts as the signaling species that binds to the receptor, resulting in an activation of intracellular G proteins and signal propagation. Hence, GPCRs are important drug targets; approximately 60% of current drugs produce their therapeutic actions by binding to GPCRs.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

The design and development of synthetic drugs is mainly focused on mimicking (as in the case of agonists) or blocking (as in the case of antagonists and inverse agonists) the action of the endogenous signaling molecule by competing at the same site (the ligand binding site) on a speciﬁc receptor. Recently, a new concept of interference with drug action at GPCRs has emerged for some receptor subclasses such as the muscarinic or adenosine receptor. This concept, namely allosteric modulation of the receptor by molecules binding at a second (allosteric) site, is thus far relatively unexplored for GPCRs, although relatively common in the family of ion channel receptors [1]. This indirect (allosteric) mechanism (i.e., the modulation of the eﬃcacy or aﬃnity of the endogenous ligand for its receptor) is the molecular basis of the therapeutic action of benzodiazepines that interact with g-aminobutyric acid A (GABAA) receptors coupled through ion channels. In contrast, there has been no therapeutic role found for directly acting agonists or antagonists on this receptor. Nevertheless, only a few drugs, such as gallamine, alcuronium, or pancuronium, are known to exert their action at an allosteric site on a GPCR [2]. The potential advantages or beneﬁts of allosteric drugs over agonists, antagonists, and inverse agonists may be elaborated as follows. It is generally found that within the GPCR family, subtypes exist that bind the same signaling molecule but have diﬀerent tissue distributions as well as functions. These receptor subtypes have often a high sequence homology, especially in the regions of the receptor that are thought to contain the ligand binding site. Thus, in most cases it has been proven diﬃcult to develop drugs that not only are highly selective for one receptor subtype but have highly controlled eﬀects on the function of that receptor and act in those tissues only where their action is desired. An allosteric drug has the following properties: 1. Has no action when binding on its own to the receptor but only modulates the actions of the naturally occurring hormone or neurotransmitter when it is released. Therefore, the temporal aspects of the natural signaling mechanism are retained and desensitization is minimized. 2. Has a deﬁned maximum eﬀect that is determined by the cooperativity associated with its allosterism. 3. Can act selectively at various receptor subtypes not only by means of its own aﬃnity but also on cooperativity. No pharmacological agent has yet exploited the latter property.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

4. Has enhancing properties that can selectively intensify a weakened signal from a speciﬁc receptor subtype, alleviating the eﬀects caused by a localized neurotransmitter deﬁcit such as in Alzheimer’s or Parkinson’s disease. Although allosteric sites have been characterized on certain biogenic amine receptors—muscarinic acetylcholine, dopamine, and a-adrenergic receptors—as well as on adenosine receptors, it is not yet known whether the presence of an allosteric site is a characteristic of only a few, all, or subsets of GPCRs. Furthermore, a relevant question is whether these sites might have a physiological regulatory role as a consequence of binding with endogenous molecules. Interestingly, there are a few recent reports showing endogenous ligands aﬀecting GPCR binding and function allosterically. The endogenous tetrapeptide Leu-Ser-Ala-Leu, released from nerve terminals upon depolarization, inhibits 5HT1B receptor binding and function at nanomolar concentrations [3,4], an eﬀect speciﬁc at 5HT1B and not at other 5-hydroxytryptamine receptors that were examined. Also, a recent study reported that binding and function of the human oxytocin receptor can be inhibited directly by nanomolar concentrations of 5hdihydroprogesterone [5]. Although the actions of the peptide and the steroid have a number of common features that make their interactions diﬀerent from those previously observed, both studies suggest an entirely unanticipated cross talk between very diﬀerent signaling mechanisms, the consequences of which are not yet known.

II. DEFINITION OF ‘‘ALLOSTERIC’’: RELATED MODELS The term ‘‘allosteric’’ was ﬁrst introduced by Monod and Jacob [6], who referred to an allosteric inhibition (of the synthesis of a tryptophan precursor by tryptophan) in describing the mechanism underlying the action of ‘‘an inhibitor that was not a steric analog of the substrate.’’ Thus, ﬁrst introduced in the ﬁeld of enzymology, the term ‘‘allosteric’’ (Greek aEEo, other, diﬀerent; jH eUeo, solid, shape) means ‘‘having a diﬀerent shape.’’ It soon referred to the presence in an enzyme of a (secondary) site of attachment for a substance that modiﬁes enzyme activity without interacting directly with the active center (primary site). The allosteric eﬀect, therefore, was attributed to a change in either the three-dimensional structure of the peptide chain or else a change in

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

conformation (allosteric transition) that aﬀects the binding of the substrate to the active site [7]. Thus, the emphasis was shifted toward the crucial role of conformational changes of proteins, as elaborated by Monod et al. in 1965 [8]. The concept extended from enzymology to ‘‘receptology,’’ ﬁrst to ion channel receptors and subsequently to GPCRs. The primary site on a receptor is thus referred to as the (classical) ligand binding site or ‘‘orthosteric site,’’ while the secondary site, or allosteric site, aﬀects binding at the primary site by inducing a conformational change in the tertiary structure of the receptor protein. The simplest model that can describe allosteric interactions at GPCRs is the ternary complex allosteric model [9]. As shown in Figure 1, according to this model two parameters deﬁne the actions of allosteric agent (X): its aﬃnity for the unoccupied receptor (Kx) and its cooperativity (a) with the ligand (A) that interacts at the primary binding site: a<1 represents negative cooperativity; a=1, no cooperativity; a>1, positive cooperativity. However, based on the concept that GPCRs are able to adopt a variety of conformations, an extended model can also be described, as shown in Figure 2. In this extended ‘‘cubic ternary complex model’’ of receptor activation and modulation, the receptor can interconvert between an active (R*) and an inactive conformation (R), each with a diﬀerent

Figure 1 Representation of the simple ternary allosteric complex model of interaction of a ligand A with an allosteric agent X at a receptor R. (From Ref. 2.)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 2 Representation of a ‘‘cubic ternary complex’’ model of allosteric interaction: R, the inactive state of the receptor; R*, the active state of the receptor; A, ligand; X, allosteric agent. (From Ref. 14.)

aﬃnity for the G protein, ligand A, and allosteric modulator (X). Relative stoichiometry of the states would be determined by the presence of G protein and agonists and modiﬁed by allosteric modulators [10].

III. MUSCARINIC AND ADENOSINE RECEPTORS Allosteric interactions on GPCRs have been observed for the muscarinic [11–13], adenosine A1 [14], a2A-adrenergic [15–17], and dopamine D2 receptor [18]. This chapter focuses only on two allosteric phenomena, as well as their potential for therapeutic exploitation: that on the muscarinic receptor and that on the adenosine receptor.

A. Allosteric Modulation on the Muscarinic Receptor The ﬁrst and best-studied allosteric site on GPCRs is that on the muscarinic receptor [9,10,12,19,20]. For the ﬁve subtypes of these receptors that have been cloned and pharmacologically deﬁned as M1 to M5, various agents have been identiﬁed that allosterically regulate selectively these

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

receptor subtypes [19,21]. Gallamine was the ﬁrst compound shown to interact allosterically with the muscarinic (M2) receptor, exhibiting negative cooperativity with antagonists [3H]NMS and [3H]QNB [22,23]. The interactions of gallamine with the M2 receptor were shown to agree with the ternary complex allosteric model in both binding and functional studies [23–25]. Since then, a number of ligands have been discovered that interact with various muscarinic receptor subtypes, conﬁrming that the allosteric site is present on all ﬁve subtypes. Furthermore, these allosteric eﬀects were shown to be truly subtype speciﬁc, depending on the nature of the allosteric modulating compound. Thus, alcuronium exerts positive copperativity with [3H]NMS at the M2 and M4 but not at the M1 and M3 receptors [26,27], while other neuromuscular junction blockers such as stercuronium, pancuronium, and dtubocurarine have been shown to exhibit their eﬀects via an allosteric mechanism speciﬁcally on the M2 receptors [28–30]. A growing number of other diverse compounds have also been shown to bind to an allosteric site on the muscarinic receptors. Among them are pirenzepine (highly selective for M1 receptor), lidocaine and verapamil (ion channel blockers), tacrine (anticholinesterase compound), batrachotoxin, and strychnine (glycine receptor antagonist) [25,31–35]. Although in the cases of gallamine and some of the other agents, a values for various ligands were all below 1, another group of compounds, such as brucine and analogues, appear to be allosteric agents exhibiting positive cooperativity at one or more muscarinic receptor subtypes [36,37]. The interest in agents positively cooperative with ACh at muscarinic receptors stems from their potential use in the treatment of cognitive deﬁcits such as Alzheimer’s disease. Brucine and analogues were shown not only to enhance the aﬃnity of ACh in radioligand binding studies for the M1, M3, and M4 muscarinic receptors but further to modulate the actions of acetylcholine in functional studies. First in GTPase and [35S]GTPgS binding assays (Fig. 3), then in cAMP production and intracellular Ca2+mobilization assays (Fig. 4), and ﬁnally in a tissue model of contraction of the guinea pig ileum strip (Fig. 5) [9]. In all cases, the activity of these analogues showed ‘‘absolute subtype selectivity’’ with variable eﬀects on the various muscarinic subtypes. The results suggested the pharmacological feasibility of selectively elevating subnormally functioning cholinergic neurons in the central nervous system (CNS) by means of an appropriate allosteric enhancer.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 3 Enhancement by brucine of ACh potency at M1 receptors in functional assays in membranes. Brucine (100 AM) increased the potency of ACh to stimulate [35S]GTPgS binding to G proteins in m1 CHO cell membranes. In this experiment, the EC50 value for ACh decreased from 2.8 AM (5) to 0.9 AM (n) without significantly affecting the basal response or maximal stimulation. (From Ref. 9.)

B. Allosteric Modulation on the Adenosine Receptor Extracellular adenosine is regarded as a local hormone that exerts numerous physiological actions in a variety of mammalian tissues. The actions of this nucleoside in the body are mediated by G-protein-coupled adenosine receptors subclassiﬁed as A1, A2A, A2B, and A3 [38]. The adenosine A1 receptor is interfaced with a Gi protein, which is negatively coupled to the adenylate cyclase–cAMP signal transduction pathway, and thus, upon activation, leads to a reduction in intracellular cAMP levels. This receptor is highly and widely expressed in not only in the CNS but also in other tissues such as fat cells, bladder, and heart [38–40]. A variety of adenosinemediated eﬀfects (hypotension, inhibition of lipolysis, analgesia) occurs via the adenosine A1 receptor, rendering it an important target for pharmacological intervention. Nonetheless, the wide distribution of adenosine

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 4 Dose–response curves for the potentiation by brucine of ACh wholecell M1 muscarinic receptor responses. (A) Brucine (104 M) enhanced the potency of ACh to increase cAMP accumulation in M1 CHO cells by 2.6-fold. (B) Brucine (100 AM) produced a 3.0-fold increase in ACh potency in Ca2+ response to ACh. (From Ref. 9.)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 5 (A) In a dose-dependent manner, N-chloromethyl brucine (CMB) enhanced the field-stimulated contractions of isolated guinea pig ileum strips. The contractions were inhibited by atropine (30 nM). (B) Histogram of the percentage enhancement of contraction produced in four independent experiments of the type illustrated in A. (From Ref. 9.)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

receptors oﬀers both opportunities and drawbacks for therapeutic intervention [38,41]. For example, A1 adenosine agonists, through their interaction with adenosine A1 receptors on fat cells, are able to reduce free fatty acid levels in the blood. Since this eﬀect sensitizes insulin’s action [42]. It may be a very useful feature in non-insulin-dependent diabetes mellitus (type II diabetes). However, serious side eﬀects occur by the concomitant bradycardia and drop in mean arterial pressure due to interference with cardiovascular adenosine receptors [43]. Various strategies have been followed to circumvent all or some of these problems, such as the development of partial agonists for that purpose [44–48]. It was shown that some of these compounds were virtually ‘‘silent’’ on the heart, while keeping a pronounced, full eﬀect on adipose tissue [49]. On the other hand, among the eﬀects of receptor-bound adenosine is the ability to protect organs, including the heart and brain, from ischemic injury [50–52]. The formation of extracellular adenosine as a breakdown product of ATP is a local phenomenon, induced by a tissue at risk (e.g., under hypoxic or anoxic conditions: heart failure, stroke, etc.). As a consequence, compounds that would increase adenosine’s concentration, and thus its tissue-protective eﬀect, might have a better therapeutic proﬁle than the agonists described earlier. Marketed nucleoside transport blockers such as dipyridamole and dilazep have already proven this concept by inhibiting the intracellular uptake of extracellular adenosine, and thereby eﬀectively increasing its concentration outside the cell [53,54]. Another interesting approach is to enhance adenosine’s action locally by means of an allosteric enhancer. In 1990, Bruns and coworkers reported on various 2-amino-3-benzoylthiophene derivatives capable of enhancing the binding and activity of reference A1 receptor agonists, such as N6-cyclopentyladenosine (CPA) [14,55]. One of these ‘‘allosteric modulators,’’ PD81,723, or (2-amino-4,5-dimethyl-trienyl) [3-(triﬂuoromethyl) phenyl]methanone (Fig. 6), has been investigated pharmacologically in greater detail by various independent research groups [56–61]. The modulator PD81,723 enhances two- to threefold the binding and function of agonists such as CPA, R-PIA, or NECA to adenosine A1 receptors [62]. As shown in Figure 7, in displacement experiments of [3H]DPCPX from the human adenosine A1 receptor (wild type), the binding curve of CPA in the presence of PD81,723 is shifted leftward; it seems that CPA binds more eﬃciently, since lower concentrations of this agonist are needed to displace the same concentration of radioligand. This

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 6 Structure of PD81,723, (2-amino-4,5-dimethyl-trienyl)[3-(trifluoromethyl) phenyl]methanone and adenosine A1 agonists/antagonists.

Figure 7 Displacement of 0.4 nM [3H]DPCPX by various concentrations of CPA from human wild-type (CHO A1) and mutant (CHO A1-mutT277A) adenosine A1 receptors in the absence (n) or presence (5) of PD81,723 (10 AM).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

‘‘enhanced’’ activity of CPA is also maintained in second messenger assays, where, for example, lower concentrations of CPA (in the presence of PD81,723) are needed for the inhibition of forskolin-stimulated cAMP production in cells bearing adenosine A1 receptors (Fig. 8). It is known that PD81,723 slows down the kinetics (dissociation) of 3H-labeled agonists such as [3H]CHA or [3H]CCPA from the receptor as shown in Figure 9; the half-life of 17 min for the dissociation of CCPA alone from the rat A1 receptor is increased to 25 min in the presence of 10 AM PD81,723 [63]. It is postulated that this compound binds to an allosteric site on the adenosine A1 receptor—which, unlike the muscarinic one, is not yet so well deﬁned—while at somewhat higher concentration it binds to the ligand binding site exhibiting antagonistic action. It is presumed that via its allosteric activity PD81,723 increases the proportion of adenosine receptors in the ‘‘active’’ (R*) conformation that has a high aﬃnity for agonists and low for antagonists and inverse agonists (Fig. 2). Not only are these eﬀects selective for the A1 receptor, but they disappear upon a mutation of the receptor at the proposed agonist binding site [62]. Threonine at position 277 on the A1 receptor is considered to interact with ribose ring of agonists, since changing it to alanine greatly decreases the aﬃnity for agonists but not for antagonists. This mutation also eliminated the activity of PD81,723 (Fig. 7), which no longer can increase the already

Figure 8 Forskolin-stimulated cAMP production of CHO A1 cells after addition of CPA in the absence (n) or presence (5) of 10 AM PD81,723.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 9 Dissociation of agonist [3H]CCPA from rat brain A1 receptors in the presence (5) or absence (n) of 10 AM PD81,723.

low aﬃnity of agonists [64–67]. This indicates that an intact agonist binding site of the receptor is required for PD81,723 to exert its allosteric action [62]. Recently we developed a series of novel PD81,723 analogues, some of which appear to be superior to PD81,723 in their enhancing activity [68,69]. The synthesis of these derivatives is relatively straightforward, as shown in Figure 10 [68–72]. The 4,5-dimethyl group and the benzoyl moiety were targets for further modiﬁcations, leading to series of 4,5-dialkyl (1a–g), of tetrahydrobenzo (1h–u) and of tetrahydropyridine (3a–g) derivatives (Fig. 10, Tables 1 and 2). These derivatives were evaluated both as allosteric enhancers of agonist binding to the rat adenosine A1 receptor and as antagonists on this receptor. Among them, a number of compounds, in particular 1b, 2e, 1j, 1n, and 1u (Fig. 11, Table 1), proved to be superior to the reference compound (PD81,723) in both enhancing activity and diminished antagonistic behavior [68]. Some structure–activity relationships of a further developed R4, R5 alkyl/cycloalkyl series (2a–o, Fig. 10, Table 1) were also investigated. This study [69] revealed structural features that favored allosteric enhancing activity, such as benzoyl lipophilic substitution and thiophene 4-alkyl substitution, while other features, such as thiophene 5-bulky substitution, favored antagonistic properties. Upon further analysis, a

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 10 Scheme of synthesis of PD81,723 analogues. Reagents and conditions: i, DMF; S8, Et3N, RT (or EtOH, S8, Et2NH, 50jC); ii, C6H6, h-alanine, HOAc; iii, EtOH, S8, Et2NH; iv, BzCl, CH2Cl2, Et3N.

Table 1 Structure and Enhancing/Antagonistic Activity of 2-Amino-3-benzoylthiophenes Analogues 1a–u and 2a–o

Compound PD81,723 1a 1b 1c

R0

R4

R5

Enhancement (%)a

Antagonism (%)b

3-CF3

CH3

CH3

100

39 (F4)

CH3 CH3 CH3

CH3 CH3 CH3

8 (F5) 80 (F19) 93 (F32)

H 3-Cl 4-Cl

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

14 (F3) 19 (F4) 41 (F6)

Table 1 Continued Compound PD81,723 1d 1e 1f 1g 1h 1i 1j 1k 1l 1m 1n 1o 1p 1q 1r 1s 1t 1u 2a 2b 2c 2d 2e 2f 2g 2h 2i 2j 2k 2l 2m 2n 2o

R0

R4

R5

Enhancement (%)a

Antagonism (%)b

3-CF3

CH3

CH3

100

39 (F4)

H 3-CF3 3-Cl 4-Cl H 2-Cl 3-CF3 3-Cl 3-I 4-CF3 4-Cl 4-Br 4-I 4-NO2 4-CH3 4-CO2CH3 4-CO2H 3,4-Cl 3-CF3 3-Cl H 3-CF3 3-Cl 3-Cl H 3-Cl 3-CF3 H H H 3,4-Cl 4-tBu 4-tBu

CH2CH3 CH3 CH2CH3 CH3 CH2CH3 CH3 CH3 CH2CH3 —(CH2)4— —(CH2)4— —(CH2)4— —(CH2)4— —(CH2)4— —(CH2)4— —(CH2)4— —(CH2)4— —(CH2)4— —(CH2)4— —(CH2)4— —(CH2)4— —(CH2)4— —(CH2)4— H CH3CH2CH2 H CH3CH2CH2 H CH3CH2CH2 H C5H9 H C5H9 H C6H11 H C6H5 H C6H5 H C6H5 H (CH3)2CHCH2 CH3 CH3CH2 CH3CH2 CH3CH2CH2 CH3 CH3 CH3 CH3 —(CH2)4—

31 112 30 97 47 73 122 93 113 131 123 128 155 34 137 44 29 151 88 67 0 99 52 57 21 38 42 7 13 69 116 125 137

(F4) (F10) (F7) (F25) (F4) (F19) (F19) (F6) (F18) (F11) (F15) (F18) (F21) (F22) (F21) (F9) (F3) (F24) (F8) (F18) (F30) (F25) (F12) (F2) (F5) (F6) (F7) (F14) (F17) (F19) (F7) (F24) (F10)

13 5 22 20 35 35 32 51 66 57 40 42 64 19 30 29 35 52 54 50 49 64 64 75 80 58 47 27 17 50 47 40

(F3) (F11) (F2) (F12) (F6) (F3) (F8) (F5) (F1) (F4) (F5) (F4) (F8) (F2) (F3) (n=1) nd (F4) (F3) (F5) (F7) (F2) (F1) (F3) (F2) (F1) (F3) (F5) (F7) (F6) (F1) (F2) (F4)

Enhancing activity (at 10 AM of test compound) is expressed as percentage of decrease (FSEM) in [3H]CCPA dissociation over control (0%) and that of PD81,723 (100%, n = 3). b Antagonistic activity is expressed as percentage of displacement (FSEM) of 0.4 nM of [3H]DPCPX by 10 AM of test compound. nd: not determined. a

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 2 Structure and Enhancing/Antagonistic Activity of 2-Amino-3-benzoyl4,5,6,7-tetrahydrothieno [2,3-c]pyridines 3a–g and 4

Compound 3a 3b 3c 3d 3e 3f 3g 4 Theophylline

R0

R1

H H H H 4-Cl 4-Cl 3,4-Cl —

H 3-Cl 4-Cl 3,4-Cl H 3,4-Cl H —

Enhancement (%)a 53 106 69 57 132 106 174 14 15

(F37) (F27) (F23) (F36) (F21) (F31) (F37) (F27) (F7)

Antagonism (%)b 67 80 52 4 60 46 51 72 56

(F5) (F1) (F2) (F2) (F0) (F2) (F0) (F2) (F5)

Enhancing activity is expressed as percentage of decrease (FSEM) in [3H]CCPA dissociation over control (0%) and that of PD81,723 (100%, n = 3). b Antagonistic activity is expressed as percentage of displacement (FSEM) of 0.4 nM of [3H]DPCPX by 10 AM of test compound. a

signiﬁcant correlation was found between antagonistic activity and hydrophobic fragment constants (k values) [73] of substituent R5 (Fig. 12), in contrast to a negative correlation with those of R 4. Finally, comparison of low energy conformations (Fig. 13) of some of the 2-amino-3benzoylthiophene derivatives (PD81,723 and 2f ) with known adenosine A1 receptor antagonists (theophylline and 8-cyclohexyltheophylline) indicated that thiophene 5-substituents (R5 ) may interact with the same lipophilic domain of the adenosine A1 receptor that accommodates 8substituents of xanthine antagonists. The separation of the two activities, antagonism and allosteric enhancement, is ultimately necessary for the development of more potent and selective allosteric enhancers for the adenosine A1 receptor with potential therapeutic applications.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 11 Concentration–effect curves for derivative 1u and PD81,723. Enhancement of 100% is expressing the maximum decrease in [3H]CCPA dissociation by the highest concentration of 1u.

Figure 12 Correlation of lipophilicity parameter (k) for substituent R 5 of compounds 1a,d–f,h, j, k, and 2a–o with their antagonistic activity. ***p < 0.0001.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 13 Structure and low-energy conformation with van der Waals surface of (a) theophylline; (b) CHT; (c) PD81,723; and (d) 2f.

IV. CONCLUSION The possibility of allosterically modulating receptors oﬀers novel pharmacological means of ‘‘ﬁne-tuning’’ receptor function. Further clariﬁcation is required with respect to whether such modulated receptors are a general feature of all or only of a subset of GPCRs and whether endogenous agents regulate via this mechanism receptor function in vivo. Finally, elucidation of the molecular mechanisms of the allosteric interactions will provide useful insights for the therapeutic exploitation of this phenomenon in the design and development of appropriate modulatory drugs. Abbreviations ACh cAMP [3H]CCPA CHO CH3CN

Acetylcholine Cyclic-3V,5V-adenosine monophosphate [3H]-2-Chloro-N6-cyclopentyladenosine Chinese hamster ovary Acetonitrile

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

CHT CMB CPA [3H]DPCPX CPT DMF Et3N Et2NH GPCR HOAc NECA NMS PD81,723 QNB R-PIA Theophylline

8-Cyclohexyltheophylline N-Chloromethyl brucine N6-Cyclopentyladenosine [3H]-1,3-Dipropyl-8-cyclopentylxanthine 8-Cyclopentyltheophylline N,N-Dimethylformamide Triethylamine Diethylamine G-Protein-coupled receptor Acetic acid 5V-(N-Ethyl)-carboxamidoadenosine N-Methyl scopolamine (2-Amino-4,5-dimethyl-3-thienyl)-[3(trifluoromethyl) phenyl]methanone Quinuclidinylbenzilate N6-[-(R)-1-Methyl-2-phenylethyl]adenosine 1,3-Dimethylxanthine

ACKNOWLEDGMENTS A. Kourounakis wishes to acknowledge support for this work from the European Commission ﬁnanced programs BIO4-CT97-5138 and QLK6CT-1999-51170.

