BURGER'S MEDICINAL CHEMISTRY AND DRUG DISCOVERY

BURGER'S MEDICIN AL CHEMISTRY AND DRUG DISCOVERY
Sixth Edition Volume 2: Drug Discovery and Drug Development
Edited by

Donald J. Abraham
Department of Medicinal Chemistry School of Pharmacy Virginia Commonwealth University Richmond, Virginia

Burger's Medicinal Chemistry and Drug Discovery is available Online in full color at www.mrw.interscience.wiley.com/bmcdd.

A Wiley-Interscience Publication

john Wiley and Sons, Inc.

BURGER MEMORIAL EDITION
The Sixth Edition of Burger's Medicinal Chemistry and Drug Discovery is being designated as a Memorial Edition. Professor Alfred Burger was born in Vienna, Austria on September 6, 1905 and died on December 30, 2000. Dr. Burger received his Ph.D. from the University of Vienna in 1928 and joined the Drug Addiction Laboratory in the Department of Chemistry at the University of Virginia in 1929. During his early years at W A , he synthesized fragments of the morphine molecule in an attempt to find the analgesic pharmacophore. He joined the W A chemistry faculty in 1938 and served the department until his retirement in 1970. The chemistry department at W A became the major academic training ground for medicinal chemists because of Professor Burger. Dr. Burger's research focused on analgesics, antidepressants, and chemotherapeutic agents. He is one of the few academicians to have a drug, designed and synthesized in his laboratories, brought to market [Parnate, which is the brand name for tranylcypromine, a monoamine oxidase (MAO) inhibitor]. Dr. Burger was a visiting Professor at the University of Hawaii and lectured throughout the world. He founded the Journal of Medicinal Chemistry, Medicinal Chemistry Research, and published the first major reference work "Medicinal Chemistry" in two volumes in 1951. His last published work, a book, was written at age 90 (Understanding Medications: What the Label Doesn't Tell You, June 1995). Dr. Burger received the Louis Pasteur Medal of the Pasteur Institute and the American Chemical Society Smissman Award. Dr.' Burger played the violin and loved classical music. He was married for 65 years to Frances Page Burger, a genteel Virginia lady who always had a smile and an open house for the Professor's graduate students and postdoctoral fellows.

vii

PREFACE
The Editors, Editorial Board Members, and John Wiley and Sons have worked for three and a half years to update the fifth edition of Burger's Medicinal Chemistry and Drug Discovery. The sixth edition has several new and unique features. For the first time, there will be an online version of this major reference work. The online version will permit updating and easy access. For the first time, all volumes are structured entirely according to content and published simultaneously. Our intention was to provide a spectrum of fields that would provide new or experienced medicinal chemists, biologists, pharmacologists and molecular biologists entry to their subjects of interest as well as provide a current and global perspective of drug design, and drug development. Our hope was to make this edition of Burger the most comprehensive and useful published to date. To accomplish this goal, we expanded the content from 69 chapters (5 volumes) by approximately 50% (to over 100 chapters in 6 volumes). We are greatly in debt to the authors and editorial board members participating in this revision of the major reference work in our field. Several new subject areas have emerged since the fifth edition appeared. Proteomics, genomics, bioinformatics, combinatorial chemistry, high-throughput screening, blood substitutes, allosteric effectors as potential drugs, COX inhibitors, the statins, and high-throughput pharmacology are only a few. In addition to the new areas, we have filled in gaps in the fifth edition by including topics that were not covered. In the sixth edition, we devote an entire subsection of Volume 4 to cancer research; we have also reviewed the major published Medicinal Chemistry and Pharmacology texts to ensure that we did not omit any major therapeutic classes of drugs. An editorial board was constituted for the first time to also review and suggest topics for inclusion. Their help was greatly appreciated. The newest innovation in this series will be the publication of an academic, "textbook-like" version titled, "Burger's Fundamentals of Medicinal Chemistry." The academic text is to be published about a year after this reference work appears. It will also appear with soft cover. Appropriate and key information will be extracted from the major reference. There are numerous colleagues, friends, and associates to thank for their assistance. First and foremost is Assistant Editor Dr. John Andrako, Professor emeritus, Virginia Commonwealth University, School of Pharmacy. John and I met almost every Tuesday for over three years to map out and execute the game plan for the sixth edition. His contribution to the sixth edition cannot be understated. Ms. Susanne Steitz, Editorial Program Coordinator at Wiley, tirelessly and meticulously kept us on schedule. Her contribution was also key in helping encourage authors to return manuscripts and revisions so we could publish the entire set at once. I would also like to especially thank colleagues who attended the QSAR Gordon Conference in 1999 for very helpful suggestions, especially Roy Vaz, John Mason, Yvonne Martin, John Block, and Hugo

x

Preface

Kubinyi. The editors are greatly indebted to Professor Peter Ruenitz for preparing a template chapter as a guide for all authors. My secretary, Michelle Craighead, deserves special thanks for helping contact authors and reading the several thousand e-mails generated during the project. I also thank the computer center at Virginia Commonwealth University for suspending rules on storage and e-mail so that we might safely store all the versions of the author's manuscripts where they could be backed up daily. Last and not least, I want to thank each and every author, some of whom tackled two chapters. Their contributions have provided our field with a sound foundation of information to build for the future. We thank the many reviewers of manuscripts whose critiques have greatly enhanced the presentation and content for the sixth edition. Special thanks to Professors Richard Glennon, William Soine, Richard Westkaemper, Umesh Desai, Glen Kellogg, Brad Windle, Lemont Kier, Malgorzata

Dukat, Martin Safo, Jason Rife, Kevin Reynolds, and John Andrako in our Department of Medicinal Chemistry, School of Pharmacy, Virginia Commonwealth University for suggestions and special assistance in reviewing manuscripts and text. Graduate student Derek Cashman took able charge of our web site, http:llwww.burgersmedchem.com, another first for this reference work. I would especially like to thank my dean, Victor Yanchick, and Virginia Commonwealth University for their support and encouragement. Finally, I thank my wife Nancy who understood the magnitude of this project and provided insight on how to set up our home office as well as provide John Andrako and me lunchtime menus where we often dreamed of getting chapters completed in all areas we selected. To everyone involved, many, many thanks. DONALD J. ABRAHAM Midlothian, Virginia

CONTENTS
1 COMBINATORIAL CHEMISTRY AND MULTIPLE PARALLEL SYNTHESIS, 1
4 APPLICATION OF

Lester A. Mitscher Apurba Dutta Kansas University Department of Medicinal Chemistry Lawrence, Kansas
2 HIGH-THROUGHPUT

RECOMBINANT DNA TECHNOLOGY IN MEDICINAL CHEMISTRY AND DRUG DISCOVERY, 81 Soumitra Basu University of Pittsburgh Center for Pharmacogenetics Department of Pharmaceutical Sciences Pittsburgh, Pennsylvania Adegboyega K. Oyelere Rib-X Pharmaceuticals, Inc. New Haven, Connecticut
5 OLIGONUCLEOTIDE THERAPEUTICS, 115

SCREENING FOR LEAD DISCOVERY, 37 John G. Houston Martyn N. Banks Pharmaceutical Research Institute Bristol-Myers Squibb Wallingford, Connecticut
3 HIGH-THROUGHPUT PHARMACOLOGY, 71

Stanley T. Crooke Isis Pharmaceuticals, Inc. Carlsbad, California
6 THERAPEUTIC AGENTS ACTING ON RNA TARGETS, 167

Steven J. Brown Imran B. Clark Bala Pandi Axiom Biotechnologies, Inc. San Diego, California

Jason P. Rife Virginia Commonwealth University Richmond, Virginia

xiii

xiv

Contents

7 CARBOHYDRATE-BASED THERAPEUTICS, 203 John H. Musser Pharmagenesis, Inc. Palo Alto, California

10 RECEPTOR TARGETS IN DRUG DISCOVERY AND DEVELOPMENT, 319

8 MEMBRANE TRANSPORT PROTEINS AND DRUG TRANSPORT, 249 Peter W. Swam Division of Pharmaceutics Ohio State Biophysics Program OSU Heart & Lung Institute Core Laboratory for Bioinformatics and Computational Biology The Ohio State University Columbus, Ohio
9 ALLOSTERIC PROTEINS AND DRUG DISCOVERY, 295 J. Ellis Bell Department of Chemistry Gottwald Science Center University of Richmond Richmond, Virginia James C. Burnett School of Pharmacy and Department of Medicinal Chemistry Institute for Structural Biology and Drug Discovery Virginia Commonwealth University Richmond, Virginia Jessica K. Bell Department of Pharmaceutical Chemistry University of California at San Francisco San Francisco, California Peter S. Galatin Donald J. Abraham School of Pharmacy and Department of Medicinal Chemistry Institute for Structural Biology and Drug Discovery Virginia Commonwealth University Richmond, Virginia

Michael Williams Department of Molecular Pharmacology and Biological Chemistry Northwestern University Medical School Chicago, Illinois Christopher Mehlin Molecumetics Inc. Bellevue, Washington David J. Triggle State University of New York Buffalo, New York
11 NICOTINIC ACETYLCHOLINE RECEPTORS, 357

David Colquhoun Chris Shelley Chris Hatton Department of Pharmacology University College London London, United Kingdom Nigel Unwin Neurobiology Division MRC Laboratory of Molecular Biology Cambridge, United Kingdom Lucia Sivilotti Department of Pharmacology The School of Pharmacy London, United Kingdom
12 LARGE-SCALE SYNTHESIS, 407

Frank Gupton Boehringer Ingelheim Chemicals, Inc. Ridgefield, Connecticut
Karl Grozinger Boehringer Ingelheim Pharmaceuticals, Inc. Ridgefield, Connecticut

13 PRINCIPLES OF DRUG METABOLISM, 431

17 DRUG ABSORPTION,

Bernard Testa School of Pharmacy University of Lausanne Institute of Medicinal Chemistry Lausanne, Switzerland William Soine Virginia Commonwealth University Department of Medicinal Chemistry Richmond, Virginia
14 METABOLIC CONSIDERATIONS IN PRODRUG DESIGN, 499

DISTRIBUTION, AND ELIMINATION, 633 Leslie Z. Benet Beatrice Y. T. Perotti University of California Department of Pharmacy San Francisco, California Larry Hardy Aurigene Discovery Technologies Lexington, Massachusetts
18 PHYSICOCHEMICAL

Luc P. Balant University Hospitals of Geneva Clinical Research Unit Department of Psychiatry Geneva, Switzerland
15 RETROMETABOLISM-BASED

CHARACTERIZATION AND PRINCIPLES OF ORAL DOSAGE FORM SELECTION, 649 Gregory E. Amidon Xiaorong He Michael J. Hageman Pharmacia Corporation Kalamazoo, Michigan
19 THE FDA AND REGULATORY ISSUES, 683

DRUG DESIGN AND TARGETING, 533 Nicholas Bodor Center for Drug Discovery University of Florida Gainesville, Florida and WAX Research, Inc. Miami, Florida Peter Buchwald WAX Research, Inc. Miami, Florida
16 DRUG DISCOVERY: THE ROLE OF TOXICOLOGY, 609

W. Janusz Rzeszotarski Food and Drug Administration Rockville, Maryland
20 INTELLECTUAL PROPERTY IN

DRUG DISCOVERY AND BIOTECHNOLOGY, 703 Richard A. Kaba Timothy P. Maloney James P. Krueger Rudy Kratz Julius Tabin Fitch, Even, Tabin & Flannery Chicago, Illinois INDEX, 783

James B. Moe James M. McKim, Jr. Pharmacia Corporation Kalamazoo, Michigan

BURGER'S MEDICINAL CHEMISTRY AND DRUG DISCOVERY

CHAPTER ONE

Combinatorial Chemistry and Multiple Parallel Synthesis
LESTER A. MITSCHER APURBA DUTTA
Kansas University Department of Medicinal Chemistry Lawrence, Kansas

Contents
1 Introduction, 2 2 History, 4 3 Solid Phase Organic Synthesis of Informational Macromolecules of Interest to Medicinal Chemists, 5 3.1 Peptide Arrays, 6 3.2 Nucleoside Arrays, 13 3.3 Oligosaccharide Arrays, 13 3.4 Lipid Arrays, 13 4 Solid and Solution Phase Libraries of Small, Drugable Molecules, 14 4.1 Purification, 23 4.2 Synthetic Success and Product Purity, 28 4.3 Resins and Solid Supports, 29 4.4 Microwave Accelerations, 30 4.5 Analytical Considerations, 30 4.6 Informetrics, 30 4.7 Patents, 30 5 Summary and Conclusions, 31

Burger's Medicinal Chemistry and Drug Discovery Sixth Edition, Volume 2: Drug Development Edited by Donald J. Abraham ISBN 0-471-37028-2 O 2003 John Wiley & Sons, Inc.
1

Combinatorial Chemistry and Multiple Parallel Synthesis

1

INTRODUCTION

The introduction of a new pharmaceutical is a lengthy and expensive undertaking. Methods which promise to shorten the time or the cost are eagerly taken up, and this is clearly the case with combinatorial chemistry and multiple parallel synthesis. Combinatorial chemistry is somewhat hard to define precisely, but generally speaking, it is a collection of methods that allow the simultaneous chemical synthesis of large numbers of compounds using a variety of starting materials. The resulting compound l i b r m can contain all of the possible chemical structures that can be produced in this manner. Multiple parallel synthesis is a related group of methodologies used to prepare a selected smaller subset of the molecules that could in theory have been prepared. The content of libraries prepared by multiple parallel synthesis is more focused and less diverse than those constructed with combinatorial technology. The primary benefit that combinatorial and multiple parallel chemistry bring to drug synthesis is speed. As with most other human endeavors, uncontrolled speed may be exhilarating but is not particularly useful. Rapid construction of compounds that have no chance of becoming drugs is of little value to the medicinal chemist. After an initial euphoric period when many investigators thought that any novel compound had a realistic chance of becoming a drug, realism has now returned, and libraries are being constructed that reflect the accumulated wisdom of the field of medicinal chemistry. Combinatorial methods have permeated and irreversibly altered most phases of drug seeking that benefits from the attention of chemists. The successful contemporary medicinal chemist must be aware of the strengths and weaknesses of these exciting new methods and be able to apply them cunningly. In the proper hands combining medicinal chemical insight with enhanced speed of synthesis is very powerful. Libraries are constructed both in solution and on solid supports and the choice between

C

AC

BC*

CC

DC

EC

Figure 1.1. A combinatorial library constructed from five reacting components.

these techniques is often a matter of personal preference, and they are performed side by side in most laboratories. For very large libraries, however, construction on resins is more practical, whereas for smaller, focused libraries, solution phase chemistry is more practical. Solid phase methods are also specially advantageous for multistep iterative processes and are notable for the comparative ease of purification by simple filtration and the ability to drive reactions to completion by the use of excess reagents. Throughout the previous decade, solid phase organic synthesis (SPOS) has dominated combinatorial chemistry, and many novel methods have been developed as a result. The main concepts in this field can be summarized in Figs. 1.1and 1.2. In Fig. 1.1, there is a hypothetical combinatorial compound library of condensation products produced by reacting every possible combination of five starting materials. This results in a library containing 25 (5 x 5) products. The library could be constructed by 25 individual reactions, with each product separate from all of the others. It could also be constructed by run-

C

AC

BC*

CC

DC

EC

Figure 1.2. A multiple parallel synthesis library
constructed from six participants.

1 Introduction

ning reactions simultaneously so that a single mixture of all 25 substances would be obtained. In this specific example, it is presumed that the reactions were run in a single step so that all compounds were produced simultaneously in a big mixture. Furthermore, it is assumed that the ideal was achieved. That is, all reactions proceeded quantitatively, and that each compound is present in the final compound collection in equal molar concentration. (This ideal is rarely achieved.) It is also assumed that BC is the only active constituent and so it is marked off with an asterisk. If the whole library were tested as a mixture, then it would be seen that it contained an active component, but one would not know which one it was. For this to happen, it is necessary that the components do not interfere with one another so false positives and false negatives are not seen. Many clever means of finding the active component expeditiously have been developed and a number of these will be illustrated later. It is clear that if the components were prepared and tested individually, 25 separate reactions would be required and the identification problem would disappear, but great strain would be placed on the bioassay and the speed of synthesis would be compromised. A perfect combinatorial library for drug discovery would only contain BC, and only a single reaction would be required, but this level of efficiency is rarely achieved by any contemporary medicinal chemical method. The real problem is to construct a compound library that is sufficiently diverse and sufficiently large that there is a high possibility that at least one component will be active in the chosen test system. This example assumes that this has been done. The reader will also readily see that with the library in Fig. 1.1, that instead of testing all of the compounds simultaneously, if one prepared and tested 10 mixtures resulting from combining the five products in each column and each row, then BC would reliably emerge as the active component because only the row C mixture and the column B mixture would be active, and the active component must be the one where the rows and columns intersect. When more than one active component is present then the problem becomes more complex.

Figure 1.2 illustrates a related multiple parallel library. This is much smaller and starts with the assumption that one knows or suspects that the best compound will terminate in component C. Five reactants are chosen to condense with C and the resultant library consists of five components. Each of these products is tested singly and BC* is quickly identified. The greater efficiency of this library compared with that of Fig. 1.1 for the purpose is obvious. Clearly, the smaller the library that succeeds in solving the problem, the more effective the process is. A perfectly efficient library would only contain BC*. The size of the library produced by either method is the product of the number of variables introduced and the number of steps involved raised exponentially. For example, starting with a given starting material (often called a centroid) and attaching four different groups of 10 sidechains to each product would produce 1 X 10 x 10 x 10 x 10 or 10,000 members. The primary advantage of either method is speed, because the products are prepared simultaneously at each step. The efficiency is also enhanced when the condensation steps involve the same conditions. The use of the library depends on the specific structures included and the purposes for which the librarv " is to be tested. The relevance of speed to drug discovery is easy to explicate. If one knew in advance the particular structure that would satisfy the perceived need, a successful compound library would only need to have one substance in it. If one has a general idea of the type of structure that would be useful, the library will have many promising compounds but still be finite in number. Quantitative bioassay of pure substances would allow one to select the most nearly perfect embodiment. If one has no idea of the type of structure that would give satisfaction, then a successful library must have a larger and more diverse number of compounds in it. Contemporary drug seeking is a complex, time consuming, and expensive process because a successful drug must not only have outstanding potency and selectivity, but it must also satisfy an increasingly long list of other structure-dependent criteria as well. The elapsed time from initial synthesis to

4

Combinatorial Chemistry and Multiple Parallel Synthesis

marketing is estimated to lie on average between 10 and 15 years and to require the preparation and evaluation of a few thousand analogs. The costs are estimated to lie between $300 and $800 million per agent. Most large firms now target the introduction of 1-3 novel drugs per year and target sales at a billion dollars or more from each. This indicates that each day of delay in the drug seeking process not only deprives patients of the putative benefits of the new drug but also represents the loss of a million dollars or more of sales for the firm! Added to this is the intense competition among big pharmaceutical companies for a winning place in the race and among small pharmaceutical companies for survival. First to market in an unserved therapeutic area can return a great profit if a sufficient number of sufferers exist who have access to the funds to pay for their treatment. The next two entries competingwith this agent can also be expected to do well. After this, success is rather more problematic because the market grows more and more fragmented. Being first to finish the race, therefore, conveys very real survival value. From an economic standpoint, it is estimated that less than 10% of products introduced repay their development costs. Those few that do must return a sufficient surplus to amortize the costs of the rest and sufficient additional funds to cover the costs of future projects and to gratify the shareholders. These imperatives have placed a premium on speed of discovery and development. The portion of this time devoted to synthesis and screening in the drug-seeking campaign is usually about 3-5 years. The enhanced speed of construction can be expected to decrease the time to market by perhaps as much as 1 year in favorable cases. While this is less than was originally hoped for when these methods were introduced, it is not trivial. Combinatorial chemistry is now such a pervasive phenomenon that comprehensive review of its medicinal chemical features is no longer possible in less than book length. Full coverage would require treatment of its impact on all of the phases of drug discovery and would exceed the space available. Thus, the remainder of this chapter will illustrate its main features and applications.

2

HISTORY

Combinatorial chemistry grew out of peptide chemistry and initially served the needs of biochemists and the subset of medicinal chemists who specialized in peptide science. Its first decade or so concentrated on oligopeptides and related molecules. It continued to evolve, however, and now permeates virtually every corner of medicinal chemistry and a major effort is underway to discover new, orally active, pharmaceuticals using these methods. Many will agree that the path leading to the present state of combinatorial chemistry essentially started with the solid phase synthetic experiments on peptides by Bruce Merrifield in 1962 (1,2). This work had immediate impact, facilitated in large part because of the essentially iterative reactions, to completion by use of reagents in excess, its susceptibility to automation, and the ease of removing detritus from the products by simple washing and filtration away from the resins. At first this extremely useful technology was employed in a linear fashion. It was probably Furka in Hungary a decade or so later who realized that the methodology could lead to simultaneous synthesis of large collections of peptides and conceived of the mix and split methods (3). Geyson made the whole process technically simpler in 1984 and produced large scale compound collections of peptides (4) and Houghton introduced "tea bag" methodologies in 1985 in which porous bags of resins were suspended in reagents (5). Comparatively few organic chemists undertook the preparation of ordinary organic substances on solid phases because the work is rather more complex when applied to non-oligomeric substances caused by greater variety of reactants and conditions required, and this work at first failed to develop a significant following. Solid phase organic chemistry was also comparatively underdeveloped and this held back the field. This changed in dramatic fashion after the publication of Bunin and Ellman's seminal work on solid phase organic synthesis (SPOS)of arrays of 1,4-benzodiazepine-2-ones in 1992 (6).Soon other laboratories published related work on this ring system, and work on other drug-like molecules followed in rapid order and the race was on. In the initial phases, solid phase or-

3 Solid Phase Organic Synthesis of Informational Macromolecules of Interest to Medicinal Chemists

5

ganic synthesis predominated, and this persisted until about 1995, when solution phase combinatorial chemistry began to make serious inroads. Until about 1997, roughly onehalf of the libraries reported were either of peptides or peptidomimetics. Subsequently libraries of drug-like small molecules have become increasingly popular. The work on combinatorial libraries has inspired the rapid development of a wide variety of auxiliary techniques including the use of reagents on solid support, capture resins, chemical and biological analysis of compound tethered to resins, informatics to deal with the huge volume of structural and biological data generated, the synthesis of a wide variety of peptide-like and heterocyclic systems hitherto prepared solely in solution, photolithographic techniques allowing the production of geographically addressed arrays on a "credit card," preparation of gene array chips, attachment of coding sequences, use of robotics, and the preparation of oligonucleotides by Letsinger in 1975 (7) and of oligosaccharides by Hindsgaul in the 1990s (8). At this moment several thousand papers are appearing each year describing the preparation and properties of compound libraries either in mixtures or as individual substances. Several books (935) and reviews (36-49) are available for the interested reader. Those of Dolle are particularly recommended because he has undertaken the heroic task of organizing and summarizing each year the world's literature on the topic. That of Thompson and Ellman is especially thorough in reviewing the literature up until 1996 from a chemical viewpoint. A great many other reviews are available, including many in slick-cover free journals that arrive on our desks weekly. In addition to these, at least three specialist journals have been established in the area. These are the Journal of Combinatorial Chemistry, Molecular Diversity, and Combinatorial Chemistry and High Throughput Screening. Another important factor leading to the explosion of interest in combinatorial chemical techniques was the development of small firms devoted to the exploitation of genetic discoveries through development of highthroughput screening methods. These firms by and large did not have libraries of com-

pounds to put through these screens and were seeking collections of molecules. Combinatorial chemistry addressed these needs. When these methods were taken up by big pharmaceutical companies, existing libraries quickly proved inadequate for the need and combinatorial methodologies clearly addressed this need as well. Just about 10 years after these seminal events, the face of medicinal chemistry has been irretrievably altered. While combinatorial chemistry has in some respects not lived up to the initial hopes, its value is firmly established and no serious firm today fails to use these methods. By the year 2002, well over 1000 libraries have been reported. Many of these include reports of the biological activity of their contents. This is remarkable considering that the field is scarcely more than a decade old!
3 SOLID PHASE ORGANIC SYNTHESIS OF INFORMATIONAL MACROMOLECULES OF INTEREST TO MEDICINAL CHEMISTS

That the solid phase synthesis of collections of peptides launched this field is not intrinsically surprising. The basic methodology existed because of Merrifield and many others. The peptide linkage has notable advantages for this work because it is relatively chemically stable; non-chiral, constructible by iterative processes amenable to automation, the products are rarely branched, possess a variety of interesting biological properties, and can be constructed in great variety. The counterbalancing defects of these compounds are that they are not easily delivered orally unless they are end capped and rather small in molecular weight, are readily destroyed by enzymatic action, and fail to penetrate into cells. The physiological reason for this is readily understood. Peptides, and other informational macromolecules, function in the body to provide specific structure or to generate signals for cells to respond to according to their sequence and architecture. It would be dangerous if they were absorbed intact from ingestion of other life forms. To prevent cellular disruption they are first digested in the gastrointestinal tract, absorbed as monomers, and then reassembled after our own genetic pattern so that they join

6

Combinatorial Chemistry and Multiple Parallel Synthesis

or supplement those already present without causing disruption in cellular architecture or function. Nucleic acids have many of the same advantages and disadvantages, and their compound libraries came into being soon after the peptides. Oligosaccharides, on the other hand, have lagged considerably behind. They possess chiral linkages whose construction must be carefully controlled, often are branched, are fragile in the presence of acid conditions, are often highly polar, and are readily digested. Controlling these processes is much more difficult and it has taken longer to conquer these problems.
3 . 1
Peptide Arrays

Peptides are prominent among the compounds of interest to biochemists but less so to most medicinal chemists for reasons explicated above. Construction of peptides of specified sequence is an iterative process whose complications largely consist of protectiondeprotection steps to ensure that side-chain functionality does not interfere with orderly amide bond formation. When made on resin beads, there are limits to the quantities that can be made in part because of the geometric restrictions caused by bead size and the need to avoid adjacent molecules from interacting with each other, and the comparative lability of the beads to aggressive reagents places limitations on the chemical conditions that can be employed. Porous beads obviously can be loaded more heavily than those whose surface only is accessible after wetting by the reagents. Many different kinds of resins and other solid supports are now available, including some that are solvent soluble depending on the solvent and the temperature. With their balance of advantages and disadvantages and the present gold standard being oral activity, peptide libraries are currently of primary value in lead seeking, in basic studies on cellular processes, or for the preparation of p a r e n t e d medications. Despite intensive study spanning several decades by some of the best minds of this generation, means of delivering thereputically significant blood levels of peptides through the oral route remain elusive. Translating the therapeutic message in peptide leads into oral non-peptide

drugs through generally applicable systematic techniques has also been elusive. The several successes that have been achieved have been primarily the result of screening campaigns or serendipitous observations, and the results have so far not revealed an underlying tactical commonality that can be exploited in new cases. Perhaps the best known of these studies has been the translation of snake venom peptides, whose injection by serpents leads to a precipitous fall in blood pressure, into truncated pseudodipeptides like captopril, and then on to enalaprilat and lysinopril, which are pseudotripeptides. The basic lesson learned from all of these studies has been that the resulting agent should be as little peptidelike as possible and not exceed the equivalent of at most four amino acid-like residues. Examination of peptide structures in light of the well-known Lipinski rules provides a rationale for what experience has shown. Beyond about four residues, the molecular weight is becoming too high, the polarity is weighted too much toward water solubility, and the hydrogenbonding inventory is excessive. Further, the compounds are excessively water soluble so that they do not pass through cellular membranes efficiently by passive diffusion. An added feature to bear in mind is that the preparation of certain medically important polypeptide drugs, such as human insulin and growth hormone, through genetic engineering methodologies, is well developed and convenient so these substances can be used in parenteral replacement therapy. Their preparation through synthetic peptide chemistry represents important achievements in peptide intellectual technology but does not satisfy a commercial need. Nonetheless, peptide compound libraries are very convenient for uncovering leads quickly for receptors where natural ligands or serendipitous drugs have not previously been found and large libraries of peptides continue to be made (50-65). The number of peptides that could in principle be made is stupefymg. For example, given that there are approximately 20 common amino acids, and allowing five post-translational modifications (and ignoring the fact that there are more of these and that there are many wholly synthetic amino acids), the avail-

developed a very diverse library on silicon wafers using photolitho- Geysen pins and wells Figure 1. With each iteration. The method requires photosensitive protecting groups and testing methodologies compatible with support-bound assay methods and the libraries are geographically coded by the position of products on the x~ axis (67). and Furka has estimated that the mass of such a library would exceed that of the universe by more than 200 orders of magnitude (3)! Were one to use just the common amino acids.1. In principle this method could produce thousands of individual peptides on a credit card-like surface. 1.1. Subsequently Fodor et al. so if one simply placed all of the potential reactants in a flask under bond forming conditions. By analogy with the number of languages that have been generated using this system.000 = 1. (Fig.) Photolytic protecting groups were employed followed by coupling the newly revealed "hot spots" with a suitable reactant. Direct methods of sequence analysis are available. they would not react at the same rate.380. After this.000 = 64. Mass spectrometry is popular as are NMR methods (involving magic angle methods on single beads). Although somewhat laborious. This method was introduced by Geysen in 1984.4.000. These methods are popular when iterative methods result in linear polymers. as is shown in Table 1. No decoding sequence needs to be attached to the beads in this kind of library. It would require an incredible effort to make a library containing even only one molecule of each. Geysen pins and wells.3 Solid Phase Organic Synthesis of Informational Macromolecules of Interest to Medicinal Chemists 7 Table 1. Edman degradation of peptides can also be performed. A convenient variation was developed for parallel synthesis in which beads were contained in porous bags and dipped into reagent solutions. These are called "tea bags. A further complication of simultaneous preparation of peptide mixtures is that individual amino acids differ greatly in their reactivity. 1. the potential number of peptides that could be made is clearly astronomical. Here a single peptide is constructed on a single type of resin. the reactions were run on resins attached to individual pins so constructed that they fit into individual wells of 96-well plates. the masks are moved as often as desired and the process repeated. Number of Possible Peptide Products as a Function of the Amino Acids Used Dipeptides Tripeptides Tetrapeptides Pentapeptides Hexapeptides Heptapeptides (20 (20 (20 (20 (20 (20 20) x 20 X 20) x 20 X 20 x 20) X 20 X 20 X 20 x 20) x 20 x 20 x 20 x 20 x 20) x 20 X 20 X 20 x 20 x 20 X 20) X 400 = 8000 = 160. the disparity between readily formed and poorly formed bonds would widen. To make 96 peptides at a time and to keep track of the products and facilitate their screening. where each peptide would be located through use of variously configured masks (Fig. the synthesis can readily be automated. the progression of peptides possible is enormous. These techniques are now widely employed for gene array amplification and identification experiments.000 = 3.000.3) (66). Synthesis of mixtures of peptides further enhanced the speed and convenience of library construction but required development of devolution methods so that active components in the mixtures could be identified.200. graphic chemistry for forming the peptides and controlled the specific place along an x-y axis.3." The identity of the peptide or peptides contained is recorded on the attached label.000 = able synthons are at least approximately equal to the number of letters in the Western alphabet. The simplest and least ambiguous method for constructing peptide libraries is the spatially separate or spatially addressed method. and the resin/products are kept separate. One way to deal with this problem is to use less than fully equivalent .

and the last group of resin piles is usually not mixed.51. one could start with the same amino acid attached to a bead. Such collections would be truly combinatorial. then each pile could be reacted in parallel subpiles with a different one of the 20 amino acids and repooled. By illustration. AD etc. 1. Keeping this one constant in the next series. each should be added in about one-half molar quantities. At the end of the reaction. If this resulted in 20 piles of beads. to the optimal sequence at all positions and produce a structure-activity relationship based on those other substances approaching its potency (GEB > IEB > HEB in the example). . This method of deconvolution is known as positional scanning. If one has. Synthesis of all of these actives as individual pure chemicals (often called "discretes") will allow the development of a structure-activity relationship.4. In general it is found that more than one active peptide is obtained. Alternatively.6). These are pooled again and thoroughly mixed. and very large libraries have been produced. amounts of each component and to allow the reactions to go to completion.5. Fodor photolithographic method. Iteration would lead. A great diversity of peptides can be constructed in short order using these methods. Detaching and testing would reveal which was the best amino acid to start with (XXB in the example shown in Fig. repeating the process with the remaining 19 would reveal the best second amino acid (XEB in the example). For example. one should have equimolar amounts of both products. No assumptions are made in pursuing the study in this manner. In split and pool synthesis (Fig. and the resulting beads are then split into equal portions and each of these groups of beads is deprotected and reacted with one of a group of different second amino acids to form dipeptides. if run in the manner described. Although each individual bead will contain a single peptide.Fodor photolithographic method riq'=J -r m I aaaaaa aaaaaa aaaaaa Combinatorial Chemistry and Multiple Parallel Synthesis - bbbbbb bbbbbb ccccccaaaaaa bbbbbb ccccccaaaaaa bbbbbb ccccccaaaaaa bbbbbb ccccccaaaaaa ccccccaaaaaa bbbbbb ccccccaaaaaa bbbbbb cccccc cccccc cccccc CCCCCC cccccc Figure 1. a lead peptide already or wishes to define a re- P P P AB t t t AC AB. AD Figure 1. if two components are employed. Pool and split method. and Lam (71). 1. This pool is again separated into equal portions and each is reacted with one of a different group of other amino acids to produce a group of tripeptide mixtures. after significant labor. Many subvariations of this process can be envisioned. split synthesis methods were developed contemporaneously by Houghton (69). This is continued until satisfied. however. each with a different amino acid attached. the final products from this methodology consist of groups of related materials. AC. Continuation would lead to 20 collections of all of the possible tripeptides. Furka (70). an initial reaction is run on a bead support to attach an amino acid as before. This would allow the subsequent testing results to be quantitatively comparable (68).

and the residual sequences not of interest in this test series can be archived and examined in future test systems. Split and mix method with positional scanning. Another advantage of the method is that the final products are not tethered so are able to assume the solution conformation dictated by their sequence or by the receptor interaction and also to interact with insoluble receptors. Figure 1. Biological response X-X-A X-X-B X-X-C Pool ldentify pool X-X-B as the more potent mixture X-D-B mixture X-E-B mixture X-F-B Step 2. keeping the remainder the same. then one can perform a more limited study by systematically varying all of the individual positions in the region in question in this manner. X-D-B X-E-B X-F-B Pool ldentify pool X-E-B as the more potent Step 3. Biological response . Although this method is somewhat wasteful. This is similar to the divide-couple-recombine method just described. Testing the 20 libraries reveals which amino acid is optimal at that position. This is repeated until all of the amino acids in the sequence being examined have been evaluated and an optimized sequence is at hand. Repeat sequence with pool X-X-B however varying the second position. Cleave and test each pool separately. but it evaluates only a single specific residue at a time. 1. . In the substitution/omission method (Fig. ldentify G-E-B as the most potent analog. one starts with a lead sequence whose activity is known and replaces sequentially each amino acid with all possible 20 analogs. gion in a lead peptide.3 Solid Phase Organic Synthesis of Informational Macromolecules of Interest to Medicinal Chemists 9 mixture X-X-A mixture X-X-B mixture X-X-C Step 1.7). I . The general experience is that potency and selectivity increases as each position is optimized and the known sequence grows (72).6. it has no preconceptions. one starts with known sequence A-B-C-D and discovers enhanced potency in the modified sequence of E-F-C-D. Repeat the sequence with pool X-E-B however varying the first position. In the illustration.

Detach and test. Substitutio~omission method. Another popular method of identifying residues is to attach a non-peptide signaling molecule orthogonally to a different attachment arm whose reactivity differs from the first each time an amino acid is attached to its arm (Fig. Vary position A with all 20 amino acids. Much ingenuity has been expended on ingenious methods of deconvolution. it is technically not possible to analyze each product.7. so . Xn-B-C-D Detaching from the resin and testing revea1s. This would seem to parallel nature's use of biological evolution to produce chemical libraries suitable for particular biological purposes.1. One can also place bar codes on the beads for convenient reading. Vary position B with all 20 amino acids. This is done at each position until the optimal sequence is detected and the relative contribution of each amino acid side-chain is clear.Combinatorial Chemistry and Multiple Parallel Synthesis Lead substance = A-B-C-D Step 1. Detach and test. 1. 1.9). The signaling or coding molecule needs to be attached to its arm using chemistry that does not interfere with the growing peptide on the other arm and does not detach either molecule prematurely. Figure 1. In the omission method. Clearly with very large libraries. Strategy b uses an external nucleophile for attachment. Various halogenated aromatic residues have also been used for this purpose. E-Xn-C-D ----c - E-B-C-D E-FC-D Detaching from the resin and testing reveals E-F-C-D to be more active than A-B-C-D. There are a variety of strategies employed in detaching the sequences from the arms (Fig. 1. Strategy a represents an internal displacement reaction and leaves no trace behind in the product of the original point of attachment. The development of these methods presents the medicinal chemist with the ability to perform a chemical evolution. The ease of identification of oligonucleotide sequences (by PCR methods) has made these popular for such coding. Step 2. Devolution by positional scanning is facilitated through testing groups of resins arranged in checkerboard rows and columns as illustrated earlier in Fig. Repeat this process with positions C and D Detaching from the resin and testing reveals E-F-C-D to be the most active sequence in this illustration. Strategy c is a reductive or hydrolytic strategy and sometimes leaves a linker trace in the product but need not do so.E-B-C-D to be more active than A-B-C-D. Subsequent libraries are much smaller so the process becomes progressively less laborious. Chemical evolution of this type is more congenial to the impatient chemist. The signal sequence carries the history of the bead and therefore codes for the history of the steps involved in the synthesis and thus for the identity of the peptide that one believes one has attached to the bead.8). one deletes one of the amino acids from a given position (most often at the end) or replaces an amino acid residue at any chosen position by an alanine or glycine and finds which omission decreases activity. Step 3. Another popular method is to embed an rf generator tuned to individual frequencies in the resin itself so that the substance on the bead can be identified by tuning to the proper frequency. This completes the product structure and usually does not leave a linker trace.

In one important recent example. Much work has been devoted to dealing with this problem but a comprehensive treatment of this complex topic is unfortunately too vast to cover here. B / \B/ C \ Product o fb D a = internal displacement b = hydrolysis or reduction c =external displacement / C\D Product o fc Figure 1. This is reasonably good for such complex work. the incorporation of unusual amino acids also began (Fig. combinatorial libraries of peptides consisted primarily of products made from linear combinations of naturally occurring amino acids. Figure 1. Here.8. This is not particularly surprising but underscores the importance of the topic when interpreting the results of screening (73). one must resort to statistical sampling instead or take it as an article of faith that each compound has been successfully prepared. These substances can be termed pseudopeptides H / A A ' . Resin with arms containing an analog and a signal sequence.5% of the contents of a library of 25. but leaves one with a sense of unease in that about 3800 wrong substances were available. Selectively detach analog and test. There are relatively few instances in the literature where a careful census has been performed from which to form an opinion on this topic. but this is conjecture not science. Those resins that contain an active material give a zone of inhibition or a color response. Selectively detach signal sequence and analyze. It is important also to note that the wrong structures were not statistically distributed among all of the targeted compounds but rather showed a bias toward certain structures. Initially.3 Solid Phase Organic Synthesis of Informational Macromolecules of Interest to Medicinal Chemists 11 Signal sequence Analog d-c-b-a- 1. to enhance the metabolic stability of such libraries. the resins are poured onto a solid surface previously seeded with a microorganism lawn or a substrate that generates a color. leading to a persistent evolutionary drift away from collections of natural peptides.10). end capping of the amino end and the carboxy end and also cyclization have been employed to stabilize these substances. 1. a statistical sampling of 7. Another method of devolution uses a method descended from early work of Pasteur for this work. Subsequently just about every devise employed in ordinary peptide work has been applied to combinatorial studies. 2. For example. The active component can then be detached for analysis and identification.9. Some detachment strategies for detachment of compounds from resins. Clearly careful rehearsal and fanatic attention to detail in the construction of the library helps. Before long.200 statin-containing pseudopeptides showed that 85% had the anticipated structure. Both of these endpoints can be detected visually and the active resins taken off of the surface with tweezers. .

For example. fluoroamides (107).11). hydroxamates (108. Later. Libraries of these compounds were very popular for a while but interest has decreased as time passes (114. halogenated ketone hydrates. phosphinates (104106). The figure illustrates some of the groups so employed. These unusual residues were predominantly incorporated either at the end or at a point where the natural peptides would be cleaved by enzymatic action. incorporation of a pepstatin residue . but may not incorporate any common amino acid components (112. Some peptide surrogates employed in libraries. the peptide linkage itself began to be modified (Fig. coumarins (110). the classical mode of stabilization against peptidase cleavage by conversion of the peptide bond NH into N-methyl subsequently evolved into the preparation of peptoids (polyglycine chains with each NH replaced by a variety of N-alkyl groups of the type that resemble the sidechains found in normal amino acids). vinylogous amides. Flirtation with other substitutes for normal peptide bonds includes preparation of libraries of polycarbamates (116).113).Combinatorial Chemistry and Multiple Parallel Synthesis Peptide H Peptoid I XHN L°KNyo 0 1LOT NHY B H I 0 Polycarbamate XHN byAydr+(NHy 0 B H 0 Vinylogous polyamide 0 XHN NHY XHN &k+ . for ex- ample. sulfonamides. 1. glycols. ketoamides. aldehydes. 103).oTy 0 B 0 Incorporation of a pepstatin residue Incorporation o f a hydroxyketone residue Figure 1. oxazoles (Ill). More recent libraries have appeared in which the overall conformation of the peptide has been mimicked so that the resulting heterocycle resembles topographically a p-turn. These moieties are of special value when the substitution takes place at a site where the processing enzyme must act and employs mechanism-based inhibitory mechanisms. hydroxyketones. and the like progressively appeared. boronates. methyleneketones.115). More specifically.nitriles. such residues as benzamidines (102.10. log). (74-101).

3. 1.3 Solid Phase Organic Synthesis of Informational Macromolecules of Interest to Medicinal Chemists 13 Peptide boronate H H H I R NHOH R H C C C Peptide hydroxamate Peptide aldehyde Peptide trifluoromethyl hydrate Peptide nitrite Peptide phosphonate Figure 1. Once again. 1. one sees the insertion of unusual peptide bond surrogates for lead seeking. 3. the side-chains project from each fourth rather than each third atom in the chains so these are not close models of amino acids. and there are many potentially competing functional groups that can be points of attachment (Fig.11. peptide trifluoromethylketone hydrates. such libraries are beginning to appear. Clearly progress is being made. In Fig. p-sheet analogs. and so on. Conformational effects and self-associations further enhance the diversity.11. The linkage is chiral and relatively hard to control. construction is iterative and bead-based automated procedures are available. the construction of carbon-carbon bonds is more difficult.2 both ribose and deoxyribose units) and be substantially smaller (Fig. 1. In these libraries.4 Nucleoside Arrays Lipid Arrays The minimum fully realized library of natural peptides would consist of 20'' components. An analogous library of nucleic acids would consist of 55 components (double this if one used Lipid libraries have largely been neglected. Such libraries. Peptides substituted with unusual carboxyl surrogate residues. For saturated fatty acids. Libraries of significant size have been constructed and evaluated (119).12a). (117).3 Oligosaccharide Arrays Construction of diverse oligosaccharide libraries is much more difficult. peptide nitriles. the bonds are acid fragile. These peptide and pseudopeptide libraries have been replaced progressively by collections containing smaller and more drug-like molecules. peptide hydroxamates. peptide phosphonates. peptide aldehydes. not surprisingly. These will be covered in their own section below. the use of post-translationally modified bases or wholly unnatural analogs increases the attainable diversity. This is at present a very active subfield of medicinal chemistry. and so on.Once again.12b). 3. Despite these complicating factors.and ureas (118)as well. Such residues include peptide boronates. One could also add to this list inclusion of p-turn mimics. are more often antagonist rather than agonist. Ste- .

This is a high hurdle. Many of the methods used for large molecules carry over but the largely non-iterative nature of small molecule synthesis is a significant complication. however. Obviously construction one at a time in the usual iterative or non-iterative fashion results in single molecules of a defined nature.-loC. roids and other polyisoprenoids are also complex to construct. the field of combinatorial chemistry and multiple parallel synthesis started with libraries of peptides. (b) Oli- gosaccharide libraries. The number of compounds that can become satisfactory drugs encompassed in this impossible collection is probably in the range of a few thousand. The current "gold standard" in small molecule drug seeking is oral activity accompa- nied by one-a-day dosing. With non-linear products a different variant is seen than experienced with large molecules (Fig. The majority of molecules that have passed have molecular weights of about 500 or so. these variations can be run in mixture or in parallel effecting a very significant time savings. so most of the effort would be wasted. three. While this evolution is still ongoing. The change brought about by combinatorial or multiple parallel synthesis methods is that the reactants are usually linear. reacting A with 10 different Bs produces 10 products (A-B. a single reaction produces a single product. or four such functional groups and given the number of possible variants. put a sig- . one sees illustrated centroids with two.12. Reaction of this with another substance produces product A-B-C. If the reaction conditions allow. this can lead to a very large library of analogs in a brief time (two functional groups with 10 variant adornments quickly results in 100 analogs. either in mixtures or as discrete compounds. it is now accompanied by a major effort to produce libraries of small. (a) Oligonucleotide libraries.13).). It has been calculated that the number of small molecules that would fall into this category is approximately 10 raised to the 62nd power! Clearly preparing all of these in reasonable time is beyond the capacity of the entire population of the earth even if they worked tirelessly.. 1. and the reaction with 10 different Cs on each of these results in 100 products (A-B. Rather large compound collections can be assembled quickly using this scheme. DRUCABLE MOLECULES As noted above. and four in 10. In time. three in 1000. Here a starting material (often called a centroid) with a number of functional groups (preferably with different degrees of reactivity-known as orthogonality) can be reacted with a variety of substituents (often called adornments) to produce a large number of analogs. It can. In the figure. unusual residues crept into the products. Hence medicinal chemical skills are still at a premium. 4 SOLID AND SOLUTION P H A S E LIBRARIES OF SMALL.-lo).000). In each case.Combinatorial Chemistry and Multiple Parallel Synthesis I O=P-0 IOH y O j base' Figure 1. drug-like molecules in library form.-. and one anticipates developments in this area. but the products can be logarithmic. For example. Combination of reactants A and B produces a single product A-B. Mixed triglycerides would seem accessible.

the polarity of another usually must be decreased. Converging methodologies address these limitations successfully. etc. linear syntheses are the most risky and produce the lowest yields. Centroid adornment. Being selective in the variations actually incorporated. Figure 1.13. Here the library was constructed by a combination of three reactants. record keeping. Clearly a great deal of thought should go into library design before the work begins. they should project into various quadrants about the centroid so as to allow a fruitful exploration of potential receptor needs. for example. budgets. and in combinatorial work. then this requires one or more of the other adornments to be made compensatingly smaller. This would fit the commonly accepted meaning of combinatorial in that all of the possible variants would be constructed. Ugi (four-component) and Passerini (three-component) reactions are very flexible and popular. Alternately. If tethered. 1. It is also helpful if.14) ( 6 ) .A notable chemical feature is the use of amino acid fluorides to drive the amide formation to completion. The molecular weight of the centroid places practical limits on the net weight that the adornments can collectively have. convey toxicity. They should not be chemically reactive. Such a li- . In hit seeking. it can be seen that centroid A should be chosen to have the smallest practical molecular weight. The net hydrogen bonding inventories. the adornments must have the usual medicinal chemical characteristics. if one is hit seeking. whereas in lead optimization progressive rigidification is often more effective. With 500 as the normal practical upper molecular weight limit. analysis. log P. If one adornment is rather large. Generally one has less control over the specific products being produced by such reactions but this is largely compensated for by the molecular diversity available in this way. It is particularly helpful if the adornments project into space into quadrants that fit precisely the needs of the receptor if one is optimizing a lead. it has functional groups remaining that can interact productively with a receptor so the weight devoted to this part of the molecule is not net loss. one often wants molecular flexibility.4-benzodiazepine-2-ones prepared on resins by Bunin and Ellman in 1992 (Fig. what affect this may have on its biological properties. and cell penetrability features set constraints on the individual and the collective nature of the adornments. Much work has been expended in addressing these potential limitations. when fully adorned. If one is rather polar. One also should consider whether this attachment will prevent the use of one of the potential adornment points.4 Solid and Solution Phase Libraries of Small. or be inordinately polar. The product should fit the modified Lipinski rules to allow for the usual molecular weight and lipophilicity creep that often accompanies analoging. the smaller each adornment can be. water solubility. it is important to consider whether the point of attachment will remain in the final analog. The more functional groups present in the centroid. The choice of benzodiazepines was inspired because of the medicinal importance of these materials and their resemblance to peptides. a smaller ("focused") library could be made that answered specific pharmacological questions but at the risk of missing an unexpected discovery. Drugable Molecules 15 Two substitutable groups Three substitutable groups Four substitutable groups d Etc. Just as in one at a time synthesis. In addition. Whether the centroid should be tethered to a solid support or free in solution must be considered carefully. the library could easily reach 1000 members (10 x 10 x 10) in short order. nificant strain on purification. If each were represented by 10 variations. and if so. The first libraries containing heterocycles recognizable as orally active drugs were the 1.

Varying the point of attachment of the hydroxyl group would lead to additional multiple . Centroids derived from molecular series that are known to be associated with good pharmacokinetics are often referred to as privileged molecules. N-alkylation ii.14. Thus. The centroid has a molecular weight of 160 when all of the available substitution points are occupied by hydrogen atoms.Combinatorial Chemistry and Multiple Parallel Synthesis CI FmocHN /O arm /O arm 1 i. /O arm Figure 1. the choice of benzodiazepines to demonstrate the potential power of combinatorial chemistry and multiple parallel synthesis was inspired. Synthesis o f benzodiazepine libraries-1. the library can contain primarily substances that have a chance to be drug. Detach Fmocdeprot. When chosen with care to convey drug-like properties and not to exceed collectively the guidelines that Lipinski has developed. This gives significant latitude for substitution. In this pioneering library. If one accepts an upper molecular weight limit of approximately 500. the final products were attached to the bead support through a phenolic hydroxyl group that remained as such in the products before testing. Such focused libraries are often more intellectually satisfying to the practitioner. brary would fit the commonly accepted definition of multiple parallel synthesis. They also can be so chosen that they allow for sub- stitution independently of each other and also to be installed without premature separation from the beads. then four variations can occupy 340 atomic mass units (500-160) so each adornment can have an average of about 85 amu. It will also be noted that the centroid chosen has amide and amine linkages that are not involved in adornment attachment and are capable in principle of interacting successfully with a receptor so the molecular weight sacrificed to the centroid may perform pharmacodynamic work. In this library the attached groups project into space at widely separated compass points around the molecule allowing a systematic exploration of receptor requirements. if the weight is evenly distributed.

As it happens. RCOCl ii. 1. these studies resulted in con- arm. Synthesis of benzodiazepine libraries-3. A somewhat more versatile synthesis of this type using stannanes and palladium acylations (Stille coupling) appeared subsequently (Fig.15.14 Figure 1.17) (122). 1.16.18) (121). The "traceless linker" technology so introduced has now become standard. testing of these agents revealed no structural svrprises. While precedent establishing. Drugable Molecules 17 1' R Similar to Figure 1. Nonetheless. One of the additional advantages of this application is that incomplete reaction occurring during the synthesis would lead to products that would not cleave from the resin and could be removed by simple filtration. somewhat gratifyingly.4 Solid and Solution Phase Libraries of Small.15). Pd(O). Indeed. of benzodiazepine analogs.16) (120). several components in these libraries were bioactive. attachment to the resin was by an amino acid ester bond. Ellman's group also developed a traceless linker sequence of a different type based on HF release of an aryl silicon link to the resin (Fig.0 SnMe3 I i. a variant of this process resulted in a traceable linker in the other aromatic ring (Fig. In this sequence. The intense study of the benzodiazepines in the empirical earlier years had apparently not missed much of significance. 1. Subsequently this bound intermediate was converted to an imine that cyclized to the benzodiazepine moiety on acid cleavage from the resin (Fig. the products were now indistinguishable from benzodiazepines prepared by usual speed analoging (USA) methods. H+ Figure 1. this was pharmacologically less NH Bpoc than completely satisfying because agents intended to penetrate well into the CNS should not usually contain such a polar substituent. Despite this. Synthesis libraries-2. and the work drew widespread attention to the promise of the methodology and was soon followed by a flood of applications to the preparation of drug-like molecules. 1. . Furthermore.

18. 125) continue to be popular. an analog emerged that was threefold more potent than captopril itself and possessed a Ki of 160 pM (Fig.23) (129.5-diones (124. 130). . Figure 1.Combinatorial Chemistry and Multiple Parallel Synthesis 1 Similar to Figure 1 . The conditions are mild in this synthesis and the yields are good (Fig. Synthesis of benzodiazepine libraries-4. in this case both in solution and in solid state.21) (127).4-benzodiazepin-2. The reaction conditions are mild enough to preserve the optical activity in this library (Fig.4-dihydropyridines.Hantzsch methodology (without the oxidative step) works efficiently on resins for this purpose. vincing proof that combinatorial chemical methods would be of dramatic use in preparingagents likely to become orally active drugs.I6 R2 Figure 1. 1. are the fluoroquinolone antimicrobial agents (Fig. (Fig. 1. Benzodiazepine libraries (123) and close analogs such as 1. From a collection of about 500 analogs. 1. The solution-based yields were superior to those obtained on the resins. Representative of another important class of drugs is the 1. 1.22) (128)! Another important class of contemporary drugs that have been made in library form. Use of peptoid starting materials and reductive traceless linker technology are features of the work of Zuckermann et al. 1.17. An interesting library of captopril analogs were prepared on resin using a split-mix iterative resynthesis deconvolution procedure. Synthesis of benzodiazepine libraries-5. For example.20) (125). one such library is prepared by a sequence involving a Borch reduction and an internal ester-aide exchange to form the seven-membered ring and then cleavage from the resin. Angiotensin-converting enzyme inhibitors are million dollar molecules.19) (126).

H+ i.4 Solid and Solution Phase Libraries of Small. [HI ii. Drugable Molecules i.6. Base ii. Amino acid ester I H I I X i. i. Borch reductive alkylation with aminoacid ester ii.19. H+ H R I 0 Figure 1. Synthesis o f benzodiazepine libraries. A iii. R'X iii.7.20. . BrCH2C02H ii. Synthesis o f benzodiazepine libraries. Anthranilic acid amidation 1 Figure 1.

22. Combinatorial chemistry plays a significant role in this work. 'one example of this class. Particular emphasis has been placed on libraries of molecules with interesting pharmacological properties. Many more examples could be covered. Synthesis of dihydropyridine libraries. Captopril library results.24). has recently been marketed as an orally effective anti-infective. 1. focused libraries explore the SAR properties of series of contemporary interest where no drugs have yet emerged or where the first of a promising series has been marketed.] Recently. but this gives a representative flavor of the field. linezolid. [The Captopril Figure 1. which after a mixed aldol reaction to the dimethylenamine. These are separated by chromatography to produce the best of a large series of analogs produced by these and other reaction se- . An example of this is the oxazolidinones. Hydrolysis of this group with formic acid produces a methyl ketone. By now most of the heterocyclic ring systems have been produced in library form. and most large firms have extensive analog programs in progress in attempts to improve on its properties (Fig. Rather extensive reviews of this work exist for the interested reader to consult.Combinatorial Chemistry and Multiple Parallel Synthesis arm. reacts with methylhydrazine to produce a mixture of the two possible methylated pyrazole isomers.21. reviews of Dolle (36-40) and of Thompson and Ellman (41) are particularly useful in this context. A Pharmacia group has shown that alteration of the morpholine function to a methylated pyrazole moiety produces a broad spectrum analog with oral activity. Palladium coupling of the iodoaromatic moiety of the starting material allows construction of the trimethylsilylacetylene side-chain. H R O H Figure 1.

After the library is screened. the needs for quantities of material for evaluation become more and more so the work usually proceeds back into the larger scale one at a time mode resembling the BC (before combichem) era.. of 80 nM.24 has the best combination of in vitro and in vivo properties in this grouping (131). D showed enhanced metabolic stability. Screening produced leads C. with an IC. In addition. Here the initial libraries are usually bigger and more diverse. of 50 nM and an - . Drugable Molecules Steps 0 Figure 1. Many analogs were then available by comparatively simple variations in the reactants employed. The product illustrated in Fig. possessing an IC. These results were very encouraging. After some adaptation to the needs of the method and rehearsal of the chemistry. First. is rather different. The compounds were prepared in high yield (82-93%)and purity in about 2 min with the aid of microwave irradiation (Fig. A library of cephalosporin antibiotic analogs was made on a solid support (basic alumina) without requiring protection-deprotection. The small molecule libraries just exemplified belong to the class called focused. possessing an IC. so it was chosen as the lead for the next phase. in drug seeking. Each succeeding library benefits from the information gained in the previous work so this can be considered the chemist's equivalent of biological evolution. First. Interestingly. 133). and B.. with an IC.25) (134). possessing an IC. Clearly. Reports of related studies have also appeared (132. A couple of examples represent the very large amount of work carried out in this manner. and D. The chemistry in libraries does not always go as intended. 1. one can operate in much the same manner after identification of a suitable hit molecule.. Screening this library in whole cells led to the identification of two main leads. The choice of materials was based on prior knowledge of the structures of other P-glycoprotein modulators. in each case a discrete molecular target was available at the outset and chemical routes were generally available. D was an unexpected by-product. of 600 nM. of 300 nM. Microwave acceleration in combinatorial chemistry is a powerful technology for enhancing reactions on solid phases.26). a 500-membered library of variously substituted imidazoles was prepared on a mixture of aldehyde and amine beads (Fig. Analoging around structure D lead ultimately to OC 144-093. however.4 Solid and Solution Phase Libraries of Small. Fluoroquinolone libraries. and useful molecules are uncovered. 1. That is. consider the discovery and progression of OC 144-093. of 150 nM.23. The third stage involved making a solutionbased library based on the structures of A and B. an orally active modulator of P-glycoprotein-mediated multiple drug resistance that has entered clinical studies. As the work progresses. The strategy required in hit seeking.. In addition to reasonable potency. subsequent refining libraries are employed that are progressively smaller and more focused. A. quences. libraries could be generated relatively quickly.. B possessed an oral bioavailability in dogs of about 35%. 1.

estimated 60% bioavailability after oral administration in humans(135. It is not a substrate for CYP3A and interferes with paclitaxel metabolism only at comparatively high doses. Further studies are in progress and it is hoped that a marketed anticancer adjunct will emerge in due course as a result of combinatorial chemistry (137). Chromat. The results are interpreted as meaning that OC 144-093 interferes with gut P-glycoprotein. a search through a company compound collection was made in an attempt to find an inhibitor of the Erm family of methyltransferases. Figure 1. In a different study.Combinatorial Chemistry and Multiple Parallel Synthesis Linezolid Strong gram +ve Moderate gram -ve Potent in vivo < Me2NCH(OMe)2 NHAc I ii.24. Oxazolidinone libraries. enhancing oral bioavailability. After IV administration. OC 144-093 does not interfere with paclitaxel's pharmacokinetic profile but elevates its area under the curve when given orally. Later biological studies in vitro and in vivo have shown that the agent enhances the activity of paclitaxel by interfering with its export by P-glycoprotein.1 3 6 ) . These bacterial enzymes produce resistance to the widely used macrolide-lincosaminide-streptogramin B an- .

4 . Using NMR (SAR by NMR) screening. starting with a very weak lead with a malleable structure. Drugable Molecules 23 I Microwave based chemistry. chemists explicate the route with formulae and often discuss the relative strengths and weaknesses of key reagents but almost never devote time to workup. From a drugability standpoint. and manifolds for filtration and for solvent removal are commercially available. Performing chem: istry on beads addresses this in that simple filtration and washing often suffices. Little is gained if one saves much time in construction only to have to give this back by tedious and repetitious purification schemes. This is not as useful if the reactions do not go to completion.25. this is often more manageable. so considerable excess of reagents and more lengthy times are often employed to drive the reactions further to completion. Column chromatography is powerful but often labor intensive and solvent consuming. Even so. This factor becomes even more demanding in combinatorial work where the need for rapid. a series of compounds including 1. the details of the workup require attention to detail in the performance and are sometimes quite challenging. focused libraries. there is a danger in this. These problems have largely been overcome. yet already most of the common drug series and hundreds of different heterocyclic classes have been prepared in library form. Cephalosporin libraries. effective workup is intensified. These were now potent in the low micromolar range. From this. and the length of the reaction sequences required.0 mM for the triazine named) (Fig.3diamino-5-thiomethyltriazine were found to bind to the active site of the enzyme. successive libraries produced analogs with quite significant potency for further exploration (138). With smaller. compounds E and F emerged. Originally the emphasis was on bead- In communicating their results. . tibiotics by catalyzing S-adenosylmethioninebased methylation of a specific adenine residue in 23s bacterial ribosomes. 1. and today the choice of beads or no beads is partly a matter of taste. the size of the libraries being made. Separation of hundreds of analogs by column chromatography would be a nightmare.dl4 against ErmC. The remainder of this chapter deals with selected examples that illustrates particular concepts and methodologies.4 Solid and Solution Phase Libraries of Small. A number of commercially available combinatorial screening libraries are peppered with such substances. This interferes with the binding of the antibiotics and conveys resistance to them. E emerged as the best substance with a Ki of 4 pM against Em-AM and 10 /. A solution phase parallel synthesis study was performed from which compound D emerged as being significantly potent. and C identified as promising for further work. 1 Purification Figure 1. Thus. The left side of analog E was fixed and the right side was investigated through a 411-membered library. Separation from solution in solid form or simple evaporation is very convenient. and analogs A. and indeed. Compounds that separate readily from polar solvents are often of very low water solubility and present difficulties in testing. and this actually slowed general acceptance of the method because few organic and medicinal chemists were familiar with the techniques needed to make small molecules on beads through non-iterative methods. Analogs were retrieved from the collection. B. albeit weakly (1. Next a 232-compound library was prepared to discover the best R group on the left side of compound D. much of the needed technology had yet to be developed and disseminated. From this.27). It is " iust a decade after this field became generally active.

A P I R4C~0. R1~OCOR2. NH40Ac. Library-based discovery of clinical candidate C 144-093. NH~OAC. Analoging H Figure 1.R=H Phase 4. NH40Ac AcOH Me C. R1COCOR2.AcOH A Phase 3.26. AcOH A R I CHO P NH2 Phase 2. R~CHO. RNH2. R = M e D.24 Combinatorial Chemistry and Multiple Parallel Synthesis Phase 1. .

X= a. H2N KD= 1.0 rnM Step 2. Analoging D Step 4.OH Y= \ / KD=0. Drugable Molecules Step 1. A library of E m inhibitors using SAR by NMR. Screen in house analogs NHz I A. Second library and screening Figure 1.Solid and Solution Phase Libraries of Small.31rnM Step 3. .27. Library synthesis and screening Step 5.

1. the requirements of combinatorial chemistry have engendered a flowering of additional resins and uses. Whereas ion exchange applications have been around a long time in the medicinal chemist's laboratory. However this is generally exceptional. In this case. More generally applicable has been the development of many reagents tethered to solid supports. The diphenylphosphine oxide product is often troublesome to remove from nontethered reactions. One can sometimes doctor silica gel. Chromofiltration methods are powerful and rapid but often require study for optimization. the compounds are in solution and the reagents are on the solids. In Fig. tethered bromine dimethylamino hydrobromide cleanly a-brominates acetophenone.28b. A great many reagents have been prepared for use in this manner and the area has been reviewed extensively (140-142). In Fig. tethered chromate cleanly oxidizes benzyl alcohol to benzaldehyde. for example. 1.28. for example. Resin tethered reagents used in combichem.28a. After amide formation. when the acid component is used in excess. resin bound diphenylphosphine is used to reduce benzamide to benzonitrile. but a few examples help clarify the many uses to which this exciting technology can be put. 1. that choice of the appropriate solvent for solutionbased multiple parallel synthesis can occasionally result in reaction mixtures from which the desired product can be isolated in pure form by suitable choice of absorbent and concentration or evaporation from the eluent (139). We have found. In Fig. More frequently automated chromatographic reverse phase methods are employed in which round the clock separations not requiring constant human supervision are available.28c. but here is readily separated leaving the clean product behind. These are particularly useful in solutionbased MPS but clearly find wide applications in other types of chemistry as well.26 Combinatorial Chemistry and Multiple Parallel Synthesis Some Examples : (a) PPh2 + Br + R-CHO CHO (b) 0 0 + NH CrC103 Figure 1. adding 1% of sodium bicarbonate to silica gel and mixing thoroughly provides a convenient way to remove reaction debris so that filtration produces pure amide on evaporation (139). . These reagents perform their intended role and then the excess reagent and the exhausted reactants can be filtered and washed away for easy product isolation. to enhance its use for these purposes. An exhaustive treatment is beyond the scope of this summary.

This can be illustrated (Fig. These and analogous reagents have been widely employed and reviewed (140-142. they are soluble. 1. The desired intermediate is present in the product mixture with a variety of reaction detritus including diarylated material. Thus. Compounds elute from these columns in the order of their decreasing fluorine content. benzaldehyde tethered to a resin is used to remove primary amines and hydrazines. are employed to remove acidic reaction products from reaction mixtures. heating a reaction mixture involving such a molecule to speed the reaction and then cooling on completion often allows the product to separate in pure or at least purified form by phase switching into the fluorinated solvent. It is obvious that many possible variants leading to a library of related molecules could be prepared by simple modifications of the reagents and substrates. and tethered trisamine removes acid chlorides. sulfonyl chlorides. The resin bound product is purified by rinsing and the reaction sequence is completed by acid release from the capture resin. .29. for example. propranolol. three involved the use of resin-based reagents. Illustration of tethered reagents in preparing propranolol. In the six chemical steps required for this preparation. This methodology is similar in concept to the well-established ion exchange methodologies. these materials can be used to remove acidic reagents or byproducts leaving the desired reaction product in the solution. KOH. Isocyanate resins are used to remove primary and secondary amines and alcohols. carbonate resins remove carboxylic acids and phenols. The use of acid and base resins to remove ionizable products and byproducts from reaction mixtures is familiar. A 0 Figure 1. however. Likewise. This is very convenient for rapid work-up of combinatorial libraries.The sequence starts with Suzuki-type coupling with a series of aryl halides. diphenylphosphine resins remove alkyl halides. Drugable Molecules MsO (Ar = a-Naphthyl) * I. In this case. Highly fluorinated organic molecules are often insoluble at room temperature in both water and in organic solvents. 1.30 wherein tamoxifen analogs are efficiently and cleanly prepared (145). Another important and powerful methodology employs capture or scavenger resins.4 Solid and Solution Phase Libraries of Small. An illustrative example of the power of this method is the resin-assisted synthesis of the adrenergic P-blockingagent. Here resin-tethered bases. outlined in Fig. At higher temperatures. 144). This device is sometimes called phase switching.29 (143). and these can be regenerated for further use.31) by application to a library of 100 mappicine alkaloid analogs. 1. silica gel columns with a fluorous phase have been introduced to facilitate separations. The desired product is captured out of this stew by use of a resin bound aryl iodide which reacts exclusively with it. L 1 \ II AcO I + 0 NMe3 ArO- Propranolol ii. An example of the preparation of a library of drug-like molecules in solution employing resin capture methodology is illustrated in Fig. The use of tethered isocyanates to remove excess amine is less familiar but readily comprehended. Recently. and isocyanates.

Three grades of products can be distinguished. Figure 1. Pure usually means greater than 95% in combinatorial work. called "practically. This chemistry is also comparatively "green" in that solvent needs can be greatly reduced and disposal of unwanted materials is simplified. Actionable quantitative biological data can be obtained from pure samples and uncertainties into selecting the most active constituents to pursue are increasingly introduced by assaying less pure material.Combinatorial Chemistry and Multiple Parallel Synthesis ___) R3-~rx Suzuki rn. This complicates analoging but is more satisfactory than basing SAR-based design on negative activity data wherein one can be significantly mislead in such cases. . disparagingly. Detagging produced the individual pure products. A lower but generally acceptable grade of purity is arbitrarily chosen at about 80%. by chromatography over immobilized receptor preparations.30. Syntheses could then be performed in mixtures and at the end the products were separated based on the number of fluorines in their tags. 147). The method is not general but is very convenient when used appropriately (146. each analog was tagged with an arm containing one of a series of a perfluorinated hydrocarbons (CnFn+l). chemists occasionally report anecdotally finding that an active component in a library has none of the intended compound in it at all. 4. and such compounds can be labeled as "practical grade." Less than this level of purity is generally unacceptable and is sometimes. The level of desired purity of the components is also a matter of debate. but the consensus is that 85% success is fine. Gel filtration is also helpful in sorting out a binding component from a mixture library containing analogs with little affmity for the pharmacological target. Illustration of resin capture use in a library of tamoxiphen analogs. today one rarely sees a separatory funnel in a combichem laboratory. inevitably one has some components in this poor state. if aggressive by-products and reagents are not present. It is also possible to remove desired reaction products. Because of the impact of these newer methods." In very large libraries where purity analysis of each component is rarely available. Indeed.2 Synthetic Success and Product Purity Only a relatively few census reports assessing the success rate of combinatorial library con- struction are available (1481.

ICI ii. polyethylene glycol (PEG) grafted resins. . Commonly employed resins are the Merrifield. such libraries have a structural bias. These include cellulose fibers. sintered polyethylene. Tag OMe * I i. and silica gels. cross-linked polyacrylamide resins. The particular advantage of this inert support is that the whole volume of the gel is available for use rather than just the surface. and PEG-based resins. Surface functionalized supports have a lower loading capacity and many types are available. Detag Tag . the chemistry may selectively favor production of more lipophilic substances so that hydrophilic examples are underrepresented. Thus. The linking functionality varies. and the like. Geltype supports are popular and consist of a flexible polymeric matrix to which is attached functional groups capable of binding small molecules. For example. Asyrn.31. Composite gels are also used. Brush polymers consist of polystyrene or the like grafted onto a polyethylene film or tube. ii. These include treated Teflon membranes. Generally these consist of cross-linked polystyrene resins.4 Solid and Solution Phase Libraries of Small. Drugable Molecules OMe I i. ii. Chrornat. BBr3 i. Illustration of fluorous phase methods in the synthesis of a mappacine library.3 Resins and Solid Supports A great variety of resins and solid supports are available for combinatorial work (151). redn. the wrong or missing structures are not statistically distributed (149). One is disturbed to note also that in a few cases where a census has been taken of very large libraries. It is hard to see how to get around this conveniently. glass. 4. kieselguhr.d' Figure 1.

With solid-state libraries. One notes however that the comparative speed and ease of molecule construction makes it possible to reduce to practice rather more examples that would have been possible in the one-at-a-time days. magic angle NMR. Rink.Combinatorial Chemistry and Multiple Parallel Synthesis CH&I Merrifield resin ~ e d Rink resin Wang resin Ellman resin Figure 1. 4. The best of these have structure drawing capacity also.4 Microwave Accelerations been put into working out analysis of single beads with mass spectrometry. This has been adapted to combinatorial methods and is even compatible with a 96-well plate format (150). This closely parallels general experience in the pre-combichem days with the excevtion that the work load is greatly magnified. An enormous effort has Patent considerations are complex in combinatorial chemistry. For example. These are illustrated in Fig. the problem is much more complex. 4. Keeping track of this is a job for highspeed computers. heat-demanding Diels-Alder reactions can take days on solid support.5 Combinatorial synthesis and high-throughput screening generate an enormous amount of data. The mass of potential data is hard to compress into a suitable format for this purpose. Some common resin types. 4. If activity is found then more detailed examination takes place. Commonly. and cherniluminescence techniques being particularly popular. and the reader can readily appreciate the kinds of chemistry that they allow. Automation is called for and hplcltof mass spectrometry is of particular value. Rather more disturbing is the increasing tendency to patent various means of making and evaluating libraries rather than focusing on their content.6 lnformetrics Molecules possessing a permanent dipole align themselves in a microwave apparatus and oscillate as the field oscillates.7 Patents Analytical Considerations Analysis of the degree of completeness and the identity of the product is simpler than that seen in solid phase work. and only minutes under microwave. Raman spectroscopy. and there are commercial packages that may be useful as well. one is usually restricted by necessity to statistical sampling and compound identification rarely goes beyond ascertaining whether the product has the correct molecular weight. Patenting of means of produc- . Wang. The fundamental purpose of patenting is to promote the useful arts and to provide protection for innovative discoveries for a period and then to share them with society in general. with very large libraries. hours in solution.32. patenting takes place comparatively late in a drug-seeking campaign and so differs little from traditional patenting. Many firms have developed their own programs for the data handling.32. Even so. This rapid motion generates intense homogeneous internal heat greatly facilitating organic reactions. 1. and Ellman types. 4. especially in the solid state.

solubility. To produce a brick wall of a given dimensions more quickly is largely a matter of buying the bricks and hiring and motivating enough skilled labor. it is . No laboratory seriously involved in the search for new therapeutic agents can afford not to employ this technology. 5 SUMMARY AND CONCLUSIONS Combinatorial chemistry and multiple parallel syntheses have transformed the field of medicinal chemistry for the better. if carried to an extreme. In drug seeking. Combichem promised to make this a reality. one has to design the bricks first and develop the technology. and penetrability. toxicity. and we no longer in the main waste time on collections of molecules that have no chance of becoming drugs. Fortunately the flow through the pipeline diminishes through increasing failure of leads to qualify for further advancement and this is helpful in reducing the magnitude of the job but the problems remaining will still be vexing. The present wedding of combichem with medicinal chemical knowledge is extremely powerful. indeed. It would be nice. if this had turned out to be true! It cannot be denied that there is some justice in this belief. Enhanced speed of construction simply produced a greater backlog of work to be done.5 Summary and Conclusions ing libraries. the next choke point in the pipeline has become animal testing. and there was a demand for more compounds. The last decade has seen a revitalization and much dramatically useful technology has been discovered. synthesis could be accomplished much more quickly than screening and evaluation of the products. Nonetheless. From the vantage point of 2002. It can readily be seen that further choke points lie distally in the pipeline and these will have to be dealt with in turn. no matter how much money and effort one is willing to throw at the problem. In addition. Some believed that there were large numbers of underexplored structural types that could be drugs if only they were prepared and screened. A significant part of these needs were met by the methods in this chapter. Part of the difficulty is that certain biological phenomena cannot be hurried. Combichem does speed the process along but does not remove the elements of uncertainty that must be overcome. producing a baby requires essentially 9 months from conception. Given the strictures placed on clinical studies and their solidification in law and custom. Generally. would have a dampening effect on the development of the field and so would inhibit the development of the useful arts. BC (before combichem) there was little motivation for enhanced speed of synthesis. In the heady early days of combinatorial chemistry one frequently heard the opinion that existing drugs were only those to which nature or good fortune had laid a clear path. With synthesis and screening back in phase. new firms were founded to take advantage of newer screening methodologies. These firms had no retained chemical libraries to screen and larger firms were reluctant to allow their libraries to be screened by outsiders. Hiring nine women will not result in producing a baby in 1 month. Some time can be saved by speeding things along the way and also by dealing with the remaining constrictions in parallel rather than simultaneously. These factors are presently under intensive examination in attempts to elucidate these properties in a similarly rapid fashion or to predict them so that favorable characteristics can be designed into chemical library members from the outset and thus largely avoid having to deal with them. The problem in shortening drug seekingis further compounded in that the problem is not akin to brick laying. Clearly space for chemical diversity is larger than space for medicinal diversity. This should be guarded against. The advent of high-through- put screening in the 1980s changed all this. but it is difficult to see how they can all be resolved in a rapid manner. For example. speculative synthesis continues to reveal important drugs. The backlogs emptied rapidly. one can now look back at what has been done in the amazingly short time that this technique has been widely explored and one can see some things more clearly now and use the methodology more cunningly. pharmacokinetics. the cruel restraints that ADME and toxicity considerations place on our chemical imagination have ruined this dream of easy and unlimited progress.

effectiveness is doing proper things. R. Beck-Sickiner and P. Eds. Merrifield.. 7. L. H. There has been a dramatic increase in investment in drug discovery during the last decade (estimated at 10% annually).. T.. New York. 85. 1999. S. 354. American Chemical Society. J. N.. 14. 12. 773 (1975). Gordon and J. Soc. 10.2000. Abelson. Pinella. Combinatorial Chemistry in Biology. Combinatorial Chemistry: A Practical Approach. Ideally one should strive to make a few more useful needles embedded in progressively smaller haystacks. 232. 114. W.. 2001. Certainly chemical novelty has largely given way to potential use. DeWitt. Springer Verlag. Today. Pless and R. 18. S. M. R. 3 . . High-throughput screening can be likened to hastening the process of finding a needle in a haystack. A. CA. 2. 16. Am.. I. San Diego. B. Acids Res. NY. Fenniri. M.. 1998. Combinatorial Library Design and Evaluation. This chemist receives approbation for hisher efforts. H. Academic Press. .. 2149 (1963). Med. Hindsgaul.-H. Janda. - REFERENCES 1 . Huntington. Diversity no longer rules. Chem. F. K. Merrifield. Houghton. Barteling. Ed. Czarnik. Part of the explanation for this is that ease of synthesis does not necessarily equate to equivalent value of the products.B. Despite all of these considerations. Academic Press. W. 21. B. The Combinatorial Index.. Bunin and J. Nova. Chem. A. Solid-Phase Organic Synthesis. Eds. H. Appel. Science. Dekker. 8. 435. 20. C. A. R. Wiley. H. (1998). 2000. Combinatorial Chemistry and Molecular Diversity in Drug Discovery. It is interesting to note also that a 100-fold increase in screening activity has not yet resulted in a corresponding increase in the introduction of new pharmaceuticals. A. 5 . Combinatorial chemistry and multiple parallel synthesis in the hands of the skillful and lucky chemist rapidly zeros in on the best combination of atoms for a given purpose. 13.. USA. New York. Unfortunately this has yet to result in a burst of new introductions. J. A. Another exculpatory factor is that much of the low hanging fruit has already been harvested and the remaining diseases are more chronic than acute and are much more complex in their etiology. 3998 (1984). N. Czarnik and S. Molecular Diversity and Combinatorial Chemistry: Libraries and Drug Discovery. J. Drug Discov. Ghose and V. New York. 1997. C.G. Dooley. J. New York. 17. Oxford University Press. K. vol. D. M. Winnacker.32 Combinatorial Chemistry and Multiple Parallel Synthesis hard to see how this phase of the drug seeking sequence can be shortened through chemical effort. Blondelle. Ed. Acad. 7 . M.. New York. C. Those who consistently come up with useless compounds will eventually be encouraged to find other work. Combinatorial Strategies in Biology and Chemistry. As with much of the points being discussed. Eds. This involves mating as well as is possible productive chemical characteristics with productive biological properties. SOC. DC. 2000. Exp. (1990).84 6. American Chemical Society. 19. L. Combinatorial chemistry can be likened to the preparation of needles. W.. Am. Proc. Furka. San Diego. 11. Combinatorial Organic Chemistry: An Educational Approach. 2002. achieving a proper balance is essential. 81.Chaiken and K. New York. R. Ellman. Viswanadhan. J. Letsinger. Nucl. Solid-Phase Organic Synthesis. 2 .10997(1992). C. and J. Washington. Heber. Cuervo.219 9. R. J. A Practical Guide to Combinatorial Chemistry. Sci.341(1986). Eds. Z. drug seeking is an exciting enterprise calling for the best of our talents and the appropriate use of high speed synthetic methods gives us a powerful new tool to use. A. WileyVCH. B. New York. Combinatorial Chemistry. Natl. DC. E. 1996. This is perhaps the combinatorial chemist's equivalent of the businessman's mantra that whereas efficiency is doing things properly. El-Basil. 1998. Kerwin. Combinatorial Chemistry: A Practical Approach. 2001.l(2002).A. 1. Meloen. A. Nature. E. New York. Wiley. If each compound in chemical libraries was carefully designed and the data therefrom carefully analyzed. 1999. Wiley. Bunin. Geysen. 4. Burgess. R.. Eds. Wong. Washington. Eds. Wang and 0. M. H. 15. Felder. Famulok. Biol. Ed. and S. then the disparity would be smaller than the present experience produces. E. Wiley. Bannwarth and E. Adv. 1996.

Sprengeler. M . Eichler. New York. 62. Ottenheijm. Gardner. Accts. 184. Johnston. E. V . Analysis. R. FL. 36. Soc.W . 1998. Baumbach. and E. Hellstrom. 51. 26. Diversity. Acc. P. Boca Raton. N.. C. Combinatorial Chemistry: Synthesis. Ostresh. Comb. K. A. I. E. 199 (1998). A. 29. Organic Synthesis on Solid Phase-Supports. Mol. Camarero. C. and R. V . 62. Gallop. D. J. 5643 (1997). F. Eds. A. Jung.235 (1999). 2000.483 (1995). Chem. K. 3815 (2000). Marquis-Omer. Houghton. 1. Z. Gallop. B. The Netherlands. and C .. Med. J. G. 53. C. Eichler and R. Mol. Johnson. A. J. 2.Voigt. New York. G. A. Weber. 61. 56. Steele. H. 32. H. D. Rees. M. von dem Bussche-Hunnefeld. 43. Nielsen.. Dolle. Chen. J. J. Appel. Zeltser. J. J. P. Muller.. Fassina. and C . Screening. Linkers and Reactants. New York. Chem. Ed. Combinatorial Peptide and NonPeptide Libraries-A Handbook. 60. R. Oxford University Press. Chem. I. Terrett. Dooley. Combinatorial Chemistry. Biopolymers. a n d L. Moos. 1998. et al. New York. H. Reed. M. I . 4. W . M. Wiley. 38. Czarnik. 3. Balkenhohl. 40. Divers. CRC Press. R. and A. S. Ed.. 1997. 24. Dekker. Analytical Techniques in Combinatorial Chemistry. A. H. Gordon. Wiley. A. Eichler and R. 34. 1. Yang... Tetrahedron. 4. R. M. and J. J. C. 1997. Kay. 51. 32. 1233 (1994). A. Kobylecki. W. Chem. Houghton. 1998. Swartz.. 42. F. R. S.96. Gordon. C . and C. M. Comb. Comb. E. W . Veerapandian. Patel. 271... 10. Bioorg. Annual Reports in Combinatorial Chemistry and Molecular Diversity. Dower. Pavia. R. Chem. Diversity. High Throughput Screen. Chem. 383 (2000). 2000. A. Zysk and W . R. 59. Miertus and G.. 2. 30. 54. S. and R. 35. Dolle and K. 37. Miller. New York. Perkin Trans. V . F. 28. Chem. Oxford University Press. Hermkens. 39. 74 (2000). Tetrahedron. New York. et al. 267. 37. R.4527 (1996). 31. 2000. P. Solid Phase Synthesis and Combinatorial Technologies. Gilon. Today. P. J.233 (1995). Dolle. Thompson and J. et al. W . Dekker. Handbook of Combinatorial Chemistry. 55. 555 (1996). Angew. Nicolaou. 3. Cornbinatorial Chemistry in Drug Design. M. K. A. Res. B. New York. E. 58. Hirschmann. Ellington. K. Edwards. Eichler. Terrett.3743 (1999). 52. 1 and 2. Dodds. et al. B. 29. Terrelt.411 (1997). 27. Blake. Houghton. M.. H. D. Dolle. S. G. Exp.. S. Ley. Stoll. Pinilla. 55. Wiley. Dolle. D. T . 132 (1996). T . Seneci. Combinatorial Chemistry and Technology: Principles. 1021 (1996).. ESCOM Science Publishers. 1997. Koblan. 53.7487. D. C .. A.2288 (1996). Comb. 35. Plant Microbe Interact. Int. vols. E. Wilson and A. 1997. Blackie Academic and Professional.477 (2001). 44. Hunden. J. and J. J. M. Blondelle. Eur. 2002. 174 (1995). Nelson Jr. L. Muir. Res. R. Combinatorial Chemistry. Ed. Methods and Applications. 2000. D. Knight. 50.31306 (1996). E. J . 171 (1998). R. et al. B. Biochemistry. Combinatorial Chemistry: Synthesis and Application. Mol. L. N. P. A. Mol. E. C. 144 (1996). H. Lucks. Tetrahedron.. Hermkens. G. New York. Mol. J. 1. New York. Divers. Ellman. 45. 46. Lohner. D. Wiley-VCH. S. Med. J. 42.. 63.. 57. 1. A. P. R. A. and D. J. 2 . vol. K. 1.233 (1998). Chem. Dekker. Houghten. Bitan. 1998. E. Mol. Chem. L. 41. R. Gonzalez. Biol. Analytical Methods in Combinatorial Chemistry. A. J.. Chem. Rev. K. and T . Blondelle. J. Structure-Based Drug Design. R. Biochem. 23. J. S. W . Muller. Wiley-VCH. 25... Houghton. Rees. A. Barrett. Lansky. Ottenheijm. J. J. Dorner. London. Gordon. Med. Wiley-VCH. New York. E. Chem.. D. 47. M.. Med. H. C. Doenvald. New York. and D. J. 1999.537 (1997). and D. Blondelle and K. Chem. M. Ayers. E. Pinilla. Org. E. 8135 (1995). S. Chem. 33. 5085 (2000). Bream. Zechel. 64. A.. J. J. C. 49. Dooley. Elolman. K. D. J. C .. P. Educ. Chem.. 121 (1996).11035 (1993). Biochemistry.. Ed. E. C. Hamilton.. 37. English. 29. Baxendale. B. A. Koppel. Fodor. R. Chem.References 22. R. 52. H. Wallace. Med. Marquardt. 2000. 48. N. Leiden. Smith. Biol. N. J. M.. J. S. 4 1 (1996). Brinker. Houchins. R. Jung. .

Comb..2239 (2000). 71.. Comb. 97.. S. M. S. 8. R. Shan. A. M. Nature. Orlowski. Andrus. E. 7. Y. P.433 (2000). Davis. Biotechniques. . Bioorg. Pharmacol. Chem. T. Quibell. Boatman. Jung. Tan. Acad. 99. S. M. G. Bioorg. Hodges. Roughton. L. Cohen.. 66. C. Egan. M. A. and S. J.. Kennedy. Salmon. G. Wipf. J. Urban. Proc. Comb. Nature.. Frankel. M. Brady. Lynas. Blake. A. T. K. Epstein.487 (1991). Dooley. J. 2.. Chem. W.. 83.11887 (1997).1011 (1998). 100. and G. Vo. 86. E. C. and I. E. Mohler. Ducruet. Pinung. Read. Z. 10. Sebestyen. M. Wipf. J. W. 77. Bioorg. Lam. K. 8. M. 42. K. Stoddard. Natl. Gong. 67. E. Pept. 92. Mosley. 282. L. C. Biochem. J. 737 (1998). Bioorg. Rossi. 65. Berk. Chem. J. 9. A. K. Boisclair. Ho and C. 72. Lindberg. D. 78.. R. Biorg.. and C.413 (1993). 75. S.. Bergnes.. Hersh. W. A. T. 354. Schunk. S. A. 76. E. and A. Pept.. P. S. Jacob-Samuel. D. Med. and S. J. Brock.2779 (1997). M. and Z. 378 (2000). Chem.1669 (1998). Bowman. K. and G. and K. J. 91.82 (1991). G. Broger. Gopalsamy and H. Med. C. Blanchard. Deal. Yang. C. Schullek. Bioorg. Ni. 0. M. 68. F. Chem. T. Kundu. Natl. Sci. L. Lett. M. Med. R. Appel. Davis. V. 173 (1998). Brady. J. R. Shea. and Q. Butler. P.650 (2000). Yokokawa. G.. and I. Chem. C. M. Ellman. H. S. H. Acad. V. L. K. Weber.401 (2000). Jin. and A. Pinilla. and J. Pal. Wells. Terstappen. Lett. Chem. et al. J. D. J. Med. Houghton. Boger. R. Choi-Sledeski. K.. C. Biophys. 88. Acad. . Pei. 2321 (1998). J. Jung. Barteling. Chem. Wasicak. S. Bioorg. K. Proc. 89. R. M. 293. E. A. and R. Qabar. 69. Enderle. Biol. Lett. Chem. Houghten. 70. M. C. W. Lett. S. M. 85. Bhandari. Med. 354. J.20 (1997). Zabel. 7.. Tucker. Witherup. J. Bianco. D.. and C. J. P. 10.. L. J.. Chem. 13. S. Tamura. Y.. J. P. Furka. C. J. Collins. M. H. and D. R. 2. H. Chem. J. Furka. 94. Med. Fodor. S. 81. Sci. et al. Wei. Turner.. Res. T. B. and D. J. 3998 (1984).37.. Farnell. J. Rizzo. Proc. A. Sun. 87.4844 (1999).. R. Chem. R. Lett. 81. A. 767 (1991). B. M. D. M. and G. Y. 705 (2000). P. Chen. A. M. 1517 (1999). 93. 716-731 (2000).412 (1992). L.. Ryder. Med. K.A. 65. V. Chem. A. J. Chaudhury. Hutchins.. C. W. 4973 (2000). Lett. M. 2309 (1998). Chem. N. C. Protein Res. E. 28759 (1998). Brady. Stryer. Lazo. Rice. Blondelle. K. S. S. Lucka. Cuervo. 94.. J. Med. Carroll. Kraas. Seneci. Chem. T. Chem. E. C. M. D. Sisko. J . Bioorg.475 (2000). D. Meara. R. Houghten. Li. J. B. Appel. R. USA. Med. R. A. Lett. Rohrer. Natl. Dolle. Birzin. 300 (1994). A. 98. 8. Chem. P. 1033 (2000). 246. Grabowska.2849 (1999). Becker.. and R. Sci. R. Bioorg. 8. T.. Jones.26589 (1998). Biol. Yokokawa. Chem. Med. Eichler. Lu. J. Knapp. Dibo. M. Bromme. F. 251. Meighan. Chem. C. Ogbu. Dotse. F. T. M. M. J. 10. Lazo and P. Chem. 101. 3. D.. S. and M. Chem. Cuervo. L. and G. Yi. Henry Jr. M. et al. C. F. 275. U. Chem. L. M. McKeever. 2. Gilliam. Schullek. H. Ryder. 96. R. M. Yuan. Hodges. Huntington.. Apletalina. N. J. Harada.2303 (1998). D. Kennedy. 2.34 Combinatorial Chemistry and Multiple Parallel Synthesis 84. Walden. Kovacs. and J. Martin. 8. J. USA. A. 95. Lett. S. Org. 4. Z. P.. Q. K. 8. Walker. Dal Cin. J. and D.. J. J. Zhao. et al. Bioorg. Lett. 90. A. 79. Pinilla. Med. Illgen. O'Brien. J. Roychoudhury.2309 (1998). Ewing. K. 41. Org. 82. D. T. 273. D. D. R. Lamango. J. Divers. Biol. F. Sebestyen.. Bauser. Kazmierski. et d . 273. Betschinger. 94. M. 74. C. S. 9. Blondelle. M. Blondelle. C. Appel. J. Lett. Anal. Jin. Bioorg. Med. C. Int. T. Biochem. B. W. Goldberg. R. M. S. J. S. Science. and K. 73. Riznik. D. Exp. L. and A. T. Commun. 1451 (2000). Reader. J.. Res. J. S. 1467 (2000). et al. Mitten. 80. Geysen. Walk. Dooley. Lee. R. Biorg.. Houghten. S. S. Warmus. Brady. Rabinowitz. T. Lumma Jr.. E. Tasker. USA. C. Y. J. Hruby. M. P. Ambudkar. J. and K. B. C. J. K.. Salvino. H. 1007 (2000). Meloen. H. Med. Solar.84 (1991). Cavallaro. J. J. L. K. Sauna. 65. and J. T. Lett. et al. Chem.. Asgedom. S. Med. Mol. Ther. Science. J. A. A. T. Comb. Dibo. J. Guo.

S. Biorg. C. J. M. 104. S. and R. S. R.4363 (1995). Proc. F. M. Krane. 4708 (1994). J.Y . Goff and R.450 (2001). R. Drug Discov.. Wyatt. Lett. et al. Denyssevych. A. Campbell. D.. Frank. Guinosso. A. Biorg. F. P. 131.. Acad. B. Chem. Sci.. M . S. C. Mayer. A. Zhang. Med. 6909 (1993). Narensingh. Rano. Chem. J. et al. Nicholson. Zuber. and E. Comb. E. M. H. L. 133. J. C. Singh. Murphy. Comb. C. Y u . Acad. 115. D. and M. 134. Revesz. Pavia. Hajduk. Biol. Vincent. Org. 266 (1996). Biol. Misra. P. 6. V . V . Krane.. Natl. Biorg. J. L. Manning. A. Commun. K. 270. 4. S.Allwine. H. R. Comb. 130. C.. Ro. al-Obeidi. C. E. B. T . J.. D.. D. Huening.. 90. Mitscher. McKinney. Chem. et al. and S. P. K. C. Med. J. J. Ellman. Kiesow. N. J. Gordeev. Ellman. C.Wilson-Lingardo. and V . Stankovic. 42. 123. D. A. Dev. 8. J. J. K. 36. Mason. 10. Dinges. A. C. S. Chem. D. Lett. Chem. K. et al. F. Tetrahedron Lett. Sarshar. Eur. Patel. S. 2. Acad. Dive. T . G. R. 6. et al. 21. J . Look. J. 112. A. J. A. Chem. Proc. Cunningham. Chem. F. Sci. 136. Jonnalagadda. Cone. Davis. Ramig. T ... R. Chem. D. Med. Chem..2599 (2000). 107. et al. 7706 (1998). V . Org.513 (2000). D. and J. S.Y . A. Ellman. Natl. Sarshar. Safarova. Lee. Cheng. F. S. Syn. C. Frank. 89 (1998). E. 10. Biochemistry. et al. 10. Curr. Ravikumar. Kaldor. A. 108. and J. Ostrem.. D. Bhushan. 8. Kiely. Jiracek. et al. 118. P. J. D. J. N. C.W . Batman... Souers. DeWitt. Bonne. et al. Med. A. C. E. A.. J. Bioorg. Lett. Nicoalau. J. Schroeder. Abeles.. R. Zuckermann. Markworth. W . Acad. B.. Bioorg. 1527 (2000). Lett. Gordon. Biorg. 27.. W .2321 (1998). Gallop.. Dive. Med. et al. J. A. E. M. T . 42. Burow. A.. 162 (2002). L. et al. J . 271.. Grega. S. K. Cotton.. Yiotakis. 60. et al. 117. 131. 1. B. M. Vassiliou. Safar. Hebert. 60. J. Med. M . Med. Drug Metab.4330 (1999). Wipf. 559 (1999). Haskell-Luevano. C. J. 1637 (2000). High Throughput Screen. USA. and M. Rice. USA. 122. Shibata. A. Med. S.-K. K. Zhang. S. Shin. Rewerts. Paikoff. and E. A. R. Med. Natl. C. J..Yiotakis. Katz. S. A. M. Med. Billedeau. Med. Patek. C. 121. Rodarte. L. K. 3243 (2001). J.. R. J... and R. R. Labaudiniere. DeWitt. R. A. J. Dixon. T . Ogbu. 139. E. J.Vickers. P. M . Keschavarz-Shokri. 8. E. Bioorg. Vincent. Martin. Cody. 19606 (1996). Salvino. Janowick. 109. Monatsh. 106. M. 38. Stringer... Chemia. Biorg...References 102. 129. Chem.. Proc. Salvino. A. USA. Lawley. Natl. et al. and L. Chem. C. R. R. Chem. et al. Med. J. M. Med. and J. J. Chem. and R. Yiotakis. A. 143 (1998). Herpin. Pure Appl. 1665 (1998). Chem. Saxena.. J. Chem. F. 4. Urban. A. Youngquist. 60. A. Devasthale. 103. 96.. Pharmacokinet. USA. Gentry. Gentry. B. . Med. Schleif. 167 (1991). F. J. and M. 0. M. M... Chem. Wagner. H. and J. E. L. G. J. T. 6006 (1995). J. C. Boojamara. 111. M. K. 138. 21701 (1995). 149 (1997). Chem. A. et d. Chem.M. Ramage. K. M. L. Comb.. 126.5742 (1995). Sci.M. 132. A. P. Dolak.. 71. Schoenen. Bioorg. G. Combs. 92. M.937 (2000). Y . Chem. M. T . C. Abbot. M. Ksaminski. J. 119. Biorg. J. H. T . and R. M.. A.4380-4387 (1999). 128.405 (1998). Lett. J. Biorg. S. Gordeev. Ellman. F. 119 (2002). Mathew. Wagner. Schkeryantz. Chem. J. Bally. J. Chem. Timkey. Sci. S. Beresini. Bioorg. L. Checler.. A. Proc. Org. Chem. D. R. Kidwai. A.. 114.. H. 120. S. 5. Cho. 120. 124. MacDonald. Plunkett and J. 125. 9. W . D. 3041 (1996). C. J. Qabar. 127. U .2603 (2000). Rotonda. P. Moran. and M . 105. High Throughput Screen. J.. R. Lecoq. T . J. R.. D. Meara. E. Lett. 10. Tang. S. V a n Kirk. S.3852 (1999). 91. 37. Biol. 1. P. S. Lett. 165 (1997). Jiracek. S. A. Barbachyn. R. T. S. E. C. J. Chem. Bunin. England. Chem.. Chem. A. J. Chem. Maeji.Med. J. Michaud.. V . Lett. 50. Rodarte. 10. 135. K. N. Guns. Fritz. J. Opin. S.883 (1998). J. 137. Rano.. 1053 (1998). P. Am. Lazo. Kim. et al. A. 116. Bastos. J. A. 113. 2513 (2000). Chem. 110. Soc. Dankwaardt. Rosenquist. J. Benbatoul. P. Zhang. Khong. Bergmeier and S. K. Peterson. Fecik. Lett. Chem. J. Plunkett. Moran. Landi and K.5744 (1995).. 4. M. J. 2937 (1995). T . F.6738 (1995).

Med. 2000. A. Jin. M. Jackson. P. J. Guo.716(2000). J. L. Albericio. et al. Sandri. Dekker. D. 4.Cork and N. 51. J. D.S. Coe. N. 141. Inc. 19. New York. H.7076(1997). Lew. Med.97 143. J. and D. 291. S. Curran. Lett. and S..713(1986). Piznik. Baxendale.. and A. Leach.5560(1994). 146. . G. I. Ley.A. R. Drewry. Org.Orena. 3815(2000). Q. Perkin Trans. Bream. Furukawa. Soc. Hird. Oderaotoshi. D. Luo. Rev.. M. Poon.36 Combinatorial Chemistry and Multiple Parallel Synthesis 140.. Today. Eds. M.. Chamberlin. P.V. et al. (1999).95 (2002). 2233(2002). J. Flynn. Comb.Dolle. P. R..408 145. 147. Chem. Y. Z. Solid-Phase Synthesis: A Practical Guide.. 151. Kvatzik.Kates and F..1766(2001).. Chem. and S. Curran and T. Science. 4. A. Org. D. SOC. Org. 144. 149. S. R. Armstrong. McMartin. 19. P. Zhang. S. J. Res. Chem. Chem. Cardillo. 62. Brown and R... S. R. 2. Chem. I.56 (2002). G.. 116. Am. D. Res. Drug Discov. Rev. Hart. 150. J. Chem. 142. Bohacek and C. D. 0. O'Brien.. E. M . W. Comb. 148. Y. (1999). M.L. R. 7.

2.2 Reporter Gene Assays.1 Homogeneous Radioisotopic Assays. 59 3.1 Following the Competition.4. 54 3. 41 3. 43 3.3.3 Homogeneous and Non-Homogeneous Biochemical Assays.2.2. 59 37 Burger's Medicinal Chemistry and Drug Discovery Sixth Edition. .4.3.3.CHAPTER TWO High-Throughput Screening for Lead ~iscovery JOHN G.1 Gene Function by Homology to Other Defined Genes.6 Screen Validation and Reagent Scale UP. Profiling. 52 3.1 Target Discovery and Validation.1. Abraham ISBN 0-471-37028-2 O 2003 John Wiley & Sons.3. 54 3. 44 3. 38 3 Lead Discoverv " Process.HOUSTON &TYN N.1.2 Assay Construction.1 Assay Design.2 Gene Function by Gene Subtraction.1. Connecticut Contents 1 Introduction. Inc.4 Cellular Assays. 54 3.4.2 Systems-Based or Focused Discovery.2 Homogeneous Non-Radioisotopic k Y % 47 3.2 Profiling Hits.4.2. 45 3. 53 3.2 Bioassay Design and Screen Construction.3 Gene Function by Expression Analysis. 50 3.5 Alternate High-Throughput Screening Techniques.1 Analyzing Screening Hits.1 Cell Proliferation Assays. and Candidate Selection. 38 2 History of Lead Discovery and Screening.2. 43 3. 42 3. 43 3. 50 3. Volume 2: Drug Development Edited by Donald J.2. 49 3.4 Hit Identification.2. 46 3.3 Constructing Compound Decks and Screening for Hits. 56 3. 59 3.2. 46 3.2.3 High-Throughput Screening.3.2. 43 3. BANKS Bristol-Myers Squibb Pharmaceutical Research Institute Wallingford.

Leading this wave of change has been the development and im~lementation of a series of innovative.4 Technology Infrastructure in High-Throughput Screening. This can be most clearly seen in the upstream phase of the R&D process starting from target discovery through early preclinical testing. this goal is set against a historical track record of less than one NCE per year. and incremental progression in chemistry to develop drugs. per company. 66 5 Summary. From penicillin in 1929 (3) to the p-blockers and Bhydroxytryptamine (5HT) antagonists of the 1970s (4). cost efficiency. The drive to have more innovative and safer drugs on the market. 2). 62 4. 2).Z). Drug candidate pipelines are usually aimed at disease areas with high unmet medical need where the commercial reward for delivering first in class and/or best in class therapies can be enormous. 63 4. companies have also set challenging goals in reducing overall drug development time to market. culminating in lead optimization and pre-clinical candidate selection. innovative drug pipeline. The R&D process is an extremely lengthy. and speed of discovery has created radical changes in the way drug discovery is carried out. the promise of major improvements in productivity. Those companies that find the quickest way to put the first and best in class products on the market in a cost-efficient way will become the clear commercial winners in such a competi- tive arena. However. pharmaceutical companies have relied on a mix of "accidental" serendipitous discovery.1 Automation in High-Throughput Screening.2 Miniaturization of Screening Assays. In this chapter. - 2 HISTORY OF LEAD DISCOVERY AND SCREENING Serendipity has been the key to the pharmaceutical industry's success over many decades. we will describe the history and process of lead discovery from initial target identification and disease validation through the lead identification process itself. An ever increasing regulatory burden and strong emphasis on drug safety ensures a series of extremely tough hurdles before any increased flow of drugs works its way through the industry's pipelines onto the market. . The last 5 years has witnessed a dramatic change in the landscape for the pharmaceutical industry. . of which less than 10% reached sales of greater than $350 million (1. Consolidation in the market place resulting from a series of megamergers has created bigger than ever R&D budgets along with matching expectations and goals for increased productivity and growth (1. and bioinfonnatics has created a new drug discovery paradigm. state-of-the-art technologies aimed at the early lead discovery process. Additionally. combinatorial chemistry. quicker than ever before. 67 1 INTRODUCTION Drug discovery is one of the most crucial components of the pharmaceutical industry's Research and Development (R&D)process and is the essential first step in the generation of any robust. compound management.high-throughput screening (HTS). It is in this area that the overall impad of technological advances in genomics. inspired scientific insight. Just keeping pace with double digit annual growth rates requires each of the top 10 pharmaceutical companies to launch at least five significant new chemical entities (NCEs) per year with blockbuster potential (greater than $500 million/year). in an ambitious attempt to exploit the maximum life of drug patents. and strongly regulated activity that has increasing attrition rates and escalating costs (1. computer-aided drug design (CADD). The full range of technologies and approaches that have been brought to bear on this complex activity will also be highlighted. However. complex. has fueled enormous interest in different ways of revolutionizing the process by which drug discovery and development is carried out.

plants have not been the only source for these medicines. Scientists such as Paul Ehrlich. In a move from serendipity to rationality. and digitalis (from foxglove leaf) also came many years. natural product drugs such as cyclosporin and lovastatin used for immunosuppression and hypocholesterolemia have also been isolated from the secondary metabolites of fungi and bacteria. Through the 1950s and 1960s. the early history of lead discovery is all about natural products and herbal remedies. the pharmaceutical industry relied heavily on chance discovery to find and develop drugs but made strong attempts to enhance this by developing a drug discovery process with the introduction of compound screening in a concerted fashion. although tedious and time consuming. and the digitalis alkaloids. Additionally. echinocandins. He found that penicillin mold could cause the lysis of staphylococcal colonies on agar plates (3). It was not until the late 18th and early 19th centuries that an analytical investigation of the active components of medicinal plants and herbal remedies was pursued.possessedactivity against pathogens such as Gram-positive bacteria and Gram-negative cocci. the use of which dates back thousands of years. Even today. natural products continue to be a viable source of new drugs for the pharmaceutical industry (12). sedation. By using specific animal tissues and developing a better understanding of the biological and physiological systems under investigation. (13). which eventually became the major constituents of many modern analgesic and cardiac medicines. along with African tribal shamen "cures. 11). and others (14). antibiotics such as the p-lactams. morphine. drug companies attempted to develop more useful tools to help facilitate the discovery process. the lack of in vitro assay technology meant that compounds had to be tested in animal models. The use of these treatments was passed down by each generation through traditional practice and word of mouth. Significantly. and penicillin and the p-lactam drugs emerged as the first of an incredible series of antibiotic medicines that have saved countless lives. It has been estimated that 119 compounds identified from 90 plants have been used as single entity medicinal agents (10). In addition. and papaverine. and dropsy (5-9). the wonderful therapeutic efficacy of penicillins was demonstrated by Chain et al. the belladonna alkaloids. did generate a remarkable number of drugs. have exemplified the idea of serial screening of compounds in an attempt to find a "magic bullet. The discovery of aspirin (from Willow bark)." This approach. In the 1940s. in some cases thousands of years. searching for an antisyphilitic. Ancient Chinese and Indian medicinal recipes. chlorpromazine. and the erythromycin macrolides all originated from microbial and fungal metabolites (5). The resulting high failure rate in these test models reduced considerably the chances of finding a molecule that could then be optimized by chemists." have included ingredients such as opium.2 History of Lead Discovery and Screening The emergence of this "accidental" approach to drug discovery has its origins in early history with traditional natural herbal remedies that were passed from generation to generation in local communities or tribes. scientists were able to piece together a far more detailed picture of the molecular parameters that impact efficacy and affinity. after these remedies were being used for pain relief. curare. The culture filtrate also. Compounds found active in these animal models usually had an unknown mechanism of action. The discovery of penicillin in 1929 by Alexander Fleming was one of the first real documented "chance" observations that eventually led to the development of a drug. the vast majority of these drugs were obtained as a result of examining the plant based on its ethnomedical use (6. making future improvements on the drug very difficult. Of course. codeine. the cannabinoids (from cannabis sativa). The analysis of the active ingredients of these herbal "cures" has supplied an enormous number of the medicines that have been used in the last 100 years. 9. This resulted in the discovery of alkaloids such as atropine. In fact. 7. The biggest limitation to this discovery approach was the relatively small number of compounds available for screening (a few hundred) and their lack of chemical diversity. . such as the benzodiazepines.

This approach heralded a golden period of drug discovery and development that produced histamine H2 receptor antagonists. where each target would be screened for 2-3 years. fully automated screening systems into pharmaceutical organizations. By the early 1990s. The range of highly specific molecules that became available also helped define a variety of new receptor subtypes that also became candidates for drug intervention. Rational design of drugs based on knowledge of the biological system being investigated allowed highly specific selective antagonists and agonists to be developed. and combinatorial chemistry libraries. The manual.000 samples through each of its 15-20 targets per year (15). A typical lead discovery organization at that time would have been expected to screen approximately 10-20. potency and other druglike properties could be "built in" to the lead structure in a coordinated. new compounds could be made based on the in vitro SAR and retested in the animal model. and the overall timelines shortened. some pharmaceutical companies quote that they have capacity to screen over 100 targets per year. with each screening campaign lasting only a matter of weeks. In this iterative process. the screening capacities rose significantly. (4) in the development of the H2 receptor antagonist cimetidine. labor-intensive process. The ability to clone and express human receptors and enzymes radically changed the way drug discovery was carried out. Concomitant advances in molecular biology. The modern screening laboratory today is a multifunctional. and partial agonists (14).High-Throughput Screening for Lead Discovery The development of in vitro assays using animal tissues became an essential sumort tool for tracking structure-activity relationships (SARI and allowing chemists to optimize structures before whole animal testing. integrated process producing a product that consists of a cohort of tractable . and microplate technology allowed biological targets to be screened that had proven to be intractable before. multiskilled environment that connects a range of discovery functions in a high capacity. HTS. advances in technologies such as lab automation. the pharmaceutical laboratory in the early 1990s was set for the first real version of highthroughput screening. in vitro bioassay design techniques. The next major change to the discovery process was seen with the first advances in molecular and cellular biology in the early 1980s. Increasing numbers of targets and compounds drove the demand for higher capacity. The integration of robotic systems and highly trained scientists with a diverse range of skill sets has allowed the development of the lead discovery process. detection systems. The emergence of functional genomics tools and bioinformatics also allowed us to identify and understand more fully the connections between specific human genes and disease. The final pieces of the jigsaw were completed with the strides made in automation of compound synthesis and the accumulation of large collections of natural and synthetic compounds in stores available for screening. liquid stores. and data capture systems allowed the first real automated versions of screening labs A to be put in place. This technology allowed optimization of lead molecules against the human version of a receptor or enzyme and allowed a deeper analysis of their physiological nature. Now in the early 2000s. Rational drug design and the pharmacological elucidation of receptor pathways were the preeminent methodology for drug discovery during the 1960s and 1970s. With all these essential technology ingredients in place. p-adrenergic receptor antagonists. This is now called ultra-high-throughput screening (UHTS) and represents a significant increase in testing capacity compared with laboratories of the late 1980s (15). planned fashion and tested. each testing over 1 million samples per target. If the lead compound subsequently failed in animal testing. These molecules could then be developed as drug candidates such as was demonstrated by Black et al. used in the 1970s and 1980s to screen compounds in low throughput gave way to new processes in the 1990s. The majority of pharmaceutical companies (16-18) have now reported applications of automation and robotics in activities such as dry compound storage and retrieval. Throughout the 1990s. This allowed a wide range of compounds to be made around a SAR hypothesis.

the initial discovery of drug candidates can be described as somewhat ad hoe and ill defined.1). Therefore. which cannot be fully captured in a chapter such as this. A prototypical early drug discovery process and the key activities in lead discovery. the array of ancillary technologies. complex process with numerous iterative sequences and feedback loops. scalable lead discovery in the modern pharmaceutical industry (Fig. chemical leads against targets of interest. in this prototypical early discove r y process. In contrast to the highly defined and strictly regulated process of drug development and manufacture. The following sections of this chapter describe the generic process of lead discovery and are followed by a more detailed analysis of some of the important functions involved in its execution.1. and processes as the central theme throughout the following section. 3 LEAD DISCOVERY PROCESS but overlapping activities that are essential for effective. has developed a surprisingly wide array of technologies and approaches to achieve the same endpoint (15-21).3 Lead Discovery Process Figure 2. However. Each pharmaceutical company. Each component of the lead discovery process is critical to the overall success and impact of a screening campaign. The disease relevance and "druggability" of the biological target and the availability of technologies to supply and test vast arrays of compounds are key factors for success. Throughout the chapter. the drug discovery process is a multi-faceted. we will keep highthroughput screening. although agreeing on the ultimate goal. 2. we will endeavor to highlight the process by which these distinct activities are brought together in an integrated process. Discovering and validating a target of interest Designing a bioassay to measure biological activity Constructing a high-throughput screen Selecting screening decks and screening to find hits Profiling hits and selecting candidates for optimization Of course. we can describe several distinct .

essential processes for signaling within and between cells and their perturbation in disease states has been the basic approach for establishing potential targets suitable for drug intervention. The elucidation of gene function and the understanding of its role in the activity of other genes can be critically important. it is hoped that particular disease mechanisms and their ensuing pathologies can be modified. There are now billions of base pair DNA sequence data and an estimated 30. The ability to search the billions of data points available through the various human genomerelated databases has rapidly evolved with the development of bioinformatics tools. Clearly. are helping to identify thousands of potential targets. bioinformatics (22). Cloning. An essential next step in the process is attributing functionality to an unknown gene. but provides useful circumstantial evidence that this target could be worth investigating in disease models. one of the more important steps in the process of developing a novel pipeline of drugs is the identification of novel genes and assessing their expression and role in various physiological and pathological states.The emphasis has now changed from finding new drugs that exploit the same wellvalidated targets to finding drugs against potentially new and innovative biological mechanisms connected to disease. .High-Throughput Screening for Lead Discovery 3. and distribution of the gene of interest are all-important next steps in qualifying the target for further discovery work. A host of genomics technologies are now available for discovery scientists to find and validate biological targets. differential mRNA display analysis. There are a number of critical methods for determining gene function some of that are similar to those used to initially identify genes. Information supporting the role of these targets in disease modulation can come from a variety of sources. For instance. Their DNA coding sequences can be aligned and compared with other known DNA sequences stored in databases to allow comparison with known genes. This type of approach has generated a number of significant and unique biological targets. Powerful computers and search algorithms allow potential genes to be identified within stretches of DNA sequence. Other methodologies can also be used to identify genes such as using phylogenetic analysis to place the novel gene in context with other genes by molecular evolution. Of course. brain-specific expression of a gene engages neuroscientists in the same way as specific expression of a novel gene in an immune cell would spike the attention of inflammation experts. differential expression of a candidate gene in disease and non-disease systems helps to focus the interest considerably. This is a significant change to past approaches where the majority of the medicines developed by the pharmaceutical industry had been targeted at only 500 known human targets (26).1 Target Discovery and Validation The drug discovery process starts with the identification. A gene that is uniquely expressed in a particular type of cell or tissue will trigger further evaluation. and to a lesser extent in the laboratories of pharmaceutical and biotechnology companies.and intercellular events. More recently. single nucleotide polymorphism (SNP) identification. Programs such as basic local alignment search tool (BLAST) are used for this type of similarity searching. expression. combined with the huge output from the human genome sequencing project. or growing evidence of. The tough questions that need to be answered at this stage of the process are related to which targets to focus on among the many thousands of potential choices. the targets have been researched and largely discovered in academic laboratories.000 human genes to be searched and investigated as potential drug discovery start points. In fact. The human genome project has generated a wealth of data around human DNA sequences and gene mapping. Traditionally. Rapid gene sequencing. genomics and proteomics techniques (23-25). and bioinformatics data mining tools are just some of the now standard techniques that are available to help identify and analyze novel genes. By pharmacologically modifying these intra. cell and tissue distribution does not solely validate a target. biological targets that are believed to be connected to a particular disease state or pathology. Basic research into understanding the fundamental.

Lexicon. and mice also have well-characterized phenotypes that will infer function. Certain animal models such as worms. If the intellectual property (IP) position on this target is also clear. By using this extensive array of genomics techniques. tissue-specific genes showing differential expression in disease and non-disease systems. the greater the chances are that the right type of targets will be selected. but no real clinical proof.3 Gene Function by Expression Analysis.3 Lead Discovery Process 3. Direct homology or knockout experimentation may have also elucidated their function. flies. then this target could rise to the top of the prioritized list for entry into the pipeline. target &d disease portfolio decision-making processes are not within the scope of this chapter. the ability to measure the activity and function of the biological target and find compounds that modulate the activity is critical.2 Bioassay Design and Screen Construction After the identification of a biological target of interest.1. By the end of this initial phase of the R&D process. Yeast models can also be used for this type of analysis. These targets may also be members of a known "druggable" target class such as GPCRs or ion channels. may keep you clear of the competition. These homologies can be as general as identifying the unknown gene as a target involved in metabolism or the homology could be as specific as identifying a specific function such as fatty acid biosynthesis (27-28). 3. This type of analysis involves extensive use of DNA microarrays and related chip technologies (31-33]. When a gene function cannot be identified by direct homology methods or when you want further proof of the importance of a particular candidate gene.. Entering a biological target into the discovery process is an extremely critical step that can affect the successful discovery of leads and the ultimate success of the drug development process. By aligning the sequence of an unknown gene and comparingit with sequences ofknown genes. but clearly play a direct role in the potential success or failure of any lead discovery program. Sequitur). the next challenge begins with the conversion of the target into a bioassay that can give a readout of biological activity. it is possible to determine function by deleting the gene (gene knockout) in an in vivo model and describing the resulting phenotype (29). The greater the knowledge of the biological systems and disease pathologies involved. from enzymes and receptors to cellular systems that represent an entire biochemical pathway or a disease process. homologies are identified that allow a tentative potential function to be assigned. The fact that a target is known and well validated reduces the discovery and development risk but raises the chances that your competitors are also working on this target. Selecting a novel target with just the earliest indications of validation. At this next stage. Consequently. Building a prioritized portfolio of disease targets and subsequently deploying resources is critical in the discovery process. which suggests they can be readily integrated into a drug discovery process (34). Overall. 3. Which disease areas are the foci for the organization? In those disease areas of interest. Tracking mRNA and protein expression as different indicators of gene activity in normal and disease tissue is now one of the more important methods for gene validation.2 Gene Function by Gene Subtraction. An example would be the use of gene knockout studies to establish whether protein tyrosine phosphatase IB is an important target for anti-diabetic drugs (30). the discovery scientist can quickly generate a list of novel. The range of potential targets is large. a list of targets will have been selected and passed into the lead discovery phase. the range of as- .I Gene Function by Homology to Other Defined Genes. what target choices are available for drug intervention? What is known about the relationship of this target to disease? All the answers to these questions and more determine the ranking order by which targets may be worked on or not. Genes specific for a particular tissue can be expressed differently under a variety of different metabolic conditions or stresses within the cell or organism.g. 3. A number of biotechnology companies offer this as a valuable service both for in vivo and in vitro systems (e.1 .1. but raises the risk of discovery and development failure as the target and drug develop through the pipeline.

If the bioassay is the basis of a highthroughput screen. or do nothing to this activity. fluorescence and fluorometers. Does the proposed HTS assay provide the desired information to progress compounds along the drug discovery pipeline? Will the hits need to be assessed in additional assays before progression into medicinal chemistry? For example. cell death. These could include criteria that would initiate a medicinal chemistry program. The process from bioassay design to HTS construction can significantly change the design and final format of an assay. What compound characteristics are required. etc. Ideally.g. then compounds can be tested in the bioassay to see if they inhibit. receptor antagonism and agonism.2. 3. a team of medicinal chemists. Are there any restrictions on the use of the assay technology? For example. are there patent restrictions on the use of any reagents or the actual biochemical technique itself.. The HTS assay designer's responsibility is to select the correct assay methodology to meet the screening objectives. If at all possible you would like to avoid de novo assay design unless there is potential for more targets within this class and the investment can have future impact. and it is often necessary to create a decision matrix to compare the potential methods. especially if one is considering screening a gene family like kinases or nuclear hormone receptors. In this assay optimization process. A variety of techniques can be applied to bioassay design depending on the nature of the biological activity being measured. or voltage changes and patch clamp. if the screen needs to find receptor agonists. K.. by some direct or indirect means. e. Each of these bioactivities can be directly or indirectly measured using reporter signals and detection systems such as radioisotopes and scintillation counters. The stability and scalability of the assay's core reagents are just two of the more important factors. A typical matrix of questions is outlined below. Here the reagents have to be made at bulk levels to allow testing of thousands of compounds in conditions where they may have to be stable for hours on an automated screening system. Once an assay has been developed that measures the biological activity of the target. a definition on how to assess the value of the screening hits. Usually there are numerous potential assay options. the assay will be optimized against an array of multifactorial parameters to provide an in vitro milieu for the bioassay. there are potential cellular functional assays that directly measure in vitro agonist . Measurement of product accumulation. It is important that assay to screen reproducibility is monitored and maintained to ensure consistent results from the different laboratories involved in the discovery process. then other crucial factors are assessed before screen construction. Do you have the relevant expertise available to build the type of assay? Modification of an existing screening protocol is the natural starting point. and HTS designers define the important objectives for a highthroughput screening campaign. which most closely resembles the physiological "normal" for the functional activity. 3.High-Throughput Screening for Lead Discovery say design techniques and types of assay available have to be correspondingly comprehensive. Regardless of the biological activity being measured and the detection system being used. in the plan- ning phase. luminescence and luminometers. such as the desired data from the screen. An assay format designed to use scarce and time-dependent labile reagents will not make the best design for a HTS environment. antagonists or agonist? 2. and a whole host of other significant parameters are modified to provide the most appropriate bioassay. and cell proliferation are just a few of the activities that could be assessed in a bioassay. 4. This approach is the basis of all compound testing for HTS and structure-activity relationship (SAR) studies in drug discovery programs. enhance. 1 . measurement of receptor mediated signaling.1 Assay Design. K.. therapeutic area biologists. measurement of enzyme substrate use. factors such as pH.

the reagents are added to a microtiter plate already containing compound and the results measured after a suitable incubation period. Is the assay method suitable for miniaturization andlor running on an automated screening system? The decision on whether to use automated screening platforms and miniaturized screening formats will be very dependent on the existing screening infrastructure. etc. It is also the most critical part of the high-throughput screening process. The majority of HTS assay designs are adapted from existing protocols. and a process for assessing the screening hits.. ionic ~ strength ~ 2. Designing and building a HTS assay can be very time consuming and is often a rate-limiting step in the screening process. Incubation conditions (defined. highly colored compounds can quench the signal from a fluorescencebased assay. and stabilization additives 3. and a clear understanding of these is critical to the interpretation of screening results. homogeneous assay formats are preferred in HTS because they can be readily adapted for automation. giving the illusion of inhibition in an enzyme assay. then the value of the results will be questionable. suboptimal substrate concentration in an enzyme assay). In a homogeneous HTS assay. room temperature. if reagents are difficult to produce then the amounts required to run a full deck HTS may indicate that a design applicable to miniaturization should have a higher priority. This is discussed later in the chapter. there are many options available to measure the biological effect of a compound or mixture of compounds in a HTS assay (35-37). A secondary assay would then be needed to distinguish between these two binding activities. proteinaggrega1. Often the existing method is not ideal for high-throughput screening. The optimization process evaluates all the important assay parameters.2 Assay Construction. Alternatively. Are all reagents readily available to perform a HTS? Will the reagents be purchased from an external vendor or produced in-house? These reagents include not only substrates and ligands.3 Lead Discovery Process 45 5. Generally. For example. Having decided on the best methodology option. non-contact piezo. fp ~and f . such as the following: The questions and criteria above are ranked in a generalized priority and can direct the assay designer to a particular design choice. syringe driven. type. but the assay design is flawed (e. 6.) 5. The planning phase should conclude with an agreed high-throughput screening assay design. 7. Additives to improve pipetting accuracy and precision 4. both strengths and weaknesses. For a particular target.g. stability.2. a different priority ranking may be more relevant de~endinc on the screen circumstances. controlled environment. medicinal chemistry may want to modify antagonists to generate agonist activity and the binding assay route would then be preferable. or modifications of assays built in therapeutic area laboratories. Examples of these homogeneous assays are described later in the chapter. and the assay will require significant optimization. 3. Is there experience in screening this type of target with the proposed assay technology? Each screening technology has its own challenges. literature methods. etc. Pipetting method (peristaltic pumps. If the probability of finding an agonist is perceived to be low. the construction phase involves the building and optimization of the assay parameters. Are there simple ways to detect and eliminate false positive data from the HTS results? Certain compounds will interfere with the readout in the HTS assay. For a given target. efficacy. 8. If the screening campaign is perfectly executed and generates hits. a clear understanding of how the reagents will be procured. Reagent purity.) . but also the enzymes/receptors/cells. ~ ~ . for example. a binding assay could be designed that would find both agonists and antagonists without discriminating between them.~ u tion.

they describe how one combines the experimental design statistics with automated liquid handling devices to optimize assay conditions. followed by a series of experiments varying other assay parameters until acceptable conditions are achieved.a plastic bead containing a fluor and coated with a lectin binds the carbohydrate extracellular . V . non-homogenous assays that involve a separation or washing step(s) are still used in several high-throughput screening approaches. These combination factor contributions to an assay signal can not be explained by the sum of their individual contributions. The interested reader is directed to a more recent publication by Haaland (38) that illustrates numerous examples of the application of experimental design. The assay signal is a composite of both contributions from individual factors as well as combinations of factors. optimized assay conditions are often a complex interaction of all the variables in a method and difficult to derive by monitoring each individual component. In cellular assays. like experimental design. is assay efficiency. membranes containing the receptor are incubated with radioactive ligand in the presence of compound in a 96-well microtiter plate. and the effect of compounds measured without the need to include a separation stage. cell number. If the compound has blocked binding to the receptor. Kinetic parameters. to optimize multivariate assay conditions. are the most commonly used and effective formats for achieving high precision. However. The advantages and disadvantages of these two approaches have been extensively discussed elsewhere (41. 7. In scintillation proximity assays (SPA). the assay must show acceptable kinetics and respond in a predictable way to known inhibitors or activators. fluorescence. absorbance.High-Throughput Screening for Lead Discovery 6. incubated. there are numerous detection techniques.3. where reagents are added. Historically. The majority of The objective of the optimization phase is to build an assay that generates a reproducible signal that will allow a statistically significant difference from the background reading. or binding constants K. this optimization process has tended to be a linear process where one or two reagents are optimized. Both biochemical and cellular screening have been automated and miniaturized. One of the common critical features of a high-throughput screening method. Specifically with HTS. Additionally..1 Homogeneous Radioisotopic Assays.2. The seminal work in this area was developed by the British biologist1 statistician Sir Ronald Fisher over 40 years ago. and passage number are just some of the potential optimization criteria tions have advantages and disadvantages. and the critical step in the assay design process is to decide what information is required and match the goal to the available options. An alternative homogeneous assay approach would be to use scintillation proximity assay (SPA) (43). the filter traps the receptor-Iigand complex. including radiochemical. In Taylor et al. This signal has to be precise. However. plating conditions. It is however possible to use statistical techniques. In a non-homogeneous format. 3.42). The major different HTS assays can be conveniently classified into biochemical techniques or cellular techniques. The reaction mixture is then transferred to a second microtiter plate that contains a filter.2. K. bound ligand is measured in a conventional scintillation counter. lower levels of radioactivity are measured by the scintillation counter. type of media. A good illustration to compare a homogeneous and non-homogeneous format is an assay that is designed to measure the interaction of a receptor with its ligand using a radioisotopic detection method. two publications detail how these statistical techniques have been used to optimize assay conditions (39. Homogeneous assays. be it biochemical or cellular. whilst maintaining throughput. kc. All of these combina- high-throughput screening detects the activity of a series of compounds using biochemical or cellular approaches. (39).40). and luminescence. and most HTS laboratories maintain capabilities for both approaches.. 3. After washing the filter to remove unbound ligand.3 Homogeneous and Non-Hornogeneous Biochemical Assays. which means that it is not easy to design assays by the traditional linear approach. Under vacuum.

2.. /3 particles (3H) 01 :Auger electrons (1251) excite the embedded f l luor in the bead and the emitted light is meaSIred. Second. such as a protein. Non-radioisotopic homogeneous methods . perpendicular to the first filter. a red shifted dye with absorbance and emission >500 nm would be preferable. there are significant numbers of compounds in a screening deck that absorb at 400-450 nm.. Only bound radioligand is physically close enough to the fluor-impregnated plastic bead to cause light to be emitted. Two of the more popular approaches are described below in more detail. Direct )lorimetric-or absorbance-based assays have ?en in use in HTS laboratories for many ?arsand are still very popular and effective.Lead Discovery Process 47 Bound radioligand * 0 Inhibitor Light 620 nm Figure 2. FP is a technique that is popular in highthroughput screening laboratories because of its homogenous format that allows the formation of a stable equilibrium and diffusion controlled kinetics. After a suita 1 sle incubation period to allow equilibrium to :cur. the ligand needs to be conjugated with a fluorescent dye. and the energy diissipates through the aqueous solution. During isotopic decay. In FP." Recently the SPA bead has been mlodified such that the emitted light is redsllifted reducing this type of interference. first through a polarized emission filter and then through a second filter. four are added to a microtiter plate C( ~mponents Well. ~dthe interested reader is directed to a used high-throughput screening web site. ligand. This type of inference will be dependent on the type of fluorophore. the emitted light is measured.3). the free fluorescent ligand is differentiated from the bound fluorescent ligand based on rotational differences. fluorescence polarization (FP) :4). One of the signifit. membrane. and the bound fluorophore will rotate very little during fluorescence lifetime.2. 2.net. From a HTS perspective. fluorescence correlation spectroscopy PCS) (45). www. For small molecular weight ligands. various considerations should be taken into account before choosing a FP assay. SPA bead. Fluorescent ligand is excited with polarized light and the emission is measured. The intensity of bound fluorescent ligand. where a variety of technology companies describe their methodology. Consemembranes containing a receptor of q'~ently. The presence of colored compounds in the screening deck could lead to bioassay interference. In a O( SCyeen designed to find receptor antagonists. Ligand in solution is too distant from tlle bead to excite the fluor. e. therefore. SIlrface ' 3. signifi- . ir: ~terest can be bound to the surface of the b c?ad. icant disadvantages of SPA is the absorption oi light by colored assay components or "( 2uench. e.g. now at the two polarized emissions. Therefore.3.2 Homogeneous Non-Radioisotopic Aslys. and fluorescence sonance energy transfer (FRET) (50). of a membrane fragment.g. the emitted light is proportional to b csund ligand (see Fig.htscreening. If the ligand is of a relatively low molecular weight. Bind this fluorescent ligand to a macromolecule. and the rotation relaxation time significantly increases. Typically. will be different. Leadseeker. The principle behind \ N o light emitted from free ligand scintillation proximity assays.e also widely used in high-throughput :reening.2). large number of techniques are now availde that support high-throughput screening. 2. and the C()mpound under investigation. Radioactive ligand binds to membrane rcxeptor-bead complex at the surface of the bt?ad. the fluorescent intensity in both planes is essentially identical because ligand rotation will occur during the fluorescence lifetime. This phenomenon then becomes the basis of the high-throughput screening assay (Fig. homogeneous time-resolved flu:ometry (HTRF) (46-49). In summary. a rc?ductionin light is proportional to the activy of the compound tested.

. Excitation at 337 nm and emission at 665 nm measures the FRET.. FP has been successfully used in a range of assays. ... and fluorescence quench... (b) Ligand bound to a macromolecule will show polarization. An antibody that specifically recognizes phosphotyrosine. . Figure 2.. (a) Ligand in free solution does not display fluorescence polarization. These properties have been commercially exploited to produce high-throughput screening assays.. Europium (Eu3+). : " Excitation E. Time gating between excitation and emission reduces intrinsic fluorescence from the signal. of 300 nm. if the fluorophore on the ligand can still rotate freely even on binding a macromolecule (the propeller effect).. SH3-SH2 binding interactions (51).. However. As described cant experimentation will be required to give accurate binding characteristics. Fluorescence is a very versatile detection mode in high-throughput screening assays because of its wide dynamic range and its potential sensitivity... benzodiazepine receptors ( 5 3 .4 gives an example of a hTRF assay for a tyrosine kinase. Eu3+ is a poor fluorophore and must be complexed in an organic framework that facilitates fluorescence.mission .. allophycocyanin. Fluorescent assay components have significantly shorter Stokes' shifts. a poor polarization signal will result. The free Eu-cryptate will emit light at 620 nm and therefore can be discriminated from the FRET emission at 665 nm by wavelength filters. Eu-cryptates have a relatively long fluorescence lifetime (>0. = 10-15 M).1 ms) compared with typical assay components (<50 ns).. The phosphotyrosyl product is directly quantified using hTRF. Often a fluorescently labeled ligand can be prepared that will compete with a radioligand in a predictable way. Time resolved fluorometry (TRF) is one technique that can minimize these types of interference. Figure 2. The Eucryptates also have a relatively large Stokes' 7 ... Streptavidin.. is covalently conjugated to allophycocyanin. This figure shows the theory behind fluorescence polarization-based assays. and HIV-1 protease inhibitors (53). .g. a bacterial protein that has a strong affinity for biotin (K. tyrosine kinase will phosphorylate an artificial biotinylated peptide substrate. However. chelate (54) or cryptate (55). and wavelength discrimination reduces signal interference. This conjugate will bind the biotinylated phosphotyrosyl peptide. assay components can cause interference with the detection of the fluorescent signal because of intrinsic fluorescence from the reagents. . forming the acceptor partner in the FRET pair. but the fluorescence emission is short lived (a few nanoseconds). Free streptavidin-allophycocyanin will emit light at 665 nm..3. e.. . This technique is based on fluorescence resonance energy transfer (FRET) from an Eu-cryptate to a red algal protein. the difference between the emission and excited wavelengths. forms the donor partner in the FRET pair.. conjugated to Eu-cryptate. .. This methodology is based on the fluorescence of certain lanthanide metal ions.. This assay is homogeneous because the FRET is measured in the presence of all the other reagents.... Bound no rotation polarized Free free rotation depolarized shift. ......Under typical biochemical assay conditions.. light absorption.. Nevertheless.. . for example. This can be discriminated from the FRET signal by time gating... .High-Throughput Screening for Lead Discovery Free rotation (b) & . In the presence of ATP and cofactors. for example.....

g. APC-allophycocyanin. UK. and ACCELLERATOR RTS Thurnall. ATP. Figure 2. To a certain extent. The reaction product is then detected (b). In hTRF.3 Lead Discovery Process @ Biotin @ Phosphate Excitation 337 nm Emission 665 nm S-Streptavidin. They may also pre-select for compounds with drug-like properties that cross membranes. Eu-Europium cryptate bound to an antiphosphotyrosine antibody. Generating reproducible data over several months can be more challenging for cellular . These approaches minimize contamination and enable a highly precise process to be constructed. The reaction (a) uses a biotinylated peptide substrate that becomes phosphorylated on tyrosine in the presence of kinase.4. thereby potentially identifying new targets. one can use a cell-based "catch all" approach to mitigate against this uncertainty. but can also be their weakness. e. and magnesium ions. FRET-Fluorescence Resonance Energy Transfer. a significant disadvantage of cellbased high-throughput screening appears at the hit assessment stage where the molecular mechanism of a compound may remain ill defined. Cellular screens may also uncover hitherto unknown sites for chemical modulation. because there are multiple points along a network of signaling pathways where compounds can elicit a response. These types of assays stand in contrast to the defined. However. The schematic shows the basis of a tyrosine kinase hTRF assay. 3. CELLMATE & SelecT. requiring investment in generating and maintaining cell lines.2. Most cellular HTS campaigns also include a cytotoxicity assay to help "flag" potentially undesirable compounds. Manchester. Cellular assays are also popular screening formats for identifying lead compounds. light absorption by assay components is one of the limitations of prompt fluorescence assays. In the presence of colored compounds. single target biochemical assays described above. This requires additional selectivity and profiling assays to help elucidate the molecular target. the readout is based on the signal ratio of Eu-cryptate and FRET emissions. Cellular assays are generally more complex and time consuming than biochemical assays. Cambridge. UK. the two signals will decrease.. earlier in this section. This is a key strength of cellular assays. but the ratio remains constant. When the precise biochemical target is unknown.4 Cellular Assays. cellular production has been industrialized and new automation options originally designed for biopharmaceuticals are now being adopted by high-throughput screening laboratories. The Automation Partnership.

f example. cell proliferation is measured. which cascad down the protein kinase second messeng pathway to trigger the production of p-12 tamase. a promoter region.5.1 Cell Proliferation Assays. or by direct measurement. either by measuring its enzymatic activity. bacterial lipopolysaccharides. The reporter is usually quantified biochemically.4. A schematic of a p-lactamase report cell line that is coupled through a seven transmel brane G-protein-coupled receptor whose secoi messenger signaling pathway operates through c clic adenosine monophosphate (CAMP). a reporter indicates the presence or absence of a particular gene product that in turn reflects changes in a signal transduction pathway. For example. because of the potential variation in cell physiology that may occur with different passage number. Reporter gene assays are routinely used in highthroughput screening laboratories. however. The plasmid is stably transfected into cells and the cell clones selected. activates adenyl cyclase. with p-lactamase direct enzyme inhibition. like Alamar Blue (56).g.High-Throughput Screening for Lead Discove assays compared with biochemical assays. as with p-lactamase or luciferase.5 s h oRS a schematic representation of a seventransmembrane G-protein-coupled recept (GPCR) that signals through CAMP. human lymphocytes can be stimulated to proliferate through a variety of stimuli. The simplest format involves incubating cells in the presence of a suitable stimuli that causes the cells to proliferate and measuring the growth using a vital stain.2. transient transfection of cells can be performed just before assay. A unique fluorescent substra CCF2 (Fig.2. Typically. Because p-lactamase is not normar present in mammalian cells. followed by scale up for the high-throughput screen. Control cell lines that contain constitutive expression of the reporter are used to ~ assess the hits from a HTS. or radioactive isotope uptake and incorporation.6) is ideally suited to hig throughput screening assays. an agonist stimulates an increase in ti intracellular level of CAMP. e. In high-throughput screening. Inside the cell. esteras rapidly convert the esters into carboxylic i . Cell proliferation assays are quick and easily automated. the growth inhibition may just reflect cytotoxicity phenomena. Figure 2. 2. Alternatively. a reporter gene plasmid is transfected into a desired cell. quality of the data.4.. In tl: assay. and compounds that modulate proliferation are readily detected. 3. and a range of new cellular assay formats are increasingly being used for high-throughput screening.The G-pl tein. and other false positi situations. stable transfection is preferred because of the higher level of reproducibility achieved in readout. selectivity. increasing t: concentration of CAMP. the amount of e zyme produced is proportional to the effica of the agonist. such as cytokines. Variants on this theme have been used for many decades. As the name suggests. Gas. However. and vectors containing reporters are readily available commercially. oxygen sensors (57). This plasmid will contain a response element that is capable of initiating transcription. Numerous techniques are commonly used for high-throughput cellular assays. which is a critical factor in the overall PKA Figure 2. 3. which stimulates the PI duction of p-lactamase through CAMPresponse e: ment. green fluorescent protein (50).2 Reporter Gene Assays. By measuring the incorporation of radioactive thymidine into DNA. Esterified cr boxyl groups allow CCF2 diffusion across t cellular membrane. etc. and a coding region for the reporter. stable cell 1i1 production can be a time-consuming proces Other types of cell lines are used to asse cytoxicity. The false positives can arise from variety of effects like membrane disruptio and indirect inference with the readout.

By stimulating a GPCR with a known agonist and then test compounds. the coumarin acts as the donor partner and fluorescein acts as the acceptor in the FRET pair. In the substrate. allowing simultaneously measurement of the two dyes. CCF2 has an emission of 518 nm and the product had an emission of 447 nm. OR) (59).6.the FRET pair is broken and the product will produce an emission at 447 nm. After cleavage by p-lactamase. Ladamase cleaves the lactam ring allowing separation of the two fluorescent pairs that in turn prevents FRET occurring. Elevated levels of CAMPand IP3 also leads to changes in intracellular calcium for which there are many methods for measuring (58). This instrument integrates com- . using membranes or whole cells. Gi.7 shows the image of cells before and after treatment with an agonist. effectively trapping the substrate within the cell. Figure 2. Molecular Devices. These can be measured in the fluorometric imaging plate reader (FLIPR. Some of the most important cell-based assays are used to find modulators of ion chan- nels or GPCRs. One useful method for HTS uses calcium-sensitive dyes. antagonist activity can be detected. High-throughput screening of GPCRs are either direct binding. or functional methods. Eugene. ids. 61). Activation of Gq-coupled receptors leads to an increase in inositol triphosphate (IP3) that can also be quantified directly. The uncleaved substrate contains two fluorophores that act as a FRET pair and yield an emission at 518 nm. Conversion of CCF2 (Aurora Biosciences) from a green fluorescent FRET substrate into a blue fluorescent product by P-ladamase. Radiometric readouts are popular in cellular HTS assay formats because absolute number of cells is not critical to the assay readout.and Gs-coupled receptors inhibit or stimulate the production of intracellular cyclic AMP that can be quantified directly in an immunoassay or indirectly through a reporter gene. The ratio of the two fluorescence signals allows quantitation of the cellular response and can improve the precision of the assay. and literature examples include the vanilloid and histamine-1 receptor (60.3 Lead Discovery Process Figure 2.

(b) the cells incubated with agonist. mixtures of compounds ca be screened. There are a nutr c ber of biotechnology companies who have d veloped screening platforms that will detec direct compound-protein interactions in higl throughput formats. See color insert. and 3Dimer sional Pharmaceuticals. 500010.neogenesi corn) (63) approach is based on technologic that rapidly separate protein-bound ligan from free ligand using chromatography an then MS to detect bound ligand. but the current form of th technology has low throughput and capacit limits so they cannot be used to screen larg numbers of compounds. Demonstration of a p-lactamase reporter cell line. Sometimes assays for biologic. Neogenesis. in the search for n e antibacterial agents. an alternative to conventional bit chemical and cellular screen needs to be use( One alternative screening approach ths does not require knowledge or analysis of th biological function of the target of choice I direct measurement of compound interactio with protein. yielding a published throughpt of 300.5 Alternate High-Throughput Screenin Techniques. (a) The unstimulated cels. Anadys Pharmace~ ticals. for thi approach. function that will allow the design a biochen ical or cellular screen. I this situation there is no known biologic. 3. In a final volume of 0. as well as calcium-sensitive dyes for ca cium channels. Ion channc modulation by compounds can also be me. A range of techniques are avai able to measure the direct binding events suc as NMR and calorimetry (63. In addition. The NEOGENESIS (www. targets cannot be conveniently designed to f with standard cellular or biochemical assa formats.7. and (c) cells treated with agonist and excess antagonist. genomic experimeni have indicated a large number of proteins thi are essential for the survival of the bacteriun but their function in the cell is unknown.High-Throughput Screening for Lead Discove~ pound addition to assay plates with direct in aging of the fluorescent event..000 cells were seeded overnight a t 37'C before being stimulated with excess agonist or antagonist for 4 h. The screenin efficiency is achieved through screening mi: tures of compounds. By usingmat encoded libraries. For example. e. To screen these types ( target. The resultant p-lactamase activity was measured after a l -h incubation with CCF2 (Aurora Biosciences) at the indicated wavelengths in an U L Analyst fluorometer. sured using FLIPR and a range of voltage-ser sitive oxindol dyes for potassium channe: (62). Novalon. These bic physical techniques can yield important b i ~ c ing information.000 compounds in a day.2.02 mL. large amounts of relatively pur protein need to be available. 64). Figure 2. .g.

See color insert.6 Screen Validation and Reagent Scale Up.. The substrate is a biotinylated peptide that is phosphorylated with ATP and the resulting phosphorylated peptide quantified using a hTRF assay format as described earlier in the chapter.. For simple enzyme assays. 3Dimensional Pharmaceuticals (www. The first challenge is to produce sufficient reagents for the screen.3dp.. The protein's melting curve is measured using a fluorescent indicator dye. The enzyme preparation should be free of other kinases. a relatively pure preparation that is devoid of competing enzymatic activity can usually be readily prepared.. com) thermofluor (65) approach measures the temperature shift in a melting curve of a protein on ligand binding. '..2.. A fluorescence dye is chosen such that it binds to the internal hydrophobic domains of aprotein. I I 1 t 30 35 40 45 50 55 60 65 70 75 80 Temperature .. \ Protein: Tmof Drug complex + ii: ..9. 2.. a tyrosine kinase domain has been requested for high-throughput screening to find enzyme inhibitors..8).. the next stage in the process is often termed the screen validation phase. The fluorescence intensity was measured over a 3080°C temperature range using 50 @ CBS (red) or dimethyl sulphoxide (blue) and 0. Having designed and optimized the con- Figure 2. competing Figure 2. This tests the robustness of an assay in high-throughput screening conditions.8. together with technological improvement to increase their screening efficiency.. pro- ditions for a screening assay.... and acceptance in HTS laboratories will depend on their record of accomplishment in discovering viable leads molecules.15 mg/mL CA.. An example of a Thermofluor ligand-binding experiment where carboxybenzene sulfonamide (CBS) binds to carbonic anhydrase (CA). By comparing the change in melting transition point with and without ligand. For example. These technologies are relatively new.9. exposing more of it internal hydrophobic core and the fluorescence signal increases... An example of a Thermofluor experiment is shown in Fig. 3.. .. Temperature tein binding can be detected (see Fig. The schematic shows the theoretical shift in the melting curve for a protein in the absence of and the presence of a drug binding.3 Lead Discovery Process t I .. 2. the protein melts. As the temperature increases.

x. + 3sJl(x2 . For example.10). standard deviation of the high signal. on different sites.XI).e. and the detection technolOD. etc. then the screening assay is validated across multiple assays. usually expressed as 2' (66): 2' = 1 . usually expressed as the coefficient of variation. the non-specific binding value. a good example is described in the development of ranitidine (67). s. a cell production schedule needs to provide a continuous supply of cells for the duration of the screen. 4. it is important to understand how pharmacological activity translates from one assay to another. and peptidase activity. This procedure allows definition of all the quality control parameters. Usually this is achieved through a combination of purification and the use of inhibitor cocktails. The 2' statistic measures the assay signal window as a fraction of the distance between the means of the distributions. the totals would be the maximum signal in the presence of ligand and the blank measured in the presence of excess competing ligand. Typically. depending on the assay. Screen validation phase tests the screening assay in a more production-like environment. 3. Process precision is measured by repeating the mini-screen on a different day. activator. 3. 2. This is also covered in detail in other chapters.3.5 will yield good screening data. Measuring the effect of a compound library in biological assay can be more complex to interpret because subtle differences exist in different laboratories. It requires rigorous attention to detail to achieve inter-laboratory precision even down to the way the compounds are solubilized and diluted. Precision for the minimum or background signal expressed as coefficient of variation. thus validating the screening process. standard deviation of the low signal. Precision of a known inhibitor. It is critical to ensure that cells that have gone through multiple and differing numbers of passages to maintain their pharmacological responsiveness profile. This is a standard fast follow or "me too" rational design approach to drug discovery and optimization. and s. When analyzing competitor's lead molecules published in the scientific or patent literature. there are a number of significant compound selection choices that depend on target knowledge andlor the ligands that bind to the target. If a series of biological assays are required to progress a chemical series toward in vivo studies. The 2' index is more critical than the signal to background in an assay (see Fig. the types of equipment. a screen that yields a 2' greater than 0. This is more difficult if the target of interest is novel and no pharmacological tools are available.1 Following the Competition. mean of high signal. Some of these choices can be relatively straightforward. 2. For example. Precision for the maximum signal. This involves screen rehearsal with a small number of compounds (i. This process should be designed to be scalable and produce sufficient material to support screening as well as additional experiments like compound concentration response curves and any selectivity assays. The assay signal window measures the distance between the distributions of the totals and the blanks. and even in large pharmaceutical companies and different countries. 3. 3.High-Throughput Screening for Lead Discovery phosphatases.. the lab bench results are replicated on the HTS robotic system. This trans-assay precision is critical for the success of an early drug discovery program. agonist. For cellular high-throughput screens. It may be that no specific chemotypes are available in the literature or from competitor . Typical values include the following: 1 . Analysis of the signal to background.2 Systems-Based or Focused Discov- ery.(3s.3 Constructing Compound Decks and Screening for Hits At this point in the lead discovery process.3. x1 mean of the low signal.. in a binding assay. Critical quality control experiments are performed at this stage in the process. the question to be asked is whether there are clear opportunities available for carving out new IP positions around novel SAR. a few thousand). If known pharmacological agents are available.

Typical control data from two screens. If the target is a known "druggable" target :h as a GPCR. By collecting and searching all the availe compounds in the company compound collection or external commercial libraries with known activity against these family's of targets.5-fold and a 2' of 0. CADD techniques can be used most effectively. the discovery of active compounds can become more effectively targeted. In some companies.10. the assay in b performed poorly. At this part of the process. In contrast. then one could take a "targeted" or stems-based" approach to lead discovery. npany publications to provide useful start. Sophisticated computer algorithms matching structural information of a biological target or ligand to search internal and external databases to find structural .86. ion channel. However. these target class compounds sets can be as small as few hundred compounds or as large as several thousands. then a number of used approaches can be taken to discover d molecules. or simple enne. with a signal to background of 3. (a) An acceptable assay was designed showing a good separation between signal and background.- ( I ) 0 100 200 300 400 Well # 0 100 200 Well # 300 400 Figure 2. if there is ormation on the nature or family of targets it are being worked on.5 and a 2' of -0. 3.!ad Discovery Process 55 m C 0) .2. and it had to be re-designed. points for chemistry.

with onward distribution to HTS laboratories or other research laboratories in the company. supplying compounds in a variety of formats and amounts to lead optimization programs. store. formatting. if possible. These virtual compound hits can. to select compounds with drug-like properties and eliminate polymers. dyes. and retrieval on demand. then in the HTS world screening one million compounds would maximize the chance of finding an active hit. The scale and industrialization of the operation often masks the incredible innovation and smart thinking that has gone into the HTS process over the last 10 years. storage. Because of the recent spate of mergers. Most large pharmaceutical companies now use advanced compound management technologies to achieve this goal (71). Unless an exclusive deal has been negotiated. High-throughput screening is positioned to have its maximum impact on the earliest part of the drug discovery process. Of course. disease biology specialists. The size and diversity of the compounds held within a company compound store largely reflects the size and diversity of the medicinal chemistry approaches a company has taken over its' history. endipitous discovery to occur. Ideally. the internally acquired and legacy compounds are analyzed before going into the compound store. these compounds are stored in large automated storage systems. random screening is that by sheer scale of numbers. some of these compound collections can be over one million compounds. Usu. The aim of this exercise is to find compounds active against the target. HTS is the process by which very large numbers of compounds (hundreds of thousands) from a variety of sources such as synthetic compound collections. The goal of the lead discovery process is to provide a cohort of chemically tractable molecules with sufficiently interesting properties against nominated biological targets. etc. The underlying premise to high-throughput. This may seem like a simplistic.3.High-Throughput Screening for Lead Discovery matches predict active site or receptor binding. and retrieval of compounds. sou can bias the ser. This will include not only compounds in the screening deck that are expected to be active against the target. This would trigger further investment of resources in lead optimization programs such as medicinal chemists. ally the traditional medicinal chemistry compounds are supplemented with libraries from commercial sources and academic collaborators. A core underpinning technology for the screening process described above is the capability to acquire. their registration. Most pharmaceutical companies have large collections of compounds derived from past and ongoing medicinal chemistry programs and samples acquired from external vendors.3 High-Throughput Screening. A typical compound management process would be intimately involved in all aspects of the drug discovery process. the compounds available from commercial and academic sources will be available to anyone who wishes to buy them. and rapidly retrieve large numbers of compounds for testing. The compound management activity covers the acquisition of compounds and libraries. In most large pharmaceutical companies. namely lead discovery. but also compounds that would not have been predicted to be active. and combinatorial chemistry libraries are tested against biological targets. Alternately. 3. A competitive . HTS teams. If the chances of finding a new chemotype against target X is one in a million. QC checking. etc. be synthesized and tested in a bioassay to confirm the authenticity of the hit (68-70). reactive intermediates. These compound management systems allowing the storage. anti-intellectual approach to drug discovery. the technique of "virtual screening" can be used to generate virtual libraries of tens of millions of compounds that can be "docked" into the binding site of a protein target of interest. but it has proven to be a very successful lead discovery paradigm for decades. focused decisions about lead discovery is not available. and biology research teams.000. although the average is more likely to be in the range of 300-500. reformatting. natural product extracts. when the type of information needed to make smart. This refined collection would then form the basis of a "solid" store where compounds exist as dry powders or films in standardized vials. tracking. most pharmaceutical companies now have at their disposal the powerful technology of HTS.

Chemoinformatic resources and a compound acquisition budget are therefore vital to the maintenance of a modern compound store. . A very common source of new compounds for high-throughput screening comes from the prodigious output of combinatorial chemistry laboratories. and then put them through the biological screens as with traditional medicinal chemistry compounds. Having compounds available in both tubes and plates is a distinct advantage for screening. The concentration at which compounds are stored very much depends on both the compound handling process. not all compounds will be stable in DMSO over protracted periods. the next challenge is to manage the requests from individual investigators both to deposit new compounds from medicinal chemistry laboratories and to ship compounds for biological experiments in a timely fashion. A few tens or hundreds of compounds from the solid compound store can be routinely weighed out using manual or automated systems and delivered to an investigator. but many biological assays can tolerate relatively high concentrations of this solvent. The real art is to build combinatorial libraries that contain "drug-like" molecules A that have biological activity. because not all compounds will readily dissolve in DMSO. Alternatively. it was refermented or re-extracted. However. In addition. The microtiter plates can be readily organized and provided to a high-throughput screening group who will test for biological ac- . a more efficient overall process can be developed. 1- ?d a '3' U- rn- l m :agoerto ive edge can be gained by developing tools to select the "best" external compounds to complement the internal compound collection and by acquiring diverse chemotypes that are absent from the current collection. It is important to have a high-throughput analytical chemistry capability to monitor the quality of the compounds. the screening process.3 Lead Discovery Process ? 1 f f e I n ?- e k e . some groups will routinely store compounds as a 3 mM stock solution in 100% DMSO. If a particular extract was shown to have the desired biological activity. and the solubility of the compounds. Having collected and organized a compound deck. the risk of compounds coming out of solution increases. a plant. there needs to be an appropriate compound refreshment process. In an attempt to improve this temporal dilemma. some companies have resorted to pre-fractionated crude natural ~roductsextracts (74). This tactic involves fractionating natural product extracts before screening to yield a series of fractions that each contains just a few components. solubilized compounds are stored as individual tubes and in microtiter plates. or a marine organism was provided to the compound store ready for screening. Here. As a compromise. Large numbers of compounds can be prepared relatively quickly using automated synthesis in either solution or on solid phase. By automating this fractionation process. one could purify or purchase purified natural product compounds directly. . The protracted timelines invariably became the "Achilles Heel" of natural products screening because the medicinal chemistry program had either advanced beyond a point where the natural product could not add significant value or the program had been terminated. Typically. This is a compromise. most cellular assays will not tolerate more than 1% DMSO without adversely affecting the cell physiology. By storing compounds at a higher concentration. and the pure compound isolated (73). Natural products have historically been a significant component of a compound deck (72). For example. a 1 m M stock solution from the compound store will only give 10 final concentration in the assay. the supply of the hundreds of thousands required for a high-throughput screen would be totally impractical using the solid store. From our experience.h 8. a crude solvent extract of a microbial fermentation.This isolation typically took weeks to months. This approach dramatically decreases the time to reveal an isolated natural product and de-couples the fractionation from the screening process. and therefore. To overcome this logistical nightmare many compound stores also contain a liquid store. therefore. screening less complex natural product extract mixtures tends to reduce the false positive rate in a high-throughput screen. The solvent of choice is usually dry dimethyl sulfoxide (DMSO). fractionated. It is beyond the scope of this chapter to detail the various combinatorial chemistries that generate screening libraries.

such as the Haystack (75). each of the 10-20 beads were separated into individual microtiter plate wells and exposed to UV. A compound was synthesized on Tentagel beads using both acid and UV sensitive linkers and encoding tags.000 compounds as dry solids and potentially 15 million compounds in a variety of liquid formats. 2. the compounds were prepared using solid phase synthesis. a modern compound inventory management process is highly integrated. and the released compound assayed for biological activity.58 High-Throughput Screening for Lead Discovery Figure 2. 77). the tag was decoded. UK (71) (Fig. Using this format. Because these compounds are likely to be scattered across hundreds of different microtiter plates. A mixture of approximately 10-20 beads per microtiter plate well was treated with acid. Certain compounds that yield the desired activity are repeat tested to measure potency. If activity was detected. This particular fully automated storage and retrieval system can store over 750.. The choice of whether to screen a full com- . i. (b) the solid storage system and the robot that handles the compound vials. a tube store allows you to prepare a customized set of compounds and supply then for screening. and in vitro toxicity or safety. An example used at Bristol-Myers Squibb is the Haystack system built by The Technology Partnership. Additionally. a standard refrigerator. requiring a combination of chemistry. Overall. identifying the monomer sequence and hence the structure. tens of thousands of compounds could be stored in a relatively small store. focused sets of compounds based around known chemotypes or chemical series are readily assembled from an automatic tube store for lower-throughput screening. tivity. There have been attempts to specifically engineer combinatorial synthesis approaches to facilitate screening that is more effective by eliminating the need for indirect compound storage and retrieval systems. This figure shows some of the components within the Haystack compound storage and retrieval system: (a) the microtiter plate handling system. These screening decks can represent the full collection of compounds in the store or tactically organized subsets (76. information technology. selectivity. Numerous companies now provide very sophisticated compound storage systems that can manage all of the operations described above. and (c) the tube picking robot placing a solubilized compound back into a tube rack. In these examples. Compound stores also contain screening decks that are available mainly for the HTS laboratories.e. If activity was detected.11). The tag encodes the order in which monomers were built onto a scaffold such that the exact monomer sequence could be identified. and production engineering skills.11.

the screening process can be carried out very rapidly (a matter of weeks). 3. It is usual to perform a high-throughput screen with a single replicate of each compound.. The results are statistical in nature and interpreted as population data. etc. toxicity.1 Analyzing Screening Hits. a statistical parameter Z' was introduced. directs further medicinal chemistry. 20-40% inhibition. Having analyzed the QC data and eliminated. Typically.3 Lead Discovery Process pound deck is dependent on the medicinal chemistry insight or structural knowledge of a particular target as discussed earlier. If an assay had a Zr of 0. From a screening perspective. Figure 2. the majority of the weak inhibitors. etc. both negative and positive. antagonists. The idea behind this type of extensive profiling is . the actual screening is now the quickest. panels of known inhibitors. respectively). 3. any compound plates that failed.. activators. the so-called "Plug and Play" approach. Of all the different stages in the lead discovery process. plates containing just DMSO. assay false positives. and Candidate Selection After ordering a copy of the compound deck in the plate format of choice." A hit that seems to be statistically significant could be explained by a variety of reasons. a variety of QC plates are inserted into the run. QC plates containing biological reagents check for drift in the assay over the course of the screen.7 and a known inhibitor was known to cause 50% inhibition at 10 km. designated as a "hit. or agonists. the percentage of inhibition could vary between 35% and 65% because of population statistics. Figure 2. During a primary screen. Interestingly. False positive results arise from pipetting errors that delivered the incorrect amount of a reagent. In this particular example. The next step is to analyze the primary screening quality control data (66). a fluorescent compound or quencher in a prompt fluorescence assay. A second round of assay(s). False positives are compounds under test directly interfere with the detection readout. detector misalignments. e. as well as true pharmacological response. 73% of the active compounds retested in the second assay.2. a and b. First. 3. The vast majority of compounds from a full deck screen has no effect. Second. then in a screen. any compounds with greater than or equal to this value are retested with replicate determinations.12. Additionally. controls. Profiling. are included on each plate. similar assays allow parallelization of assay design.4 Hit Identification.4. major companies are adding to the value of highthroughput screening by immediately profiling screening hits against a battery of selectivity. for example. were false positive compounds that interfered with the assay readout. In the assay validation section (Section 3. Having selected a particular cut-off. Ki or IC50. and there is little information on compounds that effect the target of interest. the entire screening run is analyzed. For example. The screening scientist monitors the automated system continuously for both hardware and assay performance. there are significant advantages to a systems based or target class approach. 2.g. including blank plates. and statistical analysis allows one to decide when a compound has had a significant effect. or repeat tested. Full compound deck screening is extremely useful if you wish to find new chemotypes. depending on the assay. cytotoxicity effects. are tested at multiple concentrations. Realtime data analysis allows the screening scientist to continuously monitor the performance of the screen and the robot. A third round of screening generates data from concentration response curves so potency.13 (a and b) shows the frequency histogram of the retest values and the associated scatter plot.14 shows the Zr values for this screen's quality control plates. Increasingly.4. The initial high-throughput screen is normally viewed as a population frequency histogram and as a scatter plot (see Fig. assays that indicate which compounds show the desired selectivity against other biological target or lack of cellular cytotoxicity. the screening data for a family of closely related targets is simultaneously generated for set of compounds. and safety assays (78-80).2 Profiling Hits.61. pipette error. monitor hardware performance. and or efficacy for agonists. to analyze the hits eliminates false positive results.

A further hope is that as compounds fail or succeed in the R&D process. this early profiling may start to indicate predictive in uitro profiles for each of these outcomes. TI .High-Throughput Screening for Lead Discovr % Inhibition (b) Activity 0 100000 200000 Compound number 300000 Figure 2.12. After the concentration response phase. Th is when the team described at the initial stag of the assay design process comes back t gether to decide which of the active chem types display the desired characteristics. the better the decisions will be on which compounds to progress or to terminate. that the more information scientists have about biologically active compounds. (a) A frequency histogram of the data derived from a high-throughput screen run of 340. d cisions on which hits to progress down tl drug discovery pipeline need to be made. This would provide a significant advantage to those companies trying to improve the quality of the drugs in their pip line and their chances for survival. (b) A scatter plot of the same data indicating where the cut-off was chosen to retest the compounds that showed some activity.000 compounds used to find inhibitors of 15-lipoxygenase.

(b)The same data as a scatter plot. (a) A frequency histogram of the retest data derived from the positive compounds from a high-throughput screen run of 340. There are a viwiety of reasons that include the following: [ncorrect structure when the compound was submitted to the compound store The choice of compounds to progress can vary. as well as a hig a )' 2.13. HTS processes provide in vitro data on all the major cytochrome P450 activities. and any additional information can help prioritize the future workflow. stnlctural integrity of the positive compounds should be determined: LCIMS. additional .000 compounds against 15-lipoxygenase.!ad Discovery Process 61 (a) 300 Activity of compounds against 154ipoxygenase Mean Compound number Figure 2. The activity was caused by a minor contaminant st01 Proximately 10% of the compounds will not be exactly as described in the database.

This is the point at which the initial HTS process has ended. etc. In Fig. but it helps in understanding how to drive the medicinal chemistry forward. a whole matrix of data is supplied to help drive the decisionmaking process. A QC plate was run after every 25th 384-well plate. HTS approaches and technology are back into play if the project requires further chemotypes. if the only datum available to the medicinal chemist was the potency of a compound. Chemical tractability: a more subjective flag that is based on a medicinal chemist's view on whether this is a good start for a medicinal chemistry program. red = high. BMS5 has the cleanest profile. Together with the growth of combinatorial chemistry. Drug-like properties: assessing the compound in a range of drug-like in silico models. Liability or disadvantage: this is color coded.High-Throughput Screening for Lead Discovery 0 5 1 2 3 Quality control plate number 1 2 3 4 4 Figure 2. cytotoxicity indices. The number indicates the number of positive results found. to check whether the compound was found to be active against other targets. 2. both against other targets HTS and lower-throughput screening. Mutagenesis the activity in an in vitro model of mutagenesis Cardiac Liability: activity in essential cardiac ion channels that would cause an adverse side effect in humans. This interdependence has lead to a technological revolution in high-throughput . Activity in these liability assays will not stop the progression of a compound. cardiac liability reflected through adrenoreceptor or ion channel activity. The decisions from this point are around compound optimization and continued target validation. and green = low. yellow = medium. The data are calculated from the maximum signal and the background where n = 160 for each data set. although it was 10-fold less potent than BMS1. and this chemical series was the preferred candidate for progression. BMSl would be the highest priority. In this example. many targets and compounds require HTS. Target potency activity from the HTS assay Target selectivity: search of all the screening databases. Human liver enzyme inhibition (CLM): the activity in a standard cell-based toxicity model.15.14. (a): (a): 4 TECHNOLOGY INFRASTRUCTURE IN HIGH-THROUGHPUT SCREENING Advances in genomics have caused a dramatic increase in the number of potential targets. These plots show the quality control data that came from the 15-lipoxygenase HTS expressed as 2'.

pically range from lo5-lo6 through a range ~fbiological targets in an efficient and timety manner has been one of the main drivers for automation. Detection instrumentation was also adapted to enable robots to load and unload microplates. and supplying a range of surface chemistries for specialized assays. and incubators. Automation in High-Throughput Screening 4.15. A compound profile matrix that outlines all the information on a set of compounds that we been identifiedfrom a HTS. Invariably. Robotic systems needed higher tolerances and defined dimensions to pick and place plates precisely. varying the shapc: of the well. From the late 1970s to the mid. 84 and 85.nfrastructure in High-Throughput Screening 63 BMS-I BMSP igure 2. a method moving around microplates. However. scree: ning (15. a series of detectors. see refs.1 The I !ed to screen large compound libraries that . All the high-throughput screening automation platforms tend to have the same limited number of basic operations. the types of equipment tend to be similar. Manufacturers supplied many variations on the 96-well plate theme. dispensing liquids. Automated screening systems needed accurate plate dimensions. The methods . and the way in which the screening process is integrated dictates the level of automation. and it was rlot until 1996 that a standard was recommencled by the Society for Biomolecular Scree!ning. the 96-well microtiter plate reigned supreme in many screening laboratories ( I 32). For a brief discussion on the advantages and disadvantages of automated platforms versus workstations. from manual to semiautomated to fully automated turnkey systems (83). High-throughput screening automation exists at a variety of levels. varying the color of the plate:.990s. the 96-WI 3 1 1 plates were all subtly different.81) with significant investment in tht automation and miniaturization.l.

17).ed arm or track system to move microtiter plaltes . Movements can be with an articulated robotic arm.16). Images of the Aurora Bioscience UHTS screening platform that is based around a track that moves the microtiter plates around the different workstations. or through a track that resembles a mini-production line designed to shuttle plates around the system (Fig. The range of potential detectors is dependent on the assay technology discussed earlier in the chapter. for moving microplates tend to fall into two general approaches. The incubators can range from the very sim] pie open racks of shelves to highly environmt?ntally controlled systems. The glue that puts all this together is the scheduling software tIlat controls what goes where and when. T'he scheduling software instructs the articulat. and disposes of any waste (Fig. picking and placing plates. The 96-well piezo-electric dispensing head is shown in detail.he more sophisticated systems. The liquid handling options can vary between a syringe-based system that gives higher volumetric precision to aperistaltic pump that tends to be more rapid but less precise. 2. plates containing test compounlds. an operator loatds reagents. In t. 2.16.64 High-Throughput Screening for Lead Discov Figure 2.

dated times around the various liqat the calc~ uid handle!rs. a consistency of process that can be verified. a particular detector may not be available in a format that can be integrated.17. fully automated scre!ening platform does offer continuous operat. Workstation approaches replaces the robot with a human. incubators.is can be used to monitor drift in line analy~ the assay olr whether certain wells fail to meet prescribed quality control parameters. money. and a willingness to develop the necesaary skill sets. on. and detectors. as seen with block1ed tips on a dispenser. For example.ion. Images of typical turnkey automated screening robots that use an articulated arm to move p:Lates around the screening system. and as long as . and eventually to waste. These types of error alc?rtan operator. as dat:I are generated by the detector. automated audit trial of the samples that have been tested. "a stack. 86). Additionally. The robot is placing plates into two different liquid handling devices. where plates are processed in a serial manner. Not all laboratories have the resources to build fully integrated screening platforms and support them. and safety (for further detail see Ref." in workstation approaches. Top left is a PlateMate Plus and top right is a Multidrop. plates are batched together. A limited amount of 1 artificial intelligence can be used rules-base( to aid the Iquality of the operation. This system was built internalIly by Bristol Myers Squibb engineers. A stable. Most screening laboratories will use workstation approaches in addition to fully automated platforms to enable assay flexibility. Unlike the automated platforms. For example.4 Technolog! Infrastructure in High-Throughput Screening 65 Figure 2. not all assays can be modified to work on an automated platform. Buildin] g these integrated robotic systems requires Eltrong management commitment with time.

there ha been pressure to reduce volumes even furthe: The 1536-well plate is emerging as the poter tial next step. (b) 3% well plate (25 pL well assays). The 1 pL assay is also no available in the 1536. the workstation data are comparable with that generated by a robot. reagent costs are reduced. One real advantage of a fully automated platform is where there is a need to have accurate incubation times. in the 96-we scintillation proximity assays.2 Miniaturization of Screening Assays The increasing operating costs of HTS laboratories have driven a strong interest in implementing more cost-effective ways of carrying out high-throughput screening campaigns.18). ever increasing demands to expand testing capacity and improve process efficiency while simultaneously reducing costs have pushed HTS laboratories into using 384-well plates and beyond (Fig. and automation to fit the new 384-well plate. In little over 5 years. The discussio over well density still causes many debates i screening discussion groups (87) and eve higher well densities have been proposed. The first tangible step along the miniaturization route was the introduction of the 96well microtiter plate that replaced the individual tube (82). Instrumentation companies re-invested in designing or adapting liquid handlers. HTS laboratories adopted the 384-well plate as standard.000 compound deck with average reagent and plasticware that costs $0. this often works well. allowing a fourfold increase in well density and increased screening capacity (87). A screening organization that runs 50 screens a year. Additionally. there is the need to detect the results ( a particular assay. smaller amounts of compounds are needed for the assay. Additionally. e. For example. Miniaturizing the plate format is one of the major technology solutions.18.g 9600-well plate (88). " Figure 2. The same quality control is incorporated into the workstation process. In the late 1990s. MA and Evotec OAI. Hambur! Germany) and the 3456-well microtiter plat (Aurora BioSciences. San Diego. in a kinetic assay. This scenario in a 96-well plate format generates 260. The 96-well plate became the standard workhorse in academic and industrial laboratories over the last 20 years. Th miniaturized down to 20-50 LLL minimum practical volume of 20 pL for th new 384-well plates was defined by the we shape and the need to produce an even layer ( liquid at the bottom of the plate. the scintillatio counter photomultipliers are positioned abov 7 . 4. CA) (50). and by reducing this to around 5-10 pL. low-volum 384-well microtiter plates are now commei cially available. One advantage ( the 1536-well plate in absorbance-based a! says is volume reduction while maintainin the path length. Even with the 384-well plate. we have witnessed a 10( fold reduction in assay volume and a 36-fol increase in the well density. There has been a stepwise evolution from the glass test tube and plastic Eppendorf tube into microtiter plates containing ever-increasing well densities. There are many different types of m crotiter plates that are used in miniaturized assaj for HTS. and from our experience. with square wells that allow working volume of 5-10 pL.(Corning Costar Corp Cambridge. detection systems. for example. The move to higher well densities an lower assay volumes has presented significar challenges to instrumentation companie: First. totals $5 million.High-Throughput Screening for Lead Discover the number of microtiter plates processed is acceptable.000 plates of plastic waste per year. However. (c) 1536-well plate ( pL well assays). each testing a 500. and (d) 3456-well plate (2 FL we assays) The vast majority of assays have readil volumes. (a) 96-well plate (100 pL assays). excluding waste management costs.20/well. A typical assay volume in the 96-well plate is 100200 pL. 2.