REFERENCES 1. Costa E, Guigotti A. Benzodiazepines on trial: a research strategy for their rehabilitation. Trends Pharmacol Sci 1996; 17:192–200. 2. Birdsall NJ, Cohen F, Lazareno S, Matsui H. Allosteric regulation of G-protein-linked receptors. Biochem Soc Trans 1995; 23:108–111. 3. Massot O, Rousselle JC, Fillion MP, Grimaldi B, Cloez-Tayarani I, Fugelli A, Prudhomme N, Seguin L, Rousseau B, Plantefol M, Hen R, Fillion G. 5Hydroxytryptamine-modulin, a new endogenous cerebral peptide, controls the serotonergic activity via its speciﬁc interaction with 5-hydroxytryptamine 1B/1D receptors. Mol Pharmacol 1996; 50:752–762. 4. Rousselle J-C, Massot O, Delepierre M, Zifa E, Fillion G. Isolation and characterization of an endogenous peptide from rat brain interacting speciﬁcally with the serotonergic 1B receptor subtypes. J Biol Chem 1996; 271:551–557.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

5. Grazzini E, Guillon G, Mouillac B, Zingg HH. Inhibition of oxytocin receptor function by direct binding of progesterone. Nature 1998; 392:509–512. 6. Monod J, Jacob F. General conclusions: teleonimic mechanisms in cellular metabolism, growth and diﬀerentiation. Cold Spring Harbor Symp Quant Biol 1961; 26:389–401. 7. Monod J, Changeux J-P, Jacob F. Allosteric proteins and cellular control systems. J Mol Biol 1963; 6:308–329. 8. Monod J, Wyman J, Changeux J-P. On the nature of allosteric transitions: a plausible model. J Mol Biol 1965; 12:88–118. 9. Birdsall NJ, Farries T, Gharagozloo P, Kobayashi S, Lazareno S, Sugimoto M. Subtype-selective positive cooperative interactions between brucine analogs and acetylcholine at muscarinic receptors: functional studies. Mol Pharmacol 1999; 55:778–786. 10. Christopoulos A, Lanzafame A, Michelson E. Allosteric interactions at muscarinic cholinoceptors. Clin Exp Pharmacol Physiol 1998; 25:185–194. 11. Waelbroek M. Identiﬁcation of drugs competing with d-tubocurarine for an allosteric site on cardiac muscarinic receptors. Mol Pharmacol 1994; 46: 685–692. 12. Tucek S, Proska J. Allosteric modulation of muscarinic acetylcholine receptors. Trends Pharmacol Sci 1995; 16:205–212. 13. Birdsall NJ, Farries T, Gharagozloo P, Kobayashi S, Kuonen D, Lazareno S, Popham A, Sugimoto M. Selective allosteric enhancement of the binding and actions of acetylcholine at muscarinic receptor subtypes. Life Sci 1997; 60:1047–1052. 14. Bruns RF, Fergus JH. Allosteric enhancement of adenosine A1 receptor binding and function by 2-amino-3-benzoylthiophenes. Mol Pharmacol 1990; 38:939–949. 15. Howard MJ, Hughes RJ, Motulsky HJ, Mullen MD, Insel PA. Interactions of amiloride with alpha- and beta-adrenergic receptors: amiloride reveals an allosteric site on alpha 2-adrenergic receptors. Mol Pharmacol 1987; 32:53–58. 16. Wilson AL, Womble SW, Prakash C, Cragoe EJ Jr, Blair IA, Limbird LE. Novel amiloride analog allosterically modulates the alpha 2-adrenergic receptor but does not inhibit Na+/H exchange. Mol Pharmacol 1992; 42:175–179. 17. Leppik RA, Lazareno S, Mynett A, Birdsall NJM. Characterisation of the allosteric interactions between antagonists and amiloride analogues at the human alpha2A adrenergic receptor. Mol Pharmacol 1998; 53:916–925. 18. Hoare SR, Strange PG. Regulation of D2 dopamine receptors by amiloride and amiloride analogs. Mol Pharmacol 1996; 50:1295–1308. 19. Ellis J. Allosteric binding sites on muscarinic receptors. Drug Dev Res 1997; 40:193–204. 20. Birdsall NJM, Lazareno S, Matsui H. Allosteric regulation of muscarinic receptors. Prog Brain Res 1996; 109:142–151.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

21. Lee NH, el-Fakahany EE. Allosteric interactions at the m1, m2, and m3 muscarinic receptor subtypes. J Pharmacol Exp Ther 1991; 256:468–479. 22. Clark AL, Mitchelson F. The inhibitory eﬀect of gallamine on muscarinic receptors. Br J Pharmacol 1976; 58:323–331. 23. Stockton JM, Birdsall NJ, Burgen AS, Hulme EC. Modiﬁcation of the binding properties of muscarinic receptors by gallamine. Mol Pharmacol 1983; 23:551–557. 24. Ehlert FJ. Gallamine allosterically antagonizes muscarinic receptor– mediated inhibition of adenylate cyclase activity in the rat myocardium. J Pharmacol Exp Ther 1988; 247:596–602. 25. Lazareno S, Birdsall NJ. Detection, quantitation, and veriﬁcation of allosteric interactions of agents with labeled and unlabeled ligands at G protein-coupled receptors: interactions of strychnine and acetylcholine at muscarinic receptors. Mol Pharmacol 1995; 48:362–378. 26. Tucek S, Musilkova J, Neduoma J, Proska J, Shelkovnikov SW, Vorlicek J. Positive cooperativity in the binding of alcuronium and N-methylscopolamine to muscarinic acetylcholine receptors. Mol Pharmacol 1990; 38:674–680. 27. Jakubik J, Bacakova L, El-Fakahany EE, Tucek S. Subtype selectivity of the positive allosteric action of alcuronium at cloned M1–M5 muscarinic acetylcholine receptors. J Pharmacol Exp Ther 1995; 274:1077–1083. 28. Dunlap J, Brown JH. Heterogeneity of binding sites on cardiac muscarinic receptors induced by the neuromuscular blocking agents gallamine and pancuronium. Mol Pharmacol 1983; 24:15–22. 29. Li CK, Mitchelson F. The eﬀect of stercuronium on cardiac muscarinic receptors. Eur J Pharmacol 1978; 51:251–259. 30. Waelbroeck M, Robberecht P, De Neef P, Christphe J. Eﬀects of dtubocurarine on rat cardiac muscarinic receptors: a comparison with gallamine. J Recept Res 1988; 8:787–808. 31. Roeske WR, Venter JC. The diﬀerential loss of [3H]pirenzepine vs [3H] (–)quinuclidinylbenzilate binding to soluble rat brain muscarinic receptors indicates that pirenzepine binds to an allosteric state of the muscarinic receptor. Biochem Biophys Res Commun 1984; 118:950–957. 32. Ellis J, Huyler J, Brann MR. Allosteric regulation of cloned m1–m5 muscarinic receptor subtypes. Biochem Pharmacol 1991; 42:1927–1932. 33. Waelbroeck M, Robberecht P, De Neef P, Christophe J. Eﬀects of verapamil on the binding properties of rat heart muscarinic receptors: evidence for an allosteric site. Biochem Biophys Res Commun 1984; 121:340–345. 34. Cohen-Armon M, Henis YI, Kloog Y, Sokolovsky M. Interactions of quinidine and lidocaine with rat brain and heart muscarinic receptors. Biochem Biophys Res Commun 1985; 127:326–332. 35. Cohen-Armon M, Kloog Y, Henis YI, Sokolovsky M. Batrachotoxin changes the properties of the muscarinic receptor in rat brain and heart: possible interaction(s) between muscarinic receptors and sodium channels. Proc Natl Acad Sci USA 1985; 82:3524–3527.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

36. Lazareno S, Gharagozloo P, Kuonen D, Popham A, Birdsall NJ. Subtypeselective positive cooperative interactions between brucine analogs and acetylcholine at muscarinic receptors: radioligand binding studies. Mol Pharmacol 1998; 53:573–589. 37. Gharagozloo P, Lazareno S, Popham A, Birdsall N. Allosteric interactions of quaternary strychnine and brucine derivatives with muscarinic acetylcholine receptors. J Med Chem 1999; 42:438–445. 38. DeNinno MP. Adenosine. In: Doherty A, ed. Annual Reports in Medicinal Chemistry. Vol. 33. San Diego, CA: Academic Press, 1998:111–120. 39. Jacobson KA, van Galen PJM, Williams M. Adenosine receptors: pharmacology, structure–activity relationships, and therapeutic potential. J Med Chem 1992; 35:407–422. 40. Jacobson KA, van Rhee AM. Development of selective purinoceptor agonists and antagonists. In: Jacobson KA, Jarvis MF, eds. Purinergic Approaches in Experimental Therapeutics. New York: Wiley-Liss, 1997:101–128. 41. Stiles GL. Adenosine receptor subtypes: new insights from cloning and functional studies. In: Jacobson KA, Jarvis MF, eds. Purinergic Approaches in Experimental Therapeutics. New York: Wiley-Liss, 1997:29–37. 42. Merkel LA, Hawkins ED, Colussi DJ, Greenland BD, Smits GJ, Perrone MH, Cox BF. Cardiovascular and antilipolytic eﬀects of GR79236. Pharmacology 1995; 51:224–236. 43. Mathoˆt RAA, van Schaick EA, Langemeijer MWE, Soudijn W, Breimer DD, Ijzerman AP, Danhof M. Pharmacokinetic–pharmacodynamic relationship of the cardiovascular eﬀects of the adenosine A1 receptor agonist N6-cyclopentyladenosine (CPA) in the rat. J Pharmacol Exp Ther 1994; 268:616–624. 44. van der Wenden EM, von Frijtag Drabbe Ku¨nzel JK, Mathoˆt RAA, Danhof M, Ijzerman AP, Soudijn W. Ribose-modiﬁed adenosine analogues as potential partial agonists for the adenosine receptor. J Med Chem 1995; 38:4000–4006. 45. Mathoˆt RAA, van der Wenden EM, Soudijn W, Ijzerman AP, Danhof M. Deoxyribose analogues of N6-cyclopentyladenosine (CPA) are partial agonist at the adenosine A1 receptor in vivo. Br J Pharmacol 1995; 116:1957–1964. 46. Roelen H, Veldman N, Spek AL, von Frijtag Drabbe Ku¨nzel JK, Mathoˆt RAA, Ijzerman AP. N6,C8-disubstituted adenosine derivatives as partial agonist for adenosine A1 receptors. J Med Chem 1996; 39:1463–1471. 47. Lorenzen A, Sebastiao AM, Sellink A, Vogt H, Schwabe U, Ribeiro JA, Ijzerman AP. Biological activities of N6,C8-disubstituted adenosine derivatives as partial agonist at rat brain adenosine A1 receptors. Eur J Pharmacol 1997; 334:299–307. 48. Van der Wenden EM, Carnielli M, Roelan HCPF, Lorenzen A, Von Frijtag Drabbe Ku¨nzel JK, Ijzerman AP. 5V-Substituted adenosine analogs as new

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

49.

50.

51. 52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

high aﬃnity partial agonist for the adenosine A1 receptor. J Med Chem 1998; 41:102–108. Van Schaick EA, Tukker HE, Roelan HCPF, Ijzerman AP, Danhof M. Selectivity of action of 8-alkylamino analogues of N6-cyclopentyladenosine in vivo: hemodynamic versus anti-lipolytic responses in rats. Br J Pharmacol 1998; 124:607–618. Liu G-S, Downe JM, Cohen MV. Adenosine, ischemia and preconditioning. In: Jacobson KA, Jarvis MF, eds. Purinergic Approaches in Experimental Therapeutics. New York: Wiley-Liss, 1997:153–172. Fredholm BB. Adenosine and neuroprotection. Int Rev Neurobiol 1997; 40:259–280. Knutsen LJS, Murray TF. Adenosine and ATP in epilepsy. In: Jacobson KA, Jarvis MF, eds. Purinergic Approaches in Experimental Therapeutics. New York: Wiley-Liss, 1997:423–447. Ijzerman AP, van der Wenden EM. Modulators of adenosine uptake, release and inactivation. In: Jacobson KA, Jarvis MF, eds. Purinergic Approaches in Experimental Therapeutics. New York: Wiley-Liss, 1997:129– 148. Geiger JD, Parkinson FE, Kowaluk EA. Regulators of endogenous adenosine levels as therapeutic agents. In: Jacobson KA, Jarvis MF, eds. Purinergic Approaches in Experimental Therapeutics. New York; WileyLiss, 1997:55–84. Bruns RF, Fergus JH, Coughenour LL, Courtland GG, Pugsley TA, Dodd JH, Tinney JF. Structure–activity relationships for enhancement of adenosine A1 receptor binding by 2-amino-3-benzoylthiophenes. Mol Pharmacol 1990; 38:950–958. Linden J. Allosteric enhancement of adenosine receptors. In: Jacobson KA, Jarvis MF, eds. Purinergic Approaches in Experimental Therapeutics. New York: Wiley-Liss, 1997:85–97. Kollias-Baker C, Rubl J, Dennis D, Bruns RF, Linden J, Belardinelli L. Allosteric enhancer PD 81,723 acts by novel mechanism to potentiate cardiac actions of adenosine. Circ Res 1994; 75:961–971. Musser B, Mudumbl RV, Liu J, Olson RD, Vestal RE. Adenosine A1 receptor-dependent and -independent eﬀects of the allosteric enhancer PD 81,723. J Pharmacol Exp Ther 1999; 288:446–454. Mizumura T, Auchampach JA, Linden J, Bruns RF, Gross GJ. PD 81,723, an allosteric enhancer of the A1 adenosine receptor, lowers the threshold for ischemic preconditioning in dogs. Circ Res 1996; 79:415–423. Bhattacharya S, Linden J. The allosteric enhancer, PD 81,723, stabilizes human A1 adenosine receptor coupling to G proteins. Biochim Biophys Acta 1995; 1265:15–21. Leung E, Walsh LKM, Flippin LA, Kim EJ, Lazar DA, Seam CS, Wong EHF, Eglen RM. Enhancement of adenosine A1 receptor functions by

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

62.

63.

64.

65.

66.

67.

68.

69.

70. 71. 72.

73.

benzoylthiophenes in guinea pig tissues in vitro. Naunyn-Schmiedeberg’s Arch Pharmacol 1995; 352:206–212. Kourounakis A, Visser C, de Groote M, Ijzerman AP. Diﬀerential eﬀects of the allosteric enhancer PD81,723 on agonist and antagonist binding and function at the human wild type and a mutant (T277A) adenosine A1 receptor. Biochem Pharmacol 2000; 61:137–144. Kourounakis A, Visser C, de Groote M, Ijzerman AP. Allosteric modulation of the rat adenosine A1 receptor: diﬀerential eﬀects on agonist and antagonist binding. Biochem Pharmacol 2001; 51:207–215. Townsend-Nicholson A, Schoﬁeld PR. A threonine residue in the seventh transmembrane domain of the human A1 adenosine receptor mediates speciﬁc agonist binding. J Biol Chem 1994; 269:2373–2376. Tucker AL, Robeva AS, Taylor HE, Holeton D, Bockner M, Lynch KR, Linden J. A1 adenosine receptors: two amino acids are responsible for species diﬀerences in ligand recognition. J Biol Chem 1994; 269:27900– 27906. Rivkees SA, Lasbury ME, Barbahalya H. Identiﬁcation of domains of the human A1 adenosine receptor that are important for binding receptor subtype selective ligands using chimeric A1/A2 adenosine receptors. J Biol Chem 1995; 270:20485–20490. Dalpiaz A, Townsend-Nicholson A, Beukers MW, Schoﬁeld P, Ijzerman AP. Thermodynamics of full agonist, partial agonist and antagonist binding to wild type and mutant adenosine A1 receptors. Biochem Pharmacol 1998; 56:1437–1445. Van der Klein PAM, Kourounakis AP, Ijzerman AP. Allosteric modulation of the adenosine A1 receptor. Synthesis and biological evaluation of novel 2-amino-3-benzoylthiophenes as allosteric enhancers of agonist binding. J Med Chem 1999; 42:3629–3635. Kourounakis AP, Van der Klein PAM, Ijzerman AP. Elucidation of structure–activity relationships of 2-amino-3-benzoylthiophenes: Study of their allosteric enhancing versus antagonistic activity on adenosine A1 receptors. Drug Dev Res 2000; 49:227–237. Gewald K, Schinke E, Bo¨ttcher H. 2-Amino-thiophene aus methylenaktiven Nitrilen, Carbonylverbindungen und Scwefel. Chem Ber 1966; 99:94–100. Jones G. The Knoevenagel Condensation. Vol. 15. Organic Reactions. New York: Wiley, 1967:204–599. Tinney FJ, Sanchez JP, Nogas JA. Synthesis and pharmacological evaluation of 2,3-dihydro-1H-thieno[2,3-e][1,4]diazepines. J Med Chem 1974; 17:624–630. Hansch C, Leo A. Substituent Constants for Correlation Analysis in Chemistry and Biology. New York: Wiley, 1979.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

9 Protein Misfolding and Neurodegenerative Disease: Therapeutic Opportunities Harry LeVine III University of Kentucky, Lexington, Kentucky, U.S.A.

I.

DISEASES WITH PROTEIN MISFOLDING

Alzheimer’s disease, a typically late-life dementia and the most prevalent chronic neurodegenerative disease, is pathologically characterized by the presence of insoluble h-amyloid peptide in extracellular senile and diﬀuse plaques and intracellular accumulation of neuroﬁbrillary tangles (NFTs) of hyperphosphorylated tau protein as parahelical ﬁlaments. These insoluble protein assemblies are derived from normal cellular proteins that have deposited in entities that had been recognized histologically for some 70 years before their main protein constituents were determined. A potential unifying theme has emerged in the pathology of a number of chronic neurodegenerative diseases. Improved immunological and microanalytical methods have led to the identiﬁcation of the constituents of other proteinaceous deposits associated with neurodegenerative disease. These deposits are intracellular, however, unlike those of the amyloid-beta (A h) peptides. A series of genetically dominant trinucleotide repeat diseases coding for glutamine in which the CAG repeats are expanded in a diﬀerent protein in each disease was observed to develop insoluble polyglutamine-containing inclusions, concentrated in diﬀerent brain areas

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

and undergoing cell loss depending on the protein and disease involved [1– 5] (see Table 1). The prototype is Huntington’s disease (HD) which aﬀects the largest number of people. Polyglutamine stretches can form particularly stable h-sheet structures, which are prone to aggregation [6–9]. First described in transgenic mice overexpressing exon I of Huntingtin, which contains the polyglutamine repeat [10], nuclear inclusions of Huntingtin form in susceptible regions of the brain. The same ubiquitinated inclusions are also found in human HD tissue when appropriate antibodies are used [11–14]. Although the correlation of deposits with the synaptic and cell loss of the disease pathology is imperfect, similar to the situation in Alzheimer’s disease, their appearance is consistent with a pathway of protein folding and translocation that leads to cell loss. Similar observations have been made and conclusions reached for types 3 and 7 spinocerebellar ataxia (SCA-3, SCA-7) and dentatorubral–palladoluysian atrophy (DRPLA), polyglutamine repeat diseases with cerebellar pathology [15–19]. A wide variety of insoluble proteins are associated with chronic neurodegenerative diseases (Table 2). Familial tauopathies, collectively referred to as FTDP17, are ascribed to mutations in various tau exonic or intronic sequences that alter mRNA isoform expression, resulting in insoluble ﬁbrillar deposits of the microtubule-associated protein tau on human chromosome 17. Other tauopathies have been identiﬁed, varying with respect to the brain region aﬀected and the ratio of the diﬀerent tau gene splice products deposited [20]. Progressive supranuclear palsy and Pick’s disease are classic late-onset tau deposition diseases [21,22]. Tau is a conformationally ambiguous protein that does not adopt a deﬁned structure in solution [23]. Hyperphosphorylation of tau favors conformational changes leading to rapid intermolecular h-sheet formation. This inhibits the microtubule-polymerizing activity of this microtubule-associated protein and leads to NFT formation [24]. The phosphorylation of a speciﬁc sequence on tau containing T231 facilitates binding and depletion of a prolyl isomerase, Pin1, eﬀecting its nuclear function [25]. Prion diseases resulting in encephalopathy can be transmitted between individuals within species (more rarely between species) [26–28]. A conformational variant of the normal cellular protein PrPS (PrPC) (protease-sensitive or cellular) is believed to catalyze [29] or nucleate [30– 33] conversion to the pathological form, PrPR (protease-resistant). This highly unusual nongenetic mode of transmission of an infectious agent has been strongly debated [29]. The observation of multiple examples of nucleated catalysis of aberrant polymerization of protein subunits has

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 1 Polyglutamine (CAG Repeat) Neurodegenerative Diseases Repeat number Disease

Sites of pathology

Huntington’s Striatum (medium, spiny) disease Spinocerebellar Cerebellar cortex (Purkinje cells), ataxia (SCA), type 1 brain stem SCA2 Cerebellum, pontine nucleus, substantia nigra SCA3 (Machado– Substantia nigra, globus pallidus, Joseph disease) pontine nucleus, cerebellar cortex SCA6 Cerebellar and mild brain stem atrophy SCA7 Photoreceptors and bipolar cells, cerebellar cortex, brain stem Motor neurons, dorsal root ganglia Spinal and bulbar muscular atrophy (SBMA) Globus pallidus, dentatorubral and Dentatorubralsubthalamic nuclei palladoluysian atrophy (DRPLA) a

Normal Disease 11–34

Protein

Location of disease (normal)a

36–121 Huntingtin

NI (c)

6–39

40–81

Ataxin-1

NI (n, c)

15–29

36–64

Ataxin-2

NI (c)

13–42

61–84

Ataxin-3

NI (c)

21–30 VDCCa1Asubunit 37–130 Ataxin-7

NI (memb) NI (n)

11–34

40–62

Androgen receptor

NI (n)

7–35

49–88

Atrophin-1

NI (n)

4–18 7–17

NI, nuclear inclusions; (c), normal cytoplasmic localization; (n), normal nuclear localization; (membr), normal membrane localization.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 2 Neurodegenerative Diseases with Insoluble Deposits Disease

Sites of pathology

Major protein

Deposit

Alzheimer’s disease

Neocortex, hippocampus

h Peptide; 4R, 3R tau

Multiple=system tauopathy (familial) Progressive supranuclear palsy (PSP) Corticobasal degeneration (CBD) Pick’s disease Diffuse Lewy body disease (DLB) Parkinson’s disease Multiple-system atrophy (MSA) Amylotrophic lateral sclerosis (ALS) Familial ALS Creutzfeldt–Jakob disease (CJD) New variant CJD Gerstmann–Straussler– Scheinker disease Fatal familial insomnia Kuru

Frontotemporal regions, brain stem, spinal cord Frontotemporal regions

4R tau 4R tau NF

Diffuse and senile plaques, paired helical formation, NFT NFT in oligodendroglia and neurons NFT in astrocytes and neurons

Frontotemporal regions

4R tau

NFT

Frontotemporal regions Cerebrocortical regions, substantia nigra Substantia nigra, brain nuclei Cerebellum, striatal regions

3R tau NF a-Synuclein

Paired helical formation, NFT Lewy bodies and neurites

a-Synuclein

Lewy bodies and neurites

a-Synuclein

Lewy bodies and neurites

Brain stem, spinal cord

a-Synuclein

Neuronal cytoplasm

Brain stem, spinal cord

SOD1 mutants Prion protein

Neuronal cytoplasm Extracellular deposits

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

markedly decreased the heretical ﬂavor of such concepts. Transmissible protein aggregation has also been observed with [URE3], the prion form of Ure2p, a nonchromosomal genetic element regulating nitrogen catabolism, and with [PS1], the prion form of Sup35p in Saccharomyces cerevisiae [31,34]. The speciﬁc etiology of prion diseases in mammalian systems depends on the modiﬁed form of the protein [35], of which diﬀerent variants display distinguishable conformations [36,37]. Most forms lead to a spongiform encephalopathy with marked neuronal cell loss in regions that accumulate the pathogenic protease-resistant conformer of the protein. Some of the more virulent forms of the protein expressed in Creutzfeldt– Jakob disease (CJD) are accompanied by classical intracellular amyloid plaques of PrPR. Although primarily recognized as a rare animal disease (scrapie), its appearance in the English beef herd in the 1990s and its potential for transmission to humans after a long latency caused a ﬂurry of interest in detection and treatment countermeasures [38]. a-Synuclein, a synaptic protein, is deposited in Lewy bodies and Lewy neurites in Parkinson disease [39–41], in diﬀuse Lewy body disease [42], and in the Lewy body variant of Alzheimer’s disease [43]. Multiplesystem atrophy is characterized by intracellular neuronal and glial asynuclein inclusions [44]. The role for a-synuclein in these diseases was supported by the discovery of mutant forms of a-synuclein, A53T and A30P, in familial early-onset Parkinson’s disease [40]. Like tau, a-synuclein is a conformationally ambiguous protein with little stable secondary structure in solution [45]. a-Synuclein, but not the related h- or g-synuclein, can polymerize in a nucleation-dependent fashion [46–48]. Lou Gehrig’s disease (amyotrophic lateral sclerosis: ALS) displays motor neuron deposits of hyperphosphorylated neuroﬁlament subunits in the sporadic disease. Familial ALS, some 20% of all cases of ALS, involves dominant superoxide dismutase SOD1 mutants that can form h-barrel aggregates [49–51]. To this list of protein misfolding diseases can be added rare familial amyloidoses in which the mutated proteins have the classic amyloid ﬁbril congophilic birefringence and cross-h-sheet structure (Table 3). Many of these deposits have an impact on the central nervous system (TTR, cystatin, lysozyme) as well as on other organ systems. A newly described disease, familial British dementia, is associated with the deposition of Abri, a 34 amino acid, 4 kDa peptide cleaved from a 277 amino acid precursor sequence, the last 10 amino acids of which are not normally translated [52]. Familial encephalopathy with neuroserpin inclusion bodies (FENIB) is

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Table 3 Amyloidoses Recognized by the WHO International Nomenclature Committee on Amyloidosis Precursor protein Immunoglobulin light chain Immunoglobulin heavy chain Apo-serum amyloid A protein Transthyretin h2-Microglobulin Apolipoprotein AI Gelsolin Lysozyme Fibrinogen a chain Cystatin C h-Amyloid precursor protein

Prion protein Procalcitonin Islet amyloid polypeptide Atrial natriuretic factor Prolactin Insulin Lactoferrina a

Associated disorder Plasma cell disorders Plasma cell disorders Inflammation-associated, familial Mediterranean fever Familial amyloidotic neuropathy, systemic senile amyloidosis Dialysis-associated amyloidosis Familial amyloidotic neuropathy, aortic amyloidosis Familial systemic amyloidosis Familial systemic amyloidosis Familial systemic amyloidosis Familial cerebral hemorrhage with amyloidosis Sporadic and familial Alzheimer’s disease, familial cerebral hemorrhage with amyloidosis Spongiform encephalopathies C-cell thyroid tumors Insulinoma, type II diabetes Atrial amyloidosis Prolactinomas; pituitary amyloidosis Iatrogenic amyloidosis Corneal amyloidosisa

Preliminary, awaiting conﬁrmation by WHO International Nomenclature Committee on Amyloidosis. The term amyloidosis is reserved by the committee speciﬁcally for extracellular protein deposits.

another rare hereditary dementing disorder resulting from point mutations in the neuroserpin gene [53]. FENIB is marked by unique neuronal inclusion bodies consisting primarily of abnormal aggregated neuroserpin ﬁlaments formed by a mechanism similar to that found in other familial diseases of serpin conformation, including emphysema and cirrhosis due to mutant a1-antitrypsin or thromboembolytic disease in antithrombin mutants [54].

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

The mechanisms of cell loss in these various diseases of protein deposition may diﬀer in detail, but the association of insoluble protein inclusions with the pathology suggests that interventions preventing protein misfolding and deposition may be of therapeutic utility. Alternatively, stabilization of toxic species must be avoided, since the insoluble form of the protein is one potential strategy for reducing the exposure to toxic, soluble forms. Approaches similar to those applied to blocking Ah ﬁbril formation in Alzheimer’s disease may prove fruitful with these other proteins, possibly extending to some of the same compounds being developed for AD. There are examples of this for prions [55–58] and for transthyretin [59]. The purpose of this chapter is to conceptualize the shared molecular features of protein misfolding in neurodegenerative diseases. By stressing the commonalities, rational strategies can be devised to target similar pathways that lead to cellular degeneration and eventually to clinical symptoms in these diseases. This is one way to maximize the eﬀects of progress made for the pharmaceutically attractive (relatively large patient base) neurodegenerative diseases such as Alzheimer’s and Parkinson’s for application to other serious but less prevalent neurodegenerative diseases. Such ‘‘piggyback’’ strategies may be a starting point for therapeutics that already have the appropriate bioavailability, brain penetration, and longterm safety proﬁle required for these applications. Similarly, nonneural amyloid diseases and diseases with signiﬁcant amyloid components such as type II diabetes could also be approached.

II. MECHANISMS OF PROTEIN POLYMERIZATION Protein homopolymerization is a well-studied process by which cellular structure is dynamically regulated in response to the environment and cellular metabolism. Actin and tubulin exist as nucleotide-dependent (ATP and GTP, respectively) polymers (microﬁlaments and microtubules) that rapidly elongate and shorten in a reversible manner regulated by binding proteins that can catalyze either polymerization or ﬁlament shearing. Mathematical analysis of the physical chemistry of the polymerization of these systems has deﬁned the nucleation and elongation processes and provided the theoretical basis for models describing ﬁbril assembly [60–63]. Nature has also provided evidence that small molecules, such as plant alkaloids and fungal secondary metabolites, are capable of modulating protein–protein interactions

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

(actin depolymerization: cytochalasin, podophylotoxin; tubulin depolymerization: vinca alkaloids; microtubule stabilization: taxol). Another extremely physiologically important protein polymerization reaction that has been studied quantitatively is the process of thrombin-catalyzed ﬁbrinogen fragmentation and assembly of ﬁbrillar ﬁbrin during the clotting of blood [64–66]. The larger size of the monomeric protein units in these polymers has simpliﬁed detection of the various intermediates in the assembly process. Atomic level structural resolution of these relatively large proteins has been aided by use of the modulators of polymerization [67–69]. Protein polymerization can also lead to pathological consequences. In contrast to physiologically normal assemblies, the pathological polymers are usually poorly reversible or degradable and tend to accumulate until they cause problems for the surrounding cells or tissue. The polymerization of mutant hemoglobin S inside the red blood cells of individuals aﬄicted with sickle cell anemia occurs rapidly and is modulated by the hemoglobin ligand 2,3-diphosphoglycerate. The mutation decreases the stability of the deoxygenated form of the protein, leading to exposure of hydrophobic surfaces and an increased propensity to aggregate. A model envisioning heterogeneous nucleation along the sides of the polymer and branching reactions in addition to the standard homogenous nucleation observed at the ends of growing ﬁbrils was ﬁrst described for the sickle cell hemoglobin system [70,71]. The eﬀectiveness of hydroxyurea treatments in reducing the severity of the sickle cell crisis is ascribed to stabilizing eﬀects on the mutant hemoglobin conformation [72]. Several pathological self-polymerizing systems have been biophysically characterized suﬃciently to permit identiﬁcation of protein or peptide species that could serve as molecular targets in a structure–activity relationship. These include transthyretin (TTR) [73–76], serum amyloid A protein (SAA) [77], microtubule-associated protein tau [78–80], amylin or islet amyloid polypeptide (IAPP) [81,82], IgG light chain amyloidosis (AL) [83–85], polyglutamine diseases [9,86], a-synuclein [47,48] and the Alzheimer’s h peptide [87–96]. A variety of Ah peptide assay systems have been established at Parke-Davis to search for inhibitors of ﬁbril formation that could be therapeutically useful [97]. In the search for ﬁbril formation inhibitors, the self-association to form amyloid ﬁbrils of the Ah peptides containing 40 and 42 amino acids can be treated as a coupled protein folding and polymerization process passing through multiple intermediate peptide species. The in vitro challenge is (1) to identify the various conformational forms and

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

multimeric species involved, (2) to establish their order and arrangement in parallel and/or in sequential pathways in a reaction scheme, and (3) to design assay conditions under which only one intermediate is rate limiting so that a reasonable structure–activity relationship can be determined. The in vivo relevance of a particular mechanistic scheme will eventually be assessed by the activity of bioavailable and brain-penetrant inhibitors of the deﬁned in vitro reactions in in vivo transgenic models of central nervous system Ah amyloidosis.

III. INTERMEDIATES IN FIBRIL FORMATION By analogy to the well-characterized polymerizing tubulin and actin protein systems, and consistent with experiment, h-peptide aggregation in the test tube is envisioned as nucleation event rate-limited by the formation of a multimeric intermediate from the monomeric random coil peptide in solution. Similar qualitative kinetics hold for light chain (AL) amyloid, IAPP, TTR, prion protein, tau, polyglutamine diseases, and asynuclein. Figure 1 suggests a schematic view of the process for the hpeptide separated into prenucleation, nucleation, and ﬁbril growth phases. The techniques used to characterize peptide species at the diﬀerent stages of the reaction are noted below the relevant intermediate in the ﬁgure. The presence of the early intermediates has been inferred from the kinetics of

Figure 1 Stages of amyloid aggregation: steps in protein polymerization and the techniques used to measure them for the Alzheimer’s h peptide.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Ah ﬁbril formation; an identiﬁable nucleating species has yet be isolated. Direct observation has been made diﬃcult by the small size of the h peptide, which has an eﬀective hydrodynamic radius of 4 nm [98–100], and by the apparent low abundance of nucleating species due to the low probability of their formation. Such species would be formally akin to an enzyme transition state that is usually kinetically inferred or sometimes trapped with certain kinds of inhibitor. In disaggregated, ultraﬁltered (20 nm pore size) preparations, less than 1% of the molar peptide concentration is inferred to be present as ‘‘seeds’’ or nuclei determined by the kinetics of ﬁbril formation [101]. There is, however, ‘‘hard’’ evidence for the involvement of transient species in h-peptide ﬁbril formation. Recent atomic force microscopy (AFM) [93,94,102–104] and electron microscopic observations [105,106] have characterized rope like species intermediate between nucleation and ﬁbril extension. Designated as protoﬁbrils, these species appear to anneal and to wind around each other. Such a model is consistent with oriented x-ray ﬁber diﬀraction patterns of a triple or higher helix of h sheets producing a h-helix quaternary ﬁbril structure [107–113]. These protoﬁbrils are on-pathway intermediates in amyloid ﬁbril formation [93,94] containing h-sheet structure. They are negative with respect to thioﬂavin T and apparently toxic to cultured cells [105]. A stiﬀer, more compact ﬁbril species (type I) is eventually formed from the initial type II ﬁbril, 10–20 nm high. The dominant ﬁbril form observed is dependent on the environmental conditions and the initial conformational state of the peptide [103]. Recent AFM studies indicate that amylin ﬁbrils grow bidirectionally, from both ends at roughly equal rates [114]. Branched ﬁbrils and heterogeneous catalysis along the edge of the ﬁbrils [70,71] have also been observed [94]. The variety of structures of h-peptide species observed by electron microscopy and by AFM suggest that diﬀerent surfaces would be available to bind inhibitors on each species; moreover, the ability of a given inhibitor to block the ﬁbrillization reaction should depend on the peptide species present in a particular situation. The rate-limiting species in vivo is unknown at present. It is also possible for more than one ﬁbrillization pathway to operate concurrently, depending on the in vitro and in vivo reaction conditions. A host of molecules have been claimed to inhibit hpeptide amyloid ﬁbril formation on the basis of a variety of assays for activity. The diversity of structures is represented in Figure 2. Their eﬃcacy is in general low (IC50 tens of micromolar or higher), corresponding roughly to the order of magnitude of the amount of peptide present in

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 2 Reported inhibitors of Ah aggregation: 1, nicotinamide [156]; 2, Anthranilates [59]; 3, N-alkyl-N-methylpiperidinium bromides [157]; 4, benzothiazoles [U.S. patent 6,001,331]; 5, Congo Red [142]; 6, melatonin [158]; 7, PPI-558 (Praecis Pharmaceuticals, Inc. patent WO 9628471); 8, anthracyclines (IDOX) [159]; 9, aza-anthracyclones (WO 9832754-A); 10, iminoaza-anthracyclinones (WO 9832754-A); 11, acridinones (U.S. patent 5,972,956); 12, naphthyl monoazo compounds [U.S. patent 5,955,472]; 13, porphyrins [160]; 14, naphthalenes (Japanese patent 090954222, Teijin KK); 15, rifamycins [161]; 16, rifampicins [161]; 17, alkylsulfonates/sulfates [162].

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

most assays, implying a 1:1 compound-to-peptide stoichiometry. In most cases, however, the stoichiometry for inhibition has not been determined. A high concentration of Ah peptide is generally used to overcome the unfavorable kinetics of multiple peptides interacting to form a nucleus capable of supporting the addition of monomeric peptide. Such reactions exhibit a lag phase until the nucleus is formed (Fig. 3, curve c). Inhibitors can aﬀect either the lag phase, the maximal extent of the reaction, or both (Fig. 3, curve d). Unless both the nucleation and extension reactions are monitored, inhibitors prolonging the lag phase are poorly distinguished from those blocking extension from the nucleus, thus muddying any structure–activity relationships. Quantitative treatment of the reaction has been proposed to mathematically separate the nucleation and extension reactions [62,115]. Distinguishing true nucleation from various exponential growth mechanisms is actually quite diﬃcult, requiring precise rate

Figure 3 Effect of seeding and inhibitors on aggregation reaction. The lag phase (curve c) is characteristic of reactions in which formation of nuclei for polymerization is an unfavorable process. Addition of preformed nuclei or ‘‘seeds’’ (curve a) abolishes the lag phase. Inhibitors may affect the formation of nuclei and influence either the lag phase, the extension of the nuclei changing the growth phase, or both (curve d). The inhibitor example (curve d) acts more strongly at nuclei formation than on the slope or plateau level of the growth phase.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

determinations of the ﬁrst 10% of the reaction over a range of reactant concentrations [116]. Such measurements have not been achieved with the h peptide. An extension reaction from a nucleus (Fig. 3, curve a) is pseudo–ﬁrst order in peptide concentration and thus more easily analyzed. Inhibitors of extension would be expected to decrease the reaction rate and/or extent (Fig. 3, curve b). The nature of the nucleus, however, is a variable as well. Preformed ﬁbrils can act as one kind of nucleus, adding monomers at the ends or laterally, which can give rise to more complex kinetics [70]. Branching vs linear addition reactions can be distinguished in some situations by dynamic light-scattering methods. Another type of nucleus seems to be present in solubilized aqueous peptide solutions that can pass through a ﬁlter having a pore size of 200 nm, but not 20 nm [101] and displays linear kinetics of ﬁbril extension [117]. These species are present in too low an abundance to be observed directly, though they are detectable kinetically. Distinguishing the preformed and endogenous nucleus forms is problematic. The behavior of the accretion of soluble peptide onto AD plaques in tissue sections [118] or onto sonicated ﬁbrils [119] is kinetically similar to that of spontaneous soluble nuclei. However, the endogenous soluble nuclei are not equivalent on a molecular level to ﬁbril or AD plaque nuclei, since molecules such as Congo Red inhibit endogenous soluble nuclei extension at 0.25 AM [97], while over 700 AM is required to block accretion of monomer h peptide onto ﬁbrils or plaques [119]. An assemblycompetent form of the h peptide, h10–35, has been shown to interact with ﬁbrils and plaques [120] and to adopt a particular conformation in solution as determined by NMR spectroscopy [121], although no high aﬃnity inhibitors of this accretion reaction have been reported. Alternatively, low abundance conformational forms of monomeric peptide may be the actively associating form of the peptide to endogenous seeds, accounting for the high (micromolar) amounts of peptide required in vitro for the extension reaction. These forms may be more abundant in biological systems, allowing ﬁbril formation to occur at the low bulk concentrations (nanomolar) of Ah peptides found in vivo. As a result of the confusion over the identity of nucleating h-peptide species, prenucleation events remain poorly deﬁned. A variety of methods possessing diﬀerent degrees of resolution have been employed to look at these early stages in ﬁbril formation. Chemical [122] and enzymatic (transglutaminase) [123,124] cross-linking, electron microscopy (EM) [105,106,125], AFM [93,94,102,103], ultracentrifugation, dynamic light

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

scattering, and ﬂuorescence resonance energy transfer (FRET) with modiﬁed beta peptides [89,105] have probed the oligomerization process preceding nucleus formation but have not yielded deﬁnitive structural information on the species present or the extent of their participation in nucleus formation. Diﬃculties arise from the small size (4.3 kDa) of the monomeric peptide unit, the simultaneous presence of multiple species of peptide, both conformational and association states, and their transient nature (since they rapidly form amyloid ﬁbrils as their concentrations increase). As speciﬁc inhibitors of early stages in ﬁbril growth are discovered, peptide species will be better deﬁned, particularly if the intermediates can be trapped and their structures determined. Fibril extension from nuclei preformed under deﬁned conditions has been characterized through a series of nucleus-dependent kinetic assays. The process of ﬁbril formation from a nucleus in equilibrium with soluble, mostly monomeric peptide has proved much more amenable to study than the formation of the nucleus itself. Fibrillar species are readily detected by growth in size (ﬁltration, sedimentation, static light scattering–turbidity), amyloid-speciﬁc reactivity with the optical probes Congo Red and thioﬂavin S and T, and by EM and AFM. Endogenous soluble nuclei or seeds form in aqueous solution, accumulating slowly at low temperature. Brief treatment with denaturants, organic solvents, and treatment with neat triﬂuoroacetic acid (TFA) or concentrated formic acid breaks down these seed structures, restoring the lag period of unseeded ﬁbril formation. The processes of both seed formation and ﬁbril extension are dependent on temperature and on peptide concentration, with 37jC being required for establishing equilibrium within 24 h with 30 AM h1–40. A full description of the assay system may be found elsewhere [97,117]. A 4 h reaction time is typically within the linear portion of the time course. This nucleus-dependent assay detects mainly inhibitors that are substoichiometric with the monomeric peptide, which is present at high concentration. It is relatively insensitive to inhibitors that target the monomeric peptide. Whether the inhibitors interact with the growing end of a seed or with a low abundance conformational form of the h peptide that is competent to add to the seed is diﬃcult to determine at this time. Similar dose–response curves are obtained for Congo Red as an inhibitor with either thioﬂavin T (ThT) ﬂuorescence or ﬁltration of radioiodinated peptide readouts (Fig. 4) Caveats in the interpretation of both the ThT and radiometric ﬁltration assays for the evaluation of putative inhibitors are discussed elsewhere [97].

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 4 Inhibition of nucleated fibril extension by Congo Red: h1–40 fibril formation detected by filtration of either radiolabeled peptide (circles) or thioflavin T (ThT) reactivity, (triangles) is inhibited by Congo Red with similar potency. For assay details, see Ref. 97.

The prediction that ﬁbrillization reactions proceeding via diﬀerent folding pathways governed by diﬀerent rate-limiting steps could be subject to diﬀerent modes of inhibition appears to be substantiated. The endogenously seeded type of assay identiﬁes types of inhibitor diﬀerent from unseeded assays by using light scattering or turbidity detection. With the exception of the naphthyl monoazo benzo compounds (12) and the acridinone series (11), the molecules reported in Figure 2 are ineﬀective (IC50 > 100 AM) in the presence of 30 AM h1–40 in seeded assays. In particular, short peptide sequences derived from the h16–25 amyloidogenic core of the h peptide KLVFFA are ineﬀective under the seeded assay conditions, although many modiﬁcations of this sequence have been studied [126–129], some of which (e.g., 7) are being developed as therapeutics. Inhibitors eﬀective in the seeded assay format such as Congo Red are inactive in an accretion assay onto immobilized ﬁbrils [119]. Rifampicin and daunomycin are very weakly active against accretion [130].

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

IV. CELLULAR SYSTEMS AND AMYLOID FIBRILS One of the reported biological eﬀects of h-amyloid and amylin ﬁbrils is cellular toxicity, inferred in vivo and modeled in various tissue culture systems. While amyloid ﬁbrils were initially thought to be the toxic species, it has become increasingly clear that some other entity, probably soluble oligomers of the h peptide [131,132] that are in equilibrium with ﬁbrils, are the culprit. Thus, in developing aggregation inhibitors that would be therapeutically useful, it is important to demonstrate that the nonﬁbrillar peptide species stabilized by inhibitor treatment are not toxic to cells. The selection of an appropriate cellular system is important because the resistance of cell types to the toxic eﬀects of the Ah peptide varies signiﬁcantly, often requiring industrial (50–100 AM) concentrations of peptide or the use of the nonbiological h25–35 fragment. Mixed neuronal/glial or pure neuronal embryonic hippocampal or cortical cultures would seem to be the most relevant cell type, since neuronal cell death and dysfunction are hallmarks of neurodegenerative disease like AD. Unfortunately, the embryonic primary cultures are irregularly resistant to the eﬀects of h1–42 when cell death is monitored. These cultures are heterogeneous mixtures of neuronal cell types, only some of which seem to be aﬀected by the Ah peptide. In addition, embryonic mouse neurons are not the same as the deeply diﬀerentiated cells in the brain of an 80-year-old human. Cultured PC12 cell lines have become a favorite system, with changes in MTT formazan production serving as a readout. However, the formazan deposition is not related to cell survival [133–136] and so is not reliable as an indicator of the eﬀects of amyloid on cell death. Another prominent site of deposition of h-amyloid ﬁbrils with age and in AD is within the cerebrovasculature in areas of the brain prone to parenchymal amyloid deposition [137–139]. The peptide deposits along the surfaces of the smooth muscle cells of the vascular wall, resulting in the death of those cells and their replacement by amyloid ﬁbrils, weakening the vascular wall. Endothelial cells are also aﬀected [140]. The ‘‘Dutch’’ mutation in the APP precursor protein Q22E, within the h-peptide sequence, produces a particularly ﬁbrillogenic and toxic (to smooth muscle cells) peptide associated with primarily vascular deposition of mutant peptide and hemorrhagic vessel disease [137]. Thus, in addition to neuronal cells, the brain vascular smooth muscle cells are a pathologically relevant cell type. While the source of

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

the h peptide in these deposits (brain or smooth muscle cells) is under debate, the smooth muscle cells in culture generate prodigious amounts of h peptide and accumulate C-terminal fragments of hAPP [139]. Organotypic cultures of the leptomeningeal blood vessels will accumulate exogenously applied, ﬂuorescently labeled Ah peptide [141]. The leptomeningeal vascular smooth muscle cells (VSMC) isolated from either human or canine sources have proved to be reliable indicators for h-amyloid toxicity. Overnight treatment with 10 AM h1–42 leads to deposition of ﬁbrillar peptide in ThS-positive strands onto the cell surface and apoptosis of 70 to 80% of the VSMC assessed by bisbenzamide staining of condensed nuclei. For this cell type, preformed ﬁbrils have little eﬀect on cell survival, and the added ﬁbrils remain scattered over the surface of the culture dish. Interpreting the eﬀects of amyloid-modulating compounds on h1–42-induced cellular toxicity and relating the results to in vitro aggregation inhibition is far from straightforward. A number of compounds are toxic to cells by a variety of routes. Besides interfering with aggregation, test compounds can block binding to the cell surface, internalization of h peptide, or any of a myriad of cellular events that could aﬀect expression of h-peptide toxicity. Lack of eﬀect of a compound could indicate that it is not blocking the toxic ‘‘site’’ on the peptide species, that it is not penetrating the cell, or simply that the compound is adsorbed, sequestered, or metabolized to an inactive form. In the VSMC system as in other cellular systems [142], Congo Red blocks both aggregation on the cell surface and h1–42 toxicity at 10 AM, roughly equivalent to the total added peptide concentration. For optimal eﬀect it must be added either before or along with the h peptide. Since the IC50 for an antiaggregation eﬀect on 30 AM peptide in vitro is 0.25 AM, nonspeciﬁc adsorption of the compound to cellular components and to h-peptide ﬁbrils may be mitigating the eﬀects. Congo Red and other polysulfonate/sulfate polyanions are known to displace proteins from binding sites on the cell surface [143]. Congo Red’s practical therapeutic potential is limited because it does not penetrate the cell membrane or the blood–brain barrier, and the azo linkages are susceptible to metabolism and carcinogenic liability. Such diﬃculties can be addressed by structural modiﬁcations in inhibitors that will likely also improve some of the pharmacokinetic properties in vivo. However, the connection between cellular eﬀects and desired in vivo properties of bioavailability and brain penetration is also not straightforward.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

V. ANIMAL MODELS OF BRAIN DEPOSITION OF INSOLUBLE PROTEINS The lack of animal models of Alzheimer’s disease that faithfully reproduce the human condition has signiﬁcantly retarded development of therapies that would be expected to have a signiﬁcant impact on the disease or its progression. Such models would provide a test bed for increasing conﬁdence that a given therapeutic regime would have the desired eﬀect in animals and would give insights that only experimental manipulation can provide into the disease process. Cholinomimetic approaches employed models of cholinergic deﬁcit to bring the current anticholinesterase drugs to market, even though they only modeled a subset of the pathology of AD [144]. Serious discovery and development of therapeutic agents directed at the deposition of h-amyloid ﬁbrils was put on a ﬁrm footing by the development of mice that overexpressed human APP, and depicted ﬁbrillar and nonﬁbrillar h1–40 and h1–42 in the appropriate brain regions. Various hAPP mutants responsible for early-onset familial AD have been the most eﬀective, particularly in combination with transgenic presenilin mutants, another familial AD locus. For a review of the hAPP mouse and other brain amyloidosis models see Walker [145]. Again, these are only partial models of AD, since no signiﬁcant neuronal cell death has been observed with these mice. Importantly, the Hsiao mouse, Tg 2576 (human APP695, Swedish mutation under the control of the mouse prion promoter), has been available to both commercial and academic laboratories, and thus there is a considerable shared pool of information and experience with this model. Direct comparisons with the other hAPP mice are limited because those models are not generally available. The Hsiao mouse shows robust and increasing deposition of h peptide from age 9 months onward, as well as dystrophic neurites, reactive astrogliosis, microglial activation around senile plaques, and some phosphorylated (but not tangled) tau immunoreactivity, but no detectable neuronal cell death or reduction in synaptic counts, despite plaque densities approaching that of clinical AD. An important validation of the mice that overexpress human mutant hAPP as a platform for testing therapeutics targeting h-peptide deposition has been provided by the Elan company, using their PDAPP mouse [146]. Immunization of the mice, either at an early age or after plaques had formed, resulted in clearance of immunoreactive plaques and peptide from the subjects’ brains. Although the elucidation of the mechanism explaining

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

this potential therapy has not been reported, the experiments establish that deposition of the peptide can be interrupted, and even reversal of preformed plaques is possible in this type of animal model. Human trials of the safety and eﬃcacy of the immunization protocol are to start in the near future. The plaques are thus dynamic structures, and therapeutics that interfere with Ah deposition or production would be predicted to reduce brain h-amyloid load [147]. Similar partial neurodegenerative disease animal models involving insoluble protein deposition have been developed for Huntington’s disease [148], spinocerebellar ataxia type 1 (SCA1) [149,150], and Machado–Joseph disease (MJD; SCA3) [151], all trinucleotide repeat disorders. The protein deposits, consisting primarily of the polyglutamine tract, accumulate intracellularly, eventually collecting in the nucleus, where they are thought to disrupt nuclear function both in humans and in the animal models. As for Alzheimer’s disease and the Ah peptide and tau proteins, there is considerable controversy about the relevance of the deposits to the disease process because observed neuronal cell loss does not overlap entirely with visible inclusions. The arguments in favor of relevance are that the toxic species may be smaller oligomeric species not observed by microscopy and that in some situations the deposits may serve a protective function by sequestering potentially toxic material. The Huntington’s disease models have been studied in considerable detail. The pathology bears a striking resemblance to the human disease in a number of respects, although polyglutamine overexpression is not a complete model. The Bates R6 mice expressing exon I of the human Huntingtin protein, consisting of the N-terminal 17 amino acids + a pathological number (115–156 CAGs) of glutamines + 52 more amino acids under control of the human promoter, develop age-dependent, brainregion-speciﬁc cell loss accompanied by nuclear inclusions and behavioral and motor abnormalities reﬂecting those in the human disease, leading eventually to death [152]. Time of onset, severity of the symptoms, and length of disease are dependent on the number of glutamine repeats in the observed human pathological range. A longer term mouse model with the full-length human Huntingtin under its natural promoter, which lacks the potential diabetic condition of the Bates mouse, may provide a more realistic picture of the HD process, although it would be less useful for rapid testing of potential therapies. Mice expressing high levels of human a-synuclein under the control of the human PDGFh promoter developed intracellular nuclear

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

and cytoplasmic inclusions that were immunoreactive with antibody speciﬁc for the human a-synuclein [153]. As early as 2 to 3 months of age in the transgenics, inclusions were found in the deeper layers of the neocortex, the CA3 region of the hippocampus, the olfactory bulb, and occasionally the substantia nigra, regions typically aﬀected in Lewy body disease. While nigral tyrosine hydroxylase–positive cell number was similar to nontransgenic littermates, TH-positive nerve terminals were signiﬁcantly reduced, as were TH immunoreactivity and enzymatic activity in 12-month-old animals. In the 12-month-old mice, neurological impairment similar to that found in Parkinson’s disease was demonstrable in rotorod performance. Human prion disease models have also been developed in mice [154,155]. Crossing the species barrier into an experimentally accessible animal system, the prions responsible for Creutzfeldt–Jakob disease, new variant CJD, Gerstmann–Straussler–Scheinker disease, and fatal familial insomnia produce a reproducible time-dependent neuronal degeneration leading to death.

VI. CLINICAL TRIALS FOR AD TESTING OF POSSIBLE DISEASE-MODIFYING AGENTS While testing of amyloid aggregation inhibitors against AD in human subjects is a way oﬀ, it is worth considering how such trials should be conducted to establish clinical eﬃcacy. With all the genetic and biochemical evidence that the h peptide is implicated in AD pathology, it very well may not be the only relevant pathology in all patients for this very complex disease. It remains distinctly possible that removal of all amyloid plaques and/or h peptide from the brains of AD patients will not restore cognitive function in advanced stages of the disease. Thus, a paradigm treating signiﬁcantly cognitively impaired patients to look for a leveling oﬀ in their decline or a reversal to normal may not show the desired treatment eﬀect with an aggregation inhibitor. To be fair, it may not show eﬀects with any treatment if too many neurons have died or become dysfunctional. The cholinomimetic therapies were symptomatic treatments designed to supplement function (acetylcholine) that had been lost without regard for the process that caused that functional loss. In attacking what is believed to be a fundamental process in disease progression, other measures may be needed to reverse the degeneration that has already occurred, assuming that not too much

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

damage has been inﬂicted. Disease-modifying therapies are most likely to inﬂuence progression to disease and/or delay onset of symptoms. Proper assessment of the eﬀects of a therapy will require clinical trial designs that make the appropriate measurements. It will be important to assess plaque load and/or h-peptide level in patients treated with aggregation inhibitors or other modes of reducing either brain hpeptide content or its eﬀects in addition to the classical cognitive end points. Whatever the outcome of the trials, in interpreting the results for the development of new generations of therapeutics it is important to determine whether the therapy accomplished what it was designed to do—reduce amyloid peptide deposition. If h peptide is eliminated but no therapeutic beneﬁt is observed, we should conclude that the h peptide is not the major player—at least in the patient population selected for study. The answer will be important in justifying future pharmaceutical investment as well as in guiding future research for eﬀective therapies against AD.