E. UK. The disadvantage was that it then took four times as long to read a plate. W. information technology experts. with null cell lines to remove false positives. there is a need to rapidly profile and evaluate the selectivity of compounds that are positive in a highthroughput screen. . New York. Duncan. The hits can be evaluated in this format at multiple concentrations. As mentioned earlier. the reading time is reduced to 27 h. (CCDs) provided the answer (e. Nigg a d D. 1. using charged-coupled devices. interactive process that brings together multi-disciplinary teams of chemists. To achieve nanoliter dispensing.. S. Scrip Magazine. The application of industrial automation technology to the process has increased the capacity. MA). Additionally. 5. these machines were adapted to read 384-well microtiter plates. Burger's Medicinal Chemistry and Drug Discovery. 74-84 (2000).. or 1536-well microtiter plate. J. At the top end of this scale. Imagers take the same time to read a 96-well.g. biologists. and in a range of other cell lines yielding a selectivity index. Eds. A. speed. Palo Alto. Med. " . 6. D.. In conjunction with corresponding advances in target identification. the drive to screen more compounds has fueled the need for miniaturization. e. J. fits into the drug discovery process. L. as a lead discovery tool. 20. Ganellin. 384-well. 3. Matrix Platemate (Matrix Technologies Corp. The next 10 years will no doubt bring further technological advances and creative insights to improve the drug discovery process even further. J. 10. W. Amersham. C. M. CAI (89). 4. vol.. pp.. 325 plates are used. Lewis in H. Arlington. and quality of this part of the drug discovery pipeline. A new solution needed to be found. a wealth of information is generated in a short period on exactly the same compound solution. A 500. HTS is a multi-factorial. For example. Emmett. two platforms are available: the piezo-electric inkjet dispenser and the solenoid inkjet dispenser (90-92). The overall gain in assay efficiency is dramatic. H. LEADSeeker. REFERENCES 1. C.400 96well microtiter and filter-binding plates. the ability of pharmaceutical laboratories to rapidly move from target concept to lead optimization candidates has changed dramatically over the last decade. Amersham Pharmacia.000-compound highthroughput screen using a filter binding assay format consumes approximately 10. Nature. 1536 wells per plate. Buss and R. 6. time-resolved fluorescence and for measuring light emission. automated chemistry. N.References each well to measure the emitted light. and mechanical and electrical engineers. John Wiley & Sons. From target discovery to lead optimization.. Fleming. total time taken to generate the data would be 36 days in a single plate-based scintillation counter. Lowell. Using a mask. Plants Used Medically by Indigenous Peoples. Imagers are now available for fluorescence. Seigler. Pharmaceut.35-38 (1997). Rosofsb. 1995. 236.. A. it takes approximately 10 min to measure the light from a 96-well plate. W. statisticians.g. a GPCR cell reporter assay designed to detect agonists using a p-lactamase reporter can be readily miniaturized to 2 pL in a 3456-well microtiter plate. J. Banerjee and M. P. Execut. For the LEADSeeker format using miniaturized plates. This will keep HTS approaches as a mainstream drug discovery tool for years to come. A. D. For the 1536-well assay using imaging technology. and CLIPR Molecular Devices. an intricate network of processes and decisions are required to produce a successful drug discovery campaign. 2. We have emphasized the high-throughput screening process as part of an integrated approach to hit identification and assessment. 831033. By combining this with cell toxicity assays and bio- chemical cytochrome P450 assays. Additionallv. Miniaturization facilitates the parallel processing of numerous targets simultaneously. and therefore. Exp. Br. a variety of tip-based syringe-driven devices are available. Another engineering challenge was to dispense volumes in the 20 nG1pL volume range.385-390 (1972). K. 5 SUMMARY Throughout this chapter we have described how HTS. Black. Waigh. 226-236 (1929). and data analysis. R. 5th ed. Imaging technology. Parsons. and M. Durant.

. Chichester.397-403 (1998). J. 32.. Orr-Ewing. C. John Wiley & Sons. S c r g . J. 33. T . M. Human Medicinal Agents fi-om Plants. Chem. Pezzuto. 13. P. D. 51.301302 (1999). 7. Ellis. K. 33-37 (1999). and D.Aberde. Biomol. R.1387-1392 (1998). D. 734-740 (1997). Robotics Autom. Muir. 6. J. R. J. 335353 (1999). and D. 2. D. Med. Marcel Dekker. Shu. P. 5. Fogarty. Ashman. S c r g . Banks. Robotics Autom. 1. Oldenburg. H. T . 4. Haupts. 18. A. Biomol.47-53 (2000). Today. 1217-1242 (1997). Taylor. J. T . R. Florey. S c r g . Prod. J. Farmaco. 463-480 (2000). A. Opin. J. Y-Z.. Krummenacker. et al. 245-269 (1998). N. 16. 19. Drug Discov. Wilson. Symposium Series 534. Terpstra. and J.. . 1960-1964 (2000). Lab. M. E. S. G. Denyer. Hertzberg. Lancet. Cox. 11. Bergsma. A. King. pp. Dunnington. 2. 6. B. 69-74 (2001). J. Choi. Treheme. D.261-265 (1996). M. Janzen. 901-906 (1999). A. D. D. 275-281 (1999). P. Soejarto. M. New York. S. Herrity. Wani. 1992. P. Biol. S. Biomol. W . Cook. Domanico. Pharmacol. R. C. Y . P. 27. V . Chem. G. 1.467-470 (1995). Haaland. Biol. 28. Yund. P. 8. Prystay. N. A. Houston and M. 5. M. Moore. Washington DC. Sundberg. Moore and S. 17. Opin. Upham. Karp.226-228 (1940). Z. pp. Blackstock. Curr. Drug Res.. Bushfield. Hefta. 11. Merritt. Drug Discov. 4. S. S. 37. Bingel. S. D. Chem. Tait. S. Emerick. 25. vol. Biomol. S. Xiong. Chain. J. Quinn. W . M. P. 4. 33-74. Chem. Murdock. A. Fitzpatrick. Davis. Broach. 4. Gardner. 0. Turconi. Guo. Science. E. W . Biol.. Z. Kurali. U . and A. Drug Discov. W . Today. 23. Science. Cordell. J. 35.. A. and P. Hemmila. Ruediger. A. 45. Vaidya. Farr-Jones. J. R. Jennings. K. Stewart. Scrng. J. M . J. C . Parandossh. A. K. B. 1. Prog. Scrng. 324-329 (1999). J. P.. Menius. Fox. G. A. Cordwell. L. Pope. J. D. 38. Lee. 33. Sherrill.35-44 (2000). A. Science. J. M. A. a n d M. Lane. A. G. Biol. 135-143 (2000). Trist. 3. B. Saussy. 36. Chambers. New York. J. Biomol. Curr. 270. 42. Wall. Watson. Opin. 403-406 (2000). Brown. Experimental Design i n Biotechnology. . D. 213-225 (2000). Biomol. Lsakody. Scrng. R. Binnie. W . Curr. Biomol. 183-186 (1999). . Chem. D.. 287-294 (1996). Biol. Rees.. A. R. Goetz. 30. D. and S. D.83-133 (2000). and J. Drug Discov. 72-78 (1997). S. R327-R333 (1999). Zhang. 13. Schena. K. 4. J. 0. K. Opin. World Health Organ.965-981 (1985). 26. Banks. 37. 55. 1994. F. Sigal. N. M. Lab. Evans. D. Brown. 49. D. D. S. 11. Paley. O'Neill.. 9. J. J. 61.High-Throughput Screening for Lead Discovery Phytochemical Resources for Medicine and Agriculture. 105. P. Jones aid F. 63.. B. N. Curr. Dongre and S.. L. R. S. Trends Biotechnol. K. Rep. A. Botstein. Campbell. Bull. Valler. Cordell. Kenny.. 48. 34. M. C. I.159-167 (2000). Schulz. Electrophoresis... Inc. Plenum. Lewis. Biol. R. S. M. A. and R. Wagg. Genet. Snider. Drug News Perspect. 29. P. P. Winzeler. P. .1053-1071 (1998). H. 191-204. M. Pope. Chem. 15. C.. D. Searls. and J. Chichester. Biotechnol. 287. 22. Cox. G. L. 24. Phytochemistry... 41. Lewis. E. D. G. 8. 14. Farnsworth. Biol. Chem. P. C. I. Chem.. Opin. R. J. Today. A. Sneader. K. R. John Wiley & Sons. Biomol. A. 40. 46. Burbaum and N. Fantom. 597-603 (1998). N. 25-41.. Drews. Brown and D. Heathey. Nut. J. Farnsworth. Stadel. Drug Discovery: The Evolution of Modern Medicines. B. Annu. Mander.. and T . W . R. 17. Curr. A. 416423 (2001). T . Br. American Chemical Society. Kinghorn. J. 1. S. Beecher. L. 285. S. Harris. 2. J. 5. M. A. 39. J. M. T . 201-204 (1997). Mathis. N. J. 43. 303-307 (1999). and W . S. J. 0. Silverman. M. World. 1985. A. 12. 13-19 (2001).. W . 8. R. T . Fernandes. Curr. E. 18. 301-311 (1998). J. V . U K . Lutz. Gosselin. Soejarto. 31. C. Chem. 21. R. 10. Scrng. M. I. G. S c r g . Curr. . J. J. P. 71-76 (1999). Nut. Biomol. . 125. J. Sanders. Opin. 21. 1989.. U K . and Z. Statistics: Textbooks and Monographs. J. Opin. Med. J. 20. pp. Ratti and D.277-286 (1996). Parry-Smith. 2. E. S c r g . N. Prog. B. and L. Shalon.. D. C. M. 5. P. Humphery-Smith. 44. W .. H. 56.309313 (1999). The Ethnobotanical Approach to Drug Discovery: Strengths and Limitations i n Ethnobotany and the Search for New Drugs. A. 47. 1993. Donnelly. Green.

I. Orton. M. V. Mathis. K. Roch. J. Opin. and W. L. 88. Curr. 5. I. S. A. Ranell.. Drug Discov.445-451 (2000).17-24 (2000). 6. Rink. 5. J. M. Blaney. J. D. Today. Stahl. J. Biomol. C.. D. 91. C. J. 73. Langridge. Biol. Hertzberg and A. Scrning. Biochem. Pharmacol. 77. Wilson. Zimmerman. Morrison. Inc. Schmid. 26.. 5. G. 45-81 (1993). 66. 4. Walters. R. 160-178 (1998). J. 286-293 (2000). Lett. 202. 4. D. D. . G. Shieh. Watt. pp. T. F. 125-133 (1999). Scrning. B. Alligood. . Drug Discov. J. 70. Miller. 80. Ranson. Today. 57. Today. Petrella. 141-152 (2000). Feygin. Scrng. Esbenshade. Pope. J. Oldenburg. Maffia. J. Hancock. I. Spalding. and J. Ireland. Graf. Coassin. Forgarty. K. Scott. New York. T. Biotech. E. Skolnick. S. Harding. 67. Opin. Kinder.401-410 (1999). Goldfarb.. Evans. G. 114. 87. Opin. E. I. M. 2. R. Warrior. 83. 135-143 (2000). M. R.. S.385-390 (1997). R. 10-15 (1992). J. 53. C. A. Lehn. Groebe.. 3. Sportsman. Coghlan. and W. 72. D. (Comb. Carver. Alpha. P. D. J. Kang. and D. Camacho. 52.. S. and R. 6. J.. Today. P. Today. J. S 8 4 S 9 1 (2000).Angew Chem. and M. A. Kimery. M. Y. Chem. Wildey and C. Thiericke. Bennett. J. Kiilhoma. 1999. 5. Lenz.249-257 (1999). Dev. Guarino. Jerman. A. K. 60. M.1297-1300 (2001). B. P. J-H. Drug Discov. D. 4. Today. V. R. C. 5. M. B. Biol. Hemperly. Biophys. Biol. J. 82. J. P.. Mason. Today. J. Gopalakrishnan. and M. Opin. T. D. R. Engin. and T. DrugDiscov. 89. Biomol.. Lai. Drug Discou. Banks. 85. DrugDiscov. Vlahos. C.85-91 (1998). 9. and F. T. M. 63. Today. 3. E. 570-576 (2000). 1089-1125 (1999). T. Grabley. Burns. J. S. J.. D. Today. 84.-C. S. Biol. C. J. R. N. Spencer. 68. Jindal. K. R. Masters. B. M. P. Murcko. Gilmer. Marcel Dekker. H. Zhang. 161. S61-S69 (2000). Wagner. 5. C. Curr. J. K. Combs.411-416 (1998). A. L. J . Approaches to High Throughput Toxicity Screening. J. 1997. England. 429-440 (2001). 230240 (2000). J. Bayliss. de Costa. Eur. M.. and S. and J. J. Biomol. Margolin. R. 74. 86. Banks in C.. Chem. Int... D. Chem. and T. Engl. Chung. High Throughput Screening: The Discovery of Bioactive Substances. Today. S c r g . T. 4. Tarbit and J . J. Nash. Ed. A. and M.411-419 (1999). Tisone. 59. Timmins. C. Eur. Ramm. S c r g . E. J . 1. Mol. J . Crosby. Lane. Drug Discou. J. V. Biomol. 615-617 (1998). Rose. M. T. Hamman. Biomol. 90. Biol. Biomol. I. and P. Devlin. 5. 343-349 (1998). C. 4. T. Banks. Biomol. Lobanov. D. Sullivan. K. Drug Discov. 56.4. P. Oldenburg. Brough. Biomol. 64. A.. Harker. 2. and P. R. A. Manetta. Nakayama. D. and T.31711-31719 (1994). J. D. B. K. Smart. Mere. and B. Whiteaker. F. 55. W. Mukkala. M. D. Purcell.. Springer. 54. Rodriguez.61-67 (1998). 65. B. W. 3. 81. P.5-8 (2000). Pitner. Hayes. S. L. K. 78. E.. Green. Yan. 5. Asel. J . T. E. W. Med. Biol. A. Today. R. 15-25 (1999). Charifson. 76. J. FASEB J. Oatley. M. Methods. 79. P. W. Kwasnoski. H. Berman. D. R. Drug Discov. 51. C.References 50. 4. Drug Discov. Evans. Kuntz. and P. Drug Discov. J. 71. 11. Gopalakrishnan. Cairns. J. O'Brien. Pantoliano. T. Curr. Chen. Binnie. M. A. R. J. Drug Discov. A. K. Tucker. P. J.. R. Dunn and I. Oldenburg. R. and I. G. D. Latva. M-H. Wodnicka. C.. 267. and F. J. S. M. Chem. Lemmo. K. Methods Mol. V. Clark. Myslik. Biomol. B. Stitt. Chron. 3. S. 79. S c r g . J. A. J. Drug Discov. Bradshaw. T. C. J. 4. Zhang. 51-58 (2001). Pale. 69. M.5421-5426 (2000). Salemme. Curr. J. Drug Discov. Roberts. A. Scrng. V-M. 115 (1999). Blom. Alterwill. 5.2934-2940 (1990). and M. Scrng. and G. Pub Yaler and Frances Ltd. M. H. Rose. R. Scrng. T. J. H. 5. 4. M. Rice. Herman. Pognan. 4. E. Physiol.. Chem. 67-73 (1999). 53-54 (1999). Today. A. 61. Drug Discov.). . 417. Biochem. Harrison. Today. Krystek. B. Biomol. Dee. D. M. 294-300 (2000). New York. 305-312 (2001). E. Takahashi. 9-29. J. M. P. 92. Drug Discov. Eds. Gundhorpe.266-267 (1987). 58. Rev.269288 (1982). 283-290 (1993). I. A. Negulescu. Witte. Scrng. R. M. . 3-48. Lechleiter. 363-369 (1999).55-56 (1998). Burbaum. 62. and J. 75. L. Miller. pp. Hemmila. and A. M. W. Heath. M. Biochem. D. N. E. H. Havunjian. Green. L. Ferrin. Anal. Harvey. 61. Homon. Valler and D. and A. R. A. McCabe. Pipe. P. S. D. Davis. Osbourne. Chem. Chung. Good. I. A. 26. Bioorg. Jordan. Sattler. and P.. 269. S. J.. P. U. Today.

79 Burger's Medicinal Chemistry and Drug Discovery Sixth Edition. 73 3 Pharmacology of CAMPAssays. 71 .1 Receptor Pharmacology.3 B-Arrestin Systems. Abraham ISBN 0-471-37028-2 O 2003 John Wiley & Sons. High Content STEVEN J. 76 6 HT-Flow. BROWN IMRAN B. 79 8 Summary.CHAPTER THREE Rapid. 74 3. 74 3.CLARK BALA PANDI Axiom Biotechnologies. San Diego. 72 2 Target-Directed Pharmacology. California Contents 1 Introduction. 75 5 Cellular Toxicology Assays. Inc. 73 2. Volume 2: Drug Development Edited by Donald J.2 Nondestructive Fluorescent Methods. 75 3.2 Assays and Pharmacology of Rapid Cellular Responses. 77 7 High Content Metabolic Assays.1 Introduction. 73 2. Inc. 75 4 Ion Channel Pharmacology.

3. Species-specific compound activities have been observed in drug discovery caused in part by differences in the amino acid \ Humans Isolated tissues Ease o f assay --------c Figure 3. An infamous example is provided by the discovery of a mouse G-CSF signal transduction pathway activator (1). In the next least complicated system. animal and human tissues and whole organism studies cannot be carried out with sufficient throughput to successfully drive hit to lead optimization. enzyme inhibition kinetics. Thus. enzyme kinetics. This compound was identified in a high-throughput assay in murine cell line. high-throughput compound screening. requires medicinal chemistry compound optimization driven by relevant pharmacological data. . Other examples include data obtained from pharmacological profiling of lead compounds against a broad panel of natural human cells. Pharmacological data regarding the mechanism of compound interaction with the molecular target. Systems and methods that provide high definition. medicinal chemistry efforts have not resulted in a compound active towards human G-CSFR. In vitro assays for evaluation of the adsorption and metabolism properties of compounds have been developed. test compounds. and protein-ligand binding are clearly important. adsorption. Cellfree screening eliminates cell metabolism. and protein binding issues.Cell-free systems consisting of purified target protein. toxicity) and specificity of action. It is most useful if these cells have been previously characterized for their response to known pharmacological agents.1). The role of high-throughput pharmacology in the hit to lead compound candidate identification process includes determination of efficacy and evaluation of potential compound liabilities (solubility. such as receptor agonists and antagonists. enabling toxicology studies (hit to lead). Such complex interactions. cell permeability. untransfected human cells provide the most relevant practical drug-screening systems. sequence of the target protein. are difficult to reconstruct in either cell-free systems or those employing engineered cells.Rapid.1. cells for study are often transfected with the target gene or a reporter system of gene function in a manner that creates a wide assay window but also alters the ratio of the target receptor to cellular components. Quantitative methods for evaluating the effect of pharmacological agents on receptor activation and inhibition. On the other hand. Despite active research. untransfected mammalian cells provide intact signaling pathways and other complex mechanisms not reliably reproduced in cell-free systems or in cells with overexpressed targets. The ease compound testing is inversely proportional to the complexity of the system and the physiological relevance of the assay. Pharmacological targets adapted to high-throughput methods are often poor predictors of efficacy in viuo but are used for ease of testing (Fig. such as those in multiple-protein signaling pathways. High Content Pharmacology 1 INTRODUCTION Drug discovery program progression from "hit" compounds derived from high-throughput screening to identification of a lead drug candidate suitable for an Investigational New Drug. and ligand-receptor binding are reviewed in other chapters. contentrich information about cellular responses and are robust enough to have proven useful in secondary testing pharmacological properties of compounds are reviewed in this chapter. Finally. and the necessary substrates and ligands for target protein function and signal generation are often the choice for primary.

in the HTPS laboratory.2 Target-Directed Pharmacology 2 TARGET-DIRECTED PHARMACOLOGY 2. whereas the Schild analysis of simultaneous addition of agonist and antagonist demonstrates competitive inhibition.1 Receptor Pharmacology High-throughput methods for the determination of receptor-ligand binding and inhibition have been developed. found that preincubation of L-742. Also. The agonist and antagonist can be introduced simultaneously to the cells so that both compete for receptors. increases the apparent EC50 of substance P. detected by Casensitive fluorescent dyes. These varying results stem from the time-dependent nature of . Simultaneous addition of substance P and L-742. agonist potency increases with incubation time. Apparent "non-surmountable antagonism" is observed when the system is preincubated with antagonist (3). for platebased assays. the system preincubates antagonist with the cells before introduction of the agonist.2. the cells are often grown in the microtiter plate for 1-3 days in advance of the experiment. surface plasmon resonance. functional receptor signaling and physiological cell responses are the more relevant endpoints.2). 3. In the configuration shown in Fig. however.. is perhaps the most commonly measured signal. 3. the order of addition of antagonist and agonist is known to change potency and apparent inhibition characteristics. Methods such as radiolabeled binding studies on flash plates (New England Nuclear) and the estimation of the binding of Fluoro-tagged ligands by fluorescence polarization. This introduces a source of variability because the cells can react to the different growth conditions provided in the individual wells. First. In both formats. In the HTPS system. Other parameters that change within a short time scale include intracellular pH and membrane potential. Assays and Pharmacology of Rapid Cellular Responses 2. the relative off-rate of antagonist will determine potency.694 to the cells results in an increase in the EC. Examples of receptor pharmacology across the variety of receptor families are covered below. and therefore they are more prone to evaporation. In another example Cascieri et al. Schild analysis with inhibitors pre-incubated with cells demonstrates apparent non-competitive inhibition. the cells are kept together until the moment of the test. For example. This greatly reduces variability between testing events. The HTPS uses flow-cell technology to serially introduce compounds to continuously flowing cells in a mixing chamber of the fluidics system (Fig. inhibits binding of substance P to the human tachykinin NK1 receptor.2 Probes for the fluorescent detection of rapid cellular responses have been widely used in the characterization of pharmacological agents. and typically. Receptors are often screened for ligand binding. a continuous concentration gradient of test compound can be introduced to the cells. incubation time of agonist with cells has an effect on the maximal cellular response. These endpoints have been widely used in conjunction with appropriate control compounds and/or control cell-lines to both screen compound libraries and evaluate lead doseresponse profiles. for substance P with no decrease in the maximal level of stimulation (2). The calcium response to GPCR agonists is both time and dose dependent. edges of the microtiter plate experience differences in humidity compared with center wells. and decreases the maximal level of stimulation observed. The high-throughput pharmacology system (HTPS) developed by Axiom Biotechnologies allows for the rapid and high definition characterization of immediate cellular responses. In this mode. and ELISA-type assays have all been used for high-throughput screening and rapid characterization of compounds that interfere with ligand receptor binding.694. Second. a selective morpholino tachykinin NK1 receptor antagonist. A suitable incubation time to optimize maximal response and agonist receptor potency is selected from initial time-course of response experiments. With the HTPS system. dose-response curves are typically generated by introducing different concentrations of test compounds and controls to different wells of cells that have been previously introduced into the microtiter plate. For example. This flowing system provides several advantages over the static cells in wells approaches. These methods are reviewed elsewhere. Intracellular calcium.

CAMP). many of these for drug discovery programs. A schematic diagram of the high-throughput pharmacology system developed by Axiom represent valves capable of switching between buffer and Biotechnologies. A survey of the literature will reveal a vast number of transfected GPCRs. Some GPCRs receptors demonstrate constitutive activity when expressed at high levels. It has been suggested that this effect could be used for orphan GPCR screening (5).1 PHARMACOLOGY OF CAMP ASSAYS Introduction CAMPlevels are modulated by GPCRs coupled to Gs and Gi G-proteins.The limitations are not as readily ap- parent. the concentration of a given receptor can be increased. 3 3. Such changes result in cellular responses in screening efforts that identify hits that do not translate to nontransfected cells. and other proteins that modify GPCR signaling can result in abnormal drug-receptor behavior (6). At high expression levels. Researchers have turned to molecular biology. Such results are dependent on the kinetic parameters of these systems such as the association and dissociation rates of agonist and antagonist.Expression of receptors in cells containing different G-protein. These phenomena result from the activation of G-proteins by a small portion of the GPCRs. thus increasing the cellular response (Ca flux. The length of the pre-incubation and incubation loops dictates the incubation times for the antagonists and agonists prior to entering the fluorescent detector. gene cloning. For example. Several advantages are readily apparent. 3. humans. and overexpression as a means of facilitating HTS development. High Content Pharmacology mragonlsr Concentration Incubation loop Figure 3. When an antagonist is preincubated with its receptor. The open circles (0) sample (antagonist or agonist). and most importantly.2. G-protein receptor kinases.74 Rapid.2 is very flexible and permits the researcher to readily change incubation times for both agonist and antagonist. The HTPS system depicted in Fig. The closed circles (0)are mixing chambers for the cells and compound. a lowered maximal response is detected if the antagonist dissociates slowly from the receptor compared with the agonist on rate. Indeed. which act to activate . this active fraction is large enough to cause a perturbation in the basal signal downstream from the G-protein. the behavior has been observed before for the kinetics of acetylcholine antagonists in smooth muscle (4). the physiological responses. Cells are preloaded with fluorescent dyes sensitive to physiological endpoints such as intracellular calcium or membrane potential.

Fluorescence resonance energy transfer (FRET) from fluorescein to rhodamine occurs in the holoenzyme confirmation but is ablated on CAMP binding and subunit dissociation. tissues. blocking G-protein binding and activation. Energy transfer from a hydrolyzed luciferase substrate to GFP is possible on arrestin-GFP binding to luciferase-GPCR fusion protein. called arrestin. they are not useful in following channels that desensitize or inactivate after opening. Methods for evaluating ion channel activity based on changes in the fluorescent of dyes in response to changes in membrane potential have been developed.3 B-Arrestin Systems through several different readouts to provide information about GPCR signaling. the use of these dyes is restricted to specialized low-noise equipment (11). Arrestin binding also acts as an internalization signal. including the fact that channels represent an important class of drug targets for current therapeutic agents. membrane. . B-arrestin-GFP fusion protein binding to activated adrenergic GPCRs monitored by confocal microscopy was reported. forming fluorescent and non-fluorescent dimers and higher aggregates with concomitant changes in spectral properties in response to changes in cellular potential (12). A novel methodology in which the biosensor is generated from an engineered protein has recently been reported. These methods are useful for both HTS and follow-up secondary testing. Because the change in their electrochromic properties is limited to 1-10% changes in fluorescence per 100 mV. 3. interest in developing new drugs toward specific ion channels is high. Ion channel drug discovery is hampered by the low throughput of conventional patch-clamp methodology. provided evidence for GPCR homodimerization and the absence of a B-arrestin interaction with GnRHR (9. The fast styryl dyes respond to changes in membrane potential on the millisecond time scale. and intracellular components. therefore. 3. The fmt reported involved injection of CAMP-dependent kinase labeled with the fluorescent probes fluorescein and rhodamine on the catalytic and regulatory subunits into cells. and redistribute within the cytosol. This provides a purely molecular method of detecting intracelMar CAMP levels in real time (7. and several are commercially available. After a change in membrane potential. and membrane preparations have been reported. Cyclic nucleotide-gated calcium channels demonstrated use in coupling changes in CAMPlevels with intracellular calcium. 10). While interesting. Another class of dyes is the negatively charged oxen01dyes. This approach. These are the "slow" fluorescent probes. these dyes can take several minutes to re-equilibrate and. Voltage-sensitive probes based on FRET have also been developed (13). The changes in intracellular calcium are monitored by Fura-2 fluorescence at 380 nm. Secondary testing against a panel of cell lines. Many assays for determining CAMP levels produced in cells. A family of proteins called Gprotein-related kinases phosphorylate activated G-protein coupled receptors. these molecular methods have not yet proven their worth in the drug discovery process and require stable transfection of the CAMPsensor gene into each cell line before study. With the information available from sequencing the human genome. 4 I O N CHANNEL PHARMACOLOGY Many GPCRs are down-regulated after ligand binding through the concerted action of several proteins. A second molecule. binds to the phosphorylated GPCR. previously characterized for gene expression.2 Nondestructive Fluorescent Methods Some nondestructive methods for determining CAMPlevels in living cells are known. is useful in determining the specificity of action for these compounds. A modification of this approach has been used in which Renilla luciferase is attached to'the C-terminus of the GPCR. The readout is the CAMP-sensitive FRET signal from differently colored GFP mutant proteins coupled through the CAMP-binding domain.4 Ion Channel Pharmacology or inhibit adenylate cyclase. Arrestin binding to phosphorylated GPCR has been adapted Ion channels are attractive drug targets for several reasons. 8). requiring the introduction of two chimeric proteins into cells.

. which anchors in the outer leaflet of plasma membrane of cells with the courmarin dye. which has been defined as pro- . However. LQTS results in cardiac arrhythmias. and paralysis).Rapid. Cell death can result from stimuli resulting in apoptosis. Under hyperpolarizing conditions. Assays that could help rank order "hits" from HTS and early medicinal chemistrv .g. for example. ventricular fibrillation. The FDA now recommends that all drugs be screened against the HERG channel before release to market. resulting in increased FRET. 5 CELLULAR TOXICOLOGY ASSAYS Compound toxicity is an issue that is often addressed later in the drug development process than potency. because project costs increases rapidly with time. However. cost effective. This serious and sometimes fatal cardiotoxicity has led to withdrawal of otherwise promising drugs from the marketplace (21). One such ion channel is the HERG protein. A noteworthy example is the acquired long QT syndrome (LQTS) associated with blockade of cardiac ion channels (14-16). The need exists for a sensitive. In addition. ready access to structure-activity relationships regarding compound-related toxicity would be useful in directing medicinal chemistry efforts to reduce the toxic aspects of a compound as well as improving on other desired properties. Y652 and F656. and protein synthesis (e.A large number of drugs have been found to elicit adverse side effects through inhibition of the HERG channel. cardiac arrhythmia. thereby inducing LQTS (19). Under depolarizing conditions. and it is difficult to identify the molecular target of any drug interactions. the oxonol dye equilibrates toward the cytosolic side of the plasma membrane. Developing useful assays for screening drug effects against the HERG channel has become extremely important to the pharmaceutical industry. and can lead to sudden death. Several different mutations in HERG are known to cause inherited LQTS (18). patch-clamp technology is expensive. teratogenesis). Interaction of compounds with a toxicity target can result in a wide variety of in cell phenotypes (22). These assays are useful not only in primary screening but also for secondary follow-up assay for hit compounds. the negative membrane potential will distribute the oxonol toward the courmarin dye in the outer portion of the membrane. The FRET donor is a courmarin dye linked to a phospholipid. These assays have proven to be high throughput but i f f i often suffer from decreased sensitivitv or d culty in maintaining cell lines. late stage failures caused by unexpected toxicity are undesirably expensive. the antipsychotic agent pimozide and the gastric reflux medication cisapride. low throughput. Cell culture-based assays have been developed that use fluorescent dyes sensitive to plasma membrane potential in conjunction with plate readers and flow cytometry. may be responsible for the promiscuity of drug binding by the HERG channel (20). It would be greatly advantageous . apoptosis. and physiologically relevant HERG assay. FRET assays with response times of only a few seconds provide advantages over oxonol redistribution dyes. and vasculature (leaky blood vessels). The impairment of critical cellular functions can result in systemic toxicities such as those associated with the neuromuscular system (tremor.to include some measure of compound toxicity earlier in the drug discovery process. Recent evidence suggests that two key residues.Deregulation of gene expression can result in inappropriate cell division (neoplasia. high-throughput. However. The patch-clamp method has proven very powerful in its sensitivity and ability to evaluate channels at the single molecule level. and far removed from a physiologically relevant context.efforts would save time and cost derived from pursuing leads with serious cellular toxicity liabilities. and the FRET signal is reduced. respon- sible for the rapid component of delayed rectifier K+ current in the myocardium (17). Some estimates indicate that up to one-half of all compounds under review for market approval may elicit LQTS side effects through the HERG channel. difficult. torsade de pointes. it is very low throughput. this method is not always ethical. peroxisome proliferation). High Content Pharmacology In this method a voltage-sensing oxonol dye is used as the FRET acceptor. Drug screening in humans or animals is most physiologically relevant. renal (microtubule function).

Annexin V binding to PS on the outer membrane is a characteristic of early stage apoptosis (29). Necrosis is another classical cell death pathway in which cells lose membrane integrity. Flow cytometer-based assays for several other apoptotic indicators have been developed. both indicators of apoptosis. Cellular DNA synthesis can be determined by tritium-labeled thymidine (radiometric detection) or bromo-deoxyuridine (antibody detection) incorporatio'n and thus is a direct measure of cell ~roliferation and can be related to toxic and Gostatic effects of test compounds on proliferating cells. as part of the signaling to macrophages. 24). An advantage of measuring several cytotoxicity endpoints simultaneously is that dose and mechanistic properties of moderately and highly toxic compounds are discriminated. In apoptosis. and compound metabolism may influence the response of cells to toxic agents. 31). swell and burst. Apoptotic cells are removed by macrophages in response to the signals such as the exposure of phosphatidylserine to the outside of the cell before complete loss of plasma membrane integrity. Assays that measure mitochondrial function and cell viability and growth. the characteristics of these apoptotic. and plasma membrane and nuclear blebbing (23. DNA damaging agents. thapsigargin)." Apoptosis can be initiated through TNF receptor family signaling coupled to caspase activation. phosphytidylserine (PS) "flips" from the inner to outer side of the cell membrane. necrotic. and rate and extent of thiol oxidation (24. WTS. formation of cytoplasmic vacuoles. and metabolic pathways and environmental factors form a continuum of morphological and biochemical indicators of cell death. Elevated Ca2+ levels may be important in apoptosis by activating nuclease activity. Cellular factors determining which pathway a cell follows in response to toxic exposure include caspase activity. coulter counter. Thus. factors such as cell cycle. and it would be more useful to be able to readily characterize the toxicity mechanism for hit and medicinal chemistry compounds than simply to classify cells as "alive" or "dead. are considered good early indicators of compound-induced cellular toxicity. Dye exclusion is readily measured by direct cell counting on a hemocytometer. in particular Ca2+ and H+. cell type. or flow cytometer.grammed cell death. Changes in intracellular ion levels. cell membrane integrity. active efflux mechanisms. intracellular Ca2+. Ca2+ ionophores. degree of ATP depletion. are available.and Ca2+-sensitivefluorescent probes (26-28). the term "cytotoxicity" is somewhat imprecise. levels of reactive oxygen species. DNAfragmentation assays are also used to discriminate necrotic from apoptotic cells (30. and involves a cascade of events. duration of exposure. such as trypan blue and propidium iodide (PI). reduced glutathione concentration. Other apoptosis triggers included pharmacological agents (staurosporine. Assays for reactive oxygen species and cellular free glutathione content. and a variety of chemical toxicants (24). These cellular components often elicit an inflammatory response leading to further damage to surrounding tissues. The colored species generated is proportional to the number of viable cells. and intracellular pH are useful indicators of cell-based toxicity. Together. These properties are best determined in single-cell analysis as described in the next section. Intracellular pH and Ca2+ levels are readily determined with a variety of H+. Dyes that do not cross the plasma membrane. ATP. and spill their contents into the extracellular space. membrane potential. are excluded by cells with intact membranes and are an indicator of cell viability. extent of intracellular Ca2+ increase. Additionally. and others are reductase substrates that are reduced in the mitochondria of living cells to colored formazan dyes readily detected by light absorbance. The tetrazolium salts such as XTT. 6 HT-FLOW Many of the cell-death assays described above are readily performed by flow cytometry . MTT. End labeling of the fragmented DNA followed by staining yields signal indicative of the characteristic DNA fragmentation pattern (32). 25). including caspase activation leading to cytochrome C release from the mitochondria and subsequent degradation of chromosomal DNA into distinct fragments.

Drug discovery and development follow-up to a typical HTS program requires the assaying of hundreds to thousands of compounds identified in screening. Introduction of compounds to cells in flow provides another dimension to standard flow cytometry. We have previously reported a high-throughput sample delivery system for FCM (34). and data interpretation and novel cell pharmacology assays. Early physiological respon: such as changes in pH.he FL1 histogram for a single sample. Simultaneous determination of mitochondrial membrane potential. analysis at the single cell level. 24 wells. One major disadvantage of FCM is the need to handle each sample manually and the resulting low throughput. including hiigh content toxicology data. Each vertical strip in the top pane11 from an individual well of a 96-well plate loalded with fluorescent beads. and high sensitivity. compound usage. High Content Pharmacolc (FCM). of the plate are shown. determination of light scattering properties of cells yields information that is not available in standard 96-well plate reader assays. 3. mation MoFlo FCM equipped with a Mosik:it0 autosampler. Differential analysis of cellular responses to test compounds is possible because the different cell types or differentiation states can be tagged with specific cell-surface markers. and intracellular glutathione by FCM were used to show that loss of mitochondrial membrane potential and glutathione depletion are nearly simultaneous and could not be uncoupled (33). dista handling. Hi!5hthroughput compound screening by flow CYtometry provides both a novel set of issues regarding assay cost. Forward scatter is commonly used to measure cell size and decreases early in apoptosis while side-scattered light is proportional to cellular granularity and often increases during apoptosis (34). The data were collected on a Cyto- Wells 6.1 is Figure 3. and this can be coupled to cell-type specific . Neuronal cells can be derived through differentiation of precursor cells. Each vertical stripe:in the upper panel of Fig. readouts such as light scatter. The advantage of single cell multiparameteric analyses includes the ability to test mixed cell populations. Data I collected on a Cytomation MoFlo FACS equip]ped with a Moskito autosampler. Thus. and can be combined to yield multiparametric assays that measure several complementary indices of toxicity. The bottom panel represents a onevas dimensional histogram for a single well. such as those obtained upon cell differentiation. FCM has advantages over conventional plate-based assays with regard to multiplexing. FCM also provides unique advantages for mixtures of cell types. In particular. intracellular calciuL m .3. reactive oxygen species generation. which is capable of samplinj za 96-well plate in 4 3 min.3 shows the side sc:atter for a single well and the lower panel is 1.Rapid. or cherry picked by structural similarity to original hits and generated by focused combinatorial chemistry around active compounds. and membrane potential can be monitor ed. available to the pharmacologylmedicinal chemistry drug develcOPses ment team. while not limited to FCM. Two rows. The HTPS system described earlier has been adapted for flow cytometry aind increases the information content and expands the repertoire of assays. FCM is therefore well suited to measure the multiplicity of events that occur when cells respond to toxic agents and gives a wealth of information that can be used to identify the mechanism of compound action and direct chemical optimization efforts accordingly. All of these issiles are under consideration in the developmenlt of the next generation of HTflow instrumentation. and these can be stained for the presence of neural cell adhesion molecule.

Drug pharmacokinetics are modulated in part by CYP metabolism and patients on multi-drug regimens can experience unexpected changes. Alternately. (encoded by ABCB1. Fluorescence methods for ion channels may yield new lead compounds. coupled through calcium and CAMP. Flow cytometry methods are becoming more widely used in the pharmacological testing of drug potency and toxicity. respectively. Thus. Tax01 activates the SXR receptor. PAR.Most are based on changes in the fluorescence of a known CYP substrate and specific CYP isoform inhibition in the presence of test compounds. but better methods may be needed to over come some of the problems still associated with secondary testing in this class. allowing assays to be performed on mixed cell populations. such as cyto- . SUMMARY 8 Pressure to make better decisions earlier in the drug discovery process is high. Rapid. The HT-flow system has been used to measure cellular responses to test compounds at rates of 3-4 compounds per minute as well as determine the relationship between receptor occupancy and cell response (35). is a broad specificity pump that is responsible for efflux of compounds from the intestine. Increased CYP activity can result in toxicity as well. The cytochrome P450 (CYP) enzymes metabolize xenobiotics and evolved as a defense against accidentally ingested toxic compounds. Secondary testing in nontranfected cell-lines provides the optimal balance between throughput and disease relevance and target environment. can also result in more rapid drug clearance rates. but change the ratio of receptor to regulatory molecules and may alter pharmacology. P-glycoprotein expression is regulated by the orphan nuclear receptor SXR (other names are PXR. Another protein specifically targeted for secondary pharmacological screening is the ATP- dependent pump. Taxol's efficacy is decreased by both increased excretion and metabolism through PGP and CYP2C8/CYP3A4 induction. 7 HIGH CONTENT METABOLIC ASSAYS Efforts for the rapid screening of drug adsorption metabolism have intensified as the importance and relevance of the parameters for drug discovery have become more apparent with every failed and recalled drug. modulated at least in part through xenobiotic binding to the nuclear hormone receptors such as CAR. lower serum concentration. and decreased efficacy.8 Summary markers. thus blocking their uptake. are widely practiced. CYP gene induction. Several high-throughput assays for test compound inhibition of CYPs have been reported (36-38). these molecules are known to regulate the expression of a particular cytochrome P450 Cyp2b10. drug efficacy is reduced by excretion. Drug inhibition of CYPs can lead to increased drug levels and increased risk of toxic side effects. genetically engineered cells provide signal amplification and "universal" methods for detecting receptor activation. in drug mean residence time and maximum serum levels. and this toxicity is absent in CAR knockout animals (39). high content methods are available that can facilitate decision-making." In the mouse. also known as MDRl). Reporter assays for lead compound effect on CAR and SXR activity could add useful information to drug development programs by flagging compounds for possible drug-drug interaction and metabolic liabilities. PRR or NR112). High-throughput sampling is made possible with plug-flow coupling of ambient samples and pressurized flow cytometry systems. both positive and negative. The nuclear hormone receptor CAR controls the expression-based response to a class of molecules called the "phenobarbitol-like inducers. Assay for compound associated metabolic liabilities. In addition to metabolic degradation and inactivation. increasing the expression of both Pglycoprotein and the cytochrome P450 isoforms CYP2C8 and CYP3A4 (40). P-glycoprotein is overexpressed in some tumor cells and is responsible for cancer drug resistance and is part of the blood-brain barrier where it helps protect the brain from exogenous agents. Assays for native expressed GPCRs. drug-drug interactions are important in both toxicity and efficacy. Cocaine causes acute liver toxicity in mice previously exposed to CAR agonists. P-glycoprotein. PGP.

Nut. Cancer.. Bockaert. Journot. 9. Eur. Natl. 39. Chen. 3. et al. 13. Sci. Biol. Curran. Chauret. 25-29 (2000). 14.257-259 (1998). J. 40.. 80. NY Acad. 4. Sci.Pharm. Darzynkiewicz. 24. Cytometry. Proc. Symp. et al. M . 20. 30. 97. M. et al. W .. USA. Cell. D. Rang. 920-923 (2000). Sussman. Ser... Tonini. 28. A10 (2000).584590 (2001). P. when all of these methods are in place. Klaassen in C. B. Cell. B. 249-254 (2000). 15. H. Tian. D. Photochem. S. Flow Cytometry Systems for Drug Discovery and Development. R. P.. Trends Pharmacol. Tsien. et al. and A. C .. M . Nicotera. Heart J. Synold. 28. Proc. M. Y. Sci. Darzynkiewicz. et al. Nature. 419-424 (1988). Proc. 17-19 (1999)..1232-1237 (2000). Pharmacol. 125-134 (2000). Acad. et al. 12. 325. 83-90 (2000). Sanguinetti. Biotechnol. J. 93. Nut. R. Relat. 18. J. 3921. Sigler. N. Gonzalez. et al. Screen. 21. S. 99. 16. E. 299307 (1995). Zaccolo. Ziogas. 4. Y. Sci. Farkas. J. Tsien. Soc. Ransom. Ferrando-May. et al. Green. Melamed. 313-316 (2000). J. Y. 26. et al. 8. 32. A.. Lond. Loew. 10.. 101-124 (1996). Cascieri. Pharmacol. Acad. 1996. J.90-100 (2000). Kelly and N . 12736-12743 (2001). 30. McGrawHill. Brand. et al. Endocr. R.. Plasek and K. A. T. Cell.. R. W . Mitcheson. J. R. 164. Acad. 431-439 (1999). W . S. Med. Christopoulos. 36. Montana. 3684-3689 (2000).. Roden. et al. 37. M . D. 23. 276.... Natl. and B. 5. Sci. Proc. Nut. Edwards.215-226 (1999). C. Ed. 33.12329-12333 (2000). et al. Aubry. 157-168 (1998). 81.. 812. et al. Biochemistry.. Today. M . 38. 29.327-345 (1986). 12. 1-4 (2000). et al. Kroeger. Screen. 31. Biochem. Science. et al. D.. Mol. I. 2. P.. J. G. Zipkin. 1996-2012 (1996). 2. M . 6. Tsien.. 22. USA. Finally. J. Natl.. S. X. and Z. Aliment. 276.. I..55-70 (1997). J. S... M . K. B Biol. M. J. 18. Sci. 6. 11. Biol. BioCentury. Nagai. et al. Photobiol. 281. New York. Camm. 94. et al. 21. McConkey. 35. R. J. 407.151158 (2000). Toxicol. 5... Gorczyca. et al. L. 1585-1591 (1999). 21.Venhorst. A. M. Methods Mol. 40. Soc. Biomol. 158. 8. 197-204 (1999).. 34. Drug Discov.. J. Y. Cell Biol. Gregus and C. 6. J..Rapid. High Content Pharmacology chrome P450-mediated metabolism or gene induction. 66. M . Li and Z .. Proceeding of SPIE. 13. USA. T . L. Klaassen. Ann. can be developed. Leist. 25. 19. et al. Ther. J. Macho. D. 7. REFERENCES 1. 33. 11.217-238 (1998).. A. et al. 57. J.. M .. 488-510 (1966). Cell Prolif. 91.. Gorczyca. J.2208-2212 (1996).69-73 (1999).. V .. 253-261 (1997).376-381 (2000). 17. 27. Pharmacol.. Biochem. D.. Casarett and Doull's Toxicology. Angers. and L.P. Soc. 795-803 (1995). Forman. Biomol.. Sci. Lew. Trends Neurosci. 4612-4619 (1997).. Physiol. Sanguinetti. Immunol. 28. 127-156 (1989). M. et al. Lett. J. Dussault. Wei. the challenge is the integration of these methods and their data into successful drug development from hit to lead. Eur. . J.. 97. Methods Cell Biol. 37. Circulation.. and E. Chem. and L. Anal. H. Eur.. 102.4536-4539 (1989). C. Gen. 7. 571-579 (1995). E. et al.

2 Aptamers as Diagnostics.3.6 Enzymes as Drug Targets. 96 4. 82 1.3 SELEX or In Vitro Evolution. 84 3. 94 4.1 Second-Generation Protein Therapeutics.2 Combinatorial Biosynthesis and Microbe ReEngineering. Abraham ISBN 0-471-37028-2 O 2003 John Wiley & Sons.4.1 Reagents for Screening.3 Cell-Specific Targeting. 92 4. 105 5 Future Prospects. .2 Advent of Recombinant DNA Technology.4.8 Cellular Adhesion Proteins. 88 3. 100 4. Pennsylvania ADEGBOYEGA K.4 Phage Display. 91 4.4 Future Directions. 91 4. 92 4. 83 2 New Therapeutics from Recombinant DNA Technology. OYELERE Rib-X Pharmaceuticals.7 Receptors as Drug Targets.1 Preparation of Phage Display Libraries.3 Epitope Mapping. Volume 2: Drug Development Edited by Donald J. 95 4.for Structural Biology 4.2 Phage Display Selections against Purified Proteins.3.4.2 Antibody-Based Therapeutics.5 Reagents . 89 4 Genetically Engineered Drug Discovery Tools. 94 4. Inc. 83 3 Protein Engineering and Site-Directed Mutagenesis. 93 4.4 In Vivo Phage Display. Inc.4.CHAPTER FOUR Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery SOUMITRA BASU University of Pittsburgh Center for Pharmacogenetics Department of Pharmaceutical Sciences Pittsburgh. 95 4.1 Aptamers as Drugs.1Chemistry-Driven Drug Discovery. 82 1. 94 4. Connecticut Contents 1 Introduction. 84 3. 96 4. 96 . New Haven. Studies. 106 ~ Burger's Medicinal Chemistry and Drug Discovery Sixth Edition. 88 3.

however. more fundamental questions remain. 3). During the 1980s. medicinal chemistry had clearly become the cornerstone technology of modern drug discovery. became the hallmark of new drug discovery. however. There is little doubt that the task of discovering new therapeutic agents that work potently. What determines a useful biological - . specifically. the emphasis has changed from isolation of active constituents to creation of new. This has forced the science of medicinal chemistry. the changes the industry has undergone have been profound. Since then. Systematic development of structure-activity relationships. Even with these successful compounds. The success of these two drugs sparked a realignment of chemistry-driven pharmaceutical research. From ancient times to the beginning of the last century.I Chemistry-Driven Drug Discovery The exacting process of discovering new chemical entities that are safe and effective drugs has itself undergone many changes. Receptors and enzymes were still considered as functional "black boxes" whose structures and functions were poorly understood. however.-antagonist for the treatment of ulcers (51. and without side effects has become increasingly important and coincidentallv " more difficult. In the 19209. this approach yielded many valuable compounds. As the companion chapters of this volume describe. such as the penicillins. the first efforts at understanding why and how morphine works in terms of its chemical structure were initiated. These include the natural product HMG-CoA reductase inhibitor lovastatin for the treatment of hypercholesteremia (8)and the antihypertensive angiotensin I1 receptor antagonist losartan. challenges for mass production of medicinally valuable natural products. By the mid-1960s. an understanding of the actions of drugs at the molecular level was often lacking. Although serendipitous at best. the art of rational drug design has undergone an explosive evolution. By the late 1950s. mainly alkaloids and the like. each of which was prompted by the introduction of some new technology (2. The first successful attempts at actually designing a drug to work at a particular molecular target happened nearly simultaneously in the 1970s. a selective H. treatment for illness or disease was based mainly on folklore and traditional curative methods derived from plants and other natural sources. 1 . most notably the benzodiazepine tranquilizers chlordiazepoxide and diazepam (4). and captopril. During the 1940s. The isolation and chemical characterization of the principal components of some of these traditional medicines. Within the last century. spawned the development of the modern pharmaceutical industry and the production of drugs in mass quantities. Even with these major advances in the medicinal and pharmaceutical sciences. the process of drug discovery amounted to little more than evaluating available chemical entities in animal models suggestive of human disease. synthetically optimized from a chemical library screening lead (9). potent chemical entities. were conquered. an angiotensin-converting enzyme inhibitor for hypertension (6). mechanism-targeted design and screening combined to produce a number of novel chemical entities. with the discovery of cimetidine. Advances in medical research that have provided new clues to the previously obscured etiologies of diseases have revealed new opportunities for therapeutic intervention. making use of sophisticated computational and structural methodology to help in the effort (7). Even then. even to the point at which predictions about activity might be made. This evolution from folklore to science is responsible for the thousands of pharmaceuticals available worldwide at present (1). advances in synthetic organic chemistry enabled the generation of multitudes of novel structures for broad testing into the major focus of the modern pharmaceutical industry.82 Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery 1 INTRODUCTION Discovering new drugs has never been a simple matter. once founded almost solely in near-blind synthesis and screening for in viuo effects. to become keenly aware of biochemical mechanisms as an intimate part of the development process.

i. with few exceptions. its subsequent association with an effector protein. another research avenue synergistic with existing discovery technologies is necessary. was suddenly expanded by the introduction of the first biotechnologically derived products in the 1980s. Over 400 products generated by biotechnology are estimated to be somewhere in the development pipeline or approval process.bio. has occurred in parallel with the development of contemporary medicinal chemistry (11-14). their dissimilarity to traditional medicinal agents does not end there.2 Advent of Recombinant DNA Technology protein and its identification and production have been considered the domain of molecular genetics. but also true therapeutics for both acute and chronic care. In 1985 another milestone was achieved when Genentech became the first biotechnology company to be granted approval to market a recombinant product. expression. These biotechnologically derived therapeutics are large extracellular proteins destined to be.org/er/approveddrugs.16). Although the task of drug development has now been refined into a process without major unidentified obstacles. Many of these products also satisfy urgent and previously unfulfilled therapeutic needs. to produce not only natural proteins for the treatment of deficiency-associated diseases. As in the past. provide an answer to at least one of the persistent problems of new lead discovery. and sometimes prejudices. synthesis of new chemical entities as therapeutic agents. . human growth hormone. asp.2 New Therapeutics from Recombinant DNA Technology 83 pr'operty? And how is it measured in the discovery process? The answers can determine the success or failure of any drug discovery program. receptors. to the more traditional processes of drug discovery.e. respectively. If modulation of biochemical processes by a low molecular weight compound has been the traditional goal of medicinal chemistry. However. The variety of products-from hormones and enzymes to receptors. these proteins were developed not because of the novelty of their structures. Because a comprehensive review of the genetic engineering of important proteins is well beyond the scope of this volume. injectables for use in either chronic replacement therapies or in acute or near-term chronic situations for the treatment of life-threatening indications (15. then association of a biological effect with a distinct The traditional role of the pharmaceutical industry. The traditional products of biotechnology research share few of the traits characteristic of traditional pharmaceuticals. vaccines. Their discovery hinged on recognition of a useful biological activity. this chapter will instead highlight some novel examples of advances in recombinant DNA technology. but because of the novelty of their actions. with respect to both exciting new pharmaceuticals and potential applications of recombinantly produced proteins. 2 NEW THERAPEUTICS FROM RECOMBINANT DNA TECHNOLOGY The evolution of recombinant DNA technology. Unlike most low molecular weight pharmaceuticals. 18). founded in decades of empirical success (10). and monoclonal antibodies-seeks to treat a broad . 1. and the genetic identification. A comprehensive list of FDA approved biotechnology drugs is available on the web at www. Industry estimates show the upward trend in biotechnologicallyderived drugs continuing into the new millennium. from scientific innovation to phannaceutical discovery process. The approval of recombinant human insulin in 1982 broke important ground for products produced by genetic engineering (19). or hormones. be they enzymes. These events set an entire industry into motion. because both the observation of a usehl biological property in a novel molecule and the optimization of structure-activity relationships associated with ultimate clinical candidate selection have rightfully relied heavily on practices. The application of recombinant DNA technology to the identification of proteins and other macromolecules as drugs or drug targets and their production in meaningful quantity as products or discovery tools. and production of the effector by the application of recombinant DNA technology (17. the challenge to bring the discovery of novel compounds to a comparable state of maturity remains.

allowing often even crude first attempts at producing chimeric proteins to be successful.e. despite this period of phenomenal growth for recombinant DNA-derived therapeutics. and site-specific mutations are represented by the code for the wild-type amino acid. four properties functioning in concert [i. expression. with the goal of designing and constructing new proteins with novel binding. especially antibiotics (26. amino acids are denoted by their standard one-letter codes. 3. in any protein. one of the most studied recombinant products (34b-36). fermentation procedures can be directly improved by strain optimization techniques. assuming that expression of the gene and subsequent proper folding have successfully occurred. and marine organisms by using genetically engineered reagents is becoming of special importance as more of the relevant targets identified by molecular biology operate in obscure or even unknown modes.34a). and proteins seem extremely tolerant of domain substitution.1 Second-Generation Protein Therapeutics 3 PROTEIN ENGINEERING AND SITE-DIRECTED MUTACENESIS Rapid developments in the technique of sitedirected mutagenesis have created the ability to change essentially any amino acid. once touted to be limitless. leading to few unsynthesizable mutants. protein physiology. substitutions of residues not involved in internal hydrophobic contacts are extremely well accommodated. Several surprising observations have been made during the short period that this technology has been available: amino acid substitutions lead. The art of finding new natural product-based lead compounds by screening fermentation broths. substrate specificity. and the code for the replacement amino acid (31). binding affinities (for a receptor or hormone). in general. to highly localized changes in protein structure with few global changes in overall folding. The structural diversity provided by natural products combined with the ability to test molecular biology driven biochemical hypotheses has already become an important route for the discovery of new therapeutics (28-30). metabolism. More exciting is the possibility of producing hybrid antibiotics that combine desirable features of one or more individual compounds for improved potency. for tissuetype plasminogen activator (t-PA). - specificities (either for substrate or receptor). including genetic recombination and cloning. efficacy. Throughout this chapter. plant sources.or domain-specific mutant proteins for structure-function investigations. the residue number. has instead become more realistically defined to include not only the actual recombinant products and the difficulties inherent in their production but also many spin-off technologies. including diagnostics and genetically defined drug discovery tools (20-22). The concomitant changes in protein folding and tertiary structure. bioavailability. 27). even among unrelated proteins.84 vledicinal Chemistry and Drug Discovery Application of Recombinan~tDNA Technology in ! range of clinical indications thought untreatable just two decades ago. the promise of biotechnology. or specificity.As an example. 32). clearance.. . or catalytic activity (for enzyme active site mutants) are all effects that are measurable against the "wild-type" parent. or catalytic activities (31. or even substitute or delete whole domains. Potency. bioavailability. The implications of this technology for the discovery of new pharmaceuticals lie in two areas: secondgeneration protein therapeutics and site. One particular area of traditional pharmaceutical research in which recombinant DNA technology has made a profound impact has been the engineering of antibiotic-producing organisms (23-25). fibrin affinity. and pharmaceutical formulation challenges presented by the natural protein suggest that second-generation products might be engineered to alleviate the particular problem at hand. Always an important source of new bioactive compounds. stimulation of t-PA activity by fibrin and fibrinogen. The parent proteins to which this technology has been applied extend across the range of recombinant products already approved and those in advanced stage of clinical evaluation (33. and manufacture of proteins as therapeutics involve the same problems encountered in the development and successful clinical approval of any drug. producing desired therapeutic improvements. binding The cloning. However.

(3) genetic tagging of diseases with the consequence of developing designer medications that best combat an ailment. resistance to PAI-1. and rapid hepatic elimination by receptor-mediated endocytosis (38). can be used to combine domains from different proteins to produce chimeric constructs that incorporate multiple desired properties into a single final product or reagent. protein A binding. compared with 0. Furthermore. cellmediated cytotoxicity (ADCC) toward HIV-infected cells and is efficiently transferred across the placenta of primates (43). which promises to unravel how genetic make up and variation thereof affect human response to medication (45). aided by a genetic prescreening of candidates. Dimeric constructs from human (CD4-2yl and CD4-4yl) and mouse (CD4-My2a) IgG and a pentameric chimera (CD4-Mp) from mouse IgM exhibit evidence of retained gp120 binding and anti-HIV infectivity activity. these chimeric constructs incorporate functions such as Fc receptor binding.A consensus structure combining the major domains of t-PA has been predicted based on the significant sequence homology with other serum proteins and serine proteases. all of which are imparted by the Fc portion of immunoglobulins. and (4) efficient design of clinical trials.3 Protein Engineering and Site-Directed Mutagenesis and sensitivity of the enzyme to inhibition by plasminogen activator inhibitors (PAIS)] are responsible for the localization and potentiation of the lytic reaction at a clot surface and are readily analyzed using molecular variants (37). It is becoming clear that genetic variations play critical roles in patients' response to certain medications. complement fixation. very low binding to plasminogen in the absence of fibrin. increased fibrin specificity. in an effort to overcome the short plasma half-life associated with soluble CD4.1) have been recombinantly constructed from the gpl20-specific domains of CD4 and the effector domains of various immunoglobulin classes (41. and in vivo increased potency and decreased systemic activation of plasminogen when administered by bolus dose (40). and placental transfer. more recent applications of this technology resemblance to bear a less straight-forward medicinal chemistry-driven drug discovery paradigms. For instance. rapid inactivation by PAI-1. increase in the efficiency of plasminogen activation by 500-fold in the presence of fibrin. made up of the Kringle 2 and protease domains. Both CD4-2 yl and CD4-4. BM 06. 42). Medications that were judged ineffective by traditional validation methods in a random patient population may be found beneficial to a population of patients having an overexpressed "susceptibility gene(s).25 h for rCD4. It is widely held that advances in pharmacogenetics will revolutionize drug dispensation and drug discovery processes. has been reported to have the same plasminogenolytic activity but a lower fibrin affinity compared with wild type t-PA (39). Differential expression of drug targets and or metabolic enzymes has been shown to lead to differences in efficacy and toxicity profiles of drugs in section of population that harbors this genetic variation (44).7 and 48 h. (2)an increased understanding of the molecular "uniqueness" of diseases.yl show significantly increased plasma half-lives of 6. these same recombinant techniaues . 4. moderately high affinity for fibrin. KHRR 296-299 AAAA) was demonstrated to have the combined desirable properties of decreased plasma clearance. the immunoadhesin CD4-2 yl (CD4-IgG)mediates antibody-dependent. The complexity of this structure is reflected in its functional multiplicity: efficient production of plasmin by cleavage of the R560-V561 bond of plasminogen. a recombinantly engineered t-PA deletion mutant [t-PA del (V4-E175)]. When fully realized. Another variant of t-PA (T103N. However. chimeric molecules termed immunoadhesins (Fig. pharmacogenetics. Molecular biology and its associated techniques feature prominently in bringing to birth an interdisciplinary field. with a better chance of success." . respectively. the gains of pharmacogenetics will positively impact drug discovery processes in numerous ways including the following: (1)identification of new and novel therapeutic targets.022. In addition to dramatically improved pharmacokinetics. Although the systematic changes exemplified by t-PA site-directed mutagenesis studies of medicinal chemare the rDNA equivalents istry [multiple analog synthesis for structureactivity relationship (SAR) development].

was derived from a CHI-containingCD immunoadhesin by oligonucleotide-directed deletional mutagenesis. revealed that it exists in two isoforms: constitutive COX-1 and inducible COX-2 (Fig. CH2. A prolonged use of most NSAIDs results in gastrointestinal (GI) toxicity (46). respectively. CD4. which may be debilitating enough to require hospitalization in many patients. Disulfide bonds are indicated by S-S.and IgG1-derived sequences are indicated by shaded and unshaded regions. championed by elegant biochemical and recombinant DNA studies. by Merck Frosst and Monsanto-Pfizer. especially GI tract toxicity (49). whereas COX-2 is active only at the onset of inflamma- tion (47. 48). The discovery of the COX-2 gene stimulated intensive studies aimed at verifying this hypothesis (50. thereby interfering &th the production of the protective prostaglandin products of COX-1 in addition to their pain relieve activity. It was hypothesized that selective inhibitors of inflammation associated COX-2 may produce a better drug profile and possibly avoid many side effects caused by non-selective NSAIDs. CD4 immunoadhesin consists of residues 1-180 of the mature CD4 protein fused to IgGl sequences. COX. Clinically useful NSAIDs such as aspirin.2). Structure of CD4 immunoadhesin. recombinant DNA technology is making it possible to decipher the roles of disease subtypes and genetic variability in . Alternatively. and ibuprofen exhibit their anti-inflammatory and antipyretic activity by inhibiting COX. The clinical and commercial success of Vioxx and Celebrex (1 and 2). validated COX-2 as a new target for anti-inflammation and antipyretic therapy. A recent success story in new target identification and validation is seen in the introduction of two new nonsteroidal anti-inflammatory drugs (NSAIDs). However. beginning at D216. which lacks a CHIdomain. and CYT is the cytoplasmic domain. the safety profiles of these drug showed that they do not have many toxic side effects of traditional non-selective NSAIDs (52-54). early NSAIDs inhibit both COX isoforms. and the parent human CD4 and IgGl heavy-chain molecules. Furthermore. The variable (VH)and constant (CHI.Hinge. COX-1 is always expressed and mediates the synthesis of prostaglandins that regulate normal cell functions. two highly potent COX-2 specific inhibitors. soluble rCD4. 4.and CH3)regions of IgGl heavy chains are shown.1. Vioxx and Celebrex. diclofenac. Until recently. Recent progresses in the understanding of the biology of COX. the onset of inflammation and pain has been linked to one cyclooxygenase enzyme. The immunoglobulin-like domains are numbered 1 to 4. Soluble CD4 is truncated after P368 of the mature CD4 polypeptide. expressed in Chinese hamster ovary cells and purified to >99% purity using protein A-sepharose chromatography (42). TM is the transmembrane domain. which is the first residue in the IgGl hinge after the cysteine residue involved in heavy-light chain bonding. The CD4 immunoadhesins shown.Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery The role of recombinant DNA technology in medicinal chemistry and drug discovery CD4 immunoadhesin Soluble rCD4 SS SS I I lgGl heavy chain +Fc domain + Figure 4. 51). respectively.

an overexpression of epidermal growth factor receptor-2 (HER-2) is known to occur in about 25-30% of women with breast cancer (55). e progression and response to drugs. The insightis gained from such studies are beginning to yield genetically engineered pharmaceuticals 1'or treating various human disease conditione1 including diabetes. a humanized anti-HER2 monoclonal antibody. For example. and viral diseases. (b) Binding stages of standard NSAIDs to arginine 120 to inhibit prostaglandin ynthesis by direct blockade of cyclo-oxygenase channel. (c) Differences between COX-1 and COX-2. Prostaglandin synthesis and inhibition in COX-1 and COX-2. multiple sclerosis. $ Specific blockade of COX-2. Herceptin. (a)Initial stages of prosaglandin synthesis.2.tein Engineering and Site-Directed Mutagenesis a) Prostaglandin synthesis from arachidonic acid ryr 285 (c) COX 1 and COX 2 are different COX I COX 2 Competitive inhibition b) N o n specific NSAlDs bind to COX 1 and COX 2 Time-dependent irreversible inhibition (d) COX 2 selective agents bind in COX 2 pocket COX 1 COX 2 to block entry o f arachidonic acid COX 1 COX 2 'igure 4. rheumatoid arthritis. diseam New markers that aid disease classification identified and explored as targets to are b~eing develop new and novel therapeutic strategies to bf?tter treat or manage diseases. cancer. and it is has proven . was recently introduced by Genentech to treat metastatic breast cancer.

using a technique termed homologscanning mutagenesis.. the central portion of helix 4 to the Cterminus. The alanine scan revealed a cluster of a dozen large sidechains that when mutated to alanine exhibited more than a fourfold decrease in binding affinity. Using an enzyme-linked immunosorbent assay (EL1SA)-basedbinding assay. leading to allergic reactions and reduced efficacy (59). details of which are beyond the scope of this chapter). An impressive example of a systematic search for a binding epitope is seen in the work used to define the human growth hormonesomatogenic receptor interaction (60. termed alanine-scanning mutagenesis was then applied. During the 1980s and early 1990s. Single alanine mutations (62 in total) were introduced at every residue within the regions implicated in receptor recognition. Recombination technology plays a key role in the development and commercialization of therapeutic antibodies. eight of nine antibody products on the U. and to a lesser extent. three discontinuous polypeptide determinants (the loop between residues 54 and 74. which measures the affinity of the mutant hGH for its recombinantly derived receptor.g. genetically engineered transgenic mice that can be used for production of humanized antibodies have been developed. - 3.2 Antibody-Based Therapeutics Antibody therapeutics can potentially treat diseases that can be as diverse as autoimmune disorders to cancer and infectious diseases. Whereas numerous examples of models of enzyrne-ligand complexes have been developed based on active-site modifications. Lately. complementarity-determining region (CDR) grafting and veneering techniques were established. By this analysis. the amino-termind region of helix 1) were identified as being important for binding to the receptor. monoclonal antibodies currently in development outnumber all other classes of therapeutics. because thev " were murine antibodies that induced human antimouse antibodies. Many of the residues that constitute the hGH binding epitope for its receptor are altered in close homologues. Antibodies are currently rated as an important and growing class of biotherapeutics. 58). their therapeutic uses were limited caused by immunogenicity in humans. even though the residues changed within each subset were usually distant in the primary sequence. which reduced the mouse portion of the sequence to less than 10%. such as placental lactogen and the prolactins.S. The technology uses the standard hybridoma techniques to produce fully humanized antibodies (59). The fulfillment of the many promises of pharmacogenetics is strongly hinged on the identification of unique markers that correlate genetic make-up to drug response. A second technique. In fact. (Public and private efforts are currently ongoing in identifying such informative markers.88 Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery beneficial in patients with metastatic breast cancer in which HER-:! is overexpressed (56).3 Epitope Mapping Site-directed mutagenesis technology has also been applied to one of the most challenging problems in structural biochemistry-the nature of the protein-protein interaction. The overall correct folding of the mutant proteins was deter- . a protein substrate to a protease or a hormone to its receptor) binding epitope. First. Other than vaccines. Molecular biology is and will continue to play frontline roles in the identification of these genetic markers. researchers discovered that (63) swap mutations that disrupted binding were found to map within close proximity on the three-dimensional model. 61). the next generations of antibodies were chimeric in nature with 66% human and 34% mouse produced through genetic engineering. the site-directed mutagenesis method is being extended to the formidable problem of defining the essential elements of a protein-protein (e. After the discovery of murine antibodies in 1975. 3. market are recombinant products (57. During the early days of monoclonal antibody use. segments of sequences (7-30 amino acids in length) from homologous proteins known not to bind to the hGH receptor or to hGH-sensitive monoclonal antibodies (Mab) were systematically substituted throughout the hGH structure by using a working model based on the three-dimensional folding pattern found by X-ray crystallographic analysis of the highly homologous porcine growth hormone (62).

biophysical studies. randomized mutagenesis techniques also have been directly applied to the ever-growing problem of antibiotic resistance. 2. >50-fold. localization of the regions of the protein critical to either structure or function can be accomplished.2-to 4-fold reduction. 10. 4.and Xay crystallography (66) revealed the presence f two overlapping binding epitopes (Fig. through which it causes imerization of two membrane-bound recep)rs to induce its effect.4 Future Directions In an intriguing example of what might be termed reverse small molecule design. Usig the receptor-binding determinants identied in these studies.3) n growth hormone.to 10-fold reduction. 4.to 10-fold. Application of this technique to TEM-1 p-lactamase (the . are shown: @. 4-fold increase.size-exclusion chromatography.to 4-fold. These results indicate that the omolog and alanine-scanning mutagenesis xhniques should be generally useful starting oints in helping to identity amino acid resi- dues important to any protein-protein interaction (67) and that these techniques have potential to provide essential information for rational drug design. The two sites are generally delieated by the large shaded circles.000-fold (64). and FA. and @. including calo. Using a mutagenesis technique that randomizes the DNA sequence of a short stretch (3-6 codons) of a gene. @. lined by cross-reactivity with a single set of ~nformationally sensitive Mab reagents. Map of alanine substitution in hGH disrupt binding of hGHbp a t either site 1 or site 2.3. severely limiting the useful therapeutic lifetime of most marketed antimicrobial agents (68). fluo2scence-quenching binding assay (651. This variant had an association ~nstantfor the hGH receptor that was inreased by more than 10. Sites where alanine mutations in site 1 cause changes in binding affinity for the hGHbp using an immunoprecipitation assay.metry. 3. The crystal structure mfirmed both the 1:2 hormone-to-receptor 3mplex structure and the interface residues lentified by the scanning mutagenesis maping technique. @. >lo-fold reduction. Bacterial resistance to increasingly complex antibiotics has become widespread. 4. Ed. Finally. followed by determination of the percentage of functional mutants expressed from the randomized gene. FA. a variant of human prolctin (hPRL) was engineered to contain eight iutations. Residues for which alanine mutations reduce site 2 binding are shown: Ed.to 50-fold.Protein Engineering and Site-Directed Mutagenesis Figure 4.

can be added. is still limited to the repertoire of the 20 natural amino acids encoded by DNA. (2) identification of a suppressor tRNA that can efficiently translate the amber message but that is not a substrate for any endogenous aminoacyl-tRNA synthetases. nucleophilicity..Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery Figure 4. The art of rDNA site-directed mutagenesis. In the past. The region identified as most sensitive to substitution are located either in the active site or in buried positions that likely contribute to the core structure of the protein (Fig. An early successful demonstration of this methodology involved replacement of F66 with three phenylalanine analogues in RTEM p-lactamase and subse- . a recent development of wholly in uitro translation method promises to dramatically increase the array of unnatural amino acids that can be substituted into proteins (70d). in theory. although there is a room for further development.4). with novel amino acids. aureus p-lactamase. such as increased or decreased acidity. Lines point to the position of individual libraries. but at an accelerated pace.4. Furthermore. The key requirements for successful site specific incorporation of (un)natural amino acids include the following: (1)generation of an amber (TAG) "blank" codon in the gene of interest at the position of the desired mutation.type enzyme (69). Such a library of functional mutant p-lactamases could. These requirements are individual hurdles that have been crossed in the course of development of this technology. Dark regions correspond to the position of random libraries. enzyme responsible for bacterial resistance to p-lactam antibiotics such as penicillins and cephalosporins) over a 66-codon stretch revealed that the enzyme is extremely tolerant of amino acid substitutions: 44% of all mutants function at some level and 20% function at the level of the wild. methodology for carrying out such mutations recombinantly has been successfully used. and (4) availability of a suitable in uitro protein synthesis system to which a plasmid bearing the mutant gene or corresponding mRNA and the acylated tRNA. be used to simulate multiple next generations of natural mutations. However. To effect more subtle changes in proteins. 4. Advances in nucleic acids synthesis have ensured a rapid access to aminoacylated CCA for semisynthesis of tRNAs bearing amino acids of interest (70a-70c).. Screening of new synthetic p-lactams against such a mutant library might then be used to discover compounds with the potential for increased useful therapeutic lifetime. Position of random libraries on a ribbon diagram of the homologous S. such changes were accomplished semisynthetically on chemically reactive residues such as Cys. although advancing rapidly. it has been proposed that site-directed mutagenesis using unnatural amino acids might offer the needed advantages. '(3) development of a method for the efficient acylation of the tRN&. or hydrogen-bonding characteristics without dramatically altering the size of the residue and without affecting the overall tertiary structure.

however. high flux screening of chemical. Assays to detect inhibitors of tat function by screening. which was then subjected to medicinal chemistry optimization to produce structure (4) as the ultimate clinical candidate (94). The ability to carry out large-scale. The inherent differences in these potential sources of drug design information. (91. now better enabled by use of recombinant reagents and techniques. The common practice of in vitro screeningfor enzyme activity or receptor binding using animal tissue homogenates (nonhuman. Even more to the point of human pharmaceutical discovery and design. Because Tat is necessary for HIV expression and SeAP expression is under the control of the HIV LTR.4 Genetically Engineered Drug Discovery Tools 91 quent determination of the kinetic constants k . two plasmids were cotransfected into COS cells: a gene for either Tat or the reporter gene for secreted alkaline phosphatase (SeAP) was put under the control of the HIV-1 LTR promoter (93). 75-89) also critically depends on reagent availability and consistency.Such rapid lead discovery highlights the continued importance of highly directed screening to drug discovery. natural product. 4 GENETICALLY ENGINEERED DRUG DISCOVERY TOOLS 4. especially from large combinatorially generated libraries requires that assay variations be reduced to the absolute minimum to ensure the ability to analyze data consistently from possibly millions of assay points. to screen for Tat inhibitors. because the tat trans-activator protein (one of the HIV-1 gene products) has been clearly demonstrated to regulate expression of the complete genome (90). The discovery of the HIV Tat inhibitor Ro 5-3335 and its eventual development as the analog Ro 24-7429 are successes from screening chemical libraries using recombinant reagents. 72). 74). This assay was standardized and used in high flux screening to identifjl structure (3). is the issue of species and/or tissue specificities. an inhibitor of Tat would necessarily lower the apparent alkaline phosphatase activity.1 Reagents for Screening An increasingly important application of recombinant technology lies not in new protein drug product discovery per se but in the ability to provide cloned and expressed proteins as reagents for medicinal chemistry investigations. of the mutants (70e). Subsequent applications have centered on the critical issue of the introduction of unnatural amino acid replacements into proteins. and recombinantly or synthetically derived diversity libraries (26-29. The artificial residues can probe effects on stability and folding governed by subtle changes in hydrophobicity and residue side-chain packing to a degree not possible using the 20 natural amino acids (71. In this instance. and K. Tat is a strong positive regulator of HIV expression directed by the HIV-1long terminal repeat (LTR) and as such constitutes an important and unique target for HIV regulation. and therefore nontarget) has begun to give way to the use of solid-phase or whole-cell binding assays based on recombinantly produced and isolated or cell-surface expressed reagent quantities of the relevant target protein (73. Sometimes the differences between tissue iso- . 92) presented immediate opportunities to control a key step in the HIV-1 viral replication process.

and creating molecules with . major strides have been made in carotenoid and macrocyclic polyketides production (101.5) (107).and novel chemical diversity "tagged" for detection and amplification (89).88). Natural product producing organisms have been engineered to produce enhanced levels of the desired compound compared with the wild-type (103). 4. The selective amplification afforded by PCR can also be used to identify subspecies present in tissue in especially short supply. the PCR technique can uncover and amplify sequences present in low copy number in the mRNA and offers a greater likelihood of obtaining useful. When the possibility of achieving subtype specificity. providing a distinct advantage in favor of the recombinant protein. PCR has also been applied to the generation of recombinant diversity libraries of DNA (86). The structural richness produced by this combinatorial biosynthesis approach is of the sort that will be most tasking to even the best of organic chemist. determines a particular isoenzyme as a target for selective drug action. it is of obvious importance to be able to test for the desired specificity. cornbinatorial genetic approach has been used to dramatically increase the repertoire of pharmaceutically important metabolites produced by natural producing-organisms (106). the macrocylic-aglycon portion of antibiotic erythromycin. This method is unlikely to detect cDNAs corresponding to genes expressed at low levels in the tissue from which the library was constructed. has developed into a powerful technology for drug discovery.2 Combinatorial Biosynthesis and Microbe Re-Engineering Many clinically important pharmaceuticals and initial drug candidates are derived from natural sources such as microbes and plants (96). however. Alterations of the erythromycin polyketide synthase genes have been recently demonstrated to generate a mini-library of more than 50 potential pharmaceutically useful macrolides (Fig. designed to enhance the production of known compounds or generate novel products in microbes. Particularly. plants and animals have been recently reported (for recent reviews see 97-101). 102). This consequently contributes to the dearth of derivatives of these compounds for evaluation as potential drug candidates. diagnostics. the sequence homologies and even functional characteristics can vary greatly. 4. The engineered E. first reported in 1990. the structural complexity of these drugs precludes chemical synthesis as a practical approach to commercially produce them. the source-organism of 6-dEB (104). Potential applications of such combinatorial biosynthesis methods include lead generation and optimization of existing pharmaceuticals (108). an enzymatic method for the in uitro amplification of specific DNA fragments. In contrast. Polymerase chain reaction (PCR). structurefunction studies. 4. offering yet another advantage over classic methods. coli (104). In most cases. The tools of recombinant DNA technology are now being elegantly applied in pharmaceu- tical industries to overcome these and other problems. Also.3 SELEX or In Vitro Evolution In uitro selection (or SELEX). Several examples of biosynthetic pathway engineering.92 Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery lates and recombinant reagent are small. RNA (87. Driven by the realization of the slow growth rate and difficulties in genetic manipulation of natural source-organisms. because of either tissue distribution or differential gene expression. making whole families of target proteins available for comparative studies (95). has revolutionized the search for receptor and enzyme subspecies. full-length clones. coli strain was reported to produce 6-dEB in yields comparable with the high-producing mutant of Saccharopolyspora erythraea. the biosynthetic pathways of 6-deoxyerthronolide B(6-dEB). Classic cloning requires knowledge of at least a partial sequence for low stringency screening. researchers are making progress in introduction of non-native biosynthetic pathways into genetically amenable organisms (101). Genetic manipulation of gene clusters within the biosynthetic pathways of natural products has been used to rationally design new and novel products (105). For example. Similarly. more frequently. slow generation time and low quantities of the drugs from their natural producers are usually major obstacles to contend with. has been successfully engineered into E.

Module 3 I Module 5 . A scheme of SELEX is depicted in Fig. SELEX involves the selection of RNA or DNA molecules from random sequence pools that can bind to small or large molecule or perform achosen function.R = Propyl (5)." " .+ ' % ... . or 4 (yellow).. .... Colors indicate the location of the engineered carbods) resulting from catalytic domain substitutions in module 2 (red)... . .5. module 6 (blue).* "'OH 'k..R = Prowl (6). DEBS combinatorial library. R = Ethyl (52).or modules 1. module 5 (green). . 88. R = Ethyl (49). OH "OH "'OH n (1)." " .R = Prowl (10) (11) 0 (12) (13) (14) (15) (16) " OH . 109-113). R = Propyl ' (7).Module 6 Figure 4. PCR and subcloning techniques play a major role in SELEX. R = Propyl (3). .4 Genetically Engineered Drug Discovery Tools 93 OH OH OH .. The success of aptamers as potential drugs relies on finding solutions to many of the similar issues faced .. . O H B.R = Ethyl (55). and they often bind with dissociation constants in the pico- molar range.R = Elhyl (53). . R " = Ethyl (54).." " I . .3.. 4.. 3. R = Emyl (56).R = Propyl (8) OH OH "'OH (S). 4..1 Aptamers as Drugs.R = Ethyl (501. The small or large molecule binders are called aptamers. .See color insert. ..6." %. . R = Propyl (2). novel catalytic properties (87.. 0 0 0 "'OH 0 '"OH Module 2 .R = Propyl (4).R = Ethyl (St).

4 Phage Display Figure 4. reflecting the high level specificity that can be attained by these molecules (115). 4. It also provides an opportunity to physically link genotype with the phenotype that has allowed linking DNA encoding novel functions to be selected directly from complex libraries (117. 118). an RNA aptarner was generated against theophylline with a K.Application Recombinant DNA Tr~hnology in Medicinal Chemistry and Drug Discovery Random Sequences (RNA or DNA) t In vitro selection (SELEX) (ligand binding or catalytic activity) - Transcribe into RNA I A Collect bound molecules I Discard unbound molecules t cycle 4-1 0 times I Clone sequence characterize t ficity. of 400 nM. The frequency can be . The probability of finding a ligand in a random peptide library (RPL) is proportional to the affinity of the ligand for the selector molecule and its frequency of occurrence in the library.and it showed >10. an aptamer designed to bind the vascular endothelial growth factor protein has a dissociation constant of about 100 pM (114).000 times weaker binding to caffeine.4.6. In vitro evolution or SELEX has proven to be a powerful technology that may have a variety if applications in the development of therapeutics and diagnostics. Aptamers have been shown to distinguish between protein kinase C (PKC)isozymes that are 96% identical. by conventional pharmaceutical agents.1 Preparation of Phage Display Libraries. and must reach the target molecule. Aptamers generated by SELEX have been shown to have tight binding and a high degree of specificity. Chemically modified nucleotides have addressed the concerns about biological stability. Aptamers With the identification of new therapeutic targets it has become imperative to identify newer ligands that can bind to them. but issues related to delivery and mass production still remain as concerns. 4. etc. The aptamers must have high affinity and specificity to their targets. Coupling recombinant DNA technology with phage biology. These analogs are conveniently enough substrates of T7 RNA polymerase and AMV reverse transcriptase. In fact some of them surpass the discriminating abilities of antibodies. Aptamers being significantly smaller and simpler molecules than antibodies make them attractive for diagnostic use. The aptamer technology has come a long way as far as realization of therapeutics potential is concerned because several of these molecules are in various stages of clinical trial. For example. For example. rival antibodies in terms of affinity and speci- Phage M13 and other members of the filamentous phage family have been used as expression vectors in which foreign gene products are fused to the phage coat proteins and are displayed on the surface of the phage particle.3. 4. must be biologically stable. the phage display technique has revolutionized the identification of novel peptides 'that can be used for therapeutic purposes as well as for structural studies such as epitope mapping.2 Aptamers as Diagnostics. By incorporating chemically modified nucleotide analogs such as 2'amhe or 2'-fluoro. identification of critical amino acids responsible for protein-protein interactions. which was an order of magnitude higher discrimination than offered by antibodies against theophylline (116). Schematic o f SELEX. and one way to identify such molecules is by creating a random library and selecting the members with the desired properties. two of the essential enzymes used during the aptamer selection process. The issue of biological stability of aptamers has also been addressed. modified pyrimidinescan impart stability to nucleases when incorporated into aptamers and increase affinity for the target (114).