REFERENCES 1. Davies SW, Beardsall K, Turmaine M, DiFiglia M, Aronin N, Bates GP. Are neuronal intranuclear inclusions the common neuropathology of triplet-repeat disorders with polyglutamine-repeat expansions? Lancet 1998; 351:131–133. 2. Kakizuka A. Protein precipitation: a common etiology in neurodegenerative disorders? Trends Genet 1998; 14:396–402. 3. Kim TW, Tanzi RE. Neuronal intranuclear inclusions in polyglutamine diseases: nuclear weapons or nuclear fallout? Neuron 1998; 21: 657–659. 4. Koshy BT, Zoghbi HY. The CAG/polyglutamine tract diseases: gene products and molecular pathogenesis. Brain Pathol 1997; 7:927–942. 5. Rubinaztein DC, Wyttenbach A, Rankin J. Intracellular inclusions, pathological markers in diseases caused by expanded polyglutamine tracts? J Med Genet 1999; 36:265–270. 6. Scherzinger E, Lurz R, Turmaine M, Mangiarini L, Hollenbach B, Hasenbank R, Bates GP, Davies SW, Lehrach H, Wanker EE. Huntingtinencoded polyglutamine expansions form amyloid-like protein aggregates in vitro and in vivo. Cell 1997; 90:549–558. 7. Lathrop RH, Casale M, Tobias DJ, Marsh JL, Thompson LM. Modeling protein homopolymeric repeats: possible polyglutamine structural motifs for Huntington’s disease. Proc Int Conf Intelligent Syst Molec Biol 1998; 6:105–114.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

8. Altschuler EL, Hud NV, Mazrimas JA, Rupp B. Random coil conformation for extended polyglutamine stretches in aqueous soluble monomeric peptides. J Peptide Res 1997; 50:73–75. 9. Scherzinger E, Sittler A, Schweiger K, Heiser V, Lurz R, Hasenbank R, Bates GP, Lehrach H, Wanker EE. Self-assembly of polyglutamine-containing huntingtin fragments into amyloid-like ﬁbrils: implications for Huntington’s disease pathology. Proc Natl Acad Sci USA 1999; 96:4604–4609. 10. Bates GP, Mangiarini L, Davies SW. Transgenic mice in the study of polyglutamine repeat expansion diseases. Brain Pathol 1998; 8: 699–714. 11. Perez MK, Paulson HL, Pendse SJ, Saionz SJ, Bonini NM, Pittman RN. Recruitment and the role of nuclear localization in polyglutaminemediated aggregation. J Cell Biol 1998; 143:1457–1470. 12. Gourﬁnkel AI, Cancel G, Duyckaerts C, Faucheux B, Hauw JJ, Trottier Y, Brice A, Agid Y, Hirsch EC. Neuronal distribution of intranuclear inclusions in Huntington’s disease with adult onset. Neuroreport 1998; 9:1823–1826. 13. Becher MW, Kotzuk JA, Sharp AH, Davies SW, Bates GP, Price DL, Ross CA. Intranuclear neuronal inclusions in Huntington’s disease and dentatorubral and pallidoluysian atrophy: correlation between the density of inclusions and IT15 CAG triplet repeat length. Neurobiol Dis 1998; 4:387–397. 14. DiFiglia M, Sapp E, Chase KO, Davies SW, Bates GP, Vonsattel JP, Aronin N. Aggregation of Huntingtin in neuronal intranuclear inclusions and dystrophic neurites in brain. Science 1997; 277:1990–1993. 15. Duyckaerts C, Durr A, Cancel G, Brice A. Nuclear inclusions in spinocerebellar ataxia type 1. Acta Neuropathol Berl 1999; 97:201–207. 16. Hayashi Y, Kakita A, Yamada M, Egawa S, Oyanagi S, Naito H, Tsuji S, Takahashi H. Hereditary dentatorubral–pallidoluysian atrophy: ubiquitinated ﬁlamentous inclusions in the cerebellar dentate nucleus neurons. Acta Neuropathol Berl 1998; 95:479–482. 17. Holmberg M, Duyckaerts C, Durr A, Cancel G, Gourﬁnkel An I, Damier P, Faucheux B, Trottier Y, Hirsch EC, Agid Y, Brice A. Spinocerebellar ataxia type 7 (SCA7): a neurodegenerative disorder with neuronal intranuclear inclusions. Hum Mol Genet 1998; 7:913–918. 18. Lieberman AP, Trojanowski JQ, Leonard DG, Chen KL, Barnett JL, Leverenz JB, Bird TD, Robitaille Y, Malandrini A, Fischbeck KH. Ataxin 1 and ataxin 3 in neuronal intranuclear inclusion disease. Ann Neurol 1999; 46:271–273. 19. Schmidt T, Landwehrmeyer GB, Schmitt I, Trottier Y, Auburger G, Laccone F, Klockgether T, Volpel M, Epplen JT, Schols L, Riess O. An isoform of ataxin-3 accumulates in the nucleus of neuronal cells in aﬀected brain regions of SCA3 patients. Brain Pathol 1998; 8: 669–679.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

20.

21.

22.

23.

24. 25.

26.

27.

28.

29. 30.

31. 32.

33. 34.

Goedert M. Filamentous nerve cell inclusions in neurodegenerative diseases: tauopathies and a-synucleinopathies. Philos Trans R Soc Lond B 1999; 354:1101–1118. Jellinger KA. Movement disorders with tau protein cytoskeletal pathology. Parkinson’s disease. In: Stern GM, ed. Advances in Neurology. Vol. 80. Philadelphia: Lippincott, Williams & Wilkins, 1999:393–311. Tolnay M, Probst A. Review: tau protein pathology in Alzheimer’s disease and related disorders. Neuropathol Appl Neurobiol 1999; 25: 171–187. Schweers O, Schonbrunn Hanebeck E, Marx A, Mandelkow E. Structural studies of tau protein and Alzheimer paired helical ﬁlaments show no evidence for beta-structure. J Biol Chem 1994; 269:24290–24297. Delacourte A, Buee L. Normal and pathological tau proteins as factors for microtubule assembly. Int Rev Cytol 1997; 171:167–224. Lu PJ, Wulf G, Zhou XZ, Davies P, Lu KP. The prolyl isomerase Pin1 restores the functions of Alzheimer-associated phosphorylated tau protein [see comments]. Nature 1999; 399:784–788. Foster J, Hunter N. Transmissible spongiform encephalopathies: transmission, mechanism of disease, and persistence. Curr Opin Microbiol 1998; 1:442–447. Kocisko DA, Priola SA, Raymond GJ, Chesebro B, Lansbury PT Jr, Caughey B. Species speciﬁcity in the cell-free conversion of prion protein to protease-resistant forms: a model for the scrapie species barrier. Proc Natl Acad USA 1995; 92:3923–3927. Kellershohn N, Laurent M. Species barrier in prion diseases: a kinetic interpretation based on the conformational adaptation of the prion protein. Biochem J 1998; 334:539–545. Cohen FE, Prusiner SB. Pathologic conformations of prion proteins. Annu Rev Biochem 1998; 67:793–819. Come JH, Fraser PE, Lansbury PT, Jr. A kinetic model for amyloid formation in the prion diseases: importance of seeding. Proc Natl Acad Sci USA 1993; 90:5959–5963. Lansbury PT Jr. Yeast prions: inheritance by seeded protein polymerization? Curr Biol 1997; 7:R617–R619. Saborio GP, Soto C, Kascsak RJ, Levy E, Kascsak R, Harris DA, Frangione B. Cell-lysate conversion of prion protein into its proteaseresistant isoform suggests the participation of a cellular chaperone. Biochem Biophys Res Commun 1999; 258:470–475. Lansbury PT Jr, Caughey B. The chemistry of scrapie infection: implications of the ‘ice 9’ metaphor. Chem Biol 1995; 2:1–5. Edskes HK, Gray VT, Wickner RB. The [URE3] prion is an aggregated form of Ure2p that can be cured by overexpression of Ure2p fragments. Proc Natl Acad Sci USA 1999; 96:1498–1503.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

35.

36.

37.

38.

39.

40.

41. 42.

43.

44.

45.

46.

47.

DeArmond SJ, Ironside JW. Neuropathology of prion diseases. In: Prusiner SB, ed. Prion Biology and Diseases. Vol. 38. Cold Spring Harbor, NY: Cold Spring Harbor Press, 1999:585–652. Safar J, Wille H, Itri V, Groth D, Serban H, Torchia M, Cohen FE, Prusiner SB. Eight prion strains have PrP(Sc) molecules with diﬀerent conformations [see comments]. Nat Med 1998; 4:1157–1165. Caughey B, Raymond GJ, Bessen RA. Strain-dependent diﬀerences in beta-sheet conformations of abnormal prion protein. J Biol Chem 1998; 273:32230–32235. Keohane C. The human prion diseases. A review with special emphasis on new variant CJD and comments on surveillance. Clin Exp Pathol 1999; 47:125–132. Farrer M, Gwinn-Hardy K, Hutton M, Hardy J. The genetics of disorders with synuclein pathology and parkinsonism. Hum Mol Genet 1999; 8:1901–1905. Polymeropoulos MH, Lavedan C, Leroy E, Ide SE, Dehejia A, Dutra A, Pike B, Root H, Rubenstein J, Boyer R, Stenroos ES, Chandrasekharappa S, Athanassiadou A, Papapetropoulos T, Johnson WG, Lazzarini AM, Duvoisin RC, Di Iorio G, Golbe LI, Nussbaum RL. Mutation in the alpha-synuclein gene identiﬁed in families with Parkinson’s disease. Science 1997; 276:2045–2047. Polymeropoulos MH. Autosomal dominant Parkinson’s disease and alphasynuclein. Ann Neurol 1998; 44:S63–S64. Wakabayashi K, Hansen LA, Vincent I, Mallory M, Masliah E. Neuroﬁbrillary tangles in the dentate granule cells of patients with Alzheimer’s disease, Lewy body disease and progressive supranuclear palsy. Acta Neuropathol Berl 1997; 93:7–12. Baba M, Nakajo S, Tu PH, Tomita T, Nakaya K, Lee VM, Trojanowski JQ, Iwatsubo T. Aggregation of alpha-synuclein in Lewy bodies of sporadic Parkinson’s disease and dementia with Lewy bodies. Am J Pathol 1998; 152:879–884. Spillantini MG, Crowther RA, Jakes R, Cairns NJ, Lantos PT, Goedert M. Filamentous a-synuclein inclusions link multiple system atrophy with Parkinson’s disease and dementia with Lewy bodies. Neurosci Lett 1998; 251:205–208. Weinreb PH, Zhen W, Poon AW, Conway KA, Lansbury PT Jr. NACP, a protein implicated in Alzheimer’s disease and learning, is natively unfolded. Biochemistry 1996; 35: 13709–13715. Han H, Weinreb PH, Lansbury PT Jr. The core Alzheimer’s peptide NAC forms amyloid ﬁbrils which seed and are seeded by beta-amyloid: is NAC a common trigger or target in neurodegenerative disease? Chem Biol 1995; 2:163–169. Wood SJ, Wypych J, Steavenson S, Louis JC, Citron M, Biere AL. a-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

48.

49.

50.

51. 52.

53.

54. 55.

56.

57.

58.

59.

60.

Synuclein ﬁbrillogenesis is nucleation-dependent. Implications for the pathogenesis of Parkinson’s disease. J Biol Chem 1999; 274:19509–19512. Narhi L, Wood SJ, Steavenson S, Jiang Y, Wu GM, Anaﬁ D, Kaufman SA, Martin F, Sitney K, Denis P, Louis JC, Wypych J, Biere AL, Citron M. Both familial Parkinson’s disease mutations accelerate alpha-synuclein aggregation. J Biol Chem 1999; 274:9843–9846. Durham HD, Roy J, Dong L, Figlewicz DA. Aggregation of mutant Cu/ Zn superoxide dismutase proteins in culture model of ALS. J Neuropathol Exp Neurol 1997; 56:523–530. Chou SM, Wang HS, Taniguchi A. Role of SOD-1 and nitric oxide/cyclic GMP cascade on neuroﬁlament aggregation in ALS/MND. J Neurol Sci 1996; 139 suppl:16–26. Cleveland DW. From Charcot to SOD1: mechanisms of selective motor neuron death in ALS. Neuron 1999; 24:525–520. Vidal R, Fragione B, Rostagno A, Mead S, Revesz T, Plant G, Ghiso J. A stop-codon mutation in the BRI gene associated with familial British dementia. Nature 1999; 399:776–781. Davis RL, Shimpton AE, Holohan PD, Bradshaw C, Feiglin D, Collins GH, Sonderegger P, Kinter J, Becker LM, Lacbawan F, Krasnewich D, Muenke M, Lawrence DA, Yerby MS, Shaw CM, Gooptu B, Elliott PR, Finch JT, Carrell RW, Lomas DA. Famillial dementia caused by polymerization of mutant neuroserpin. Nature 1999; 401:376–379. Carrell RW, Lomas DA. Conformational disease. Lancet 1997; 350:134–138. Baures PW, Peterson SA, Kelly JW. Discovering transthyretin amyloid ﬁbril inhibitors by limited screening. Bioorg Med Chem 1998; 6: 1389–1401. Caughey WS, Raymond LD, Horiuchi M, Caughey B. Inhibition of protease-resistant prion protein formation by porphyrins and phthalocyanines. Proc Natl Acad Sci USA 1998; 95:12117–12122. Peterson SA, Klabunde T, Lashuel HA, Purkey H, Sacchettini JC, Kelly JW. Inhibiting transthyretin conformation changes that lead to amyloid ﬁbril formation. Proc Natl Acad Sci USA 1998; 95:12956–12960. Tagliavini F, McArthur RA, Canciani B, Giaccone G, Porro M, Bugiani M, Lievens PM, Bugiani O, Peri E, Dall’Ara P, Rocchi M, Poli G, Forloni G, Bandiera T, Varasi M, Suarato A, Cassutti P, Cervini MA, Lansen J, Salmona M, Post C. Eﬀectiveness of anthracycline against experimental prion disease in Syrian hamsters. Science 1997; 276:1119– 1122. Oza VB, Petrassi HM, Purkey HE, Kelly JW. Synthesis and evaluation of anthranilic acid-based transthyretin amyloid ﬁbril inhibitors. Bioorg Med Chem Lett 1999; 9:1–6. Oosawa F, Akakura S. Kinetics of Polymerization. Thermodynamics of the Polymerization of Protein. New York: Academic Press, 1975:41–55.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

61. Oosawa F, Higashi S. Statistical thermodynamics of polymerization and polymorphism of protein. In: Snell FM, ed. Progress in Theoretical Biology. New York: Academic Press, 1967:79–164. 62. Bishop MF, Ferrone FA. Kinetics of nucleation-controlled polymerization. A perturbation treatment for use with a secondary pathway. Biophys J 1984; 46:631–644. 63. Flyvbjerg H, Jobs E, Leibler S. Kinetics of self-assembling microtubules: an ‘‘inverse problem’’ in biochemistry. Proc Natl Acad Sci USA 1996; 93:5975–5979. 64. Endres GF, Ehrenpreis S, Scheraga HA. Covalent bonding in the reversible polymerization of ﬁbrin monomer. Biochim Biophys Acta 1965; 104:620–623. 65. Bark N, Foldes Papp Z, Rigler R. The incipient stage in thrombin-induced ﬁbrin polymerization detected by FCS at the single molecule level. Biochem Biophys Res Commun 1999; 260:35–41. 66. Mosesson MW. Fibrinogen and ﬁbrin polymerization: appraisal of the binding events that accompany ﬁbrin generation and ﬁbrin clot assembly. Blood Coagul Fibrinol 1997; 8:257–267. 67. Schutt CE, Myslik JC, Rozycki MD, Goonesekere NC, Lindberg U. The structure of crystalline proﬁlin-beta-actin. Nature 1993; 365:810–816. 68. Weeds AG, Gooch J, McLauglin P, Pope B, Bengtsdotter M, Karlsson R. Identiﬁcation of the trapped calcium in the gelsolin segment 1–actin complex: implications for the role of calcium in the control of gelsolin activity. FEBS Lett 1995; 360:227–230. 69. Lowe J. Crystal structure determination of FtsZ from Methanococcus jannaschii. J Struct Biol 1998; 124:235–243. 70. Ferrone FA, Hofrichter J, Eaton WA. Kinetics of sickle hemoglobin polymerization II. A double nucleation mechanism. J Mol Biol 1985; 183:611–633. 71. Ferrone FA, Hofrichter J, Eaton WA. Kinetics of sickle hemoglobin polymerization. I: Studies using temperature jump and laser photolysis techniques. J Mol Biol 1985; 183:591–610. 72. Eaton WA, Hofrichter J. The biophysics of sickle cell hydroxyurea therapy. Science 1995; 268:1142–1143. 73. Lashuel HA, Lai Z, Kelly JW. Characterization of the transthyretin acid denaturation pathways by analytical ultracentrifugation: implications for wild-type, V30M, and L55P amyloid ﬁbril formation. Biochemistry 1998; 37:17851–17864. 74. Lai Z, Colon W, Kelly JW. The acid-mediated denaturation pathway of transthyretin yields a conformational intermediate that can self-assemble into amyloid. Biochemistry 1996; 35:6470–6482. 75. Kelly JW, Colon W, Lai Z, Lashuel HA, McCulloch J, McCutchen SL, Miroy GJ, Peterson SA. Transthyretin quaternary and tertiary structural

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

76.

77.

78.

79.

80.

81.

82.

83. 84.

85.

86.

87.

changes facilitate misassembly into amyloid. Adv Protein Chem 1997; 50:161–181. Bonifacio MJ, Sakaki Y, Saraiva MJ. ‘‘In vitro’’ amyloid ﬁbril formation from transthyretin: the inﬂuence of ions and the amyloidogenicity of TTR variants. Biochim Biophys Acta 1996; 1316:35–42. Yamada T, Kluve Beckerman B, Liepnieks JJ, Benson MD. Fibril formation from recombinant human serum amyloid A. Biochim Biophys Acta 1994; 1226:323–329. Friedhoﬀ P, Schneider A, Mandelkow EM, Mandelkow E. Rapid assembly of Alzheimer-like paired helical ﬁlaments from microtubule-associated protein tau monitored by ﬂourescence in solution. Biochemistry 1998; 37:10223–10230. Schweers O, Mandelkow EM, Biernat J, Mandelkow E. Oxidation of cysteine-322 in the repeat domain of microtubule-associated protein tau controls the in vitro assembly of paired helical ﬁlaments. Proc Natl Acad Sci USA 1995; 92:8463–8467. Wilson DM, Binder LI. Free fatty acids stimulate the polymerization of tau and amyloid beta peptides. In vitro evidence for a common eﬀector of pathogenesis in Alzheimer’s disease. Am J Pathol 1997; 150:2181–2195. Kayed R, Bernhagen J, Greenﬁeld N, Sweimeh K, Brunner H, Voelter W, Kapurniotu A. Conformational transitions of islet amyloid polypeptide (IAPP) in amyloid formation in vitro. J Mol Biol 1999; 287:781–796. Charge SB, de Koning EJ, Clark A. Eﬀect of pH and insulin on ﬁbrillogenesis of islet amyloid polypeptide in vitro. Biochemistry 1995; 34:14588–14593. Wetzel R. Domain stability in immunoglobulin light chain deposition disorders. Adv Protein Chem 1997; 50:183–242. Helms LR, Wetzel R. Speciﬁcity of abnormal assembly in immunoglobulin light chain deposition disease and amyloidosis. J Mol Biol 1996; 257:77–86. Stevens PW, Raﬀen R, Hanson DK, Deng YL, Berrios Hammond M, Westholm FA, Murphy C, Eulitz M, Wetzel R, Solomon A, et al. Recombinant immunoglobulin variable domains generated from synthetic genes provide a system for in vitro characterization of light-chain amyloid proteins. Protein Sci 1995; 4:421–432. Hollenbach B, Scherzinger E, Schweiger K, Lurz R, Lehrach H, Wanker EE. Aggregation of truncated GST-HD exon 1 fusion proteins containing normal range and expanded glutamine repeats. Philos Trans R Soc Lond B Biol Sci 1999; 354:991–994. Burdick D, Soreghan B, Kwon M, Kosmoski J, Knauer M, Henschen A, Yates J, Cotman C, Glabe C. Assembly and aggregation properties of synthetic Alzheimer’s A4/beta amyloid peptide analogs. J Biol Chem 1992; 267:546–554.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

88. Fraser PE, McLachlan DR, Surewicz WK, Mizzen CA, Snow AD, Nguyen JT, Kirschner DA. Conformation and ﬁbrillogenesis of Alzheimer A beta peptides with selected substitution of charged residues. J Mol Biol 1994; 244:64–73. 89. Garzon-Rodriguez W, Sepulveda-Becerra M, Milton S, Glabe CG. Soluble amyloid Abeta-(1–40) exists as a stable dimer at low concentrations. J Biol Chem 1997; 272:21037–21044. 90. Soto C, Castano EM, Kumar RA, Beavis RC, Frangione B. Fibrillogenesis of synthetic amyloid-beta peptides is dependent on their initial secondary structure. Neurosci Lett 1995; 200:105–108. 91. Soto C, Castano EM, Frangione B, Inestrosa NC. The alpha-helica to beta-strand transition in the amino-terminal fragment of the amyloid betapeptide modulates amyloid formation. J Biol Chem 1995; 270:3063–3067. 92. Evans KC, Berger EP, Cho CG, Weisgraber KH, Lansbury PT Jr. Apolipoprotein E is a kinetic but not a thermodynamic of amyloid formation: implications for the pathogenesis and treatment of Alzheimer disease. Proc Natl Acad Sci USA 1995; 92:763–767. 93. Harper JD, Wong SS, Lieber CM, Lansbury PT Jr. Assembly of A amyloid protoﬁbrils: an in vitro model for a possible early event in Alzheimer’s disease. Biochemistry 1999; 38:8972–8980. 94. Harper JD, Lieber CM, Lansbury PT Jr. Atomic force microscopic imaging of seeded ﬁbril formation and ﬁbril branching by the Alzheimer’s disease amyloid-beta protein. Chem Biol 1997; 4:951–959. 95. Harper JD, Lansbury PT Jr. Models of amyloid seeding in Alzheimer’s disease and scrapie: mechanistic truths and physiological consequences of the time-dependent solubility of amyloid proteins. Annu Rev Biochem 1997; 66:385–407. 96. Jarrett JT, Lansbury PT Jr. Amyloid ﬁbril formation requires a chemically discriminating nucleation event: studies of an amyloidogenic sequence from the bacterial protein OsmB. Biochemistry 1992; 31: 12345–12352. 97. LeVine H III. Screening for pharmacologic inhibitors of amyloid ﬁbril formation. Methods Enzymol 1999; 309:467–476. 98. Lomakin A, Teplow DB, Kirschner DA, Benedek GB. Kinetic theory of ﬁbrillogenesis of amyloid beta-protein. Proc Natl Acad Sci USA 1997; 94:7942–7947. 99. Shen CL, Scott GL, Merchant F, Murphy RM. Light scattering analysis of ﬁbril growth from the amino-terminal fragment beta(1–28) of beta-amyloid peptide. Biophys J 1993; 65:2383–2395. 100. Tomski SJ, Murphy RM. Kinetics of aggregation of synthetic betaamyloid peptide. Arch Biochem Biophys 1992; 294:630–638. 101. Lomakin A, Chung DS, Benedek GB, Kirschner DA, Teplow DB. On the nucleation and growth of amyloid beta protein ﬁbrils: detection of nuclei

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

102.

103.

104.

105.

106.

107.

108.

109.

110. 111.

112.

113. 114.

115.

and quantitation of rate constants. Proc Natl Acad Sci USA 1996; 93:1125–1129. Stine WB Jr, Snyder SW, Ladror US, Wade WS, Miller MF, Perun TJ, Holzman TF, Kraﬀt GA. The nanometer-scale structure of amyloid-beta visualized by atomic force microscopy. J Protein Chem 1996; 15:193– 203. Harper JD, Wong SS, Lieber CM, Lansbury PT. Observation of metastable Abeta amyloid protoﬁbrils by atomic force microscopy. Chem Biol 1997; 4:119–125. Kowalewski T, Holtzman DM. In situ atomic force microscopy study of Alzheimer’s beta-amyloid peptide on diﬀerent substrates: new insights into mechanism of beta-sheet formation. Proc Natl Acad USA 1999; 96:3688–3693. Walsh DM, Hartley DM, Kusomoto Y, Fezoui Y, Condron MM, Lomakin A, Benedek GB, Selkoe DJ, Teplow DB. Amyloid h-protein ﬁbrillogenesis. Structure and biological activity of protoﬁbrillar intermediates. J Biol Chem 1999; 274:25949–25952. Walsh DM, Lomakin A, Benedek GB, Condron MM, Teplow DB. Amyloid beta-protein ﬁbrillogenesis. Detection of a protoﬁbrillar intermediate. J Biol Chem 1997; 272:22364–22372. Blake CC, Serpell LC, Sunde M, Sandgren O, Lundgren E. A molecular model of the amyloid ﬁbril. CIBA Found Symp 1996; 199:6–15; discussion 15–21, 40–16. Inouye H, Fraser PE, Kirschner DA. Structure of beta-crystalline assemblies formed by Alzheimer beta-amyloid protein analogues: analysis by X-ray diﬀraction. Biophys J 1993; 64:502–519. Li L, Darden TA, Bartolotti L, Kominos D, Pederson LG. An atomic model for the pleated beta-sheet structure of Abeta amyloid protoﬁlaments. Biophys J 1999; 76:2871–2878. Sunde M, Blake C. The structure of amyloid ﬁbrils by electron microscopy and X-ray diﬀraction. Adv. Protein Chem. 1997; 50:123–159. Sunde M, Serpell LC, Bartlam M, Fraser PE, Pepys MB, Blake CC. Common core structure of amyloid ﬁbrils by synchrotron X-ray diﬀraction. J Mol Biol 1997; 273:729–739. Nguyen JT, Inouye H, Baldwin MA, Fletterick RJ, Cohen FE, Prusiner SB, Kirschner DA. X-ray diﬀraction of scrapie prion rods and PrP peptides. J Mol Biol 1995; 252:412–422. Lazo ND, Downing DT. Beta-helical ﬁbrils from a model peptide. Biochem Biophys Res Commun 1997; 235:675–679. Goldsbury C, Kistler J, Aebi U, Arvinte T, Cooper GJ. Watching amyloid ﬁbrils grow by time-lapse atomic force microscopy. J Mol Biol 1999; 285:33–39. Wall J, Schell M, Murphy C, Hrncic R, Stevens FJ, Solomon A.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

116. 117. 118.

119.

120.

121.

122. 123. 124.

125.

126.

127.

128.

Thermodynamic instability of human E6 light chains: correlation with ﬁbrillogenicity. Biochemistry 1999; 38:14101–14108. Ferrone F. Analysis of protein aggregation kinetics. Methods Enzymol 1999; 309:256–274. Naiki H, Nakakuki K. First-order kinetic model of Alzheimer’s betaamyloid ﬁbril extension in vitro. Lab Invest 1996; 74:374–383. Maggio JE, Stimson ER, Ghilardi JR, Allen CJ, Dahl CE, Whitcomb DC, et al. Reversible in vitro growth of Alzheimer disease h-amyloid plaques by deposition of labeled amyloid peptide. Proc Natl Acad Sci USA 1992; 89:5462–5466. Esler WP, Stimson ER, Ghilardi JR, Felix AM, Lu Y-A, Vinters HV, Mantyh PJ, Maggio JE. Ah deposition inhibitor screen using synthetic amyloid. Nat Biotechnol 1997; 15:258–263. Esler WP, Stimson ER, Ghilardi JR, Lu YA, Felix AM, Vinters HV, Mantyh PW, Lee JP, Maggio JE. Point substitution in the central hydrophobic cluster of a human beta-amyloid congener disrupts peptide folding and abolishes plaque competence. Biochemistry 1996; 35:13914–13921. Lee JP, Stimson ER, Ghilardi JR, Mantyh PW, Lu Y-A, Felix AM, Llanos W, Behbin A, Cummings M, Criekinge MV, Timms W, Maggio JE. 1H NMR of Ah amyloid peptide congeners in water solution. Conformational changes correlate with plaque competence. Biochemistry 1995; 34:5191–5200. LeVine H III. Soluble multimeric Alzheimer beta(1–40) preamyloid complexes in dilute solution. Neurobiol Aging 1995; 16:755–764. Dudek SM, Johnson GV. Transglutaminase facilitates the formation of polymers of the beta-amyloid peptide. Brain Res 1994; 651:129–133. Rasmussen LK, Sorensen ES, Petersen TE, Gliemann J, Jensen PH. Identiﬁcation of glutamine and lysine residues in Alzheimer amyloid beta A4 peptide responsible for transglutaminase-catalysed homopolymerization and cross-linking to alpha 2M receptor. FEBS Lett 1994; 338:161–166. Nybo M, Svehag SE, Holm Nielsen E. An ultrastructural study of amyloid intermediates in A beta(1–42) ﬁbrillogenesis. Scand J Immunol 1999; 49:219–223. Poduslo JF, Curran GL, Kumar A, Frangione B, Soto C. Beta-sheet breaker peptide inhibitor of Alzheimer’s amyloidogenesis with increased blood–brain barrier permeability and resistance to proteolytic degradation in plasma. J Neurobiol 1999; 39:371–382. Pallitto MM, Ghanta J, Heizelman P, Kiessling LL, Murphy RM. Recognition sequence design for peptidyl modulators of beta-amyloid aggregation and toxicity. Biochemistry 1999; 38:3570–3578. Tjernberg LO, Lilliehook C, Callaway DJ, Naslund J, Hahne S, Thyberg J,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

129.

130.

131.

132.

133.

134.

135.

136.

137.

138.

139.

140.

Terenius L, Nordstedt C. Controlling amyloid beta-peptide ﬁbril formation with protease-stable ligands. J Biol Chem 1997; 272:12601–12605. Ghanta J, Shen CL, Kiessling LL, Murphy RM. A strategy for designing inhibitors of beta-amyloid toxicity. J Biol Chem 1996; 271: 29525–29528. Howlett DR, Perry AE, Godfrey F, Swatton JE, Jennings KH, Spitzfaden C, Wadsworth H, Wood SJ, Markwell RE. Inhibition of ﬁbril formation in beta-amyloid peptide by a novel series of benzofurans. Biochem J 1999; 340:283–289. Lambert MP, Barlow AK, Chromy BA, Edwards C, Freed R, Liosatos M, Morgan TE, Rozovsky I, Trommer B, Viola KL, Wals P, Zhang C, Finch CE, Kraﬀt GA, Klein WL. Diﬀusible, nonﬁbrillar ligands derived from Abeta(1–42) are potent central nervous system neurotoxins. Proc Natl Acad Sci USA 1998; 95:6448–6453. Roher AE, Chaney MO, Kuo YM, Webster SD, Stine WB, Haverkamp LJ, Woods AS, Cotter RJ, Tuohy JM, Kraﬀt GA, Bonnell BS, Emmerling MR. Morphology and toxicity of Abeta-(1–42) dimer derived from neuritic and vascular amyloid deposits of Alzheimer’s disease. J Biol Chem 1996; 271:20631–20635. Liu Y, Schubert D. Cytotoxic amyloid peptides inhibit cellular 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MMT) reduction by enhancing MTT formazan exocytosis. J Neurochem 1997; 69:2285–2293. Patel AJ, Gunasekera S, Jen A, Rohan de Silva HA. h-Amyloid-mediated inhibition of redox activity (MMT reduction) is not an indicator of astroglial degeneration. Neuroreport 1996; 7:2026–2030. Isobe I, Michikawa M, Yanagisawa K. Enhancement of MTT, a tetrazolium salt, exocytosis by amyloid beta-protein and chloroquine in cultured rat astrocytes. Neurosci Lett 1999; 266:129–132. Abe K, Saito H. Amyloid beta protein inhibits cellular MTT reduction not by suppression of mitochondrial succinate dehydrogenase but by acceleration of MTT formazan exocytosis in cultured rat cortical astrocytes. Neurosci Res 1998; 31:295–305. Van Nostrand WE, Melchor JP, Ruﬃni L. Pathologic amyloid betaprotein cell surface ﬁbril assembly on cultured human cerebrovascular smooth muscle cells. J Neurochem 1998; 70:216–223. Wisniewski HM, Frackowiak J, Mazur Kolecka B. In vitro production of beta-amyloid in smooth muscle cells isolated from amylid angiopathyaﬀected vessels. Neurosci Lett 1995; 183:120–123. Frackowiak J, Mazur Kolecka B, Wisniewski HM, Potempska A, Carroll RT, Emmerling MR, Kim KS. Secretion and accumulation of Alzheimer’s beta-protein by cultured vascular smooth muscle cells from old and young dogs. Brain Res 1995; 676:225–230. Crawford F, Soto C, Suo Z, Fang C, Parker T, Sawar A, Frangione B,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

141.

142.

143.

144. 145. 146.

147.

148.

149.

150. 151.

152.

153.

Mullan M. Alzheimer’s beta-amyloid vasoactivity: identiﬁcation of a novel beta-amyloid conformational intermediate. FEBS Lett 1998; 436:445–448. Prior R, D DU, Frank R, Prikulis I, Pavlakovic G. Experimental deposition of Alzheimer amyloid beta-protein in canine leptomeningeal vessels. Neuroreport 1995; 6:1747–1751. Lorenzo A, Yankner BA. Beta-amyloid neurotoxicity requires ﬁbril formation and is inhibited by Congo Red. Proc Natl Acad Sci USA 1994; 91:12243–12247. Sadler II, Smith DW, Shearman MS, Ragan CI, Tailor VJ, Pollack SJ. Sulphated compounds attenuate beta-amyloid toxicity by inhibiting its association with cells. Neuroreport 1995; 7:49–53. Knopman DS. Current pharmacotherapies for Alzheimer’s disease. Geriatrics 1998; 53:S31–S34. Walker LC. Animal models of cerebral beta-amyloid angiopathy. Brain Res Brain Res Rev 1997; 25:70–84. Schenk D, Barbour R, Dunn W, Gordon G, Grajeda H, Guido T, Hu K, Huang J, Johnson-Wood K, Khan K, Kholodenko D, Lee M, Liao Z, Lieberberg I, Motter R, Mutter L, Soriano F, Schopp G, Vasquez N, Vandevert C, Walker S, Wogulis M, Yednock T, Games D, Seubert P. Immunization with amyloid-h attenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nat Lond 1999. Wagner SL, Munoz B. Modulation of amyloid ß-protein precursor processing as a means of retarding progression of Alzheimer’s disease. J Clin Invest 1999; 104:1329–1332. Davies SW, Turmaine M, Cozens BA, DiFiglia M, Sharp AH, Ross CA, Scherzinger E, Wanker EE, Mangiarini L, Bates GP. Formation of neuronal intranuclear inclusions underlies the neurological dysfunction in mice transgenic for the HD mutation. Cell 1997; 90:537–548. Klement IA, Skinner PJ, Kaytor MD, Yi H, Hersch SM, Clark HB, Zoghbi HY, Orr HT. Ataxin-1 nuclear localization and aggregation: role in polyglutamine-induced disease in SCA1 transgenic mice [see comments]. Cell 1998; 95:41–53. Klement IA, Zoghbi HY, Orr HT. Pathogenesis of polyglutamine-induced disease: A model for SCA1. Mol Genet Metab 1999; 66:172–178. Ikeda H, Yamaguchi M, Sugai S, Aze Y, Narumiya S, Kakizuka A. Expanded polyglutamine in the Machado–Joseph disease protein induces cell death in vitro and in vivo. Nat Genet 1996; 13:196–202. Lione LA, Carter RJ, Hunt MJ, Bates GP, Morton AJ, Dunnett SB. Selective discrimination learning impairments in mice expressing the human Huntington’s disease mutation. J Neurosci 1999; 19:10428– 10437. Masliah E, Rockenstein E, Veinbergs I, Mallory M, Hashimoto M, Takeda A, Sagara Y, Sisk A, Mucke L. Dopaminergic loss and inclusion body

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

154.

155.

156.

157. 158.

159.

160. 161.

162.

formation in a-synuclein mice: implications for neurodegenerative disorders. Science 2000; 287:1265–1269. Weissman C, Raeber AJ, Shmerling D, Aguzzi A, Manson JC. Knockouts, transgenics, and transplants in prion research. In: Prusiner SB, ed. Prion Biology and Diseases. Vol. 38. Cold Spring Harbor, NY: Cold Spring Harbor Press, 1999:273–305. Raeber AJ, Brandner S, Klein MA, Benninger Y, Musahl C, Frigg R, Roeckl C, Fischer MB, Weissmann C, Aguzzi A. Transgenic and knockout mice in research on prion diseases. Brain Pathol 1998; 8:715–733. Salomon AR, Marcinowski KJ, Friedland RP, Zagorski MG. Nicotine inhibits amyloid formation by the beta-peptide. Biochemistry 1996; 35:13568–13578. Wood SJ, MacKenzie L, Maleeﬀ B, Hurle MR, Wetzel R. Selective inhibition of Abeta ﬁbril formation. J Biol Chem 1996; 271:4086–4092. Pappolla M, Bozner P, Soto C, Shao H, Robakis NK, Zagorski M, Frangione B, Ghiso J. Inhibition of Alzheimer beta-ﬁbrillogenesis by melatonin. J Biol Chem 1998; 273:7185–7188. Merlini G, Ascari E, Amboldi N, Bellotti V, Arbustini E, Perfetti V, Ferrari M, Zorzoli I, Marinone MG, Garini P, et al. Interaction of the anthracycline 4V-iodo-4V-deoxydoxorubicin with amyloid ﬁbrils: inhibition of amyloidogenesis. Proc Natl Acad Sci USA 1995; 92:2959–2963. Howlett D, Cutler P, Heales S, Camilleri P. Hemin and related porphyrins inhibit beta-amyloid aggregation. FEBS Lett 1997; 417:249–251. Tomiyama T, Shoji A, Kataoka K, Suwa Y, Asano S, Kaneko H, Endo N. Inhibition of amyloid beta protein aggregation neurotoxicity by rifampicin. Its possible function as a hydroxyl radical scavenger. J Biol Chem 1996; 271:6839–6844. Kisilevsky R, Lemieux LJ, Fraser PE, Kong X, Hultin PG, Szarek WA. Arresting amyloidosis in vivo using small-molecule anionic sulphonates or sulphates: implications for Alzheimer’s disease. Nat Med 1995; 1:143– 148.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

10 Uncoating and Adsorption Inhibitors of Rhinovirus Replication Guy D. Diana ViroPharma, Inc., Exton, Pennsylvania, U.S.A.

Adi Treasurywala Pfizer Central Research, Groton, Connecticut, U.S.A.

I.

INTRODUCTION

Rhinoviruses are responsible for approximately 50% of infections resulting in the common cold [1]. These infections are caused by over 100 distinct serotypes, which vary by exhibiting minor or major changes in structure. The virus is divided into a major group, consisting of approximately 90%, and a minor group, diﬀering by the mode of attachment of the virus to the cell. The members of the major group of serotypes have been shown to bind to domain 1 and 2 of ICAM-1 [2 – 5], while the minor serotypes appear to have a binding preference for the human low density lipoprotein receptor (LDLR) [6]. An eﬀective antirhinovirus agent would be expected to be active against the majority of serotypes, since at any time one may become infected by any of the 100+ serotypes. The compounds in the series shown in Figure 1 have demonstrated broad spectrum antirhinovirus activity against both the minor and major group of serotypes [7 – 12]. These compounds have been shown to inhibit uncoating of the major group [13] and to block adsorption of the minor serotypes to the cell [14]. This chapter describes our eﬀorts to determine the mode of binding

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 1

General structure of isoxazole series of antipicornavirus agents.

of these compounds and to enhance the activity by utilizing x-ray crystallography coupled with traditional structure –activity methodology.

II. CAPSID BINDING COMPOUNDS In 1985 Dr. Michael Rossmann and his colleagues determined for the ﬁrst time the three-dimensional structure of a human rhinovirus [15]. Their studies, performed with human rhinovirus type 14 (HRV-14), revealed the structure as an eicosahedron consisting of four proteins designated VP1, VP2, VP3, and VP4 forming a protomeric unit, combined to form a ﬁvefold axis of symmetry (Fig. 2). The surface of the capsid

Figure 2 The three-dimensional structure of HRV-114 consisting of four viral proteins, VP1, VP2, and VP3; Vp4 is pointing toward the center of the capsid protein and is not visible.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 3 The orientation of disoxaril in the binding site of HRV-14.

protein contains a canyon that was shown to be the cell receptor binding site [2]. Subsequently, the structure of several additional rhinovirus serotypes was determined [16 –19]. Although these rhinoviruses share the same general structure described for HRV-14, the latter appears to be distinctly diﬀerent from other rhinoviruses, particularly with respect to the sequence similarity. Following the elucidation of the structure of HRV-14, x-ray studies were performed on two members of the series of compounds shown in Figure 1, disoxaril and WIN52084 [20]. The purpose of this study was to elucidate the nature of the binding of these compounds to the capsid protein. Disoxaril was shown to bind in a hydrophobic pocket below the a depression referred to as the ‘‘canyon,’’ with the oxazoline ring in the ‘‘toe’’ region of the binding pocket. The isoxazole ring resides in the ‘‘heel’’ below the area designated as the pore (Fig. 3). The nitrogen of the isoxazole

Figure 4 Structure of WIN52084.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

ring was within 3.5 A˚ of asparagine 219, suggesting the possibility of hydrogen bonding. The phenyl ring was in a stacking conformation with tyrosine 128 and tyrosine 152. WIN52084, however, was bound in the opposite orientation, with the isoxazole ring in the ‘‘toe.’’ Subsequently, it was determined that only two additional compounds that were examined were bound in the same orientation as WIN52084 (Fig. 4) [21]. This observation led to the following conclusions: Analogues with a seven-carbon chain connecting the phenyl and isoxazole rings, and with a substituent on the oxazoline ring, were bound with the isoxazole ring in the toe of the hydrophobic pocket. All other analogues, regardless of the length of the connecting chain, were bound in the opposite orientation.

A. The Nature of the Binding Site The inﬂuence of substituents connected to the oxazoline ring on the binding orientation of these molecules was intriguing. Since the carbon to

Figure 5 Homologues of WIN52084 illustrating an entaniomeric effect. The asymmetric center on the oxazoline ring is designated by asterisk.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

which the alkyl substituents are attached is asymmetric, both enantiomers of WIN52084 as well as a homologous series of compounds were evaluated against HRV-14 [22] (Fig. 5). In each case, the S isomer was considerably more inhibitory than the R, which suggested an enantiomeric eﬀect. Examination of WIN52084 in the pocket clearly showed that the S-methyl group was in close proximity to a hydrophobic pocket formed by Leu106 and Ser107 (Fig. 6). To further analyze the interactions of the methyl group of the two comformers in the binding site, an energy proﬁling study was performed. With the x-ray crystal structure of the S isomer of WIN52084 in the virus pocket serving as a starting point, a window consisting of all residues within 8 A˚ of any atom was excised from the starting structure. After charges had been set on the atoms of the resulting pocket and drug according to a method in Chem-X [23], and after the hydrogen atoms had been removed, the intermolecular van der Waals energy was calculated via a 6 –12 function for conformations resulting from the rotation of the oxazoline ring about the bond connected to the phenyl ring, in increments of 10j. A plot of this function versus the rotation angle

Figure 6 WIN52084 bound to HRV-14. The methyl group on the oxazoline ring is pointing toward a hydrophobic pocket formed by Leu106 and Ser107.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

showed mainly two peaks at – 90j and 100j. A repeat of this calculation with the R conformer resulted in a ﬂat valley between 30j and approximately 120j. Similar results were obtained with the R- and S-ethyl compounds, which showed an even more dramatic pattern that was signiﬁcant because the ethyl homologue was more potent (Fig. 7). These results suggested that the twist angle about the two rings could be an important factor in determining biological activity. It is possible that the conformation with the appropriate twist angle may be imposed by the nature of the binding pocket and that maximum interaction with the hydrophobic pocket formed by Leu and Ser may also be of importance [20 – 22].

B. Aliphatic Bridge The x-ray studies on several analogues in this series of compounds showed that the chain connecting the isoxazole and phenyl rings adopts a bowed

Figure 7 Plot of energy vs torsion angle from an energy profiling study resulting from rotating the oxazoline ring of the S isomer of WIN52084 about the phenyl ring.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

conformation when bound to HRV-14. It had been assumed that ﬂexibility of the chain was critical for binding and biological activity. Dynamic studies by Dr. Andrew McCammon with WIN52084 in HRV-14 revealed considerable motion of the aliphatic chain during an observation lasting for 10 ps (Fig. 8) (Dr. Andrew McCammon, University of Houston, personal communication). This result posed several questions regarding the importance of ﬂexibility vs rigidity of the chain. Would a conformationally rigid chain oﬀer enhanced hydrophobic interactions and consequently improved binding, or are there other factors in the binding process that would require a ﬂexible chain? To address these issues, several compounds with rigidity incorporated into the chain were synthesized; their activity against HRV-14 and HRV-1A examined (Fig. 9) and the compounds modeled in the respective binding site [24]. WIN54954, which

Figure 8 Molecular dynamics of WIN52084 in HRV-14 during a 10 ps run, illustrating the movement of the chain. (Courtesy of Andrew McCammon, University of Houston.)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 9 Table comparing the activity of the E and Z olefin and butyne analogues of WIN54954.

has been clinically evaluated, was used as a comparator. The Z oleﬁn demonstrated a two- to threefold reduction in activity in comparison to WIN54954, while the E isomer showed a threefold enhancement in activity. The potency of the butyne analogue was more than fourfold greater than that of WIN54954 against HRV-14 and was comparable to that of the E isomer.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

C. Modeling of Conformationally Restricted Analogues The structures shown in Figure 9 were constructed using WIN54954 as a template, since its x-ray conformation in HRV-14 had been determined. The resulting structures were subjected to the Tripos force ﬁeld (Maximin 2), using Sybyl version 5.41, with default settings. Rotatable bonds in the alkyl ether chain were deﬁned, and the structures were ﬂexibly ﬁtted to WIN54954, in virus-bound conformation, for insertion into the HRV-14 binding site (Fig. 10). The optimized ﬁtted structures were inserted into each serotype by replacement of virus-bound WIN54954. Since the drugbound conformation of the virus binding site with several of the compounds had been determined, revealing only minor variations in compound structure, insertion of the modeled compounds into the binding site conﬁguration, derived from WIN54954, appeared reasonable. Two interesting observations emerged from this study. The acetylene analogue, which was more than fourfold more potent than WIN54954

Figure 10 Overlay of energy-minimized structures of the E and Z isomers and WIN54954.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

against HRV-14, was inactive against HRV-1A, and there was a dramatic diﬀerence between the activities of the E and Z oleﬁns against HRV-14. Because x-ray studies had shown that the HRV-14 binding site was longer than the 1A site, the results of this study supported the premise that the activity is dependent on the length of the molecule. The butyne was modeled in HRV-14, causing no serious steric interactions. However this was not the case in HRV-1A, where the chlorine atom appeared to interact with Ile125. The diﬀerence in activity of the E and Z oleﬁns against HRV-14 was explained by examining the relatively low energy virus-bound conformations. The result of an overlay of WIN-54954 (based on x-ray crystallography data), minimize E- and Z-oleﬁnic structures and the butyne analogue, suggested that the E isomer showed a reasonable ﬁt while the Z isomer did not. Furthermore, when the Z isomer was inserted into the HRV-14 pocket, unfavorable interactions occurred. The very high minimal inhibitory concentration (MIC) values for the Z isomer against HRV-14 and HRV-1A may reﬂect a slow kon in both cases. The conformational space accessible to the isoxazole of the E and Z oleﬁns, the butyne, and the three-carbon chained homologue of WIN 54954 by conformational sweep graph and for the Z oleﬁn disclosed a signiﬁcant inaccessible region of space, while the butyne, E oleﬁn, and alkane do not show this deﬁcit. Consequently, binding to this site may be dependent on conformational permissibility in this region that is required for entry into the pocket. These results suggested that the activity of these compounds against the two serotypes is strongly dependent on the ﬂexibility of the hydrocarbon chain and the ability of the molecule to ﬁt into the conformational space of both pockets.

III. PHENYL STACKING Thus far all the compounds that were examined bound to HRV-14, with the exceptions noted, are oriented with the phenyl ring in a stacking mode with Tyr128 and Tyr152. Aromatic –aromatic interactions have been shown to be quite common in protein – protein interactions [25 – 31], and in many cases have displayed [32,33] an electrostatic component. Furthermore, such interactions would be expected to contribute extensively to the binding energy [34]. To determine the nature of the aromatic stacking interactions, an energy proﬁling study was performed by twist-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

ing the phenyl ring about the carbon oxygen bond and examining a number of parameters such as heat of formation and electronic energy. Figure 11 shows the torsion angle between the oxazoline and phenyl rings for each compound after energy minimization and ﬂexible ﬁtting. These studies were performed with a variety of substituents in the position ortho to the ether. It was anticipated that if electrostatics were involved in the stacking of these rings, the correlation of the results of the proﬁling study with antiviral activity should relate to the physical and electronic properties of the substituents. There was no correlation between energy maxima or minima and size or electronic nature of the substituent, however, nor do the results correlate with biological activity. We concluded that electrostatics play no part in the stacking; rather, the interactions appear to be hydrophobic. In addition, these results suggest that a planar orientation of the phenyl rings is preferred. The lack of an electrostatic eﬀect associated with the phenyl –phenyl interactions may be due to the inability of the phenyl ring of these compounds to adapt a true end-surface orientation as a result of space constraints within the pocket.

Figure 11 Torsion angle between the oxazoline and phenyl rings obtained from minimized structures fitted to the x-ray structure of WIN54954.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

IV. HYDROGEN BONDING The initial x-ray study with disoxaril and WIN52084 in the binding site revealed that asparagine 219 was within hydrogen-bonding distance of the nitrogen of the isoxazole or oxazoline rings. Several pieces of information suggested that this type of binding contributes negligibly if at all to the total binding energy. Lau and Pettitt [35] examined whether the close approach of the asparagine and isoxazole ring, which had been observed crystallographically, was indeed an attractive event. By selectively computing the pairwise attraction of the hydrogen of the asparagine 219 and the nitrogen of the isoxazole ring, which could conceivably be involved in the hydrogen bond, and disregarding the contribution of this energy to the overall energy of the system, the researchers were able to predict that the potential hydrogen bond was inconsequential. In addition to the computational studies that argued against the existence of a hydrogen bond with Asn219, further evidence was obtained by site-directed mutagenisis of the asparagine in question to an alanine (Dr. Daniel C. Peaver, Sterling Winthrop Inc., personal communication). Conﬁrmation of the mutation was accomplished by sequencing. A comparison of the sensitivity of the mutant with the wild type showed that no change in sensitivity had resulted from the removal of the hydrogen donor potential. Consequently, these ﬁndings were in complete agreement with the results reported by Lau and Pettitt. Although the evidence presented strongly suggests the lack of contribution of Asn219 to the binding energy, examination of the x-ray result of HRV-14-bound compounds revealed the presence of a water molecule in the vicinity of the isoxazole ring and hydrogen-bonded to the backbone of Leu106, Ser107, and Asn219 (Fig. 12) [36]. A similar hydrogen-bonding network has been seen in HRV-50 (Dr. Vincent Giranda, Sterling Winthrop Inc., personal communication). This observation could shed some light on the relative activity of other heterocyclic replacements for the isoxazole ring.

A. HRV-14 Model Development The extensive data generated from x-ray studies with HRV-14 permitted the development of a model that could deﬁne the properties required of this class of compounds for antiviral activity [37]. This model was dependent on the orientation and x-ray conformational data for compounds bound to the viral pocket. Some assumptions were made based on earlier results and on rules generated for predicting compound orientation. For example, it

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 12 Hydrogen bonding network involved in the binding of WIN54954 analogues to HRV-14.

was assumed that all the compounds included in the study that had not been examined by x-ray crystallography behaved in a predictable manner. Compounds were divided into two groups of seven compounds each. One group whose conformations were known (Fig. 13) demonstrated various levels of activity against the virus. The second group consisted of inactive compounds with related structures (Fig. 14). In the absence of conformational data for this group, one of the active compounds was used as a template for these compounds. A SYBYL (version 5.0) database was created. All the structures were overlaid in the position found in the binding site (Fig. 15). Volume maps were then calculated for the Boolean ‘‘union’’ of all active and inactive compounds, which were then overlaid, and the excess volume occupied by the inactive compounds, in comparison to the active compound (Boolean minus), was calculated (Fig. 16). A similar procedure was followed for the excess volume for the actives (Boolean plus). These combined results revealed that inactive compounds displayed excessive bulk around the phenyl ring. Although some bulk is desirable in this area to enhance hydrophobic interactions, excessive bulk, which leads to

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 13 Compounds active against HRV-14, which were used in the development of the volume map model.

steric interactions, leads also to inactivity [38]. Conversely, space occupancy in the pore area of the binding site was found to contribute to good biological activity. To reﬁne this model qualitatively, the binding of several of these compounds was subjected to a CoMFA (Comparative Molecular Field Analysis) [39]. This program examines electrostatic and steric parameters

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 14 Compounds inactive against HRV-14.

and through a partial-least-squares analysis determines the correlation of these eﬀects with biological activity. Either rigid structures or ﬁxed conformations are required to carry out the analysis. Eight compounds, which had been used in the volume map study and whose binding conformations were known, were employed. These compounds also had a reasonable spread of activity against HRV-14. In addition, they oﬀered some degree of structural diversity. Van der Waals radii for atoms were taken from a standard Tripos force ﬁeld. Charges were calculated by the AMI method by single-point calculations on the receptor-bound conformation of the drug molecule. Point charges on the hydrogen atoms were not collapsed onto the atom to which they were bound but were left on the hydrogen atoms. Log p values were calculated using the MedChem software package (version 3.54). All these parameters, in addition to the CoMFA ﬁeld values designated by *, were used in the quantitative structure – activity (QSAR) analysis (Fig. 17).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 15

Overlay of active and inactive compounds.