To identify the specific ligand for the receptor.4 Genetically Engineered Drug Discovery Tools enhanced by constructing libraries with large sequence diversity and flexibility of the insert. The selection led to the identification of two peptides. Interestingly. Small molecules are traditionally good binders of active site clefts or allosteric regulatory sites that are often buried within the enzvme struc" ture. The advantage of a peptide-based enzyme inhibitor is that it can theoretically sample the entire exposed surface of the enzyme with more contact points than a small molecule inhibitor. Additionally they demonstrated that the isolated peptides can also be useful for identifying small molecules inhibitors of the target enzyme in high-throughput screens. The phage particles that display the desired molecules are mainly isolated by binding to immobilized target molecules or by affinity chromatography. Another example is isolation of ligands for the thrombin receptor. Recently. Sometimes this may also pose problem during selection because of the chemical complexity of available targets leading to non-specific binding. a key regulator of the coagulation cascade with a high degree of specificity and potency (120). a known agonist for the thrombin receptor was used to elute the bound phages. Selection from a phage display library of peptides that will bind to a known target has its advantage in that specific ligands can be identified in the presence of a complex milieu of biomolecules. it has been established that the chances of finding a good ligand increases by screening more libraries.4. One of the first peptides to be isolated from a selection based on whole cells is an antagonist for the unknown plasminogen activator receptor. Therapeutic applications include identification of specific cell surface proteins in diseased cells as diagnostic agents and gene delivery agents.4. A wide variety of selectors have been used that ranges from organs in whole animals to cell surface proteins and studies with enzyme.2 Phage Display Selections against Purified Proteins. 4. antibody. Peptides that bind to purified cell surface receptor proteins or other therapeutically relevant proteins can be isolated from phage display libraries. This selection was carried out on transfected COS-7 and SF-9 insect cells that overexpressed the receptor. Furthermore. isolated a peptide that non-competitively inhibits the activity of serine protease factor VIIa (FVIIa). or receptor proteins. One of the classes of therapeutically relevant targets is enzymes. Extensive structure-function characterization showed that can home in on specific cell types have enormous potential both in basic research and therapeutic applications (121). Dennis et al. Peptides that will selectively "exosite" that is distinct from the active site. one of which acted as an agonist with the ability to activate the thrombin receptor. But the libraries that produce higher affinity ligands are created by deliberately introducing predefined structural features or structural constraints like disulfide bridges. Selective inhibition of individual proteases of the coagulation cascade may have enormous therapeutic value but is extremely difficult. Various protein and peptides including complex peptide libraries have been displayed on phage. obtained peptide inhibitors against six of the seven different enzyme classes that they tested (119). Apparently the inhibition was via an allosteric mechanism. They can therefore act as both drug molecules and as reagents in drug discovery. which can be long and unconstrained with the goal of making multiple contacts with the selector molecule. Hyde-Der et al. The latter approach can lead to identification of hitherto unknown cell surface receptors or target molecules. The other peptide could immunoprecipitate the thrombin receptor .3 Cell-Specific Targeting. the selection was executed on human platelet cells that naturally express the thrombin receptor. 4. Phage display libraries can also effectively identify enzyme inhibitors. The selection of cell-targeting peptides may involve targeting a known cell surface molecule or screening whole cells without a priori knowledge of the chemical nature of the cell surface.

4. Pasqualini and Ruoslahti demonstrated for the first time that an in vivo phage display can specifically target organs (122). functional enzymes and receptors) with potential application to drug discovery fall into a number of general categories: enzymes (with catalytic function). to the identified peptides. Recent advances in NMR techniques. structural biology also has used physical techniqueenuclear magnetic resonance (NMR) spectroscopy and X-ray crystallography-to its advantage in the study of proteins as drug targets. Rather than exhaustively catalog further examples. a common chemotherapeutic agent. The brain-targeting phage accumulated in the brain tissue 4-9 times better compared with the kidney-targeted phage.5 Reagents for Structural Biology Studies ing study. The isolated phages were injected back into the mice and the selectivity of the phages was measured. One of the motifs. the next sections will highlight instances in which combinations of directed specific assays and structural biology studies have aided in non-protein drug discovery. from enzymes and hormones to receptors and antibodies. soluble proteins or cell-surface expressed.possibly relevant to the design of synthetic ligands or inhibitors. homed in to several tumor types (including carcinoma. Three main peptide motifs were identified that targeted the phages into the tumors. sarcoma. CDCRGDCFC (embedded RGD motif). Recombinantly produced reagents (accessible as either purified.4 in Vivo Phage Display. and melanoma) in a highly selective manner. to determine the complete three-dimensional structure of proteins. in particular the methodology surrounding mechanismbased enzyme inhibition (131a and 131b) and the market success of various enzyme inhibitors such as captopril. In combination with molecular genetics. Inclusion of such structural biology results into the more traditional synthesis-driven discovery paradigm has become a recognized and important component of drug design (130). models for new drugs. offer dramatic improvements in spectral resolution and interpretation (127. 4. This has tremendous potential for gene therapy.128). thereby acting as a true antagonist. or in concert.6 Enzymes as Drug Targets A large number of enzymes have been cloned and expressed in useful quantities for biochemical characterization. In a pioneer- 4. The above examples demonstrate vividly the power of recombinant technology for developing techniques and agents that directly advances medicinal chemistry. where a major hurdle is delivering the gene (124. The variety of studies undertaken using these structural biology techniques spans the range of proteins of interest. brain and kidney. The isolated tissues were homogenized and the bound phages were reamplified in E. These two techniques can be used independently. and the two targeted organs. Phage libraries were injected into the circulation of nude mice bearing breast carcinoma xenografts. and binding proteins (with cellular adhesion properties). The advent of rational drug design paradigms. and discovery tools (126). receptors (with signal transduction function). coli. In another exciting study of in vivo selection of phage display libraries was used to isolate peptides that home exclusively to tumor blood vessels (123). were isolated. Identification of differences in the results from comparative studies on the same protein can reveal important structural or dynamic information (129). the efficacy of doxorubicin was increased and the toxicity was markedly decreased. and the targeting was specifically inhibited by the cognate peptide. Two pools of peptide-phage libraries were injected into mice through the tail vein. especially multidimensional heteronuclear studies. Several dominant peptide motifs were identified. By conjugating doxorubicin. Phage display has also been effectively used to identify peptides that will achieve gene delivery by targeting cell surface receptors to augment receptor-mediated gene delivery. 125). have made enzymes of . A synthetic peptide inhibited the uptake of the corresponding phage in the brain showing the specificity of the isolated peptide. 4. The mice were killed.- 96 Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery form membrane extract and promoted aggregation of platelets.

the natural substrate (139). 135a. The ability of rDNA technology to expedite access to quantities of a specific enzyme in a situation in which some indication of specificity would eventually be required of the final inhibitor is no where more evident than in the case of the retroviral aspartic HIV-1 protease (HIV-1 PR) (143d). are used extensively in the treatment of neoplastic and infectious disorders. especially to predict or explain specificity.but some of the conclusions made based on the enzyme-inhibitor binding interaction have been challenged by a crystal structure of human recombinant DHFR complexed with folate. 1331.. 4. but gave a 10-fold increase in K. reduced form (NADPH). a 24-fold increase in Km for dihydrofolate. Inhibitors of DHFR. Many enzymes linked to pathologies or known to regulate important biochemical pathways have been cloned for subspecies differentiation and/or access to human isotypes. phosphodiesterase (134. From among the multitude of potential points of intervention into viral replication of the HIV-1 genome. Comparisons of the conformations of the conserved human and mouse DHFR side-chains revealed differences in packing. but not the host proteases. The F34S mutant. mutations within the sequence GRD/N (positions 86 to 88 in HIV-1 PR)-a highly conserved domain in the retroviral proteases but not present in cellular aspartic proteases-were found to be completely devoid of proteolytic activity. 137a-137c) families. The search for important tertiary structural differences between HIV-1 PR and known eukaryotic proteases began by determination of the X-ray crystal structure (Fig. an enzyme essential for growth and replication at the cellular level. The rational basis of enzyme-inhibitor interactions. These results helped to pinpoint major differences among DHFRs of various species and thus suggest ways to design new and more species-specific inhibitors that would preferentially target pathogen versus host DHFR.7) of recombinantly expressed material at 3 A resolution (144). The mutant F31L (human F to E. this enzyme was identified as a viable target for antiAIDS drugs because mutation of the active site aspartic acid (D25) effectively prevents processing of retroviral polyprotein. 136. Site-directed mutagenesis studies confirmed the importance of this observation. specifically the availability of the target enzyme in quantity for screening. Such compounds would be expected to be more potent chemotherapeutics exhibiting less toxicity in humans. cyclooxygenase (COX) isozymes. however. showed greater differences: a 3-fold reduction in Km for nicotinamide adenine dinucleotide phosphate. coli thymidilate synthetase such as (5) attests to the viability and potential cost-effectiveness of this rational design approach (142. 138). In addition to the residues DTG at positions 25 to 27. Similar results were found for the F31S mutant. producing immature.7 A) material helped locate side-chains and resolved . The design and refinement of inhibitors of E. Some of the observed species selectivities for these inhibitors have been explained in terms of distinctive structural differences at the binding sites of the chicken and E.143a-143c). and an 80. In contrast to the DHFR investigations for which the goal is refinement.4 Genetically Engineered Drug Discovery Tools all types more reasonable laboratory tools and therefore accessible targets for medicinal chemistry efforts. non-infective virions.000-fold increase in Kd for MTX. is among the most intensely active areas of structural biology. 135b). Subsequent crystallographic studies on both synthetic (at 2. for which there was also a 10-fold increase in Km for dihydrofolate and a 100-fold increase in Kdfor MTX. suggesting that phenylalanines 31 and 34 make different contributions to ligand binding and catalysis in human DHFR (141). a 3-fold reduction in k. One of the most studied therapeutic targets is dihydrofolate reductase (DHFR). and the phospholipase A. coli L mutation) gave equivalent Ki values for inhibition by TMP. most notably the antifolates methotrexate (MTX) and trimethoprim (TMP). (135c. for dihydrofolate (140). coli enzymes (137d. problems in de novo design of inhibitors require more fundamental help. Also. potentially pinpointing a site critical for design of specific inhibitors capable of recognizing the viral. important advances have been made in the molecular biology and target validation of various classes of enzymes including protein kinase C (132. most noticeably the orientation of F31.8 A) and recombinantly expressed (at 2.

The structure of native HIV-1 protease drawn as a ribbon connecting the positions of a-carbons.(Y/F)P as a consensus cleavage site for HIV-1 PR. Far more useful for inhibitor design purposes.154). One such inhibitor.7. However. with the pseudo-twofold axis perpendicular to the page.NHl-nL]Q)R-NH. The lower structure is a top view.Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery some ambiguities in the dimer interface region (145. represents a view along the substrate binding cleft. fermentation broths. . only a limited number have been shown to inhibit effectively viral proteolytic processing and syncytia formation in chronically infected T-cell cultures (153. From this information. From among the large numbers of peptides identified as HIV-1 PR inhibitors. The short interval from identification of the enzyme as a target from among the possibilities presented by the HIV-1 genome to accessing material for assay and structural purposes has obviously hastened the determination of the viability of HIV-1 PR inhibitors as AIDS therapeutics and also has provided an excellent example of structurally driven rational drug design. Peptide sequences derived from specific retroviral polyprotein substrates and inhibition by pepstatin and other renin inhibitors identified (S/T)P. SGN(F+[CH2N]P)IVQ. which both inhibits HIV-1 PR and exhibits effective and noncytotoxic antiviral activity in chronically infected cells at nanomolar concentrations (155). whereas Ac-TI(nL+[CH. and designed inhibitors using HIV-1 PR cleavage of synthetic pseudosubstrates. complexes of four structurally distinct inhibitors bound to HIV-1 PR were solved (148-151).8) was developed (143). 146). from which a generalized closest contact map (Fig. The upper structure.P. as with other peptidomimetic structures such as inhibitors of another aspartyl protease. additional studies with both recombinant and syn- Figure 4. in which the pseudo-twofold axis relating one monomer to the other is vertical and the plane of the page. 4. their transformation into potential drugs will require additional synthetic work. renin (156). thetic material have yielded automated robotics assays for screening of chemical libraries. The most advanced peptidomimetic compound is structure (6). a model of the substrate binding site was proposed (147). was used in co-crystallization studies. With the functional role and tertiary structure of the protease determined.has been used as an affinity reagent for large-scale purification of recombinant HIV-1 PR (152).

Hydrogen bonds between a prototypical aspartic protease inhibitor (acetylpepstatin) and HIV-1protease. C115A. 0 (6) . Using recombinant CyP as the standard. The search for a common mechanism of action of the immunosuppressive drugs cyclosporin (CsA) (7)and FK-506 (8) highlights another possibility in which the drug was discovered by screening in cellular or in vivo models.. The residues labeled 25-50 are from monomer A.:wit 0 . Flap 0 b 5 0 \ ' NY ANl-?"49f G1~48 NH' ' ( \ / ) ". and C161A) were shown to retain full affinity for . 159). The residues are labeled at the C-P position (C-a for glycine). Using the active molecules.4 Genetically Engineered Drug Discovery Tools 0 Active site loop 0 Inhibitor 0 Val2 . and those labeled 1-6 are with acetyl pepstatin. but the exact mechanism or site of action is unknown (157).' 0. but the hypothesis was soon challenged by evidence that showed that the binding interactions are peptide-sequence specific (160) for both proteins. OH 0 . cyclophilin (CyP) and the FK binding protein (FKBP) were identified as the specific receptors for cyclosporin and FK-506. four cysteine-toalanine mutants (C52A. C62A.NH ' " o Figure 4. Sta4 . The biochemical mechanism of inhibition was first proposed to involve a specific covalent adduct between inhibitor and its rotamase (159). those labeled 225-250 are from monomer B. These binding proteins were discovered to be distinct and inhibitor-specific cis-trans peptidyl-prolyl isomerases that catalyze the slow cis-trans isomerization of proline peptide bonds in oligopeptides and accelerate the ratelimiting steps in the folding of some proteins (158.O Alas . respectively.8.

The immunosuppressant. NMR studies of [8. such as hypertension and nephrotoxicity. but the elucidation of these molecular-level mechanisms will allow the development of more highly tailored and potent immunosuppressive agents (166. X-ray crystallographic and computational modeling studies have been carried out on both CsAICyP and FK-506PKBP complexes to attempt to fully determine a structural basis for activity (163. displayed by the immunophilin protein. The complex seems also to exert two subsequent effects: halting DNA translation in T-lymphocyte nuclei and inhibition of IgE-induced mast cell degranulation. however.9). of CsA were designed by LoRusso et al.9-[13C]FK506 bound to recombinant FKBP. 4.Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery CsA and equivalent rotamase catalytic activity.and calmodulin-dependent serinelthreonine phosphatase. It was identified that the immunosuppressive effect of CsA is unconnected with its hypertensive effect. indicating that the cysteines play no essential role in binding or catalysis (161). 167a). The exact signal transduction mechanism that triggers the immunosuppressive response in T-cells. suggesting that the a-ketoamide of FK-506 serves as an effective surrogate for the twisted amide of a bound peptide substrate (162). (181). wherein the likely mechanism of inhibition also is noncovalent. 4.7 Receptors as Drug Targets Even more so than with enzymes. causing inhibition of the phosphatase (Fig. remained un- known until an elegant set of experiments identified calcineurin. In the case of FKBP. Toward this goal. molecular genetics has been primarily responsible for the identification of functional receptor sub- . this is vital information that may be of use in the design of CsA derivatives devoid of these side effects. experiments that sought to probe the relationship between the immunosuppressive effect and the common side effects. a calcium. then effectively functions as the critical element that binds the pentapartite complex together. Numerous further NMR. These in vitro observations must necessarily be confirmed by further in vivo work. and a complex of calcineurin with calmodulin as the binding targets of the immunophilin-drug complexes (165). 164).

a subtype-specific heterologous stable expression system in Chinese hamster ovary (CHO) cells suitable for screening potential subtype-specific ligands was developed. and M. es. From this assay. FKBP-FK506 also binds calcineurin. Success in the case of cimetidine. Endogenous ligands J I for cellular regulation? 2. a setive H. calcineurin (CnA. From this work. At this point. or lack thereof.The two subtypes M. pirenzepine. The classic tissue-binding armacological methods that made distincns on the basis of ligand selectivity have n supplemented. General protein foldases involved in protein synthesis and trafficking? 3. Any studies that profile e in vitro receptor subtype-specificity of mpounds can theoretically help identify pontial in vivo side effects of the compounds if e association of subtype to effect is known or uspected. suggesting that studies using pirenzepine on tissue homogenate have failed to distinguish adequately among the subtypes (173). for pirenzepine and were later confirmed by molecular cloning to be distinct gene products m1 and m2. CnB. .respectively (167b-169). and inhibits. One of the earliest examples of the role of olecular biology in discerning receptor subtype roles was the case of the muscarinic cholinergic receptors (MAChRs). Other biological I I I I C a M functions? - \ I I I I I I I I I I C a M Figure 4. previously thought to bind to M1 only. FK506-binding protein complxes with FK506 or rapamycin (Rapa). A subunit. The target of FKBP-raparnycin is unknown but is presumed to be different from cacineurin. Three additional muscarinic receptor genes (m3 to m5) were subsequently isolated (170-172). made it clear at the design of specific ligands for at least e receptors is a task amenable to medicichemistry.9. Schematic representation of immunosuppressant-immunophilin complex interactions.and an almost equivalent (to MI)binding affinity for M3 and M4. and in most cases suplanted. CUM. This complex binds to.camodulin) in a calcium-dependent manner. B subunit. CyPs bind CsA to form a complex in which both components undergo change in structure. had been defined pharmacologically by their affinity. by further subtyping made possible cross-hybridization cloning using the own receptor genes.-receptor antagonist. Similar breakthroughs have been realized across the rest of the family of signal transduction G-protein-coupled receptors. just the indication of more specific profile with fewer side effects is ough to help choose one compound over anher for preclinical development. was found to have only a 50-fold reduced affinity for M2.4 Genetically Engineered Drug Discovery Tools I Physiological function I I I I lmmunosuppression -. 1.

tachykinin.. mutagenesis studies on the &-adrenergic ceptor have localized the intracellular I mains involved in (1) the coupling of the recl tor to G proteins (178). (2) homologc desensitization by p-adrenergic receptor nase @-ARK) (179). A chimeric muscarinic chol .-adrenergic receptor. For example. because in addition to the muscarinics.10). itself cloned and a pos ble target for down-regulation inhibit! (180). the primary structures of the adrenergic (a.102 Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discov~ CHO CHO Human a-adrenergic receptor COOH Figure 4. serotonergic.10. The receptor protein is illustrated as possessing seven hydrophobic regions each capable of spanning the plasma membrane. thus creating intracellular and extracellular loops as well as an extracellular amino terminus and a cytoplasmic carboxyl terminal region. and PJ. p. 4. The exceedingly high homology among the large family of G protein-coupled receptors also has allowed the development of three-dimensional models of the proteins to aid in drug refinement (177). (3) heterologous desensitization CAMP-dependent protein kinase (181). All of these display the now-familiar homology pattern of seven membrane-spanning domains packed into antiparallel helical bundles (Fig. Topographical representation of primary sequence of human /3.. and rhodopsin receptors have been determined (174176). a.. a typical G-protein-coupled receptor.and an extracellular domain with conserved c teine residues implicated in agonist liga binding (182).

12). whereas the clearance receptor (ANP-C) completely lacks the necessary intracellular domains for signal transduction through the guanylate cyclase pathway (186)..Genetically Engineered Drug Discovery Tools (a) R + L R*L 7 R Desensitized R*L aGDPBy CZGDP + Pi Effectors Phosphodiesterase Phospholipase C Adenylyl cyclase Phospholipase A2 Ion channel /dm Figure 4. The receptor is desensitized by specific phosphorylation (-P).11. Receptor G-protein-mediated signal transduction. it may be possible to discriminate among the receptor . fixing them in an activated form.. 4. The diverse signal transduction functional roles of the many G-proteins to which these receptors are coupled (Fig.11) also makes them viable drug targets (184). Pertussis toxin (PTX) blocks the catalysis of GTP exchange by receptor. of GTP for GDP bound to the a-subunit of a G-protein. 4. has been significantly clarified by the cloning of three receptor subtypes.Because the NPs have differential. Data from experiments performed with C-ANP. The py-heterodimer may remain associated with the membrane through a 20-carbon isoprenyl modification of the y-subunit. affinities for their corresponding receptors (189) and because both agonism (190) and antagonism (191) of the GC activity have been demonstrated in vitro using ANP analogs.which can catalyze the exchange specific ligand (L). The a-atrial natriuretic peptide (a-ANP) reeeptor (NPA-R)and the brain natriuretic peptide (BNP) receptor (NPB-R) contain both protein kinase and guanylate cyclase (GC) do- mains. the regulatory systern that ads to balance the renin-angiotensinaldosterone system (1851.Cholera toxin (CTX) blocks the GTPase activity of some a-subunits. indicates that the clearance receptor (NPC-R)may be coupled to the adenylate cyclase/cAMP signal transduction system through an inhibitory guanine nucleotide regulatory protein (188). (a) Receptor (R) associates with a stabilizingan activated form of the receptor (R*). rgic:&adrenergic receptor engineered to acvate adenylyl cyclase (a second messenger system not coupled to MAChR agonism) also has helped identify which intracellular loops may be involved in direct G-protein interactions (183). Activated a-subunits (aGTP) and py-heterodimers can interact with different effectors (E). This system defines the first example of a cell surface receptor that enzymatically synthesizes a diffusible second messenger system in response to hormonal stimulation (187) (Fig. as determined by both sequence homologies and catalytic activities. The complicated biochemical pharmacolow of natriuretic peptides.(b) The G protein cycle. but not absolute. which revealed the functional characteristics of a new paradigm for second messenger signal transduction through guanylate cyclase.

. which aided in the identification of both the biochemically isolated guanylin (193) and the cloned proguanylin (194) versions of the endogenous natural ligand. The importance of access to human cloned receptors continues to be underscored as receptor binding plays an increasingly critical role in modern drug discovery (214). This system is presumed to play a role in water retention through cGMP modulation of the CFTR chloride ion channel (195). GR. These striking differences could be reversed by change of a single transmembrane domain residue (T355N). AR. multiple members of the interleukin cytokine receptor family (2111. the natural receptors were found to bind 5-HT identically. Proposed allosteric modulation of guanylate cyclase by a-ANP is schematically illustrated by a change in shape of the intracellular domain and a thicker arrow to denote an increase in guanylate cyclase-specific activity with greater production of the second messenger cGMP. In the comparison of rodent versus human analogs of the 5-hydroxytryptarnine receptor subtype 5-HT. These include epidermal growth factor (EGFR). insulin (INSR). multiple members of the steroid (ER.Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery pied ANP-A receptor is shown on the left with a basal rate of cGMP synthesis (indicated by a thin arrow). platelet-derived growth factor (PDGFR) receptors and related tyrosine kinases (1971. Homology between the NP receptors and another guanylate cyclase firmly identified the latter as the elusive heat-stable enterotoxin receptor St(a)-R (192).4. PR. / GTP cGMP + PPi GTP cGMP + PPi GCs to obtain more subtle structure-activity information for the design of selective NP analogs.5-triphosphate(IP3)-bindingprotein P400 (204). To make the case for using cloned human receptors for drug discovery even stronger. inositol 1. insulin-like growth factor-1 (IGF-lR). dramatic evidence that minor amino acid sequence variations inherent in species variability can produce profoundly different pharmacological effects was recently provided in two instances. A similar . tumor necrosis factor receptors 1 and 2 (198). and retinoid (RARa.. kianate-subtype glutamate receptor (205). follicle-stimulating hormone receptor (206). human y-interferon receptor (203). but they differed profoundly in their affinities for many serotonergic drugs. which effectively rendered the two receptors pharmacologically identical (215). The number of receptors of biological significance cloned and expressed for further study continues to grow at an exponential rate (196). The effect of ligand binding to the amino-terminal extracellular domain is shown on the right. p. and y. MR). subtypes of the GABAAbenzodiazepine receptor complex (199-2021. and subtypes of the glutamate (212) and adenosine (213) receptor families. thyroid hormone (TRa and p). and RXRa) receptor superfamily of nuclear transcriptional factors (207-210).

13. A soluble form of human ICAM-1 effectively inhibits rhinovirus infection at nanomolar concentrations (219). General polypeptide structure of integrins..These and many similar findings emphasize the critical importance of the availability of human proteins as potential gets for drug action and as sources of strucpotent and selective therapeutic agents. a highly related and widely expressed group of ap-heterodimeric membrane proteins (217). and replacement by glycine (C575G) generated a mutant receptor that could bind RU486. bWause of the presence of a glycine at position 575.13). subunit Figure 4. it is processed into two polypeptides that remain disulfide bonded to one another. 4. avidity for RGDcontaining peptides-will allow their individual roles in specific disease pathophysiolog~es to be ascertained and provide the means to develop integrin-specific antagonists for a multitude of uses. The a-subunit of the integrins is translated from a single mRNA.8 Cellular Adhesion Proteins understanding of the molecular processes govern cell localization in various patho- logical conditions has been significantly expanded because of the cloning and expression of some of the major cellular adhesion proteins. The avaiiability of the individual members of this heterodimer superfamly-charaderized in structure (Fig.. Likewise.b. also known as GPII. 4-13]. especially in the integrin family. the plate- domains 9 . As the pivotal interaction in the adhesion of leukocytes to activated endothelium and tissue components exposed during injury (Fig. . The interaction in the antigen-receptor crosslinking adhesion of T-cells mediated by the intercellular adhesion molecule (ICAM-1) and the lymphocyte function-associated molecule (LFA-1) was clarified by the cloning and expression of ICAM-1. the integrin a. the major cell surface receptor for rhinovirus (218). 4.4 Genetically Engineered Drug Discovery Tools 105 study comparing human to chicken and hamster progesterone receptors showed that the steroidal abortifacient RU486 (9) shows an- in the m n i s t activity in humans but other species. the P-subunit is tightly folded by numerous intrachain disulfide bonds. Both the chicken and hamster receptors have a cysteine at this position. and in some cases.. Of importance to antithrombotic drug discovery efforts.III. Other members of the integrin family have been also successfully cloned (220). by gross function. The a-subunit contains a series of short-sequenceelements homologous to known calcium binding sites in other proteins. mutation of the human receptor (G575C) abrogated RU486 binding (216). The a-subunit and the P-subunit contain typical transmembrane domain that is thought to traverse the cell memebrane and bring COOH-termini of the subunits into the cytoplasmic side of the membrane. inhibition of ICAM-l/LFA-l bindingrepprime internention point resents a for new anti-inflammatory drugs. and in some cases.

Lymphocyte and neutrophil trafficking. for which tissue-specific up-regulation of receptor population may successfully compete with other cholesterol-lowering agents (236. bohydrate analogs and noncarbohydrates (228. Molecular modeling of E-selectin based on antibody mapping and homology ta mannose binding protein (230) also suggests drug design possibilities based on proposed analogous structural interactions of the selectin with structure (10) (231). ment of integrin-mediated inflammatory processes that follow (232). RGD peptides. regulation of inducible or tissue-specific gene expression has been an obvious but elusive target for pharmacological intervention (233-235). The carbohydrate itself. tion sites on mucin-like glycoproteins such as Spg50. is known to occur through specific cell surface receptors and ligands. The carbohydrates are likely displayed at multiple 0-glycosyla. sequencing. Different receptors and ligands are expressed a t different time points during inflammatory processes. was successfully expressed as the functional heterodimer. For example. 229). such as shock and adult respiratory distress syn. which match the inflammatory cell to the right target. and helping in the manage. allowing control of acute inflammatory processes. The tools to monitor such events are now available. Molecular cloning of the murine HR now called L-selectin revealed that the receptor contains a lectin (carbohydrate binding) domain that is responsible for the binding event (227). as in the case of the low density lipoprotein receptor.237). the first step in the development of an inflammatory response. and cloning techniques (225). selectin discovered bv " a combination of bio. a novel endothelial ligand for L. have been shown to inhibit thrombus formation and platelet aggregation in animal and human clinical trials (223. chemical isolation. is the common endpoint in platelet activation.106 Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discover) let fibrinogen receptor (221).at specific times might help break a spiraling acute inflammatory cycle. specifically the tetrasaccharide sialyl Lewis X (10). initiating the plateletplatelet cross-linking through fibrinogen that is responsible for thrombus formation. however. In par- . from seconds to hours. FUTURE PROSPECTS 5 Molecular genetics is only now beginning ta identify new targets for drug action. drome (ARDS). A number of compounds. showing that prior association of the endogenous subunits is necessary to produce the cell surface complexes (222). which target circulating lymphocytes to specialized targets provide one such opportunity for intervention (225. Surface expression of GPIIJII. 226). The selectin familv " consists of three cell surface receptors that share this affinity for carbohydrate ligands. These protein recognition signals-previously termed homing receptors (HR) and now uniformly called selectins-are membrane-bound proteins. 224). and organic mimetics. including disintegrin snake venoms. Inhibition of selectin-mediated cellular trafficking . offers a viable starting point for drug design as a structural lead for both car.

Chem. Med. Prog.. 24. Res. REFERENCES 1. Chem. pp. Res. C. and J... There is little doubt among medicinal chemists that effective collaboration between chemistry and biology is not only needed but is actually growing in importance in drug design (245. 5. Eds. computational analysis. Liebenau. Annu. A. Sneader. B. J. Price. Pergamon Press. A. The ability to test for inhibition easure the effects of cies-specific agents against relevant pharcological targets in animal models has also eered gene knockt animals. and will be for some time. unknown. Mutschler and E. 149-191 (1992). Ondetti and D. 18. Pergamon Press. V. Taylor. Trends in Medicinal Chemistry. G. . M. 1. pp. H. M.397-400 (1988). R. M. G. 9. Rev. J . and J. pp. J. Johnson. Taylor. in conjunction with structural biology. Eds. Eds. a protein overexssed in most cancer cells.. The first FDA apIsis pharmaceuticals in nt of CMV retinitis in enasense. Wexler. VCH Verlagsgesellschaft. in C. 1987. 83-103. Drug Res. Drugs. Taylor. Lowe and P. Sammes. J. 1990. R. 1. 1. Burger. the d area of drug development might qualify e such meeting place for medicinal chemand molecular biology where the trend is plication of genetic engineertechniques to biochemical and pharmacocal problems will facilitate the discovery of novel therapeutics with potent and selective actions.429454 (1986). Sterbach. B. 1. 1990. another ann. P. B. Hansch. Timmermans. another genetic marker for atheroscleotic disease. vol. P. 6. is in late-stage unct cancer-treating g (240b). W. and medicinal chemistry (247). 5. 81-98. 1990. Oxford. Heel. A. L. Duncia. The eventual extent of the impact that molecular biology will have on the drug discovery process is. Eds.. L. Rep. Eds. A. and P. Comprehensive Medicinal Chemistry. chemistry and biology. Henwood and R.. K. pp. in E. Sammes. 22. vol. J. Sammes. Hobart. Hansch. A. vol. Winterfeldt.307-316 (1983). and J. M.229-266 (1978). C. R. Med. [are] r apart even as they discover e common ground" (244). 1-5. 10. Pergamon Press. John Wiley & Sons. 1982. G. P. New York. Ganellin in J. in C. An en more direct method to interfere with ecific recognition elements olecules designed to inhibit protein expresn at the level of translation of mRNA into e undesired protein (239). Cushman. Lednicer. S.246). T. Comprehensive Medicinal Chemistry. R. P.el. The power of molecular genetics to provide r drug discovery is The prospects for vering the molecular etiology of a disease tate or for gaining access to a disease-relet target enzyme or receptor are already beessful mapping of states at previously inaccessible or unmbinant DNA o a fully integrated component s is inevitable remarked that es. J. W. Freter. in screening d evaluation procedures. Carini. C. in C. 3. New York. 12. Seydel. Wong. D. J. B.355-361 (1981). Oxford. Chronicles ofDrug Discovery. Bindra and D. is a target for sedown-regulation (238). Oxford. A.. 4. 8. 11. 36. computer-assisted molecular modeling. 7. K. However. 2. lipoprotein(a). the reality of recombinant protein therapeutics offers the assurance that this same technology. Pharm.. Med. Inc.. 1-38.. Comprehensive Medicinal Chemistry. W.. vol. However. such as the CFTR-defective ouse for cystic fibrosis (241). will help make possible better therapies for those diseases already controllable and new therapies for diseases never before treatable. Hansch. Chiu.

E. 1.Annu. Today.632-637 (1983). J. Hynes and D. Rep.. 5.. Copsey and S. Med. J. C. C. 1988. USA. J. L.. D. S. Rep.H. L. W . 22. Annu. Y . London. 184. Vane.. T.. 111-120 (1988). Venuti in T . 267. Livingston.Y . Grossbard. et al.. Forsch. 56. 3-6. M. C. and D. Sci. L. Bennett. I. 1989. 28. L. Edu. Nat. 236.. J. Wang. Stockton Press. N. 2.129-138 (1992). Nisbet and J. J. Chem. S. Oxford. J. Ed. Trends Pharmacol. and S. P. C. and G. Harvey. 24. 42. L. 47. 282. A. Krause and P. R. et al. 53. J. Anderson. 29.647-662 (1995). F. W . J. 46. 1189-1208 (1989).. Henry. K. J. Taylor. 79.. Science. 455-479. Nisbet. 217-241.. Eds. Med. M. Venuti. Natl. D. Haemostas. Exp. and W . 25. Jouanneau i n C. 783-792 (2001).. E. Chem.1921-1928 (1999).201211 (1990). Taylor. Byrn. Johnson. Genetically Engineered Human Therapeutic Drugs. 45. S. 1990. 525-531 (1989). Kaiser. M. E. and K. R. W . 20. Delnatte. 68-70 (1989). P. S. Borell. pp. D. 48. Pater. and L. Hansch and J. 17. 213-221 (1989). Pergamon Press. Cancer Invest. 165-185 (1994). Friedmann. Oxford.. Eds. Hotchkiss. 409454. Andersen. 1. Nature.259-269 (1991). Nelson. T.. 43. B. Drug Res. 7384-7388 (1992). Capon. E. Martin. C. 21. Rev. P. 33. 23. Annu. 89. 8. 313319 (1993). New Drugs from Natural Sources. A. Biochim.97-120 (1998). 137-146. and M. D. J. 37-42. Nut. Arthritis Rheum. 32. 27. New York. 337. J. D.. 37. Kohnert. Holtzman. J..667-670 (1990). and S. S. 1999. (2001). R.. Szkrybalo. Fosslien.361-363 (1987). M.. Inc.. Nature. Jensen. Med. 54. E. Rahrnan. B. New York. Hutchinson. 52. G. Eng. Quertermous.. DiPippo.1990. 50. 52.632-637 (1989). vol. J. J. J . Proc. Engl. Eng. pp. A. Matsueda. Vane. Kiefer. C. Prod. 23. Schneider. Graham. Med. 133-168. N. Drug Res. N. W . Prot. P. Langman. Delnatte. pp. 13. J. 26. Toxicol. Rev. Sci. Kutter. 53. 317319 (1991). 557-567 (1988).149-157 (1986). Med. Med. Pharmacol.215-216 (1994). M. A. Slamon. 19. J. R. and Y . 1198.883-891 (1996). Botting. F.91-121 (1990). J. 1. 243. N.... Paoni.. Chem. Rep.. Ed. Traunecker. Ann. Science. Biophys. 21. GouldRothberg.. R. Comprehensive Medicinal Chemistry. 1591-1602 (1998). Katz and C. Y . 30. Annu. Higgins. 12. Rininger. Pergamon Press. 67-81 (1998). 31. J. 833-841 (2001). 25. Knowles. and D. Portanova. Chem.M. Dillman. and V . J. 34b. Waterman. Tanswell. Stockton Press. L. 49. et al. Ogez. F.. Comprehensive Medicinal Chemistry. vol. 334. D. J . Med. 21438-21445 (1992). Wurm. Nature. News. B. Bennett.344. 38. Stone. M. 282. R. P.108 Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery 12. Austel. 38. Annu. Hutchinson. 24. Eng. IBC Technical Services. K. Angew. Toxicol. Chem. Watson. H. 367. Rep. A. Badillo. JAMA. J. J. D. U. JAMA. Cometta. Refino. Chem. Turk. Coornbes. vol. pp. 41. S. Otten. C. Annu. Nature. and B. 339. 34a. 929-934 (1989). Rep. Acta. 1988. 5. Eds. T .929-937 (1989). Prod. J. S. Y . 40. A. N. Drug Discov. L.. Hylands and L. Hla and K.913-922 (1988). Stern. 19. A. M. F. E. Annu. Gordon i n D. S.93-100 (1992). 13-22 (1990).. Res. 219.1929-1933 (1999). 12. 39.. Chem. Marcel Dekker. Hansch and J .Forsch. Pharm. 39. G. Molecular Genetic Medicine. 70. 631-620 (1989). J.. J. M. A. et al. Weaver. Res. 18. . D. Stutzmann-Engwall. Clin. 1991. J. F. Simon. S. 50. Bakhle. E. Chem. 44. S. P. C. New York. 51-56 (1989). Hentschel i n D. Med. R. R. Ross. J. 4. Karjalainen.. B. Pharmacol. D. Keyt.1252-1258 (1987). Bost and A. J. 27. Higgins and W . Simon. Pharmazie. 26. 30.. pp. J. Arzneim. 560-568 (2000). Lab. Zhang. vol.. Arzneim. Haber.. Adams in C. R.V . Academic Press. Swetly i n Y . P. 28. Biol. New York. 51. 39.. 0. 15. 32. New Engl. 55. Medicinal Research Series. 14. Med. Science. C. 35. Rep. Sci. 57. Williams. Eds. Chem. 36. Runge. and R. 41. Westley. 16.. Genetically Engineered Human Therapeutic Drugs. M . Hauck. Prot.. Thromb. Haigwood. Chem. Kubinyi. L. Sharnick. L. pp. Modern Drug Research: Path to Better and Safer Drugs. Hutchinson.. H. Acad. Med. C. 27.

S. 69. A. J. J . G. Solas. Toole. 244. D. Vos. 643463. Vos... C. H . Marks. Nishikawa. Soc. Science. Bentle. S. M. Tarpley. J. D. Krebs. Nature. 82-84 (1991). Cunningham. Mulkerrin. H . J. R. Med. Prot. M . Sci. H. B.29-44 (1992). Biol. Mulkerrin. D. 84.S. 453475 (1990).5381-5383 (1992). N. 564566 (1992). J.. Schultz. J. R. A. Wells. G. D. Simon. Ellman. K. 505-510 (1990). Wells. and T . Nature. AlDS Res. and A. Gelfand. Biochem. Proc. E. D. Hersch. Henner. Hruby. T . B. Griffith. 74. . S. Hsu. 14. IBC Press. 70a. Hoogenboom.. Acad. Kazmierski. D. A. 8. Sci. McPherson.197-200 (1992). Drewe and R. H. Mulkerrin. (2001). B. CZin. Annu. Hasler. A. T. J.821-825 (1991). Houghten. Oliver and A. J. V . K. 1992. R. R. pp. J. Immunol. Sninsky. 12. L. Science. Mendel. 151-363 (1986). Natl. 81-85 (2002). Stryer. Scott. and J. C. Science. S. 257. L. Cunningham.. IBC Technical Services Ltd.1081-1085 (1989). C. . 1798- 73. 72. A. D. J.306-312 (1992). Kollman. Trends Biochem.J. Blondelle. W . Sci. G. Struct. 346. and P. 49. K. and D. E. H. Bass. M. Ultsch. Meloen. M . S. Nature. 12. M. Methods Enzymol. D. 28. 1992. Wells. Tomari. A. C. S. Wells. A. H. Res.3998-4002 (1984). 256.C. P. L. Schultz. Knapp. J . Szostak. and J. C. Green. Science. Bourell. 254. Science. B. 244. Chem. Richrnan. Chem. Pitts. Anthony-Cahill. Med. M. Ultsch. Noren. Slice. 818-822 (1990). R. AIDS Res. 267. J . Science. Chem. Fuh. Kossiakoff. L. S. 20. Inoue. Yokogawa. Fodor. I. 61. 182-188 71. A. J. 243. 249. H . Science. H. J. Dooley. and J. 89. H. K. 61.301-336 (1991). S. J. Pavia.. 751-755 (2001). H. Erlich. Neubauer.. Burch and D. Natl. Cunningham.3111-3115 64. Strehl. B. Barteling. London. and G. 62. and P. Sci. Science. M . S. M . 334. J. H. D. D. Moos. Acad. USA. and P. 5. Sci. Botstein. 141-147 (1991). Noren. Res. M . 5. 315-324 (1993). C. P. MA. Read. Suzuki. T . Hum.. W . Ed. L. Dayringer. Pinilla. 93-100. C. Lowman. S. C. Science. 38. Nature. Lam. Latham. Geysen. H. D. Acad. T . Hormone Res. Southborough. S. Weighous. Schutt. Curr Opinion Biotech. G. 265. Brochier. M. Natl. Bock. Palzkill and D. Potash. Kyle. J. Chamow. 55. Tuerk and L. Greene.J. D..Wells. M . Walkinshaw. 67. W . Steinmetz i n Rapid Functional Screening for Drug Development. M . G. P. Mendel. H. and L. M. J . Shieh. M. Science. Nau. 188-194. and P. J. R.58. Ellman. Sharma. A. W .. Oyelere.J. Wells. Biol. 354. Cunningham and J. Fuh. 59. M . C. 4 . Pharm.84-88 (1991). 255. S.253-275 (1992).. T . Vermaas. A. Batcheler. and A. Hum. G. L. Abdel-Meguid. USA. 4168-4171 (1996). 255. 8.400570d. J. Volsky. Sherman. Acad. Petteway.9367-9371 (1992). M . Light. 16007-16010 (1992). Shimizu. D. . C. Proc. 365-366 (1990). A. L. 1461-1465 (1990). Schultz. Natl. 89. Coombes. 81. Smith. A. S. A. Annu. M . R. LernerProc. 251. C. G. 60. Oliver and A. Prod. H . B. 48. and D. Trans. Nature. J.. 7Oc.. Nature Biotech.241-245 (1992). Pirrung. E. 202.Y. Z. P. F. R. Salmon. W . A. Tett Lett. Nat. Rep. E. R. M . D. E. K. J. Fellows in J. Griffin. J. R. Oyelere. A.. Ng. Org. Violand. Ultsch. G.. Mendel. McFarland. Chang. 317- W . 252. Rev. 525-531 (1992). 65. 1643-1651 (1991). G. 345. Wolff and A. Retroviruses. Gold. E. and J. A. C. J . Appel.. Lu. J. Pathol. 17. and B.1064-1073 (1992). B. Ellington and J . D. Recent Prog. Prog.507-516 (1989). 275-278 (1989). 70b. Klausner. Jhurani. T . C. J. Science. J . T .. H. 355. M. M. Winter. L. Brenner and R. Cunningham. C. Function Genet. D. A. M. Chadd and S. C. D. Org. E. and M. Ellman. Schultz. USA. A. A. Q. C. Retroviruses. N. Gnffiths. Science. Cuemo. USA. and S. S. Chem. Powell. A. and R. and J.. Y. T . 767-773 (1991). and J . R. E. M. A. G.. Ferguson. G. 19. W . B. D. Proc. Evans. 247.1330-1336 (1989). W . S. Anthony-Cahill. L. et al. New Drugs from Natural Sources. 66. Chem. M. Bass. D. A. and J. 7Oe. Rev. 68. and W . Vos. Ueda.

Natl. 258. Philadelphia. A. A. and L. Jellinek. J.. Opin. 36. Hutchinson. W. M. Kramer. Thamchaipenet. Quan. M. Ellington.549-554 (1995). Jucker. S. M. Burg. R. C.1078-1082 (1992). M. Science.R. S. Biochemistry. Neilsen. R. 116.1390-1399 (1991). 76-83 (1998). 111. P. 364-367 (1996). and G. Chem. 35. Chem. Jensen-Pergakes. 110. 109. A. Challiss. Admiraal. P. L. 264. Mol. M. M. D. A. Tizard. J. Rando. J.. C.E. E. E.3205132054 (1994). Natural Product Medicine-A Scientific Guide to Foods. J. A. . M. Cane. Rohlin. 11. Sandmann. T . Marderosian. R.M. 7. 98. Chem.. Lett. Science. C. Rouslahti. Curr. K. 114. J. and D. Sci. M. Annu. Plant. 55. J. 279. PA. Phurm. Chem.. 505409 (1998). and P.M. Matsumura. A.. C. Clore.. Oh. Polisky. C. 133. 11.. Pasqualini. S. 282. P.11231136 (2001). Chem. and C. 40. Acad. Prog. Janjic. Gonzalez. Biotechnol. Larocca. Mason. Good. Pruzanski. 104. 11. Curr. 132. 137b. Santi. Beebe. 1988. 128. 2. Jayasena. 136. Science.285-292 (2002). and N. Med. Hayes. Exp. Biol. M. and E. Kloss. Parker. D.1790-1792 (2001). Erickson and S. E. 135a. 127. 119.16-21 (2000). E. Curr. Miller. D. Sidhu. Fu.1846-1851 (1999). 36. Pharmacol. Microbiol. Stickley. 683-695 (1995). Ravey. Robinson. D. Brown. Crocker and R. P. 101. H. and B. D. Sci. Baird. Curr. D. Co. Biol. E. Baird. 567-597 (2001). W . Gilliland. Albrecht. 791-795 (1999). S.. Biotech.. Bell. H. Drugs Today. L.. Conrad. Opin. Science. R. Gustafsson. Science. 6. 1174011745 (1999). 269. C. Biol. A. McGray. Chow. Microbiol. F. and M. Opin. 96. 45.377-380 (1998). R. Nature.. F. Rev. A. C . 124. Santi. J. Thera. 121. C . 125. Hopwood. Opin.Gronenborn.. Nature. Biotech.. 607-614 (1992). Ebert-Khosla. Acad. S.I.519-535 (1999). et al.. 32053218 (2001).3. Chem. H. Appl. Ellington. Baltz. C. H. T . Science. 126. I . M. I. M . D. R. Feistner. Kuntz. Khosla. D. L. C. G. R. Opin. 95. Biol. Biol. Biotech. Keranen.. 120. R. 263. and R. (1997). J. 123.Witte. S. Gill. 74. W. and A. W . A. Drugs Today. 97. A. S. S. G. D.19-27 (1991). 103. Proc.. 2937-2945 (1991). (Tokyo). Osborne. Chem. R.. 137a.. R. and A. Lazarus. USA.961-964 (1992). 118. Biotechnol. D.J. D. D. 11. and A. Opin. C. D. Science.. and R. G. B. H. 101. 137c. Hum. A. Chem. Biol. 27. Liao. K. 4. Science. Jenison. Xue. Shaanan. Curr. 117. Opin. 252. Green. and G. Jenne. Patents. 102. 5-9.610-616 (2000). C. Gene Ther. Curr. 21. D. 263283 (2001). Fesik. Science. 129..93-99 (2000). Fesik. 105. 134. J. Chartrain. Dennis. 108. G. Rouslahti. Biotech. Salmon. Betlach. Kealey. 2. A. Natl. 100-105 (2000). Proc. and J. Johansen. J. K. Seilhamer. 255-261 (2000). A. Suri. R. Y. and E. 40. 1425-1429 (1994). G. 1628-1650 (1999).63-68 (1998). McDaniel. Murakami and I. C.271-289 (1992). B. Davies. V . Townley.9513-9521 (2001). B. M. Ashley. Sci. W . Proc. Chem. L. USA. M. Opin...1. 39-60 (2000). M. Opin. Betlach. B. G. Acad. S. S. 107. Larocca. Ther. and C. D. Pfeifer. B. Khosla. and A. Veerapandian. Med. 106. Chem. Buckland. Liu. Cohen. J.330335 (2001). Barr. 100. 131. Pardi. and G. M. J. R.45-57 (2002). J. 380.2393-2399 (1998). R. C. L. A. 111-142 (1984). Nishizuka. G. Hoess. Trends Pharmacol. 17-25 (2000). J.S. McDaniel. Newton. Sci. Drugs and Cosmetics. Khosla. G. Curr Pharm Des. C. J. Famulok and A. J. K. Clin. Q.320-327 (1998). 264. 209-214 (2000). H. Nicholson. Kudo. 375. et al. Curr. Trends Microbiol. 131b. M. Lehr. 130. USA. V .. L.. B. C. P. 122. Arap. R. 4. K. 34. Roberge. Gill. Marshall. L. 257. 112. 99.. Pepinsky. Stief. Schmidt-Dannert. A. and B. Vadas.. 291. Biotech. S. N. C. Shahid. Hyde-DeRuyscher. J. Chem. P. Biochem. 113. 96. 135c. Gramajo. D.110 Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery 96.C. Biol. Biol. A. K. Curr. F. 9. Martin.R. Lipid Res. Sun. S.. Clore. G. A. D. Hession. Rep. Pierce. 233. and C. Rev. Pate1 and A. J.W . 131a. 12. Capper and L.5335-5338 (1989). Misawa. Johnson. 115. S. 135b. Ashley..5678-5775 (1989). 7. 257. Walsh. 2. M. 167-197 (2001). Chem. Pasqualini and E. Johnson. Natl. Cane. Curr. 853-859 (1986).

. Hayano. Acad. Matthews. E. Bertino. R. Pharmacol. N. W. Weber. et al. J .. C. Nature. Russo. Albers. E. L. 264. 140. Res. Michaelson. 156. Blakley. C. H. 34. and C.. Sci. Stein. S... 904-917 (1996). U. Biochem. Commun. L. Science. E. J. and A. Chem. W. Med. S. and T. 168. 164. B. Barlocco. J. and J. and D.454-456 (1990). T. G.. 45. and A. A. Miller..1925-1934 143a. Natl. Kuntz. R. and J . Science.G. 260. Lang. Schreiber.M. I. Darke. 358-361 (1990). Rich. A. S. Buckley. E. Jones. Winslow. Bonner. Delchamp.3923-3929 (1987). Ramachandran. Br.. et al.. Meek. M. D. Wittmann-Liebold. Bartlett.A. G. et al. 66. et al. A. 143d. Sokolovsky. and F. J. K. Greenlee. J. L.. Prot. 616-621 146. C. J. T. 403-410 (1988). Sigal. McQuade. L. R. K. 527-533 151. R. Appleman. D.473-475 (1989).269-281 (2000). Biophys. Buckley. Villafranca. Chen. D. J. 248. Marzoni.J. Friedmann. S. Ashkenazi. R. M. K. V. D. Peralta. J. Andre. A. Galat. 342. W. 247. Skalka. Chem.6. and D. 476-478 (1989). J. 143b. 174. A. Smith. Palmer. R. K. and F. 161.R. et al. 237. 66. S. N. M. H. 147. L. Chem. J. 823-826 (1991). Buckley.. 251. Oefner. Stout. M. 141. 169. T. T. Biol. I. 87. and W. S. J. F. R. and P. EMBO J. J. H. R... K. Lapatto. 39. Tomasseili. Science. Shoichet. S. Life Sci. 139. 928-931 (1989). Proc. Freisham. Traber. R. J. Ruegg. Nature. L. D. Peralta. Brann. R. H. Liu. Matthews. M. 248. Moss. 337. T. Heinrikson. Wuthrich. 9 (Suppl). Widmer. 245.. R. Mol. von Freberg. Drug Res. Webber. et al. 173. Klebe. A. Science.. Schneider. P. H. 2305-2314 144. 170. Wlodawer. R. 167b. 164. Res. Srirnatkandata..90-92 (1990). P. Schreiber. Biochemistry. I. 265. J. N. H. Schreiber. W. C.. B. L. J . S. Wlodawer. K. 343. 163. Kiefiaber. Sawyer.D. W. W. 249. 157. J. Acad. and C. Volz. Eur. Harrison and R. 38. Erickson. T. Science. Swain. J. K. 18. M. Sheridan. J.283-287 (1991). McKeon. D. J. J. L. and M. Ward. H. N. 150. and S. R. Med. Schreiber. Adv. Tondi. 162. Takahashi. Gallion and D. P. G. 6-11 (1988). 20786-20795 142. Moomaw. et al. 155. J. R. M. G. Capon.4645-4650 (1989). and N. 10. M. Young. D. M. Rev. M. Science. Rosen. and S. Sci.527-532 (1987). 5. C. Filman. L. Chem. Leis. C. M. D. Biol. 171. A. Walsh. Weissman. Miller. J . J. J . 165. 29. Science. Bonner. S. K. 1. Garsky. D. Webber. Gritsman. Schmid. Ramachandran. J.377-385 (1988). D. Heimbach. 1831-1851 (1989). Farmer. P. R. Winkler. I. Janson. B. Green. Fischer. Eng. Proc. Huff. Capon. J. G. 8805-8809 152. Biochem. 166. Ashkenazi. G.3813-3816 (1990). C. X. T. Liu. J. Santi. J. Navia. 863866 (1990). S. Biochemistry. 118. Mei. Liu. A. E. D. J. Ringe. L. 28. F. and J. and H. J. A. V. Chem. Hulme. Appelt. 87. Rev.. A. J.T. Science. Miller. et al. 254. A. C. S. Birdsall.137d. Tarpley. Wider. Med. 172. A. Suzuki.. Nature. A. I. Howland. 337. Science. Natl. Bacquet. A. Schweitzer. Appelt. D. 246. W. J. M. Cell.299-302 (1989).M. C. USA. J. W. K. W. 615-620 145. D.. 34. J. V. Toxicol. Lewis. P. and F. 1607-1617 (1999).391-397 (1992). Nakatsuka. R. Herrmann. T. Jaskolski. Costi. White. J. E. A. Skutella. 78. Chem. Cell. and W. M. W.. Biochemistry. Burridge. 633673 (1990). Bolin. Biol.. M. W. Chem. J.. I. 260. L. I.2304-2308 (1990). Lane. Nature. 159. A. Rinaldi. K. 138. A. 381-391 (1985). J. Young. B. I. C..885-892 (1996). M. M. 30. Science. USA. D. K. Pecorari. Prendergast. Med. Biol. B. A. E. 143c. K. 431-509 (1989). Smith. Morse. L. Stroud. Yamamura. Dixon. Roberts. C. F.T. Matthews. M. T. V. Roeske. Neuron. Passaquin. Sci. M. Braun. C. A. 143. . Winslow. et al. A. 158. Pharmacol. 167a. M. 337. Venkataraghavan. 160.. T. 14209-14219 (1990). 1149-1152 149. Kraut. J... 955- 153. 148. D'Arcy. Nature. B. Booth. et al. Trends Pharmacol. L. 173-276 (1990).. J. Med. and M. S. M. J. G.381-391 (1985). Standaert. 154. Weber. M. Kent. K. Varney. T. Smith. Kathardekar.953-954 (1991). N. Fitzgerald. and N. Wlodawer.807815 (1991). Annu. Brann. Welsh.-M.

G. K. S. V . 191. 204. Sci. 342. S.2030920313(1989).. Chen.3135-3143(1989). P. G. O'Dowd. Science. F. Rev. and T . Bigogne. 197. Pharmacol. Kishimoto. L. 1469-1480(1990). Duffin. Fok. P. Sci. P.. C. Dixon. S. Chang. L. 12. G. Moore. P. Chem. M . Rogers. and E. A.889-895 (1988). Trends Pharmacol. Sieghart. J. Conneely. Currie. and D. D. Garbers. Molec. Biochem. S. Smith. Evans. J . J. Santi.2854 (1989). A. 33. Lefkowitz. A. Rev. Chem. and R. W . Clin. T . 11. F. J. Biol. R. Rev. M S . Rep. Strader. Gofflo. L. 341. Science. Furuichi. W . 342. 13. J. Cheo. 12. A. Wagner. W . R. D. Hamra. Tobin. 264. I. 192. 246. Bennett. and P. P. Benovic. Lowe. B. 180. F. S. M. L. Caron. M. 86. G.343-348 (1989). B. Jung. Parker. Dong. T .802-808 (1991)... A. Seeburg. Science. K. 200. 1827-1830(1990).941-948 (1990). V . F. Nature. 265. O'Dowd.V . and K. J. A. and R. G. 212. R. Today.Abbott. Gilman. Jones. H . Garbers. J. 13... R. Tzukerman. G.C. D. Lowe. et al. 183. and K. G. Cell. and H. A. 208. Inouye.Sprengel. M. 213. 214. Picard. Chem. O'Shea-Greenfield. Nikolics. Res. B. J. and S. USA. 8.30113015 (1989).. M .166-120 (1991). J. Anand-Srivastava.32-38(1989).Bovy.V.C. J. . Green. Santiso-Mere. 264. Lefkowitz. 176. Biochemistry. M . V . J . R. FEBS Lett. 265. D.151-153(1992). N. D. Med. M. and K.. K. Pritchett. Clark. 61.235240(1989).525530(1990). M . 216. J. W . Pharmacol. M. 68-72(1989). B. K. Godowski and D. D. Nature. USA.. Chambon. G. 161-163(1992). J. M.1377-1384(1989). 264. 10. Jin. 190. Oksenberg. L. J. E. Dohlman. Schlessinger. 43-48(1989). R. Lefkowitz. W . Birnbaumer.407411(1989). A. W . L. Yuen. 38. 215. Bovy. G. L. and D. R. T . S. Trends Pharmacol. Acad. S..112 Application of Recombinant DNA Technology in Medicinal Chemistry and Drug Discovery 174. D. 196. Chem. 193. M . 245. Garcia. 189. R. 181. R. Rev. Stiles. 205.2657-2664(1987). Yoshikawa.. 248. Hnatowich. R. Stone. J. S. J. Med. Schultz. 0. 615649(1987). P. P. Proc. H . Gautam. M . J . J.. Goeddel. A.. Chem. Galen. H. Science. 13. 4. V .1389-1392 (1989). Quan. Cantin. L. Gleeson. Tartaglia and D. T . D. Krueger. and D.6429-6482 195. Strathmann. and M . M . Chang. and K. Goeddel. S. Annu. 147-184 (1991). Garbers. 207. 263.20415-20421(1989). 202. and P. G.291300(1992). 185. Grayson. C. A. 206. M . Clevenger. 36. C. and N.. Kambayashi.-F. S.. Mirzadegan. M.597-603 (1992). A. 89.and P. Nakajima. Biol. 169 (1992). Benhamou. Stefanos. Goeddel. Goeddel. 30. Sprengel. Nakanishi. Annu. 199. G. Mukhin. P. S. Chem. 175. K. and M . Sairam. R. Natl. Lewis. P. S. Cell. 198. W . Taga. P. D. Regan. McDonnell. L. Biochem. Biol. Trends Pharmacol. 360. J. J. Horuk. Chem. and C. and B. Gardner.9266-9270 (1989).. Biol. 10. Ullrich and J . Marsters. and S. E.. Guidotti..593-597(1992). E. C. Lerouge.Yuen. Power. Kumar. S. P.-S. Pharmacol. 210. B. J. Miyawaki.. B. M. J. Schultz. Vergezac. Med. Ashkenazi. G. R. C. 211. D. 27. Werner. and C . Gronemeyer. 194. Hollmann. Chao. S. Mol. Ross. Ueda.. 179. Wada. 1165-1172 (1993). Peska. J. J. H. 56. C. Zmmunol. Med. E. Biol. C.. 182. 240. 188. F. O'Malley. 8566-8572 (1990). Mikoshiba. 177. Science. B. R. S. Seeburg. Sci. D. Caron. 203. EMBO J. Akira. Williams. M . 13 (Suppl). Acad. Science. S. Benovic. 63. M. 178. T .. T . Toxicol. R. M. M . 258.675-705 (1990). Sci. desauvage. Friedman. Res. R.1065-1072 (1994). 186. 252. S. M . Nature. Dana. Nature.. Annu. Johnson. Caron. 267. and D. H . Pharmacol..A. Olsen and A. K. Caron. DeBlasi. Fraser. Kato. Hellmiss. and D. Y . Humblet and T . Seeberg.15985-15992 (1988). Hellmiss. Segaloff. Heinemann. Chem. 184. Science. O'Connell. M . Chem. 4. 206-209 (1992). 203212(1990). D. Rev. Wong. A. Maeda. J. Biol. Goeddel. S. 265. Y . 187. F. M . Havlik. and R.. Luddens. Peroutka. T . Trends Pharmacol. 255. E. Sci. J . K. T . 26. 258. J. Endocrinol. Simon. and A. F. A. 423-471 (1992). A. Res. EMBO J. 318323(1992). M. Michaels. Proc.6219-6224 (1990). M . Leader.115-142 (1990). desauvage. Natl. Biol. 947-951 (1992). 201. J ... Singh. Burnier. F M E B J. Biol. Braun. P. 209. R. (1992). G. F. Lefkowitz. Lohse. G. L. Sci.

Stratowa. R. 44. Phillips. Chem. 28.. Chem. 344. Milligan. Curr. Med.79-88 L.Knight. Scarborough. Blackburn and T. Zeelen. D.Marlin.425-434 ( 1 9 9 0 ) . 257. . J. 133. L. A. Consortium. S.Osborn. Cell. Annu. Griffiths. B. Fischer.207-211 241. and B. 0. H. Opin. Chem. Trends Phamacol. Maniatis. R. S. A. et al.43. F.Schreiber. C. Hurley. 1. L. Merluzzi.3013 1 0 ( 1 9 9 2 ) . I. Edu. W. 2 3 8 .313-378( 1 9 8 9 ) . Cell Biol. K..1083-1088( 1 9 9 2 ) . F. K. 65. 26.Utermann. J. H.. K. Rep. 244. 988..215-227 ( 1 9 9 2 ) . W. N. Ther. G. Hynes.. Annu. 6888-6891 ( 1 9 8 7 ) . D. Biochemistry. D.. Weis. G. 27.30.359-362 ( 1 9 9 1 ) . and J. Goodbourn.246. Ther.J. Karlsson. S. 22.. Annu. News. Musser. Napier. Cell. Springer.Hirschmann.904-910 ( 1 9 8 9 ) . 70-72 ( 1 9 9 0 ) . 62. Adv. Bodary.. 303-317 ( 1 9 8 9 ) . S.1304-1351 ( 2 0 0 1 ) . E. A. Jabs and P. K.. Jakubowski. Rep. Chem. Latour. Engl. 22-32 ( 1 9 9 2 ) . 1237-1245 ( 1 9 8 7 ) .964-969 ( 1 9 9 2 ) . Malouf. Springer. Annu. 404408( 1 9 9 9 ) . 242c. et al. 235. W. Med.L.7A-8A( 1 9 8 7 ) . A. Biochim. Ophthalmol. D. Cell. M. Science. Sci. D.187-219 ( 1 9 8 9 ) . Smith. Chem. 237.287-296 ( 1 9 9 1 ) . R. 360. 2 4 7 .18859-18862 ( 1 9 8 9 ) . Gadek. Cell.V. 2 4 5 . ( 2 0 0 1 ) . 0 . Rep. Science. 258. Am. 70.. Nature. J.552-556 ( 2 0 0 2 ) .M. Koller. 240c. T . F.R. Angew Chem. Mol. Lasky. A.Matteucci and N. 246. Ruoslahti and M. Lasky. 69.472 243.Banerjee. C. Science. J.Erbe. Chem. C. Immunol.491-497( 1 9 8 7 ) . 27. J . Pharmacol. Med. 1278-1301 ( 1 9 9 2 ) .. I. Rep. R. H.Hendrickson.D..J.D. Kim. Biophys Acta.1045-1055 ( 1 9 8 9 ) . Med. Int. J. D.65. and J. Rep. 346. and V. 12.D.Frankel and P. Kornberg. 27. Science. W. 119. D.G. 240b. Nature. Bischofberger.860-921 2 4 2 a . Staunton. T.. 242b.Coombes and M. 1 ( 1 9 8 9 ) . Cell. Rosen. I . J. A. Nature. Trends Pharmcol. ( 1 9 9 2 ) . 2 3 9 . Drickamer.8. T. ( 1 9 9 3 ) . J. E.99-108 ( 1 9 9 3 ) . 409. Nature. D. K. 56. 127-134 ( 1 9 9 2 ) . Science. H. D. Annu. Scien'ce.A. Med.11-25 ( 1 9 9 2 ) . McLean. Yednock and S.264. and W. Chem. 291.J. Sci. Venter. 238. F.Charo.717-719 ( 1 9 9 1 ) . et al. A.311-320 236. E. and D. N. Bouchere. N. 0. Schneider. 236. T .. Eng.30. S. 26. Snouwaert. Chem. C. -2 7 2 ( 1 9 9 1 ) . M. A.3-6 ( 1 9 9 0 ) . G. L. 2 6 5 J. A. Sall. Chem. P. A. M..A. Biotechnology. 105-107 ( 1 9 9 0 ) . Sommergruber. R.. J. Brigman. A. and R.Pierchbaucher.. Rep. Med. Biol. Annu.. ( 1 9 8 7 ) .Smithies. H. Evans. S. Med. 240a. Catapano. Shea and J. L.

1 Definition.1 5'-Capping.5 Binding and Effects of Binding to Nonnucleic Acid Targets. 122 4 Molecular Mechanisms of Antisense Drugs.4 Activation of Double-Strand RNases.1.2.3 Strategies to Induce Transcriptional Arrest. CROOKE Isis Pharmaceuticals.2 Inhibition of 3'-Polyadenylation.1. 123 4. Volume 2: Drug Development Edited by Donald J.1.1 Inhibition of 5'-Capping. 116 2 History. 121 3. 117 2.1 Polynucleotides. Inc. 116 1. 119 2. 126 4.2 Inhibition of Splicing.1. 120 3.1.1. 128 rger's Medicinal Chemistry and Drug disco^'cry S i x t h Edition. 121 3. 118 2. 122 3. Inc.1 Occupancy-Only-Mediated Mechanisms. 120 2.6 Terminating Mechanisms. 124 4. 122 3.2 Occupancy-Activated Destabilization. 120 3. 125 4. 117 2.3 RNA Structure.1 Positive Demonstration of Antisense Mechanism and Specificity.2. 125 4.1.1.1 Factors That May Influence Experimental Interpretations. 117 2.4 Disruption of Necessary RNA Structure.1 Oligonucleotide Purity.4 Variations in i n Vitro Cellular Uptake and Distribution. 121 3.8 Kinetics of Effects.1.5Combinatorial Approaches to Oligonucleotide Therapeutics. 123 4.2 The Antisense Concept.2 Oligonucleotide Structure.2 Recommendations.1. 124 4. 120 3 Proof of Mechanism. 123 4. 126 4. 123 4.3 Translational Arrest.CHAPTER FIVE Oligonucleotide Therapeutics STANLEY T. 120 3. 122 3. Abraham ISBN 0-471-37028-2 O 2003 John Wiley & Sons. . 125 4.4 Ribozymes. 121 3.7 Effects of "Control Oligonucleotides". Carlsbad.6 The Medicinal Chemistry of Oligonucleotides.1. California Contents 1 Introduction.2.3 Other Mechanisms.1. 122 3.3 Activation of RNase H.2.

Antisense technology exploits oligonucleotide analogs (typically 15-20 nucleotides) to bind to cognate RNA sequences through WatsonCrick hybridization.5 Other Toxicities. 139 6. antisense technology has matured.1 Definition During the past decade.8. the emerging data showing the activity of a number of antisense drugs in clinical trials.8. and the overwhelming evidence from studies in animals. has never before been considered in the context of drug-receptor interactions.2 ISIS 3521.3 in Vivo.8. 149 7. WatsonCrick hybridization. 155 1 1 Acknowledgments. 142 6.Oligonucleotide Therapeutics 5 Antisense Technology as a Tool for Gene Functionalization and Target Validation.8.2. Thus antisense technology represents a "new pharmacology.5.1 Acute and Transient Toxicities.8.7.3." oligonucleotides.7.1 Hybridization.1 Nuclease Stability.2 Enhanced Cellular Uptake. 143 6.7.1 ISIS 2302. 150 7. Further.3 Pharmacokinetic Properties.3 G3139. 141 6. 132 6.6.4.3. 143 7.1 Introduction. 142 6. 129 6.2 Topical Administration.8. 138 6. 143 6.4 Sugar Modifications.3. 128 6 Characteristics of Phosphorothioate Oligonucleotides. 143 7. 134 6.8 Toxicology.8. 132 6. 144 7.2 in Vivo Pharmacological Activities.9 Conclusions. 136 6.1 Molecular Pharmacology.6 Pharmacological Properties. 148 7.5.3. 135 6. 142 6.8.3. 147 7.5 in Vivo Effects.6. 143 6.1 Complement Activation.8. 138 6.3 Coagulation. 138 6. the potential of antisense as a therapeutic technology is now better appreciated. 153 8 2'-0-(2-Methoxyethyl) Chimeras: SecondGeneration Antisense Drugs. .4 ISIS 2503. 147 7.5 In Vivo Pharmacokinetics.3.3.1 Aerosol Administration. 141 6.5. Finally.1 2'-Modified Oligonucleotides with Ability to Activate RNase H. resulting in the destruction or disablement of the target RNA.1 Nuclease Stability.1 in Vitro.153 9 Conclusions.4.4. antisense technology is generally accepted as a broadly useful method for gene functionalization and target validation. to address its role in gene functionalization and target validation as well as therapeutics.1 Pyrimidine Modifications. Today.2 Purine Modifications.4. 144 7.3 Summary.8. The basis of the drug-receptor interaction.2.3. 139 6. Before the advent of antisense technology.6 Summary. 149 7. 139 6. 152 7. 143 7 The Medicinal Chemistry of Oligonucleotides.3 RNA Cleaving Groups.5 ISIS 5132. 156 1 INTRODUCTION 1. The purposes of this review are to provide a summary of the progress in the technology. 155 10 Abbreviations. messenger RNA (mRNA).4 Platelet Effects. 129 6. 139 6. 149 7." The receptor. 139 6.2 Interactions with Proteins.4. postbinding events such as recruitment of nucleases to degrade the receptor RNA had never been considered from a pharmacological perspective.7. 133 6. 146 7. 139 6.2 Heterocycle Modifications. and to consider the limitations of the technology and a few of the many questions that remain to be answered. with Vitravene's approval by regulatory agencies around the world (making it the first antisense drug to be commercialized). had never been considered as a potential binding event for drugs and put into a pharmacological context.2 Genotoxicity.2 Subchronic Toxicities. 132 6.4 RNase H Activity.4 Human Safety.7.3 Oligonucleotide Conjugates. 129 6.7 Clinical Activities. no medicinal chemistry had been practiced on the putative "drugs.4. 135 6.8.5 Backbone Modifications. 143 6.2 Cytokines. 134 6.4 in Vitro Cellular Uptake. 134 6. 152 7.

The mechanisms resulting in the toxicities are still not clearly understood. and metabolism of the receptors for these drugs. As with any of the class of drugs. In short. the molecular mechanisms by which might induce antitumor effects were and numerous other studies ed to demonstrate antitumor activities In contrast to studies on DNA as a theratic agent. Although poly r1:poly rC was shown to have tiviral and antitumor activities in animals. Much of the effort focused on the ability polynucleotides to induce interferon (5) and e most thoroughly studied polynucleotide this regard is double-stranded polyriboosine:polyribonocytidine (poly r1:poly rC) ). and T-cell activation. antibody production. vivo pharmacokinetic properties of poly r1:poly rC were extensively studied. 7). poly r1:poly rC and other polynucleotides designed to induce interferon failed to demonstrate substantial antiviral or anticancer activities at doses that did not produce unacceptable toxicities. a mismatched poly r1:poly rC 12U resulting from mispairing of the duplex (16). Methods to stabilize.A key to understanding antisense technology is to consider it in a pharmacological context. resulting in fever. Polycationic substances such as DEAE-dextran (diethylyaminoethyl dextran). complexed with polylysine in the presence of carboxymethyl cellulose. anti-RNA antibodies asglomerular nephritis and cennervous system (CNS) toxicities were inent. In rats. Thus. ver. Pain. platelet depletion. Complex formation with polycations altered pharmacokinetic properties but did not enhance efficacy and probably exacerbated toxicities. DNA from several sources dismor activity and the activity was ported to vary as a function of size. substantially more work has n reported on RNA and polyribonucleoes. e compound produced substantial toxicities animals and humans that limited its utility. Poly r1:poly rC was shown to have potent tiviral and antitumor activities in vitro and vivo and these activities were generally corlated with interferon induction (6).2 The Antisense Concept Clearly. the antisense concept derives from an understanding of nucleic acid structure and function and depends on Watson-Crick hybridization (18). It is essential to consider the future in the context of advances in antisense biology and medicinal chemistry that result in improved pharmacological be- 2. Although the initial efforts that focused on polynucleotides that stimulate a variety of immunological events are not directly applicable to the more recent focus on specific effects of oligonucleotides. in- cluding humans. and convulsions were also reported (7-15). although it has been reported to have broader activities and lower toxicities and is still in development. function. Ampligen has properties similar to those of poly r1:poly rC.1 Polynucleotides Before the evolution of effective transfection ethods and an understanding of molecular hniques. enhance cellular uptake. they have provided a base of toxicological information on the effects of polyanionic compounds and guidance in design of toxicological studies. was studied in patients with cancer and found to be very toxic (12. 2. see Ref. DNA and RNA were ads potential therapeutic agents. the demonstration that nucleic acid hybridization is . n mice. In rabbits. This compound has been shown to induce interferon and to activate 2'5'-adenosine synthetase (17). arguably. hypotension. Poly r1:poly rC. poly r1:poly rC was immunotoxic. r example. and alter the in. 13). poly L-lysine. and histones were shown to complex with poly r1:poly rC and other polynucleotides and alter all of these properties (7). A second polynucleotide that has been studied extensively is ampligen. poly r1:poly rC lethal at low doses and in mammals. base and secondary structure (1-3). it is essential to consider the effects of antisense oligonucleotides in the context of dose-response curves. It is essential to understand the structure. fever was dose limiting review.

13 nucleotides long. that was complementary to a sequence in the respiratory syncytial virus genome. 34). 24). methylphosphonates (21). Miller and Ts'o initiated studies on the neutral phosphate analogs. Pursuit of several strategies has resulted in significant progress (for review. In this publication. interest in oligonucleotide therapeutics intensified. Considerable progress in creating chemical motifs capable of binding to duplex DNA with high affinity and specifici- . the Food and Drug Administration.and 5'-terminal modifications and showed evidence of antiviral activity. when advances in several areas provided technical solutions to a number of impediments. Polynucleotides were reported to form triple helices as early as 1957. Though less precisely focused on the therapeutic potential of antisense oligonucleotides. the explosive growth in availability of viral and human genomic sequences provided the information from which "receptor sequences" could be selected. see Ref. Thus.3 Strategies to Induce Transcriptional Arrest An alternative to the inhibition of RNA metabolism by way of an antisense mechanism is to inhibit transcription by interacting with double-stranded DNA in chromatin. in fact. triple-stranded structures can be formed between a third strand composed of pyrimidines or purines that interact with a homopurine strand in a homopurine-homopyrimidine strand in a duplex DNA. is the synthesis in 1969 of phosphorothioate poly r1:poly rC as a means of stabilizing the polhu- cleotide (25).Oligonucleotide Therapeutics feasible (19) and the advances in in situ hybridization and diagnostic probe technology (20) lay the most basic elements of the foundation supporting the antisense concept. see Ref. Despite . triple-stranding strategies. Although there was initially considerable debate about the value of triple-stranding strategies vs. 30-33). they discussed possible sites for binding in RNA and mechanisms of action of oligonucleotides. these authors reported the synthesis of an oligodeoxyribonucleotide. However. 34). until recently.the observations of Miller and Ts'o and Zamecnik and colleagues. the first clear enunciation of the concept of exploiting antisense oligonucleotides as therapeutic agents was in the work of Zamecnik and Stephenson in 1978. attracted essentially all of the attention. forming the foundation of oligonucleotide therapeutics. Of the two most obvious binding strategies for oligonucleotides binding to double-stranded nucleic acids. With the demonstration that homopyrimidine oligonucleotides could indeed form triplex structures (35-37). Because antisense drug design requires an understanding of the sequence of the RNA target.there was little debate that much work remained to be done to design oligonucleotides that could form triple-stranded structures with duplexes of mixed sequences. antisense approaches (38). The inception of the third key component (medicinal chemistry). Subsequently. Triple strands can form by non-Watson-Crick hydrogen bonds between the third strand and purines involved in Watson-Crick hydrogen bonding with the complementary strand of the duplex (for review. interest in triple-strand approaches to inhibit transcription heightened. 2. With these advances forming the foundation for oligonucleotide therapeutics and the initial studies suggesting in vitro activities against a number of viral and mammalian targets (28. Their focus on methyl phosphotriester-modified oligonucleotides as a potential medicinal chemical solution to pharmacoki* netic limitations of oligonucleotides presaged a good bit of the medicinal chemistry yet to be performed on oligonucleotides. They suggested that this oligonucleotide could be stabilized by 3'. More important. strand invasion and triple-strand formation. 22). The development of methods for synthesis of research quantities of oligonucleotide drugs then supported antisense experiments on both phosphodiester and modified oligonucleotides (23. interest in antisense research was quite limited until the late 1980s. and groups at the National Institutes of Health. and the Worcester Foundation investigated phosphorothioate oligonucleotides (26-29). the work of Miller and Ts'o and their collaborators during the same period helped establish the foundation for antisense research and reestablish an interest in phosphate backbone modifications as approaches to improve the properties of oligonucleotides (21.

41). for example. To achieve potential therapeutic utility.. 44-46). 57). Additionally. In fact. to increase nuclease resistance and support retained ribozyme activity (59-61). The ribozyme .c. and both ham- merhead and hairpin ribozymes. progress in developing the basic tools with which to evaluate the potential of sequence-specific interactions with DNA has been reported. and other non-ribozyme antisense-based mechanisms are emerging as alternatives. In either case.4 Ribozymes Ribozymes are RNA molecules that catalyze biochemical reactions (54). Perhaps. but thorough comparative dose-response curves were not reported (58). Further. has been shown to be feasible if analogs with sufficient affinity can be synthesized. an alternative approach to obstructing transcription by formation of triple strands. the pharmacokinetics of a relatively nuclease-stable hammerhead ribozyme were determined after intravenous i v . Thus. Modifications including phosphorothioates and nucleoside analogs have been demonstrated to be incorporable in many sites in hammerhead ribozymes. the value of a ribozyme relative to that of an antisense inhibitor that recruits a cellular enzyme such as RNase H or a double-strand RNase to cleave the target an RNA would be defined as the difference in either specificity or potency achieved by the ribozyme compared to that of the antisense inhibitor vs. progress in developing sequence-specific minor groove binders has been reported (52) (for review.2 History 119 ties supportive of binding to sequences other than polypyrimidine polypurine traits has been reported. Strand invasion. In addition to advances in the chemistrv " of triplex formation. Modifications such as 7-deazaxanthine and 2'-aminoethoxy were reported to enhance triplex formation (40. Ribozymes cleave single-stranded regions in RNA through transesterification or hydrolysis reactions that result in cleavage of phosphodiester bonds (55). several RNA catalytic motifs. Substantial progress has also been reported with regard to the synthesis and testing of nuclease-resistant ribozyme drugs. although much remains to be done. a number of studies have shown results in cells consistent with triplex formation in chromosomal DNA (39. have been identified (56).g. most important. group I introns. as discussed later.) administration to mice. . A comparison of a ribozyme and a phosphorothioate oligodeoxynucleotide to the 5'-transactivator region (TAR) of HIV showed a slight improvement in activity. 43). see Ref. To date. to block DNA polymerases (47-49). RNase P. sGomelysin. 2. Ribozyme-forming sequences have been incorporated into plasmids and administered. 53). Thus. or intraperitoneal (i. and to inhibit binary of a number of proteins to DNA (50). splicing inhibition.1. RNase H has proved to be a remarkably robust mechanism and double-stranded RNase activation. Additionally.To date. They have been shown to inhibit transcription initiation and elongation. as ribozyme gene therapy (e. ~ R N levels A in knee joints of rabbits after intra-articular injection (62). nucleic acids have been shown to be of particular value for triplex interactions because of the relatively high affinity of this modification. two approaches have been taken. subcutaneous (s. in effect. the object is to take advantage of the specificity of hybridization-based interactions and couple that to improved potency that might derive from the ability of the ribozyme to cleave the target RNA. peptide nucleic acid (PNA) has been shown to form on triplestranded structures in isolated DNA and in mouse cell chromosomal DNA (39). 5'-propionylmodified nucleosides have been shown to enhance triplex formation (42) (for review. PNAs have been shown to have verv " high affinity and be capable of strand invasion of double-stranded DNA under some conditions (51). Pe~tide . To date no comparative data from in vivo experiments have been reported. the costs and limitations imposed by the structural requirements to effect ribozymic activity. see Ref.Dimeric PNAs have been created that are reported to form PNAIDNAIPNA triplexes and these have been shown to induce genetargeted mutations in streptolysin-o permeabilized cells (39). data that address this issue are limited. see Ref. For example. modified relatively nuclease-resistant ribozvmes were reported to decrease the target.p.

2. Currently.1. a ribozyme designed to inhibit hepatitis C virus replication is in early clinical trials. bone marrow. These can then be screened for potential activities without knowledge about the cause of the disease or the structure of the target. studies from many laboratories demonstrated that capped oligonucleotides were relatively rapidly degraded in cells (70-72). Clearly. type of terminating mechanism. 56). substantial progress has been reported with regard to both ribozyme gene therapy and synthetic ribozyme drug therapy. 3.Oligonucleotide Therapeutics was well absorbed after i. or s. phosphorothioate oligonucleotides can be pre- . As previously mentioned. with careful evaluation of mechanism of action. sugars. Both fully modified oligonucleotides and oligonucleotides "capped" at the 3'.5 Combinatorial Approaches to Oligonucleotide Therapeutics 2. backbones. However. including novel bases. Ribozyme and antisense gene therapy has been used fairly widely in vitro to determine the biological roles of various genes (for review. see Ref.c. perhaps the most interesting of the initial modifications were the phosphate analogs. 74). a variety of modifications were rapidly tested. see Ref. see Ref. must be generated. we understand essentially nothing about the interplay of these factors. Thus. cholesterol (75). kidney.p. Many of these modifications have proved to be quite useful and are progressing in testing leading to clinical trials. Additionally. At present. the challenges gene therapy faces must be met and ribozymes must be shown to be of more value than simply expressing antisense genes. Clearly. and chimeric oligonucleotides have been tested (for review. an enormous range of modifications. Nor were point modifications with intercalators that enhanced binding to RNA (73.6 The Medicinal Chemistry of Oligonucleotides Because it was apparent almost immediately that native phosphodiester oligodeoxy.1 Oligonucleotide Purity. substantial hurdles remain before sufficient data derived from animal and human studies with multiple ribozymes define the potential of the approach. Nevertheless. the ultimate biological effect of an oligonucleotide will be influenced by the local concentration of the oligonucleotide at the target RNA. 78). distributed to liver. data in animals and humans for numerous ribozymes.The potential advantage of a combinatorial approach is that oligonucleotide-based molecules can be prepared to adopt various structures that support binding to nonnucleic acid targets as well as nucleic acid targets. for ribozyme gene therapy to be broadly applicable. To validate synthetic ribozymes. 2'-0-(2-Methoxyethyl) chimeric antisense inhibitors are now in clinical trials. In addition to challenges associated with gene therapy. Clearly. the 30-min elimination half-life must be lengthened with new modifications and again the value of a ribozyme versus a much simpler antisense approach must be defined. Since the initial approaches to modifications of oligonucleotides. the concentration of the RNA.and/or 5'-termini with phosphorothioate or methylphosphonate moieties were tested (69). the phosphorothioates (68) and the methylphosphonates (21).or ribonucleotides are unsatisfactory as drugs because of rapid degradation (67). 77) sufficiently active or selective to warrant broad-based exploration. This ribozyme designed to inhibit vascular endothelial growth factor (VEGF) receptor synthesis has been reported to be in clinical trials. and the rates of the events that result in termination of the RNA's activity. 56).1 Factors That May Influence Experimental Interpretations At least two methods by which oligonucleotides can be created combinatorially have been published (64-66). the rates of synthesis and degradation of the RNA. and other tissues. and displayed an elimination half-life of 33 min (63). 3 PROOF OF MECHANISM 3. or poly L-lysine (76. conjugates. identifying optimal sites in target RNAs for ribozyme binding and the colocalization of the antisense transcript and the target RNA are issues of concern about which progress has been reported (for review. dosing.