Structural alignments were obtained from x-ray crystallographic analysis. The backbone residues within 20 A˚ of any atom of the compound were included in this study. This eﬀectively created a cube 20 A˚ on a side, which was divided into grid points 1 A˚ apart. A hydrogen atom and a proton as a probe were used to sample each grid point for both electrostatic and steric eﬀects. The data were tabulated and cross-validated along with the physical parameters by means of a partial-least-squares method, with the following results: good correlation of MIC with CoMFA data and good predicative capabilities in the case of steric properties (Fig. 18). No meaningful correlation was seen with electrostatic parameters, either taken in combination with steric factors or evaluated alone. A regression analysis using all the values shown in Figure 17 revealed no contribution of any parameters, other than the CoMFA ﬁeld, to the activity of these compounds. In addition to the QSAR data, which resulted from this program, three-dimensional contour maps were generated for both steric and elec-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 16 Overlay of volume maps of active compounds and inactive compounds.

trostatic ﬁelds. Cutoﬀs were used to contour together points where the correlations were highest and positive and those that were highest and negative. Although the shapes of the maps coincide with the shape of the pocket, the structure of the macromolecules was not part of the calculations. The visual results displayed by the contour maps qualitatively agree with the QSAR results; that is, there is no signiﬁcant correlation between electrostatics and biological activity (Fig. 19), despite a strong correlation between the steric ﬁelds and activity, as predicted. Although a moderate positive eﬀect was seen in the vicinity of the aromatic ring, in general, this model predicts that excessive bulk in this area negatively correlates with biological activity. These results are in agreement with the conclusions empirically generated from the volume map study and also conﬁrm the lack of electrostatics involved in the phenyl–phenyl stacking interactions, which had been observed earlier.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 17

CoMFA coordinates.

B. Model Development Based on Small-Molecule, Low-Energy Conformations Thus far, the model development discussed has been one based on x-ray conformations. Considering that there are over 100 serotypes, the discovery of broad spectrum antirhinovirus agents would require considerably more three-dimensional virus structures. We have investigated the possibility of simply using energy-minimized small-molecule conformations exclusive of the virus structures [40]. The compounds in question were constructed in SYBYL and minimized by means of Maximin. By using a template for spatial referencing (Fig. 20) that represented one of the more potent compounds against HRV-14, it was possible to employ the program Superimpose (SYBYL version 5.0) to algorithmically overlay the molecules based not on conformational similarity but rather on shape. Volume maps were then constructed as already described. Figure 21 compares the volume maps created by this method with those from x-ray structures in HRV-14. The diﬀerence maps clearly show that within certain limits, increasing chain length increases activity. Although there is a space-ﬁlling requirement for activity, exceeding the appropriate distribution or extent of bulk results in inactive compounds. These ﬁndings essentially duplicate those obtained from the preceding method.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 18 Regression equation obtained from partial-least-squares data crossvalidated in Figure 17.

This procedure was repeated for HRV-1A. Duplicate maps were generated by means of both procedures and clearly show that shorter molecules, as measured from the phenoxy to isoxazole moieties, are more active. Molecules with the correct degree and placement of bulk in the middle of the volume are also more active. These encouraging results suggest that this method can be applied to other serotypes without giving consideration to their three-dimensional structures.

C. Application of Model Development to Drug Design The results of the model development for HRV-14 and HRV-1A demonstrated that the problem of drug design is complicated by the diﬀerence in the dimensions of the binding sites, at least in the case of these two serotypes. One solution to this problem is to prepare a compound that

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 19 Three-dimensional contour map generated from CoMFA analysis of electrostatic fields showing correlation with antiviral activity. Red indicates a strong correlation. The grid map is a result of an analysis using a probe atom (charge 0). All charges on the molecules were calculated using the AM1 Hamiltonian without geometry optimization.

has some degree of ﬂexibility and would be accommodated by the binding sites in both serotypes. We chose to examine the homologous series shown in Figure 22. The three-carbon bridge structure demonstrated good activity against HRV-1A but poor activity against HRV-14. As the ﬂexible side chain is increased, a concomitant improvement in activity is seen against HRV-14, with optimum activity observed against both serotypes with the three-carbon side chain. This avenue was pursued because further testing of the three-carbon chained analogues against 100 rhinovirus serotypes indicated that these compounds exhibited a broader spectrum of activity, suggesting that perhaps HRV-14 is not representative of the majority of serotypes.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 20 Minimized structure serving as a template for spatial referencing in the model development using energy-minimized small molecules.

D. The Development of a Clinical Candidate Biological activity is not the only criterion required for drug development, as anyone who has been involved in this area is aware. Potency, toxicity, bioavailability, metabolic stability, and plasma half-life are only a few of the critical issues that must be addressed. Although satisfactory potency and spectrum activity had been achieved with WIN54954, which has been clinically evaluated, this compound lacked metabolic stability and consequently displayed a short half-life. It became clear that the oxazoline ring was metabolically unstable and was responsible for the generation of crystalurea with disoxaril and for a drug-induced rash with WIN54954, accompanied by a short plasma halflife. Consequently, a replacement for the oxazoline ring was sought, which would be metabolically stable and would demonstrate satisfactory bioavailability. After examining several heterocyclic replacements, the 4-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 21 Comparison of volume map for HRV-14 generated from x-ray data (left) and small-molecule energy-minimized structures (right).

methyltetrazole analogue with a three-carbon linker (Fig. 23) appeared to provide good chemical stability and improved biological activity in comparison to WIN54954 [41]. However, when this compound was administerd to dogs, hepatotoxicity was observed which was attributed to metabolic instability. Further modiﬁcations resulted in the synthesis of the 5-methyl 1,2,4-oxadiazole analogue, which was selected as a possible

Figure 22

Homologous series of compounds.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 23 Structure of the 4-methyltetrazole analogues.

development candidate based on potency and spectrum of activity [42]. To address metabolic stability, however, a monkey liver microsomal assay was established by means of which the half-life, the extent of metabolism, and the nature of the metabolic products could be determined [43]. Initially, WIN54954 was incubated at 37jC with a liver microsomal mixture for 30 min and the incubate was extracted with hexane. The extracts were analyzed by high performance liquid chromatography (HPLC), which revealed 18 metabolic products. When the oxadiazole analogue was subjected to the same conditions, two major peaks, metabolites A and B, were observed by HPLC (Fig. 24), in addition to six minor ones. The rate of metabolism was similar to that of WIN54954, however, with a half-life of 27 vs 20 min.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 24 HPLC spectrum resulting from the incubation of the oxadiazole with monkey liver microsomes.

The question at this point was whether modiﬁcations could be made to the oxadiazole molecule to enhance metabolic stability and achieve comparable activity. This approach required knowledge of the site of metabolism and the nature of the metabolic products. This information was obtained from ion mass spectrometry. The identity of these products was determined by comparing the fragmentation pattern of metabolites A and B with the parent compound and the corresponding daughter ions (Fig. 25). Analysis of the metabolic products indicated that hydroxylation occurred to a greater extent (30%) on the methyl group attached to the isoxazole ring than to the methyl group on the oxadiazole ring (10%). The methyl group in this postion was replaced with a triﬂuormethyl group to prevent hydroxylation. The result of the incubation of this compound indicated that although this position was protected, three metabolic products were produced; in addition, the half-life was not substantially diﬀerent from the parent compound. A similar replacement on the oxadiazole ring (Fig. 26) not only prevented metabolism at this position but also protected the entire mole-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 25 Biotransformation of WIN 61893 and WIN 64172.

Figure 26 Metabolism of the trifluoromethylisoxazole analogue.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 27

Structure of second properties pleconaril.

cule, resulting in two minor metabolites, and substantially increased the half-life. In addition, pleconaril has exhibited a broad spectrum of antipicornavirus activity and has shown good bioavailability (Fig. 27) and is undergoing clinical trials for upper respiratory rhinovirus infections.

V. CONCLUSIONS x-Ray crystallography has added a new dimension to antirhinovirus drug design. It has enabled us to examine the molecular interactions within the compound binding site and to better understand the mechanism of binding. We have been able to devise a model based on x-ray crystallography that qualitatively describes properties of molecules that are beneﬁcial for antirhinovirus activity. Also, by comparison to a volume map based on x-ray conformations, we have developed a comparable model based on small-molecule energy-minimized structures exclusive of x-ray data. Finally, we have been able to apply our results to the synthesis of compounds active against both HRV-1A and HRV-14. Aside from the design aspects, we have dealt with the more practical considerations such as metabolic stability and bioavailability, which have led to a clinical candidate. Now certain unanswered mechanistic questions can be addressed. How does the drug enter the binding site? Is there a recognition site, which may explain some anomalous results that remain a mystery? Hopefully,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

future work in this regard will eventually lead to an understanding of the binding process and its relationship to biological activity.

REFERENCES 1. Gwaltney JM Jr. Principles and Practices of Infectious Diseases, 2nd ed. New York: John Wiley & Sons, 1985. 2. Abraham G, Colonno RJ. Many rhinovirus serotypes share the same cellular receptor. J Virol 1984, 51:814 – 817. 3. Greve JM, Davis G, Meyer AM, et al. The major human rhinovirus receptor is ICAM-1. Cell 1989; 56:839 – 847. 4. Staunton DE, Merluzzi VJ, Rothlein R, Barton R, Marlin SD, Springer TA. A cell adhesion molecule ICAM-1 is the major surface receptor for rhinoviruses. Cell 1989; 56:849 – 853. 5. Uncapher GR, Dewitt CM, Colonno RJ. Human rhinovirus serotype families contain all but one human rhinovirus serotype. Virology 1991; 180:814 – 817. 6. Hofer F. Members of the low density lipoprotein receptor family mediate cell entry of a minor group common cold virus. Proc Natl Acad Sci USA 1994; 91:1839 – 1842. 7. Otto MJ, Fox MP, Fancher MJ, Kuhrt MF, Diana GD, McKinlay MA. In vitro activity of WIN 51711, a new broad spectrum antipicornavirus agent. Antimicrob Agents Chemother 1985; 27:883 – 886. 8. Diana GD, McKinlay MA, Otto MJ, Akullian V, Oglesby CJ. [[(4,5-Dihydro-2-oxazolyl)phenoxy]alkyl]isoxazoles. Inhibitors of picornavirus uncoating. Med Chem 1985; 28:1906 – 1912. 9. Diana GD, McKinlay MA, Brisson CJ, Zalay ES, Miralles JV, Salvador UJ. Isoxazoles with antipicornavirus activity. J Med Chem 1985; 28:748 – 752. 10. Diana GD, Otto MJ, McKinlay MA. Inhibitors of viral uncoating. Pharmacol Ther 1985; 24:287 – 297. 11. Diana GD, Oglesby RC, Akullian V, et al. Structure-activity studies of 5-[[4(4,4-dihydro-2-oxazolyl)phenoxy]alkyl]-3-methylisoxazoles: inhibitors of picornavirus uncoating. J Med Chem 1987; 30:383 – 388. 12. Fox MP, Otto MJ, Shave WJ, McKinlay MA. Prevention of rhinovirus and poliovirus uncoating by WIN 51711, a new antiviral drug. Antimicrob Agents Chemother 1986; 30:110 – 116. 13. Fox MP, McKinlay MA, Diana GD, Dutko FJ. Binding aﬃnities of structurally related human rhinovirus capsid binding compounds are related to their activities against human rhinovirus type 14. Antimicrob Agents Chemother 1991; 30:110 – 116. 14. Pevear DC, Fancher MJ, Felock, et al. Conformational changes in the ﬂoor of the human rhinovirus canyon blocks adsorption to Hela cell receptors. J Virol 1989; 63:2002 – 2007.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

15. Rossmann MG, Arnold E, Erickson JW, Frankenberger EA, et al. Structure of a human common cold virus and functional relationship to other picornaviruses. Nature 1985; 317:145 – 155. 16. Badger, Minor I, Kremer MJ, et al. Structural analysis of a series of antiviral agents complexed with human rhinovirus 14. Proc Nat Acad Sci USA 1988; 85:3304 – 3308. 17. Oliviera MA, Zhao R, Lee WM, et al. The structure of human rhinovirus 16. Structure 1993; 1:51 – 68. 18. Kim S, Smith TJ, Chapman MS. Crystal structure of human rhinovirus 1A (HRV 1A). J Mol Biol 1989; 210:91 – 111. 19. Giranda VL. Structure-based drug design of antirhinoviral compounds. Structure 1994; 2:695 – 698. 20. Smith TJ, Kremer MJ, Luo M, Vriend G, et al. The site of attachment in human rhinovirus 14 for the antiviral agents that inhibit uncoating. Science 1986; 233:1286 – 1293. 21. Badger J, Minor I, Kremer MJ, et al. Structural analysis of a series of antiviral agents complexed with human rhinovirus 14. Proc Natl Acad Sci USA 1988; 85:3304 – 3308 22. Diana GD, Otto MJ, Treasurywala AM, et al. Enantiomeric eﬀects of homologues of disoxaril on the inhibitory activity against human rhinivirus-14. J Med Chem 1988; 31:540 – 544. 23. Chem-X, developed and distributed by Chemical Design Limited, Oxford, England. Geistiger J. Marsili M. Iterative partial equalization of orbital electronegativity – a rapid access to atomic charges. Tetrahedron 1980; 36:3219 – 3228. 24. Mallamo JP, Diana GD, Pevear et al. Conformationally restricted analogues of disoxaril. J Med Chem 1992; 35:4690 – 4695. 25. Burley SK, Petsko GA. Aromatic-aromatic interactions: a mechanism of protein structure stabilization. Science 1985; 229:23 – 28. 26. Burley SK, Petsko GA. Weakly polar interactions in proteins. Protein Chem 1988; 39:125 – 192. 27. Levitt M, Perutz MF. Aromatic rings act as hydrogen bond receptors, J Mol Biol 1988; 201:751 – 754. 28. Singh J, Thornton JM. The interaction between phenylalanine in proteins. FEBS Lett 1985; 191:1 – 6. 29. Anderson DE, Hurley JH, Nicholson H, Baase WA, Mathews BW Hydrophobic core repacking and aromatic-aromatic interaction in the thermostable mutant of T4 lysozyme Ser 117-Phe. Protein Sci 1993; 2:1285 – 1290. 30. Karlstrom G, Linse P, Wallqvist A, Johnson B. Intermolecular potential for the H2O-C6H6 and the C6H6-C6H6 system calculated in an ab initio SCF CI approximation. J Am Chem Soc 1983; 105:3777 – 3782. 31. Fowler PW, Moore GJ. Calculation of the magnitude and orientation of electrostatic interactions between small aromatic rings in peptides and pro-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

32.

33. 34. 35.

36. 37.

38.

39.

40.

41.

42.

43.

teins: implications for angiotensin II. Biochem Biophys Res Commun 1988; 153:1296 – 1300. Serrano L, Bycroft M, Fersht AR. Aromatic-aromatic interactions and protein stability. Investigation of double mutant cycles. J Mol Biol 1990; 218:465 – 475. Pawliszyn J, Szczeskiak MM, Scheiner S. Interactions between aromatic systems-dimers of benzene and S-triazine. J Phys Chem 1984; 88:1726 – 1730. Weiner SJ, Kollman PA, Nguyen DT, Case DA. An all atom force ﬁeld for simulations of proteins and nucleic acids. J Comput Chem 1986; 7:230 – 252. Lau WF, Pettitt MB. Selective elimination of interactions: a method for assessing thermodynamic contributions to ligand binding with application to rhinovirus antivirals. J Med Chem 1989; 32:2542 – 2547. Giranda VL, Russo G, Felock P, Draper T, Diana G, Guiles J, Oglesby R, Long M, Pevear DC. Submitted. Diana GD, Treasurywala AM, Bailey TR, Oglesby RC, Pevear DC, Dutko FJ. A model for compounds active against human rhinovirus-14 based on X-ray crystallography data. J Med Chem 1989; 33:1306 – 1311. Diana GD, Cutcliﬀe D, Oglesby RC, Otto MJ, Mallamo JP, Akullian V, McKinlay MA. Synthesis and structure-activity studies of some disubstituted phenylisoxazoles against human picornavirus. J Med Chem 1989; 32:450 – 455. Diana GD, Kowalczyk P, Treasurywala AM, Oglesby RC, Pevear DC, Dutko FJ. CoMFA analysis of interactions of antipicornavirus compounds in the binding pocket of human rhinovirus-14. J Med Chem 1992; 35:1002 – 1008. Aldous DJ, Diana GD, Treasurywala AM, Jaeger EP, Pevear DC. Progress in the development of an antirhinovirus model based on small molecule conformations, Poster presentation 28 presented at the American Chemical Society Meeting, San Diego, CA, 1994. Woods MG, Diana GD, Rogge MC, Otto MJ, Dutko FJ, McKinlay MA. In vitro activity and in vivo activities of WIN 54954, a new broad spectrum antipicornavirus drug. Antimicrob Agents Chemother 1989; 33:2069 – 2074. Diana GD, Volkots D, Nitz TJ, et al. Oxadiazoles as ester replacements in compounds related to disoxaril: antirhinovirus activity. J Med Chem 1994; 37:2421 – 2436. Diana GD, Rudewicz P, Pevear DC, et al. Picornavirus inhibitors: triﬂuoromethyl substitution provides a global protective eﬀect against hepatic metabolism. J Med Chem 1995; 38:1355 – 1371.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

11 Profiles of Prototype Antiviral Agents Interfering with the Initial Stages of HIV Infection Erik De Clercq Rega Institute for Medical Research, Katholieke Universiteit Leuven, Leuven, Belgium

I.

INTRODUCTION

The initial stages of the human immunodeﬁciency virus (HIV) infection could be deﬁned as the steps of the viral growth cycle that precede the integration of the proviral DNA into the host cell genome. These stages occur during the acute phase of the HIV infection, that is, when the virus has invaded new cells. Once the proviral DNA has been integrated into the host genome, the host cell and all its progeny cells can be considered to be persistently or chronically infected. Expression of the integrated viral genome will follow the classical ﬂow of gene expression: that is, transcription, translation, and post-translational modiﬁcations under the concerted regulatory action of both cellular and viral factors. This chapter reviews prototypes of single chemical entities that interfere with the initial stages of the acute HIV infection, at steps that are predominantly, if not solely, determined by speciﬁc viral proteins. Thus, the compounds interacting with these steps in the HIV replicative cycle may be expected to display a reasonably high speciﬁcity in their mode of action. The targets that could be envisaged for such chemotherapeutic attack are the following: (1) virus adsorption, involving the

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

viral envelope glycoprotein gp120, (2) virus – cell fusion, involving both viral glycoproteins gp120 and gp41, (3) viral uncoating, involving the viral capsid proteins, (4) the substrate (dNTP) binding site of the viral reverse transcriptase, and (5) an allosteric non – substrate binding site of the HIV-1 reverse transcriptase. HIV inhibitors interacting with these targets have been the subject of some earlier reviews [1 –4]. The chemokine receptors CXCR4 and CCR5 used as coreceptors by X4 and R5 HIV-1 strains are not discussed here; for reviews on inhibition of HIV infection by these receptor antagonists, see Refs. 5 and 6. The compounds are highlighted from the following viewpoints: antiHIV potency and selectivity, mechanism of action, antiviral activity spectrum, clinical or therapeutic potential, and risk of resistance development.

II. VIRUS ADSORPTION INHIBITORS: POLYANIONIC SUBSTANCES Various polyanionic substances (viz., polysulfates, polysulfonates, polycarboxylates, and polyoxometalates) have been reported to block HIV replication; for a review on the polysulfates, see Ref. 7. These substances inhibit HIV-induced cytopathicity at a concentration of 0.1 to 1 Ag/mL, while not being toxic to the host cells at concentrations up to 2 or 5 mg/mL, thus achieving selectivity indexes of approximately 10,000 [7]. The target of interaction for the polysulfates would be the V3 loop of the viral gp120 glycoprotein [8 –10]. This loop contains a highly basic region with which the polyanionic substances could interact electrostatically. Thus, polyanions such as dextran sulfate may be assumed to block virus adsorption by shielding the viral envelope glycoproteins [8]. Alternatively or additionally, polyanionic substances may also interact with the cellular CD4 receptor [11], thus preventing the viral envelope gp120 from anchoring to the outer cell membrane. Depending on their molecular weight, the nature of their anionic groups, and the density/distribution of their negative charges, the polyanionic substances exhibit an activity spectrum that extends to several enveloped viruses other than HIV: among the retroviruses, SIV (simian immunodeﬁciency virus); among the herpesviruses, HSV (herpes simplex virus) and CMV (cytomegalovirus); among the orthomyxoviruses, inﬂuenza A; among the paramyxoviruses, RSV (respiratory syncytial virus); and toga-, ﬂavi-, arena-, bunya-, and rhabdoviruses. Among the diﬀerent HIV strains, rather striking diﬀerences have been noted with regard to

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

susceptibility to polyanionic substances (e.g., dextran sulfate) [12], and this diﬀerential susceptibility may be related to diﬀerences in the composition of the viral glycoprotein portions with which the compounds interact. Because of their broad activity spectrum, encompassing various enveloped viruses, polyanionic substances may be of practical utility in the prophylaxis and/or therapy of a number of important virus (e.g., HIV, HSV, CMV, RSV, inﬂuenza A) infections. Yet, there is little, if any, evidence for the in vivo eﬃcacy of these compounds following either parenteral or topical administration. Polyanions, and dextran sulfate in particular, are poorly absorbed upon oral administration [13], and, in addition, sulfated polysaccharides are notorious for their anticoagulant activity. However, these problems can be overcome by the appropriate chemical modiﬁcations (Fig. 1). Thus, h-cyclodextrin sulfate becomes orally bioavailable following substitution of benzyl groups at either C-2 or C-6 of the sugar residues mCDS71 [14] and mCDS11 [15], respectively, and heparin loses anticoagulant activity when acylated at the C-3 position of the sugar rings [16]. These favorable features (oral bioavailability, loss of anticoagulant activity) were obtained without impairing the anti-HIV activity of the products (mCDS71, mCDS11, or O-acylated heparin). The polyanionic substances may be expected to yield their greatest promise when put in contact with the virus under the conditions that

Figure 1 Modified sulfated polysaccharides: (A) O-acylated heparin (m = 2, 4, 6, . . .) and (B) mCDS71 (a modified h-cyclodextrin sulfate).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

mimic as closely as possible the in vitro situation, where these compounds have proved so clearly eﬀective. Further studies should address such issues as delivery forms, route(s) of administration, and time of treatment (with respect to the virus infection), before an appropriate candidate compound is submitted to clinical trials. Polyanionic substances are not known to lead to resistance development, although, as mentioned, diﬀerent HIV strains may diﬀer markedly ab initio in their susceptibility to this class of compounds.

III. VIRUS–CELL FUSION INHIBITORS: LECTINS, ALBUMINS, AND TRITERPENE DERIVATIVES Because of their interference with the interaction between the viral envelope gp120 glycoprotein and the cellular CD4 receptor, polyanionic substances not only inhibit virus adsorption to the cells but also block syncytium (giant cell) formation between the HIV-infected (gp120+) cells and uninfected (CD4+) cells. Since syncytium formation results in a selective destruction of the CD4+ cells, this syncytium formation may play an important role in the pathogenesis of AIDS (a hallmark of which is a progressive decline of the CD4+ cells). There are a number of compounds known to block syncytium formation without (markedly) aﬀecting virus binding to the cells. These compounds may therefore be assumed to directly interfere with the virus – cell fusion process, that is, fusion between the viral envelope and the outer cell membrane. The compounds that have been postulated to inhibit virus – cell fusion include the following: mannose-speciﬁc lectins (i.e., from Listeria ovata, Hippeastrum hybrid, Cymbidium hybrid, and Epipactis helleborine) and N-acetylglucosamine-speciﬁc plant lectins (i.e., from Urtica dioica) [17,18]; a derivative from polyphemusin, a peptide that is highly abundant in hemocyte debris of the horsehoe crab Limulus polyphemus [19]; succinylated human serum albumin, Suc-HSA [20], and aconitylated human serum albumin, Aco-HSA [21] (Fig. 2); and triterpene (i.e., betulinic acid) derivatives (Fig. 3) [22]. Research into the eﬃcacy of a number of natural products that have been described as anti-HIV agents is reviewed in Ref. 23. Despite their widely varying origin and structure, these diﬀerent classes of compounds seem to be targeted at the virus –cell fusion process, although the exact mechanism by which the compounds inhibit fusion,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 2 Modified human serum albumins from HSA: succinylated human serum albumin (Suc-HSA) and aconitylated human serum albumin (Aco-HSA).

and the target amino acid sequences (at gp120 and/or gp41) with which they putatively interact, remain to be elucidated. Most of the compounds inhibit HIV replication at concentrations of 0.1 to 1 Ag/mL and some (Aco-HSA and betulinic acid) are even eﬀective within the concentration range of 0.01 to 0.1 Ag/mL [21,22]. For the plant lectins [17,18] and modiﬁed serum albumins [20,21], the inhibitory eﬀects on HIV replication correlated closely with their inhibitory eﬀects on syncytium formation, which corroborates the hypothesis that their anti-HIV activity is due to inhibition of virus –cell fusion. Whereas the plant lectins are inhibitory to HIV-1, HIV-2, and a number of other (enveloped) viruses (viz., CMV, RSV, inﬂuenza A), at

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 3 Triterpene derivative: betulinic acid RPR103611: N V-{N-[3h-hydroxylup-20(29-ene-28-oyl]-8-aminooctanoyl}-L-statine.

concentrations well below the cytotoxicity threshold, the succinylated and aconitylated albumins inhibit only HIV, HIV-2 being (much) less susceptible to these compounds than HIV-1 [20,21]. Betulinic acid is even more restricted in its antiviral activity spectrum in that it is active only against HIV-1, and not even all HIV-1 strains [22]: speciﬁcally, betulinic acid is not active against the NDK strain of HIV-1. This must point to a highly speciﬁc molecular site for the interaction of betulinic acid [22] and should help in deciphering the target amino acid sequences (at gp120 or gp41) for this compound. The clinical potential of the fusion inhibitors in the therapy and/or prophylaxis of HIV infections remains a subject for further study. Since these compounds directly interfere with syncytium formation, they should be able to block HIV infections generated by both free virus particles and HIV-infected cells. It is not known how readily the virus may become resistant to this class of compounds. For betulinic acid, it has been ascertained that some HIV-1 strains (e.g., NDK) may be resistant ab initio.

IV. VIRUS UNCOATING INHIBITORS: BICYCLAMS Bicyclams (Fig. 4) consist of two cyclam (1,4,8,11-tetraazacyclotetradecane) units tethered via an aliphatic (i.e., propylene, as in JM2763) or aromatic bridge [i.e., phenylenebis(methylene), as in JM3100]. While the bicyclam JM2763 inhibits HIV-1 and HIV-2 replication at a concentration of 0.1 to 1 Ag/mL [24], the bicyclam JM3100 does so at a hundredfold lower concentration, that is, at a concentration that is more than

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 4 Bicyclam derivatives (A) JM2763 and (B) JM3100, each consisting of two cyclam (1,4,8,11-tetraazacyclotetradecane) moieties tethered via a propylene (JM2763) or phenylenebis(methylene) bridge (JM3100).