This in influences the optimal length of an oligocleotide needed to achieve maximal affinity cause in structured RNA the optimal affinty is determined by the difference between e binding energies required for a nucleotide invade a duplex and those gained per nucleby binding of the oligonucleotide. RNA tructure produces asymmetrical binding ites that then result in very divergent affinity onstants. Phosphorothioate oligo- nucleotides tend to bind to many proteins and those interactions are influenced by many factors. Finally.1. RNA is structured. and excretion. our laboratory reported that different synthetic and purification procedures resulted in oligonucleotides that varied in cellular toxicity (72) and that potency varied from batch to batch. it is obvious that a generalization that all phosphorothioates are taken up by all cells in uitro primarily by receptormediated endocytosis is simply unwarranted. even after careful consideration of all in uitro uptake data. such as the type of medium. unpublished results. in fact. Before that time.2 Oligonucleotide Structure. metabolism. Within the phosphorothioate class of oligonucleotides. distribution.Moreover. this has been the case for only the past several years.g. but not linearly.. Furthermore. it should be studied in several cell lines. given the numerous proteins that interact with RNA that undoubtedly influence binding. At present. Antisense 3. synthetic methods were evolving and analytical methods were inadequate. stretches of guanosine residues) are prone to adopt more complex structures (80). new synthetic. resulting in more stable oligonucleotide duplexes than would be expected based on rules derived from work with oligodeoxynucleotides (Freier.861. Furthermore. undoubtedly they complicated earlier studies. uptake varies as a function of length.1. before an oligonucleotide could be said to be inactive in vitro. and very little is understood about these ternary interactions. 85).82). We now understand that certain sequences (e. Cell type has a dramatic effect on total uptake. purification. and pattern of subcellular distribution. one cannot predict in uivo pharmacokinetics of the compounds (85. The oligonucleotide chemical class obviously influences the characteristics of uptake as well as the mechanism of uptake. depending on many conditions (72. Given the foregoing.3 Proof of Mechanism pared consistently and with excellent purity (79). degree of confluence. In fact. They may induce non-antisense effects that can be mistakenly interpreted as . although it may be absolutely correct that receptor-mediated endocytosis is a mechanism of uptake of phosphorothioate oligonucleotides (87).The potential to form secondary and tertiary structures also varies as a function of the chemical class.1. Although there are no longer synthetic problems with phosphorothioates. with each new analog class. Uptake varies as a fundion of sequence. this is only a fraction of the story. he structure of the RNA has a profound inence on the affinity of the oligonucleotide d on the rate of binding of the oligonucleoide to its RNA target (81.1. ries have clearly demonstrated that cells in tissue culture may take up phosphorothioate oligonucleotides through an active process. 1990). there are now several lines of evidence in animals and humans that.5 Binding and Effects of Binding to Nonnucleic Acid Targets.4 Variations in In Vitro Cellular Uptake and Distribution. The effects of binding can influence cell uptake. More important. 3. Studies in several laborato- oligonucleotides are designed to be single stranded.3 RNA Structure. extrapolations from in uitro uptake studies to predictions about in uiuo pharmacokinetic behavior are entirely inappropriate and. However. depending on the position of oligocleotide in that structure (82-84). higher affinity 2'modified oligonucleotides have a greater tendency to self-hybridize. 3. Tissue culture conditions. and the presence of serum.88-90). Thus. there is no unifyng hypothesis to explain these differences. it is obvious that conclusions about in uitro uptake must be very carefully made and generalizations are virtually impossible. and that the uptake of these oligonucleotides is highly variable. can all have enormous effects on uptake (85). and stability in cells is also influenced by sequence (85. 3. and analytical challenges are encountered. For example. kinetics of uptake.

Unfortunately. cell type. Thus. In brief. 3. Northern blot. high concentrations of drugs. and metabolism of an oligonucleotide in cells. it is difficult to know whether non-antisense mechanisms would be as dominant in vivo or result in antiviral activity. This has resulted in considerable confusion. we know little to nothing about the potential biological effects of such "controls" and the more complicated a biological system arid test. and influenza virus. then. oligonucleotide chemical class. and similar pharmacokinetic considerations in animals.6 TerminatingMechanisms. Fortunately. essentially no data bearing on such interactions are currently available. extrusion. ination rates of the drugs tested (see below). RT-PCR. because variations in terminating mechanism may result in significant changes in antisense potency and studies have shown significant variations from cell type to cell type in vitro. 3. The sensitivity of each virus to non-antisense effects of oligonucleotides varies. For RNase H-activating oligonucleotides.8 Kinetics of Effects. Finally. it is essential to positively demonstrate effects consistent with an antisense mechanism. Oligonucleotides may interact not only with proteins but also with other biological molecules. binding to proteins can certainly have toxicological consequences. the rates of distribution. By inhibiting RNase H.1. a demonstration that closely related isotypes are unaffected should be included. In these assays.Oligonucleotide Therapeutics antisense or complicate the evaluation of whether the pharmaceutical effect is the consequence of an antisense mechanism.2 Recommendations amply demonstrated that oligonucleotides may employ several terminating mechanisms.2. the non-antisense effects have been so dominant that identifying oligonucleotides that work through an antisense mechanism has been difficult. 3. the mechanism of that activity should be explored carefully before concluding that the effects of the "control oligonucleotide" prove that the activity of the putative antisense oligonucleotide are not the result of an antisense mechanism. Unfortunately. Unfortunately. protein binding may inhibit the antisense activity of some oligonucleotides. Obviously. and cells are often coincubated. Ideally. the rates of uptake into cells. It has been 3. . herpes simplex viruses. RNase protection assays. for HIV. 3. more careful kinetic studies are required and more rational in vitro and in vivo dose schedules must be developed. cytomegaloviruses. depending on the nature of the virion proteins and the characteristics of the oligonucleotides (91. in the past several years. such as the rate of synthesis and degradation of the target RNA and its protein. The predominant terminating mechanism is influenced by RNA receptor site. In particular. Many rate con- stants may affect the activities of antisense oligonucleotides. at present. An especially complicated experimental situation is encountered in many in vitro antiviral assays. and probably many other factors (93). or transcriptional assay analyses showing selective loss of the target RNA are the ideal choices and many laboratories are publishing reports i n vitro and in vivo of such activities (94-97). and such interactions like those with proteins will be influenced by the chemical class of oligonucleotide studied. it is essential that the terminating mechanism be well understood.7 Effects of "Control Oligonucleotides". Nevertheless. when a control oligonucleotide displays a surprising activity.92). such as lipids or carbohydrates. for proof of mechanism. many more careful dose-response and kinetic studies have been reported and in general they demonstrated a relatively slow onset of action and a duration of response consistent with the elim-. Until more is un- derstood about how antisense drugs work.1 Positive Demonstration of Antisense Mechanism and Specificity.1. viruses. the more likely that "control" oligonucleotides may have activities that complicate interpretations. Given the artificial character of such assays. the following steps are recommended.1. A number of types of control oligonucleotides have been used including randomized oligonucleotides. our understanding of terminating mechanisms remains rudimentary.

inhibition of 5'-capping of mRNA could be an effective antisense mechanism. Evaluate specificity and therapeutic indices through studies on closely related isotypes and with appropriate toxicological studies. Perform sufficient pharmacokinetics to define rational dosing schedules for pharmacological studies. The effects of RNase H-competent oligonucleotides were compared to those of oligonucleotides that do not serve as RNase H substrates. The major conclusions from this study were. no reports of antisense inhibitors of capping have appeared. To date. Use RNase H protection assays and transcriptional arrays to provide broader analyses of specificity where the assays have been validated. To this end. 3. or the phosphate.conclusions about mechanisms of action were drawn without using appropriately modified oligonucleotides.1. A key step in Classic competitive antagonists are thought to alter biological activities because they bind to receptors. and phosphorothioate backbones (31. numerous analogs have been identified that do not support RNase H cleavage.4 Molecular Mechanisms of Antisense Drugs 1. They showed that 2'-MOE phosphorothioate oligonucleotides could correct aberrant beta-globin splicing in a cell-free system.2 lnhibition of Splicing. When control oligonucleotides display surprising activities. which may be attributable to the inaccessibility of the 5'-end of mRNA before capping. 4 MOLECULAR MECHANISMS OF ANTISENSE DRUGS 4. Fortunately.1 lnhibition of 5'-Capping. oligonucleotides that bind to sequences required for splicing may prevent binding of necessary factors or physically prevent the required cleavage reactions. Directly demonstrate the proposed mechanism of action by measuring the target RNA and/or protein. Similar observations (104). These can be the intermediary metabolism of most mRNA molecules is the excision of introns. Perform careful "gene walks" for all RNA species and oligonucleotide chemical classes. methylphosphonate. 4. 6. however. the antisense agents must be modified sufficiently not to support RNase H cleavage. 7. 5. or replacement of the sugarphosphate backbone (98). These "splicing" reactions are sequence specific and require the concerted action of spliceosomes. in a number of earlier ~ublications. Conceptu- ally. Kole and colleagues (102-104) were the first to use modified oligonucleotides to inhibit splicing. were recorded in a cellular svstem " In our laboratory. we have attempted to characterize the factors that determine whether splicing inhibition is effected by an antisense drug (105). that most of the earlier stud- . Then the effects of antisense drugs designed to bind to various sites were characterized. Activities have been reported for anti-c-rnyc and antiviral oligonucleotides with phosphodiester. determine the mechanisms involved. other nucleic acids. Perform careful time courses before drawing conclusions about potency. This would then result in inhibition of the production of the mature mRNA. Consequently.1 Occupancy-Only-Mediated Mechanisms classified into modifications of sugar moiety. 8. 4. thereby preventing agonists from binding the inducing normal biological processes. 99-101). Perform careful dose-response curves in vitro using several cell lines and methods of in vitro delivery. first. 2. To create antisense inhibitors that clearly work through non-RNase H mechanisms.1. a number of luciferase-reporter plasmids containing various introns were constructed and transfected into HeLa cells. or other factors required for essential steps in the intermediary metabolism of the RNA or its se by the cell. 5'-Capping is critical in stabilizing mRNA and in enabling the translation of mRNA (99). Correlate the rank-order potency in vivo with that observed in vitro after thorough dose-response curves are generated in vivo. Unfortunately. Binding of oligonucleotides to specific sequences may inhibit the interaction of the RNA with proteins. 9. 4.

oligonucleotides designed to bind to highly conserved sequences in 5. RNA adopts a variety of three-dimen- for which many oligonucleotides have been designed is translational arrest by binding to the translation initiation codon. in only a relatively few studies have the oligonucleotides been shown to bind to the sites for which they were designed. it is surprising that occupancy-based disruption RNA has not been more extensively exploited.1.1. antisense-mediated alternative splicing is a potentially powerful tool with which to investigate this important source of biological diversity. Again. Numerous oligonucleotides sional structures induced by intramolecular hybridization. to demonstrate that RNase H is not involved in effects observed. We synthesized a number of oligonucleotides designed to disrupt TAR. Furthermore. Studies with PNA analogs have shown that these analogs can inhibit translation in cellfree systems. the TAR element. A number of examples purporting to employ this mechanism have been reported. However. the 3'-splice site and branchpoint are usually the best sites to which to target the oligonucleotide to inhibit splicing. vesicular stomatitis virus (VSV) (761. and ribonucleoproteins that participate in the intermediary metabolism and activities of RNA species.8 S RNA inhibited protein synthesis in rabbit reticulocyte and wheat germ systems (121). Similar results were reported for antisense agents designed to bind to Bclx pre-mRNA (107. A mechanism with 2'-modifications have also been studied and have been shown to inhibit translation when targeted to 5'-UTR or the translation initiation codon (118). Although designed to induce relatively nonspecific cytotoxic effects. given the potential general activity of the mechanism. Second. For morpholino oligomers. As an example. 4. Target RNA species that have been reported to be inhibited include HIV (28). 115). However. 2'0-methoxyethyl oligonucleotides redirected splicing of IL-5 receptor pre-mRNA (106). Several studies have now demonstrated antisense-mediated redirection of slicing of an endogenous cellular mRNA. the most common of which is the stem loop.g. and inhibit TAR-mediated production of a reporter gene (66). antisense activity has been reported in both cell-free and cellular assays (116. Third. Oligonucleotides designed to bind to a 17-nucleotide loop in Xenopus 28 S RNA required for ribosome stability and protein synthesis inhibited protein synthesis when injected into Xenopus oocytes (120). and a number of normal cellular genes (110113). 4.. translation arrest represents an important mechanism of action for antisense drugs. They are used to provide additional stability for RNA and as recognition motifs for a number of proteins. two other examples are noteworthy. general rules useful in disrupting stem-loop structures were developed as well (84). it is necessary to use antisense drugs that do not form duplexes that are RNase H substrates (e. but to date no data have been reported from cellular studies (114. These structures play crucial roles in a variety of functions. The positioning of the initiation codon within the area of complementarity of the oligonucleotide and the length of oligonucleotide used have varied considerably. 117). we designed a series of oligonucleotides that bind to the important stem loop present in all RNA species in HIV. optimal inhibition is effected by binding at the 5'-cap in RNAs that have significant 5'-untranslated regions (119).Oligonucleotide Therapeutics ies in which splicing inhibition was reported were probably the result of nonspecific effects. and other data that support translation arrest as the mechanism have not been reported. Thus. nucleic acids. In conclusion.4 Disruption of Necessary RNA Structure. and to create focused therapies for diseases caused by disorders is splicing. Similarly. showed that several indeed did bind to TAR. fully modified 2'-oligonucleotides). less effectively spliced introns are better targets than those with strong consensus splicing signals. unfortunately. disrupt the structure.3 Translational Arrest. In one study. 108). and recent studies on several compounds have provided data that unambiguously demonstrate that this mechanism can result in potent antisense drugs. n-myc(l09). . Thus.

1 5'-Capping. Several oligonucleotides that bind near the cap site have been shown to be active. 4. interactions in the 3'-terminal region of pre-mRNA could inhibit polyadenylation and destabilize the RNA species. In our laboratory. It also is important in binding to the nuclear matrix and transport of mRNA out of the nucleus. many other regulatory features and mechanisms will be identified.2 Inhibition of 3'-Polyadenylation. subcellular distribution. Pre-mRNA is transcribed from a gene. In fact. it presents an interesting target. Theoretically.2.__-Degradation CAP \ 3' 2 - Nucleus : mRNA CAP J Translation in Figure 5.4 Molecular Mechanisms of Antisense Drugs Transcription +Transcriptional arrest 7 Capping1 polyadenylation Splicing CAP + AAAA CAP A+ . in no published study have the oligonucleotides been shown to bind to the sequences for which they were designed.2. It is likely that. 3 Effects on Catabolism o f mRNA Translational arrest 4. we have designed oligonucleotides to bind to 5'-cap structures and reagents to specifically cleave the unique 5'cap structure (123). Although there are a number of oligonucleotides that interact in the 3'-untranslated region and .( + vLI ' Effects on Anabolism o f ---__ ----________ ___-_ ---. and transport. various processing events. How- 3'-untranslated regions of pre-mRNA molecules are sequences that result in the posttranscriptional addition of long (hundreds of nucleotides) tracts of polyadenylate. This stabi- ever. Because the structure of the cap is unique and understood. Polyadenylation stabilizes mRNA and may play other roles in the intermediary metabolism of RNA. 5. presumably by inhibiting the binding of proteins required to cap the RNA. again. the synthesis of SV40 T-antigen was reported to be most sensitive to an oligonucleotide linked to polylysine and targeted to the 5'-cap site of RNA (122). In the lizes pre-mRNA and is important for the stability of mature mRNA. These studies demonstrated that 5'-cap-targeted oligonucleotides were capable of inhibiting the binding of the translation initiation factor eIF4a (124).1. A number of structural features of RNA are known to influence stability.1). For example. A key early step in RNA processing is 5'-capping (Fig.2 Occupancy-Activated Destabilization RNA molecules regulate their own metabolism. in no published study has this putative mechanism been rigorously demonstrated. 4. as RNA intermediary metabolism is better understood.

136). oligonucleotides composed of wings of 2'-methoxy phosphonates and a five-base gap of deoxyoligonucleotides bind to their target RNA and activate RNase H (141. the concentration of the enzyme in the nucleus is thought to be greater and some of the enzyme found in cytoplasmic preparations may be attributed to nuclear leakage. However. 138). Additionally.Changes in the sugar influence RNase H activation as sugar modifications that result in RNA-like oligonucleotides. chimeric molecules have been studied as oligonucleotides that bind to RNA and activate RNase H (141. For example. In addition to 5'- capping and 3'-adenylation. no study to date has reported evidence for alterations in polyadenylation (125).142). phosphorothioates are excellent substrates (74. 134). it may play other roles in the cell and is found in both the cytoplasm and the nucleus (131).and 3'-untranslated regions of mRNA that affect the stability of the molecules. Again. In addition. It has been identified in organisms as diverse as viruses and human cells (130).3 Activation of RNase H RNase H is a ubiquitous enzyme that degrades the RNA strand of an RNA-DNA duplex.It .142).and 5'-terminal sequences in Rous sarcoma viruses was active (126). inasmuch as the 2'methoxy analog of ISIS 1939 was much less active. backbone modifications influence the ability of oligonucleotides to activate RNase H. there are clearly other sequences in the 5'. in addition to destabilization to cellular nucleolytic activity. ISIS 1939 is a 20-mer phosphorothioate that binds to and appears to disrupt a predicted stem-loop structure in the 3'-untranslated region of the mRNA because the intercellular adhesion molecule (ICAM) is a potent antisense inhibitor. composed of methylphosphonate deoxyoligonucleotide wings and phosphodiester gaps. Recent studies in our laboratories provide additional. a single ribonucleotide in a sequence of deoxyribonucleotides was shown to be sufficient to serve as a substrate for RNase H when bound to its complementary deoxyoligonucleotide (143). studies in our laboratories have shown that sequences in the 3'-untranslated region of RNA molecules are often the most sensitive (127-129). Despite the information about RNase H and the demonstration that many oligonucleotides may activate RNase H in lysate and purified enzyme assays.2. At least two classes of RNase H have been identified in eukaryotic cells. In contrast. Although RNase H is involved in DNA replication. albeit indirect. Multiple enzymes with RNase H activity have been observed in prokaryotes (130). Zamecnik and Stephenson reported that 13-mer targeted to untranslated 3'. Against several RNA targets. In fact. 2'-fluoro or 2'-methoxy do not appear to serve as substrates for RNase H (133.140). were compared to full phosphodiester oligonucleotides (145). Furthermore. However. activation of RNase H (see below) is also involved in the activity of ISIS 1939 (94). Methylphosphonates do not activate RNase H (137. insights into these questions. Alterations in the orientation of the sugar to the base can also affect RNase H activation as a-oligonucleotides are unable to induce RNase H or may require parallel annealing (135. for example. 4. The precise recognition elements for RNase H are not known.139. For example. However. 4. relatively little is yet known about the role of structural features in RNA targets in activating RNase H (146-148). it has been shown that oligonucleotides with DNA-like properties as short as tetramers can activate RNase H (132). Oligonucleotides conjugated to an acridine derivative and targeted to a 3'-terminal sequence in type A influenza viruses were reported to be active. ISIS 1939 is a 20-mer phosphorothioate complementary to a sequence in the 3'-untranslated region of ICAM-1 RNA (94). That it is possible to take advantage of chimeric oligonucleotides designed to activate RNase H and have greater aMinity for their RNA receptors and to enhance specificity has also been demonstrated (144.3 Other Mechanisms. RNase H-mediated cleavage of target transcript was much more selective when deoxyoligonucleotides. there are a number of antisense drugs that may work by these mechanisms.Oligonucleotide Therapeutics display antisense activities. In one study. direct proof that RNase H activation is the mechanism of action of oligonucleotides in cells has until very recently been lacking. 145). it is likely that.

placing a positive charge (e. Very recently. we now have the necessary tools to begin to explore the roles of human RNase H proteins in biological and pharmacological processes and to begin to develop drugs designed to interact with them more effectively. Using a series of noncleavable substrates and Michaelis-Menten analyses. but has properties similar to those described for human RNase H type 2 (151.A more direct demonstration that RNase H is likely a key factor in the activity of many antisense oligonucleotides was provided by studies in which reverse-ligation PCR was used to identify cleavage products from bcrabl mRNA in cells treated with phosphorothioate oligonucleotides (149). Finally. but to date the expressed protein has not been shown to be active (154). It is a double-stranded. A 2'-methoxy analog of ISIS 1939 displays higher affinity for the RNA than that of the phosphorothioate. Further. is stable in cells. human RNase H2 has been cloned.g. Given the emerging role of chimeric oligonucleotides with modifications in the 3'. we cloned and expressed human RNase HI. the level of target RNA has no effect on the potency of antisense inhibitors (156). 152). Any 2'-modification in the antisense drug at the cleavage site inhibited cleavage. In one such study on E. Thus. and inhibited by Mnt2 in the presence of Mg'2. we reported that the enzyme displays minimal quence specificity and is processive. 2'-propoxyamine) in the antisense drug reduced affinity and cleavage. The Kd for the RNA duplex was 1. When a chimeric oligonucleotide with 2'-modified sugars in the wings was hybridized to the RNA. ISIS 1570. studies focused on understanding the effects of various modifications on the efficiency of the enzyme(s) are also of considerable importance.and 5'-wings designed to enhance affinity for the target RNA and nuclease stability and a DNAtype gap to serve as a substrate for RNase H. and the Kd for single-stranded DNA was 942 pM. we were able to evaluate both binding and cleavage. coli RNase H. In subsequent studies. We also examined the effects of antisense oligonucleotideinduced RNA structures on the activity of E. Any structure in the duplex substrate was found to have a significant negative effect on the cleavage rate. coli RNase H1 is a double-stranded RNA binding protein.. the initial site of cleavage was the nucleotide adjacent to the methoxy-deoxy junction closst to the 3'-end of the RNA substrate. and the impacts of these interactions on RNase H in more detail (150). We showed that E. Finally. It is likely that ISIS 1939 destabilizes the RNA and activates RNase H.4 Molecular Mechanisms of Antisense Drugs inhibits ICAM production in human umbilical vein endothelial cells and Northern blots demonstrate that ICAM-1 mRNA is rapidly degraded.a result of the number of mol- . the enzyme could cleave RNA only in an RNA-DNA duplex. The id rate of cleavage increased as the size of he DNA gap increased and the efficiency of he enzyme was considerably less against an RNA target duplexed with a chimeric antisense oligonucleotide than against a full DNAtype oligonucleotide (149). coli RNase H1 (150). In contrast. inhibited by higher concentrations. 33kDa molecular weight RNA binding protein and exhibits unique positional and sequence preferences for cleavage (153). In contrast. we recently demonstrated that in the range of 1-150 copies of RNA per cell. we evaluated the interactions of antisense oligonucleotides with both structured and unstructured targets. The enzyme is stimulated by low concentrations of Mg+'. and this was explained by steric hindrance imposed by the RNA loop traversing either the minor or major grooves or the heteroduplex. an 18-mer phosphorothioate that is complementary to the translation initiation codon of the ICAM-1 message. cleavage of selected sites was inhibited entirely.6 pM. depending on the site in the mRNA at which they bound (94). the K. we reported on the effects of several mutations introduced into human RNase H1 on the activity of the enzyme (155). for a DNA duplex was 176 pM. Recently. The protein is homologous to E. but inhibits ICAM-1 protein production much less potently than does ISIS 1939. but significant charge reduction and 2'-modifications were tolerated at the binding site. at least with regard to RNase Hinduced degradation of targeted RNA. Additionally. Thus. two oligonucleotides that are capable of activating RNase H had different effects. inhibited production of the protein but caused no degradation of the RNA. coli RNase HI.

Figure 5.Oligonucleotide Therapeutics 2 hours 2 days 2 days 1 week 1-2 weeks { Identify target gene sequence $ J 1 \ Design antisense inhibitor f J Synthesis o f antisense inhibitor + Cell culture screening Lead inhibitor o f target gene Animal model ( + J 1 1 J Validated genomic target for development by partner Figure 5. This may be an important step forward because it adds to the repertoire of intracellular enzymes that may be used to cleave target RNAs. The processes and timelines required to generate optimal antisense inhibitors using automated systems. In this figure. 4. we have shown that mammalian cells contain enzymes that can cleave doublestranded RNAs (157). other factors must be rate limiting. Figure 5. the RNA is displayed 5' in the translated region to the 3'-untranslated region and each bar rep- . antisense technology has proved to be a versatile. The antisense inhibitors are then screened in vitro by the use of an automated reverse transcriptase/polynuclease chain reaction (RT-PCR).With these advances. In addition.3 shows a typical 96-well-plate screen designed to determine the optimal site for antisense effects.2 provides a general scheme for rapid creation and evaluation of antisense inhibitors. antisense inhibitors to 80 sites in the target RNA are synthesized and screened in TNFa-receptor 1 positive cells. effective tool for these purposes (158). 5 ANTISENSE TECHNOLOGY AS A TOOL FOR GENE FUNCTIONALIZATION A N D TARGET VALIDATION The sequencing of the human and other genomes has resulted in intense interest in the development of tools with which to determine the roles of various gene products and to de- termine whether they are appropriate targets for drug therapy. it is now possible to create antisense inhibitors to hundreds to thousands of genes per year. In this assay. These are coupled to in-line automated analytical methods that provide quality assurance.4 Activation of Double-Strand RNases By the use of phosphorothioate oligonucleotides with 2'-modified wings and a ribonucleotide center. The advances in automation include rapid small-scale synthesizers that can synthesize antisense inhibitors in a 96-well-plate format. the effects of a variety of control oligonucleotides are evaluated. in this case for TNFa-receptor 1. ecules of antisense drug per cell exceeding the number of copies of RNA by several orders of magnitude. Thus. Antisense: a drug discove~y and genomics tool: identical process.2. and because chimeric oligonucleotides 2'-modified wings and oligoribonucleotide gaps have higher affinity for RNA targets than do chimeras with oligodeoxynucleotide gaps. With advances in automation.

including in vivo evaluation in various animal models. When one considers the desired attributes for tools for gene functionalization and target validation. This reduction in Tm per nucleotide is virtually independent of the number of phosphorothioate units substituted for phosphodiesters. In the past decade. association rates for oligonucleotides that display acceptable affinity constants are sufficient to support biological activity at therapeutically achievable concentrations. efficient. phosphorothioate oligodeoxynucleotides must typically be 17. For example.1 a phosphorothioate oligodeoxynucleotide for RNA is approximately 0. several antisense inhibitors can be selected for further study. a phosphorothioate oligodeoxynucleotide has a T . and is much more resistant than the parent phosphorothioate to nucleases (159). so they provide direct insights into the types of responses to be expected from acute interventions with drugs that affect the target. it is readily apparent that antisense technology meets these criteria. 83). the phosphorothioate modification is quite attractive. depending on the structure of the RNA.to 20-mer in length and that invasion of double-stranded regions in RNA is difficult (66. given that the AT. and structure-specific binding events.2% lower per unit (163). Association rates of phosphorothioate oligodeoxynucleotide to unstructured RNA targets are typically lo6-lo7 M-' s l. Having automated antisense. it is rapid. at Isis Pharmaceuticals alone we have created antisense inhibitors to nearly 1000 genes. for a phosphodiester or phosphorothioate oligodeoxynucleotide 18 nucleotides in length with the mismatch centered (163). one of the oxygen atoms in the phosphate group is replaced with a sulfur. Thus. determined by RT-PCR.3% to -l. the class that has resulted in the broadest range of activities and about which the most is known is the phosphorothioate class.O°C.2 Interactions with Proteins Hybridization I1 The hybridization of phosphorothioate oligonucleotides to DNA and RNA has been thoroughly characterized (79. Interestingly. Nonspecific binding to a wide variety . Antisense inhibitors are gene specific. respectively. 144. is chiral at each phosphorothioate phosphodiester.5% less per nucleotide than for a corresponding phosphodiester oligodeoxynucleotide. This modification clearly achieves the objective of increased nuclease stability. Phosphorothioate oligonucleotides were first synthesized in 1969 when a poly(rlrC1phosphorothioate was synthesized. sequence context has some influence. Compared to RNA and RNA duplex formation. independent of oligonucleotide length or sequence (81. each of which may have different characteristics and effects. Antisense inhibitors are pharrnacological agents. Antisense inhibitors are versatile in that they can be used for both in vitro and in vivo studies.6 Characteristics of Phosphorothioate Oligonucleotide resents the level of the target RNA. of Phosphorothioate oligonucleotides bind to proteins. The T . However.7 or 12. This means that to be effective in vitro. value of about -2. and cost effective. can vary from -0. a recent study using phosphodiester oligonucleotides coupled to fluoroscein showed that hybridization was detectable within 15 min after microinjection into K562 cells (165). - ' 6. and other factors (163). in general. 6 CHARACTERISTICS OF PHOSPHOROTHIOATE OLIGONUCLEOTIDES Of the first-generation oligonucleotide analogs. The specificity of hybridization of phosphorothioate oligonucleotides is. In this class of oligonucleotides. clearly interacting in several sites and resulting in significant reductions in the target RNA. from this perspective.8"C reduction in Tm. Based on these data. depending on sequence. a T-C mismatch results in a 7. 6. The interactions with proteins can be divided into nonspecific. so we are confident that the technology can be used for any gene. 164). Association rates to structured RNA targets can vary from lo2to 10sM-' s-'. Said another way. slightly greater than that of phosphodiester analogs. The resulting compound is negatively charged.160-162). site of binding in the structure. 83. Antisense inhibitors can also be used to dissect the roles of splice variants and to identify novel gene functions. sequence-specific.

3. The results are expressed as percentage control RNA determined by RT-PCR analyses. TNF R1 screen. Results from a screen to identify the optimal antisense inhibitor to tumor necrosis factor a receptor 1 (TNFa-Rl).2'-Methoxyethyl chimeric antisense inhibitors were synthesized to 80 sites in the target RNA and their effects on TNFa-R1. . The antisense inhibitors that result in the maximum (smallest bars) are then evaluated with careful dose-response curves and mg with control oligonucleotides to ensure the effects are specific.lsis # ( U n k h k - Figure 5. RNA were evaluated after incubation with TNFa-R1-expressingcells. The effects of the antisense inhibitors are displayed schematically from 5'UTR to 3'UTR right to left on the figure.

Like other oligonucleotides. Phosphorothioate oligonucleotides have been reported to be competitive inhibitors for HIV-reverse transcriptase and inhibit RT-associated RNase H activity (175. At higher concentrations. 174). we were unable to reproduce the results (for review. However. In these studies. Additionally. 169). These compounds are slowly metabolized by both endo. oligonucleotides containing runs of guanosines can form tetrameric structures called G-quartets. 160). 176).6 Characteristics of Phosphorothioate Oligonucleotides 131 of proteins has been demonstrated. phosphorothioate oligonucleotides may interact with a wide range of proteins through several types of mechanisms. Moreover. self-complementary oligonucleotides are avoided. In conclusion. In this study.and exonucleases and inhibit these enzymes (160. However. and toxicologic properties of these molecules. They have been reported to bind to the cell surface protein CD4 and to protein kinase C (PKC) (177). In our laboratories. Again. the K lop5M for bovine serum albumin and 2 to 3 X lop4 M for human serum albumin. the oligonucleotides appear to be competitive antagonists for the DNA-RNA substrate. More recently. They may also complicate studies on the mechanism of action of these drugs. phosphorothioates can adopt a variety of secondary structures. For example. However. phosphorothio- . Phosphorothioates also bind to RNase H when in an RNA-DNA duplex and the duplex serves as a substrate for RNase H (172). little is known about the affinities. In contrast. very little is known about these binding events. Exemplary of this type of binding is the interaction of phosphorothioate oligonucleotides with serum albumin. and may obscure an antisense activity. it has been reported that phosphorothioates bind to an 80-kDa membrane protein that was suggested to be involved in cellular uptake processes (87). Despite this inhibition. For example. 167). or structure specificities of these putative interactions. we have shown potent. Furthermore. The affinity of such interactions is low. non-sequence-specific inhibition of RNA splicing by phosphorothioates (105). in this study. 173. These interactions may influence the pharmacokinetic. At present. and these appear to interact with a number of proteins with relatively greater affmity than that of unstructured oligonucleotides (80). As a general rule. warfarin and indomethacin were reported to compete for binding to serum albumin (168). again. Additionally. interactions with 30. we have shown extensions of only 2-3 nucleotides. no competition between phosphorothioate oligonucleotides and several drugs that bind to bovine serum albumin was observed. Phosphorothioate oligonucleotides can interact with nucleic acid binding proteins such as transcription factors and single-strand nucleic acid binding proteins. Phosphorothioates interact with nucleases and DNA polymerases. Clearly. to avoid duplex formation between oligonucleotides. However. other structures that are less well understood can also form. 171). The K. much more work is required before definitive conclusions can be drawn. Phosphorothioates have been shown to be competitive inhibitors of DNA polymerase a and p with respect to the DNA template. in experiments in our laboratory.and 46-kDa surface proteins in TI5 mouse fibroblasts were reported (170). binding and competition were determined in an assay in which electrospray mass spectrometry was used. sequences. in a study in which an equilibrium dissociation constant was derived from an assay using albumin loaded on a CH-Sephadex . The inhibition of these enzymes appears to be competitive and this may account for some early data suggesting that phosphorothioates are almost infinitely stable to nucleases. several studies have suggested that phosphorothioates might serve as primers for polymerases and be extended (140. pharmacologic. presumably by binding as a single strand to RNase H. see Ref. in a range similar to that of aspirin or penicillin (166. if possible. value for albumin is approximately 200 phi. a full explanation as to why no longer extensions are observed is not available. Various viral polymerases have also been shown to be inhibited by phosphorothioates (140). value ranged from 1 to 5 X column. the oligonucleotide-toenzyme ratio was very high and thus the enzyme was inhibited. phosphorothioates inhibit the enzyme (149. and noncompetitive inhibitors of DNA polymerases y and 6 (172).

and uptake has been identified (93). However. advances in extraction. For example. 189).and exonucleases. whether the cells are transformed or are virally infected. 196.4 in Vitro Cellular Uptake 6. This would obviously obscure antisense effects on this target. the stability of phosphorothioate oligonucleotides to nucleases is probably a bit less than initially thought. media and sequence. 6.186).1 Nuclease Stability. a variety of labeling techniques have been used. 93. In some cases. . In our laboratories. 197). 3'. Finally.and 5'-terminus as a means of enhancing stability have proved to be unsuccessful.3. uptake of phosphorothioate oligonucleotides into a prokaryote. 188. The pattern of metabolites suggests primarily exonuclease activity. 6. 187). a tritium exchange method that labels a slowly exchanging proton at the C8 position in purines was developed and proved to be quite useful (180). Numerous studies have shown that phosphorothioate oligonucleotidesdistribute broadly in most cells once taken up (93. Again. either uniformly (179) S-labeled or base-labeled phosphorothioates are preferable for pharmacokinetic studies. strategies in which oligonucleotides are modified at only the 3'. Although several studies have suggested that receptor-mediated endocytosis may be a significant mechanism of cellular uptake. however. Cationic lipids and other approaches have been used to enhance uptake of phosphorothioate oligonucleotides in cells that take up little oligonucleotide in vitro (193-195). and length of the oligonucleotide (93). Very recently. I important role in the metabolism of oligonucleotides. Similarly. No obvious correlation between the lineage of cells. there are substantial variations from cell type to cell type. 192). Consequently. It is also influenced by cell type.172. a method that added radioactive methyl groups through S-adenosyl methionine was also successfully used (181). with perhaps modest contributions by endonucleases. however. Thus. 172. and detection methods have resulted in methods that provide excellent pharmacokinetic analyses without radiolabeling (182). nor are the factors that result in differences in uptake of different sequences of oligonucleotide understood.and 5'-modified oligonucleotides with phosphodiester backbones have been shown to be relatively rapidly degraded in cells and after administration to animals (90. the data are not yet compelling enough to conclude that receptor-mediated endocytosis accounts for a significant portion of the up take in most cells (87).although quite stable to various nucleases. Other approaches to enhanced intracellular uptake in vitro have included streptolysin D treatment of cells and the use of dextran sulfate and other liposome formulations as well as physical means such as microinjections (93. given the high concentrations (that inhibited nucleases) of oligonucleotides that were employed in the early studies. Again. phosphorothioate oligonucleotides are degraded slowly by cells in tissue culture with a half-life of 12-24 h and are slowly metabolized in animals (183.185. These are probably less satisfactory than internally labeled compounds because terminal phosphates are rapidly removed by phosphatases and fluorescently labeled oligonucleotides have physicochemical properties that differ from those of the unmodified oligonucleotides. a number of lines of evidence suggest that. Uptake is time and temperature dependent.3 Pharmacokinetic Properties To study the pharmacokinetics of phosphorothioate oligonucleotides. in many cells and tissues. In fact.93. has been reported as has uptake into Schistosoma Mansoni (190. 3'or 5 -32P end-labeled or fluorescently labeled oligonucleotides have been used in either in vitro or in vivo studies. 184). separation. endonucleases play an Phosphorothioate oligonucleotides are taken up by a wide range of cells in vitro (72. are competitive inhibitors of nucleases (67.Oligonucleotide Therapeutics ate oligonucleotides were reported to enhance lipopolysaccharide-stimulated synthesis or tumor necrosis factor (178). The principal met- abolic pathway for oligonucleotides is cleavage by endo. 183. Phosphorothioate oligonucleotides. significant differences in subcellular distribution between various types of cells have been noted. cell-culture conditions. Consequently. Vibrio parahaemoyticus. 191).

Although no direct evidence of base excision or modification has been reported. 205). the brush border membrane. One study of prolonged infusions of a phosphorothioate oligonucleotide to humans has been reported (206).Characteristics of Phosphorothioate Oligonucleotides In Vivo Pharmacokinetics osphorothioate oligonucleotides bind to sem albumin and a-2 macroglobulin. after an radermal dose 3. After intradermal injection in humans. the authors suggested that reabsorption might be mediated by interactions with specific proteins in the brush border membranes. and similar data have been reported by others (174. infusions at a dose of 0. approximately 70% f the dose was absorbed within 4 h and total stemic bioavailability was in excess of 90% 00). at a dose of 6 mgkg to mice. provides a repository r these drugs and prevents rapid renal excreon. kidney. In this study. intact oligomer may be und in urine (174..The data suggested that the oligonucleotides are filtered by the glomerulus. oligonucleotides nistered intravenously a t doses of 15-20 saturate the serum protein binding ca(199). Studies in our labosuggest that in rats. bone marrow.The pharmacokinetic properties observed were consistent with our results as described above. autoradiographic analyses showed drug in renal cortical cells (181). 200. therefore. five patients with leukemia were given 10-day i. Additionally. nucleosides that are degraded by normal metabolic pathways.198). but other tissues display small quantities of drug (200. Because serum protein binding is saturable. and within renal tubular epithelial cells (204). 182). suggesting that there is uptake from the basal side also. 156 h). the oligonucleotide is accumulated in a nonfiltering kidney. the proximal convoluted tubule. a 0-mer phosphorothioate.and endonucleases that result in shorter oligonucleotides and.g. Clearly. Liver. skeletal muscle. The distribution into the kidney has been studied more extensively and drug was shown to be present in Bowman's capsule. Moreover. No evidence of significant penetration of the blood-brain barrier has been reported. then reabsorbed by the proximal convoluted tubule epithelial cells. In a very thorough study. In one study.v. at higher doses.6 mglkg of 14C-ISIS2105. but not fully characterized (174). administration is extremely rapid. The oligonucleotides were internally labeled with 3H-CH3by methylation of an internal deoxycytidine residue using Hhal methylase and S-[3Hladenosyl methionine (181). ultimately. The apwent affinity for albumin is quite low (200pill) and comparable to the low affinity binding observed for a number of drugs (e. 198. from renal medulla. terminal half-life from liver. Phosphorothioates distribute broadly to all peripheral tissues. elimination of drug relatively rapidly from liver compared to that from many other tissues (e. the potential for conjugation reactions and extension of oligonucleotides by these drugs serving as primers for polymerases must be explored in more detail. Blood and plasma clearance is multiexponential.203). Phosphorothioate oligonucleotides are rapand extensively absorbed after p a r e n t e d inistration. 20-nucleotide phosphodiester and phosphorothioate oligonucleotides were administered i. pirin. as would be expected (202). penicillin) (167. The rates of incorporation and clearance from tissues vary as a function of the organ studied. and similarly. in this report.g. Metabolism is mediated by exo. Serum proin binding. 203.v. with liver accumulating drug most rapidly (20% of a dose within 1-2 h) and other tissues accumulating drug more slowly. 62 h. Distribution of phosphorothioate oligonucleotides from blood after absorption or i.v. 168. a larger molecular weight radioactive material was observed in urine. Elimination half-lives reportedly varied from 5. with a terminal elimination half-life from 40 to 60 h in all species except humans. these are theoretical possibilities that may occur. in rats. absorption of ISIS 2105 was similar to that observed in rats (201). Subcutaneous administration to rats and monkeys results in somewhat lower bioavailability and greater distribution to lymph nodes.9 to . In addition. 203). We have reported distribution half-lives of less than 1 h. Clearance of phosphorothioate oligonucleotides is attributed primarily to metabolism (200.05 mg kg-' h-'.. and skin accumulate the highest percentage of a dose. For example. where the terminal elimination half-life may be somewhat longer (201).

5 mgkg was 2.Target reduction in the lung has been demonstrated (214) and these drugs have been shown to distribute broadly to all cell types in the lung after aerosol administration. However. and immunohistochemical approaches have shown that these drugs are localized in endopromal convoluted tubular cells. Doses from 0. whereas that of 0.Clearance from plasma. the study showed that subcellular distribution also varied. Autoradiographic. In addition to the pharmacological effects that have been observed after phosphorothioate oligonucleotides have been administered to animals (and humans). was dose dependent. 2000). Clearly. and cells in the skin and liver (204.06 to 2. were determined.18 and 26. Phosphoro- thioate oligodeoxynucleotides have been shown to be attractive for inhalation delivery to the lung and upper airway (211-213).3 Summary. fluorescent.5. pig.5. Phosphoro- thioate oligonucleotides formulated in a very sim e formulation have been shown to . endothelial. no intact drug was found in urine.0 mgkg were studied and the peak plasma concentrations were shown to increase linearly with dose.~ lcream penetrate normal mouse. 6. In summary. was shown to accumulate throughout the dermis and epidermis after topical administration and to result in a positive trend in the primary endpoint.07 mL min-' kg-'. Nearly 50% of the administered radioactivity was recovered in urine. but most of the radioactivity represented degrades. Overall. do not cross the bloodbrain barrier.71 h. Essentially.5 ~g1mL. and hepatocyte cell population varied and that as doses were increased. unpublished observations. Further. that is. in a phase IIa study in patients with plaque psoriasis.28 mL min-' kg-'. and are eliminated primarily by nuclease metabolism. infusion at a dose of 0. In a more recent study in which the level of intact drug was carefully evaluated using capillary gel electrophoresis. distribute broadly to many peripheral tissues. Plasma clearance of total radioactivity was biexponential with initial and terminal elimination half-lives of 0. the two most recent studies differ from the initial report in several facets. the pharmacokinetics of ISIS 2302. Perhaps more compelling and of more longterm value are studies recently reported showing the distribution of phosphorothioate oligonucleotides in the liver of rats treated intravenously with the drugs at various doses (210). In contrast. however. in ICAM-1 inhibitor.5. with 30% of the radioactivity being intact drug. the behavior of phosphorothioates in the plasma of humans appears to be similar to that in other species. respectively. In fact. not differences between compounds. Metabolites in urine included both higher and lower molecular weight compounds. once daily or . a peak plasma concentration of 295. Urinary recovery of radioactivity was reported to be 30-60% of the total dose.7 days. B71.1 Aerosol Administration. and human skin and to penetrate and accumulate in human psoriatic skin grown on nude mice. ISIS 2302.Oligonucleotide Therapeutics 14. degradation was extensive and intact drug pharmacokinetic models were not presented. Further. see Ref.2 Topical Administration. the most likely explanation is related to the evolution of assay methodology. 215).209). after a 2-h infusion. 208.v.1 mgkg. studies have demonstrated a substantial accumulation of drug throughout the dermis and epidermis and reduction of targets such as ICAM-1. induration (Kruger. with the 2 mg/kg dose having a clearance of 1. these drugs were shown to be well tolerated at doses up to 12 mg/kg (212). This study showed that the kinetics and extent of the accumulation into Kupffer. a number of other lines of evidence show that these drugs enter cells in organs. the distribution changed. In short. pharmacoki- netic studies of several phosphorothioates demonstrate that they are well absorbed from p a r e n t e d sites. a 20-mer phosphorothioate oligodeoxynucleotide. with the 2 mg/kg dose resulting in peak plasma concentrations of intact drug of about 9. Although a number of factors may explain the discrepancies. no intact drug was found in the urine at any time (207).8 ng/mL was observed at the cessation of the infusion. 6. and B72 (for review. Thus. 6. Moreover. when GEM81 (a 25-mer phosphorothioate oligodeoxynucleotide) was administered to humans as a 2-h i. various bone marrow cells.

160. arising from the misinterpretation of an activity as being attributed to an antisense mechanism when it is attributed to non-antisense effects. However. 217). they can be delivered to the lung by aerosol and topically administered. phosphorothioate oligodeoxynucleotides are expected to induce RNase H-mediated cleavage . direct analyses of target protein or RNA. to a large extent. To serve as a substrate for RNase H. Phosphorothioates have also been shown to have effects inconsistent with the antisense mechanism for which they were designed. and induce degradation of RNA by RNase H (28. Although significant progress has been made in developing general rules that help define potentially optimal sites in RNA species. others are attributed to nonspecific interactions with proteins. 6. this remains an empirical process that must be performed for each RNA target and every new chemical class of oligonucleotides. Some of these effects are attributed to sequence or are structure specific. This results in a requirement of at least 15-17 nucleotides to have sufficient affinity to produce biological activity (164).100.Again. Although the similarities between oligonucleotides of different sequences are far greater than the differences. creating oligonucleotides with the highest affinity per nucleotide unit is pharmacologically important. Thus.6. As discussed later. one useful way to think of antisense oligonucleotides is as competitive antagonists for self-complementary regions of the target RNA. Within the phosphorothioate oligodeomucleotide chemical class. phosphorothioate oligodeoxynucleotides are relatively competitively disadvantaged. examples of biological activity have been reported for only three of these mechanisms. Although multiple mechanisms by which an oligonucleotide may terminate the activity of an RNA species to which it binds are possible. in that many oligonucleotides bind to the gp120 protein (80). Antisense oligonucleotides are designed to bind to RNA targets through Watson-Crick hybridization. and progress in achieving acceptable oral bioavailability continues.6 Characteristics of Phosphorothioate Oligonucleotides every other day systemic dosing is feasible.6 Pharmacological Properties 6. dependent on terminating mechanism and influenced by the chemical class of the oligonucleotide. 218). Antisense oligonucleotides have been reported to inhibit RNA splicing. Selection of sites at which optimal antisense activity may be induced in an RNA molecule is complex. Moreover. In this context. These effects are particularly prominent in in vitro tests for antiviral activity because high concentrations of cells. Obviously.1 Molecular Pharmacology. and a comparison of the aMinity of the oligonucleotide to a complementary RNA oligonucleotide is the most sensible comparison. effect translation of mRNA. a sugar moiety in the oligonucleotide that induces a duplex conformation equivalent to that of a DNA-RNA duplex and a charged phosphate are required (38). the potential for confusion. viruses. and oligonucleotides are often coincubated (92. Human immune deficiency virus (HIV) is particularly problematic.of the RNA when bound. Each RNA appears to display unique patterns of sites of sensitivity. additional studies are required before determining whether there are subtle effects of sequence on the pharmacokinetic profile of this class of drugs.0% T . a duplex between RNA and a "DNA-like" oligonucleotide is required. these data simply urge caution and argue for careful dose-response curves. per unit) (216). other factors. Specifically. . is certainly not limited to antiviral or just in vitro tests (219-221). the mechanism that has resulted in the most potent compounds and is best understood is RNase H activation. studies in our " laboratory have shown antisense activity can vary from undetectable to 100% by shifting an oligonucleotide by just a few bases in the RNA target (94. such as overrepresented . and inclusion of appropriate controls before drawing conclusions concerning the mechanisms of action of oligonucleotidebased drugs.Without question. In addition to protein interactions. Because RNA can adopt a variety of secondary structures through Watson-Crick hybridization. in that the affinity per nucleotide unit of oligomer is less than that of RNA (> -2. many chemical approaches that enhance the affinity of an oligonucleotide for RNA result in duplexes that are no longer substrates for RNase H.125).

Again. and behavioral effects in animals with chemical lesions.Again. Additionally. the antisense oligonucleotide inhibited intimal thickening and cyclin-dependent kinase activity. Nevertheless. However. local administration of a phosphorothioate oligonucleotide designed to inhibit N-myc resulted in reduction in N-myc expression and slower growth of a subcutaneously transplanted human tumor in nude mice (225).2 In Vivo Pharmacological Activities. This laboratory also reported the selective reduction of dopamine type 1 receptor and RNA levels with the appropriate oligonucleotide (227). for only a few of these reports have sufficient studies been performed to draw relatively firm conclusions concerning the mechanism of action. and scrambled oligonucleotides were reported (228). direct observations ofAT-1and AT-2 receptor densities in various sites in the brain after administration of different doses of phosphorothioate antisense. Injection of antisense oligonucleotides to synaptosomal-associated protein-25 into the vitreous body of rat embryos reduced the expression of the protein and inhibited neurite elongation by rat cortical neurons (230). Because many of these studies have been reviewed previously. only one dose level was studied. 160. suffice it to say that antisense effects of phosphorothioate oligodeoxynucleotides against a variety of targets are well documented (79. However. whereas a scrambled control did not (229). the effects of the oligonucleotide were suggested to be attributable to a non-antisense mechanism (220). Intraventricular injection of antisense oligonucleotides to neuropeptide-y-yl receptors reduced the density of the receptors and resulted in behavioral signs of anxiety (226). 78). 173. antisense oligonucleotides administered intraventricularly selectively inhibited dopamine type 2 receptor expression. A relatively large number of reports of in vivo activities of phosphorothioate oligonucleotides have now appeared.222). sense. but no effect by a control oligonucleotide. In this study. see Ref. Antisense oligonucleotides administered intraventricularly have been reported to induce a variety of effects in the CNS. Similarly. Given the variability in cellular uptake of oligonucleotides.Oligonucleotide Therapeutics sequences of RNA and unusual structures that may be adopted by oligonucleotides. In studies in rats. a Northern blot analysis showed a significant reduction in c-myb RNA in animals treated with the antisense compound. careful evaluation of dose-response curves and clear demonstration of the antisense mechanism are required before drawing conclusions from i n vitro experiments.6. the variability in potency as a function of binding site in an RNA target and potential non-antisense activities of oligonucleotides. whereas a control oligonucleotide had no effect (224). In a recent study. A phosphorothioate oligonucleotide designed to inhibit c-myb production and applied locally was shown to inhibit intimal accumulation in the rat carotid artery (223). Similar effects were reported for phosphorothioate oligodeoxynucleotides designed to inhibit cyclin-dependent kinases (CDC-2 and CDK-2). Controls included randomized oligonucleotides and the observation that no effects were observed on doparnine type 1 receptor or RNA levels (142-144). an antisense oligonucleotide designed to bind to NMDA-R1 receptor channel RNA inhibited the synthesis of these channels and reduced the volume of focal ischemia produced by occlusion of the middle cerebral artery in rats (226). can contribute to unexpected results (80). numerous well-controlled studies have been reported in which antisense activity was conclusively demonstrated. so much remains to be done before definitive conclusions are possible. I will review in some detail only a few reports that provide sufficient data to support a relatively firm conclusion with regard to mechanism of action. dopamine type 2 receptor RNA levels. intraventricular administration of phosphorothioate antisense oligonucleotide resulted in a decrease in tryptophan hydroxylase levels in the brain. 6. documenting activities after both local and systemic administration (for review. Similar observations were reported in studies on AT-1 angiotensin receptors and tryptophan hydroxylase. Local effects have been reported for phosphorothioate and methylphosphonate oligonucleotides. 140. in rats. In a series of well-controlled studies. Consequently. Aerosol administration to rabbits of an antisense phosphorothioate oligodeoxynucleotide designed to inhibit the production of .

and histamine-induced bronchoconstriction (214). In a similar series of studies. In the second study. A number of studies from our laboratories that directly examined target RNA levels. after either prolonged subcutaneous infusion or intermittent subcutaneous injection (233). In one study.v. Neither a control nor an oligonucleotide complementary to bradykinin B2 receptors reduced adenosine A1 receptors' density. spleen. Again. and it is clearly necessary to include a range of controls and to evaluate effects on target RNA and protein levels and control RNA and protein levels directly. it is difficult to argue with the conclusion today that some . The drug was shown to inhibit the development of leukemic colonies in the mice and to selectively reduce bcr-abl RNA levels in peripheral blood lymphocytes. myb mRNA levels appeared to be selectively reduced (234). there are now a number of studies with sufficient controls and direct observation of target RNA and protein levels to suggest highly specific effects that are difficult to explain by any mechanism other than antisense. house dust mite allergen. As would be expected. inhibited the growth of human melanoma in mice. bone marrow. although the oligonucleotides complementary to bradykin in B2 receptor mRNA reduced the density of these receptors. Monia et al. dose of 10-15 mg/kg. In addition to local and regional effects of antisense oligonucleotides. Similar results with other antisense oligonucleotides were shown in another in uiuo tumor model. a growing number of well-controlled studies have demonstrated systemic effects of phosphorothioate oligodeoxynucleotides. liver. demonstrated highly specific loss of human Craf kinase RNA in human tumor xenografts and antitumor activity that correlated with the loss of RNA (237. Thus. it is possible that the effects on the RNA levels were secondary to effects on the growth of various cell types. at a dose of 1 mglday for 9 days to immunodeficient mice injected with human leukemic cells. lungs. The effects lasted at least 24 h after a dose and a clear dose-response curve was observed with an i. A phosphorothioate oligonucleotide designed to inhibit human PKC-a expression selectively inhibited expression of PKC-a RNA and PKC-a protein in human tumor cell lines implanted subcutaneously in nude mice after intravenous administration (236). selectively inhibited expression of PKC-a RNA in mouse liver without effects on any other isotype. reducing PKC-a RNA levels in the liver by 50% 24 h after a dose (235). Single and chronic daily administration of a phosphorothioate oligonucleotide designed to inhibit mouse protein kinase C-a (PKC-a). target protein levels. Finally. and the endpoint measured. the potency of these effects vary depending on the target. In conclusion. there is a growing body of evidence that phosphorothioate oligonucleotides can induce potent systemic and local effects i n uiuo. as well as the route of administration and the time after a dose when the effect is measured. although it is of obvious importance to interpret in uivo activity data cautiously. Several recent reports further extend the studies of phosphorothioate oligonucleotides as antitumor agents in mice. More important. a single injection of a phosphorothioate oligonucleotide designed to inhibit cAMP-dependent protein kinase type 1 was reported to selectively reduce RNA and protein levels in human tumor xenografts and to reduce tumor growth (239). a phosphorothioate oligonucleotide antisense to the protooncogene myb. the organ. Oligonucleotides to the NF-KBp65 subunit administered intraperitoneally at 40 mg/kg every 3 days slowed tumor growth in mice transgenic for the human T-cell leukemia viruses (232). Expression of interleukin 1 in mice was inhibited by systemic administration of antisense oligonucleotides (231).p. In these studies.b t 6 Characteristics of Phosphorothioate Oligonucleotides adenosine A1 receptor has been reported to reduce receptor numbers in the airway smooth muscle and to inhibit adenosine. effects on RNA and protein levels were highly specific and observed at doses lower than 6 mg/kg. 238). A large number of control oligonucleotides failed to show activity. However. and brain (96). and pharmacological effects using a wide range of control oligonucleotides and examination of the effects on closely re- lated isotypes have been completed. a phosphorothioate oligonucleotide directed to inhibition of the bcr-abl oncogene was administered i.

The duration of effect after a month of dosing was in excess of 6 months.1 lSlS 2302. see Ref. Further.Oligonucleotide Therapeutics effects have been observed in animals that are most likely primarily ascribed to an antisense mechanism. In the initial placebo-controlled study in steroid-dependent patients. It has been demonstrated to reduce ICAM-1 levels in a number of organs in a variety of animal models and to result in potent anti-inflammatory effects (for review. Europe. The 0. In approximately one-half the patients. study was equivalent to that used for the rheumatoid arthritis study.7. ISIS 2302 has been evaluated in a number of clinical trials in patients with inflammatory diseases (for review.053) in favor of increasing ISIS 2302 concentrations with regard to the primary endpoint. 241). In a subsequent 300 patient trial. In this study. The study showed that ISIS 2302 achieved very high dermal and epidermal concentrations that increased as the concentration in the cream was increased. 1.0. see Ref. with no increase in toxicity.v. 247). using a variety of dose schedules at doses as high as 30 mg kg-' day-' and the drug was well tolerated. and 2. 242). Additional phase I11 trials at higher doses are in progress.v. 6. again the drug was shown to be extremely well tolerated.1 shows the results reported for the drug in these patients with advanced chemoresistant malignancies. This drug has been evaluated as a single agent in a variety of solid tumors.2 lSlS 3521. A population pharmacokinetic study " demonstrated that heavier patients and women achieved greater exposures to the drug and that in the patients in the upper two quartiles of drug exposure. and topical routes as a potential therapeutic for patients with psoriasis. Additionally. ISIS 2302 has been evaluated most extensively in Crohn's disease. Again. serial colonoscopic biopsies demonstrated reduction of ICAM-1 protein in the small bowels of treated patients (245). Table 5.5 and 2. Further.v. all doses were well tolerated. the dosing for the i. The most advanced experience is a phase 1/11 study in 53 patients with non-small cell carcinoma of the lung in combination with carboplatin and taxol. has been shown to be safe and effective in the treatment of this disease after intravitreal injection and has been approved for sale by regulatory agencies in the United States.0 mg/kg i. ISIS 3521 has also been evaluated in combination in a number of chemotherapeutic agents in a number of solid tumors.7. the drug was applied topically in a simple cream formulation once a day for 3 months. ISIS 3521 is a phosphoro- thioate oligodeoxynucleotide inhibitor of PKCa (for review.7 Clinical Activities Vitravene. Additional studies with topical ISIS 2302 are in progress. 240). The patients were followed for a total of 6 months. see Ref. 6. ISIS 2302 is a phosphoro- thioate oligonucleotide inhibitor of human ICAM-1. At the end of the treatment Reriod an approximately 30% reduction on psoriasis activity index score (a measure of psoriasis activity) was reported (244). with 2% and 4% creams given dermal concentrations similar to those at which pharmacological effects were observed in animals. but only 17 patients were evaluated. The drug was evaluated in a randomized placebocontrolled phase IIa trial in 43 patients with moderate to severe rheumatoid arthritis (243). induration. several concentrations of ISIS 2302 were evaluated in patients with moderate plaque psoriasis. In this trial the effects of 0. there was a trend (P= 0. In this study.5. ISIS 2302 resulted in a substantial reduction in symptoms and a dramatic reduction in steroid use and reduction of ICAM-1 levels in the small intestines of patients treated and evaluated with serial colonoscopies. Again. compared to placebo.0 mgkg doses resulted in prolonged improvement in rheumatoid arthritis activity. see Ref. three times weekly for 1 month were compared to placebo. ISIS 3521 was associated with a substan- . 6. skin biopsies were taken. ISIS 2302 was evaluated both by i. a phosphorothioate oligonucleotide designed to inhibit cytomegalovirus-induced retinitis. in which patients were treated with placebo or 2 mgtkg every other day for 2 or 4 weeks. All doses were extremely well tolerated. the drug was very well tolerated. and South America (for review. the drug produced a statistically significant in: crease in complete remissions compared to placebo (246).

The effects seemed to be dependent on the phosphorothioate backbone and independent of sequence or 2'-modification. In a study in which murine B-lymphocytes were treated with phosphodiester oligonucleotides. Also. see Ref. In our laboratory. Based on these data. Furthermore. given that little in vitro testing has been performed and no data from long-term studies of oligonucleotides are available.8. 6.4 lSlS 2503. As a general rule. empirical data must be generated. ISIS 2503 is a phosphoro- thioate oligodeoxynucleotide designed to inhibit Ha-ras.8. ISIS 2302. see Ref. G3139 is also being studied in combination with chemotherapeutic agents. with a few exceptions. we have ported to be mitogenic for lymphocytes in vitro (259).6 Characteristics of Phosphorothioate Oligonucleotides tial improvement in survival (16 months) compared to that of historical controls (248).8 Toxicology 6. As with any new chem- evaluated the toxicities of scores of phosphorothioate oligodeoxynucleotides in a significant number of cell lines in tissue culture.1).' tion was induced by oligonucleotides with unmethylated CpG dinucleotides (263). It has been shown that phosphorothioate oligonucleotides stimulate B-lymphocyte proliferation in a mouse splenocyte preparation (analogous to bacterial DNA). Additionally. bacterial DNA species have been re- ical class of therapeutic agents. ISIS . oligodeoxynucleotides (30-45 nucleotides in length) were reported to induce interferons and enhance natural killer cell activity (260). It is currently completing phase I1 single-agent and combination studies in patients with a variety of solid malignancies (for review. ISIS 2105. Collectively. We performed mutagenicity studies on five phosphorothioate oligonucleotides. nor is it the case with regard to in vivo immune stimulation (see below). 6.1 In Vitro.5 lSlS 51 32. the oligonucleotides that displayed natural killer cell (NK)-stimulating activity contained specific palindromic sequences and tended to be guanosine rich. In this study. these observations indicate that nucleic acids may have broad immunostimulatory activity. G3139 is a phosphorothioate oligodeoxynucleotide designed to inhibit BCL. 6.7. In phase I studies.. concerns about genotoxicity cannot be dismissed. 247). This clearly is not the case with regard to release of IL-la from keratinocytes (262). In a Phase I study with decarbazine in patients with malignant melanoma. 6. Clearly. 192). This has been extrapolated to suggest that the CpG motif may be required for immune stimulation of oligonucleotide analogs such as phosphorothioates. a large phase I11 program is in progress. both human keratinocytes and an in vitro model of human skin released interleukin 1-a when treated with 250 @ to 1 mM of phosphorothioate oligonucleotides. given the limitations in our understanding about the basic mechanisms that might be involved. levels in peripheral blood cells were reduced by (33139 (249). In the latter study. We also have evidence of enhanced cytokine release by immunocompetent cells when exposed to phosphorothioates in vitro (262). 100 Polynucleotides and other polyanions have been shown to cause release of cytokines (8).This drug is being evaluated in a number of other malignant diseases and in a number of other combinations. BCL.7.7.2 Genotoxicity. B-cell activa. and the response may underlie the observations of lymphoid hyperplasia in the spleen and lymph nodes of rodents caused by repeated administration of these compounds (see below) (261). It has demonstrated single agent activity in nonHodgkin's lymphoma. no significant cytotoxicity is induced at concentrations below 100 pM oligonucleotide. this drug was shown to reduce C-raf kinase levels in peripheral blood cells and displayed activity in patients with ovarian cancer (251). It too is completing phase I1 evaluation (for review. 6. no significant effect on macromolecular synthesis is observed at concentrations below (188. a 21% response rate in 14 patients was reported (250). In the non-Hodgkin's lymphoma trial. 247). It has been reported to be well tolerated and preliminary suggestions of activity have been reported (see Table 5.3 C3139. ISIS 5231 is a phosphoro- thioate oligodeoxynucleotide designed to inhibit C-raf kinase.

infusion tiw. One with skin recurrence 15 months after start of treatment Stable disease for 6 months 1 months Partial response: time to progression (TTP):1 Marker only disease with 75% J in CA-125. 3 weeks out of 4 Colon Renal Ovarian Pancreatic Renal Sarcoma Pancreatic Colon Mesothelioma Melanoma Lymphoma 41. TTP: 10 months Stable disease for 10 months Stable disease for 9 months Stable disease for 10 cycles Stable disease for 9 cycles Stable disease for 8 cycles Stable disease for 6 cycles Melanoma: stable disease for 7 cycles Complete response: 1/17 patients 1 1 stable disease (including 2 minor responses) 6/17 patients with improved lymphoma symptoms 7/16 patients with & Bcl-2 protein in PBMC.Table 5.i.1 Single-Agent Antitumor Activity in Clinical Trials of Antisense Oligonucleotides OligoITarget ISIS 3521PKC-a! Schedule 2 h i. or lymph node Ref.i.i. 1 1 2 day c. infusion 32 48 .v.v.. bone marrow. 2 weeks out of 3 31 G3139lbcl-2 24 h c. Concurrent J C-raf expression in peripheral blood mononuclear cells 97% decrease in CA-125 with stable evaluable disease.42 40 ISIS 2503Ma-ras 14 day c.c. TTP: 7 months 80% J in CA-125 with stable measurable disease Stable disease for 7 months. 31% decrease of CEA Concurrent & C-raf expression in peripheral blood mononuclear cells Stable disease for 9 months. infusion tiw. 3 weeks out of 4 21 day c. TTP: 7 months Marker only disease with 40% J in CA-125. repeated weekly 14-day s.. 3 weeks out of 4 ISIS 5132lC-raf Non-small cell lung Ovarian 12 2 h i.i..v.v. 3 weeks out of 4 Tumor Lymphoma Efficacy Results Complete response in 2 of 2 patients with low grade disease: One continuing 30+ months after start of treatment..v.v.

and found them to be nonmutagenic at all concentrations studied (264. cytokine release. Consequently.3.2 Inhibition of Clotting. The other factors contributing to cardivascular collapse are not yet fully elucidated. A. than that of other potential mechanisms (e. The threshold concentration for activation of complement in the monkey is 40-50 pg/mL phosphorothioate oligodeoxynucleotide and. However.1 Acute and Transient Toxicities 6. ISIS 14803.8.1 Complement Activation. attributed to activation of the alternate pathway. it appears that monkeys are substantially more sensitive than humans to complement activation. phosphorothioate oligodeoxynucleotides activate complement in a concentration-dependent fashion. Moreover. oxidation of the phosphorothioate backbone to the natural phosphodiester structure would also yield nonmutagenic (and probably nontoxic) metabolites. For most viruses. The mechanism of complement activation is currently believed to be attributable to an interaction with factor H (169). However. which presumably would be nongenotoxic. alteration of the activity of growth factors.7. viruses have evolved specialized enzyme-mediated mechanisms to achieve integration. new analogs that deviate significantly more from natural DNA would be even less likely to be integrated. it is possible that phosphorothioate bonds could be hydrolyzed slowly.6 Characteristics of Phosphorothioate Oligonucleotides 5132.8. it would seem that the risk of genotoxicity resulting from genomic integration is no greater. and it is unlikely that enzymes such as integrases. monkey serum demonstrates a dramatic difference. These effects can be reduced by chemical modifications and formulations that reduce plasma protein binding (169).Presumably.of human serum to inhibition of complement activators. 169). but is absolutely related to peak plasma concentration (269).3. manifested as an increase in activated partial thromboplastin . variable but potentially dangerous levels of complement activation are observed. Studies in humans have avoided peak plasma concentrations that would induce complement. 6. gyrases.1. and Vitravene. However. it has been suggested that dosing monkeys that are restrained may result in exacerbating cardiovascular events. metabolism of phosphorothioate oligodeoxynucleotides by base excision would release normal bases. nonspecific effects on membranes that might trigger arachidonic acid release. once the threshold is reached. it is likely to be extremely rare.g. However.3. Finally. or inappropriate intracellular signaling). second concern that has been raised about possible genotoxicity is the risk that oligonucleotides might be degraded to toxic or carcinogenic metabolites. and probably less. Although integration of an oligonucleotide into the genome is conceivable. In monkey serum. phosphorothioate oligodeoxynucleotides induce a transient.3 In Vivo 6. complement activation is actually inhibited at higher concentrations of oligonucleotide (169). 265). It human serum. One possibility is that an oligonucleotide analog could be integrated into the genome and produce mutagenic events.8. but complement activation is necessary but not sufficient to produce the observed cardiovascular effects (for review. releasing nucleoside phosphorothioates that presumably would be rapidly oxidized to natural (nontoxic) nucleoside phosphates. preliminary studies in our laboratory have shown that phosphorothioate oligodeoxynucleotides are generally poor substrates for DNA polymerases. Two mechanisms of genotoxicity that may be unique to oligonucleotides have been considered. is relatively insensitive to sequence. In all species studies. see Ref. Rapid in- fusion of phosphorothioate oligodeoxynucleotides to nonhuman primates can result in cardiovasulcar collapse (266-268).. of course. apparently self-limited. 6. These hemodynamic effects are associated with complement activation. A comparison of the effects on comdement activation in hu* man vs. peak plasma concentration-related inhibition of clotting. This is thought to be attributable to the sensitivity . Activation of the complement cascade.8. and topoisomerases (that have obligate DNA cleavage as intermediate steps in their enzymatic processes) will accept these compounds as substrates. viral DNA integration is itself a rare event and. oligonucleotides with modified bases andlor backbones may pose different risks. Similarly.

subcutaneously. occasional increases in liver function enyzmes are noted. immune stimulation is far less ~rominentthan that in the rodent.8.8. lymphoid hyperplasia. Intravenous doses have ranged from 0. At Isis. appear to be self-limited.doses .e. we have had the opportunity to study approximately 10 phosphorothioate oligonucleotides in humans. in the rodent but not the monkey. In contrast. an intravitreally administered drug. on occasion. 265). In the ro- dent. Modifications of the base. - 6. Occasionally. Immune stimulation in the rodent is a property common to all phosphorothioate oligodeoxynucleotides. and have not resulted in bleeding diatheses in either animals or humans. a 20-nucleotide phosphorothioate antisense inhibitor of ICAM-1. intradermally. Subchronic administration of doses as low as 10 mg kg-' day-' in rodents results in splenomegaly. although the effects are directly proportional to the length of the phosphorothioate oligodeoxynucleotide (169). Similarly. and the backbone can reduce immune stimulation in the rodent. but these occur at high doses and are not associated with histopathological changes (169).283). For example. The effects on clotting are transient. doses necessary to produce immune stimulation have not been identified. In fact. Most patients have been treated with multiple .8. In primates.8. Also. the A sugar. Similar data were observed with all the phosphorothioate oligodeoxynucleotides administered on this . 169). see Ref.3. ISIS 5132.Increases in aPTT are minimally affected by sequence. and X.Oligonucleotide Therapeutics time (aPTT) (270-273). Rodent immune stimulatory motifs include palindromic sequences and CG motifs (282. an inhibitor of C-raf kinase was dosed from 0.4. and many have been treated for prolonged periods (i. At least one factor contributing to these species differences is that the optimal rodent immune stimulator sequence is simple and immune stimulation is induced by all sequences. For example. Ma.0 m&g and no meaningful increases in complementsplit products were observed. With 2-h infusions at a dose of 2 mgkg. 6. 280.4 Human Safety. in clinical trials. The mechanism of aPTT increase appears to be an interaction with the intrinsic tenase complex (274. In the monkey. and topically.000 doses. Effects on clotting can be ameliorated by chemical modifications such as 2'-0-methoxyethyl substitution (169). intravenously.2 Subchronic Toxicities 6. perhaps associated with complement activation (for review. a PKCa inhibitor gave equivalent data and both of these drugs were studied in patients with various malignant diseases. We have studied antisense drugs administered intravitreally.in vivo immune stimulation. these in vitro studies are reasonably predictive of the relative potencies of phosphorothioates in moving . in the primate. the optimal sequences are different and more complex (284) and the response is much less promiscuous with regard to sequence.. 275).8. we have studied more than 2000 patients treated with more than 70. Other toxicities noted in animals are mild and occur infrequently at therapeutic doses. is commercially available around the world. 5'methycytosine and 2'-0-methoxyethyl containing oligonucleotides display substantially reduced potency for immune stimulation in the rodent (169).5 m&g every other day to as much as 30 mg kg-' dayp' as a continuous infusion. but potency varies substantially as a function of sequence. despite evaluation of numerous oligonucleotides at a variety of doses and schedules (169). the most prominent toxicity is immune stimulation. 6.5 to 6. and diffuse multiorgan mixed mononuclear cell infiltrates (276-279).3. ISIS 3521. This effect is also frequently a confounding variable that must be evaluated with regard to the mechanism of action of pharmacological effects (for review.1 Immune Stimulation. Vitravene.1 Complement Activation. see Ref. These effects are reminiscent of stimulator effects observed on isolated splenocytes (261. After systemic administration.2.2 Other Toxicities. 281).2. in the monkey transient thrombocytopenia is observed. 6. no increases in complement-split products were observed in more than 300 patients with inflammatory diseases treated with ISIS 2302. multiple months). This interaction is complex and involves effects on multiple clotting factors including factors VIIIa. single-cell hepatolyte neurons is observed and this has been related to immunostimulation (265).3.

and interference with clotting limit the dosing and may limit the use of these agents in the treatment of some diseases. phosphorothioates must be dosed every other day. or other organ performance been observed (285). When phosphorothioate oligodeoxynucleotides have been administered by continuous i.4. with chronic subcutaneous administration. and were not associated with bone marrow effects (253. when these drugs were given as 24-h infusions.9 Conclusions Phosphorothioate oligonucleotides have perhaps outperformed many expectations.. IL-IR. and TNFa were observed. are not significantly orally bioavailable. 24 mg kg-' dayp1) as a continuous i. Pharmacodynamically. in many experiments.2 Cytokines.These effects appear to be more prominent after the first dose.v. 285 and 286).4. local lyrnphoadenopathy has been dose limiting (285). significant increases in IL-6. Most of the time. self-limited increases in activated partial thromboplastin time (aPTT).. These effects were more frequent during the first course of therapy. 7 THE MEDICINAL CHEMISTRY OF OLIGONUCLEOTIDES 7. the higher end of the pharmacological dose-response curve is lost. see Refs. ISIS 5132. they have relatively low affinity per nucleotide unit. All phosphorothioate oligodeoxynucleotides studied in normal volunteers and in patients have resulted in transient. 287.8. 6.3 Coagulation. may display dose-dependent pharmacokinetics.4 Platelet Effects.At present. Thus. once these often correlated with flulike syndromes (285). Pharmacokinetically. Nevertheless. 289). In none of the patients treated did we observe any evidence of bleeding. these compounds appear to display satisfactory therapeutic indices for many indications. and ISIS 2503. Despite these increases in complement-split products. Further. the intense focus on the medicinal chemistry of oligonucleotides dates perhaps to no more than 5 years before the preparation of this chapter.8. and . only a small number of patients experienced mild fevers and myalgias a t these very high doses.4. When given as 21-day antitumor infusions at doses as high as 10 mg kg-' day-'.8. 6. ISIS 3521. but probably involve margination. Although the synthesis of modified nucleic acids has been a subject of interest for some time. 6. do not cross the blood-brain barrier.v. At higher concentrations. 6. the precise roles of each of the cytokines and complement activation in the clinical signs and symptoms (myalgia and fever) at high doses are not defined. the antisense mechanism has been directly demonstrated as the hoped-for selectivity. This means that longer oligonucleotides are required for biological activity and that invasion of many RNA structures may not be possible. a PKCa inhibitor. In contrast. The effects are dose and peak plasma concentration dependent (285). At very high doses (i. were not obviously dose-related. significant increases in complement-split products were observed at doses of 18 mg kg-' day-' and greater (285). kidney. these compounds inhibit RNase H as well. no increases in complement-split products were observed (257. infusion in patients with malignant diseases. phosphorothioates clearly have significant limits.8. They display attractive parenteral pharmacokinetic properties. 6. .1 Introduction The core of any rational drug discovery program is medicinal chemistry. Toxicologically. The mecha- nisms for these effects are not clear.4. activation of complement. Further. a Ha-ras inhibitor. a C-raf kinase inhibitor. Three phosphorothioate oligodeoxynucleotides.e. have been thoroughly characterized in patients with malignant diseases and dosed with long-term infusions. They have produced potent systemic effects in a number of animal models and. infusion.5 Other Toxicities. platelet counts returned to normal while dosing was continued and no bleeding was observed.7 The Medicinal Chemistry of Oligonucleotides schedule in a variety of patients (for review. clearly the release of cytokines. At no dose studied have significant effects on liver. transient thrombocytoplastin has been observed in a few patients (285). so the effects on aPTT have not proved to be a problem in the clinic and humans appear to behave similarly to other animals with regard to this side effect. 288).

values approximately equal to DNA with T that of natural base pairs (296). in Further.6"C/modification).1 Pyrimidine Modifications. phenothiazine. sugar. the occasional mispairing with G and the potively large number of modified pyrimidines tential that the oligonucleotide might degrade have been synthesized and are now incorporated into oligonucleotidesand evaluated. phenoxazine.298). Binding studIn contrast. termed phenoxazine. value of +5. = 1. The subjects of medicinal chemical programs include approaches to create enhanced affinity and more selective affmity for RNA or duplex structures. another tricyclic cytosine analog This selectivity is a result of weaker hydrogen bonding and increased steric bulk of the in which an aminoethoxy moiety is attached to 2-thiocarbonyl group (293. and phosphate moieties of oligonucleotides and oligonucleotide conjugates have been reported. unpublished nucleotides.2. resulting from a highly preorganized RNA-like C3'actions. was found to hybridize well to DNA and. Although the stability of duplexes may be enhanced by incorporating 7.and support RNase H ac291). Consequently. oligoribonucleotides with 2'-0-methyl-2-thiouridine shown to hybridize with guanine and to enexhibited a AT. Phenoxazine has been resubstituent). Oligonucleotides with this modported to increase the cellular permeation of ification also exhibit better hybridization disoligonucleotides and alter RNase H cleavage crimination for the wobble U-G base pair forsites (301). cellular uptake and distribution.Oligonucleotide Therapeutics have have been reported. the rigid phenoxazine scaffold. the G-clamp. In a different study. Tetrafluorophenoxazine is less selecendo conformation (attributed to the combitive in hybridization because it also hybridizes nation of 2-thio modification and 2'-0-Me with adenine (300). Oligonucleotides containing and C6. 1995). Modifications in the base. C5. and in vivo tissue distribution. the scope of medicinal chemistry has recently expanded enormously. including ha7. However. metabolism. and tetrafluorophenoxazine have been synthesized fact.4). The 5-heteroarylpyrimidines were also the analogs that displayed interesting properties is incorporated here. The and release toxic nucleosides analogs cause principal sites of modification are C2. These and other nucleoside analogs 5-propynylpyrimidine modifications have been have recently been thoroughly reviewed (290. an hancement of 2-FCImodification. concern (290). although the biological data to support conclusions about synthetic strategies are only beginning to emerge. Additionally. C4. these base modifications were results.5"C/modification hance thermal stability by extended stacking interactions and increased hydrophobic interwhen hybridized against RNA (2921. enhanced nuclease stability. When incorporated into oligo1. several modifications in the ies demonstrated that a single incorporation 4-position that have interesting properties of the "G-clamp" enhanced the binding affin- Consequently.5°C/modification(Swayze et al. a very brief summary of tivity. Inasmuch as the C2 shown to increase the stability of duplexes position is involved in Watson-Crick hybrid(218. 5. However. oligonucleotides containing C2 alkylported for the tricyclic 2'-deoxycytidine aqamodified pyrimidines have shown unattraclogs. the ability to cleave nucleic acid targets. value of and evaluated. a fluorescent base has been incorporated into oligonucleotides and shown to enhance DNADNA duplex stability (297). A large number of modifications at the C5 position have also been reported. A bicyclic and an N4-methoxy analog of cytosine were shown to hybridize with both purine bases in . .2 Heterocycle Modifications logenated nucleosides. depending oligonucleotide containing Pthiothymidine on the positioning of the modified bases (299).. 294) (Fig. and clearance. A rela5-halogenated uracil-containing nucleosides. These helix-stabilizing properties have been shown to be further improved with mation compared to the normal U-A base pair. shown to enhance the duplex stability (AT. A more dramatic influence was reization. 4-Thiopyrimidine~ been incorporated into oligonucleotides with no significant negative effect on hybridization (295). recent studies have shown destabilization in the normal U-A base pair formation and stabilization of the wobble U-G base pair for Cthiouridine (294). even better to RNA with a AT. exhibiting an entive hybridization properties.

The tethered amino group may act as a hydrogen bond donor and interact with the 0 6 of guanine on the Hoogsteen face of the base pair. Thus the increased affinity of a G-clamp is presumably attributed to the combination of extended base stacking and an additional specific hydrogen bond.4. Selected pyrimidine modifications. whereas the G-clamp oligonucleotide maintained similar activity (303). This is the highest affinity enhancement attained so far with a single modification. As expected. Importantly. The propynyl oligonucleotide was four times less active than the control oligonucleotide.. the gain in helical stability does not compromise the specificity of the oligonucleotide for complementary RNA.. G clamp ity of a model oligonucleotide to its complementary target DNA or RNA with a AT. There is no structural proof yet available for this hypothesis. The RNase H activation properties of eleven 5-propynyl modifications compared to a single Gclamp substitution have been assessed in another oligonucleotide sequence in terms of percentage cleavage of complementary RNA with nuclear extracts of HeLa cells. modifications in the C6 position of pyrimidines are highly duplex destabi- .7 The Medicinal Chemistry of Oligonucleotides 145 0 I 0 2-thio-T 2-thio-2'-OMe-U I OMe 0 I 0 4-thio-dU 4-thio-U I OH I I 0 I 0 I 0 I Phenoxazine dC Tetrafluoro phenoxazine dC Figure 5. value of up to 18°C relative to that of 5-methyl cytosine (302).

However. C2 modifications have usually resulted in destabilization. coli RNase H activity by 2.2. oligonucleotides containing N2-imidazolylpropylguanine and N2-imidazolylpropyl-2aminoadenine moieties showed a remarkable enhancement of binding affinity when hybridized to complementary DNA (306). N1 modifications of the purine moiety have resulted in destabilization of the duplex (305). However.. 7. value by 1-2°C per modification. coli RNase H-induced degradation of RNA targets (304). This modification reduced E. 5.. when incorporated into oligonucleotides. they usually have resulted in destabilization of duplexes.5 times. Similarly. possibly because of increased steric crowding in the minor groove (149). Generally.146 Oligonucleotide Therapeutics Figure 5. This modification showed . Oligonucleotides with an aminopropyl group placed at the N2-position of 2'-deoxyguanosine have also been studied. there are a few exceptions where a purine modification had a stabilizing effect.5.2 Purine Modifications. Selected purine modifications. lizing (304).Oligonucleotides containing 6-aza pyrimidines have been shown not only to reduce the T. but also to enhance the nuclease stability of oligonucleotides and to support E. A brief summary of some of these analogs is discussed below (Fig. Although nu- merous purine analogs have been synthesized.5).