100,000-fold lower than the cytotoxic concentration [25]. In primary T4 lymphocytes and monocytes (macrophages), JM3100 inhibits HIV-1 replication at concentrations lower than 1 ng/mL [25]. The bicyclams represent the only retrovirus inhibitors that have been postulated to interfere with the viral uncoating process. This assumption has been based on ‘‘time of addition’’ experiments where the compounds (JM2763 and JM3100) were found to act at a stage following virus adsorption but preceding reverse transcription; and, since the compounds did not prove inhibitory to syncytium formation (JM2763) [or were inhibitory only at a concentration substantially higher than that required for inhibition of HIV replication (JM3100)], their target of action could be tentatively identiﬁed as a viral uncoating event. This hypothesis was then corroborated by ‘‘uncoating’’ experiments in which the viral RNA, recovered from HIV-infected cells that had been exposed to the compounds, was monitored for sensitivity to ribonuclease A: the viral RNA was protected against degradation by RNase A, as could be anticipated if the uncoating (i.e., dissociation of the capsid proteins from the viral RNA) had been impeded [24,25]. Current investigations are attempting to determine with which viral (capsid) proteins, and which amino acid residues of their target proteins, the bicyclam interact. The antiviral activity spectrum of the bicyclams is clearly diﬀerent from that of other anti-HIV agents in that the bicyclams are equally eﬀective against HIV-1 and HIV-2 but less eﬀective or not active against SIV (in human T lymphocytes). Given its high selectivity index ( >100,000) in vitro, the bicyclam JM3100 oﬀers great potential for the treatment of HIV-1 and HIV-2 infections in humans. Although the bicyclams may under some conditions select out drug-resistant variants from clinical HIV-1 (i.e., HE) strains, it has otherwise proved diﬃcult, or

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

even impossible, to generate resistance to the compounds (i.e., JM3100) after repeated passages in cell culture [25].

V. REVERSE TRANSCRIPTASE INHIBITORS INTERACTING WITH THE SUBSTRATE BINDING SITE A. Dideoxynucleoside Analogs In addition to the three anti-HIV agents [AZT (zidovudine), DDI (didanosine) and DDC (zalcitabine)] that have been formally approved by the U.S. Food and Drug Administration for the treatment of HIV infections, several other 2V,3V-dideoxynucleoside (ddN) analogues (Fig. 5), including 3V-ﬂuoro-2V,3V-dideoxy-5-chlorouridine (FddClUrd) and 2V,3V-didehydro-

Figure 5 Dideoxynucleoside (ddN) analogues: 2V,3V-dideoxycytidine (DDC), 3Vazido-2V,3V-dideoxythymidine (AZT), 3V-fluoro-2V,3V-dideoxythymidine (FLT), 2V,3V-didehydro-2V,3V-dideoxythymidine (D4T), 3V-thia-2V,3V-dideoxycytidine (3TC), 3V-thia-2V,3V-dideoxy-5-fluorocytidine (FTC), and 2V,3V-dideoxy-L-cytidine (L-DDC).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

2V,3V-dideoxythymidine [D4T (stavudine)], have been reported to inhibit HIV replication (for review, see Refs. [26 – 28]). In particular, FddClUrd appears to be an attractive candidate for further development, since it is much less toxic to the host cells than AZT and most other ddN analogues [29,30]. Also ranking among the most promising ddN analogues are 3Vthia-2V,3V-dideoxycytidine [3TC (lamivudine)] and 3V-thia-2V,3V-dideoxy5-ﬂuorocytidine (FTC), which are actually more active in their (– )-h- or L-isomeric form than in the (+)-h- or D-isomeric form [31,32]. All ddN analogues, including 3TC and FTC, act in a similar fashion; that is, following intracellular phosphorylation to their 5V-triphosphate form, they serve as competitive inhibitors/alternate substrates of the reverse transcriptase (RT) reaction, thus leading to chain termination, as has been clearly demonstrated with AZT [33]. The anti-HIV activity of ddN analogues is critically dependent on their intracellular phosphorylation, the ﬁrst phosphorylation step being the most crucial. For some compounds (viz., 2V,3V-dideoxyuridine) and in some cells (viz., monocytes/macrophages), the nucleoside kinase activity of the cells may be inadequate to satisfactorily accomplish the ﬁrst phosphorylation step; and thus prodrugs, including aryl methoxyglycinyl derivatives [34] and bis[S(2-hydroxyethylsulﬁdyl)-2-thioethyl] esters [35] have been designed that deliver the 5V-monophosphate form intracellularly, bypassing the ﬁrst phosphorylation step. The antiviral activity spectrum of the ddN analogues should, in principle, extend to all retroviruses as well as hepadnaviruses [i.e., hepatitis B virus (HBV)], since HBV, like retroviruses, replicates through an RNA template-driven RT process. Indeed, various ddN analogues (particularly, the L-enantiomeric forms 3TC, FTC, and L-DDC) have been shown to inhibit HBV replication [36 –38]. Consequently, 3TC is, at present, pursued as a potential drug candidate for the treatment of both HIV and HBV infections. Prolonged AZT therapy of HIV-infected individuals leads to a reduction of virus sensitivity to the drug [39]. This reduced sensitivity, generally termed ‘‘resistance,’’ appears to be based on the following mutations in the HIV-1 RT [40,41]: 41 Met ! Leu, 67 Asp ! Asn, 70 Lys ! Arg, 215 Thr ! Phe/Tyr, and 219 Lys ! Gln. Of these mutations, the 215 Thr ! Tyr mutation has been the most frequently detected among AZT-resistant HIV isolates from patients under prolonged AZT therapy [42]. The 74 Leu ! Val mutation is responsible for resistance to DDI [43], and the 184 Met ! Val mutation confers resistance to 3TC, FTC, DDC, and DDI [44 – 46].

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

The mutations at position 74 (Leu ! Val) and 184 (Met ! Val) of the HIV-1 RT do not lead to cross-resistance to AZT. Nor would the 215 Thr ! Tyr mutation lead to cross-resistance to 3TC, FTC, DDC, or DDI. In fact, the mutations at positions 74 and 215 seem to counteract each other, and so do the mutations at positions 184 and 215. Based on this ‘‘mutually counteracting mutation’’ principle [47], drug combinations could be envisaged that, if combined, might counteract emergence of resistance to one another: namely, combinations of AZT with either DDI, 3TC, FTC, or DDC. As will be explained further, these two-drug combinations may be extended to three-drug or four-drug combinations, following the addition of one or more of the HIV-1-speciﬁc nonnucleoside RT inhibitors (NNRTIs).

B. Acyclic Nucleoside Phosphonates Acyclic nucleoside phosphonates (ANPs) (Fig. 6) may be regarded as analogous to the ddN monophosphates, thus allowing us to circumvent the ﬁrst phosphorylation step required for the intracellular activation of the compounds. After they have been taken up as such by the cells, the acyclic nucleoside phosphonates (PMEA, PMEDAP, PMPA, PMPDAP, FPMPA, and FPMPDAP) are converted intracellularly to their respective diphosphate form (PMEApp, PMEDAPpp, PMPApp, PMPDAPpp, FPMPApp, and FPMPDAPpp) and, in such form they interact as competitive inhibitors, alternate substrates, or chain terminators with the reverse transcriptase [48 –50]. PMEA and its congeners are more eﬀective in vivo than could be predicted from their in vitro potency. While less potent as an antiretrovirus agent than AZT in vitro, PMEA proved clearly superior to AZT when the two drugs were compared for their eﬀectiveness in vivo, in mice infected with murine Moloney sarcoma virus [51,52]. PMEA was also shown to be eﬀective against various other retrovirus infections, including Friend leukemia virus (FLV), Rauscher leukemia virus (RLV), and LPBM5 (murine AIDS) virus infection in mice, feline leukemia virus (FeLV) or feline immunodeﬁciency virus (FIV) infection in cats, and SIV infection in macaque (rhesus) monkeys (for review, see Ref. 53). In the latter model [54], again PMEA proved far superior to AZT in suppressing several parameters of the disease. The antiviral activity spectrum of PMEA, PMEDAP, and their congeners is not conﬁned to retroviruses but also extends to hepadnaviruses (e.g., HBV). PMEA has proved eﬀective against duck HBV infec-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 6 Acyclic nucleoside phosphonates (ANPs): 9-(2-phosphonylmethoxyethyl)-adenine (PMEA) and -2,6-diaminopurine (PMEDAP), (R )-9-(2-phosphonylmethoxypropyl)-adenine (PMPA) and -2,6-diaminopurine (PMPDAP), (S )-9-(3fluoro-2-phosphonylmethoxypropyl)-adenine (FPMPA) and -2,6-diaminopurine (FPMPDAP), and the bis(pivaloyloxymethyl) ester of PMEA [Bis(pom)-PMEA].

tion in both duck hepatocytes and Pekin ducks [55]. For PMEA and PMEDAP, but not for PMPA, PMPDAP, FPMPA, or FPMPDAP, the activity spectrum also extends to herpesviruses (e.g., HSV, CMV). This would make PMEA and PMEDAP particularly attractive as therapeutic modalities in AIDS patients, since they might be useful not only for the treatment of the underlying HIV infection but also for the therapy/ prophylaxis of the intercurrent HSV or CMV infections.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Another attractive feature of PMEA, and, in fact, all ANPs, is prolonged antiviral action, lasting for several days, or even one week or longer, after a single-dose administration. This long-lasting antiviral action may be related to the long half-life of the active metabolites (e.g., PMEApp) within the cells and may permit infrequent (e.g., weekly) dosing of the ANPs in the prophylaxis and/or therapy of (retro)virus infections. Little is known on how readily or rapidly retro- or herpesviruses may develop resistance to the ANPs. In the in vitro and in vivo experiments done so far with PMEA, PMEDAP, or any of the other ANPs, resistance development did not seem to occur, but further studies are needed to address this issue. Since the ANPs are only slowly taken up by the cells and poorly absorbed following oral administration, some eﬀorts have been directed toward the development of prodrugs (esters) that would be better taken up by the cells. These eﬀorts have yielded the bispivaloyloxymethyl [bis(pom)] derivative of PMEA (Fig. 6) [56]. Bis(pom)-PMEA shows a cellular uptake increased more than a hundredfold, as well as ﬁvefold better oral bioavailability than the parent compound [57]. Both PMEA (given intravenously) and bis(pom)-PMEA (given perorally) are now in clinical trials in patients with AIDS.

VI. REVERSE TRANSCRIPTASE INHIBITORS INTERACTING WITH A NONSUBSTRATE BINDING SITE: NON-NUCLEOSIDE REVERSE TRANSCRIPTASE INHIBITORS The identiﬁcation of the HIV-1-speciﬁc non-nucleoside reverse transcriptase inhibitors (NNRTIs) as a separate class of HIV inhibitors was heralded by the discovery of the tetrahydroimidazo[4,5,1-jk][1,4]benzodiazepin-2(1H )-one and -thione (TIBO) derivatives (Fig. 7) [58,59] and 1-(2-hydroxyethoxymethyl)-6-(phenylthio)thymine (HEPT) derivatives (Fig. 8) [60,61]. The ﬁrst TIBO derivatives (R82150, R82913) were the ﬁrst NNRTIs [58] postulated to act as inhibitors of HIV-1 RT [59]. For the HEPT derivatives it became evident that they also interact speciﬁcally with HIV-1 RT after a number of derivatives (i.e., E-EPU, E-EBU, and E-EBU-dM) had been synthesized that were more active than HEPT itself [62,63]. Following HEPT and TIBO, several other compounds, i.e., nevirapine, pyridinone, and bis(heteroaryl)piperazine (BHAP), were

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 7 Tetrahydroimidazo[4,5,1-jk][1,4]benzodiazepin-2(1H )-one (TIBO) derivatives (A) R82913 and (B) R86183 (with a chlorine substituted in the 9- or 8position, respectively).

Figure 8 (A) 1-(2-Hydroxyethoxymethyl)-6-(phenylthio)thymine (HEPT). (B) 5Isopropyl-1-(ethoxymethyl)-6-benzyluracil (I-EBU, MKC-442).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

described as HIV-1-speciﬁc RT inhibitors (for a review on the HIV-1speciﬁc RT inhibitors, see Refs. 28 and 64). The HEPT and TIBO derivatives were discovered as the result of a systematic evaluation for anti-HIV activity in cell culture. They were later found to achieve their anti-HIV-1 activity through an interaction with the HIV-1 RT. In contrast, nevirapine, pyridinone, and BHAP emerged from a screening program for HIV-1 RT inhibitors. The anti-HIV-1 activity of these compounds was subsequently conﬁrmed in cell culture. Like the HEPT and TIBO derivatives, the 2V,5V-bis-O-(tert-butyldimethylsilyl)-3Vspiro-5VV-(4VV-amino-1VV,2VV-oxathiole-2VV,2VV-dioxide)-pyrimidine (TSAO) derivatives (Fig. 9) [65,66] and a-anilinophenylacetamides (a-APA) (Fig. 10) [67] were discovered through the evaluation of their anti-HIV activity in cell culture. Subsequently, they were found to act as speciﬁc inhibitors of HIV-1 RT. Yet other compounds have been found to inhibit HIV-1 replication through a speciﬁc interaction with HIV-1 RT (i.e., quinoxaline S-2720 [68], 5-chloro-3-(phenylsulfonyl)indole-2-carboxamide [69], dihydrothiazoloisoindolones [70] and a number of natural substances (e.g., calanolide A and inophyllums, from the tropical rain forest trees Calophyllum lanigerum and Calophyllum inophyllum, respectively) [71,72]. All these and yet other compounds could be considered to be NNRTIs. The most potent among the NNRTIs, some of the HEPT derivatives (E-EBU-dM) [63] and a-

Figure 9 2V,5V-Bis-O-(tert-butyldimethylsilyl)-3V-spiro-5W-(4W-amino-1W,2W-oxathiole-2W,2W-dioxide)pyrimidine (TSAO) derivatives TSAO-T, TSAO-m3T, and TSAO-e3T.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 10 a-Anilinophenylacetamide (a-APA) derivatives (A) R18893, (B) R88703, and (C) R89439.

APA derivatives (R89439) [67], inhibit HIV-1 replication at a concentration of approximately 1 ng/mL, that is, 100,000-fold below the cytotoxicity threshold. While the ddNs and ANPs must be converted intracellularly to their 5V-triphosphates (ddNTPs) or diphosphate derivatives before they can interact as competitive inhibitors/alternate substrates with regard to the natural substrates (dNTPs), the NNRTIs do not need any metabolic conversion to interact, noncompetitively with respect to the dNTPs, at an allosteric, non –substrate binding site of the HIV-1 RT. Through the analysis of NNRTI-resistant mutants, combined with site-directed mutagenesis studies, it has become increasingly clear which amino acid residues are involved in the interaction of the NNRTIs with HIV-1 RT, and, since the conformation of the HIV-1 RT has been resolved at 3.0 A˚ resolution [73], it is now possible to visualize the binding site of the NNRTIs [74]. The antiviral activity spectrum of the NNRTIs is limited to HIV-1, probably because only HIV-1 RT contains a pocket site at which the

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

NNRTIs may bind. The high speciﬁcity displayed by the NNRTIs in their binding to HIV-1 RT signals that it should, a priori, be relatively easy for the enzyme (and the virus) to escape the inhibitory eﬀects of the NNRTIs through mutations of the amino acid residues that either are directly involved in the binding of the NNRTIs or contribute to the conﬁguration of the pocket that is ideal for NNRTI binding. From pilot studies carried out in the clinic with the NNRTIs TIBO R82913 [75] and pyridinone L-697,661 [76], it appears that the compounds are well tolerated and do not cause toxic side eﬀects. Most of the HIV-1 isolates obtained from the patients treated with TIBO R82913 appeared to be as sensitive to the compound as wild-type virus; only two HIV-1 variants were isolated, showing a sensitivity that was reduced 20-fold or more than 100-fold, the latter being caused by a mutation (Tyr ! Leu) at position 188 of the RT [77]. In fact, the latter mutation was lost upon passaging the virus in vitro in cord blood lymphocytes. Following treatment of the patients with pyridinone L-697,661, drug-resistant HIV-1 variants appeared that contained mutations at the RT positions 103 (Lys ! Asn) and 181 (Tyr ! Cys) [76]. HIV-1 resistance to NNRTIs rapidly arises following passage of the virus in cell culture in the presence of the compounds. The 181 Tyr ! Cys mutation is most commonly seen, and it leads to resistance, or at least to reduced sensitivity, to most of the NNRTIs (i.e., TIBO, HEPT, nevirapine, pyridinone, BHAP, TSAO, a-APA) [78 – 84]. The 188 Tyr ! His mutation is associated with resistance to TIBO [85], but not nevirapine [82]. The 103 Lys ! Asn mutation is associated mainly with resistance to TIBO and pyridinone [78,85]. The 100 Leu ! Ile mutation is associated mainly with resistance to TIBO [85,86]. The 106 Val ! Ala mutation mainly leads to resistance to nevirapine and HEPT [83,84,87]. The 138 Glu ! Lys mutation is responsible for resistance to TSAO [88,89]. The 190 Gly ! Glu mutation accounts for resistance to quinoxaline [68], while also leading to a dramatic reduction in RT activity [90]; and the 236 Pro ! Leu mutation is responsible for resistance to BHAP [91]. The rapid emergence of drug-resistant HIV-1 mutants under selective pressure of the HIV-1-speciﬁc RT inhibitors has been generally viewed as a limitation for, if not an argument against, the clinical usefulness of these compounds. Yet, several aspects of virus – drug resistance, particularly with respect to the NNRTIs, remain to be addressed before the problem of resistance can be fully assessed. For example, how pathogenic are drugresistant variants in comparison to wild-type virus? How readily are such drug-resistant variants transmitted from one person to another? Do virus-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

resistant variants persist when the drug is withdrawn, or do they readily revert to the wild type? Assuming that the development of drug resistance may indeed compromise the clinical usefulness of the NNRTIs, how might this problem be prevented or circumvented? If resistance develops to one of the NNRTIs, treatment could be switched to any of the other NNRTIs to which the virus has retained sensitivity. For example, 5-chloro-3-(phenylsulfonyl)indole-2-carboxamide [69] is active against the HIV-1 strains that, because of the 103 Lys ! Asn mutation or 181 Tyr ! Cys mutation, have acquired resistance to various other NNRTIs (i.e., TIBO, nevirapine, pyridinone, BHAP). The a-APA derivative R89439 [67] is active against the 100 Leu ! Ile mutant, which is resistant to the TIBO derivatives R82913 and R86183. Within the TIBO class, a minor chemical modiﬁcation, the shifting of the chlorine atom from the 9-position (R82913) to the 8-position (R86183), suﬃces to restore activity against the 181 Tyr ! Cys mutant [92]. Similarly, pyridinone L-702,019, which diﬀers from its predecessor L-696,229 only by the addition of two chlorine atoms (in the benzene ring) and substitution of sulfur for oxygen (in the pyridine ring), remains remarkably active against HIV-1 mutants containing the 103 Lys ! Asn or 181 Tyr ! Cys mutation [93]. In some instances resistance to one of the NNRTIs may even be accompanied by hypersensitivity to others: the 236 Pro ! Leu mutation, which causes resistance to BHAP, confers 10-fold increased sensitivity to TIBO, nevirapine, and pyridinone [91]. The 181 Tyr ! Cys mutation, which is responsible for resistance to most NNRTIs, has been found to suppress the 215 mutation (Thr ! Phe/ Tyr), which is responsible for resistance to AZT [94], and, vice versa, the 181 Tyr ! Cys mutation can be suppressed by AZT, which thus means that the mutations at positions 181 and 215 counteract each other. Yet other mutations have proved to counteract each other: 236 Pro ! Leu vs 138 Glu ! Lys, and, as mentioned, 215 Thr ! Phe/Tyr vs 184 Met ! Val, and 215 Thr ! Phe/Tyr vs 74 Leu ! Val [47]. Based on the resistance mutations that counteract each other, combinations of diﬀerent drugs could be envisaged—namely, combinations of AZT with either TIBO, a-APA, HEPT, nevirapine, or pyridinone—and these two drug combinations could be extended to three- or four-drug combinations by the addition of another ddN analogue (such as 3TC) and/or another NNRTI (such as BHAP or TSAO). What would seem to be an attractive approach to the prevention of resistance development is the ‘‘knocking-out’’ strategy [95]. If NNRTIs,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

such as BHAP (U-88204 or U-90152), are used from the start at a suﬃciently high concentration (i.e., 1 or 3 AM, respectively), they completely suppress virus replication [96,97], with the result that the virus is ‘‘knocked out’’ and does not have the opportunity to become resistant. If U-90152 is combined with AZT, the concentrations of the individual drugs can be lowered to achieve total virus clearance [97]. Five NNRTIs (TIBO, HEPT, nevirapine, pyridinone, and BHAP) have been shown to ‘‘knock out’’ HIV-1 in cell culture when used at concentrations (1 –10 Ag/mL) that are nontoxic to the cells [95]. That the virus was really knocked out, and thus the cell culture cleared (‘‘sterilized’’) from the HIV-1 infection by the NNRTIs, was ascertained by two successive rounds of 35-cycle PCR (polymerase chain reaction) analysis, which failed to reveal the presence of any proviral DNA [95]. Thus, when used at ‘‘knocking-out’’ concentrations, the NNRTIs may be expected to eﬀect a long-lasting suppression of HIV-1 replication. This ‘‘knocking-out’’ phenomenon could be obtained at lower concentrations if the NNRTIs were combined with each other, or with any of the ddN analogues (i.e., AZT), particularly if selected on the basis of the ‘‘mutually counteracting mutation’’ principle.

VII. CONCLUSION An acute HIV infection can be blocked at any of the following stages of the infection: virus adsorption, virus – cell fusion, viral uncoating, and reverse transcription. At the reverse transcriptase (RT) level, chemotherapeutic intervention could be envisaged at either the substrate or a non – substrate binding site. Polyanionic substances (i.e., sulfated polysaccharides) prevent virus adsorption; plant lectins, succinylated (or aconitylated) albumins, and triterpene (i.e., betulinic acid) derivatives interfere with virus – cell fusion; bicyclams inhibit viral uncoating; 2V,3V-dideoxynucleosides (ddNs) and acyclic nucleoside phosphonate analogues, following intracellular conversion to their phosphorylated derivatives, interact with the substrate binding site of the RT; and the nonnucleoside reverse transcriptase inhibitors (NNRTIs) are targeted at a non –substrate binding site of HIV-1 RT. Some of these compounds (viz., bicyclams) and, among the NNRTIs, some of the HEPT and a-APA derivatives, were found to inhibit HIV-1 replication at concentrations (f1 ng/mL) that were 100,000-fold or more below the cytotoxicity threshold. As a rule, it may be postulated that the more speciﬁc the antiviral action, the more likely the development of virus – drug resistance; hence, NNRTIs, which engage in a highly

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

speciﬁc interaction with HIV-1 RT, rapidly lead to the emergence of drugresistant virus strains. To prevent such drug-resistant virus strains from emerging, several strategies could be envisaged, the most attractive being the combination of several drugs at concentrations high enough to ‘‘knock out’’ the virus from the start. This ‘‘knocking-out’’ phenomenon has been achieved with the NNRTIs, regardless of whether combined with any of the ddN analogues, and it may be extended to combinations of drugs that interact at targets other than the reverse transcriptase.

ACKNOWLEDGMENTS The original investigations of the author are supported by the Biomedical Research Programme of the European Community, the Belgian Nationaal Fonds voor Wetenschappelijk Onderzoek, the Belgian Fonds voor Geneeskundig Wetenschappelijk Onderzoek, the Belgian Geconcerteerde Onderzoeksacties, and the Janssen Research Foundation. I thank Christiane Callebaut for her dedicated editorial assistance.