7-Deazaas the principal nuclease in serum is a 3'-exoguanine was similarly destabilizing. 7-iodo-. 2-6-Diaminoputargeting the 3'-UTR of murine C-raf mRNA. Inasmuch atively hydrophobic nature (311). A convenient method substituted bases have yielded improved nuof synthesizing this modification has been reclease resistance when incorporated in oliported from our laboratories (309).3. but when nuclease. the 7-iodo 7-deamodifications include 3'-conjugates (305. Of the gonucleotides.No. coli RNase H activity (149). some C8in the oligonucleotides. In many cases. of the base. binding affinity dramatically (AT. This increase in deoxyadenosine exhibited a two. and 7-cyano-7-deaza-2-amino-2'-deoxyadeno. against complementary RNA for enhanced binding properties against both single substitutions and smaller increases per DNA and RNA targets and was used for conjugating other functionalities (307). Internal modifications of base. The increase in T. No.1 Nuclease Stability.6). the data supporting some of the claims and other antisense-related properties.323). 1991. 1993.Thus. N6are limited and generalizations are not possible imidazolylpropyl-adenine-modified oligodebased on the data presently available (Fig. . Inosine has been shown to have 7.3'-3' linked modification. reasons for increased binding affinity nor the When the 7-propyne and 7-iodo nucleosides were incorporated into antisense sequences sequence dependency of the binding affinity enhancement was explored. destabilizing (290). oxynucleotides enhanced E. these nuclease-resistant modifications and conjugates require sines showed increases of 3-4°C per modifica- . but because it ifications have been reported to enhance the can air and stack with all four normal DNA stability of oligonucleotides in serum (305). (312). only the 3-deaza adenosine analog has been shown to have no negative effect on hybrid7. 2687). Such nuclease-resistant modifications. sugar. 3201. on occasion. Interestingly. 4579. as a general rule. Incorporation of 7-deazainosine serum and. (2) increased stacking interacoligonucleotides (327). and 3'-loop oligonucletion. zaguanine residue was incorporated into olicationic modifications (316-318). L-nucleotides and analogs nine) (312).3151. introduction of more and backbone have also been reported to enthan one modification in a nucleobase may hance nuclease stability at or near the modicompensate for destabilizing effects of some fied nucleoside (305). and conjugated oligonucleotide tends to be greater this was considered to be attributed to its relas bulkier substituents are added. = 10. zwitterionic gonucleotides and shown to enhance the modifications in the 2'-position (319. the stability of a into oligonucleotides was destabilizing. godeoxynucleotides with 7-propynyl-. a-anomeric oligonucleotides (3241.to threefold binding affinity is independent of either the increase in potency over that of unmodified presence or the absence of any 2'-modification controls in vitro (313). Numerous 3'-modlittle effect on duplex stability.3 Oligonucleotide Conjugates ization. In the incorporation for multiple substitutions relapreceding instances. 5. bases.7 The Medicinal Chemistry of Oligonucleotides tion in T. it enhanced hybridizations lization. it is not surprising that 5'-modifica8-aza-7-deazaguanine was incorporated into tions have resulted in significantly less stabioligonucleotides. and (3) favorable pK. value was at(322. it behaves as a universal base and creBoth neutral and charged substituents have ates an ambiguous position in an oligonucleobeen reported to stabilize oligonucleotides in tide (310). Agrawal. Oliotides (Khan. rine has been reported to enhance hybridizathe sequences with three and four substitution by approximately 1°C per modification tions of the 7-propyne-7-deaza-2-amino-2'when paired with T (308). In contrast. tributed to (1)the hydrophobic nature of the 3'-3' terminal linkages (325. but seem to be somewhat 3-position-substitutedbases reported to date.326). neither the structural tive to those of unmodified control sequences. Although conjugation of various functionalModifications at the C6 and C7 positions ities to oligonucleotides has been reported have likewise resulted in only a few interest(314) to achieve a number of important objecing bases from the viewpoint of hybridization tives.O"CI geometrically altered linkages such as 2'modification compared to that of 7-deazagua5' linkages (321).

there may be a loss in binding affinity to the target RNA because of these changes with geometrically altered linkages. cellular association. However. kid- ney. Furthermore. with perhaps the most compelling data relating to phosphorothioate oligonucleotides.Oligonucleotide Therapeutics Cholesterol at the 8-end H Figure 5. In fact. even a 5-nucleotide-long phosphodiester gap in the middle of a phosphorothioate oligonucleotide resulted in sufficient loss of nuclease resistance to cause complete loss of pharmacological activity (164).3.Moreover.and 5'-termini of a phosphodiester oligonucleotide did not markedly enhance the nuclease stability of the parent compound in mice (90). or plasma (166). neither a 5'-cholesterol nor 5'-C18 amine conjugate altered the metabolic rate of a phosphorothioate oligodeoxynucleotide in liver. . a clear objective has been to improve cellular uptake of oligonucleotides (72. a phosphorothioate oligonucleotide with a 3'hairpin loop was reported to be more stable than its parent in rats (328). 3'-modification of a phosphorothioate oligonucleotide was reported to enhance its stability in mice relative to that of the parent phosphorothioate (329). Thus. cholesteryl. and others) modified oligonucleotides have been reported to improve nuclease stability. 7.6. Selected 5'-oligonucleotide conjugates. 264). and binding affinity (328). Inasmuch as the mechanisms of cellular uptake of oligonucleotides are still very poorly understood.2 Enhanced Cellular Uptake. The demonstration that modifications may induce nuclease stability " sufficient to enhance activity in cells in tissue culture and in animals has proved to be much more complicated because of the presence of 5'-exonucleases and -endonucleases. In our laboratory. blocking the 3'. In mice. 3'-modifications and internal point modifications have not provided sufficient nuclease stability to demonstrate pharmacological activity in cells (183). the medicinal chemistry approaches have been largely empirical and based on many unproven assumptions. Thiono triester (adamantyl. 3'-modifications may enhance the stability of the relatively stable phosphorothioates sufficiently to be of value. additional synthetic steps and modifications to the oligonucleotide synthesis protocol or monomer synthesis. Although oligonucleotides have been shown to be taken up by a number of cell lines in tissue culture. In addition.

Both compounds increased the fraction of an i. It also seems likely that the effects of various conjugates on cellular uptake may be affected by the cell type and target studied. nearly three. may not affect the factors required for enzyme activity. In addition. 7. albeit in low efficiencies (333). especially at the oligonucleotide termini. the only significant change in the toxicity profile was a slight increase in effects on serum transamineses and histopathological changes. relatively otide conjugates were recently reported to act as artificial ribonucleases. studies in this area have not been systematic and.3. To date. Additionally. polycyclic aromatic chromophores such as phenazine (PZN) and dipyridophenazine (DPPZ) have been evaluated in E. More recently. and psoralens have been used extensively to effect a crosslinking of oligonucleotide and the target few studies have been reported in vivo. The effects of these modifications on cellular uptake have been assessed using fluorescent. and simple alkyl chains have been conjugated to oligonucleotides at various sites in the oligonucleotide. liver. the conjugation of lipophilic substituents to enhance membrane permeability has been a subject of considerable interest. Oligonucle- RNA. polycationic substitutions and various groups designed to bind to cellular carrier systems have been synthesized. porphyrins. there is precious little information about the changes in physicochemical properties of oligonucleotides actually effected by specific lipid conjugates. Interestingly. A cholesterol modification was reported to be more effective than the other substituents at enhancing uptake.338). Acridine-conjugated oligonucleotides targeted to P-globin mRNA were shown to be better inhibitors of P-globin synthesis than the unmodified oligonucleotides (335) in a microinjected Xenopus oocytes assay as a result of RNase H activity. lanthanides and complexes thereof have been reported to cleave RNA by way of a hydrolytic pathway.4 RNase H Activity. For example. coli RNase H cleavage of the target strand when ' conjugated to oligonucleotides (337. indicative of . the data reported to date are insufficient to draw firm conclusions about the value of such approaches or structure-activity relationships (SARs) (332).7 The Medicinal Chemistry of Oligonucleotides Because phosphodiester and phosphorothioate oligonucleotides are hydrophilic. The activities of short alkyl chains. The cholesterol conjugate resulted in more than 80% dose accumulation in the liver. Oligonucleotide conjugates may play a distinct role in improving RNase H activity because conjugation of small molecules. Phospholipids. very few systematic studies have been performed. or kidney (166). cholic acid.5 In Vivo Effects. or radiolabeled. adamantine. 7.3. cholesterol. Recently. Unfortunately. Although many compounds have been synthesized. Neither conjugate enhanced stability in plasma. The 3'-conjugate performed better than the 5'-conjugate.3. bolus dose found in the liver. this treatment may lead to translation arrest. 7. oligonucleotides or by measuring pharmacological activities. and porphyrin-conjugated oligonucleotides were compared in one study (330). we have studied cholic acid conjugates of phosphorothioate deoxyoligonucleotides or phosphorothioate 2'-0methyl oligonucleotides and observed enhanced activity against HIV and no effect on the activity of ICAM-1-directed oligonucleotides (331). a novel europium complex was covalently linked to an oligonucleotide and shown to cleave 88% of the complementary RNA at physiological pH (334). From the perspective of medicinal chemistry. In principle. at present.to fivefold. cholesterol and cholesterol derivatives.v. Conjugation of chemically reactive groups such as alkylating agents. fluorescein. by RNase H compared to the nonconjugated parent (336). The 3'-conjugated antisense oligomers promoted faster hydrolysis of the target RNA than the unmodified compound and a stabilizing interaction between the ligands and the enzyme has been proposed. photoinduced azides. Cholesterol-conjugated oligonucleotides targeted to an mRNA fragment of Ha-ras oncogene were able to promote a greater extent of target RNA hydrolysis.3 RNA Cleaving Croups. daunomycin. The properties of 5'-cholesterol and 5'-C18 amine conjugates of a 20-mer phosphorothioate oligodeoxynucleotide have been determined in mice.

OMe 2'-OMe. highly electronegative fluoro or fluoro alkyl moieties. However. the content of the following discussion is restricted to a summary of the main events in this area (Fig.7. there is a direct correlation between affinity and potency. alkoxy alkyl groups. A 5'-cholesterol phosphorothioate conjugate was also recently reported to have a longer elimination half-life. Of these. Uniform modifications at the 2'-position have been shown to enhance hybridization to RNA.4 Sugar Modifications The focus of second-generation oligonucleotide modifications has centered on the sugar moiety. and to induce greater liver toxicity in rats at a high dose of 50 mglkg (340). A growing number of oligonucleotides in which the pentofuranose ring is modified or replaced have been reported (343). 2'-Opropyl Figure 5. which covers this topic in great detail (341). and another review has been published. They hybridize in parallel fashion to single-stranded DNA and RNA and are nuclease resistant. slight liver toxicity associated with the cholesterol conjugate (339). and sterically bulky methylthio derivatives. 7. Chimeric oligonucleotides containing 2'-deoxyoligonucleotide gaps (for RNase H activation) with 2'-modified wings have been shown to be more potent than parent molecules (144). These modifications include lipophilic alkyl groups. to be more potent. amphipathic amino-alkyl tethers. 2'-0Arninopropyl (2'-0-AP) modification exhib- . In oligonucleotides. and in some cases. Other sugar modifications include a-oligonucleotides. they have been reported to be oligonucleotides designed to inhibit Ha-ras expression.Oligonucleotide Therapeutics R= -F 2'-F. 2'-fluoro . the pentofuranose sugar ring occupies a central connecting manifold that also positions the nucleobases for effective stacking. 5. and hexapyranosyl oligonucleotides (343). 2'-O(2-methoxyethyl) 2'-OPr. a symposium series has been published on the carbohydrate modifications in antisense research. to enhance nuclease resistance (343). Recently.Therefore. A growing number of oligonucleotides in which the C2'-position of the sugar ring is modified have been reported (332. which summarizes the antisense properties (342). 2'-Omethyl Me -oaO/ -O*M~ 2'-OMOE. acetamide groups. a-oligonucleotides have been most extensively studied. as can be seen. carbocyclic oligonucleotides. positively charged polyamines.333).7). Selected 2' modifications. All these oligonucleotides support RNase H and. intercalators.

(2) additional solvation of the alkoxy substituent in water.5-Anhydrohexitolnucleic acid) (Cyclohexene nucleic acid) Locked nucleic acid Figure 5. The 4'-oxygen has been replaced with sulfur. we have compared the pharmacokinetics of 2'-0-pro- pyl-modified phosphodiester and phosphorothioate deoxynucleotides (182). nuclease resistance. Other modifications of the sugar moiety have also been studied.. In mice the 2'-0-propyl phosphorothioate was too stable in liver or kidney to measure an elimination half-life. Oligodeoxynucleotides containing 4'-a-C-aminoalkylthymidines have been reported to have acceptable RNA binding properties and the hybrids formed between this class of oligomers and their complementary RNAs were reported to be substrates for E.8. Incorporation of 2'-0-methoxyethyl into oligonucleotides increased the T.coli RNase H and human RNase H present in HeLa cell nuclear extracts. the 2'-0-propyl modification increased lipophilicity and nuclease resistance.7 The Medicinal Chemistry of Oligonucleotides 0 0 HNA 1 LNA I (1. Some new sugar modifications. 2'-0-Methyl-substituted phosphorothioate oligonucleotides have recently been reported to be more stable than their parent compounds in mice and to display enhanced oral bioavailability (328. 5. and (3) preorganization of the sugar moiety into a C3'-endo conformation (351). hexose-containing oligonucle- . including other sites such as the 1'-position and the 4'-position as well as more substantial modifications. In addition.Also.8). except the kidney. excellent review articles have appeared in the last few years on the synthesis and properties of C2'-modified oligonucleotides (333.3443471. several other 2'-0-alkoxy modifications have been reported to enhance the affinity (350). The analogs displayed tissue distribution similar to that of the parent phosphorothioate. The only difference in toxicity between the analogs was a slight increase in renal toxicity associated with the 2'-0-propyl phosphodiester analog (339). Similarly. As expected. Interestingly.l°C/modification when hybridized to the complement RNA. However. However. Although a single substitution of a 4'-thio-modified nucleoside resulted in destabilization of a duplex. In a similar manner. much less is known about the antisense effects of these modifications (93). 349). incorporation of two 4'-thio-modified nucleosides increased the affinity of the duplex (352). The beneficial effects of a C2'-substitution on the antisense oligonucleotide cellular uptake. and crystallographic studies involving this modification suggest that the interference of the positively charged 2'-0-substituent with the metal ion binding site of the exonuclease may be the reason for effective slow down of the nuclease degradation. whereas the affinity of the phosphorothioate for albumin was enhanced. and binding affinity have been well documented in the literature. The increase in affinity with these modifications was attributed to ( 1 ) the favorable gauche effect of side chain. The 2'-0propyl phosphodiester did not significantly bind to albumin. in which the 2'-0-propyl phosphodiester was remarkably stable. a ited very high nuclease resistance in in vitro studies. value by l. the rates of RNase H cleavage were found to be slower than those of natural substrates (348).. More substantial carbohydrate modifications have also been studied (Fig. the 2'-0-propyl phosphodiester was much less stable than the parent phosphorothioate in all organs.

5. However.2"C/modification when hybridized to both the complement RNA and DNA. Suffice it to say that numerous modifications have been made that replace phosphate. 2'-FANA has been shown to activate RNase H when hybridized to target RNA and cleave it. translation arrest. This claim is questionable on the basis of conformational and steric factors associated with this modification. The nuclease stability of ANA and 2'-F-ANA oligonucleotide phosphodiesters is not sufficient to use them as antisense drugs. Uniformly 2'- major groove of the helix. 7. compared to those of the corresponding hybrids formed by ANA. ANA) and 2'-deoxy-2'-fluoro-D-arabinonucleic acid (2'F-ANA)] are substrates of RNase H (358. the rates of RNase H compared to those of the natural analog have not been evaluated. 5.9). enhance stability. Both ANA and 2'-F-ANA are expected to have C2'-endo conformations.10). However. as shown by circular dichroism studies. compared to that of the control DNAIRNA hybrids. One of the important advantages of the high affinity backbone modifications is their use in RNase H-independent mechanisms of action. retain hybridization.exhibit Tmstabilization of 1. This destabilization (AT. recent reports show that hybrids of RNA and arabinonucleic acids [the 2'stereoisomer of RNA based on D-arabinose instead of the natural D-ribose (2'-arabino-OH. phosphorothioate DNA.O"C/modification)is presumed to derive from steric interference by the -C2'-OH group. A closely related analog with the cyclohexene-containing oligonucleotides. Because these modifications are now being evaluated in uitro and i n uiuo. which is oriented into the Substantial progress in creating new backbones for oligonucleotides that replace the phosphate or the sugar-phosphate unit has been made. and potentially improve the pharmacokinetics (Fig.1 2'-Modified Oligonucleotides with Ability to Activate RNase H. Because of this process. For a review of the backbone modifications reported to date. 1. and enhance stability. 93 and 355. resulting in cleavage of the RNA strand. PNA oligomers have been shown to bind to singlestranded DNA and RNA with extraordinary . Replacement of the entire sugar-phosphate unit has also been accomplished and the oligomeric compounds produced have displayed very interesting characteristics. possibily because of the poor binding affinity of ANA for RNA. The 2'F-ANAI RNA duplex had higher values of T .Oligonucleotide Therapeutics otides were created and found to have very low affinity for RNA (353). When hybridized to RNA they form an A-type helix similar to that of RNA-DNA hybrids. Furthermore. However. see Refs. such as inhibition or modulation of splicing. called cyclohexene nucleic acids (CeNA) (354). Several of these modifications have been reported to enhance hybridization (355). ANAIRNA hybrids had a lower Tmvalue. and DNA (AT. = . this analog is able to activate E. the toxicity properties of arabino nucleosides may be a matter of concern. coli RNase H . 7. ANAIRNA hybrids are poor substrates for RNase H. However. Bicyclic sugars have been synthesized with the hope that preorganization into more rigid structures would enhance hybridization. This modification has also been claimed to activate RNase H (357)as a mixed oligomer with DNA and not as a chimera involving deoxy gaps. However.5 Backbone Modifications modified oligonucleotides are not capable of activating RNase H in an oligonucleotide/RNA hybrid.4. This modification can interconvert easily between C3'-endo-like and C2'-endo-like conformations because of a low energy barrier associated with the cyclohexene ring system. causing local deformation (unstacking).5-anhydrohexitol-based nucleic acids (HNA) exhibit greater than 2OC/modification for RNA hybridization but fail to activate Rnase H in RNA-HNA hybrids.359) (Fig. a preliminary assessment should be possible shortly. The objectives of these programs are to improve hybridization by removing the negative charge. and disruption of necessary RNA structure. = + O. Replacing the ara-2'-OH group by a 2'-F atom resulted in an increase in duplex melting temperature.FC/modification compared to that of DNA).l. A bicyclic sugar analog connecting 2'-oxygen to 4'-carbon by way of a methylene bridge has been termed locked nucleic acids (LNA) and shown to have improved hybridization (greater than 4-5"C/modification) to both DNA and RNA (356). alter charge.

in the past 5 years. the length of the antisense oligomer could be shortened from 20 to 15 nucleobase units to obtain a comparable effect (360). enormous advances in the medicinal chemistry of oligonucleotides have been reported.6 Summary In summary. Modifications at nearly every position in oligonucleotides have been attempted and numerous potentially interesting analogs have been identified. A recent demonstrated that antisense PNAs alter splicing of the IL-5R-a pre-mRNA efficiently in a fashion similar to their 2'-0-MOEodified counterparts of the same sequence en electroporated in cells.9. 8 2'-0(2-METHOXYETHYL) CHIMERAS: SECOND-GENERATION ANTISENSE DRUGS The 2'-0-(2-methoxyethyl) (Fig. 7. using PNA as the splicing modulator. They have been shown to be able to invade some double-stranded nucleic acid structures. 5. and higher target binding affinity. O 0 W F B 0 ::!\OW 0 F 1 2'-F-RNA Figure 5. it is clear that a wealth of new chemicals is available for systematic evaluation and that these studies should provide important insights into the SAR of oligonucleotide analogs. Peptide nucleic acid oligomers were shown to be able to act as antisense and transcriptional inhibitors when microinjected in cells (47). and 2'-F-ANA. PNA oligomers appear to be quite stable to cleases and peptidases as well. PNA oligomers with lysine ends have better water solubility. Structures of RNA. PNA oligomers can form triple-stranded structures with DNA or RNA.L.11) substitution represents a significant advance in an- . less aggregation. 2'-F-RNA. then. O 0 W OH B 0 O W \!:: 0 RNA OH 0 ANA rz.8 2'-0(2-Methoxyethyl) Chimeras: Second-Generation Antisense Drugs ". affinityand high sequence specificity. Although it is far too early to determine which of the modifications may be most useful for particular purposes. ANA. Moreover.

one of which is a phosphorothioate oligodeoxynucleotide.to 15-fold increases in potency both in vitro and in vivo. the modified oligonucleotide was at least fivefold more potent (363). see Ref. tisense therapeutics and the culmination of more than a decade of progress in antisense technology. Numerous chimeric 2'-0-(2-methoxyethyl) oligonucleotides have not be& studied extensively. the other a 2'-O-(2-methoxyethyl)chimera. administration in a rat heart allograph model. for review. They Figure 5. Although such oligonucleotides typically display affinity for target RNA that is several orders of magnitude greater than that for phosphorothioate oligodeoxynucleotides.. Although oligonucleotides in which every nucleotide contains a 2'-042methoxyethyl) modification have proved of great value when used for occupancy-only-mediated mechanisms (e. PNA peptide nucleic acid without (A) and with Lys end (B). have been used most broadly in chimeric structures designed to serve as substrates for RNase H.11. showed that in vitro the 2'-O-(2-methoxyethyl) chimera was approximately fivefold more potent.v.g. induction of alternative splicing. after i. 361). Similarly. they are less attractive substrates for RNase H (for review. Consequently. For example. Structure of 2'-0-MOE: 2'-0-(2-methoxyethyl) RNA.Oligonucleotide Therapeutics Figure 5.10. a direct comparison of two antisense inhibitors of C-raf kinase with the same sequence. 362). Similar data were reported from a direct comparison of antisense inhibitors of survivin both in vitro and in vivo in a human tumor . they typically display only 5. see ref.

372). a 2'0-MOE chimeric antisense inhibitor of TNFa. unpublished observations. Pharmacological studies of antisense oligomers modified with 2'-0(2-methoxyethyl) chemistry have recently been reviewed (365). 2'-0-(2-methoxyethyl) chimeras have been shown to be less proinflammatory (169. oral bioavailability is potentially feasible (215). systemic bioavailability in excess of 20% was observed in all three species (215). dogs. 10 ABBREVIATIONS aPTT AT-1 AT-2 c-AMP CeNA activated thromboplatin angiothromin receptor 1 angiothrombin 2 cyclic AMP cyclohexene nucleic acid CMV cytomegalovirus . and monkeys. Studies in animals have recently been extended to humans. so within the next few years we should have some sense of the near-term potential for oral delivery of antisense inhibitors in humans. ISIS 104838 produced dramatically lower local inflammation than that of first-generation antisense drugs. we will be in the position to perform progressively more sophisticated studies and to understand more of the factors that determine whether an oligonucleotide actually works through an antisense mechanism. In mice. In rodents. Although significant progress in achieving acceptable oral bioavailability of antisense inhibitors has been reported. a 2'-0-(2-methoxyethyl) chimera. 366. 2001). Initial clinical trials in which a variety of solid dose forms containing ISIS 104803. a TNFa antisense inhibitor. With the development of 2'-0-MOE phosphorothioates. Finally.10 Abbreviations 155 xenograft model (364). After repeated subcutaneous dosing in humans. Because the gut has a very high level of nucleases contributed by both the host and bacteria resident in the GI tract. two key barriers to oral bioavailability have been identified and partially overcome: presystemic metabolism and penetration across the gastrointestinal (GI) mucosa. Furthermore. Thus. see Ref. 2001) and to have an elimination half-life in plasma of approximately 30 days (Geary. metabolism of phosphorothioate oligodeoxynucleotides in the gut occurs much too rapidly to support adequate oral bioavailability (370). ISIS 104838.5%). 2'-0-(2-methoxyethyl) chimeras are substantially more stable than phosphorothioate oligonucleotides. Chimeric modifications of the oligonucleotides have been shown to enhance stability and oral bioavailability (328. the elimination halflife is nearly 30 days in plasma and several tissues (314. 366-368). more convenient and better locally tolerated subcutaneous administration may be feasible. was shown to reduce TNFa secretion at doses at least 10 times lower than would be expected in humans by first-generation antisense drugs (Dorr. rats. 371). unpublished observations. After - intrajejunal administration of several 2'-0MOE-modified oligonucleotides in the presence of penetration enhancer-containing formulations. Initial studies with solid dose forms containing penetration enhancers resulted in significantly less systemic bioavailability (5. 9 CONCLUSIONS Although many more questions about antisense remain to be answered than are currently answered. in the liver. much remains to be accomplished and key clinical studies in progress will help determine how feasible oral delivery is (for review. We should also have the opportunity to learn a great deal more about this class of drugs as additional studies are completed in humans. progress has continued to be gratifying. as more is learned. elegant correlations between phannacokinetic and phannacodynamic effects have been reported to show a duration of pharmacological action of nearly 20 days in the mouse (369). 215). Clearly. the permeability of the GI tract to antisense inhibitors has been significantly enhanced by using formulations containing penetration enhancers such as bile acid salts and fatty acids. clearly therefore more progress is required before solid dose forms with adequate (10-20%) oral bioavailability are available (215. Perhaps more important. These studies demonstrate the feasibility of once-a-month dosing. 367). Based on the progress. and various penetration enhancers are in progress. and monkeys.

C. 5749-5763(1987). 1983. A. D. 25. New York. R. B. A. L. 579-588(1972). USA. and M.. Crick. C. G. J. Krivit. 4. 149. C. J. . Young. S. 4 . W. 26. 11. J. Biol. Clin. D. Pathol. 84.1137-1139(1969). S. S.. P. Miller. Proceedings of the Symposium on Molecular Biology.. Ts'o. Nucleic Acid Res. 202. Suhadolnik. Development of Target-Oriented Anticancer Drugs. Response Mod. Field. A. Lee. 1971. Schlick. Yu. Prog. Springer-Verlag. Am. Chamberlin. 23. De Clercq. Proceedings of the Symposium on Molecular Biology. G. Pitha. S. P. J. Kerr.. Hammond.261-266(1990). Acad. TX. and P. M. Coo.J . P. J. 3. G. Gillespie. Simon. Beers. J. School of Graduate Studies. F. 13. New York. 213. R. Mol. R. S. Homan. and P. Mol. M. M. Eckstein. Stevenson. B. H.. Zendzian. Glick and A. C. Cohen.pp. R. 9 . E. Bettens. Clin. 15. Watson and F. E. 0. Response Mod. A. Gillespie and S. Carter.676678(1967). 0. Appl. Glick. C. W. Foon. Woerner. 14. Response Mod. G. Stewart. Reichenbach. Herberman. 6.Shinozuka. 189. Res. H. P. Houston. M. 1972. H. W. Alvarado-Urbina.829-842 20. J. R. Science.. Lancet. J. and J . 10.. 4. Mitsuya. 27. Liu. and D. J . P. Salim.270-274(1981). 997-998(1965). K. CA. Hewetson. Zon. Smalley. New York. CA. 23. 737 (1953). M. Krown. 165. and K. Colby. L. F. 2 . P. H. Science. Rotem. J. F. Quinnan. R. R. 230. Biol. Nature. P. and T. W. 1971. A. M.4. and M. 1. Art Levin and Mano Manoharan for thoughtful review and additions to several sections of the manuscript. Caruthers. 11. J. C.2338-2341 (1967). Greene. Duck. Science. 5 . Pellegrino. E. F. Raven Press. Ogilvie. T. D. D. University of California. Preliminary Studies on Genetic Engineering: The Uptake of Oligonucleotides and RNA by Novikoff Hepatoma Ascites Cells. 171. P. 64. Colby. 24. C. M. S. and W. Biol. P. J. N. I. 15. 16. Levy. REFERENCES 1 . Response Mod. Springer-Verlag.Oligonucleotide Therapeutics Ha-ras HIV HPV HSV ICAM IL IL-la LNA mRNA NK PKC ply rC PNA poly rI PTHrP Tm ATm TAR vsv Harvey ras human immune deficiency virus human papillomavirus herpes simplex virus intercellular adhesion molecule interleukin interleukin l a locked nucleic acid messenger RNA natural killer protein kinase C poly ribocytosine peptide nucleic acid poly ribinosine parathyroid hormone-related peptide thermal melting temperature change in Tmrelative to DNA transactivator response element vesicular stomatitis virus 11 ACKNOWLEDGMENTS I thank Drs. L. Proceedings of the Symposium on Molecular Biology.449-456 (1985). Ts'o. C. Lancet. 12. A. and G. R.. E.305-320 (1971). 13. Carter.. Schoenberger. A.628-633(1985). Crooke. 8. Nucleic Acids Res. K. Sathe. CA. R. (1965). La Jolla. M. 1971. A. Brodsky. V. and W. Biol. Miller.. Biochem.281-285 (1985). Carter and P. 23. Proc. Isaacs. Zon. and H. 214. Ruffmann. Cancer Res. J. Oettgen. Springer-Verlag. and R. E. 18. Adamson. Barrett. M. Goldberg. Strayer. Biol. and K. Glick and A.. 7 . Baylor College of Medicine. J. D.1286-1292(1987). A. R. Biol. P. Reitz. Toxicol. I. 7706-7710 (1987). Duesberg. Marcus-Sekura. Brodsky. and Z. 17. S. Natl. L. 12. B. University of California. D. K.4897-4906 (1974).1-32(1971). C. Levy. J.. Spiegelman. Gillen. G. New York. Pharmacol.650-655 (1985). H. Sci. H. W. Thanks also to Bella Aguilar and Marcia Southwell for converting the manuscript into readable material. A. B. Biol. 27. J.531-537(1985). C. H. J. 113(1963). 21. Chirigos. R. R. G. H. Abrarns. Broder. and S. D. La Jolla. Matsukura. Lampkin. University of California. C. Nature. 22. Levine. A. Science. F. La Jolla. 2.640-649(1985). Biochemistry. Merigan. 19. 4. Yu. Prog. Thompson and D. M. Bender.

A. G. J . M .-L. 7749-7760 (1987). T . 2001. J . J . Glazer. G. and E. 203. Goodchild. Krawczyk. 83. Cancer Res. Gagel. Noble. and P. 645-650 (1987). W . Hassman. M . Bacon. Bi0chemistry7 373 17714-17725 41. Lacoste. Egholm. 15. L. and Applications.2787-2791 (1986). M . C. Biophys. S. 62. Science. and M . E. W. 21312138 (1994). Sarin. F. USA. Biol. and L. M . Koizumi and E. P. Decout. E. Koppelhus. F. Civeira. Biol. Dutschman. Biochem. 35. Proc. 30. T . Nielsen. 46. Praseuth. Wang. 56.V . Mergny. Glazer. A. M . Chem. Nielsen. A. Reddy. Sci. 22595-22601 (1995). 50. Flory. Prasmickaite. Praseuth. Jarvis. D. Mirabelli and S. P. A.. G. 51. and P. 83.-C. Antisense Research and Applications.297-302 (1989). Antisense Drug Technology: Principles.157 28. H .468-471 (1998). S. 60.802-805 (1996). Cooney. 55. 1993. T . A.. and C. Baird. M .5507-5511 (1988). E. Seidman. C. Egholm. R. and P. G.. E. Zamecnik. 0. 391. Strategies. M. M . Faruqi. Pluznik.. A. Carter. S. and P. New York. A. USA.3911-3914 (1999). S. R. Antisense Research and Applications. Pavco. Liu. Marcel Dekker. N. Chem. H . Biophys. Babiss. Neckers. G. Civeira. pp. Lyman. R. Flint. J. C. Biol. D. A. 0. M . M. M . M . Egholm. R. J . Mol. 49. A. and P. M. A. Proc.. Saragosti. pp. Buchardt. New York. USA. 85. E. Antisense Research and Applications. 226-238 (1996). Marel-Bellan. W . C.63342. 7-35. 36. Wincott. Glazer.. M . Nature. Matteucci. Berg. J. Antisense Research and Applications. E. S. F. N. 25. F. Thomson. J . M. F. A. 61. B. P. andP. Wickstrom. pp. Wickstrom. L. J. S. L. A. and N. M . 1309. J. Levy.45638. Sci. 59. 25. P. N. Glazer. J. Levy. D. Perkins. 58. P. Watt. 31. Loke.. Cadilla. Natl. G. Josey. New York. Sarin. and L. Ebbesen. P. P. Boyd. Ts'o. 59. J. Nielsen. 258. 197-200 (1993). FL. 363-373. T . Sci. Le Doan. Turner. 0. D. pp. 54. T. 271. Helene. Parthasarathy. M . D. K. Miller. Proc. D. Acta. M . Hogan. Habhoub. FL. Natl. Pritchard. Zachar. L. Antisense Technology: Principles. A. Goodchild. 2001. Lesch. Cell. L. Goodchild. R. 445-449 (1987). W . Burli and H . Moser. Faruqi. 59. Natt. Thuong. In Vitro Cell D e w Biol. 241. M . J. Bruckenstein. W . Seidman. Acad. Lacroix. Jahnke. Hogset. E. Buchardt. Biochem. 21. Marcel Dekker. D. Nucleic Acids Res. and E. 5145-5150 (1991). H. 52. Science. Proc. Acad. A. T . Biochemistry. L. h a d . 38. P. Fodstad. 1398-1403 40. F. M. H. M . 26. 32.. 0. E. 95. Hovig. and R. 47. L. A. Heidenreich. Nature. Rev. P. Natl. 270. CRC Press. E. Uhlenbeck. Nielsen. Glazer. M. and 0. Wickstrom. Gao. and P. Postel. K. P. Chem. G. 34. Ricca. Bisi. Nucleic Acids Res. J. Boca Raton. D. Biochim.. F. and P. G. Szewczyk. Natl. Helene. Blommers. M. 30. Sci. Nucleic Acids Res. Heikkila. Rossi. 44. Cote. L. M. Stein. P. Au. 833-857. M . Seidman. X. Schwab. Zamecnik. 889-898 (1994). J . and 0. Grigoriev. 2001. Cheng. Ventura. Strategies and Applications. D. Science. 33. Nucleic Acids Res. 43. Acad.543468 (1990). and B. C. Reddoch. 269. 85. M . M . K. D. Cech. Wang. J. M. D. R.2167-2173 (1997). USA. Crooke. H . Ohtsuka. Peffer. Glazer. G. CRC Press. E. Strategies and Applications. D. 238. FL. Perrouault.. 1993. 48.. J. S. 37. H. Berg. L. Glazer. S. E. Guieysse. J. C. A. R. . Nucleic Acids Res. S. and S. CRC Press. Res. E. 1993. Matteucci. L. Fahrenholz. S. F. U. A. White. Annu. T . and P. Antisense Technology: Principles. Dervan. and P. Commun. 11521-11526 (1989). FL. J. P. Beigelman. 4241-4248 (1998). E. E. p. Czernuszewicz. 1993. Gonzalez. L. M . G. Olsen. S. 1893-1901 (1999). Biochemistry. 25. M. L. Eckstein. and F. J. Maelandsmo. C. Sun. Smith. M . C. Thornton. Macris and P. Boca Raton. and Y. 15. pp.. Boca Raton. G. Agrawal. A.1759-1768 (1995). 14811485 (1992). Bonham. pp. S. J . H .. C. Agrawal. 264. Boca Raton. 45. C. Cohen. M . E. Benseler. 328. S. Franck. E. 53. Moser and P. 39. C.Wang and P. Aurelian. 57. H . C. 887-905. Marcel Dekker. J . Dervan. E. Hanvey. R. Berg. Science. E. J. P.Wang. 7079-7083 29. Biol. B. Cuenoud. S. CRC Press. J.

63. Lemaitre. 89. USA. and P. B. U. Boca Raton. Calabretta. Sci. 1945-1954 (1993). 81. 91. Iozzo. Biol. G. USA. Sci. Chem. A. Helene. E. 1. W . Milhaud. Deliv. Ecker. Crooke. 79. L. Antisense Research and Applications. USA. 3474-3478 (1989). J. Neckers. P. Subasinghe.. J. FL. Delarue. 67-82. McKay. T . Nicolaides. S. C. 72. P. Kant. Sandberg.4751-4757 (1990). Klimkait. Proc. Cohen. 80. Loreau. Zounes. 76. R.. Natl. S. Crooke. Crooke. T X .Wyatt. L.. Helene. Ther. E. 818-822 (1990). Schrier.Mori. Germany. Nature. Agents Chemother. A. T . S. G. 1993. M . T . Z. Gorey-Feret. P. P. T . Tidd and H. Thuong.. M . Lancelot. S.. K. Antisense Nucleic Acid Drug Dev. 74. H . Lima. 505-510 (1990). R. C. Bennett. P. E. and D. Sands. J. Sci. Crooke. Zhang. K. D. T . Crooke. Bayard. Freier. A. 17. Antisense Research and Applications. Cowsert. Skorski. Ecker. R. 1993. S. M . N. S. Ther. Ecker. S. J . 249.W . pp. L. B. A. 1-50. M. 18. 521-533. Acad. B. Toulme. Bao. Chem. Exp. Sarin. T . M . D. H. A. Proc.Oligonucleotide Therapeutics HuntIII. M . R. and S. Y . Science. Science. 297-320 (1995).97-102 (1986). 91. M. Chiang.. Cook. M . Acad. Baker. A. Natl. T . T . Zounes. 267. Wincott. Zhang. Boca Raton. Biochem. USA. M . J. J. T .. V . R. Hobbs. M . 145-146 (1994). Anderson. Acad. P. Markham. Pharmacol. pp. F. J. USA. Chidester. 81. 95. Natl. Sci. Cummins. Parry. N . Reynolds. USA. DeRiel. M . 31. and A. M . Natl. and C. Lebleu. and T . Acad. L. Nucleic Acids Res. 3359-3368 (1991). P. Sasmor.754-758 (1996). W . 37. J. CRC Press. J. 138. Troung. Gene. R. R. J. T . J. S. Sci. Therapeutic Applications of Oligonucleotides. and S. Lima. and D. Drug Target. J.. 84. 70. Walder. L. Roberson. M. Owens. 1137-1140 (1969). and D. Cohen. J. Nieborowska-Skorska. F.Vickers. Wickstrom. Crooke.44994503 (1994). 97. Gallegos. P. Agrawal. Rayner. P. Jr. Ikeuchi. Imbach. 69. Jr. F. M. M. Condon. S. R. T . 7-35. A. h a d . 271-277 64. and L. J. V . Shea. Biochemistry. Proc. Walder. J. Proc.. M . Monia. A. Stein. DeBaets. 269. Natl. M. 1356-1360 (1994). Landes. Nakanishi. P. Weeks. Acad. Leonetti. S. C. M . Antisense Res Dev. Natl. 93. C. h a d . and J. 87. G. and S. Anticancer Drug Des. T . CRC Press. Gewirtz. L. CRC Press. J. De Clercq. G. Crooke and C. 1181-1190 (1993). R. 94. Manoharan. 72. Crooke. Berlin. T . Methods. L. D. 67. 9. 96. 68. Szczylik. M . D. D. Cossum. Dagle. Pharmacol. Sun. Zounes. 1993. T . Dagle. Acad. FL. Science. M . Toulme. W . FL. L. Loke. 66. C. Bennett. Zhang. P. J. 82. 91. Gagnor. Biophys. K. and B. Buckheit. Iversen..323-332 (1988). J. T . Merigan. 4. G. Cancer. G. L. F. D. J. D. Springer-Verlag. Ellington and J. S. Antisense Research and Applications. T . 11 (1991). K. 6. A. 75. Buckheit. Proc. Hijiya. B. A. and B. and J . Vickers. and M . 60. F. . Acad.18162-18171 (1991). J. P. J. Truong. L. J. 346. D. S. 609646 (1991). Nucleic Acids Res. Natl. K. Cazenave. Asseline. Boca Raton. S. S. T . Tanaka. D.648-652 (1987). W . C. S. 4 .4255-4273 (1989). Austin. 65. Nucleic Acids Res. D. 47. N. L. A. Zon.. and F. 257. Lemaitre. 78. Ecker. R. 269. F.. Exp. E. A. Natl. M . J. M . 90. A. Antisense Research and Applications. J. 1998. 86. Z. and D.6553-6556 (1989). Antisense Res. 85. N. 1993. Rayner. Tuerk and L. 83. Cossum. M. Mol. R. W . Davis. C. Proc. S. J. pp. Gold. Markham. Warenius. Biol. 91. J. Szostak. 93.89-94 (1994). M . Mirabelli. (1999). X.. B. and B. 4.. Grimm. B. F. Letsinger. Antimicrob. L. Montenay-Garestier. CRC Press. W . P. Owens. Jenison. Pharmacol. and K. 13. M . C. Ho. USA. Sci. Herlyn. J. Vickers. 19. H. A. J . T . Y. pp. Cocuzza. 165. B. 1995. 266. Bruice. P. Pavco. Dellinger. 958-961 (1992). USA. Br. Dev. Weeks. Eckstein. M. Proc. Sci. Proc. Shea. R. D. 71. 73. Ecker. A. 16416-16424 (1994). H. Freier. Handbook of Experimental Pharmacology: Antisense Research and Application.4504-4508 (1994). pp. M . 88. C. W . Driver. Ratajczak. Chan. Thuong. Imbach. 12055-12061 (1992). M.343-350 (1989). Freier. N. Boca Raton. FL. and C. A. F. 84.636-646 (1995). J. Bouhana. L. 86. Sci. P. 86. 92. J. F. Lebleu. S. 386-400. J. N. Stein.. Dean. 77. M . A. and C. C. P. Neckers.3297-3301 (1984). Crooke. Subasinghe. R.. M. p. R. R. and T . Freier. L. C.

179192 (1979). Manrow. 276. F. Proc. Monia. Blood. 17445-17452 (1996). M . Dean. Lesnik. and P. J. Chem. Taguchi. Kole. Condon. J. 1985. Sproat. Bennett. USA. R. Gross. Proc. and I. L. p. 521-529 (1991). S. Matthews.18162-18171 (1991). Mol. D. and R. A. 33783379 (1993). Proc. Baker. A. Sci. P. A. and L. Biol. R. S. Immunol. Biol. T . Miraglia. 68686872 (1989). S. N. 267. and R. Irnbach. 266. S. Biotechnol. B. 137. Sarin. C. Walker. B. A. USA. Acta. A. B. F. Cold Spring Harbor Laboratory Press. F. P. Weller. D. Kilgannon. and R. D. Maciag. M . L. 831-841 (1993). Y. B. Biochim. Roeder. Acad. Lerner. and M . P. 88. D. Acta. N. B. R. and D. M. Cold Spring Harbor. CRC Press. P. Chiang. P. Helene and J . 1. E. 0. Zamecnik and M ... C. T . Zounes. R. D. L. J. K. 271. Shiramizu. P. 48. G. Al-Nasser. Natl. Goodchild. Chem. Mol. Cook. R. Z. Agi-awal. 5409-5415 (2000).. 1489. and C. Biol. Cancer Res. P. Ackerman. Kole. Lu. Hagedorn. S. Nelson. Sburlati. Sci. H. B. Berger. B. Dirksen. Ahmed. Zerial. E. Thuong. H.3263-3269 (1990). Summerton. and L.11495-11499 (1992). Natl. Zounes.. and P. H. Zheng. Hodges and S. and P. 4569-4581 (1988). p. Lesnik. Natl. Beijer. 335. C. Ikegaki. Crooke. 249. J. Monia. 25. W . Paulauskis. Acad. 164. 16411-16417 (2001). Natl. USA. Kole. E. J . Biomed.63-70 (1997). T. Lamond. 1097-1100 (1999). and G. C. Freier. P. 36. and N . 265. K.. A. Weller.Whitesell. Knudsen and P.. J . Anticancer Drug Des. Dominski and R. 5107-5114 (1989).. 8. Stevenson.. M . Huber. M . Neckers. 265. 494-500 (1996). B. Boca Raton. E. 15. Acad. J. M . Johnson. Cidlowski. Goldschmidts. A. Fishelevich. FL. J. L. C. Green. Rayner. C.. S. 280-284 (1978). 86. E. Sambade. T . Sci. Natl. T . J. L. Dominski and R. 17. J. Rayner. Natl.. J. Grignini. Baker.. K. Crooke. Sahai. P. A. E. J. Am. J . 16. and T .. M. Lima. A.3-18 (1999). 93. J. T .A. and D. Biol. Q. Saxena and E. Nucleic Acids Res. Baker.8673-8677 (1993). and S. 7445-7454 (1994). Miller. Kennett. Leonetti. Smith. Acad. R. Acad. Neuner. Z. D. Morvan. Le Coutre. Rosolen. Med. J.2428-2430 (1990). L. J. J. Dean. Kulka. Antisense Nucleic Acid Drug Dev. 50. Biochim. Sasmor.9909-9919 (1987). 17. Karras. H. F. Y . Sci. B. and C. A. Loke. Imbach. N. Proc. and B. Proc. Pharmacol. 90. 1989. and T . Sci. Thuong. N. Nucleases.3758-3762 (1989). Stein. Kole. and C. and P. Hagedorn. Chem. L. Antisense Nucleic Acid Drug Dev. Voulalas. McGraw. C. Bhatia. 3373-3386 (1989). Freier. 88. J.. F. Mologni. 24. Boca Raton. L. Roth. Gonzalez. C . MonteneyGarestier. Nielsen. Biol. F. Nucleic Acids Res. Ts'o. Kobzik. M . USA. Helene. B.. R.. Ryder. P. C. Cummins. Cell.. L. S. Gagnor. E. S. M. Biol..rences B. Mercatante. K. S. S. and J. 7. M. Bortner. H . Gambacorti-Passerini. K. H . 48. Chem. L. S. B. R. 6316-6322 (1990). Sci. Cooper. A. T . Taylor. Chem. R.. 115. F. C . M. M . Nazar.. C. Acad. Chem. 253-257 (1991). Cancer Commun. L. . R. D. Nielsen. 17. T .. Sci. F. 6. Oligodeoxynucleotides: Antisense Inhibitors of Gene Expression. F. W . M.225-232 (1989). D. NY. Nut.. B. Oligodeoxynucleotides: Antisense Inhibitors of Gene Expression.905-918 (1995). 1989. C. K. H . pp.. McKay. and C. Elela. U. Westermann. FL. Nucleic Acids Res. P. L. and U . 211-241. P. Kawasaki. 14. Chem.1411-1417 (1996). A. Huang. S. Magrath. I.. P.. Crum. Meade. 83.. C. Nucleic Acids Res. and B. Donis-Keller. 75. 86. M. Taylor. Neckers. Zamecnik. K. Soc. P. Aurelian. J. Zhang. USA. 12840-12844 (1996). Walker. Hoinkis. Asseline. Vasanthakumar and N. L. Biol. 4143-4146 (1986). Biol. CRC Press. D. USA. Marchesi. J . Sierakowska. G. Nucleic Acids Res. Proc. McManaway. USA. A. Biophys. Lebleu. Lot. L. N. 7. J. 272. Baker and B. Chem. R. 133-134 (1998). Acad. S. A. Q. J. J. 11994-12000 (1997). 1570-1574 (1990). M. Casper. S. C. Chem. Maier. Proc. Nucleic Acids Res. Fotedar. Bennett. D. M. Wyatt. 808-811 (1990). ChengFlournoy. Lancet. Toulme. Bertazzoli. H. Crouch and M. M . J. Science. S. Partridge. A.85-93 (1989).-L. Natl. Redner.

151. Crooke. . New York. Kornburst. L. W. Lima. and S. Anderson. H. Paoletti. Acta. Cohen. R. Temsamani. C. CRC Press. B. V. K.. Zuckerman. C. F. 275. Chem. A. Hawley and I. R. and S. L. 153. (1992). Natl. and S. Imbach.. Zhang. Boca Raton. Crooke. 166. Templin. Eder and J . 882-886 (1992). 139. Burger's Medicinal Chemistry and Drug Discovery. Ther. Ann. Antisense Nucleic Acid Drug Dev. Frank. B. Miraglia. X . 131-139 (1995). Lemaitre. D. T . 725-730(1989). J. Cazenave.-Y. Rev. pp. Nucleic Acids Res. 67-82. 11538-11543(1998). 154. 170. Biol. K. J . W . Han. Antisense Research and Applications. Pharmacol. Xu. 5. 174. Frank. Malvy. 79. J. Dominski. D. Proc. Pharmacol. and S. Walder and J.Giles and D. Antisense Research and Applications. C. Boca Raton. S. L. 41. P. W . 149. Albert. Biochem. Science. 17. Exp. and S. Guinosso. Biol. Stein. F. R. Brakel. P. 22. Natl. J. S. K. M . and B. T . D. Hunt. Tivel.53-61 (1998).329-376(1992).W . CRC Press. P. 223-229 (1992). and S. Hall. L. T . Lima and S. Iversen. W . 36. J . Anticancer Drug Des.453-461 (2000). P. T . Monia.. C.. Watt. Crooke. Crooke. R. A. 4 ..M . Cheng. Braunshofer-Reiter. and R. 274.. Lima. 95. Grimm. Crooke. 1. USA.. E. CRC Press. 275. Schleich.5011-5015 (1988). S. 15.14514-14522 (1993). H. Oligodeoxynucleotides: Antisense Inhibitors of Gene Expression. Natl. S. 266. 312. F. W u . R. C. Stein and Y. and C. E. 1489. Rayner. A. 427-499. F. Conklin. K. E. J. Acids Res.19954-19962 165. Kawasaki. Sokol. Nucleic Acids Res.. P. M . F. 150. 156. Proc. Griffey. 144. Exp. and A. 1989. A. 205. USA. 5247-5254 (1994). C. Sci. Srinivasan. Cummins. 155. F.J. Tidd. A. D. J. Cheng. A. C. L. FL. A. Biol. FL. A. Y . 267. 169. Nucleic Acids Res. W . and M. 162. CRC Press. H. Biochemistry. M . P. J. Crooke. S. Griffey. Lebleu. Ther. 155-166 (1991). Boca Raton. 6.Mirabelli.. F. Greig. A.. Toulme. S. 164. 8.. T . Lemonidis. L. and P. Biochemistry. Dev. T . Freier. and R. John Wiley & Sons..p. 863-900.19-30(1999). Chem. L. 145.. and J . B. 20. 268. Antisense Res. and J. Natl. 32. Proc. D. H. 175. Lima. 168. 1995.. 152. Cowsert. 158. 138. Miraglia. Lima. Antisense Nucleic Acid Drug Dev. Minshull and T . J . Monia. Cheng. Proc. A. Strategies and Applications. M. Busen. pp. J. Cook. Chem. Biol. K. 1004-1012(1993).12872-12877 (1998). Crooke. 166. Biol. and S. M. B. MaherIII. W . R. S. pp. R. L. Zounes. M . H. Ecker. C. P. T . S. Crooke. Antisense Technology: Principles. L. 10. P. S.28270-28278(1999). M . H . Graham. 277. Graham.7253-7262(1989). P. Biol. Dervan. F. McGee. 172. 147.W . Quartin. R. S. Monia. S. J. M. T . 160. 245. Gewirtz. J. Neilson.M. M. and P.599-608(1995). Bennett. 146. A. L. J . S. Gonzalez. S. T .-C. E. Nucleic Acids Res. T . Lima and S. Science. F. 140. FL. Bertrand. Crooke. H. W . C. J. U. S. C.6472-6479(1991). Levin. J.. Gao. T . K.9193-9204 143. USA. 85.10419-10436(1987). Acad. and S. Kole. Wintersberger. Henry. Toxicol. Johnston. M. T . Acad.36957-36965(2000). T . 647-661(1991). M. Wetmur.Oligonucleotide Therapeutics 137. FL. 261. Ther. B. Thenet. Chem. 1993. F. C. C. 157. Majumdar. Sci. Lesnik.. Guinosso. Walder. and Y . W . pp. 201-267. 176. Exp. H. J. Sci. J. New York. R. J. Han. 148. Nucleic (1989). Antisense Nucleic Acid Drug Discow. W u . G. M . Biochim. Tang. 173. Manoharan. Nucleosides Nucleotides. D. Sasmor. 142. 10. Gibson. Acad. Vol.27513-27516 (1997). Crooke. J . M . 1993. Pharmacol.Furdon.. J. Wold. Y . F.-S. T . Tewary. 95. J . Egan. J. S. Graham. Storm. C. Sci. 14.113(1969). 763-770(1992). Gagnor. and F. M . Z . 10. Agrawal. Bennett andL. Brooks. 88. 1993. Crooke. P. Antisense Research and Applications. Acad. and S. BioTechnology. 171. M. 28. p. Crooke. 1340-1346(1989). USA. Crooke.6433-6451 (1986). Freier. H. J . J. B. M . J. Monteith.. C.. W u . 2001. D. 272.. Marcel Dekker.185-195(1996). Wilson. Chem. Boca Raton. Crooke. Lu. Mol. J. 159. 6. 167. 7595-7599 (1991). B. Gao. Chem. Crooke. 163. J. C. Freier.. 141. Pharmacol. 161. S. Biophys. Cohen. and J. 390-398(1997). Lesnik. Y . Cooke.. J. L. Broder. S.923-937(1996). 462-473 (1995). T .. P. S. Walder. Nucleic Acids Res. 17. M . Joos and W . Morvan. Miller. Pharmacol. and W .

and J.. Agents Chemother. R. Coleman. and G. H. Yan. L. Sci. Henry. D. Mini. Bacon. S.. Ruddon. E. Lesnik. T . K. 58. M. H. Cossum. Owens. J. R. R. J. Boca Raton. E.. 5. Chiang. Ther. P. Freier. Grillone. Clin. Shoemaker. Kidney Int. Ther. G. T . G. E. J.. and S. Crooke. J. M. S. T . and R. Zounes. 4. Tewary. Iversen. 5 . M.182-193 (2000). L. and S. W . Zhang. Neckers. Lee. and D. Ther. Cummins.1023-1033 (1992). K. L. A. H. Crooke. J. Mol. Cauchon. and M.. Cocuzza. Tendolkar. A. L. 1993. S. S. K. Arneson. Sands. 1181-1190 (1993). P. D. Antimicrob. Nucleic Acids Res. P. J. Campbell. Martin.. Boca Raton. M. M. Lab. M. Antisense Research and Applications. J. Ther. J. A. Ther. Zon. J..3318-3322 (1995). 549-553 (1994). Hobbs. 5. Papucci.85-102 (1993). Draper. Dew. Dev. Markham. Pirruccello. Templin. Pharrnacol. Saag. B. Hoke. T .W . P. C. K. G. J. Mol. Natl. K. Chan. Sci. Kanamaru. A. Dellinger. J. Monteith. J. C. J. Liposome Res. Ther. Stein. Truong. J. R. Marx. and R. L. 77. J. M. Y . Cowsert. Stecker. Griffey. S. C.V. Takakura. Moore. J. M. Kornbrust. 1181-1190 (1993). S.. D. Lu. Z. J. Markham. 171-177 (1993).177-183 (1996). 277. G. Butler. K. Exp. Z. Pharmacol. Low.Watanabe. Jiang.. P.. S.. D. Antisense Nucleic Acid Drug Dev. T . Giles. M. Hashida. and W . G. J. A. F. Bennett. H . P. Grimm. E. R. Matson. K. M . M. Takakura.. Iverson. CRC Press. J. S.. P. Garrett. FL. M. Owens. Antisense Research and Applications.447-458 (1998). Iversen. Ecker. Schleich.. Miyao. Yoshida. C. R. M. Y . Truong. Shahinian. 56. Agrawal. Opin. and C. and S. Mirabelli. D. Pharmacol. S. Pharmacol. R. 2. T . Antisense Res Dev. 686-693 (1991). Curr. Sharp. 10. T . A. J. Klotman. H. and G. M. K. S. Methods. Wongwit. Acquir. Guinosso. B. M.. Temsamani. L. Driver.. 47. M. . Crooke. B. P. Biochem. Gonzalez. Chiang. Bayever. Zon. F. E. Graham. N.435-441 (2000).. R. 20. 6. 6. Crooke.259267 (1990). Bennett. Burckin. Manoharan. 45. Pharmacol. Brooks. Copeland.123-129 (1995).932-943 (1994). D. Crooke. and S. V .. 267. M. 1462-1469 (1995). H. and R. M. Drugs. J. Antisense Nucleic Acid Drug Dev. Wickstrom. Freier.367-381 (1993). Antisense Res Dev. Tao. R. Pharmacol. S. Quattrone. Walz. P. Crooke. 2 5 3 1 (1995). 41. P.J. A. Cummins. K. Leeds. A. Pazirandeh. Exp.5743-5748 (1991). K. G. 9. 90. Owens. E. Spiller. J. L. D. T . Ther. Fitchett. J. Kessinger. 1993. Pharmacol. Bennett. Anticancer Drug Des. R. R. Capaccioli. H.. FL. L. M. Leeds. 451-460. G. M . L. P. D. 92. T . Pharm. J. Boca Raton. Geary.383-390 (1993). M . and M. M. Maslova. R. and E. Anticancer Drug Des. 3737-3743 (1993). G. C. M. N. R. 3. H. Pharmacol. Hartmann. CRC Press. A. and C. R. S.89-94 (1994). Fox. Dev. and R. Antisense Research and Applications. Martin. Clin. L. T . Wang. Yoneda. M. Campbell. F. Burckin. Hum. 531-538 (1992). S. M. M. S. and P. 1045-1048 (1993). B. M. Chan. Med. Pal. and C. M. 641-646 (1994). Jiang. L. L.379-388 (1997). Mumbauer. R. E. B. 427-449. Hoke. and R. Kopp. J. L. Shea. Antisense Res. Zuckerman. Sezaki. Shea. Tivel. K.. Trainor. S. I. A. and P.44-53 (1995).V . Z. Chrisey. Levin.. Sasmor. Bruggeman. USA. Graham. A. and A. A. Syndr. Crooke. T . M. pp. CRC Press. Geary. Gorey-Feret. Liu. L. and S. Arnin.923-937 (1996). Gorenstein. Rappaport. Barr. Sasmor. W .429-438 (1996). S. Mirabelli. Stecker. J. A. J. Conklin. K. Invest. L. M. Crooke. Cooper. J. Cossum. A. Sasmor. J. Immune Defic. R. M. M. Proc. Graham. P. P. M. S. Lemonidis. 269. Cummins. and A.rences C. S. Exp. and D. 267. T . Endres. Greig. Chidester. 3. Coffman. L. I. Anticancer Drug Des. Diasio. Y u . J. 471-492. 21. 1993. J.. Crooke. Hanss. Pharmacol. J.. G. S. T . 286. S. FL.Y . Acad. D. F. Exp. Ecker. 2. 531-538 (1991). B. Leeds. Markham.115-121 (1995). M. and E. Dellinger.. Biophys. Tidd. Bishop. R. D.. M. S. 6. Fratkin. Nair. Exp. Krug. pp. R. D. Nucleic Acids Res. F. Truong. Y . Cossum. Levin. pp. Mahato.. Mol. Akiyama. Invest. Schiavone. Schechter. R.. Z . T . and S. Neckers. 37. H. Antisense Res. Hashida. F. Crooke. J. J. C. A. Retrovirol. J. J. 19. M. Crooke.. A. J. T . M. A. Antisense Res. H. M. C. Waller-Fontaine. Shea. Agrawal. L. 3.

New York. McKay. Neckers. M.528-533 (1995). T. 59. L. C. Muller. Fredriksson and U.238-244 (1978). Shanahan. Lemoine. 234. Wagner. Regul. J. 221. B. 236. Bennett. M. 216. and R. A.. 223. A. Tillman. Kisner. C. Simons. M. T. 217. W. Nielsen. Koob. Fisher. Arthritis Rheum. Antisense Technology: Principles. Nut. Herlyn. 212. and B. and R. K. S.4051-4055 (1995). 245. Dev.-W. Ross. Weinbach. G. Acad. New York. Exp. D. Nicklin. Lott. M. J. H. Staple. Ayala. Abe. J. A. 773-794. A. A. FL. K. F. S. L. Monia. 6. Doan. M. Hardee. L. T. 16-24 (1994). Wagner. A. Kramer. and W. Nature. Rosen.. 1993. Bennett. K. R... 725-738. Hu. Washington. F. USA. S. DC. S 92. Froehler. Graham. B. Burch and L. Sarmiento. L. Higgins. Catsicas. M. Bennett. and K. M. and B. Yee. Strategies and Applications. Science.. 259. and C. 238..528-531 (1993). F. Miki. Rosolen. Strategies and Applications. Geiger. Bowen-Yacyshyn. P. 232. Butler. Science. CRC Press. 1149-1156 (1997). Antisense Therapeutics: Principles. and Y. Zhang. 291-318. Cummins. C. G.. 104. USA. I. A. J. J. G. Nyce. G. Levin. G.9901-9905 (1993). Shanahan. Ther. 239. Marcel Dekker. 2001. Weiss. W. Langer. L. J. S. DeKeyser. B.USA. F. T. Crooke. J. Invest. M. Sasmor. E. N. Nature. Aberg. 36. S170 (1999). 271. R. R. Perez. 171-182 (1995). J. F. W. R. Neckers. J..34993507 (1996). Whitesell. M. Invest. c i . 215. J. and A. Taguchi. M. B. D. J. Lewis. 237. A. K. Expert Opin. M. Antisense Technology: Principles. Barton and N. and L. Cooper. C. Boca Raton.. Antisense Research and Applications. Y. 91. 222. Shanahan. Martin. Jones.1133-1142 (1998). I. Biophys. E. A. Strategies and Applications. J. pp. R. J. Yacyshyn. J. D. Tami. Cancer. Pierce. and S. 243.. Tami. Zon. Gastroenterology. 59. D. M.445-448 (1993). J. Nyce and W. T. Hijiya. K.Oligonucleotide Therapeutics 211. P. Med. Qian. Pettersson. R. Br. Holstad. Axelsson. 91. Sci. J. Marcel Dekker. 240. K. Strategies and Applications. R. G. G. Giddings. 2. CRC Press. 668-675 (1995). C. H. M. 218. Derrnatologica. A.-P. Acad. N. Yacyshyn. Knowles. DeRiel. J. 71. Shanahan. S. F. Okazaki. C. P. P. Zhou. 233. Herrera. Chern.. 231. Natl. 679-685. New York. A. pp. and M. F. 2001. Johnston.1462-1470 (1994). Bready. M. Drugs. Proc.A. 241. Kurokawa. L. Z.813-818 (1995). M. D. R. Catsicas. Dean. W. Regul. Monia. Burgess. J. Ratajczak. D. Rosenberg. L. A. Proceedings of the Annual Meetings of American Gastroenterological Association and American Association for the Study of Liver Diseases. Miraglia. M. Derrnatol.. K. Lima and S. and D. Fabbro. C. Proc. I. S. Edelman. Res. 229. 149-187. Marcel Dekker. A.11762-11766 (1994). 230. Cancer Res. Templin. Kumada. Antisense Drug Technology: Principles. B. Pharrnacotherapy. 13. pp. 246. Pept. Geary. Cohen. McKay. Acad. K. L. 198. M. 67-70 (1992). Antisense Nucleic Acid Drug Dev. 157. Boca Raton. Metzger. P. Pich. J. Biochemistry. Fabbro. Bennett and S. W. Zhang. 88. 114. 271. and W. Nyce. L. F. Hertl. A.1190-1196 (1991). B. E. McCarthy. Natl. Biochern. M. Matteucci. Katz. Farrell. McComas. Mahan. 795-832. L. M. 390398 (1997). M. Gewirtz. T. pp. Kisner. Jr. Proc. Grenningloh. 721-725 (1997). V. Sci. Wahlestedt. Marcel Dekker. M. 242. C. Pept. L. M. Dorshkind. F. Bennett. 364. R. G. Williams. P. Antisense Res. Natl. 1453314540 (1996). FL. Biol. Blackburn. M. 244. W. W. T. V. Med. L. S.. Proc. 2001. &ad. Gyurko. J . and D. and L. Phillips. N. D. J. 214. S. F. Johnston.-L. L. Osen-Sand. New York. Grillone. R. Nesterova and Y. and M.177-188 (1998).44994503 (1994). E. T. and W. 10. 163-170 (1995). M. 225. Arnbuhl. Kant.359-368 (2000). W. and S. F.429-437 (1995). and L. Takuwa. C. 219. 235. J. J. 1510-1513 (1993). 1997. 220. New York. W. 227. Maksymowych. 213. Coleman. J. B. Woloschuk. Goldman. 228. 1. M. L.Pharrnacol. Dean and R. E. 1. Meister. B. 260. Moulds. pp. F. Strategies and Applications. Ziel. Cho-Chung. R. J. H. Martini. F. J. Bayewitch. U. pp. 385. 90. pp. Isenberg. . Antisense Technology: Principles. J. 2001. F. Geiger. 171-193. Howard. F. Takuwa. K. Tami. Nagel. Clin. M. Marcel Dekkey. Gutierrez. Yacyshyn. 42. Nut. Jewell. J. 58. Narayan. Zhou. USA. Simonet. Heilig. Hutchings. Cornrnun. Muller. and D. R. L. 1996. 359. 2001. and C. Natl. Therapeutic Modulation Cytokines. Nature. Cook. M. Invest. Sci. Crooke. J. K. 226. D. R. M. 224. 343350 (1991).. N.

Toxicology. P. S. J. Ognoskie. and B. Pharmacol. Bayever. J. L. Clin. Vasilevskaya. and K. J . Kornbrust. R. B. A. W . J. D. Vasilevskaya. S. R. M . 281. J. A. 1128-1131 (1992). Sandler. 92. Geary. 739-759. 250. P.. New York. and D. and B. Advani. Richards. Raynaud. and W . 3977-3982 (1999). Nicklin.. Antisense Research and Application. Cunningham. A. Tidmarsh. 395-408 (1997). Holmlund. S. Clin. Geary. Krieg.3977-3982 (1999).. 17. J. Cassella. Monia. T . Ther. 374.Wallace. J . L.-S. J. Kornbrust. J. M.. Bruce. 16. E. Nemunaitis. and A. Kwoh. K. T . T .. J. R. Clin. Monia. 365-375 (1998). R. J. J. J. J. O'Dwyer. 4. Pisetsky. R. 252.. Stevenson. 1626-1631 (2000). Dorr. Butler. Gallagher. I. D. Messina. Oncol. Henry. and S. C. Crooke. M. K. Mitchell. J. and P. J. J. Ther. Henry. A. Schechter. Halsey. 5. 1999. J. Clin. Gilkeson. R. C. P.. G. A. 5. 3. J. T . J. P. T . S. and D. Holmlund. Hobson.. Teasdale. P. Clin. Monia. G. S. Kwoh. D. Craig.. J. 18. G. Eckhardt. Bolte. Waldschmidt. F. 12. Nature. T . L.. P. Matson.G. A. and T . M. Kuramoto. A. Oncol. Sikic. Johnston. Novotny. Jpn. Fisher. J. B. Nemunaitis. J . H. 2001. Kornbrust. Crooke. Leeds. Fife. T . Pharmacol. Grillone. L. J. Levin. Becker. J . Smith. M. Cancer Res. A. Stevenson.Y . Dorr. M. Monia. B.and S. B. Monia. GA. P.. C. A. A. J. M . Friedland. A. and D.1306-1312 (1996). A. C. and E. Cossum. 641-646 (1994). G. Marcel Dekker. P. Kraynak. Sci. P. J. Dorr. G. Barr. M. R. Y i . Dorr. Kataoka.3271-3276 (1993). .Yano. 54. Nemunaitis. Yuen. W . L.-K. J. Pangburn. J. Giclas. Auletta. A. Cornish. Dorr. and P. R. Phillips. Perez. Toole. Henry.3586-3595 (1999). L. M. L. Novotny. Kwoh. I. and J. pp. and D. and E. J. Kwoh. Toxicol. and G. V . 101-107 (1994). J. D. BerlinIHeidelberg. Yamamoto. Cassella. Holmlund. A. S. di Stefano. Proceedings of the American Society of Clinical Oncology. J. J . J.. 99-107 (1994). A. Levin. 278. Johnston. Bayley. 11. J. 0 ' Dwyer. 17. Iversen. J. Clin Cancer Res. Sikic. S. and L. R. M. Halsey. Commun. Kwoh. B. Schiller. Larkin. Geary. C. Baba. J. Nicklin. and M. Sledge. A. J. T . P. T .T . Strategies and Applications.. J. A. Gallagher. Jansen. Geary. 7. C. and K. Monteith. Graham. Galbraith. Ther. Lum. S. J. Fisher. T . Dorr.. 3357-3363 (1999). Life Sci. 57. 59. Lan. Antisense Technology: Principles. San Francisco. P.201-206 (1994). Makino. Henry. Klinman. S. Atlanta. N. 2227255. A. wacheck. A. G. T . D.-S. Dorr.16171625 (1998). Cunningham. Auletta.. Crooke. M. Monteith. Pisetsky and C. 261. R. J. T. Henry. M. Levin. B. Levin. Proceedings of the 35th Annual Meeting of the American Society of Clinical Oncology. M. Mitchelson. F. Holmlund. 248. D. J. Pharmacol. Battiato. 1759-1764 (1991). 120. Blood. Giclas. J. F. M. K. Cancer Res. L. Kwoh. E. 85-93 (1996). Arneson. Anticancer Drug Des. Appl. O'Dwyer. Yao. Dorr.. Leeds. Agrawal.-C. and D. W . Sheehan and H. S. 81. Gordon. and R. Monia. 239-247 (1993). 0 . P. P. 1999. 56. 1998. Waters. 5. A. J. and A. D. 145-155 (1997). Clin. Kimura. J. A. M. 6. Auletta. F. Garrett. 253. R. K. Blood. Yao. S351-S490 (1994).503-510 (1997).247. I. Clin.. M. Geary. Bock. S. Fratkin. W . 249. CancerRes. Yuen. Holmlund. R. A. J. A. 58. A. and D. Clarke. E. F. S. Leeds. Leeds. 1812-1823 (2000). Springer-Verlag. 2001. 251. K. Atlanta. Koretzky. and S.810-816 (1997). 260. and J. B. Monteith. W . 46. S. D.. Holmlund. Narayanan... J. Proceedings o f AACR. 140. Bishop. A. Geary. Toxicol. Bazemore. Reich. Tokunaga. Bruce. J. P.V . M. A. A. CA. Kornbrust. J. Heere-Ress. Zmmunol. Proceedings o f the 35th Annual Meeting of the American Society o f Clinical Oncology. Geary. A. S. 1145-1153 (1997). B. 83. D. T . Griffin. M. H. Sarmiento. M. Cancer Res. Deu. and F.546549 (1995). Res. 254. Cooke. Stevenson. C. Henry. R. P. 169-214. Yamamoto. Antisense Nucleic Acid Drug Deu. Leung. U. V o n Hoff. S. Holmlund. S. R. P. 2000. R. Exp. Holmlund. S. Pharm. Deu. 56. J . T . Kwoh. Antisense Res. Tendolkar. Holmlund. Pharmacol. Exp. J. 147. P. Templin. Gallagher. Cotter. Cassella. Leeds. P. J . Dorr. J. P. 4. J.. P. Yao. Ambler. pp. GA. Holmlund. and B. J . Nucleosides Nucleotides. Kornbrust.Yu. Geary. D. M. Oncol. Pharmacol. P. Antisense Res. A. J. Webb. F. A. A.

M. Chem. L. Ognoskie. Chem. Wagner.10373-10383 (1989). Manoharan. J.5540-5544 (1994). Matteucci. J. M. and P. 944-953 (2000). and B. H. J. Castro.. 285. T. Marcel Dekker. M. 286.Oligonucleotide Therapeutics 279. S. J. 1998. Holmlund. E. Am. Kwoh. 98. Strategies and Applications. Sekine. E. Antisense Nucleic Acid Drug Dev. R. Von Hoff. M. 297. Dorr. 17. D. Zounes. and A. Smith. Weeranta. Seio.. Pribble. Sproat. V. Nemunaitis. J. Geary. F. and B. Kraynak. Clin..-Y. 273-288. J. 38. 37. Soc. Bioorg. Bharadwaj. Antisense Research and Application. J. Johnston. 281. Lett. K. S. 1993. Testa. F. W. T. J. J. L. Chem. and M. Berlinmeidelberg. Chen. P. D. Med. R. A. 117. T. 7955-7976 (1998). D. J. 35. 298. 294. Flanagan. J. and R.3873-3874 (1995). Nemunaitis. D. Chem. 243-262.3586-3595 (1999). Wagner. 733-738 (1991). 308. Boca Raton. M. pp. Beijer. 303.. and I. D. R. I. 21. T. Seela. Manoharan. M. D. Mohan. W. G. Manoharan. G. New York. D. G. M. 164. Halsey. H. Davis.. K. Oncol. 310.7119-7128 (1985). J. Y. 16655-16662 (1999). J. A. Cook. S. 1998. Nucleosides Nucleotides. 17. K. Matteucci. D. CO. 13. D. M. H. 300. 26.8531-8532 (1998). 304. 295. A. Ohtsuka. 1119-1129 (1996).. 305. Toxicologist. pp. and 0. Acevedo. 2001. R. H. 293. J. Tygrett. Nut. Nucleic Acids Res. Lipsky. A. J. L. Templin. Turner. Brown. Glover. S. Monia. 312. P.3350-3357 (1998). 283. C. J. Ognoskie. Liang. Springer. and A. Immunol. 76757678 (1996). Nucleic Acids Res.. Antisense Technology: Principles. T. J. New York. J. Antisense Research and Applications. W. and M. 31973203 (1993). Mohhn. J. Lin. Pisetsky.. Disney. Y. Soc. Harmann and A.. M. 3513-3518 (1999). and M. 32. Lin. Eckhardt. 54. 1272-1280 (1997). Lemonidis-Farrar. P. Wang. Matteucci.. T. and D. Nikiforov and B. Krieg. 303-349. E. 311. M. 290. Sci.. Clin. F. Kwoh. Denver. FL. 292. A. G. D. Yuen. Oncol. . T. and J. K. H.. B. M. 1993. M. N. D. Freier. and S. 291. 19. Waldschmidt. Am.. J. L. A. D. Tetrahedron. Cummins. J. Lesnik. Kwoh. 42. Dorr. Matteucci.395-408 (1997). Olson. Dorr. K.. M. Nucleic Acids Res. H. Y. Nucleic Acids Res.3586-3595 (1999). Antisense Research and Applications. and E. 282. Schorr. J. K. Geary. Crocke. V. M. J. A. Gonzalez. Chem. pp. Balow. Bharadwaj. Geary. Freier. Chem. and S. Boca Raton.. Matteucci. 302. Am. M. Herdewijn. S. pp. 17. 17.. and P. Krieg. J. Aboul-ela. Cook.. FL. Lin and D. Schiller. 2001. Sanghvi. 307. K. N. N. Sikic. 13.17951798 (2000). CRC Press. Terhorst. Proc. Med. J. 215-218 (1994). A. 5. D. Turner. and P. USA. Ramzaeva. Manoharan. R. Antisense Technology: Principles. I. F. Biotechnol. Wolf. M. Monia. Connolly. 120. M.. 301. M. H. 1998. D. J . Froehler. K. Am. A. M. Symons. Imura. 18. A.-Y. J. D. Henry. Sanghvi. Dorr. Soc. K. D. Kreig. Marcel Dekker. Eckhardt. Reich. 113. D. M. R. C. L. M. 391-470. Cummins Lendell. 12. 280. and P. 116. D. and M. M. Kierzek. Shohda. R.Anticancer Drug Des. Ramasamy. J. Immunol. Richards. T. P. 10.30493052 (1995). D. D. R. Kisner. 96. 314. Gonzalez. A. K. Grant. S OC. J. Levin. Vickers. Nemunaitis. and A. A. M. Nucleic Acids Res. Los Angeles. and M. 313. Nishioka. Kraynak. T. G. H. 284. Bioorg. Zounes. S. 4313-4322 (1991). CA. M. 1203-1204 (1999). S. Cook. H. Iribarren. 160. R. A.-Y. 120 (2000). R. S. 289. P. 297-310 (2000). C. 269-290.. Bruce. A. and Y. Wada. F. Bruce.. B.. R. Monteith. S. Tetrahedron Lett. G. Lin. Von Hoff. and D. Okamoto. 48-52 (1999). Natl. 296. Dorr. R. 287. A. Hoke.. T. M.. R. Davis. S. P. Tetrahedron Lett. Jr. Ross. Acad. Proceedings of the American Society of Clinical Oncology. Cook. F. Tinoco. Lesnik. Lin and M. R. Holmlund. 309. B.. J. C. Sasmor. Biochemistry. Springer-Verlag.8927-8938 (1985). C. Grant. Holmlund. Graham. and R. Kierzek. Y. and P. pp. A. Richards. M. S. Strategies and Applications. W. Griffey. R.870-876 (1998). Tetrahedron Lett. Martin. J. Rarnasamy. Invest. 288. 306. Kumar and D. B. Nucleic Acids Res. Gutierrez. An. 25. P. 10. Inoue. K. SantaLucia. T. Clin. Jones. R. H. CRC Press. Flanagan. R. Nucleic Acids Res.3851-3854 (1991).. Fisher. G. Garcia. 299. Proceedings of the 33rd Annual Meeting of the American Society of Clinical Oncology. J. Lett.

326. Andrade. Maier. Lamond and B.45-46 (1999). Amirkhanov. 348.. Lett. Org. Agrawal. Reed. E. 320. M. Ecker. Owens. R. Monia. Dev. Pieles. S. Habus. M. Tetrahedron. J. Zhang.. L. D. Boutorine. H. K. S.. Fishback. Weiler. J. and R. Yu. 39. Freier. J. 346. A. K. and G. Jiang. Manoharan. B. A. Res. and P. N. 545-556 (1995). 340. Vichier-Guerre. 122. Biophys. J. R. Manoharan. Manoharan. Chaix. 2. A. J. D. Shuto. R. Ossipov. E. Kessler. and R. and P. J. Nucleosides Nucleotides. S. and T. Bioorg. Deu. T. and R. Miyagoshi. American Chemical Society. B. FL. and S. 339.593-606 (2001). Froehler. Padmapriya. T. Uddin. and B. Saison. A.. 24. Yan. E. I. S. K.l-14 (1997). 2. 2422-2432 (2000). and B. E. D. J. Manoharan. Saison-Behmoaras. Johnson. C. R. M. Seliger. Moser.References 315. G. L. R. Wheeler. USA. J. in press. 332. and S. 20. Tang. A. X. A. Ueno. P. Lefebvre. P. B. 4717-4736 (1987). Ser. I. M. 4.-J. FEBS Lett. Tetrahedron Lett.. Freier. Frank. Schmitt. G. M. Imbach. A. Jiang.. C. N. H. 349. Lu. and P.. K. Haner. J. B. Blocker. 327. 1. P. Adams. A. P. M. K. A. F. J. L. 351-362. Lett. R. 329. Y. Kanazaki. 303-349.571-576 (1995). D. Damha. J. M. 9-18 (1994). G. Med. Antisense Res. Manoharan. M. Jr. Carbohydrate Modifications in Antisense Research. P. CRC Press. Parmentier. Cook. Boca Raton. Mori. Luu. Med. G. A. Guzaev.5361-5368 (1994). D. Lu. Pharmacol. K. Org. Guinosso. Acta. J. M. B. D. and C. P. J.. Sci. Biophys. Vickers. M. 344. S. C. A. Cai. J. 50. K. Biochim. Chem. Bennett. 33. Shaw. E. Sproat. Borkow. 50. S.307-314 (1992). Chem. Greig. 338. 404-410 (1995).. K. and L. Gall. Cazenave. Pompon. Kubert. G. Z. Martin. Cook. R.. Soc. J. 6. H"usken. Bird.185-190 (1994). H. Pradeepkuman. Chim. Desjardins. 1819-1822 (2000). FL. 324.. Crooke. K. Tetrahedron. Sanghvi and P. Z. W. Yu. Diasio. I. 12. X. M. Proc.. R. Biochemistry. Acta. H. J. Brown. U. R. P. 1993. Mata. 318. S. Am. Damha. Rojko. M. Antisense Res. Roesch. Antisense Research and Applications. Tivel. J. J.827-832 (1996). Acad. . J . Helene. Cox. 325. E. 331. B. H.. Akagi. 7478-7486 (1998). Marfey. E. P. T.. C. Zhao. Nucleic Acids Res. Nucleic Acids Symp. 83. Zhao. R. Gamper. and H. 366-374 (1995). Martin. S. and P. 5100-5109 (1996). S. pp. 78. Moser. Toulme. Temsamani. N. Hall. 37463750 (1986). Ortigao. Haener. and H. D. R. C. 7877-7885 (1994). M. P. B. Biol. T. H. T. and S. and J. 328. 57.Yumoto. Zamaratski. J. Chem. A. L. 325. Teramichi. Meyer. L. Iyer. 747-750 (1991). Zhang. Chem. Chem. Antisense Research and Applications. Marky. 343. A. 342. Wancewicz. 1053-1060 (1994). I. 1489. I. 345. Iyer. F. Zhang.. and S. Eur. R. J. 1489. Cummins. Manoharan. J. S. Zhang.123-127 (1993). J. Zon. Cook.2. Anticancer Drug Des. ACS Symposium Series580. W. 350. pp.. E. Drug Target. and C. Cook. A. H. Owens. and J. Loreau. Vasquez. D. Cowsert. S. 1993. Lesnik.-P. A. P. 117-130 (1999). 323. H. N. M. F. Biochim. Graham. M. V. 334. 277-284 (1993). 21. and J..477485 (1995). L. Scholler. D.. Diasio. 10. Med. 3647-3650 (1995). D. D. Hall. M. 19. 333. A. Nucleic Acids Res. M. 321. Ross. Wasner. Lett. and A. G. A. 50. P. L. Chem. D. Kent. T. Urata. Huet. Harman. A. 273 (1991). and M. and M. Arion. Habus. J. 1994. Biochem. CRC Press. I. Sproat and A. Pharmacol. Parniak. 36. D. J. Bioorg. Noronha. Agrawal. 341. 28. Henry.469-477 (1991). 145-150 (1993). E. Matsuda. R. Biochem. C. Breslauer. Iyer. 486-504 (1995). Tinsley. J. Amirkhanov. A. J. 42. 3995-3998 (2000).. 316. T. Chattopadhyaya. Biochem. Tetrahedron Lett. Agrawal. 319. DC.. Acta. Samaratski. J. Iversen. Zendegui. Dolle. M. 15..37. Kitchen. 117-130 (1999). Zuckerman. Sasmor. Natl. Boutorine. J . Manoharan. 347. H. and P. Tamburin. Nucleic Acids Res. 4. S. V. Boca Raton. R. Z. Hogen. S. 336. and M. Vasquez. 337. and M. Chattopadhyaya. 322. Griffey. S. Nucleic Acids Res. Z. H.. Mohan. M. Y. D. Virosco. A. Thuong. H. Helene.. Agrawal. J.. Nucleic Acid Ther. Lamond. Prakash. E. Godard..M. M. 317. H.. Giannaris. De Mesmaeker. Lett. 3. Froehlich. D. Acc. P. Manalili. Kawasaki. Biochem. J. Lesnik. Washington. 335.. Montenarh. H. 330. Helv. 2.

P. 1-7 (2001). Y u . M. 1994. Graham. C. Damha. N.5633-5638 (2000). A. Monia. Strategies and Applications. L. J.535-539 (1999). Exp. and A. and M . F. Lok. 39. P.898-904 (2001). A. Soc. A. B. and J . New York. SpringerVerlag. J. Krishnamurthy. B. 18. Antisense Technology: Principles. Salmi. Jakobsen. Kela. Exp. 78. Skouv. Z. Geary. Sasmor. Moore. J. Viazovkina. M. P. H. 296. Henry..468-479 (2000). S. E. Pharmacol. J. Watanabe. 355. R. 1998. Wengel. Agrawal. T . 656-661 (2000).Oligonucleotide Therapeutics 351. pp. J. Conklin. Strategies and Applications. Exp. M. D.. R. C.. Jiang... Carbohydrate Modifications in Antisense Research. J. L. Res.. J. M . K. New York. C. M. H. Diasio. et al. pp. Barascut. 517-539. American Chemical Society. and M . M. Bennett. Bellon. Nickel. Sanghvi and P. Kole and D. Pharmacol. and N. B. Liu. B. BerlinIHeidelberg. Lu. M . S. Karras. Truong. Froeyen. and J. 352. 2001. B. L. Minton. Broberger. Rosemeyer. Nut. 369. C.Khatsenko. Chem. Pitsch. Koshkin. Inamati. 361. . A. I. Sasmor. S. 301-310 (1999). M . Stepkowski. B. and M.-N. 155-182. T . J. Parniak. and C. R. New York. V . K. Strategies and Applications. A.. Natl. N. DC. H. Wang. 97. Stecker. K.929-939 (1995). Y . J. Egli. Chen. 372. R. M . Wang. Jacobsen. Minasov. 68-79. USA. 2001. C. P. Ossipov. Lai. Marcel Dekker.. S. Borkow. T . Johnston. Stecker. E. DC. Lescrinier. Washington. New York. A. J. Am. E. C. American Chemical Society. 141-168. Sioufi. Monia. C. V . 0. Arion. 356. pp. T . Lu. D. F. Wilds. R. Craig. Myers. S. Qu. Acc. 7050-7062 (2000). 70. 362. Good. Jaun. Wancewicz. Ther. Washington. Butler. Damha. 360. 319338. Leydier. 49. 6. W . H . Levin. Z. Ther. I. Seela. G. 122. and B. Monteith. M. Biol. Brooks. S. Johnsson.. 366. Miyake. Exp. Cook. Geary. pp. H. Chi. 296. A. pp. Zhang. C. Transplantation. S. P. Galbraith. Noronha. Herdewijn. Pharmacol. E. J. Marcel Dekker. J. Nicklin. Wilds. T . Luyten. A. Nut. 32. and M . 354. A. and M . L. Chem. 365. M . Ther. and M. Dean. 1-28. Proc. Parniak. Crooke. Hokfelt. Bunker. 40. Truong. J. Gleave. Manoharaon. Struct. Wendeborn. F. Watt. Ren. Acad. V a n Aerschot. Bolli. K. Pharmacol. Cook. 357. K. S. Crooke and M . L. M . D. Lesnik. Am. Wahlestedt. 363. 367. A. Biochem. 371. A.Carbohydrate Modifications i n Antisense Research. 364. 7853-7859 (2001). J. M . A. Antisense Research and Application. Porreca. T . Cooper. W . and P. G. S. J. Soc. pp. S. 359. Arion. 2001. H. J. Tian. 2001. D. Marcel Dekker. ACS Symposium Series580. Freier. G. Monia. A. R. S. K. Manoharan. J . Noronha. Chem. Ther. Helv. Cook. Oerum.1621-1635 (1995). F.8595-8602 (2000). Schloenvogt. 353. Hendrix.. Lesueur. Biochemistry. A. Pharmacol. Acta. B. Bennett. M. Zhang. J. Biochemistry. F. M . M . Crooke. J. 370. 358. 1994. T . P. 292. S. Sengel. I. K. Brukner. Z. Tereshko. Dean. Maier. X. 12976-12977 (1998). Conklin. L. Phillips. C. Sci. Teplova. N. Strategies and Applications. and S. M . Manoharan. Manoharan. Antisense Technology: Principles. R. Antisense Technology: Principles.862-867 (2000).890-897 (2001). B. Antisense Technology: Principles. J. and A. M. Mercatante.. Zellweger.Verbeure. Biotechnol. K. 298. Levin.-E. B. D. L. Chim. M . R. R. 120. Burchkin. P. Marcel Dekker. G. 368. and J. D. T. Holzner.