REFERENCES 1. De Clercq E. New perspectives for the chemotherapy and chemoprophylaxis of AIDS (acquired immune deﬁciency syndrome). Verh K Acad Geneeskd Belg 1992; 54:57 – 89. 2. De Clercq E. Anti-HIV agents interfering with the initial stages of the HIV replicative cycle. In: Morrow WJW, Haigwood NL, eds. HIV Molecular Organization, Pathogenicity and Treatment. Amsterdam: Elsevier Science Publishers, 1993:267 – 292. 3. De Clercq E. The emerging role of fusion inhibitors in HIV infection. Drug Dev Res 1999; 2:321 – 331. 4. De Clercq E. Inhibition of HIV infection by bicyclams, highly potent and speciﬁc CXCR4 antagonists. Mol Pharmacol 2000; 57:833 – 839. 5. De Clercq E, Schols D. Inhibition of HIV infection by CXCR4 and CCR5 chemokine receptor antagonists. Antiviral Chem Chemother 2001; 12 (suppl 1):19 – 31. 6. De Clercq E. New developments in anti-HIV chemotherapy. Curr Med Chem 2001; 8:1543 – 1572. 7. De Clercq E. Anti-HIV activity of sulfated polysaccharides. In: Yalpani M, ed. Carbohydrates and Carbohydrate Polymers, Analysis, Biotechnology,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

Modiﬁcation, Antiviral, Biomedical and Other Applications. Mount Prospect, IL: ATL Press, 1993:87 – 100. Schols D, Pauwels R, Desmyter J, De Clercq E. Presence of class II histocompatibility DR proteins on the envelope of human immunodeﬁciency virus demonstrated by FACS analysis. Virology 1992; 189:374 – 376. Callahan LN, Phelan M, Mallinson M, Norcross MA. Dextran sulfate blocks antibody binding to the principal neutralizing domain of human immunodeﬁciency virus type 1 without interfering with gp120-CD4 interactions. J Virol 1991; 65:1543 – 1550. Batinic D, Robey FA. The V3 region of the envelope glycoprotein of human immunodeﬁciency virus type 1 binds sulfated polysaccharides and CD4derived synthetic peptides. J Biol Chem 1992; 267:6664 – 6671. Schols D, Baba M, Pauwels R, Desmyter J, De Clercq E. Speciﬁc interaction of aurintricarboxylic acid with the human immunodeﬁciency virus/CD4 cell receptor. Proc Natl Acad Sci USA 1989; 86:3322 – 3326. Schols D, Pauwels R, Witvrouw M, Desmyter J, De Clercq E. Diﬀerential activity of polyanionic compounds and castanospermine against HIV replication and HIV-induced syncytium formation depending on virus strain and cell type. Antiviral Chem Chemother 1992; 3:23 – 29. Lorentsen KJ, Hendrix CW, Collins JM, Kornhauser DM, Petty BG, Klecker RW, Flexner C, Eckel RH, Lietman PS. Dextran sulfate is poorly absorbed after oral administration. Ann Intern Med 1989; 111:561 – 566. Moriya T, Saito K, Kurita H, Matsumoto K, Otake T, Mori H, Morimoto M, Ueba N, Kunita N. A new candidate for an anti-HIV-1 agent: modiﬁed cyclodextrin sulfate (mCDS71). J Med Chem 1993; 36:1674 – 1677. Otake T, Schols D, Witvrouw M, Naesens L, Nakashima H, Moriya T, Kurita H, Matsumoto K, Ueba N, De Clercq E. Modiﬁed cyclodextrin sulphates (mCDS11) have potent inhibitory activity against HIV and high oral bioavailability. Antiviral Chem Chemother 1994; 5:155 – 161. Baˆrzu T, Level M, Petitou M, Lormeau J-C, J. Choay J, Schols D, Baba M, Pauwels R, Witvrouw M, De Clercq E. Preparation and anti-HIV activity of O-acylated heparin and dermatan sulfate derivatives with low anticoagulant eﬀect. J Med Chem 1993; 36:3546 – 3555. Balzarini J, Schols D, Neyts J, Van Damme E, Peumans W, De Clercq E. a(1-3)- and a-(1-6)-d-mannose-speciﬁc plant lectins are markedly inhibitory to human immunodeﬁciency virus and cytomegalovirus infections in vitro. Antimicrob Agents Chemother 1991; 35:410 – 416. Balzarini J, Neyts J, Schols D, Hosoya M, Van Damme E, Peumans W, De Clercq E. The mannose-speciﬁc plant lectins from Cymbidium hybrid and Epipactis helleborine and the (N-acetylglucosamine)n-speciﬁc plant lectin from Urtica dioica are potent and selective inhibitors of human immunodeﬁciency virus and cytomegalovirus replication in vitro. Antiviral Res 1992; 18:191 – 207.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

19. Nakashima H, Masuda M, Murakami T, Koyanagi Y, Matsumoto A, Fujii N, Yamamoto N. Anti – human immunodeﬁciency virus activity of a novel synthetic peptide, T22 ([Tyr-5,12,Lys-7]polyphemusin II): a possible inhibitor of virus – cell fusion. Antimicrob Agents Chemother 1992; 36:1249 – 1255. 20. Jansen RW, Molema G, Pauwels R, Schols D, De Clercq E, Meijer DKF. Potent in vitro anti – human immunodeﬁciency virus-1 activity of modiﬁed human serum albumins. Mol Pharmacol 1991; 39:818 – 823. 21. Jansen RW, Schols D, Pauwels R, De Clercq E, Meijer DKF. Novel, negatively charged, human serum albumins display potent and selective in vitro anti – human immunodeﬁciency virus type 1 activity. Mol Pharmacol 1993; 44:1003 – 1007. 22. Mayaux J-F, Bousseau A, Pauwels R, Huet T, He´nin Y, Dereu N, Evers M, Soler F, Poujade C, De Clercq E, Le Pecq J-B. Triterpene derivatives that block entry of human immunodeﬁciency virus type 1 into cells. Proc Natl Acad Sci USA 1994; 91:3564 – 3568. 23. De Clercq E. Current lead natural products for the chemotherapy of human immunodeﬁciency virus (HIV) infection. Med Res Rev 2000; 20:323 – 349. 24. De Clercq E, Yamamoto N, Pauwels R, Baba M, Schols D, Nakashima H, Balzarini J, Debyser Z, Murrer BA, Schwartz D, Thornton D, Bridger G, Fricker S, Henson G, Abrams M, Picker D. Potent and selective inhibition of human immunodeﬁciency virus (HIV)-1 and HIV-2 replication by a class of bicyclams interacting with a viral uncoating event. Proc Natl Acad Sci USA 1992; 89:5286 – 5290. 25. De Clercq E, Yamamoto N, Pauwels R, Balzarini J, Witvrouw M, De Vreese K, Debyser Z, Rosenwirth B, Peichl P, Datema R, Thornton D, Skerlj R, Gaul F, Padmanabhan S, Bridger G, Henson G, Abrams M. Highly potent and selective inhibition of human immunodeﬁciency virus by the bicylam derivative JM3100. Antimicrob Agents Chemother 1994; 38:668 – 674. 26. De Clercq E. HIV inhibitors targeted at the reverse transcriptase. AIDS Res Human Retrovir 1992; 8:119 – 134. 27. De Clercq E. Hamao Umezawa Memorial Award Lecture: An odyssey in the viral chemotherapy ﬁeld. Int J Antimicrob Agents 2001; 18:309 – 328. 28. De Clercq E. New developments in anti-HIV chemotherapy. Farmaco 2001; 56:3 – 12. 29. Balzarini J, Van Aerschot A, Pauwels R, Baba M, Schols D, Herdewijn P, De Clercq E. 5-Halogeno-3V-ﬂuoro-2V,3V-dideoxyuridines as inhibitors of human immunodeﬁciency virus (HIV): potent and selective anti-HIV activity of 3V-ﬂuoro-2V,3V-dideoxy-5-chlorouridine. Mol Pharmacol 1989; 35:571 – 577. 30. Van Aerschot A, Herdewijn P, Balzarini J, Pauwels R, De Clercq E. 3VFluoro-2V,3V-dideoxy-5-chlorouridine: most selective anti-HIV-1 agent among a series of new 2V- and 3V-ﬂuorinated 2V,3V-dideoxynucleoside analogues. J Med Chem 1989; 32:1743 – 1749.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

31. Schinazi RF, Chu CK, Peck A, McMillan A, Mathis R, Cannon D, Jeong L-S, Beach JW, Choi W-B, Yeola S, Liotta DC. Activities of the four optical isomers of 2V,3V-dideoxy-3V-thiacytidine (BCH-189) against human immunodeﬁciency virus type 1 in human lymphocytes. Antimicrob Agents Chemother 1992; 36:672 – 676. 32. Schinazi RF, McMillan A, Cannon D, Mathis R, Lloyd RM, Peck A, Sommadossi J-P, St Clair M, Wilson J, Furman PA, Painter G, Choi W-B, Liotta DC. Selective inhibition of human immunodeﬁciency viruses by racemates and enantiomers of cis-5-ﬂuoro-1-[2-(hydroxymethyl)-1,3-oxathiolan-5-yl]cytosine. Antimicrob Agents Chemother 1992; 36:2423 – 2431. 33. Huang P, Farquhar D, Plunkett W. Selective action of 3V-azido-3V-deoxythymidine 5V-triphosphate on viral reverse transcriptases and human DNA polymerases. J Biol Chem 1990; 265:11914 – 11918. 34. McGuigan C, Pathirana RN, Balzarini J, De Clercq E. Intracellular delivery of bioactive AZT nucleotides by aryl phosphate derivatives of AZT. J Med Chem 1993; 36:1048 – 1052. 35. Puech F, Gosselin G, Lefebvre I, Pompon A, Aubertin A-M, Kirn A, Imbach J-L. Intracellular delivery of nucleoside monophosphates through a reductase-mediated activation process. Antiviral Res 1993; 22:155 – 174. 36. Doong S-L, Tsai C-H, Schinazi RF, Liotta DC, Cheng Y-C. Inhibition of the replication of hepatitis B virus in vitro by 2V,3V-dideoxy-3V-thiacytidine and related analogues. Proc Natl Acad Sci USA 1991; 88:8495 – 8499. 37. Furman PA, Davis M, Liotta DC, Paﬀ M, Frick LW, Nelson DJ, Dornsife RE, Wurster JA, Wilson LJ, Fyfe JA, Tuttle JV, Miller WH, Condreay L, Averett DR, Schinazi RF, Painter GR. The anti-hepatitis B virus activities, cytotoxicities, and anabolic proﬁles of the ( – ) and (+) enantiomers of cis-5ﬂuoro-1-[2-(hydroxymethyl)-1,3-oxathiolan-5-yl]cytosine. Antimicrob Agents Chemother 1992; 36:2686 – 2692. 38. Lin T-S, Luo M-Z, Liu M-C, Pai SB, Dutschman GE, Cheng Y-C. Antiviral activity of 2V,3V-dideoxy-h-L-5-ﬂuorocytidine (h-L-FddC) and 2V,3V-dideoxyh-L-cytidine (h-L-ddC) against hepatitis B virus and human immunodeﬁciency virus type 1 in vitro. Biochem Pharmacol 1994; 47:171 – 174. 39. Larder BA, Darby G, Richman DD. HIV with reduced sensitivity to zidovudine (AZT) isolated during prolonged therapy. Science 1989; 243:1731 – 1734. 40. Larder BA, Kemp SD. Multiple mutations in HIV-1 reverse transcriptase confer high-level resistance to zidovudine (AZT). Science 1989; 246:1155 – 1158. 41. Kellam P, Boucher CAB, Larder BA. Fifth mutation in human immunodeﬁciency virus type 1 reverse transcriptase contributes to the development of high-level resistance to zidovudine. Proc Natl Acad Sci USA 1992; 89:1934 – 1938. 42. Mayers DL, McCutchan FE, Sanders-Buell EE, Merritt LI, Dilworth S, Fowler AK, Marks CA, Ruiz NM, Richman DD, Roberts CR, Burke

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

43.

44.

45.

46.

47. 48.

49.

50.

51.

52.

DS. Characterization of HIV isolates arising after prolonged zidovudine therapy. J AIDS 1992; 5:749 – 759. St Clair MH, Martin JL,Tudor-Williams G, Bach MC, Vavro CL, King DM, Kellam P, Kemp SD, Larder BA. Resistance to ddI and sensitivity to AZT induced by a mutation in HIV-1 reverse transcriptase. Science 1991; 253:1557 – 1559. Gao Q, Gu Z, Parniak MA, Cameron J, Cammack N, Boucher C, Wainberg MA. The same mutation that encodes low-level human immunodeﬁciency virus type 1 resistance to 2V,3V-dideoxyinosine and 2V,3V-dideoxycytidine confers high-level resistance to the ( – )enantiomer of 2V,3V-dideoxy-3V-thiacytidine. Antimicrob Agents Chemother 1993; 37:1390 – 1392. Tisdale M, Kemp SD, Parry NR, Larder BA. Rapid in vitro selection of human immunodeﬁciency virus type 1 resistant to 3V-thiacytidine inhibitors due to a mutation in the YMDD region of reverse transcriptase. Proc Natl Acad Sci USA 1993; 90:5653 – 5656. Schinazi RF, Lloyd RM Jr, Nguyen M-H, Cannon DL, McMillan A, Ilksoy N, Chu CK, Liotta DC, Bazmi HZ, Mellors JW. Characterization of human immunodeﬁciency viruses resistant to oxathiolane – cytosine nucleosides. Antimicrob Agents Chemother 1993; 37:875 – 881. De Clercq E. HIV resistance to reverse transcriptase inhibitors. Biochem Pharmacol 1994; 47:155 – 169. Balzarini J, Hao Z, Herdewijn P, Johns DG, De Clercq E. Intracellular metabolism and mechanism of anti-retrovirus action of 9-(2-phosphonylmethoxyethyl)adenine, a potent anti – human immunodeﬁciency virus compound. Proc Natl Acad Sci USA 1991; 88:1499 – 1503. Balzarini J, Holy´ A, Jindrich J, Dvorakova H, Hao Z, Snoeck R, Herdewijn P, Johns DG, De Clercq E. 9-[(2RS)-3-Fluoro-2-phosphonylmethoxypropyl] derivatives of purines: a class of highly selective antiretroviral agents in vitro and in vivo. Proc Natl Acad Sci USA 1991; 88:4961 – 4965. Balzarini J, Holy´ A, Jindrich J, Naesens L, Snoeck R, Schols D, De Clercq E. Diﬀerential antiherpesvirus and antiretrovirus eﬀects of the (S) and (R) enantiomers of acyclic nucleoside phosphonates: potent and selective in vitro and in vivo antiretrovirus activities of (R)-9-(2-phosphonomethoxypropyl)-2,6-diaminopurine. Antimicrob Agents Chemother 1993; 37:332 – 338. Balzarini J, Naesens L, Herdewijn P, Rosenberg I, Holy´ A, Pauwels R, Baba M, Johns DG, De Clercq E. Marked in vivo antiretrovirus activity of 9-(2phosphonylmethoxyethyl)adenine, a selective anti – human immunodeﬁciency virus agent. Proc Natl Acad Sci USA 1989; 86:332 – 336. Balzarini J, Sobis H, Naesens L, Vandeputte M, De Clercq E. Inhibitory eﬀects of 9-(2-phosphonylmethoxyethyl)adenine and 3V-azido-2V,3V-dideoxythymidine on tumor development in mice inoculated intracerebrally with Moloney murine sarcoma virus. Int J Cancer 1990; 45:486 – 489.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

53. Naesens L, Balzarini J, De Clercq E. Therapeutic potential of 9-(2-phosphonylmethoxyethyl)adenine (PMEA) as an antiviral drug. Rev Med Virol 1994; 4 :147 – 159. 54. Tsai C-C, Follis KE, Sabo A, Grant RF, Bartz C, Nolte RE, Benveniste RE, Bischofberger N. Preexposure prophylaxis with 9-(2-phosphonylmethoxyethyl)adenine against simian immunodeﬁciency virus infection in macaques. J Infect Dis 1994; 169:260 – 266. 55. Heijtink RA, De Wilde GA, Kruining J, Berk L, Balzarini J, De Clercq E, Holy´ A, Schalm SW. Antiviral activity of 9-(2-phosphonylmethoxyethyl) adenine (PMEA) on human and duck hepatitis B virus infection. Antiviral Res 1993; 21:141 – 153. 56. Starrett JE Jr, Tortolani DR, Hitchcock MJM, Martin JC, Mansuri MM. Synthesis and in vitro evaluation of a phosphonate prodrug: bis(pivaloyloxymethyl)-9-(2-phosphonylmethoxyethyl)adenine. Antiviral Res 1992; 19:267 – 273. 57. Srinivas RV, Robbins BL, Connelly MC, Gong Y-F, Bischofberger N, Fridland A. Metabolism and in vitro antiretroviral activities of bis(pivaloyloxymethyl) prodrugs of acyclic nucleoside phosphonates. Antimicrob Agents Chemother 1993; 37:2247 – 2250. 58. Pauwels R, Andries K, Desmyter J, Schols D, Kukla M-J, Breslin HJ, Raeymaeckers A, Van Gelder J, Woestenborghs R, Heykants J, Schellekens H, Janssen MAC, De Clercq E, Janssen PAJ. Potent and selective inhibition of HIV-1 replication in vitro by a novel series of TIBO derivatives. Nature 1990; 343:470 – 474. 59. Debyser Z, Pauwels R, Andries K, Desmyter J, Kukla M, Janssen PAJ, De Clercq E. An antiviral target on reverse transcriptase of human immunodeﬁciency virus type 1 revealed by tetrahydroimidazo-[4,5,1-jk][1,4]benzodiazepin-2(1H)-one and -thione derivatives. Proc Natl Acad Sci USA 1991; 88:1451 – 1455. 60. Baba M, Tanaka H, De Clercq E, Pauwels R, Balzarini J, Schols D, Nakashima H, Perno C-F, Walker RT, Miyasaka T. Highly speciﬁc inhibition of human immunodeﬁciency virus type 1 by a novel 6-substituted acyclouridine derivative. Biochem Biophys Res Commun 1989; 165:1375 – 1381. 61. Miyasaka T, Tanaka H, Baba M, Hayakawa H, Walker RT, Balzarini J, De Clercq E. A novel lead for speciﬁc anti-HIV agents: 1-(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine. J Med Chem 1989; 32:2507 – 2509. 62. Baba M, De Clercq E, Tanaka H, Ubasawa M, Takashima H, Sekiya K, Nitta I, Umezu K, Walker RT, Mori S, Ito M, Shigeta S, Miyasaka T. Potent and selective inhibition of human immunodeﬁciency virus type 1 (HIV-1) by 5ethyl-6-phenylthiouracil derivatives through their interaction with the HIV-1 reverse transcriptase. Proc Natl Acad Sci USA 1991; 88:2356 – 2360. 63. Baba M, De Clercq E, Tanaka H, Ubasawa M, Takashima H, Sekiya K, Nitta I, Umezu K, Walker RT, Mori S, Ito M, Shigeta S, Miyasaka T.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

64.

65.

66.

67.

68.

69.

70.

71.

72.

Highly potent and selective inhibition of human immunodeﬁciency virus type 1 by a novel series of 6-substituted acyclouridine derivatives. Mol Pharmacol 1991; 39:805 – 810. De Clercq E. HIV-1-speciﬁc RT inhibitors: highly selective inhibitors of human immunodeﬁciency virus type 1 that are speciﬁcally targeted at the viral reverse transcriptase. Med Res Rev 1993; 13:229—258. Balzarini J, Pe´rez-Pe´rez M-J, San-Fe´lix A, Schols D, Perno C-F, Vandamme A-M, Camarasa M-J, De Clercq E. 2V,5V-Bis-O-(tert-Butyldimethylsilyl)-3Vspiro-5VV-(4VV-amino-1VV,2VV-oxathiole-2VV,2VV-dioxide)pyrimidine (TSAO) nucleoside analogues: highly selective inhibitors of human immunodeﬁciency virus type 1 that are targeted at the viral reverse transcriptase. Proc Natl Acad Sci USA 1992; 89:4392 – 4396. Balzarini J, Pe´rez-Pe´rez M-J, San-Fe´lix A, Velazquez S, Camarasa M-J, De Clercq E. [2V,5V-Bis-O-(tert-Butyldimethylsilyl)]-3V-spiro-5VV- [4VV-amino-1VV,2VVoxathiole-2VV,2VV-dioxide) (TSAO) derivatives of purine and pyrimidine nucleosides as potent and selective inhibitors of human immunodeﬁciency virus type 1. Antimicrob Agents Chemother 1992; 36:1073 – 1080. Pauwels R, Andries K, Debyser Z, Van Daele P, Schols D, Stoﬀels P, De Vreese K, Woestenborghs R, Vandamme A-M, Janssen CGM, Anne´ J, Cauwenbergh G, Desmyter J, Heykants J, Janssen MAC, De Clercq E, Janssen PAJ. Potent and highly selective human immunodeﬁciency virus type 1 (HIV-1) inhibition by a series of a-anilinophenylacetamide derivatives targeted at HIV-1 reverse transcriptase. Proc Natl Acad Sci USA 1993; 90:1711 – 1715. Kleim J-P, Bender R, Billhardt U-M, Meichsner C, Riess G, Ro¨sner M, Winkler I, Paessens A. Activity of a novel quinoxaline derivative against human immunodeﬁciency virus type 1 reverse transcriptase and viral replication. Antimicrob Agents Chemother 1993; 37:1659 – 1664. Williams TM, Ciccarone TM, MacTough SC, Rooney CS, Balani SK, Condra JH, Emini EA, Goldman ME, Greenlee WJ, Kauﬀman LR, O’Brien JA, Sardana VV, Schleif WA, Theoharides AD, Anderson PS. 5-Chloro-3-(phenylsulfonyl)indole-2-carboxamide: a novel, non-nucleoside inhibitor of HIV-1 reverse transcriptase. J Med Chem 1993; 36:1291 – 1294. Mertens A, Zilch H, Ko¨nig B, Scha¨fer W, Poll T, Kampe W, Seidel H, Leser U, Leinert H. Selective non-nucleoside HIV-1 reverse transcriptase inhibitors. New 2,3-dihydrothiazolo[2,3-a]isoindol-5(9bH)-ones and related compounds with anti-HIV-1 activity. J Med Chem 1993; 36:2526 – 2535. Kashman Y, Gustafson KR, Fuller RW, Cardellina JH, McMahon JB, Currens MJ, Buckheit RW Jr, Hughes SH, Cragg GM, Boyd MR. The calanolides, a novel HIV-inhibitory class of coumarin derivatives from the tropical rainforest tree, Calophyllum lanigerum. J Med Chem 1992; 35:2735 – 2743. Patil AD, Freyer AJ, Eggleston DS, Haltiwanger RC, Bean MF, Taylor PB,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

Caranfa MJ, Breen AL, Bartus HR, Johnson RK, Hertzberg RP, Westley JW. The inophyllums, novel inhibitors of HIV-1 reverse transcriptase isolated from the Malaysian tree, Calophyllum inophyllum Linn. J Med Chem 1993; 36:4131 – 4138. Jacobo-Molina A, Ding J, Nanni RG, Clark AD Jr, Lu X, Tantillo C, Williams RL, Kamer G, Ferris AL, Clark P, Hizi A, Hughes SH, Arnold E. Crystal structure of human immunodeﬁciency virus type 1 reverse transcriptase complexed with double-stranded DNA at 3.0 A˚ resolution shows bent DNA. Proc Natl Acad Sci USA 1993; 90:6320 – 6324. Nanni RG, Ding J, Jacobo-Molina A, Hughes SH, Arnold E. Review of HIV-1 reverse transcriptase three-dimensional structure. Implications for drug design. Perspect Drug Dis Des 1993; 1:129 – 150. Pialoux G, Youle M, Dupont B, Gazzard B, Cauwenbergh GFMJ, Stoﬀels PAM, Davies S, De Saint Martin J, Janssen PAJ. Pharmacokinetics of R82913 in patients with AIDS or AIDS-related complex. Lancet 1991; 338:140 – 143. Saag MS, Emini EA, Laskin OL, Douglas J, Lapidus WI, Schleif WA, Whitley RJ, Hildebrand C, Byrnes VW, Kappes JC, Anderson KW, Massari FE, Shaw GM, the L-697,661 Working Group. A short-term clinical evaluation of L-697,661, a non-nucleoside inhibitor of HIV-1 reverse transcriptase. N Engl J Med 1993; 329:1065 – 1072. Vandamme A-M, Debyser Z, Pauwels R, De Vreese K, Goubau P, Youle M, Gazzard B, Stoﬀels PA, Cauwenbergh GF, Anne´ J, Andries K, Janssen PAJ, Desmyter J, De Clercq E. Characterization of HIV-1 strains isolated from patients treated with TIBO R82913. AIDS Res Hum Retrovir 1994; 10:39 – 46. Nunberg JH, Schleif WA, Boots EJ, O’Brien JA, Quintero JC, Hoﬀman JM, Emini EA, Goldman ME. Viral resistance to human immunodeﬁciency virus type 1-speciﬁc pyridinone reverse transcriptase inhibitors. J Virol 1991; 65:4887 – 4892. Richman D, Shih C-K, Lowy I, Rose J, Prodanovich P, Goﬀ S, Griﬃn J. Human immunodeﬁciency virus type 1 mutants resistant to nonnucleoside inhibitors of reverse transcriptase arise in tissue culture. Proc Natl Acad Sci USA 1991; 88:11241 – 11245. Mellors JW, Dutschman GE, Im G-J, Tramontano E, Winkler SR, Cheng Y-C. In vitro selection and molecular characterization of human immunodeﬁciency virus-1 resistant to non-nucleoside inhibitors of reverse transcriptase. Mol Pharmacol 1992; 41:446 – 451. De Vreese K, Debyser Z, Vandamme A-M, Pauwels R, Desmyter J, De Clercq E, Anne´ J. Resistance of human immunodeﬁciency virus type 1 reverse transcriptase to TIBO derivatives induced by site-directed mutagenesis. Virology 1992; 188:900 – 904. Sardana VV, Emini EA, Gotlib L, Graham DJ, Lineberger DW, Long WJ,

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

83.

84.

85.

86.

87.

88.

89.

90.

Schlabach AJ, Wolfgang JA, Condra JH. Functional analysis of HIV-1 reverse transcriptase amino acids involved in resistance to multiple nonnucleoside inhibitors. J Biol Chem 1992; 267:17526 – 17530. Balzarini J, Karlsson A, De Clercq E. Human immunodeﬁciency virus type1 drug-resistance patterns with diﬀerent 1-[(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine derivatives. Mol Pharmacol 1993; 44:694 – 701. Byrnes VW, Sardana VV, Schleif WA, Condra JH, Waterbury JA, Wolfgang JA, Long WJ, Schneider CL, Schlabach AJ, Wolanski BS, Graham DJ, Gotlib L, Rhodes A, Titus DL, Roth E, Blahy OM, Quintero JC, Staszewski S, Emini EA. Comprehensive mutant enzyme and viral variant assessment of human immunodeﬁciency virus type 1 reverse transcriptase resistance to nonnucleoside inhibitors. Antimicrob Agents Chemother 1993; 37:1576 – 1579. Balzarini J, Karlsson A, Pe´rez-Pe´rez M-J, Vrang L, Walbers J, Zhang H, O¨berg B, Vandamme A-M, Camarasa M-J, De Clercq E. HIV-1-speciﬁc reverse transcriptase inhibitors show diﬀerential activity against HIV-1 mutant strains containing diﬀerent amino acid substitutions in the reverse transcriptase. Virology 1993; 192:246 – 253. Mellors JW, Im G-J, Tramontano E, Winkler SR, Medina DJ, Dutschman GE, Bazmi HZ, Piras G, Gonzalez CJ, Cheng Y-C. A single conservative amino acid substitution in the reverse transcriptase of human immunodeﬁciency virus-1 confers resistance to (+)- (5S)-4,5,6,7-tetrahydro-5-methyl6-(3-methyl-2-butenyl)imidazo[4,5,1-jk][1,4]benzodiazepin-2(1H)-thione (TIBO R82150). Mol Pharmacol 1993; 43:11 – 16. Balzarini J, Karlsson A, Pe´rez-Pe´rez M-J, Camarasa M-J, Tarpley WG, De Clercq E. Treatment of human immunodeﬁciency virus type 1 (HIV-1)-infected cells with combinations of HIV-1-speciﬁc inhibitors results in a diﬀerent resistance pattern than does treatment with single-drug therapy. J Virol 1993; 67:5353 – 5359. Balzarini J, Velazquez S, San-Fe´lix A, Karlsson A, Pe´rez-Pe´rez M-J, Camarasa M-J, De Clercq E. Human immunodeﬁciency virus type 1-speciﬁc [2V,5V-bis-O-(tert-butyldimethylsilyl)-h-D-ribofuranosyl]-3V-spiro-5VV-(4VV-amino-1VV,2VV- oxathiole-2VV,2VV-dioxide)-purine analogues show a resistance spectrum that is diﬀerent from that of the human immunodeﬁciency virus type 1speciﬁc non-nucleoside analogues. Mol Pharmacol 1993; 43:109 – 114. Balzarini J, Karlsson A, Vandamme A-M, Pe´rez-Pe´rez M-J, Zhang H, Vrang L, O¨berg B, Ba¨ckbro K, Unge T, San-Fe´lix A, Velazquez S, Camarasa M-J, De Clercq E. Human immunodeﬁciency virus type 1 (HIV-1) strains selected for resistance against the HIV-1-speciﬁc [2V,5V-bis-O-(tert-butyldimethylsilyl)-3V-spiro-5VV-(4VV-amino-1VV,2VV-oxathiole-2VV,2VV-dioxide)]-h-D-pentofuranosyl (TSAO) nucleoside analogues retain sensitivity to HIV-1-speciﬁc nonnucleoside inhibitors. Proc Natl Acad Sci USA 1993; 90:6952 – 6956. Kleim J-P, Bender R, Kirsch R, Meichsner C, Paessens A, Riess G. Muta-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.