193 9 Lessons Learned from RibosomeIDrug Complexes. 177 5.3 Refined View of Macrolide Binding. 170 4. 170 4.1 Mechanism of Action. 169 4 Streptomycin. 189 8. 188 8. 198 . 186 8 Erythromycin and Analogs. 171 5 Neomycin and Gentamicin Type Aminoglycosides.1 Mechanism of Action. 185 7.5 Subclass. 188 8. 168 2 RNA as a Drug Target. 176 5.2 Structure-Activity Relationship. 177 5.1 Mechanism of Action.6 Subclass. 194 10 Acknowledgments.2 Structure-Activity Relationship. 185 7.1 Mechanism of Action.2 Structure-Activity Relationship of Rings I and I1 of Both Subclasses. 184 7 Tetracycline and Analogs.4 Structure-Activity Relationship of Ring I11 of the 4. 183 of Rings 1 5.2 Structure-Activity Relationship.3 Structure-Activity Relationship 1 1and IV of the 4. 168 3 Ribosome Structure and Function. 183 6 Aminoglycoside Toxicity.Contents 1 Introduction.

and initially served as a complement to the penicillins. the details of streptomycin binding were finally revealed with the report of the crystal structure of the streptomycinl30S complex (10). highthroughput screening. function. more than 50 years after the discovery of streptomycin. there have been several other structural rePo& of antibiotics that bind to ribosomal RNA. natural product screening. In addition to those studies. the design of inhibitors of protein function has been aided by the structural probing of inhibitor/protein complexes since the late 1960s. many of which are human therapeutics. These structures now provide the basic knowledge needed to explain the molecular function of those drugs and to design novel drugs similar to them. and a variety of drug-ribosomes complexes at good resolution (10-14). it was clear that streptomycin inhibited protein synthesis (171. Since then. and mechanism has continued unabated to the present day. which are useful against gram-positive infections (16). in complex with ribosomal subunits (10-14. The antibiotic streptomycin provides an excellent example of why the design of novel RNA binding therapeutics has lagged behind the drug discovery process in general. Study ofribosome structure. and in 2000. Its complexity has served as a tantalizing biochemical problem for hundreds of scientists over several decades and it is arguably the most studied enzyme (4). Identification of discrete ribosome particles occurred in the early 1950s. Some 25 years after structural biology was initially brought to bear on the problem of drug design. clear evidence that streptomycin interacted extensively with RNA was provided (191. but not until 1961 were suggestions made that streptomycin inhibited ribosome function (18). X-ray crystallography was being used to investigate enzymehnhibitor complexes a decade before that (2). streptomycin entered clinical use in 1949 as a treatment for gram-negative infections. The first RNAcontaining enzyme to be discovered was the ribosome in the early 1950s. Several questions remain to be answered. and definitive proof that ribosomes carried out protein synthesis came in 1955.Therapeutic Agents Acting on RNA Targets 1 INTRODUCTION Structure-aided drug design began in earnest in the 1970s. progress that has recently been punctuated with reports of numerous crystal structures of both subunits at atomic resolution (5-8). medicinal chemists have in hand similar toolkits with which to aid in their search for novel drugs that target RNA structure and function. and proteomics to identify drug candidates against existing and novel targets (3). Therefore. medium resolution structures of whole ribosomes (9). and lessons learned from those structures continue to contribute to the design of drugs in the present day. As early as 1948. Consider the maturity of drug design against protein targets compared with similar efforts against RNA targets. 2 RNA AS A DRUG TARGET Can drugs be found that specifically target RNA? For rRNA in the ribosome the answer is . The advent of recombinant DNA technologies and fast computers provided the last pieces of the infrastructure necessary to permit structural biology to play a regular role both in explaining drug action and in development of new drugs. Drugs that bind to RNA are half a centuljr old. (1)Do small ligands bind to RNA in fundamentally different ways than they do to proteins? (2)Are there enough drugiRNA complexes available to serve as a foundation for future drugs design? (3)Are there qualities unique to rRNA that qualify it as a drug target that would preclude other RNAs from being suitable drug targets? (4)Can any RNA motif confer sufficient specificity to be the sole in vivo target. Modern day drug discovery relies on structural biology and molecular modeling. Isolated and shown to have strong antibiotic properties (15). numerous enzymes were available in sufficient quantity for most of the 20th century to allow biochemists to design and test inhibitors of protein function. but the intentioned design of RNA-targeted drugs is new. genomics. An early example at that time Howwas against dihydrofolate reductase (1). 20). ever.In 1987.

Drug discovery efforts are underway against other viral targets as well (26). when a neomycin-specific binding element was engineered into the 5' untranslated region of a reported gene. is about twice the mass of the small ribosomal subunit. In E.3 Ribosome Structure and Function easy. except that they are slightly larger. it has recently been suggested that in bacteria mRNA can act as a direct sensor of small molecules such as thiamin to control their expression (31). Both subunits together have a molecular weight of about 2. they are all natural products that were discovered as fermentation products from soil microbes. In large part the success or failure of many of those analogs can be explained by the recent crystal structures of antibiotic/ribosome complexes. such as oncogenes. neomycin Blgentamicin-like aminoglycosides. Although no compound has yet been elevated to the status of clinical drug candidate. 16s rRNA. Once key drug-target interactions have been identified. and macrolides (e. For example. expression of mRNA that encodes for thiamin-synthesizing proteins is down-regulated by the presence of thiamin itself. which is 120 nucleotides in length. In this example. and 31 proteins (Ll-L31). In bacteria. Considerable effort is underway to characterize the function of the numerous mRNA structural elements known to exist and to identify others. tobramycin. whose velocity is 30s. the expression of that gene could be regulated by neomycin. whereas a similar aminoglycoside. Toxicology and pharmacokinetics of drug candidates are equally important. the large subunit. coli. whose sedimentation velocity is 50S. Some parts of . the spliceosome. binding remains the defining event in drug action. First.g.and it is integral to other RNAI protein enzymes (e. In theory. the 50s subunit is composed of two RNA molecules. in all cases except for streptomycin these efforts have led to clear improvements.. within untranslated regions of a given mRNA there exist numerous structured elements integral to the expression of that mRNA. It serves to preserve genetic information (retroviruses) or to transport it temporarily (mRNA). 3 RIBOSOME STRUCTURE AND FUNCTION Ribosomes are found in all cells and are responsible for translating mRNA into protein. it is a validated target that has been exploited therapeutically for more than 50 years. failed to affect gene expression (30). Other important considerations pertaining to these drugs are addressed elsewhere in this series and will not be addressed here (32). Second. and 23s rRNA. because they have more components and the average size of all components is larger. it serves as a co-factor in protein synthesis (tRNA). tetracyclines. which binds directly to those mRNAs. the expression of a particular mRNA could be regulated through the binding of drug-like compounds to those elements (27-29). erythromycin).5 x lo6 Da and are composed of roughly one-third protein and two-thirds RNA.g. Each of these drugs had similar beginnings. it is the core of the ribosome (rRNA).. The quest for drugs that specifically downregulate expression of individual genes. which is 2904 nucleotides in length. telomerase). 5 s rRNA. Ribosomes are composed of two subunits. Nevertheless. and 21 proteins (SlS21). is ongoing at the proof-of-principle stage. The 30s subunit (hereafter referred to as 30s) is composed of a single 1542-nucleotide RNA molecule. Exhaustive searches have been conducted for semisynthetic analogs with improved therapeutic qualities. Arguments in favor of targeting other RNA molecules have been made elsewhere and only the most salient points will be recapitulated here (21-23). The search for RNA-binding drugs that inhibit vital protein-RNA interactions in HIV replication has been underway for about a decade (25). This chapter will address the binding of four classes of drugs: streptomycin. RNA function in biology is diverse (24). Eukaryotic ribosomes are essentially identical to prokaryotic ribosomes. At least two pieces of evidence suggest that this is possible in vivo. to give a few examples. it is then possible to search for other analogs that retain those key interactions yet also satisfy other necessary drug requirements. of which binding affmity is only a single component. many structurally diverse compounds have been identified that bind tightly to their RNA targets. Drug development is a multi-factorial endeavor.

streptomycin binds tightly to a pocket involving nucleotides from four regions of 16s rRNA. transfer of the P-site tRNA to the E (or exit)-site. all have guanidino groups at the C1 and C3 position except for bluensomycin. which is linked to a streptose ring.and P-sites.361. Streptomycin binds to 30s in the ram state and consequently stabilizes that state. The streptomycin/30S complex has recently been solved by X-ray crystallography (10). Independent functions can be ascribed to the two subunits. 30s exists in one of two conformations termed the ram state and the restrictive state (10. which tend to be stripped-down versions of prokaryotic ribosomes. numerous cofactors are required for initiation. a common convention is to number the rRNA nucleotide sequence and helices according to those found in E. streptomycin was brought into service to treat Gram-negative and some Gram-positive infections. and termination (for a detailed description please see ref. After each round of peptide bond formation. which contains a carbamoyl group at the C1 position instead. or translocation. In the ram state. crosslinking. fast translation and hyper-accurate. a translocation step takes place that involves the movement of the mRNA through the ribosome. and mutagenesis data. which in turn is linked to an N-methylglucosamine. which results in errorprone translation and consequently the accumulation of miscoded proteins and the reduction of proteins with the correct sequence. Peptide bond formation takes place on the 50s subunit within the peptidyl transferase center. peptide bond formation within the peptidyl transferase center. and methyl alcohol can substitute for the methyl at the C4' position (34). . leading to hyper-accurate translation. chloroplasts. 4. tRNAs in the P-site and the A-site span both subunits and couple the two events. Certain eukaryotic organelles (mitochondria. elongation. and plastids) also have ribosomes.1. coli. Some variation is tolerated within the streptose ring: the aldehyde can be reduced to an alcohol. No matter the source. while others are not. it is still frequently used as a component of a multi-drug mixture to treat multi-drug resistant cases of tuberculosis. The mechanics of this two-state switch have been described (10). The ram state (or ribosomal ambiguity) was identified many years ago by correlating the presence of certain 16s rRNA and ribosomal protein mutations with error-prone translation. 6. nucleotides 912-910 in helix 27 of 16s rRNA base pair with nucleotides 885887.1 Mechanism of Action STREPTOMYCIN Streptomycin and other antibiotics that are close analogs (Fig. Mutations or ligands that stabilize the 912-9101 888-890 pairing restrict tRNA binding. whereas in the restrictive state 912-910 base pair with nucleotides 888-890. whereas decoding of the mRNA takes place on the 30s subunit within the decoding A. Conversely.1) all contain a streptamine aminocyclitol. Shortly after its discovery. mutations that support the restrictive state exhibit hyper-amate translation.Considerable modifications of the glucosamine moiety are also tolerated. The binding of streptomycin is stabilized by a variety of salt bridges and hydrogen bonds between streptomycin and backbone phosphates and through hydrogen bonds to Lys45 of protein S12. 4 its greatest contribution coming as the primary agent to treat tuberculosis (35). 33). Its use has waned over the years because bacterial resistance against streptomycin has emerged and less toxic agents have become available. A balance between sloppy. and that is the convention that will be followed here. and transfer of the A-site tRNA to the P-site. Of the 10 natural streptomycin-like compounds listed in Fig.Therapeutic Agents Acting on RNA Targets ribosome structure are highly conserved across all domains of life. slow translation must be struck for organisms to be successful. with Streptomycin binds to the 30s ribosomal subunit and disrupts normal protein synthesis both by weakening proofreading and by interfering with initial tRNA selection. As predicted by chemical protection. Mutations or ligands that stabilize the 912-910/885-887 pairing promote indiscriminant tRNA selection. Ribosome function is complex. 6. Nonetheless. Most antibiotics target one of the listed steps: decoding at either the A-site or the P-site.

Although no analogs of streptomycin exist in which the C5 position is altered. dihydrostreptomycin. Important measurements such as binding affinity for 3 0 s were rarely carried out.' but rings I1 and I11 (streptobiosamine) together are inactive. theobserved activity (or lack of activity) is assumed to result from a mechanism of action identical to that of streptomycin.2 Structure-Activity Relationship Initial understanding of streptomycin structure-activity relationships can be gleaned from comparing the activities of the component rings in groupings. most of the C1 guanidino modifications investigated are tolerated. The site most often exploited for streptomycin modification is the 3' aldehyde moiety of ring 11. However. Modifications of the C6 hydroxyl group are poorly tolerated. modifications would be tolerated. Synthetic analogs in which the C1 guanidino group has been replaced with 1-N-[(S)-4-amino-2-hydroxybutyryl] or 1-N-[(S)-4-guanidino-2-hydroxybutyryl] have been evaluated (38). methylation of the C1 guanidino group or substitution of that group with chemically similar moieties is usually tolerated. Phosphorylation. this in part explains the high tolerance of ring I11 for modifications (37). Figure 6.4. as in bluensomycin. The C1 guanidino group forms a salt bridge ate of C1490. in which this hydroxyl group forms hydrogen bonds with Lys45 of the ribosomal protein S12 and the phosphate of G1491. When this guais replaced with a carbamoyl oup. These compounds were tested on a variety of organisms. a universally conserved nucleotide. Methylation of 'Unless otherwise stated. Within the streptamine ring. if any. activity refers to the ability of an agent to stop bacterial growth in uitro. bluensomycin is about 10-fold streptomycin for most strepmycin-sensitive organisms. Overall. Figure 6.2b illustrates many of the aldehyde conversions . as there are no direct contacts between its hydroxyl group and the ribosome. In all cases. Phosphorylation or adenylation of the hydroxyl group at C6 abolishes activity (41). which cannot be altered. The consequences of modifying the hydroxyl group at either C2 or C5 are unInspection of the streptomycin/30S crystal structure and sequence comparisons of 16ssequences from numerous organisms generally explain these observations (10). This analog was later isolated from fermentation sources (Streptomyces humidus) (44). Therefore. These bonds clearly must be maintained for full activity. some of which were resistant to streptomycin.2 highlights permissible and nonpermissible alterations to streptomycin and analogs. the reported activity was lower than for dihydrostreptomycin (which has activity similar to streptomycin) and similar to or somewhat lower than for bluensomycin. forms a bifurcated hydrogen bond with 02' and 0 3 ' of A914. in general. contained an alcohol in place of the streptose aldehyde (43). any change to the guanidino group at C3 abolishes all activity (38-40). Examples described in the text illustrate that such assumptions are not always the C1 guanidino group is also tolerated and the streptomycin/30S crystal structure supports this. Methylation of the C3 guanidine would result in the disruption of one of these hydrogen bonds. however. adenylation. Rings I and I1 alone are sufficient for antibiotic activity. The 6-deoxydihydrostreptomycin analog was synthesized and tested in hopes of circumventing the action of C6. the reason is apparent from the streptomycin/30S structure. an analogous interction can take place and is therefore moder.o as active as dihydrostreptomycin (42). or removal of the C6 hydroxyl group causes a strong reduction in activity. the streptose ring. but this compound was only l/lo-l/. removal of that group greatly reduces activity (42). the structure of the complex suggests that very few. because a salt bridge to the phosphate of C1490 can be retained. The first semisynthetic analog of streptomycin. The C3 guanidino group. Similar restrictions are not expected for C2 modification. Other modifications would similarly disrupt the multiple hydrogen bonds involving the C3 guanidino group.OH modifying enzymes found in some streptomycin-resistant bacteria. Replacement of the C1 guanidino group with a 2-imidazolin-2-ylamino substituent was inactive (40). because that hydroxyl group is also involved in a hydrogen bond with Lys45 of S12.

The amino deriva- tive and short-chain alkylamine derivatives remained active up to the hexylamine analog.1. The heptaamine derivative was not investigated. as active as dihydrostreptoonly about 1 mycin. with surprising results (47-50). 46). This sharp inflection prompted one .6-a-D- Mannose 0-CO-NH2 NH-CNH-NH2 NH-CNH-NH2 CHO CHO CHO CH3 CH3 CH3 CH3 CH3 CH3 H H H Ashirnose CO-CH20H a-D-mannose1. Conversion of the aldehyde to an acid or reduction to a methyl abolishes all activity (45. However. the hexylamine derivative is / 1 . Longer alkylamines (octyl and above) were nearly as active as dihydrostreptomycin. which have been synthesized and tested. activity diminished with increasing chain length. . The structures of streptomycin and several natural analogs of streptomycin.6-a-DMannose AC4437 = 5'-hydroxystreptomycin lacking ring Ill Figure 6. The aldehyde oxygen forms a hydrogen bond with a phosphate oxygen of G527.Therapeutic Agents Acting on RNA Targets Streptomycin Dihydro- NH-CNH-NH2 NH-CNH-NH2 NH-CNH-NH2 NH-CNH-NH2 NH-CNH-NH2 CHO CH2-OH CHO CHO CHO CH3 CH3 CH2-OH CH3 CH2-OH CH3 CH3 CH3 H CH3 H H H H H H Streptomycin 5'-hydroxyH H Streptomycin N-DemethylStreptomycin Mannosidohydroxystreptomycin Dimannosidostreptomycin Bluensomycin Ashimycin A Ashimycin B H a-D-mannoside NH-CNH-NH2 CHO CH3 CH3 H a-D-mannose1. Numerous groups have investigated conversion of the aldehyde to its amino derivative and a variety of alkylamine derivatives.

Dashed arrow points to modifications that results in compounds that are active by an unknown mechanism. Dashed lines indicate possible hydrogen bonds (some of which are salt bridges when suitably reinforced with favorable electrostatic potentials).Adenylation Phosphorylation J All modifications Figure 6. (b) Ring 11. arrows with an X point to non-permissible modifications.2. (a)Ring I. Arrows point to permissible modifications. top) with various modifications tested for activity (bottom). coli numbering. (c) Ring 111. Molecular interactions between streptomycin and 30s (E. .

) .2.Therapeutic Agents Acting on RNA Targets Figure 6. (Continued.

phosphorylation a-D-mannose (=DM) a-DM-1.2.6-a-DM 2"'-carboxy-xylo-furanose (ashimose) Removal Figure 6. (Continued.Streptohydrazid was tested long before the mechanism of action of streptomycin was known (the mechanism of isoni- . was synthesized and found to be at least as active as combined therapy using both streptomycin and isoniazid (51). This compound termed streptohydrazid. another prominent anti-tuberculosis drug. the Another alkylamino derivative tested was a conjugate of streptomycin and isoniazid.) up to consider the mechanism of action of ggested that the short-chain alkylamine an- t ribosomally unrelated mechanism.

However. The logic is sound: both analogs should retain streptomycin-like binding affinity. 5 NEOMYCIN AND GENTAMKIN TYPE AMINOCLYCOSIDES Shortly after the discovery of streptomycin. suggesting that modifications at this position might be tolerated. for reasons that are not clear. As expected. The structure of paromomycin.Therapeutic Agents Acting on RNA Targets azid is not fully understood). neither of which are essential to activity (37. Two semisynthetic analogs. they never reached clinical status.) Neomycin B is the prototypical 4. Correlation of the streptomycinl30S crystal structure with the various streptomycin analogs that involve aldehyde modification suggests that only a hydrogen bond between the aldehyde oxygen and a protonated phosphate oxygen (or a salt bridge for the amino derivatives) must be maintained. Reduction in binding only occurs when the modification becomes too large to be accommodated within the binding pocket. (Although many antibiotics contain aminoglycoside rings and are often referred to as aminoglycosides. all of ring I11 can be dispensed with and cause only modest reductions in activity. in complex with rRNA. A cursory inspection of the streptomycinI30S structure suggests that the methyl group at C4' contributes little to ribosome binding. none of which abolish activity. because it is the most effective agent in this class. The gentamicins are the most commonly used 4.6-linked 2-deoxystreptamine aminoglycoside antibiotics were found in fermentation products of actinomycetes (Fig. Indeed. Several active natural streptomycin analogs. members of the 4.6-subclasses and will be abbreviated AG. Yet. Semisynthetic derivatives of this position are absent.5-linked AG and is composed of four rings. such as 5'-hydroxystreptomycin and AC4437. a close analog of neomycin B. but presumably it has streptomycin-like function.5-linked subclass. has been solved multiple times.and the 4. 54). The only exception to this is when streptomycin is oxidized to streptomycinic acid. the general term "aminoglycoside" will be used here to specifically refer to the 4. isoniazid-like function. were synthesized to circumvent common streptomycin resistance (53. The mechanism by which streptohydrazid works is not known. The two subclasses have their first two rings in common and work by the identical mechanism of binding to the decoding A-site of 30S. Other poorly conserved nucleotides within this loop also form part of the AG binding pocket and provide the basis for organismal specificity (58). it was reasoned that a conjugate of the two might act synergistically. once by NMR and twice by X-ray crystallography (10. or some combination of the two. Some streptomycin-resistant bacteria harbor genes that encode proteins that either phosphorylate or adenylate the C 3 hydroxyl group of streptomycin (41). 13. deleterious modifications to this ring are possible.56). A limited number of natural modifications occur at C2". thus causing misreading of mRNA (57). The 4. Although the possibility of forming the required hydrogen bond exists. are hydroxylated at C5' (34. Some glycosylations have been observed at C 4 in natural streptomycin analogs. 59). 6. these analogs worked well against common bacterial strains and better than streptomycin against many streptomycin resistant strains. yet remain active against bacteria that express enzymes that phosphylate and adenylate the C3".6-linked AGs and are composed of only three rings.3) (55. 3-epidihydrostreptomycin and 3"-deoxydihydrostreptomycin. 52). NMR was also used to solve the structure of gentamicin Cla in complex with its target RNA sequence (60).5. The orientation of binding is the same-rings I and I1 from . yet not be substrates for inactivating enzymes. the analog is inactive. 52). presumably because of the electrostatically unfavorable close approach of the carboxylic acid to a phosphate that would occur on binding. These structures clearly describe the important modes of binding for the 4. Biochemical probing has firmly established the binding of these agents to the major groove of an asymmetric loop composed in part of several absolutely conserved nucleotides (19).5-linked and 4.6-linked subclass constitutes the majority of the clinically useful agents. yielding somewhat less active antibiotics.OH. Ring I11 (glucosamine) makes two direct contacts with 30S.

w NH2 Gentamicin Cla (member of 4. however. peptide bond formation will occur regardless of whether or not the correct tRNA occupies the A-site. A1493. 5. The ultimate consequence of AG binding remains the same. In the correct presentation a helical minor groove stabilizes the "loopedout" conformation of A1492. because of alternate linkages that exist between rings I1 and 111.5 Neomycin and Centamicin Type Aminoglycosides / . Incorrect tRNAs do not elicit the same response and eventually diffuse away from the A-site with- Although the two subclasses of the A-site binding aminoglycosides are chemically distinct. Nucleotides A1492 and A1493. For both classes.1 Mechanism of Action out the formation of a new peptide bond.5 subclass) Figure 6. they do share rings I and I1 in common and up to that point bind 30s in identical ways (60). the conformation which serves as a signal for peptide bond formation to occur. this ring alone is insufficient for antibiotic activity.6 subclass (Fig. and G517. the tRNA and mRNA nucleotides involved in the codonlanticodon interaction form a regular helix conformation. Therefore. and consequently normal minor groove geometry. 5.6 subclass) Neomycin 6 (member o f 4.5 or 4. However.the remaining portions of the two subclasses take different trajectories. ring I of neomycin alone is minimally active as an antibiotic (41). Although this class is grouped by the inclusion of 2-deoxystreptamine. Sev- . When the correct tRNA occupies the A-site.3. Removal of ring I from all aminoglycosides results in inactive compounds. 6. it displaces A1492 and A1493 into the "looped-out" conformation. without regard for whether or not a particular modification was observed in the 4. ring I is the only essential ring. play pivotal roles in discriminating between cognate codon/anticodon interactions and non-cognate interactions. both subclasses bind identically. Therefore.6 AG binds to 16s rRNA.4a).2 Structure-Activity Relationship of Rings I and II of Both Subclasses The molecular details by which AGs cause miscoding were recently elucidated (61). when either a 4. 30s is no longer able to discriminate between cognate and noncognate tRNAs when the tRNA binds to the A-site (57). along with G517. discussion of ring I and ring I1 modifications of both subclasses will be treated at the same time.5 subclass or the 4. The structures o f ?omycinB and gentamicin Cla.

6 subclass. Replacement of C6'-OH with an m i n e or a short-chain alkylamine is tolerated or improves activity. (d) Paromomycin ring IV. era1 naturally occurring antibiotics contain only ring I and ring11 (paromamine and neamine are examples). The arrow with a ? points to a modification that was shown to be inactive. The ring oxygen and C6'-OH . but is not readily explained by the crystal structure. which when acetylated.6 linked aminoglycosides and 30s (E. Molecular interactions between 4.5 and 4. abolishes antibiotic activity (41). NHCH3 NHCH2CH20H / HO Phosphorylation H Figure 6. top) with various modifications tested for activity (bottom).4. Some AGs have an NH. (b) Paromomycin ring 11. (e) Gentamicin Cla ring 111. but all clinically active agents contain four rings for the 4.5 subclass and three rings for the 4. (a) Paromomycin ring I. Arrows point to permissible modifications. coli numbering. (c) Parornomycin ring 111. at C6'. arrows with an X point to non-permissible modifications.Dashed lines indicate possible hydrogen bonds (some of which are salt bridges when suitably reinforced with favorable electrostatic potentials).Therapeutic Agents Acting on RNA Targets ring I I - Stabilizing intramolecular interaction Acetylation CH3 ' u c .

An important observation is - that paromomycin remains active against protozoa despite the fact that in protozoa nucleotide 1408 is a guanosine. because neither of the original hydrogen bonds can be satisfied. G substitutes for A at this position. branching at .5 Neomycin and Centamicin Type Aminoglycosides various water-mediated contacts LIU Acetylation Figure 6. But an amino group at C6' (as in neomycin B) isn't active. an analogous pseudo-pairing isn't possible. and this the prokaryotic specificity of antibiotics (58). In eukaryotes.) (or C6'-NH2) form a psuedo-trans WatsonCrick pair with A1408 (13). With a G in "" position 1408. Finally.4. (Continued. because at least one hydrogen bond is preserved.

62). Because an amino at this position should make a productive salt bridge. All active aminoglycoside compounds have a hydroxyl at C4'. 3'-deoxy and 3'-epi analogs are not substrates for such resistance enzymes. Phosphorylation at this position abolishes all activity and forms a mode of resistance (41. because branching a t this position doesn't disrupt the pseudo-trans pairing to G1408 (41). in an effort to preserve the hydroxyl phosphate hydrogen bond while eliminating the possibility of inactivation by phosphorylating enzymes (63). derivative tested may be inactive for other reasons. hydroxyl groups at C4' form a hydrogen bond to the phosphate of A1493. or are unsaturated at C4' (41).4. the particular C4'. are C4'-deoxy. The methoxy derivative is not tolerated at C4' (411. at least up to one carbon. however. presumably because of steric clash with the phosphate of Al493. Recognition that all 3'-deoxyderivatives lose a productive interaction that presumably lowers their affinities for 16s rRNA prompted the design of 3' ketokanamycin A.Therapeutic Agents Acting on RNA Targets I I ring I I HO CI Phosphorylation OH - 3" 2" Figure 6. It is .NH. and they remain active against some bacteria resistant to aminoglycosides (41). The 3' hydroxyl group in paromomycin forms a hydrogen bond to the phosphate of A1492.) C6' is permissible. (Continued. An amino at the C4' position surprisingly abolishes activity.

However. it was more active than kanamycin A against E. therefore only the most illustrative examples will be described here. Although only a limited number of highly active derivatives were found. Shortening or lengthening the carbon chain by " more than one carbon unit drastically reduces activity.64. coli harboring the gene for AF'H(3')-Ia. Conversion of kanamycin A to include 4-amino-2(S)-hydroxylbutyrylamide (AHBA) at C1 yielded an aminoglycoside (amikacin) that was effective against many aminoglycoside resistant bacteria with little reduction in activity against aminoglycoside sensitive bacteria (41). clear structure-activity relationships of other acylated products are observed. which retains the same hydrogen bonding to A1492. a large variety of modifications are tolerated at C1-NH. as does moving the hydroxyl .Success with amikacin prompted an exhaustive search for other C1 derivatives with improved efficacy over parental aminoglycosides.) expected that the 3' keto derivative will exist in equilibrium with a hydrated variant (3'gemdiol analog) and serve as a substrate for phosphorylation.Although the keto analog was substantially less active than kanamycin A against E.ite by resistance enzymes [henJ the 2'-epi configuration a1bolbr con fjshesactivity (41.Using AHBA as a basis of comparison for all N-1 acyl modifications.62). coli. thereby regenerating the 3' ketokanamycin A analog. N-acylation of C1-NH. is the most explored type of modification at this position. Only hydroxyl and amino groups are toler. (Continued. (41. nis E. when expressed. is Ring I1structurally defines this class of anP %iioticsas the 2-deoxystreptamine aminogly% - cosides and has given way to modification more frequently than the other rings (Fig. 6.4b).6 subclass to stabilize the conforma$on of the antibiotics in the active state.ated at C2' because an intramolecular hydrogen bond must be satisfied between ring I and iring I11 of the 4. phosphorylates C3'-OH.65). It is impossible to catalogue all of the reported C1 modifications.4. Several important commercial aminoglycosides emerged from such efforts.5 Neomycin and Gentamicin Type Aminoglycosides Adenvlation t mannose-0 5"' epi Figure 6. In 60th subclasses other substitutions at the C2' !position would disrupt this hydrogen bond. a resistance gene which. the phosphorylated product would undergo spontaneous elimination of the phosphate moiety.5 subclass and ring I and ring Hofthe 4.

it is clear that AHBA makes highly selective. (Continued. (The observation that . yet unknown contacts with 30s. because the same moiety is found in the naturally occurring aminoglycoside butirosin and was successfully introduced into gentamicin B. Substituting the a-hydroxyl group with an amino group abolished all activity. The a-deoxy derivative retains partial activity. yielding the active compound isepamicin (41). With the exception of unsubstituted amidino and guanidino groups.Therapeutic Agents Acting on RNA Targets NH2 / / NHR (R is large) 0 Figure 6. any change to the terminal arnine generally reduces activity or abolishes activity altogether.4. The positive qualities of C1 AHBA modification may be universal to all 2-DOS containing aminoglycosides that bind in the decoding A-site pocket. Taken together.) group away from the a position or inverting the stereochemistry of the a carbon. whereas a fluorine replacement retains full activity.

S include a hydrogen bond with N7 of 1494 and stabilizing electrostatic interacons with the phosphates of G1494 and 1493. In h subclasses. which can be either an amino or a hydroxyl group (41). this doesn't preclude the 2" hydroxyl to amine substitution.6 subclass is an alternative to rings I11 and IV of the 4.3 Structure-Activity Relationship of Rings Ill and IV of the 4.4d) (41).g. lividomycin) (41). th AHBA-like alkane and C1-NH.. as do virtually all her modifications at this position (62). A co-strucof amikacin (or one of the other N-1 A aminoglycosides) bound to 30s or exnsive molecular modeling is required before ore insight into the positive attributes of A can be realized. Only two direct interactions between 30s and ring I11 of paromomycin are observed in the crystal structure: hydrogen bonds between the 5"-hydroxyl group and N7 of G1491 and between the 2"-hydroxyl group and N4 of C1407. and substitution at C2"' (411.however. 5. Connection to ring IV is made from C 3 of ring I11 to Cl'" of ring IV. (10). cetylation of C3-NH. Related efforts to find novel aminoglycoside analogs sought to replace both rings I11 and W. the observation of both stereoisomers at C5"' (41). The 3 amino group forms hydrogen . Replacement of ring IV with a short alkyldiamino tail yielded a compound with activity identical to that of neomycin B (68).6 s. dihyoxypropyl kanamycin analogs showed near efficacy despite being structurally dis(66. suggesting that this intramolecular raction isn't necessary for activity (41). direct interactions between C1-NH. The NMR structure of the gentamicin A complex reports a hydrogen bond beeen the C5 hydroxyl group and ring I.) Considerably fewer C1-NH.5 Subclass The chemical structure and the trajectories of ring I11 in each of the two subclasses diverge. Extensive modifications are possible within ring IV of the 4. In the 4. In the 4. this explains why so any modifications are easily tolerated at this ition. and it makes numerous direct contacts with 30s (Fig. 5. conversion to the deoxy ana(at either position C5 or C6) is tolerated ). so from this point the structure-activity relationships of the two subclasses will be treated separately. The 2" hydroxyl is within hydrogen bonding distance of 0 6 of G1405 and 0 4 of U1406. substitution of this type in her configuration is tolerated. The ogen bond that this amino group forms h U1406 is evidently a crucial interaction.4~). These interactions are presumably aintained in the above derivatives. which is permissible (41). For example.6 Subclass Ring I11of the 4. Group 4. suggesting that these latter groups are too ) large (Fig. by resistance enes abolishes activity. Chlorination or phosphorylation of C5" abolishes activity. which be satisfied by deoxy analogs of the 4. What isn't known is why some modtions are better than others.6 subclass ere is generally a free hydroxyl at C5.5-aminoglycosides are inactivated on adenylation of the C6"' amino group (Fig.5 Neomycin and Centamicin Type Aminoglycosides same modification is optimal in both the -subclass (butirosin) and in the 4. alkyl derivaes have been synthesized and tested for acvity. 6. The nueotides A1492 and A1493 are displaced on g binding. ring I11 is usually a ribose (41).5 subclass.6-subs (amikacin and isepamicin) gives further eight to the supposition that rings I and I1 of th subclasses bind in the same manner. Ring IV has only a few natural modifications: connection to mannose through C 4 (e. Conversion of the C6" amine to a hydroxyl is tolerated (68). whereas C5" deoxy and C5" amino derivatives are active. but it seems that considerable struca1 diversity can be tolerated. whereas for the 4. 67). e only information that exists for modificans at C2 regards substitution with a hyoxyl group (41).5 subclass there is generally a free oxyl at C6. 4 ~(41).5 subclass. 6. the best compounds only had intermediate activities between neamine (ring I and I1 only) and neomycin B (69). which provides a large pocket acent to C1-NH.5 subclass of aminoglycosides in agreement with the limited number of ribose contacts. 6 .4 Structure-Activity Relationship of Ring Ill of the 4.

Therapeutic Agents Acting on RNA Targets 5' C-G 1 4 0 5 G-C 5' C-G G-C 5' C-G G-C 5' C-G G-C U C A U 1495 G U C G U G A A U C A U U C A U G A Figure 6.be at work. (3) Mitochondrial DNA from individuals with the A1555G polymorphism can be co-transfected into pO HeLa cells.G C-G 5' E. only the N-hexanoyl-N-butylarnino derivative was essentially inactive against all bacteria tested. including streptomycin. Some natural variations exist at C4". It is speculated that this polymorphism makes the A-site of 12s rRNA (mitochondrial) look more like the A-site of 16s rRNA (bacterial) (Fig. For example. Nonstreptomycin aminoglycosides are also renally toxic (16). A wide variety of natural and synthetic variations is permissible at C5".holds that for at least some individuals aminoglycosides bind to mitochondrial ribosomes and shut down protein synthesis (72). An extensive study of C6" modifications of kanamycin B revealed that halogeno. 6. 6 AMlNOCLYCOSlDE TOXICITY All aminoglycosides. 71). A A C-G 1 4 1 0 A. One of the two most complete theories states that redox-active aminoglycoside/iron complexes catalyze the formation of free radicals. A hydrogen bond exists between the hydroxyl at C4" and a phosphate oxygen of U1406. Multiple hypotheses exist to explain aminoglycoside ototoxicity. (1) Reports of several pedigrees exist. The underlying causes of aminoglycoside ototoxicity are unresolved (reviewed in Ref. a hydroxyl group (in either configuration) is usually present but sometimes is replaced by a methoxy (41). The predicted lack of interaction between C5" substituents and 30s explains why a variety of small substitutions is permissible. At this position restrictions on the size and geometry of a substituent limits other allowable modifications. which are destructive to the hair cells of the inner ear (71). have the potential to produce irreversible vestibular and cochlear intoxication (16). 30s A-site sequences of ribosomes from various sources. amino. The NMR structure of gentamicin Cla (which is 5 deoxy) indicates that any modification at C 5 would be directed away from 30s (60). Mechanisms of the severe ototoxicity associated with aminoglycoside use are only partially known. alkylthio.will be discussed here. The other theory . which lack mitochondrial DNA altogether. Because of its ribosome involvement only the second theory . This polymorphism conspicuously occurs in the A-site of 12s rRNA.5. which correlate hypersensitivity to aminoglycosides with a single polymorphism in the mitochondrial genome (A1555G) (72-76). causing these cells to become sensitive to streptomycin intoxication (78). azido. and alkoxy substitutions are all well tolerated until they become very bulky (70). suggesting mitochondrial involvement (77).u C-G 5' G A A C C C A C-G U-A 5' - A C C C-G C-G U-A 5' - Human-cytoplasmic Mitochondrial wild type Mitochondrial mutant bonds to the N7 and a phosphate oxygen of G1405 that must be maintained for activity. coli C A U-A A. (2) The phenotype of sensoneural hearing loss or deafness due to aminoglycoside toxicity is the same as that found in most mitochondrial encephalomyopathies. indeed. (4) Significant uptake of tritium labeled aminoglycosides has . which is analogous to the gentamicin/neomycin binding site in bacterial ribosomes. The arrows indicate the point of variation between the 1555 polymorphic mitochondrial sequences. but none are com- pletely satisfactory.5).u 1 4 9 0 C. multiple mechanisms may . but circumstantial data suggest that aminoglycoside ototoxicity results from unintended interactions with human mitochondrial ribosomes. but the smaller substituents retained good activity. and at C6" for C 5 branched compounds (41).

7 TETRACYCLINE AND ANALOGS I 3 A direct model of aminoglycoside ototoxicity exists for neomycin-like and gentamicinlike compounds. The abundance of natural. Chlortetracycline was the first tetracycline to be isolated. streptomycin does interact with the backbone of A1491 (10). previous observations of tetracycline structure and activity can be rationalized from recent tetracyclinel30S crystal structures (11. either directly or indiredly. active tetracyclines. As was found for the aminoglycosides.41 to 1. in 1948. (6) Paromomycin has been observed to shut down mitochondrial activity in Leishmania understood. provides a rich collection of compounds from which to build meaningful structure-activity relationships.12). The Berlin group refined the positions of six tetracycline binding sites. However. with significant localization to mitochondria (79). (5)In vitro measurements demonstrate that gentamicid neomycin aminoglycosides bind poorly to the human mitochondrial decoding A-site. . 12). and greater chemical stability have produced superior drugs (821.7 Tetracycline and Analogs been observed in proximal tubular kidney cells and within sensory hair cells. lower toxicity. known as glycylcyclines. It is possible that the A1555G polymorphism potentiates streptomycin biding against mitochondrial ribosomes. Other common tetracyclines.86). In individuals with the A1555G polymorphism (analogous to A1490 in the E. one of the two mismatches is converted to a Watson-Crick base pair ((31410-G1490). Extensive efforts to identify tetracycline analogs from both natural and semisynthetic sources that possess greater efficacy. Ring I of both subclasses stacks atop the Watson-Crick base pair formed by C1409 and G1491. from Streptomyces aureofaciens (83). If the direct model is correct. Numerous reviews exist that report on the biosynthesis and use of tetracyclines (82. ranging in occupancies from 0. 6.0. In vitro binding studies of common aminoglycosides and both mitochondrial RNA sequences confirm that aminoglycosides bind the A1555G polymorphic sequence more tightly than the wildtype (79). 7. it nonetheless provides the basis for tetracycline nomenclature (Fig.which may help restore the stacking interaction between ring I and 1409 and 1491. one of which resides in the decoding A-site and seems to be responsible for the antibiotic effect of tetracycline (19).1 Mechanism of Action Biochemical probing identified multiple tetracycline binding sites within 30S. This base pair and the adjacent base pair form mismatched pairings ((31409 C1491 and C1410 A1490) in human mitochondrial ribosomes. coli small subunit rRNA). Although tetracycline was not the first agent of its class discovered. Two groups have independently solved tetracyclinel30S structures by X-ray crystallography (11. A second weakness is that tissue specificity for aminoglycoside toxicity is poorly . such as minocycline and doxycycline. but that is true for all competing theories of aminoglycoside ototoxicity. - Tetracyclines were the first broad-spectrum antibiotics and have been used successfully for decades to treat both gram-positive and gram negative bacterial infections (82). were isolated from Streptomyces sources in subsequent years. are superior to all existing clinically useful agents against several types of tetracycline-resistant bacteria (87-89).84. then there is optimism that agents with enhanced specificity for the bacterial sequence will necessarily be less ototoxic. in which aminoglycosides bind to mitochondrial ribosomes containing the A1555G mutation more tightly than they do to the wild-type ribosomes. but newer semisynthetic analogs may be on the way. such as oxytetracycline and tetracycline. Members of a new tetracycline class. but bind to the RNA containing the A1555G polymorphic sequence with an affinity similar to that of RNA with the bacterial sequence (80). coupled with extensive synthetic alterations. One obvious weakness in the theory of mitochondrial involvement in aminoglycoside intoxication is that streptomycin doesn't share a binding site with gentamicirdneomycin-like aminoglycosides.6). and the second pair is thought to disrupt the stabilizing stacking interaction between ring I and nucleotides 1409 and 1491.

as are the hydroxyls at C3. 7. allowing the correct codon-anticodon interaction to take place. tion of the A-site tRNA is blocked by tetracycline. Maintenance of the ketones at C1 and C11 is essential. and the subsequent hydrolysis of GTP At this point a necessary rotaby EF-Tu (11). inhibition of protein synthesis and the unproductive hydrolysis of GTP. or aromatization of ring A or C destroys all tetracycline activity (85). disrupting aromaticity in ring D.As with most ribosome targeting antibiotics.5 A (12). C. Removal of the dimethylamino group at C4 diminishes in vitro activity and abolishes in vivo activity (85). In this scenario.6. A refinement of this model has the elongation factor TuItRNA complex docking correctly into the A-site. and C12a. The amido group at C2 is rarely substituted. C10.Neither the hydroxyl nor the methyl at C6 is essential for activity. and D are generally not tolerated. breaking any ring. The structures of tetracycline and several analogs of tetracycline. All of these groups are involved in an extensive network of RNA binding interactions composed of hydrogen bonds and shared coordination to Mg2+. one of which. Essentially all alterations to the general tetracycline ring skeleton are deleterious. Tetracycline occupies a pocket formed by helices 31 and 34 of 16s rRNA (11). The two sites identified by the Cambridge group are a subset of the six sites reported by the Berlin group. rings A and B contain sites of unsaturation (Fig. Only the structure of the H311H34 tetracyclinel30S complex will be discussed here.7).Therapeutic Agents Acting on RNA Targets Tetracycline Chlortetracycline Demeclocycline 6-Dernethyltetracycline (Dernecycline) Oxytetracycline 6-Deoxytetracycline Minocycline Doxycycline R H CI CI H H H H ' Fi2 Fi3 OH O H OH OH OH H H H N(CH3)2 CH3 CH3 H H CH3 CH3 H CH3 R4 H H H H OH H H OH Figure 6. D. is aromatic. with both groups identifying tetracycline bound in a pocket formed by 16s rRNA helices 31 and 34 as the dominant and most relevant site. the con- . tetracycline extracts two payments from the cell. although other tetracycline binding sites could contribute to ribosome inhibition. Reduced binding is expected with the removal of this group. leading to the ejection of tRNA from the A-site without peptide bond formation. C12.2 Structure-Activity Relationship Tetracyclines contain four fused rings (A-Dl. Either chemical entity is able to hydrogen bond to the ribose of C1195. but a 2-acetyl moiety has been observed (85). It has been accepted for some time that tetracycline interferes with proper tRNA binding to the A-site (57). to a resolution of 4. 6. but extensive modification is possible elsewhere. no direct ribosome interaction is observed for the C6 substituents.4A (11). which they refined to 3. it interacts electrostatically with the phosphate of G966. Modifications to the functional groups on ring A and the bottoms of rings B. while the Cambridge group identified two tetracycline sites.

.

8. Indeed. the C5-C9 edge of tetracycline is directed away from any nearby 30S. which confers resistance to macrolides. hydroxyl and methyl.to 16-member monocycle lactone ring. These negative qualities provided the impetus to search for erythromycin alternatives. Several analogs (i. an amino sugar. It is a relatively broadspectrum antibiotic with low toxicity. A new class of macrolide analogs. the iso- Macrolides bind to a region of 50s near the peptidyl transferase center and block the progress of the nascent peptide through the exit channel (14). These derivatives. remain active against a wide variety of tetracycline-resistant bacteria. no hydrogen bonds are formed between tetracycline and any nucleobases-sequence selective binding comes from the proper presentation of sequence-independent groups. azithromycin and clarithromycin) were noticeable improvements in all regards except resistance. and the C8 methoxy derivative are possible. and C9. and virtually all substitutions at C6 are permitted. The remaining perimeter of tetracycline makes no close interactions with 30s. The dominant mechanism of resistance is through methylation of A2058 in domain V by the erm family of methyltransferases. The second most common form of erythromycin resistance comes from active efflux encoded for by mef genes (93). show great promise because of their increased activity both overall and against many erythromycin resistant bacteria (92-94). Also. Extensive hydrogen bonding networks are observed between the drug and backbone phosphate and ribose oxygen atoms. The structures explain virtually all of the absolutely required tetracycline groups. known as glycylcylines. bacteria were generally cross-resistant to erythromycin and its analogs. lincosamide. and widespread bacterial resistance (32.Therapeutic Agents Acting on RNA Targets tacts are to RNA only. is common. productive interactions with 30s. uncertain pharmacokinetics. even for groups as large as benzylmercaptan and glycosides (85). C8 methoxy derivatives are occasionally observed and are active (85). the first clinically used macrolide. Drug binding encompasses interactions to nucleotides from domain V (directly) and domain I1 (allosterically) of 23s rRNA (95).8). Both electron donating and electron withdrawing groups are tolerated at C7 (85). indicating that the amido substitutions are making additional unique. Cross-resistance arises from the sharing of a common binding site that includes interaction with A2058. that were improvements over erythromycin in one or more of the above categories.e. C9 amido substitutions yielded compounds with enhanced ribosome binding that escape recognition by certain efflux pumps (87-89). and cladinose (Fig. Of particular interest here is the fact that many of these derivatives have greatly enhanced ribosome binding. A l l attempts to modify tetracycline groups involved in the intricate interactions detailed above are deleterious. Numerous substitutions at C9 are also permissible. a collection of structurally diverse antibiotics that are collectively referred to as MLS. Erythromycin. Natural macrolides contain a highly substituted 12.. 8 ERYTHROMYCIN AND ANALOGS lation of numerous other macrolides quickly followed and the search for others continues today. 82). both from natural sources and from semisynthetic modifications. was isolated in 1953 from Saccharoplyspora erythraea (formerly Streptomyces erythreus) (91). clinical isolates . More work will be needed before the identities of these new interactions are revealed. Recently. Tetracycline binding has two interactions that are unusual drugrRNA interactions: binding through a shared Mg2+. but it is also characterized by poor chemical stability. referred to as an aglycone. Erythromycin is a member of the 14atom macrolide family and contains two sugar moieties: desosamine. referred to as the ketolides. antibiotics (57. 6. Removal of one or both of the common C6 substituents. and aromatic stacking between ring D and C1054.1 Mechanism of Action Isolation of macrolides began in the 1950s with the discovery of pikromycin (90). Similar to streptomycin. which explains why extensive modifications at C6. and group B streptogramines. C7. which coordinates several oxygens from tetracycline and phosphate groups from the ribosome. 93). to which one or more deoxysugars are attached.

9a-c). and both drugs block the exit channel. the stal structure and structure-activity relanships of erythromycin demonstrate that ly a handful of these potential contact nts are actually involved in direct contact g. sistant to erythromycin have recently been ported to contain 23s rRNA and ribosomal tein mutations (93). a ring common to 14-membered macrolides. Structure-Activity Relationship crolides have a rich assortment of chemical features through which tight. Therefore. or m aglycones. with the 14-atom class servthe basis for most clinically relevant mac- rolide antibiotics. This limited engagement by e regions of erythromycin has allowed the struction of diverse subgroups of macroe antibiotics. by virtue of a disaccharide in place of desosamine. clarithromycin. have both medical and veterinary use. selective molecrecognition of 50s can exist. although there is extensive evidence to support a similar mode of binding for the 16-membered ring macrolides. The 12-atom aglycone macrolides exhibit poor activity. only the 14-membered macrolides will be addressed here in detail. extensive modifican of the ring skeleton is possible. The structures of the macrolides: erythromycin. Notably. whereas 16-member aglycones. . 14. 6. Isolated macrolides typically have 12. Structural information exists only for the 14membered macrolides. Tylosin and erythromycin compete for the same binding site on 50S. enough differences between the mechanisms of action of the two classes exist to caution against extrapolating the 16-membered macrolides from existing 14-membered macrolide/50S structures. such as josamycin and tylosin.8. and azithromycin. tylosin and other 16-membered macrolides also inhibit peptide bond formation (96). however.8 Erythromycin and Analogs (desosamine) 1' \ Erythromycin CH3 Clarithromycin CH3 '33 Azithrornycin Figure 6.

the cladinose ring seems to make hydrophobic contacts with 30s. and the C12 hydryoxyl group forms a hydrogen bond to the . Erythromycin seems to be anchored to domain V of 23s rRNA by only a few hydrogen bonds and a salt interaction: the C6 hydroxyl forms a hydrogen bond to N6 of A2O62. arrows with an X point to non-permissible modifications.9-hemiketal or the 8.9. system of U2609 (14). For example. coli numbering) with various modifications tested for activity. (b and c) Several modifications tested.Therapeutic Agents Acting on RNA Targets O H 0+ ' (cladinose) Figure 6. there are also regions of erythromycin that make only minor contributions to binding. Clearly.12-spiroketal. Dashed arrow points to a modification that is not permissible unless corresponding changes are made elsewhere.TI. Under acidic conditions erythromycin can undergo spontaneous intramolecular ring cyclizations. The synthesis and isolation of the biologically active . the C2' desosamine hydroxyl group forms a hydrogen bond to N1 of A2058 and either N1 or N6 of A2059. Nevertheless. they may also contain the 6. the products of which are inactive (96-99). (a) Erythromycin/50S interactions.9-hemiketal.9. the C3' dimethylamino group is close to the phosphate of G2505.9. Dashed lines indicate possible hydrogen bonds (some of which are salt bridges when suitably reinforced with favorable electrostatic potentials).9-anhydro-6. such as the region encompassed by C7-Cll. other important non-bonding interactions must contribute in a large way to erythromycin binding. Molecular interactions between erythromycin and 50s (E. All contain either the 6. Arrows point to permissible modifications.

The synthesis of 11. Conversion of the C9 ketone to 0-substituted oximes and amino groups yields active compounds and removes the possibility of forming the above inactive cyclization products (101. because removal of this ring without other compensating changes elsewhere abolishes activity. (Continued. This interaction must account for a significant amount of binding energy. 104). The two sugars that form part of erythromycin are important for activity in most macrolide antibiotics. The lack of directional interactions. Erythromycin 9-oxime undergoes a Beckman rearrangement to give the ring-expanded 15member aglycone. Tricyclic analogs are also active (106).8 Erythromycin and Analogs "0 1' (desosamine) CH3 Figure 6. such as hydrogen bonds involving 2 . but the ring itself stacks atop the base equivalent to C2610 in E. agrees with the observation that methylation of ring hydroxyls is of little consequence to .0 and 4 . which is further converted to the 9-"azalide" (azithromycin) on reduction and N-methylation (103.9-bicyclic derivatives. 102). coli.9.12-bicyclic-epoxy macrolides established that it is the presence of the 6.12-N-substituted carbamates has generated many new useful compounds currently under clinical investigation (105).0 . Hydrogen bonding and strong electrostatic interactions are absent between cladinose and 50S.The above modifications are varied and sometimes extreme. 108). yet all of them are tolerated except for the formation of 6.9-bicycle that leads to inactive macrolides (100). all can be rationalized through inspection of the erythromycinl50S structure. cladinose can be replaced or dispensed with when other compensating modifications are made that restore ribosome binding (107.12-carbonate and 11.) 9.

as are several C4" derivatives (109. the desosamine ring is important for activity. illustrating that some of the derivatives compensated for the cladinose ring in terms of ribosome binding but not its ability to induce antibiotic resistance. In complex with 50S. unlike cladinose. those antibiotics remained active against MLSB-resistantStaphylococcus aureus and ef- flux-resistant Streptococcus pneumoniae.. Numerous 3-0-acyl derivatives of erythromycin were synthesized and Some of observed to be active antibiotics (111). (Continued. The cladinose ring seems to be an essential determinant for the induction of the inducible form of MLSBresistance. desosamine makes nu- . 110).9. an observation that prompted the investigation of analogs in which the cladinose was removed or replaced by other moieties.Therapeutic Agents Acting on RNA Targets I (desosamine) Figure 6. Like the cladinose ring.) activity. the C2" methoxy derivative is active. aromatic rings) showed the best activity. no synthetic alternatives have been identified. Substitutions that would be expected to stack most strongly (i.e. Other modifications to cladinose are also possible. further supporting the notion that the primary binding energy of cladinose is from stabilizing nonbonding interactions (1 11).

117. Numerous 16-membered macrolides (e. instead of the monosaccharide desosamine alone (115). The authors of this study speculated that the newly introduced 6-0-aryl groups enhance the overall interaction of those drugs with domain I1 of 23s rRNA. specifically with either of the extensions. Inspection of both of those structures suggests that the same hydrogen bond can be preserved for the 6-0-aryl derivatives. A pikromycin biosynthetic analog was produced that was p-glycosylated at the 2' position of desosamine. Because disruption of the hydroxyVA2058 interaction is enough to reduce the efficacy of erythromycin by several orders of magnitude. a disaccharide that contains desosamine. Methylation of both hydroxyl groups is tolerated. 108). 6.g. tylosin) have mycaminose. but is not ideal (111). Numerous 6-0-aryl derivatives yielded agents with superior antibiotic activity for erythromycin. and 11. ketolide. The ketolides.and 16membered macrolides prove to be identical.12-carbamate analogs that have been shown to have improved binding to domain I1 nucleotides.3 Refined View of Macrolide Binding At the resolution limit of the erythromycin1 50s structure. The cladinose ring is absent and the C3-OH is oxidized to a ketone. If true. Ribosomes containing these mutations bind macrolide antibiotics less well. Nevertheless.10) (108).8 Erythromycin and Analogs merous specific interactions: the C2' hydroxyl is within hydrogen bonding distance of N1 and N6 ofA2058 and N6 of A2O59. Modification of C6-OH of the aglycone is not only possible.2 A) between A752 of domain I1 and U2609 of domain V.OH/U2609 putative interaction seeming to be more favorable. 8. with the C12.which the 11. MLS. such as the recently approved drug telithromycin. Removal of the cladinose ring.12-carbamate arm could interact simultaneously with both nucleotides or an indirect mechanism by which ketolide interaction with one nucleotide subsequently affects the internucleotide interaction. Biochemical probing implicates nucleotide 752 of domain I1 as being involved with ketolide binding. which means that both types of modifications could not exist simultaneously. and the dimethylamine forms a salt bridge to the phosphate of G2505. ordered water molecules. as discussed above. Methylation of this hydroxyl group yields an antibiotic (clarithromycin) with superior pharmacokinetics (82). more work will needs to be done before definitive conclusions can be drawn about where the aryl arm extends with respect to 50s. reduces overall binding affinity. If the structure-activity relationships of the 14. it is preferable. 116). which suggests either a direct mechanism by .12-carbamate derivatives (106.12-carbamate functional groups (telithromycin) or the C6-0-aryl groups (ABT 773) (107. this also affirms the importance of the C2' hydroxyl group (114). which when present would disrupt hydrogen bonding between the C2' hydroxyl and the adenosine (14). which help mediate many RNA-ligand interac- . which can be more than compensated by the inclusion of N-linked arms that extend from either the 11. 118). again illustrating the importance of the C2' hydroxyl interactions.. laboratory isolates resistant to kelotides but not to erythromycin suggest additional domain V involvement as well (108. it is clear that the pocket into which the desosamine binds needs to be explored further before adequate structure-activity relationships can be realized.-resistant strains of bacteria produce a methyltranseferase that converts A2058 to an N6JV6-dimethyladenosine (112). Clinical isolates of Helicobacterpylori and other pathogenic bacteria with a limited number of rRNA operons resistant to clarithromycin sometimes contain either a common A2058 +G mutation or a rarer A2058 + C mutation in 23s rRNA (113). are a new class of erythromycin analogs (Fig. The hydroxyl groups from both C11 and C12 are within hydrogen bonding distance of U2609. this means that these modifications may occupy the same site as N-linked moieties from recent 11. yielding a compound devoid of antibiotic activity. In both the erythromycin/50S and the clarithromycin/50S structures the C6-0 is within hydrogen bonding distance of the N6 of A2062. The erythromycin/50S crystal structure reveals a close interaction (3. that interaction coupled with any negative steric clash that may be present must account for a significant portion of the overall binding energy.

12-carbarnate moiety (118). With the possible exception of the 11. Therefore. Macrolides such as erythromycin bind 50s by occupying one large pocket formed by nucleotides within the central loop of domain V and then branch out into other neighboring pockets by way of external sugar groups or. in the case of the ketolides.10. then there are a t least four pockets. by extensions from either the 11.12-carbamate or the C6-OH. At least three satellite pockets exist: the desosamine pocket. a global model of binding is emerging. ABT 773 dependent from each other.12-carbamate and C6-OH extensions.11). such as intricate hydrogen bonding networks. are not observed. The structures of the ketolides: telithromycin and ABT 773. remain to be identified.Therapeutic Agents Acting on RNA Targets 1' CH3 Telithromycin (formerly HMR 3647) Figure 6. 9 LESSONS LEARNED FROM RIBOSOME/DRUC COMPLEXES tions.12-extensionl60-aryl pocket. It seems that extensions from the C6-OH occupy the same pocket as do extensions from the 11. the cladinose pocket. interactions with these satellite pockets are essentially in- Reports of the individual drug/ribosome complexes are important both in terms of describing the mechanism of action and binding of those particular drugs and in terms of cataloguing and comparing the various modes of . all of which can be maximized for antibiotic activity (Fig. and the 11. pruning of one interaction can be compensated for by growth into another pocket. but the finer details. And while a quantitative accounting of the interactions that govern macrolide binding is absent at this time. but placement of large features like the core aglycone ring and the external cladinose and desosamine rings can be positioned with confidence (14). gross features of erythromycin binding are sound. If they do not. 6.

but there is substantial variation in how this is accomplished. like the drugs discussed above. An overall view of macrolideketolide binding. Two drugs. 6. adding further to the collective knowledge of drug/RNA interactions. Previous sections described the : binding of four classes of drugs in detail. P2 (C2610) binding involved (10-14). are drawn in Fig. each brings knowledge of previously unknown modes of interaction. and erythromycin contain sugars that stack atop the aromatic nucleotide bases (10. Several arms extend from the cupies the P1 binding pocket and makes interactions to multiple nucleotides including A2058. Extensions from cladinose extend into another pocket. Therefore. P2'. Cladinose occupies the P2 binding pocket and stacks with C2610. Other antibiotic/ RNA complexes have been solved (8-12). Ketolide antibiotics have either 11. aromatic rings from tetracycline and pactamycin stack with nucleotide bases. whereas compounds such as hygromycin B bind to a pocket within an expanded major groove through several hydrogen bonds to nucleotide bases but with no phosphate contacts (10). chloramphenicol and tetracycline. Their structures plus the structure of linezolid.12. 60).11.1 lists several ribosome binding antibiotics and one novel compound. gentamicin. Paromomycin and gentamicin share the same pocket within a major groove and make numerous contacts to both backbone phosphates and nucleotide bases (10. but also illustrated other important features of drug binding. Streptomycin binds to a site that ties together multiple helices almost exclusively through hydrogen bonds and salt bridges to backbone phosphates.14. such as hydrogen bonds to nucleobases and stacking of drug sugars atop base pairs (59. both extensions seem to interact with a pocket involving A752. 14). a recently approved synthetic antibiotic. Table 6. it shouldn't r surprise anyone that the observed modes of binding were very different. Shared metal coordination is not a common interaction between drug and proteins.s 9 Lessons Learned from Ribosome/Drug Complexes I Figure 6. given i their level of chemical diversity. a macrolide analog presumably cannot p r have both types of extensions. interact with rRNA through Mgt atoms that are coordinated by both the drugs and the RNA phosphate oxygens (11.60).59. Pactamycin illustrates the latter phenomenon in reverse: a ribose sugar from 16s rRNA stacks atop one of pactamycin's two aromatic rings (11). linezolid .60). seem to rely heavily on hydrophobic interactions to provide necessary binding energy. Earlier NMR structures of aminoglycosideIRNA complexes confirmed the predicted importance of electrostatic interactions between drug amino groups and RNA phosphate groups. Some antibiotics. For example. whereas paromomycin.12 carbamate extensions or aryl extensions from C6. such as erythromycin and chloramphenicol. it remains to be seen if this is a common phenomenon for RNAIantibiotic complexes or not.

even in tightly packed structures.12. It is interesting and not surprising to note that the two most hydrophobic compounds. Numerous high-resolution protein/RNA complexes exist in the crystallographic database. gentamicin. but if this trend holds true. As a group these LogP values tend to be lower (more negative) than the LogP values for most drugs.6. such as Coulombic forces. Structures of several other antibiotics whose 3D structures have been solved in corn-' plex with either 30s or 50S. In a more electrophilic environment the magnitude of some non-bonded interactions. all interact extensively with backbone phosphates through amino groups.1 to 4. paromomycin. The structure of linezolid in complex with 50s has not been solved. but others.8 for erythromycin for fully protonated drugs. but may also reflect the more hydrophilic nature of the surrounding RNA. and streptomycin. seem from their crystal structures to engage in mostly hydrophobic interactions with the ribosome. whereas the most hydrophilic compounds. Estimated LogP values range from -7. will be dampened. is extensively hydrated. and the novel synthetic antibiotic linezolid. such as the hydrophobic effect. which may reflect the type of interactions that are observed between antibiotics and RNA. Comparison between forces that mediate proteidprotein association . (119). will be enhanced. then linezolid should interact with the ribosome pri- marily through hydrophobic interactions. whereas the core of proteins is not. and the log of their estimated noctanol/water partition coefficients (LogP). RNA.Therapeutic Agents Acting on RNA Targets / OH Paromomycin HO HO H2N Pactamycin IYCH3 H3C HO kb$:o CH20H HN p H 3 OH O%&~*NH~ HO \i Hygromycin B OH Chloramphenicol 0 H3C HO CH3 Clindamycin Linezolid Figure 6.2 for paromomycin to 2. erythromycin and clindamycin. which also inhibits ribosome function (120-123). hygromycin B. hinting that the relative values that contribute to ligandlreceptor associations may be different for RNA and protein targets. The values for the same compounds in their neutral forms range from -4.

9 Lessons Learned from RibosomeIDrug Complexes

197

Table 6.1 "Rule of 5" Values for Selected RNA Binding Antibiotics
Compound Chloramphenicol Clindamycin Erythromycin Gentamicin 1Ca Hygromycin B Linezolid Padamycin Paromomycin Streptomycin Tetracycline Average: LogPa -1.689 0.826 2.874 -5.622 -5.619 2.011 0.495 - 7.212 -5.26 -2.688 -2.1884

NH + OHb
3 4 5 8 11 1 7 13 14 6 7.2

MW
323 410 734 449 527 337 531 600 581 444 493

N + 0"
6 7 14 12 16 7 12 18 19 10 12.1

Alertd 0 0 1 1 1 0 1 1 1 1

"LogP calculated using HINT (135).All alkylamineswere treated as ammonium ions, which at least for paromomycin and gentamicin 1Ca underestimates LogP. bSumof OH and NH H-bonds donors (132). "Sum of 0 and N H-bond acceptors (132). dComputationalalert according to rule of 5:O, no problem:l, poor absorption or permeation likely (132).

and those that mediate protein/RNA association may help in explaining global features of RNA molecule recognition (124-131). Ample evidence exists to confirm that RNA-binding ligands can be found that have drug-like affinity and specificity against a variety of RNA targets. A question that remains to be answered is whether similar agents can be applied to a wide range of targets, including cellular targets where a balance between drug solubility and cellular permeability is more important. Membrane permeability and drug solubility are intricately linked; the more permeable a membrane is to a drug, the less soluble the drug needs to be (132). The rule of 5 states that poor absorption or permeability is more likely when there are more than 5 Hbond donors, more than 10 H-bond acceptors, the molecular weight is greater than 500, and the calculated LogP is greater than 5 (133). This simple mnemonic is a set of criteria which most, but certainly not all, drugs obey. In the drug discovery process they are often used as a filter to reduce the number of potential candidates before costs of synthesis and testing are incurred (134). Here the rule of 5 will be used to predict whether or not any of the drugs in Table 6.1 would be acceptable drugs for a cellular target. Or put another way, are the chemical properties required for RNA binding mutually exclusive of the chemical properties required for good membrane permeability, absorption, and solubility?

Three of the antibiotics listed in Table 6.1, chloramphenicol, clindamycin, and the synthetic drug linezolid, are expected to have absorption and permeation qualities similar to the vast majority of drugs. All of the compounds that failed had either too many hydrogen bond acceptors, too many hydrogen bond donors, or both. Only two compounds in the failed set, gentamicin Cla and tetracycline, had molecular weights under 500. Many compounds were borderline; in particular, tetracycline has six hydrogen bond donors instead of the limit of five, but otherwise satisfies the rule of 5. Note that many active tetracycline derivatives satisfy the rule of 5. Can drugs be developed that act on nonribosomal RNA? Several facts highlighted within this chapter should provide optimism that such drugs can be found. The structural data that exist for antibiotic/ribosome complexes illustrate a wide range of intermolecular interactions that are sufficient to selectively inhibit RNA function; other tightly binding ligands against RNA targets, such as ligands against HIV genomic RNA, demonstrate similar specificity. And on balance, antibiotic/RNA binding is more hydrophilic in nature than druglprotein complexes. However, some agents, such as erythromycin, make use of extensive hydrophobic interactions, and for three compounds, chloramphenicol, clindamycin, and linezolid, chemical features required for RNA binding seem to

Therapeutic Agents Acting on RNA Targets

intersect with chemical properties required for drug-like absorption and permeability. The semisynthetic library, assembled largely in the absence of structural information, is rich for several classes of antibiotics; in most cases improvements in drug performance were possible, indicating that, even in the absence of structural data, promising lead compounds can be optimized enough to afford bona fide therapeutic agents. Note added in proof- In the final stages of preparing this chapter, an article has appeared from the groups of Moore and Steitz describing the structures of several macrolide drugs in complex with the 50s ribosomal subunit (136). Readers interested in macrolide antibiotics are encouraged to visit this article. It contains new information not presented within this chapter.
10

ACKNOWLEDGMENTS

7. F. Schuenzen, A. Tocilij, R. Zarivach, J.Harms, M. Gluehmann, D. Janell, A. Bashan, H. Bartels, I. Agmon, F. Franceschi, and A. Yonath, Cell, 102,615-623 (2000). 8. J. Harms, F. Schuenzen, R. Zarivach, H. Bashan, S. Gat, I. Agmon, A. Bartels, F. Franceschi, and A. Yonath, Cell, 107,679-688 (2001). 9. J. H. Cate, M. M. Yusupov, G. Z. Yusupov, T. N. Earnest, and H. F. Noller, Science, 285,20952104 (1999). 10. A. P. Carter, W. M. Clemons, D. E. Brodersen, R. J. Morgan-Warren, B. T. Wimberly, and V. Ramakrishnan, Nature, 407,340-348 (2000). 11. D. E. Brodersen, W. M. Clemons, A. P. Carter, R. J. Morgan-Warren, B. T. Wimberly, and V. Ramakrishnan, Cell, 103, 1143-1154 (2000). 12. M. Pioletti, F. Schlunzen, J. Harms, R. Zarivach, M. Gluhmann, H. A d a , A. Bashan, H. Bartels, T. Auerbach, C. Jacobi, T. Hartsch, A. Yonath, and F. Franceschi, EMBO J., 20, 1829-1839 (2001). 13. Q. Vicens and E. Westhof, Structure, 9, 647658 (2001). 14. F. Schlunzen, R. Zarivach, J. Harms, A. Bashan, A. Tocilj, R. Albrecht, A. Yonath, and F. Franceschi, Nature, 413,814-821 (2001). 15. A. Schatz, E. Bugie, and S. A. Waksman, Proc. Soc. Exp. Biol. Med., 55,66-69 (1944). 16. H. F. Chambers and M. A. Sande in J. G. Hardman, L. E. Limbird, P. B. Molinoff, R. W. Rudden, and A. G. Gilman, Eds., Goodman & Gilman's The Pharmacological Basis of Therapeutics, 9th ed., McGraw-Hill, New York, 1996, pp. 1103-1153. 17. R. J. Fitzgerald, F. Bernheim, and D. B. Fitzgerald, J. Biol. Chem., 175, 195 (1948). 18. C. R. Spots and R. Y. Stanier, Nature, 192,633 (1961). 19. D. Moazed and H. F. Noller, Nature, 327,389394 (1987). 20. P. Nissen, J. Hansen, N. Ban, P. B. Moore, and T. A. Steitz, Science, 289,920-930 (2000). 21. N. D. Pearson and C. D. Prescott, Chem. Biol., 4,409-414 (1997). 22. T. Hermann, Angew Chem. Int. Edu., 39, 1890-1905 (2000). 23. J. Gallego and G. Varani, Acc. Chem. Res., 34, 836-843 (2001). 24. J. A. Doudna, Nature Struct. Biol., 7(Suppl), 954-956 (2000). 25. M. L. Zapp, S. Stern, and M. R. Green, Cell, 74, 969-978 (1993).

The author thanks the Deafness Research Foundation and the American Cancer Society (IN-105),which supported some of this work. I am grateful to Ms. Heather O'Farrell for proofreading the manuscript and to Prof. Glen Kellogg for helpful discussions.
REFERENCES 1. D. A. Matthews, R. A. Alden, J. T. Bolin, S. T. Freer, R. Hamlin, N. Xuong, J. Kraut, M. Poe, M. Williams, and K. Hoogsteen, Science, 197,
2. 452-455 (1977). D. J. Abraham in T. J. Perun and C. L. Propst, Eds., Computer-Aided Drug Design: Methods and Applications, Marcel Dekker, New York, 1989, pp. 93-132. D. J. Abraham, Ed., Burger's Medicinal Chemistry, 6thed., vol. lA, Wiley-Interscience, New York, 2002. P. B. Moore in R. F. Gesteland, T. R. Cech, and J . F. Atkins, Eds., The RNAWorld, Cold Spring Harbor Laboratory Press, New York, 1993, pp. 119-135. N. Ban, P. Nissen, J. Hansen, P. B. Moore, and T. A. Steitz, Science, 289,905-920 (2000). B. T. Wimberly, D. E. Brodersen, W. M. J. Clemons, R. Morgan-Warren, C. von Rhein, T. Hartsch, and V. Ramakrishnan, Nature, 407, 327-339 (2000).

3.

4.

5. 6.

References

h
26. A. J. Collier, J. Gallego, R. Klink, P. T. Cole, S. J. Harris, G. P. Harrison, F. Aboul-Ela, G. Varani, and S. Walker, Nut. Struct. Biol., 9, 375380 (2002). 27. S. J. Suscheck, W. A. Greenberg, T. J. Tolbert, and C.-H. Wong, Angew Chem. Znt. Edu., 39, 1080-1084 (2000). 28. J. B.-H. Tok, J. Cho, and R. R. Rando, Biochemistry, 38, 199-206 (1999). 29. L. Varani, M. G. Spillantini, M. Goedert, and G. Varani, Nucleic Acids Res., 28, 710-719 (2000). 30. G. Werstuck and M. R. Green, Science, 282, 296-298 (1998). 31. G. D. Stormo and Y. Ji, Proc. Natl. A c d . Sci. USA, 98,9465-9467 (2001). 32. G. D. Wright in D. J. Abraham, Ed., Burger's Medicinal Chemistry, 6th ed., vol. 4, Wiley-Interscience, New York, 2002. 33. V. Ramakrishnan, Cell, 108,557-572 (2002). 34. W. Piepersberg in W. R. Strohl, Ed., Biotechnology of Antibiotics, 2nd ed., Marcel-Dekker, New York, 1997. 35. G. L. Mandell and W. A. Petri in J. G. Hardman, L. E. Limbird, P. B. Molinoff, R. W. Rudden and A. G. Gilman, Eds., Goodman & Gilman's The Pharmological Basis of Therapeutics, 9th ed., McGraw-Hill, New York, 1996, pp. 1155-1174. 36. 0. Jerinic and S. Joseph, J. Mol. Biol., 304, 707-713 (2000). 37. R. L. Peck, R. P. Graber, A. Walti, E. W. Peel, C. E. Hoffhine, and K. Folkers, J. Am. Chem. SOC., 6 8 , 2 9 3 1 (1946). 38. S. Umezawa and T. Tsuchiya in H. Umezawa and I. R. Hooper, Eds., Aminoglycoside Antibiotics, Springer-Verlag, Berlin, 1982, pp. 37-110. 39. T. Usui, T. Tsuchiya, and S. Umezawa, J. Antibiotics, 31, 991-996 (1978). 40. D. L. Delaware, M. S. Sharma, B. S. Iyengar, W. A. Remers, and T. A. Pursiano, J. Antibiotics, 34,251-258 (1986). 41. K. E. Price, J. C. Godfrey, and H. Kawaguchi in D. Perlman, Ed., Structure Activity Relationships Among the Semisynthetic Antibiotics, Academic Press, New York, 1977, pp. 239-395. 42. T. Tsuchiya, T. Kishi, S. Kobayashi, Y. Kobayashi, S. Umezawa, and H. Umezawa, Curb. Res., 104, 69-77 (1982). 43. R. L. Peck, C. E. Hoffhine, and K. Folkers, J. Am. Chem. Soc., 68,1390-1391 (1946).
44. S. Tatsuoka, T. Kusaka, A. Miyake, M. Inoue, H. Hitomi, Y. Shiraishi, H. Iwasaki, and M. Imanishi, Chem. Pharm. Bull., 5, 343-349 (1957). 45. J. Fried and 0 . Wintersteiner, J. Am. Chem. S OC., 69, 79-86 (1947). 46. H. Heding, Tet. Lett., 1969,2831-2832 (1969). 47. W. A. Winsten, C. I. Jarowski, F. X. Murphy, and W. A. Lazier, J. Am. Chem. Soc., 72,39693972 (1950). 48. H. P. Treffers, N. 0. Besler, and D. C. Alexander, Antibiotics Ann., 1953,595-603 (1953). 49. H. Heding, Acta Chem. Scand., 26,3251-3256 (1972). 50. H. Heding and A. Diedrichsen, J. Antibiotics, 28,312-316 (1975). 51. F. C. Pennington, P. A. Guerico, and I. A. Solomon~, J. Am. Chem. Soc., 75,2261 (1953). 52. M. Awata, N. Muto, M. Hayashi, and S. Yaginuma, J. Antibiotics, 39, 724-726 (1986). 53. T. Tsuchiya, S. Sakamoto, T. Yamasaki, and S. Umezawa, J. Antibiotics, 35, 639-641 (1982). 54. H. Sano, T. Tsuchiya, S. Kobayashi, M. Hamada, S. Umezawa, and H. Umezawa, J. Antibiotics, 25, 978-980 (1976). 55. S. A. Waksman and H. A. Lechevalier, Science, 109,305-307 (1949). 56. M. Murase, T. Wakazawa, M. Abe, and S. Kawaji, J. Antibiotics, A14, 156-157 (1961). 57. E. F. Gale, E. Cundliff, P. E. Reynolds, M. H. Richmond, and M. J. Waring, The Molecular Basis ofAntibiotic Action, 2nd ed., Wiley-Interscience, London, 1981, pp. 402-547. 58. M. I. Recht, S. Douthwaite, and J. D. Puglisi, EMBO J., 18,3133-3138 (1999). 59. D. Fourmy, M. I. Recht, S. C. Blanchard, and J. D. Puglisi, Science, 274, 1367-1371. 60. S. Yoshizawa, D. Fourmy, and J. D. Puglisi, EMBO J., 17,6437-6448 (1998). 61. J. M. Ogle, D. E. Brodersen, W. M. Clemons, M. J. Tarry, A. P. Carter, and V. Rarnakrishnan, Science, 292,897-902 (2001). 62. J. Davies and G. D. Wright, Trends Microbiol., 5,234-240 (1997). 63. J. Haddad, S. Vakulenko, and S. Mobashery, J. Am. Chem. Soc., 121,11922-11923 (1999). 64. T. Yamasaki, Y. Narita, H. Hoshi, S. Aburaki, H. Kamei, T. Naito, and H. Kawaguchi, J. Antibiotics, 44, 646-653 (1991). 65. H. Hoshi, S. Aburaki, S. Iimura, T. Yamasaki, T. Naito, and H. Kawaguchi, J. Antibiotics, 43, 858-872 (1990).

Therapeutic Agents Acting on RNA Targets

66. K. Richardson, K. W . Brammer, S. Jevons, R. M. Plews, and J. R. Wright, J. Antibiotics, 32,973-977 (1979). 67. K. Richardson, S. Jevons, J. W . Moore, B. C. Ross, and J. R. Wright, J. Antibiotics, 30,843846 (1977). 68. P. B. Alper, M. Hendrix, P. Sears, and C.-H. Wong, J. Am. Chem. Soc., 120, 1965-1978 (1998). 69. W . A. Greenberg, E. S. Priestley, P. S. Sears, P. B. Alper, C. Rosenbohm, M. Hendrix, S.-C. Hung, and C.-H. Wong, J. Am. Chem. Soc., 121,6527-6541 (1999). 70. A. V a n Schepdael, J. Delcourt, M. Mulier, R. Busson, L. Verbist, H . J. Vanderhaeghe, M. P. Mingeot-Leclercq, P. M. Tulkens, and P. J. Chaes, J. Med. Chem., 34,1468-1475 (1991). 71. J. Schacht, Ann. NYAcad. Sci., 884, 125-130 (1999). 72. R. A. Casano, Y . Bykhovskaya, D. F. Johnson, M. Hamon, F. Torricelli, M. Bigozzi, and N. Fischel-Ghodsian, Am. J. Med. Genet., 79, 388-391 (1998). 73. X. Estivill, N. Govea, E. Barcelo,C. Badenas, E. Romero, L. Moral, R. Scozzri, L. D'Urbano, M. Zeviani, and A. Torroni, Am. J. Hum. Genet., 62,27-35 (1998). 74. A. Pandya, X. Xia, J. Radnaabazar, J. Batsuuri, B. Dangaansuren, N. Fischel-Ghodsian, and W . E. Nance, J. Med. Genet., 34, 169-172 (1997). 75. M. Shohat, N. Fischel-Ghodsian, C. Legum, and G. J. Halpern, Am. J. Otolaryngol., 20, 64-67 (1999). 76. S. Kupka, T . Toth, M. Wrobel, U . Zeissler, W . Syzfter, K. Szyfter,G. Niedzielska, J. Bal, H. P. Zenner, I. Sziklai, N. Blin, and M. Pfister, Hum. Mutat., 19, 308-309 (2002). 77. D. C. Wallace,Annu. Rev. Biochem., 61, 11751212 (1992). 78. K. Inoue, D. Takai, A. Soejima, K. Isobe, T . Yamasoba, Y . Oka, Y . Goto, and J. Hayashi, Biochem. Biophys. Res. Commun., 223, 496501 (1996). 79. D. Beauchamp, P. Gourde, and M. G. Bergeron, Antimicrob. Agents Chemother., 35, 2173-2179 (1991). 80. K. Hamasaki and R. R. Rando, Biochemistry, 36,12323-12328 (1997). 81. M. Maarouf, Y . de Kouchkovsky, S. Brown, P. X . Petit, and M . Robert-Gero, Exp. Cell Res., 232,339-348 (1997). 82. J . E. Kapusnik, M. A. Sande, and H. F. Chambers i n J. G. Hardman, L. E. Limbird, P. B.

Molinoff, R. W . Rudden and A. G. Gilman, A. G., Eds., Goodman & Gilman's The Pharmological Basis of Therapeutics, 9th ed., McGraw-Hill, New York, 1996, pp. 1123-1153. 83. J. H. Boothe and J. J. Hlavka i n Kirk-Othmer Encyclopedia o f Chemical Technology, vol. 3, 4th ed., New York, Wiley, 1992, p. 331. 84. I. S. Hunter and R. A. Hill i n W . R. Strohl, Ed., Biotechnology of Antibiotics, 2nd ed., MarcelDekker, New York, 1997, pp. 659-682. 85. R. K. Blackwood and A. R. English in D. Perlman, Ed., Structure-Activity Relationships Among the Semisynthetic Antibiotics, Academic Press, New York, 1977, pp. 397-426. 86. D. Schnappinger and W . Hillen, Arch. Microbiol., 165,359369. 87. J. Bergeron, M. Ammirati, D. Danley, L. James, M. Norcia, J. Retsema, C. A. Strick, W.-G. Su, J. Sutcliffe,and L. Wondrack, Antimicrob. Agents Chemother., 40, 2226-2228. 88. P. J. Petersen, N.V . Jacobus,W . J.Weiss, P. E. Sum, and R. A. Tests, Antimicrob. Agents Chemother., 43, 738-744. 89. P.-E. S u m and P. Petersen, Bioorg. Med. Chem. Lett., 9, 1459-1462. 90. H. Brockman and W . Henkel, Naturwissenschaften, 37,138-139 (1950). 91. J. M. McGuire, R. L. Bunch, R. C. Anderson, H. E. Boaz, E. H. Flynn, H. M. Powell, and J. W . Smith, Antibiot. Chemother., 2,281-283 (1952). 92. A. Bonnefoy, A. M. Girard, C. Agouridas, and J . F. Chantot, J. Antimicrob. Chemother., 40, 85-90 (1997). 93. R. Leclerq, J. Antimicrob. Chemother., 48, 9-23 (2001). 94. D. J. Hoban,A. K. Wierzbowski, K. Nichol, and G. G. Zhanel, Antimicrob. Chemother., 45, 2147-2150 (2001). 95. M. Dam, S. Douthwaite, T . Tenson, and A. S. Mankin, J. Mol. Biol., 259, 1-6 (1996). 96. S. M. Poulsen, C. Kofoed, and B.Vester, J. Mol. Biol., 304,471-481 (2000). 97. P. Kurath, P. H. Jones, R. S. Egan, and T . J. Perun, Experientia, 27,362 (1971). 98. C. Vinckier, R. Hauchecorne, T . Cachet, G. V a n der Mooter, and J. Hoogmartens, Znt. J. Pharmaceutics, 55, 67-76 (1989). 99. R. J. Pariza and L. A. Freiberg, Pure Appl. Chem., 66,2365-2368 (1994). 100. D. J. Hardy, R. N. Swanson, N. L. Shipkowitz, L. A. Freiberg, P. A. Lartey, and J. J. Clement, Antimicrob. Agents Chemother., 35, 922-928 (1991).

rences

J.-C. Gasc, S. G. d'AmbiBres, A. Lutz, and J.-F. Chantot, J. Antibiotics, 44, 313-330 (1991). E. H. Massey, B. S. Kitchell, L. D. Martin, and K. Gerzon, J. Med. Chem., 17,105-107 (1974). G. M . Bright, A. A. Nagel, J. Bordner, K. A. Desai, J. N. Dibrino, J. Nowakowska, L. Vincent, R. M. Watrous, F. C. Sciavolino, A. R. English, J. A. Retsema, M. R. Anderson, L. A. Brennan, R. J. Borovoy, C. R. Cimochowski, J. A. Faiella, A. E. Girard, D. Girard, C. Herbert, M. Manousos, and R. Mason, J. Antibiotics, 41, 1029-1047 (1988). S. Djokic, G. Kobrehel, N. Lopotar, B. Kamenar, A. Nagl, and D. Mrvos, J. Chem. Res.152153 (1988). Y . J. W u and W . G. Su, Cum. Med. Chem., 14, 1727-1758 (2001). Z. Ma, R. F. Clark, A. Brazzale, S. Wang, M. J. Rupp, L. Li, G. Griesgraber, S. Zhang, H. Yong, L. T . Phan, P. A. Nemoto, D. T . W . Chu, J. J. Plattner, X. Zhang, P. Zhong, Z. Cao, A. M. Nilius,V . D. Shortridge, R. Flamm, M. Mitten, J . Meulbroek, P. Ewing, J. Alder, Y . S. Or, J. Med. Chem., 44,4137-4156 (2001). C. Agouridas, A. Denis, J.-M. Auger, Y . Benedetti, A. Bonnefoy, F. Bretin, J.-F. Chantot, A. Dussarat, C. Fromentin, S. G. D'AmbriBres, S. Lachaud, P. Laurin, 0. Le Martret, V . Loyau, and N. Tessot, J. Med. Chem., 41,4080-4100. S. Douthwaite and W . S. Champney, J. Antimicrob. Chemother., 48, 1-8 (2001). P. B. Fernandes, W . R. Baker, L. A. Freiberg, D. J . Hardy, and E. J. McDonald, Antimicrob. Agents Chemother., 33, 78-81 (1989). S. T. Waddell,G. M. Santorelli, T . A. Blizzard, A. Graham, and J. Occi, Bioorg. Med. Chem. Lett., 8, 549-554 (1998). T . Tanikawa, T . Asaka, M. Kashimura, Y . Misawa, K. Suzuki, S. Morimoto, and A. Nishida, J. Med. Chem., 44,4027-4030. C. Abad-Zapatero, P. Zhong, D. E. Bussiere, K. Stewart, and S. W . Muchmore i n X . Cheng and R. Blumenthal, Ed., S-AdenosylmethionineDependent Methyltransferases: Structures and Functions, World Scientific Publishing, River Edge, NJ, 1999, pp. 199-225. B. Vester a n d S. Douthwaite, Antimicrob. Agents Chemother., 45, 1-12 (2001). L. Zhao, D. H. Sherman, and H. Lui, J. Am. Chem. Soc., 120,9374-9375 (1998). S. M. Poulsen, C. Kofoed,and B.Vester, J. Mol. Biol., 304,471-481 (2000).

L. T . Phan, R. F. Clark, M. Rupp, Y . Or, D. T . W . Chu, and S. Ma, Organic Lett., 2, 2951-2954 (2000). S. Douthwaite, L. H. Hansen, and P. Mauvais, Mol. Microbiol., 36, 183-193 (2000). G. Garza-Ramos, L. Xiong, P. Zhong, and A. Mankin, J. Bacteriology, 183, 6898-6907 (2001). S. J. Brickner, D. K. Hutchinson, M. R. Barbachyn, P. R. Manninen, D. A. Ulanowicz, S. A. Garmon, K. C. Grega, S. K. Hendges, D. S. Troops, C.W . Ford, and G. E. Zurenko, J. Med. Chem., 39,673-679 (1996). P. Kloss, L. Xiong, D. L. Shinabarger, and A. S. Mankin, J. Mol. Biol., 294,93-101 (1999). L. Xiong, P. Kloss, S. Douthwaite, N. M . Andersen, S. Swaney, D. L. Shinabarger, and A. S. Mankin, J. Bacteriology, 182,5325-5331 (2000). U . Patel, Y . P. Yan, F.W . Hobbs, J. Kaczmarczyk, A. M. Slee, D. L. Pompliano, M. G. Kurilla, and E. Y . Bobkova, J. Biol. Chem., 276,3719937205 (2001). W . S. Champney and M. Miller, Curr. Trends Microbiol., 44,350-356 (2002). K. Nagai, Curr. Opin. Struct. Biol., 6, 53-61 (1996). R. N. De Guzman, R. B. Turner, and M. F . Summers, Biopolymers, 48, 181-195 (1998). A. R. Ferre-D'Amare and J. A. Doudna, Ann. Rev. Biophys. Biomol. Struct., 28, 57-73 (1999). Y . Muto, C. Oubridge, and K. Nagai, Curr. Opin. Biol., 10, R19-R21 (2000). J. R. Williamson,Nat. Struct. Biol., 7,834-837 (2000). J. M. Perez-Canadillas and G. Varani, Curr. Opin. Struct. Biol., 11,53-58 (2001). P. B. Moore, Biochemistry, 40, 3243-3250 (2001). V . Ramakrishnan and P. B. Moore, Curr. Opin. Struct. Biol., 11, 144-145. C. A. Lipinski, J. Pharmacol. Toxicol. Methods, 44,235-249 (2000). C. A. Lipinski, F. Lombardo, B. W . Dominy, and P. J. Feeney, Adv. Drug Deliv. Rev., 46, 3-26 (2001). W . P. Walters and M. A. Murcko, Adv. Drug Deliv. Rev., 54, 255-271 (2002). G. E. Kellogg and D. J. Abraham, Eur. J. Med. Chem., 35,651-661 (2000). J . L. Hansen, J. A. Ippolito, N. Ban, P. Nissen, P. B. Moore, and T . A. Steitz, Mol. Cell, 10, 117-128 (2002).

CHAPTER SEVEN

Carbohydrate-Based
JOHN H. MUSSER
Pharmagenesis, Inc. Palo Alto, California

Contents
1 Introduction, 204 2 Carbohydrate-Based Drugs, 204 2.1 To Market-Last Decade, 205 2.1.1 Diabetes, 205 2.1.2 Gastrointestinal System, 206 2.1.3 Central Nervous System, 206 2.1.4 Infection, 207 2.1.5 Thrombosis, 209 2.1.6 Inflammation, 210 2.1.7 Enzyme Replacement, 211 2.2 Therapeutics in Development, 211 2.2.1 Gastrointestinal System, 212 2.2.2 CNS, 213 2.2.3 Infection, 213 2.2.4 Thrombosis, 215 2.2.5 Inflammation, 216 2.2.6 Cancer, 216 2.2.7 Lysosomal Storage, 218 2.3 Agents in Research, 219 2.3.1 Diabetes, 219 2.3.2 Gastrointestinal System, 219 2.3.3 CNS, 219 2.3.4 Infection, 221 2.3.5 Inflammation, 223 2.3.6 Cancer, 225 3 Carbohydrate Chemistry, 227 3.1 Chemistry, 227 3.1.1 Challenges and Opportunities, 227 3.1.2 Solution Phase, 227 3.1.3 Enzymatic, 229 3.1.4 Solid Phase, 230 3.1.5 Glycoprotein, 231 3.1.6 Carbohydrate Library, 231 3.2 Analysis, 232 3.2.1 Mass Spectrometry, 232 3.2.2 NMR, 232 3.3 Natural Products Isolation, 232 3.4 Multivalency, 233

Burger's Medicinal Chemistry and Drug'Discovery Sixth Edition, Volume 2: Drug Development Edited by Donald J. Abraham ISBN 0-471-37028-2 O 2003 John Wiley & Sons, Inc.

Carbohydrate-Based Therapeutics

3.4.1 Viral Inhibitors, 233 3.4.2 RNA Target, 233 3.4.3 Toxin Inhibitors, 234 3.4.4 Mechanism of Multivalency, 234 3.5 Heparin, 235 3.5.1 Synthesis, 235 3.5.2 FGF, 235 3.5.3 Heparin Analytical Tool, 236 4 Glycobiology, 236 4.1 Driving Forces, 236 4.1.1 Human Genome, 236 4.1.2 Therapeutic Glycoproteins, 237 4.1.3 Structure-Function, 237

4.2 Glycoconjugates, 237 4.2.1 Glycolipids, 237 4.2.2 Proteoglycans, 238 4.2.3 Glycoproteins, 238 4.3 Immunogens, 239 4.3.1 Immunoglobulins, 239 4.3.2 Hemagglutinins, 240 4.4 Ledins, 240 4.4.1 C-Type Lectins, 240 4.4.2 Mannan-Binding Lectin, 241 4.4.3 Siglecs, 242 4.5 Health and Disease, 242 5 Summary-Continue the Pace, 242

1

INTRODUCTION

A decade has elapsed since we first wrote-"by the 21st century, research on carbohydrates will emerge as significant new approach to drug discovery (I)." Indeed, significant progress has been made toward the realization of this statement. The pharmaceutical impact of this class of biomolecules can be seen in the carbohydrate-based therapeutics recently approved for marketing. For example, the trend toward producing lower molecular weight heparinbased products has culminated in the approval of a pentasaccharide for preventing venous thromboembolism after major orthopedic surgery. Also, in the antiviral arena that has traditionally resisted the introduction of new medicines, not one but two new monosaccharide glycomimetic drugs that block the ability of influenza viruses to exit infected cells by binding to neuraminidase were registered for the treatment of flu. The pharmaceutical impact of carbohydrate-based therapeutics can also be seen in the agents now under evaluation in pre-clinical and clinical development. For example, synthetic carbohydrate chains, modeled on tumor cell antigens but modified to stimulate an immune response, are being used to develop novel cancer vaccines. In research, incremental increases in the capability of research tools have made the localization and functional characterization of glyconjugates possible. Hints at how to solve the many carbohydrate synthesis problems can be gleaned from the developments made in solid and solution phase as well enzymatic

chemistry. Most importantly, carbohydrate chemists and glycobiologist have forged links that have allowed carbohydrate-based therapeutics to enter the mainstream drug discovery process. Thus, all the above factors have combined to clearly demonstrate the importance of carbohydrates as a source for new drug discovery.

2 CARBOHYDRATE-BASED DRUGS

The carbohydrate-based drug section is divided into three parts. The first part contains drugs that have made it to the market in the last decade. An emphasis is placed on new product introductions both because there has been significant progress in the commercialization of carbohydrate-based therapeutics and the product approval process is a very stringent measure of success. The second part of this section deals with therapeutics in development. Although not as stringent as marketed products, the introduction of new therapeutics into clinical trials is not trivial. Also, showing safety first, then efficacy, is a demanding task, and there are failures. A review of carbohydrate-based therapeutics that failed in development is as informative as those that have succeeded. The final part is focused on carbohydrate agents in research. Indeed, it is interesting times for carbohydrate research because there are both promising ideas and nake approaches, and only time will tell which research will lead to new carbohydrate-based drugs.

2 Carbohydrate-Based Drugs

2.1 To Market-Last Decade

What marketed carbohydrate-based drugs means in the context of this chapter is that the indicated drugs have been approved by government regulatory agencies during the last decade in one or more countries as safe and efficacious medicines for human use. It also means the drug's chemical structure or the drug's mechanism of action is related to carbohydrate chemistry or glycobiology. In the last decade, a number of new carbohydrate-based therapeutics have been approved for human use including both new chemical entities (usually low molecular weight) and new biological entities (usually high molecular weight). This section is divided by molecular weight and organized into therapeutic classes. The therapeutic classes are grouped as follows: diabetes, gastrointestinal, central nervous system, infection, thrombosis, inflammation, and enzyme replacement.
2.1.1 Diabetes. Diabetes is a condition

where there is too much glucose in the blood. Therefore, it is not surprising that carbohydrate-based drugs would be active in a disease The tetrasaccharide acarbose (1)is used as an adjunctive in non-insulin-dependent type 2 diabetes and is also of benefit in insulin-dependent type 1 diabetes when used in conjunction is derived from with insulin (2). Acarbose (1) the fermentation process of the fungus Actinolanes utahensis and is orally active. It works intestinal a-glucosidases, resulting in lower blood glucose levels after meals. Thus, acarbose (1)is an effective treatment for post-prandial hyperglycemia. is The mechanism of action for acarbose (1) informative as to how carbohydrate-based

drugs may work effectively. In a typical reaction in the intestine, the polymeric sugar chains are broken down into individual monosaccharide units by the action of a-glucosidase. This enzyme recognizes specific residues in the polymeric sugar chain and hydrolyzes the inter-glycoside linkage through the formation of an oxonium ion intermediate. A subsequent hydrolysis reaction of the intermediate releases the non-reducing end monosaccharide that becomes available for quick absorption in the intestine and release in the blood. Acarbose (1) is a stable transition state inhibitor of this reaction. Its structure resembles the transition state structure of the polysaccharide, although it lacks the ability to form a stable oxonium ion intermediate. A carbose (1)has a 105-fold higher affinity than typical substrates. Thus, it binds the a-glucosidase enzyme, incapacitating it from its action on its normal substrates. This prevents the release of monosaccharides from polysaccharides in the stomach and intestine. The normal polysaccharides that are usually broken down in the gastrointestinal tract are eliminated in feces. There are gastrointestinal side effects with acarbose (1)treatment including pain, diarrhea, and flatulence. The drug is not recommended for people with limited kidney function, and high doses cause abnormal liver enzyme values. Therefore, the maximum dosage should not exceed 100 mg, three times per day. Voglibose (2) is anti-diabetic carbohydratebased drug, and like acarbose (I),it is an a-glucosidase inhibitor for prevention of postprandial hyperglycemia. Clinical studies have shown that voglibose (2) limits daily fluctua-

OH
,d N//,Q,

CH3 -

HO'""'
OH

OH

;

* ',

Carbohydrate-Based Therapeuti

tions in blood glucose level and decreases the amount of insulin produced by the pancreas. Voglibose (2) is well tolerated and has been proven to be safe for further clinical use in combination with other drugs for diabetes, such as glibenclamide (3). In rats with type 2 diabetes, the combination of voglibose and a pioglitazone seems to control blood glucose levels and improves the ability to regulate future glucose levels by restoring the function of damaged islet cells in the pancreas (4). Structurally, voglibose (2) is a synthetic monosaccharide that may offer advantages over the tetrasaccharide, acarbose (1). However, when compared side by side clinically, 1-day administration of acarbose (1)and voglibose (2) at currently recommended clinical doses demonstrated that acarbose (1) was more effective in sparing endogenous insulin secretion than was voglibose (2) (5). Another azasugar that has been used for diabetes treatment is miglitol (3) (6). Like acarbose (1)and voglibose (21, it is an oral a-glucosidase inhibitor for use in the management of non-insulin-dependent diabetes mellitus. Structurally, miglitol (3) is a desoxynojirimycin derivative, and this family of

compounds has antiviral and anticancer pro] erties (vide infra). Nevertheless, this age1 has anti-diabetic properties that are similar 1 both volglibose (2) and acarbose (1).
2.1.2 Gastrointestinal System. Because hi

torically carbohydrates were considered t function as merely energy storage, such 2 starch, it is ironic that some newly approvc therapeutics are used to heal the gastrointe tinal system. Dosmalfate (4) is an oral cytoprotecti~ agent that is used for healing gastric, duodl nal, and esophageal ulcers (7). An addition; indication is to prevent lesions caused by tl. chronic intake of non-steroid anti-inflammatoi drugs. Its mechanism of gastric protection dependent on multiple factors including tl modulation of prostaglandins. Dosmalfate (4) reported to have a coating effect, including ant r acid diffusion activity, but there are not a anti-secretory properties. It was well tolerate in the clinic.
2.1.3 Central Nervous System. Carboh:

drate-based drugs acting on the central ne:

2 Carbohydrate-Based Drugs

-

vous system (CNS) is surprising because, in : general, saccharides are water-soluble, and drugs that are CNS active usually have appreciable lipid solubility. Apparently, a way to solve the problem is to mask the free hydroxyl groups of the saccharide with alkyl groups. Topiramate (5) is a monosaccharide diacetonide with a sulfamate group that has activ-

Symptoms include acute onset of decreased visual acuity and/or ocular pain.
2.1.4 Infection. Carbohydrate-based drugs that treat infection include antibiotics and antiviral~. In this section are agents to treat sepsis and those that act as vaccines. Also, no attempt is made to cover all carbohydrate-based anti-infectives because of the number involved and because antibiotics are covered more thoroughly elsewhere (10). Aminoglycosides are a diverse group of carbohydrate-based antibiotics that are used to treat a wide range of human infections. Aminoglycosides include streptomycin, kanamycin, and dibekacin. These compounds inhibit protein synthesis in susceptible bacteria, although resistance and the risk of serious side effects, such as ear and kidney damage, have lessened their use. In view of the current concerns over the global rise in antibiotic-resistant microorganisms, there is a renewed interest in the new aminoglycosides that are approved for the effective treatment of resistant microorganism infections. The aminoglycoside, arbekacin (6), is a derivative of dibekacin and is effective against most resistant strains. It showed remarkable antimicrobial effects on methicillin-resistant Staphylococcus aureus (MRSA), a leading cause of nosocomial infections. Since its introduction, arbekacin (6) has been clinically used as one of the most effective antibiotics in the treatment of MRSA infections (11).

ity as a novel anti-convulsant (8), and more recently, it was approved to treat epilepsy. As one of the new generation of anti-epileptics, it is used to treat partial-onset seizures as adjunctive therapy in children. In 2001, topiramate (5) was also approved as adjunctive treatment for primary, generalized, tonicclonic seizures in which a temporary loss of consciousness and muscle control occurs in young adults and children. Off-label prescribing of topiramate (5)as a migraine preventive is known, and this is of concern especially in light of a side effect consisting of acute myopia-associated glaucoma that has been reported in patients receiving topiramate (5)(9).

Carbohydrate-Based Therapeutics

In contrast to antibiotics, viral diseases present an especially difficult challenge to the pharmaceutical industry. This is evident simply because there are not enough good antiviral drugs available. The discovery and development of novel antiviral therapeutics with new mechanisms of action is a notable event. Thus, it is especially significant that the carbohydrate chemistry of sialic acid was used in combination with neuriminidase glycobiology to produce not one but two new antiviral therapeutic drugs for the treatment of influenza. Neuraminidase cleaves terminal sialic acid (7) residues from carbohydrate moieties on

the surfaces of host cells and influenza virus envelopes. The enzymatic process promotes the release of progeny viruses from infected cells (12). Neuraminidase inhibitors are analogs of sialic acid (7). They work by binding to the active, catalytic site of neuraminidase that protrudes from the surface of influenza viruses. Viral hemagglutinin binds to the intact sialic acid residues, which results in viral aggregation at the host cell surface and a reduction in the amount of virus that is released to infect other cells (13). The sialic acid analogs or glycomimetics, zanamivir (8) and oseltamivir (9), are members of a new class of antiviral agents that selectively inhibit the neuraminidase of both influenza A and B viruses. Zanamivir (8)is an inhaled anti-viral drug for the treatment of uncomplicated types A and B influenza, the two types most responsible for flu epidemics (14). Patients need to start treatment within 2 days of the onset of symptoms, and the drug is less effective in patients whose symptoms do not include a fever. Zanamivir (8)is a powder that is inhaled twice

a day for 5 days, and it is not recommended in patients with severe asthma or chronic obstructive pulmonary disease. In contrast to zanamivir (8), oseltamivir (9) is an oral anti-viral drug for the treatment of uncomplicated influenza in patients whose flu symptoms have not lasted more than 2 days (15, 16). This product treats types A and B influenza; however, the majority of patients in the United States are infected with type A. E f f i c a c y of oseltamivir (9) in the treatment of influenza in subjects with chronic cardiac disease and/or respiratory disease has not been established. Oseltamivir (9) is also approved for the prevention of influenza in adults and adolescents older than 13 years. Efficacy of oseltamivir (9) for the prevention of influenza has not been established in immune-compromised patients. In the treatment of influenza, further comments on the mechanism of action of neuraminidase inhibitors are warranted. Neuraminidase inhibitors occupy the active site of neuraminidase by binding to it more readily than sialic acid (7), the sugar residue it normally cleaves (17). Based on structure-activity relationship of analogs, it is apparent that sialic acid (7) is held in the active site through its glycerol and carboxylate groups, which

&
k
i

2 Carbohydrate-Based Drugs

1
L

i
r
I
I

i
i

!

jg
t

'
2

form hydrogen bonds with amino acids in the active site. Zanamivir (8)adds other hydrogen bonds by replacing the hydroxyl group of a sialic acid derivative with a large, positively charged guanidine moiety, which forms strong attachments to two negatively charged amino acids at the bottom of the target protein cleft. Oseltamivir (9) is a prodrug that is mnverted to free acid in the body, and the resulting molecule retains the carboxylate bonds found in sialic acid (7). It also makes use of a hydrophobic group that apparently binds better than a glycerol of sialic acid (7) that it replaces. The antibiotics zanamivir (8)and oseltamivir (9) are outstanding examples of new carbohydrate-based therapeutics. Historically, the only therapeutics for sepsis were antibiotics. People suffering from sepsis are unable to regulate inflammation and coagulation, leading to tissue damage, organ failure, and sometimes death. Trauma, surgery, burns, or illnesses like cancer and pneumonia can trigger the condition. Sepsis is second to heart attack as the leading cause of death with intensive care unit patients. Drotrecogin alfa is a genetically engineered version of the human activated protein C molecule, which helps balance the forces behind sepsis, including inflammation, coagulation, and suppression of fibrinolysis (18). Drotrecogin alfa differs from the natural form of activated protein C by specialized chemical attachments composed of complex carbohydrates that are not found in other biotech treatments such as insulin. Protein C is a serine protease and naturally occurring anticoagulant that plays a role in the regulation of homeostasis through its ability to block the generation of thrombin production by inactivating factors V a and VIIIa in the coagulation cascade. Human protein C is made in vivo primarily in the liver as a single polypeptide of 461 amino acids. The protein C precursor molecule undergoes multiple post-translational modifications. The reasons why drotrecogin alfa is included as a carbohydrate-based therapeutic is not only because it is a glycoprotein but also because to be fully active the core protein C must be glycosylated by the addition of four Asn-linked oligosaccharides. Other post-translational modifications required for activity include the addition of nine y-carboxy-gluta-

mates and one erythro-0-hydroxy-Asp, and the removal of the leader sequence and the dipeptide Lys 156-Arg 157. Without such posttranslational modifications, protein C is nonfunctional. Drotrecogin alfa is a good example of the role saccharides may play in the activity of glycoproteins. Several new glycoconjugate vaccines have recently been approved, including those made from Pneumococcal, Haemophilus B, and Typhoid Vi capsular polysaccharides. This type of carbohydrate-based drug involves carbohydrate natural products isolation and serves as an example for synthetic oligosaccharide glycoconjugates in research and development. Hib vaccines as a class have virtually eliminated bacterial meningitis from the United States, most European countries, and a growing number of developing nations (19). The bacterium primarily responsible is Haemophilus influenza type b (Hib). Four different vaccines for Hib are licensed in the United States. All these vaccines are composed of Hib capsular polysaccharide conjugated to a protein carrier. All four vaccines use a different carrier. Polyribosyl ribitol phosphate (PRP) capsule is an important virulence factor that renders Hib resistant to phagocytosis by neutrophils in the absence of specific anti-capsular antibody. Antibody directed against PRP capsule of Hib is primarily responsible for host resistance to infection. The Hib vaccine is composed of the purified, capsular polysaccharide of Hib. Antibodies to this antigen correlate with protection against invasive disease. Prevnar is a glycoconjugate heptavalent pneumococcal vaccine for young children made from the polysaccharide capsule of the pneumococcus (20). Although older vaccines made from pneumococcus capsular polysaccharide are available for adults, young children cannot make antibodies to the capsule. By coupling the pneumococcal capsular polysaccharide with a protein carrier, a vaccine was created that will trigger an infant's immune system to produce antibodies.
2.1.5 Thrombosis. Heparin was the first polysaccharide-based drug to find widespread application in humans. Isolated from mammalian tissues as a complex mixture of glycosami-

'

Carbohydrate-Based Therapeuti

noglycan (GAG) polysaccharides, heparin has been used clinically since 1937 to treat thrombosis because of its powerful anti-coagulant activity. Heparin acts by enhancing the ability of antithrombin (AT) to inactivate thrombin and factor Xa, enzymes that promote coagulation (21). Low molecular weight (LMW) heparins are produced by chemical or enzymatic depolymerization of heparin, and have a chain length of 13-22 saccharides with a mean molecular weight of 4-6.5 kDa. LMW heparins act through AT inhibition of factor Xa, where only the pentasaccharide high-affinity binding sequence is required (22). Although lacking full anti-coagulant activity, LMW heparins are at least as effective as heparin in the prevention of deep venous thrombosis (DVT) and subsequent venous thromboembolism (VTE) (23). LMW heparins have more predictable activity than heparin, longer duration of action, do not require the measurement of anticoagulation, and can be self administered subcutaneously. However, LMW heparins cannot be effectively reversed by protamine, and duration of action is prolonged in renal failure. In higher doses, they have been shown to be equally effective with intravenous heparin in unstable angina, pulmonary embolism, and ischemic stroke. LMW heparins include dalteparin (24), enoxaparin (251, nadroparine (261, ardeparin (27), and danaparoid (28, 29). The progression from the isolated mixture heparin to the depolymerized LMW heparins has culminated in a synthesized single chemical entity that can act like heparin and the LMW heparins with fewer side effects. Fondaparinux (10)is the first synthetic, selective factor Xa inhibitor. This pentasaccharide agent is significantly better than the Na03S0

LMW heparin, enoxaparin, for preventir VTE after major orthopedic surgery (30). TI show thi results from a meta-analvsis " fondaparinux (10)is associated with an ove all risk reduction versus enoxaparin for tI prevention of venous thromboembolism. Of a the LMW heparins, enoxaparin is widelym r garded as the treatment standard for VT prophylaxis. Fondaparinux (10) is an injec able solution for the prevention of DVT and the only antithrombotic agent approved in t i United States for hip fracture surgery. DVT is caused by a blood clot forming in deep vein in the leg. It may cause pain an swelling and enlargement and discoloration I the veins. The clot can grow in size and bloc other veins. In the most serious thrombot event, portions of the clot may break away an travel through the veins to the lungs, leadir to a life-threatening pulmonary embolism. Fondaparinux (10)exhibits antithrombot activity, which is the result of AT-mediate selective inhibition of factor Xa. By selective: binding to AT, fondaparinux (10)activates t k innate neutralization of factor Xa by AT. Th neutralization interrupts the blood coaguli tion cascade and thus inhibits thrombin fo: mation and thrombus development. Note that heparin, LMW heparins, an c fondaparinux (10)all consist of long linear s quences of alternating hexosamine (D-gh cosamine or D-galactosamine) and hexuron: acid (D-glucuronic acid or L-iduronic acic units carrying sulfate substituents at varioc positions.
2.1.6 Inflammation. Chemically related t

heparin is the GAG, hyaluronic acid. It is natural component of the extra-cellular mi trix found in connective tissues. Clinical inver tigations have demonstrated the safety and e

HOllnnm..

3
HO

NaOzC

Na03S0 '"""OH l

~HO l

o

~
5

l

L
= :
5

~
HO
s

~
%

~

~

NHSO~N~

Na03S0

HO 5 NHS03Na

0S03Na

NHS03Na

(10)

; 2 Carbohydrate-Based Drugs

ficacy of hyaluronic acid in the treatment of osteoarthritis of the knee and other large joints. In addition to restoring the elasticity and viscosity of the synovial fluid, hyaluronic acid modulates acute and chronic inflarnmation processes both in animals and human beings (31). Specifically, hyaluronic acid interacts with endogenous receptors such as CD 44, ICAM-1, and RHAMM and may play an important role in controlling a variety of cellular behaviors, such as the migration, adhesion, and activation of pro-inflammatory cells, chondrocyte maturation or differentiation, and matrix synthesis in the cartilage microenvironment (32). A highly purified, high molecular weight, high viscosity injectable form of hyaluronic acid (Orthovisc) is available to improve joint mobility and range of motion in patients suffering from osteoarthritis of the knee (33). Orthovisc is injected into the knee to restore the elasticity and viscosity of the synovial fluid.
2.1.7 Enzyme Replacement. In contrast to

all the drugs discussed thus far, carbohydrate enzyme replacement therapeutics is different in that the enzyme drugs are not primarily structurally related to carbohydrates, but play an important role in saccharide or glycoconjugate processing. Coincidentally, the replacement strategy requires high affinity receptormediated uptake and delivery to lysosomes that is regulated by specific oligosaccharide side-chains of these enzymes. Carbohydrate enzyme replacement therapy is the administration of exogenous enzymes to patients that have defective saccharide-processing enzymes that result in an accumulation of harmful metabolic products. Metabolic diseases often have genetic causes and some of these diseases involve saccharide or glycoconjugate metabolites. Two examples are Gaucher's and Fabry's disease. Gaucher's disease is a glycolipid storage disease and results from genetic mutations that can either slow or prevent the breakdown of certain glycolipids. Patients with Gaucher's disease are born with a deficiency in the enzyme glucocerebrosidase that results harmful quantities of a fatty substance called glucocerebroside to accumulate in the spleen, liver, lungs, and bone marrow.

Fabry's disease is an inherited genetic disorder caused by deficient activity of the lysosomal enzyme a-galactosidase A. In patients with Fabry's disease, harmful quantities of globotriaosylceramide accumulate in the kidney, heart, nervous system, and blood vessels. Imiglucerase is a recombinant form of the enzyme glucocerebrosidase and it is approved to treat type 1 Gaucher's disease. The therapy is highly effective, but it requires a 2-h intravenous infusion as often as three times a week and is expensive. Agalsidase alfa is a human a-galactosidase A produced by genetic engineering technology in a human cell line. Patients can receive agalsidase alfa every other week in a 40-min intravenous infusion at home rather than in a hospital setting. Although Table 7.1 does not contain all the new carbohydrate-based drugs introduced into the market in the last decade, it should be representative. The choice of what to include in Table 7.1 is also arguable. Nevertheless, it is impressive that in the last decade, over 20 carbohydrate-based drugs have been approved.
2.2
Therapeutics in Development

In the past decade there has been significant progress in development of carbohydratebased therapeutics. Also, new methods for the large-scale production of carbohydrates and their analogs are allowing the thorough evaluation of their safety and efficacy in human trials. Challenges that face the development of carbohydrate-based therapeutics continue to include the need to procure sufficient amounts of complex carbohydrates, which is often difficult and expensive, whether they involve synthesis or isolation from natural sources. Additional challenges include the low bioavailability of orally delivered carbohydrates and the inability of many animal models to provide data relevant to human disease. In contrast to the challenges, advantages in favor of carbohydrate-based therapeutics include low toxicity and high structural diversity. This section, like marketed drugs above, is divided by molecular weight and organized into therapeutic classes. The therapeutic classes are grouped as follows: gastrointestinal system, central nervous

212

Carbohydrate-Based Therapeutics

Table 7.1 To Market-Last Decade
Generic Name (Brand Name) Acarbose (1)(Precose, Glucobay, Prandase) Voglibose (2) (Basen, Glustat) (AO-128) Miglitol(3) (Glyset, BAY m 1099) Dolsarnate (4) (Flavalfate, F3616M) Topiramate (5) (Topamax) Arbekacin (6) (Habekacin) Zanamivir ( 8 ) (Relenza) GG167 Oseltamivir (9) (Tamiflu) Pneumococcal heptavalent vaccine (Prevnar) Haemophilus b (ActHIB, OmniHIB) Typhoid Vi (Typhim Vi) Drotrecogin alfa (Xigris) Dalteparin (Fragmin) Enoxaparine (Lovenox) Nadroparine (Fraxiparine) Ardeparin (Normiflo) Danaparoid (Orgaran) Fondaparinux (10) (Ariixtra) Hyaluronic acid (Orthovisc) Imiglucerase (Cerezyme) Agalsidase alfa (Replagal) Indication Diabetes Diabetes Diabetes Gastric ulceration Anti-convulsant, anti-epileptic Anti-microbial Company Bayer AG Takeda and Abbott Bayer Faes Ortho-McNeil, Johnson & Johnson Meiji Seika Japan in 1994 US. in 1996 Spain in 2000 U.S. in 1996,1999 Japan in 1997 US. in 1999 Hoffmann-La Roche Gilead Wyeth SmithKline Beecham, Pasteur Merieux Pasteur Merieux Lilly Pharmacia Upjohn Aventis France in 1986 Wyeth Organon Organon and SanofiSynthelabo Anika, Zimmer Europe Genzyme Transkaryotic Therapies U.S. in 1997 U.S. in 2000 U.S. in 2001 Europe and Canada in 1998 US. in 1994 Europe in 2001 U.S. in 1999 Approval Country and Year

Conjugate vaccine Conjugate vaccine Conjugate vaccine Sepsis Anti-coagulant, antithrombotic Anti-coagulant Antithrombotic Anti-coagulant, antithrombotic Anti-coagulant, antithrombotic Anti-coagulant; antithrombotic Anti-coagulant; antithrombotic Viscoelastic supplement Gaucher's disease Fabry's disease

France in 1996 U.S. in 1995 US. in 2001 US. in 2000

system, infection, cancer, inflammation, and lysosomal storage. The major difference from the marketed drugs is that the diabetes class is not represented. With diabetes, it is not clear why no new mechanism drug is being developed to follow up the success of a-glucosidase inhibition. In place of diabetes, cancer therapeutics is added with a focus on cancer vaccines that are in clinical development.

2.2.1 Gastrointestinal System. Oral cyto-

protective agents like dosmalfate (4) have not been followed in development presumably because the mechanism of action is not well understood. In place of cytoprotection in gastrointestinal drug development is motilin antagonism. Mitemcinal (11) is an agonist of motilin, a peptide hormone that plays a role in the con-

2 Carbohydrate-Based Drugs

21 3

tractile movement of the gastrointestinal tract (34). Mitemcinal(l1) is an erythromycin A derivative that is being studied in phase I1 clinical trials to assess its effect on the recovery of gastrointestinal motility. Gastrointestinal motor-stimulating activity is a side effect of erythromycin, and other compounds based on erythromycin, such as EM574, are in development (35).
2.2.2 CNS. Because CNS active drugs tend

surprising that a glycolipids-based drug would be in development. What seems to be the problem, however, is glycolipid drug supply. The development the glycolipid ganglioside GM1, called Sygen (12), for treating Parkinson's disease was reported (36). This complex molecule is currently isolated from natural sources and studies into its large-scale synthesis are underway. A phase I1 clinical trial for acute spinal cord injured patients was started in 1997. Results from single center studies, as well as efforts in experimental spinal cord injury and stroke, indicated that Sygen (12) promoted improvement in neurological status. However, there were problems. Early forms of the drug were often impure and there were several severe complications from the use of the European version of the drug. Also, higher doses of Sygen (12) were associated with increased risks in that there was a small incidence of allergies and an increase in cholesterol that was related to being on the drug for a period of time (37).
2.2.3 infection. Anti-infective carbohydrate-

to have appreciable lipid solubility, it is not

based therapeutics in development include antibiotics and antivirals. The emergence of multi-drug resistant bacteria has increased the need for new antibiotics or modifications of older antibiotic agents. The modification of known antibiotics may be a productive way to combat multi-drug resistant bacteria.

given in a once-a-day dosage. no further information became available." The other silica-linked oligosaccharide drug. SC-48334 (15)was not found to be of any ma- . allowing for use by the elderly and young children. showed that the balance between efficacy and safety did not justify further development of the product. Doses of peramivir (14) ranging from 100 to 400 mg/ day produced a dose-dependent. The total synthesis everninomycin may allow semi-synthetic variants of this antibiotic to be made that are active versus multi-resistant bacteria (39). which is produced by Shigella dysenteriae. Everninomycin (13). was for the treatment of C. and can be made into a liquid form. gastrointestinal anti-infective therapies were advanced to clinical trials. Both compounds were discontinued in development. Synsorb Pk was developed for the treatment of E. The use of avilarnycin as a growth promoter has created a reservoir of isolates with decreased susceptibility to everninomycin (13)before this antibiotic was finally approved for human use (38). silica-linked. however. Indeed. may have been a victim of antibiotic overuse. Phase I11 clinical studies. can cause serious clinical complications in humans infected by these organisms. Shiga toxin. coli infections to prevent hemolytic uremic syndrome. difficile-associated diarrhea. carbohydratebased. Another candidate. a clear dose response is seen for the primary endpoint (area under the curve versus viral titer). coli. Peramivir (14) is a neuraminidase inhibitor that is reported to have both an excelient safety profile and effective against influenza A and B. Synthetic Gb3 analogs covalently attached to insoluble silica particles can competitively adsorb toxin from the gut and can serve as a starting point for the adsorbent carbohydrate drugs.an intravenous antibiotic. Perarnivir (14) at 400 mg once daily for 5 days results in a highly significant reduction in mean viral titer area under the curve and duration of viral shedding compared with placebo. further development of SC-48334 (15) (N-butyl-deoxynojirimycin) for HIV infection was discontinued. virological efficacy response compared with placebo. and the high incidence of complications such as acute kidney failure in children correlates with the expression of Gb3 in the - pediatric renal glomerulus. The target E. was terminated based on a phase I1 clinical trial data in which NE-1530 showed no significant benefit compared with placebo (40). NE-0080. Another oligosaccharide. It is orally active. avilamycin. two orally delivered.although promising. Synsorb Cd. Despite promising early data on activity. The functional toxin receptor on mammalian cells is the glycolipid Gb3 [a-DGal(1 +4) P-D-Gal(1 +4) P-D-Glc(1 + O-ceramide)]. has been used as a growth promoter for food animals in the European Union. and the homologous Shiga-like toxins (SLTs) of E. coli variant produces verotoxin. NE-1530. was in development for hospitalized patients with resistant gram-positive infections. Trials in AIDSIARC patients indicated that the drug had no significant effect on p24 antigen levels or CD4 cell counts (42). which is structurally similar. the toxin in "hamburger disease. Development of the pediatric ear infection drug. was being evaluated as a treatment for stomach ulcers and other gastric complaints caused by the bacterium Helicobacterpylori. In a clinical challenge study.214 Carbohydrate-Based Therapeutics Everninomycin (13). but after phase I trial indicated the safety (41).

which appears on the surface coat of E.4 Thrombosis. Neutralase (heparinase I) is being developed as a replacement to protamine that is currently used to reverse the effects of heparin. which binds to heparin to form large macromolecular complexes. coli 0157 triggered a strong protective response in healthy volunteers (45). fragments.2. for the treatment of lysosomal storage disease (see Section 2. No further clinical development has been forthcoming with Celgosivir (16). coli 0157. neutralase enzymatically breaks heparin into small. Celgosivir (16) demonstrates inhibitory effect on glucosidases and has antiviral activity against HN-1 and against murine leukemia virus. coli 0157. The vaccine has two components. particularly the elderly and young children who are most vulnerable to infection. 2. mostly inactive. Preliminary efficacy measurements showed good results for one-third of the patients. 80% of the 87 volunteers had developed large numbers of antibodies that killed laboratory cultures of E. In contrast to protamine. a carbohydrate called the 0-specific polysaccharide.2. and a protein from Pseudomonas aeruginosa. Within a week of vaccination. it is being developed under a different name. The enzymatic cleavage jor benefit in reducing viral load in a phase I1 trial (43). Although (15) did not find use in HIV. Pre-clinical studies indicate that neutralase can also inactivate the anticoagulant effects of the low LMW heparins and the synthetic pentasaccharide fondaparinux (10) (see . stable results for one-third.i I 2 Carbohydrate-Based Drugs kL. By cleaving heparin. A prototype vaccine against E. the bacterium that causes pneumonia. heparinase I inactivates heparin's anticoagulant effects. vevesca.7). Celgosivir (16) was tested alone or in combination with AZT orally in AIDS patients with CD4+ counts between 100 and 500 cells1 at specific sites in heparin could be important for anticoagulation therapeutics. The response may be rapid enough to protect people once an outbreak has begun. This is an example of how understanding the glycobiology and using this knowledge can be used to discover efficacy in another indication. and rather poor results for the other one-third of the patients (44).

a-1. Bonds between P-selectin and PSGL-1 primarily mediate the rolling phase of the adhesion cascade. PSGL-1 binds to a selectin-P subtype on platelet cells. diarrhea with or without bleeding. and neutrophil membranes.2. This fusion peptide is active as a competitive inhibitor of PSGL-1. The 0-linked glycans displayed by PSGL-1 must have two specific post-translational modifications. A major ligand for P-selectin is P-selectin glycoprotein ligand-1 (PSGL-11. PSGL-1 is a 240-kDa homodimer consisting of two 120-kDa polypeptide chains. Until then. in the initial inflammatory response. A partial form of recombinant human PSGL-1 has been covalently linked to IgG to give rPSGL-Ig (47). cylexin was the most advanced selectin-binding inhibitor.5). inhibiting P-selectin may also prove to be useful for stopping the spread of tumors.2. OP2000 is an oligosaccharide product derived from heparin with antithrombotic and anti-inflammatory properties and is in development for inflammatory bowel disease (IBD). and product development has been discontinued for MI. The structure of functional PSGL-1 includes the sLex oligosaccharide.6 Cancer. The development indications for rPSGL-Ig are not only in inflammation but also cancer. The trial was negative. In addition. OP2000 is a product of the chemical cleavage of heparin and has the comparatively low molecular weight of 2. Clinical pharmacology of heparin in ulcerative colitis support the idea that heparin can safely induce remission in IBD patients. Several carbohydrate-based therapeutics that have now reached clinical trials seek to estab- .Carbohydrate-Based Therapeutics Section 2. often causing recurrent abdominal pain. causing red blood cells to stick to leukocytes and create blood clots. which affects the lowest portion of the small intestine. Cylexin showed no benefit over placebo in a 138-patient trial for the treatment of reperfusion injury in infants undergoing cardiopulmonary bypass to facilitate the surgical repair of life-threatening heart defects. 2. As an inhibitor of neutrophil-endothelial cell adherence. and this ligand is expressed on the surface of monocyte. which results in inflammation of the large intestine. and threonine residues typical of mucin glycoproteins. and in an animal model. 2. pretreatment with rPSGL-Ig reduces both the thrombo-inflammatory and neointimal proliferative responses (48). P-selectin is a major component in the early interaction between platelets and neutrophils. multi-center placebo-controlled. Cylexin was studied for preventing reperfusion injury in infants (46). IBD is a group of chronic inflammatory disorders of the intestine of unknown cause. serine. Metastasizing tumor cells recognize P-selectin on platelets and use them as a protective shield against immune system cells.3-fucosylation and a-2. fever. Clinical observations suggest that IBD may result from increased clotting activity.5 kDa. The PSGL-1 gene en- codes a transmembrane polypeptide rich in proline.3-sialylation. A double-blind randomized.5 Inflammation. and endothelial cells. Two forms of IBD are Crohn's disease.1. cramps. The goal of a cancer vaccine is to make the host immune system responsive to antigens characteristic of cancer cells. dose-ranging phase I1 study on the efficacy and safety of recombinant rPSGL-Ig in patients with acute myocardial infarction (MI) was assessed by positron emission tomography. Inhibiting P-selectin may prevent not only ischemia reperfusion injury but also the subsequent clotting events that can re-occlude a vessel that's just been opened. OP2000 is a potent anti-clotting agent like other LMW heparins. Investigators have observed evidence of increased clotting in the bowel and other organs during flares of IBD. and fatigue. lymphocyte. rPSGL-Ig was in early clinical development for the treatment of ischemiareperfusion injury. and ulcerative colitis. to be active as a counter-receptor for P-selectin. Neutralase is able to neutralize these agents because it cleaves the recognition site for anti-thrombin. Clinical trials with cylexin were halted based on disappointing results from its ongoing phase II/III clinical trial. The development of carbo- hydrate-based vaccines designed to activate Tcells targeting of cancer cells has been pursued for some time.

Current best therapy is surgical removal of high-risk melanoma and treatment with Interferon alfa. and antibody-dependent cellular cytotoxicity (54).! S E b i 2 Carbohydrate-Based Drugs lish the molecular vaccine design in which a cell surface antigen activates T-cells (49). mucin-1. The cell surface gangliosides GM2 and GD2 are expressed on several cancers. it showed a dose-dependent inhibition of tumor growth in the BC1 syngeneic rat mammary adenocarcinoma (56). IgM antibodies bound to the cancer cells. In pre-clinical studies. An effective treatment for the deadly skin cancer melanoma is clearly needed. complement-mediated cytotoxicity. GM2 is immunogenic. GMK vaccine is capable of producing a specific immune response against a defined and chemically purified melanoma antigen. and some patients showed evidence of cancer cell death (52). in a group of 27 women with metastatic breast cancer. Other carbohydrate-based drugs in development to treat cancer include a sulfated oligosaccharide. globo H. is being developed for several other cancer indications. The development of GMK and MGV cancer vaccines seems to be on hold. Theratope is a cancer vaccine that consists of modified tumor glycopeptide antigens for patients with advanced breast cancer. The vaccine uses a carbohydrate antigen known as STn. PI-88 is a sulfated oligosaccharide. The vaccine was established as safe and capable of inducing specific antibody response to globo H. part of a larger antigen. globo H raised blood levels of IgM antibodies. sarcoma. and neuroblastoma. and. The antigen is attached to Keyhole Limpet Hemocyanin (KLH) protein to increase the immune response to the vaccine. Melanoma patients have a 50-80% chance of relapse without adjuvant treatment. and other epithelial cancers. gastric. the immune response to the vaccine based on synthetic globo H-protein conjugate also recognized naturally occurring cancer-derived glycoprotein. GD2 is immunogenic when coupled to KLH and mixed with the adjuvant QS21. Earlier studies in humans provide strong evidence of the development of an anti-tumor T-cell response after immunization with a cancer-associated carbohydrate antigen (50). Interferon-treated patients were 50% less likely to experience a return of their disease and 52% less likely to die than were GMK-treated patients (55). In 16 of these patients. lymphoma. and the expressed levels of these antigens are magnified in the context of such tumors. and an azasugar. Globo H occurs on cell surface glycoproteins of breast. The globo H vaccine is being tested in clinical trials to determine whether it will block the reappearance or progression of cancer. PI-88 also inhibits heparinase secreted by a variety of cells. is found on cancer cells and is an antigen for vaccine design. thus prevent- . and induction of anti-GM2 antibodies correlates with improved survival in stage 1 1 1 melanoma patients with no evidence of disease after surgery. GAG mimetic that inhibits growth factors and heparanase produced by tumors and has antiangiogenic and anti-metastatic properties. as in the case of Theratope. including colorectal. The hexasaccharide. AntiGM2 antibodies in melanoma patients induced by the GMK demonstrate binding to melanoma cells. and in a rat model. Chemists have synthesized globo H. GMK vaccine was compared with high-dose Interferon alfa 2b in high-risk melanoma patients who had undergone surgery. Similarly. attached it to KLH to augment the antigen's immunogenicity (51). small-cell lung. The GMK vaccine is in development for melanoma and contains the ganglioside GM2 linked to the carrier protein KLH and combined with the adjuvant QS-21. Thus far. which has as an endpoint of increased survival. Theratope is in phase I11 human trial. Another vaccine. including melanoma. PI-88 retarded the growth of primary tumors by inhibiting angiogenesis. More importantly. This inhibition is thought to be the result of its ability to inhibit endothelial cell proliferation by preventing basic fibroblast growth factor (bFGF) and vascular endothelial growth factor (VEGF) from binding to the receptors on endothelial cells. MGV. prostate. found on breast cancer cells. prostate cancer patients who relapsed after surgery or radiation therapy are being treated with the globo H tumor vaccine (53). a pectin derivative.

7) indicates that the patients were successfully maintained on oral therapy alone during the study period. as a surrogate of activity in patients. OGT 918 is a small molecule drug and as such can be given orally.2. the side effects and long-term toxicitv " must be determined.1. the dose limiting effect of GCS-100 is pulmonary insufficiency. N-butyldeoxynojirimycin) is in clinical development for the treatment of Gaucher's and Fabry's diseases. GD0039 (17) abrogates much of the toxicity of commonly used chemotherapeutic agents in both healthy and tumor-bearing mice. GD0039 (17) is in phase I1 clinical age diseases (LSDs) are genetic errors of metabolism resulting primarily from the absence of an enzyme whose target is a substance to be lowered in cellular tissues. Both small molecule enzyme inhibitors and large molecule enzyme replacement approaches are being taken to treat LSDs.Carbohydrate-Based Therapeutics ing the degradation of extra-cellular matrix and the release of growth factors.7). Carbohydrates have demonstrated an increasing role in tumor biology. such as Fabry's and Gaucher's diseases (discussed in Section 2. such as macrophages and splenocytes. The trial will also gather information about PI-88's ability to inhibit the blood vessels formation in tumors. GD0039 (17) blocks pulmonary colonization by tumor cells and stimulates components of the immune system. A multi-center phase I1 clinical trial to further investigate the efficacy and safety of PI-88 in patients with multiple myeloma was started.7 Lysosomal Storage. The carbohydrate GCS-100 is a pectin derivative that binds to the galectin 3 receptor. is used to examine the role of glycan structures in metastasis (58). Phase I11 results with aldurazyme warranted filing for marketing approval in 2002. It will be of interest in the future to contrast the success of enzyme replacement that requires infusion with orally delivered drugs for the treatment of Gaucher's and Fabry's diseases.1. Furthermore. Analysis of a 6-month study to investigate the potential of OGT 918 treatment in Fabry's patients who had been receiving enzyme therapy (see agalsidase alfa. is a biochemical marker related to the disease. or in most cases. shortened lifespan. The build-up of these substances causes a loss of function in one or several crucial areas of the body and may result in mental and physical disability. 2. Pectin is the methylated ester of polygalacturonic acid. Aldurazyme is an enzyme replacement therapy for the treatment of MPS I disease. A partial list of LSDs includes glycolipidosis disorders. Those Gaucher patients that reached the 24-month time point have continued into their third year of therapy. Lysosomal stor- sidase 11. Oligosaccharide moieties of cell-surface glycoproteins are involved in recognition events associated with cancer metastasis. Oral drugs are clearly more convenient. with a known correlation between the expression of the galectin 3 receptor and the propensity for metastatic spread in pre-clinical models (57). an inhibitor of Golgi a-manno- trials in metastatic renal cancer and 5-FU resistant colorectal cancer. GD0039 ( 1 7 ) . and mucopolysaccharidosis (MPS). MPS I is a genetic disease caused by the deficiency of . Long-term data from type 1 Gaucher patients at 24 months show a progressive improvement on the results previously obtained after 12 and 18 months of treatment with OGT-918. OGT-918 (15) (vevesca. however. Chitotriosidase activity. GCS-100 is able to inhibit the metastatic spread of tumor lines to the lung and is able to shrink large tumors. In pre-clinical models. such as MPS I and VI diseases. Section 2. Phase IIa data in failed metastatic pancreatic and colorectal cancer patients warranted proceeding into larger studies. GCS-100 is in phase I1 clinical trials for pancreatic and colorectal cancer. In pre-clinical toxicology studies.

causing them to change or lose function. including the inactivation of certain enzymes. Specific binding molecules are many and include polyglycosylceramide (611. 15 of 26 carbohydrate-based therapeutics (58%)are still in development. and joint deformities.which are implicated in the development of diabetic kidney disease. the enzyme was well tolerated. In contrast. The complete genome sequence of H. and there were no drug-related serious adverse events and no significant allergic reactions to the infusions. During the course of the disease.3. however. GAGs are only partially broken down. reflecting a complex interrelation with the environment.1 Diabetes. at least four of the therapeutics in development will go on to be successful drugs. Long-term observation of the development process in the pharmaceutical industry indicates that roughly one of every six therapeutics (17%) submitted as investigational new drug applications go on to be filed successfully as New Drug Applications with the U. more carbohydrate-based therapeutics in development will be discontinued. 2. 3DG is also a proven factor in the development of advanced glycation end products (AGES). 3DG is a highly reactive molecule that can cross-link proteins. Arylsulfatase B (N-acetylgalactosamine 4-sulfatase) is a deficient enzyme in patients with the genetic defect disease MPS VI.aggressively advertise "products in the R&D pipeline" without disclosing specific molecular information that describe the potential products.3. starting with diabetes research that continues to attract new carbohydrate-based agents. impaired vision and hearing. and carbohydrate residues build relentlessly in the lysosomes of cells. 2. the build-up of GAGs results in inhibited growth and mental development. Urinary GAG excretion was reduced by a mean of 70% in the high dose group and 55% in the low dose group. They contain no nucleic acid. has emerged as the causative agent in chronic active gastritis and peptic ulcer disease (60).1 and 2. H.2 Carbohydrate-Based Drugs aa-iduronidase.The bacterium is unique in its carbohydrate-binding complexity. The following is organized in the approximate order as Sections 2. About 3000-4000 patients in developed countries have MPS I.3 difficult.3 CNS. Large pharmaceutical companies that are a traditional source of new drugs. if the industrial average holds. arteriosclerosis. an enzyme normally required for the breakdown of GAGs. In a phase I trial. have a mass of 30 . arylsulfatase B. Aryplase is a specific form of the recombinant human enzyme. leading to severe disability and early death.2 . small biotechnical companies that have focused on these technologies . Areas of special interest in carbohydrate drug research include the treatment of Helicobacter pylori gastritis and prion-induced CNS disease. Prions are composed exclu- Research agents are the source of new drugs. including the early publication of structures. pylori-infected patients with promising results (63). The human Agents in Research specific gastric pathogen. The normal breakdown of GAGs is incomplete or blocked if the enzyme is not present in sufficient quantity. pylori reveals a surprisingly large part coding for outer surface proteins. Clearly. pylori. however. In Table 7. Patients with MPS I are usually diagnosed in childhood and get progressively worse. Research on carbohydrate-based agents that may have use as drugs is increasing. and sialic acid conjugates (62). the exogenous supply of arylsulfatase could breakdown the stored GAGs and cellular function could be restored.S. and other diseases. getting specific structural information is more sively of a single sialoglycoprotein called PrP. have not wholly endorsed carbohydrate chemistry or glycobiology. The accumulation of GAGs in the lysosomes of the cell cause MPS I. DYN 12 is a small molecule that reduces plasma concentrations of 3-deoxyglucosone (3DG) in diabetic rats by 50%. Amadorase or the closely related fructoseamine-3-kinase is the putative enzyme responsible for the formation of most of the 3DG in the body (59). 2. Also like a-L-iduronidase replacement. reduced cardiovascular and pulmonary function. 2. Food and Drug Administration. Recently. Urinary G A G excretion is a biomarker for MPS.3.2 Gastrointestinal System.2. Like MPS I. a sialylated receptor analog was evaluated in clinical trials with H.

breast cancer Breast. Elan Cytel Wyeth Biomira Memorial SloanKettering Cancer Center Progenics. Medigen Safescience GlycoDesign Oxford GlycoSciences BioMarin. Genzyme BioMarin Synsorb Pk Synsorb Cd Peramivir (14) (RWJ270201) SC 48334 (15)(N-Butyldeoxynojirimycin) Celgosivir (16) (MDL 28574. Inflammatory bowel disease Reperfusion injury Heart attack Theratope (Sialyl Tn Ag conjugate vaccine) Globo H conjugate vaccine GMK (GM2 KLHIQS-21 conjugate vaccine) MGV (GM2lGD2 KLH QS21 conjugate vaccine) PI-88 GCS 100. N-butyldeoxynojirimycin) Alpha -L. prostate cancer Malignant melanoma Colorectal. breast cancer Fabry's. DRG-0202) E. Kitazato Development Phase I1 Phase I1 Phase I1 Discontinued (phase 111) Discontinued (phase 11) Discontinued (phase I) Discontinued 1 1 ) (phase 1 Discontinued (phase 111) Phase I 1 1 Discontinued (phase 11) Discontinued (phase 11) Phase I Phase I1 Phase I 1 1 Discontinued (phase 11) Discontiued (phase 11) Phase I11 Phase I Sygen (12) (GM-1) Everninomicin (13) (Ziracin.220 Carbohydrate-Based Therapeutics Table 7.SCH 27899) NE-1530 Fidia Schering-Plough Neose Neose Synsorb Synsorb Johnson &Johnson. Incara. small-cell lung.2 Therapeutics in Development - Name Generic (Code) Mitemcinal(l1) (GM611) Indication Gastroparesis. coli 0157 infection Anti-coagulation reversal Antithrombosis. Gaucher's disease MPS I MPS VI Discontinued (phase 111) Discontinued (phase 11) Phase I1 Phase I1 Phase I1 Phase I11 Phase I11 Phase I1 . cancer Multiple myeloma. BioCryst Searle Aventis NICHHD BioMarin Opocrin. BristolMyers Squibb Progenics. gastric. BristolMyers Squibb Progen. colorectal. spinal cord injury Antibiotic Pediatric ear infection Helicobacter pylori infection E. adenocarcinoma Renal. coli 0157:H7 Hemolytic uremic syndrome C. Tumor angiogenesis Pancreatic cancer.iduronidase (Aldurazyme) Arylsulfatase B (Aryplase) Metastatic coloredal. coli 0157 vaccine. Heparinase I (Neutralasel Deligoparin (OP2000) Cylexin HIV E. (GBC-590) GD0039 (17) (Swainsonine) Vevesca 15 (OGT 918.dificile associated diarrhea Anti-viral Company Chugai Takeda. gastroesophageal reflux Chronic gastritis gastroesophogeal reflux Parkinson's disease.

Multiple strains of S. causes illnesses that range from minor skin abscesses to severe pneumonia. A conformational change in the PrP from an a-helix to a P-sheet induces the same a-helix to P-sheet conformation change in the other PrP protein. led to the isolation of a promising vaccine antigen. poly-N-succinyl-p-1-6-glucosamine (PSG) (68). It may provide tools in studies of the molecular basis of the binding to the minor groove of DNA. Variations in the complex carbohydrate structure in the CNS could account for selective targeting in particular areas of the brain. and then selectively cleave the glycosidic linkage to vancosamine. Purified PSG injected into rabbits produced large amounts of PSG antibodies that persisted for at least 8 months. This protein polymerizes into rods possessing the structural characteristics of amyloid. One chemical approach is to sequentially pro- tect the hnctional groups in the molecule according to their reactivities. keratin. However.kDa. inhibit the neurotoxicity of amyloid fibrils (65). has renewed interest in this class of agents (67). and are composed of 145 amino acids. and Down's syndrome. Sulphated GAGS. and the most prevalent species. CreutzfeldtJakob. Usually.the last effective antibiotic against S. Coynebacteria. Clostridia. Modifying vancosamine residue on vancomycin (18)could increase the antibiotic's activity against vancomycin-resistant strains. bone. Amyloid protein is deposited in many human C N S diseases including Alzheimer's. In the absence of evidence for any significant covalent modifications that can distinguish PrPC (a-helix)from PrPSc (P-sheet).such as heparin. involving a convergent synthesis in which preassembled sugar units are attached in specific order to the aglycone. A recent synthesis of olivomycin A (19). and chondroitin. Vancomycin (18)is a rela- tively small glycoprotein derived from Nocardia orientalis. The result is a protected pseudoaglycon. aureus vaccine. thereby determining the rate of PrPSc formation and the specific patterns of PrPSc deposition. aureus. and infections of the heart. This is a permanent conformational change and induces replication. Listeria. bloodstream. aureus are now antibiotic resistant. The goal is to use sulfoxide glycosylation chemistry to reattach vancosamine derivatives and produceglycosylated vancomycin (18)analogs with increased antibiotic activity and activity against resistant microorganisms (66). which lack one of the hydroxyl groups found in most sugars. Olivomycin A (19)is a member of the aureolic acid family and is derived from Streptomyces. Also important is the development of methods for synthesizing and manipulating the unusual monosaccharides found in the aureolic acid family. When PSG antibodies were injected into mice and ex- .3. and joint. The P-sheet-forming peptides aggregate to form amyloid fibrils. and the amyloid fibrils kill thalamus neurons through apoptosis. aureus infections are clearly needed. including a few strains that are partially resistant to vancomycin (IS).focus has shifted to the difference in secondary and tertiary structure between the two forms of the prion protein. and an addition of a phosphatidylinositol glycolipid at the C-terminal. Glycosylation may modify the conformation of PrPC (64). New treatments for S. One approach is to discover an S. Pathogens from the Staphylococcus genus are found in both the hospital and the community. aureus. the consideration that bacteria grown under laboratory conditions may have different antigens than the same bacteria in actual infection. PrP undergoes several post-translational events to become the prion protein including glycosylation at amino acids 181 and 197. . Glycosylation could also affect the affinity of PrPC for a particular conformer of PrPSc. antigens isolated from bacteria grown in the laboratory are new vaccine candidates. which is well known for its activity against Gram-positive bacteria including Streptococci. 2. meningitis. This seems to be mediated by the inhibition of the polymerization of the PrP peptide into fibrils. and Bacillus.4 Infection. These findings provide a basis for using sulphated polysaccharides for CNS therapy. The methodology can be used for synthesis of aureolic acid analogs that may prove useful as cancer chemotherapeutic agents.

Current therapy includes interferon and nucleoside inhibitors (lamivudine) treatment. 20) inhibits hepatitis B virus ability to construct its M envelope protein in human liver cells. however with nppolySA. was conjugated to the immunogenic carrier protein KLH and mixed with the immune adjuvant QS-21(69). nppolySA. Although serogroup A and C vaccines exists. PolySA or the modified version. SCLC patients were vaccinated. after treatment ceased.none of the animals developed an infection. does alter the ability of the virus to be + . and it is able to stimulate an immune response. using the standard woodchuck model for HBV with the woodchuck hepatitis virus. the polySA expressed by serogroup B meningococci makes this pathogen invisible to immune surveillance. N-nonyl deoxynojirimycin (NNDNJ. . treatment with NN-DNJ (20) . relapses are common and the use of immunization against polySA antigenic targets on SCLC cells could improve outcomes.. sera from five of six patients in the nppolySA vaccine group also exhibited reactivity against serogroup B meningococci. viral titers returned to pretreatment levels. and research is being directed toward the generation of vaccines that may have use in these diseases. NN-DNJ (20). Intereptingly. including strains partially resistant to vancomycin (la). Not unexpectedly.Carbohydrate-Based Therapeutics posed to eight different strains of S. Replacing the N-acetyl chemical moiety on polySA with an N-propionyl moiety gives nppolySA. all patients developed antibodies against polySA. Some patients developed strong reactivity against the polySA positive SCLC cell line H69. Apparently. An interesting observation is that small cell lung cancer (SCLC) and N. however. In vivo. and the resulting morbidity and mortality is significant. induced a loss of viremia (70). In vitro. Hepatitis B infection occurs worldwide. meningitidis both have polysialic acid (polySA) cell surface antigens. SCLC is responsive to therapy. aureus. minor cellular carbohydrate processing disruption (6%) results in a greater than 99% reduction in the secretion of hepatitis B virus. . in a dose-dependent manner.After achieving a major response on standard therapy. Serologic analysis indicated no reactivity in the unmodified polySA vaccine.

and P-selectins are highly homologous. including ar- thritis. and reperfusion injury. Presumably the mechanism involves an interaction with a cellular trafficking protein and an improper folding of the altered hyper-glycosylatedenvelope protein. and P-selectins. E-. microbial infection. Each contains a domain similar to calcium-dependent lectins or C-lectins. allergies. and several complement binding protein-like domains (72). Selectins are a family of glycoproteins that are involved in the adhesion of leukocytes to platelets or vascular endothelium (71). it was no surprise that sLex serves as a ligand also for P-selectin (75). recirculation. and autoimmune disease. Because the C-type lectin domains of E. ' . allergic.5 Inflammation. Adhesion is an early step in leukocyte extravasation in which sequelae includes thrombosis. and inflammatory. 2. and inflammation. Adhesion of cells in the immune system is clinically important when considering such indications as cancer. Research to define the native carbohydrate ligands for each selectin receptor continues (73). L-.Carbohydrate-Based Drugs secreted in an infectious form. Three protein receptors. are assigned to the selectin family based on their cDNA sequences. The carbohydrate ligand for E-selectin on leukocytes contains the sLex epitope (74).3. an epidermal growth factor-like domain.

2. Moreover. whereas E-selectin requires presence of only sLex group. macrocyclic sLex analog exhibits a dramatic enhancement (100 times) relative to sLex (82). The sLex group and sulfated tyrosine residue are necessary for Pselectin to bind the cells. After capture.80). L-selectin is the smallest of the vascular selectins. mimics of the tetrasaccharide have been synthesized . the physiological ligand for P-selectin (81).Carbohydrate-Based Therapeutics Another P-selectin glycoprotein ligand is PSGL-1. this adhesion molecule contributes greatly to the capture of leukocytes during the early phases of the adhesion cascade (77). and L-selectin can be recognized. important chemical features for ligand binding to E-. sulfation of sLex does increase its affinity for Lselectin (79. and is constitutively expressed on granulocytes. L-selectin is shed from the leukocyte surface after chemoattractant stimulation. Another major ligand for L-selectin is a sulfated form of sLex. sialyl6-sulfo Lewis X (78). In fact. The cell adhesion mediated by L-selectin and sialyl 6-sulfo Lewis X triggers the action of chemokines and activates lymphocyte integrins. a 74. sulfated oligosaccharides have a higher affinity for proteins than the corresponding non-sulfated precursors. L-selectin is important for lymphocyte homing and adhesion to high endothelial cells of post-capillary venules of peripheral lymph nodes. leading the cells to home into lymph node parenchymas. the blocking of selectin-ligand binding with glycomimetics is a seductive approach to the discovery of novel therapeutic agents. P-. L-selectin interacts with three known counter receptors or ligands. This determinant is expressed on HEV in human lymph nodes where it recruits naive Tlymphocytes expressing L-selectin. In general.to 100-kDa molecule. Considering the forgoing discussion. Similarly. a chemoenzymatic route was used to generate sLex -substituted glycopeptides to elucidate the critical binding epitope of PSGL-1. MAdCAM-1. a 220-kDa glycoprotein found on leukocytes (see Section 2. monocytes. With knowledge of the importance of sLex functional groups for complexation. and many circulating lymphocytes. GlyCAM-1. In a P-selectin inhibition assay.5) (76). By combining the structural data obtained from the identification of the natural selectin ligands with synthetic structure-activity information. PSGL-1 has a sLex group on its carbohydrate side-chains and contains at least one sulfated tyrosine residue near the N-terminus. and CD-34. Structure-activity data on sLex analogs have suggested important chemical features for selectin binding.

in the initial development of such a vaccine. and needed for in vivo experiments in models if that the presentation mode of the vaccine conacute inflammation such as reperfusion injury struct could be simplified because a KLH carand rejection of organ transplants.l-glycosidicbond formation through form. That is. malignant transforma(89). For 1 i : ! a oligosaccharides and glycoconjugates have example. Like other carbohydrate-based antigens. a multi-antigenic.and L-selectin (85). carrier structure. and skin cancers vaccines. A new method coproteins are often presented in clustered for the 1. more closely approximate a tumor cell surface. Basic research on the strucproduction of the carbohydrate P-1. It is thought that tumor-associated glyagainst P. In general. colon. the cell surface carbohydrates. an antigenic carbohydrate apcoupling of protected trimethylsilyl-P-D-galac. The results in mice showed that the clusLarge-scale synthesis of the lead compounds is tered presentation is strongly antigenic. serious regulatory issues have been raised. The three carbohydrate antigens were linked together to form a single molecule that in turn was conjugated to a KLH carrier protein (87). The pothan 50 times more active at blocking E-selecx tin than sLe (84). Synthetic seem normal. particularly if vaccine comA rational design and synthesis of (3-0pounds can be synthesized that resemble the carboxymethy1)-P-D-galactopyranosyl-a-D. 2.6-GlcNAc-branchedN-glycans. MBR1. To described. resulting not only Golgi enzyme required in the biosynthesis of in unusual multiple antigens.2 Carbohydrate-Based Drugs 8 t [ that are potent selectin antagonists (83).surface environment of transformed cancermannopyranoside (22) provided a disacchaous cells.The mutant mice deficient in Mgat5 tural antigens on tumor cell surface. but also in a seP-1. toside and a-mannosyl fluoride in the presand the resulting carbohydrate-amino acid ence of boron trifluoride-etherate is also motif is repeated in a series of residues. rial clustering of antigens. this strategy is the basis for the convergent total synthesis of a tumor-associated mucin motif (86). With Mgat5 knockout mice. For example. These data underscore the tential of inducing active immunity with a tremendous progress made in generating effifully synthetic carbohydrate-based vaccine is cacious glycomimetics. The effects of epitope clustering. and Lewis y antigens was prepared.3. values in the nanomolar range..6 Cancer.However.pendage is attached to a serine or threonine. they react differently . x Certain tetravalent sLe -glycans are exa carbohydrate construct was synthesized tremely potent inhibitors of L-selectin-depenthat presents Lewis y on three consecutive dent lymphocyte traffic to sites of inflammaamino acids. now possible. tion with IC. rier was not needed. In the pursuit of optimal carbohydrate-based anticancer vaccines. however. The gene Mgat5 encodes N-acetylglution of cells is accompanied by an increase in cosaminyltransferase V (GlcNAc-TV). and adjuvant on antibody responses x ride that is five times as active as sLe in to Lewis y conjugates were examined in mice binding to E-selectin and is also effective (88).6-GlcNActure of antigens has stimulated new concepts branched N-glycans is elevated in human mathat are being converted into experimental lignancies of breast. there is obThe observation of unusual carbohydrate served a suppression of cancerous tumor antigens led to the synthesis of glycoconjugrowth and the spread of tumor cells to the gates that mimic unusual carbohydrate struclung (90). a tetrasaccharide mimic is greater been shown to trigger humoral responses in murine and human immune systems. unimolecular glycopeptide containing the Tn.

6-GlcNAcbranched N-glycans ManLev (22) Indication Diabetes Anti-bacterial Anti-bacterial. sialic acids are abundant terminal components of oligosaccharides on mammalian cell-surface glycoproteins and are synthesized from the six-carbon precursor N-acetylmannosamine. poly-N-succiny1-p1-6 glucosamine) Polysialic acid. On conjugation with the ricin toxin. anti-cancer Anti-bacterial Mechanism Amadorasel fructoseamine-3kinase inhibition Cell wall biosynthesis Inhibition DNA minor groove binding Vaccine Company/Institution Dynamis Therapeutics Princeton University University of Michigan Harvard Medical School NIAID Memorial SloanKettering Cancer Center IgX Oxford Hepatitis Memorial SloanKettering Cancer Center UC San Diego Memorial SloanKettering Cancer Center University of Toronto. KLH QS-21 N-nonyl deoxynojirimycin (20) (NN-DNJ) Tn MBRl Lewis y KLH Galactopyranosyl-a D. The killing of ManLev (23)-labeled HeLA cells with ricin is depen- . Mgat5 gene Toxins or immune recognition Anti-inflammatory from other mice when exposed to a powerful gene that causes cancer. cancer Vaccine Inhibition of glycoprotein folding Multi-antigenic. unimolecular vaccine Leukocyte traffic Lewis y clustering vaccine GlcNAc-TV modulation. For example. keto-carbohydrate-based technology is ad- vancing from in vitro to in vivo experiments (91). indicating that this carbohydrate structure plays a role in promoting the growth and spread of cancer. Because cell surface ketones are rare in cells. Briefly.6GlcNAc-branched N-glycans or modulate GlcNAc-TVcould be used to either to dampen T-cells in autoimmune disease or to sensitize the immune system in the treatment of cancer and infections. New drugs designed to mimic P-1. When HeLa cells in culture are incubated with N-levulinoyl-nmannosamine (ManLev 231. GlycoDesign UC Berkeley Meningitis. nppolySA. The Mgat5-deficient mice are resistant to the induction in breast cancer growth and metastasis to the lungs compared with other mice. the ketone ends up on a cell surface. the cell dies. aureus vaccine. 3 Agents in Research Name Generic (Code) DYN 12 Vancomycin (18) analogs Olivomycin A (19) S. Experimental work on a tumor-targeted.mannopyranoside (21) Lewis y epitope p-1. Ketone-containing monosaccharides can serve as substrates in the oligosaccharide biosynthesis (92).226 Carbohydrate-Based Therapeutics Table 7 . (NSG. The Mgat5 gene in cancers promotes cell movement and directs involvement of carbohydrate chains in cancer growth. cells incubated with a ketone-containing monosaccharide produce ketone-tagged cell-surface oligosaccharides. conjugation with keto-selective nucleophiles is possible.

Glycosides that have a trans relationship of the C-1 and C-2 substituents are synthesized by taking advantage of neigh- . Saccharides of glycoconjugates found on cell surfaces can modulate cellular interactions and are under investigation as research agents. protein sequences are also linear and can be easily determined. and manipulated through a combination of recombinant DNA technology and solid phase synthesis. In a similar fashion.1 Chemistry To fully realize the potential of carbohydratebased therapeutics. The creation of the glycosidic linkage between carbohydrate monomers is a strategically important step in oligosaccharide synthesis and requires stereocontrol of the newly formed linkage. saccha- Carbohydrate-based therapeutics is now established. Nature has evolved very sophisticated mechanisms for making exact polypeptides and polynucleotides. considering a relatively small tetrasaccharide and using only nine common monosaccharides found in humans. the shear number of possible isomers becomes difficult to calculate. and supply carbohydrates is needed.1. and the research and development pipeline for these agents is well outlined.2 Solution Phase.3 summarizes carbohydrate based agents in research. rides are a challenge to synthesize because monosaccharides have both regiochemistry and stereochemistry in both open and closed ring forms.1 Challenges and Opportunities. Also. either a-or p-covalent bonds can link the resulting molecules. Using this method. all of which promises a robust future for carbohydrate drugs. 3. Likewise. Carbohydrate biosynthetic pathways are often amenable to interception with synthetic saccharides. In general. If monosaccharides are combined. Traditionally. saccharide agents or their glycomimetics can modulate carbohydrate biosynthetic and processing enzymes that can block the assembly of specific saccharide structures. The one characteristic of carbohydrates that most vividly demonstrates both the opportunities and challenges of carbohydrate chemistry is the potential of this class of biomolecules to form diverse structures. carbohydrates do possess precise structures that are on the surface of complex large molecules and that affect function. carbohydrates have been prepared by synthesis or natural product isolation. oligosaccharide synthesis requires multiple selective protection and deprotection steps. The inevitable comparison of polysacchrides to other biopolymer types helps put the challenge into perspective. new chemical methods to discover. These precise carbohydrate structures are targets of isolation and synthesis. saccharides are made in an apparent random fashion with a diverse set of enzymes competing to produce very diverse products. corresponding glycomimetics or xenobiotic natural products. the linkages may be linear or branched. it may be possible to make cells susceptible to toxins or immune recognition. there is greater than 15 million possible isomers that can be assembled (93). produced. Modulators of carbohydrate recognition and biosynthesis can reveal the biological functions of the carbohydrate epitope and its target receptors. 3. carbohydrates of low-to-intermediate complexity. 3.1. For example. including multiple branch points with a given monosaccharide. 3 CARBOHYDRATE CHEMISTRY With oligosaccharides. Nucleic acids are made in a linear manner through chemical and biological synthetic techniques. characterize. And if many monosaccharides are covalently bonded. Table 7.3 Carbohydrate Chemistry B 2 I dent on the number of ketones expressed on the cell surface. Although random in appearance. In contrast.

however. Other coupling functional groups include phosphites. the solvent.Carbohydrate-Based Therapeutics Activation 0 Lewis Acid 1 Scheme 7 . and the reaction temperature. Two major advances used to streamline the chemical synthesis of oligosaccharides are one-pot reactions (101) and polyrner-supported synthesis (see Section 3. Benzyl ethers are most commonly used as blocking groups in combination with temporary acyl protecting groups. thioglycosides. Trichloroacetimidates and glycosyl sulfoxides are now commonly used for coupling (95. and pentenyl glycosides (97-99). With acyl type protecting groups at C-2.1).2trans products (Scheme 7.2-cisglycosides. boring group participation from the 2 position. Note that the protecting groups used influence the chemical reactivity of the glycosyl donors and acceptors. 1 .2-cis and 1.4) (102). The method uses a new sulfonium reagent that activates glycals for direct addition of the acetamido group and coupling with a variety of glycosyl acceptors.2-cisglycosides.1) is one of the most widely used techniques (94). the reaction proceeds through an acyloxonium cation intermediate that reacts with a nucleophile to give 1. The multi-functional nature of the carbohydrate molecules requires the use of protecting groups on hydroxyls or other functional groups not involved in the reaction to provide an unambiguous synthesis. the stereoselectivity depends on the chemical reactivity of the glycosyl donor and acceptor. A single vessel method that allows easier access to this key class of oligosaccharides is now available (100). changing a single protecting group at a remote position may result in changes in stereoselectivity and yields. a prerequisite is the use of non-participating blocking groups at the 2 position. however. and in some cases. The Koenigs-Knorr method of coupling glycosy1 halides (X = halogen in Scheme 7.1. Carbohydrates with an N-acetyl group at the 2 position are found in many important biological structures. Note that N-silylated amide must be used to carry out the nitrogen-transfer process and use of unprotected primary amides as nitrogen nucleophiles does not seem to be effective. whereas temporary protecting groups are required on functional groups that must be manipulated at a later stage of the synthesis. One promising approach at developing automated synthesis uses single-vessel reaction schemes that take advantage of the reactivity . Alternative leaving groups have been developed that are more stable. the strength of the promoter used for the reaction. For the synthesis of 1. methods for synthesizing such sugars require many steps and often yield complex mixtures of stereoisomers.2-trans stereochemistry in the product. The relative instability of the sugar halide necessitates the construction of the saccharide from the reducing end. whereas the oxocarbonium ion is expected to give mixtures of the 1. Blocking groups are required for functional groups that are not involved in the whole synthetic scheme. Using a non-participating substituent. phosphates. does not necessarily result in the formation of 1. 96).

3 Enzymatic.3 Carbohydrate Chemistry 8 Scheme profile of different protected monosaccharides to determine the outcome (103). glycoconjugate. If the necessary enzyme is available. By using disaccharide thioglycosides as reactants in the linear scheme. the predominance of a desired target compound is assured.1. including the cancer antigen globo H hexasaccharide (see Section 2. a reactivity database on protectedp-methylphenyl thioglycosides was generated and used as the basis of a computer program that selects the best reactants for one-vessel synthesis of oligosaccharides (104).6) (105). Chemical synthesis and enzyme-based routes are complementary.2. 3. such as glycosyl fluorides of the opposite anomeric configuration. A novel exception to generalities above is demonstrated in the stereospecific synthesis of sucrose (Scheme 7. and can form stereochemical and regiochemical reaction products with rate acceleration. mutant glycosidases can be employed in connection with novel donor substrates. operate at room temperature under neutral aqueous conditions.and stereo-control. Although some enzymes will act on alternative substrates. However. the desired bond can be formed. The synthesis was accomplished by remote tethering the two ends of the donor part so that only one possible coupling stereochemistry is pos- sible. the generality of remote tethering for ordinary oligosaccharide syntheses is not clear. In comparison. This approach has been used in the synthesis of a large number of oligosaccharides. and one anomeric center has the capability of isomerization. Natural and non-natural saccharide building blocks can be assembled with natural or non-natural linkages. The reactivity of a monosaccharide is highly dependent on the protecting groups and the anomeric-activating group used. Work is needed to design a complete set of building blocks for use in the synthesis of most bioactive saccharides. . As an exception.108). branch points are incorporated. These reactions are typically performed in solution. Chemical synthesis offers exceptional flexibility. enzymes can be used to effect glycosylation with absolute regio. Many reagent enzymes accept unnatural substrates. Thus. chemical synthesis provides the means to generate any oligosaccharide. and are used to produce both oligosaccharides (109) and polysaccharides (1 10). The problems faced by the sucrose synthesis are that the two monosaccharides are connected through tertiary anomeric centers.2) (106). enzymatic transformations are performed on unprotected carbohydrates. or glycomimetic. The key to this approach is to have extensive quantitative data regarding the relative reactivities of different protected monosaccharides. Enzymes are useful re- agents in oligosaccharide synthesis (107. often with high efficiency. Both glycosyltransferases and glycosidases are employed in the enzymatic synthesis of oligosaccharides. By adding substrates in sequence from most reactive to least reactive. Of particular value.

and mucin glycopeptides containing common core structures have been synthesized (128). Incorporation of additional monosaccharides using glycosyltransferases is a novel method of glycoprotein synthesis (129). For example. Many oligosaccharides are linked covalently to peptides. one approach for solid phase carbohydrate synthesis has actually incorporated coupling chemistry into an automated solid phase oligosaccharidesynthesizer (123). is used to transfer an oligosaccharide group from a water-soluble polymer to ceramide. such as monosaccharide halides and p-nitrophenyl glycosides. 3. A problem with attempting to couple large oligosaccharide side-chains to a polypeptide is low yield. In addition. Several laboratories have used protected glycosylated amino acids as building blocks for automated solid phase peptide synthesis to generate glycosylated peptide fragments reminiscent of natural glycoproteins (125). The acceptor amino acid can be part of the evolving chain in solid phase peptide synthesis and reacted with a saccharide to produce the target glycopeptide. The endoglycosidase. glycosylated amino acids containing one to three sugars have been incorporated in solid phase synthesis of glycopeptides (127). 117). Many resins can be used. For example. and the substrates for these enzymes are the nucleotide-activated monosaccharides that are also not readily available and must be prepared from monosaccharides through enzymatic or biological methods (111).Alternatively. amino . Although glycosidases use available substrates. Solution phase supports have the problem of product recovery.This enzyme has also been used for the synthesis of bicyclic carbohydrates (115).4 Solid Phase. Water-soluble supports have been used in the enzymatic synthesis of pseudo-GM3 (118). yields are lower.Carbohydrate-Based Therapeutics The major problem with enzymatic synthesis is generating the required catalysts and substrates. Other novel donor substrates used in enzymatic reactions include oxazolines for the synthesis of artificial chitin (112) and 6-0x0-glycosides for the synthesis of N-acetyl-D-lactosamine derivatives (113). Modern 9-fluorenyl methoxy carbonyl (FM0C)based peptide synthesis methods are sufficiently mild that the oligosaccharide remains intact throughout the synthesis.1. Reaction on solid phase requires soluble enzymes that are either recovered for recycling or discarded. Chemical linkage of saccharides to peptides is affected in several ways. glycopeptide segments prepared on solid phase. The productivity from these efforts is evident from the diversity and complexity of oligosaccharides that have been synthesized. An alternative approach is to couple the substrate to a watersoluble polymer that can be removed from solution by precipitation of the polymer or size exclusion chromatography. and the integration of oligosaccharides into glycopeptides has started. including sequences containing up to 12 residues (122). Few glycosyltransferasesare readily available. This enzymatic tool provides an efficient new method for polymer-supported synthesis of glycosphingolipids (1 19). Other novel enzymatic examples include the synthesis of cyclic imine saccharides through the use of the fructose diphosphate aldolase (114). ceramide glycanase. and polystyrene (116. Another example. which may be attributed to steric factors. Newly developed glycosylation methods and advances in solid phase organic chemistry increased interest in solid phase oligosaccharide synthesis in the 1990s (121). there is the use of glycosyltransferase-regenerating beads that are potentially applicable to automated oligosaccharide synthesis (124). can be condensed using serine proteases. A method that overcomes the steric problem was developed in the case of coupling a high mannose-type pentasaccharide to a variety of tripeptides (130). a glycopeptide' derived from the mucin-like leukocyte antigen leukosialin was made using trisaccharide-amino acid as a building block (126). silica. including Sepharose. Solid phase synthesis of oligosaccharides and glyconjugates similarly offers powerful advantages for oligosaccharide synthesis in comparison with conventional methods because it circumvents multiple purification steps needed in traditional solution syntheses (120). polyethylene-based resins. polyacrylamide. Although of limited scope.

and tunicamycin (144).5 Clycoprotein. In preparing a large number of resin-bound carbohydrates in a combinatorial fashion. Solution phase split-and-pool methods are capable of generating diverse libraries containing large numbers of compounds. Progress on this front has resulted in the generation of oligosaccharide libraries (134. two carbohydrates that exhibited a higher affinity than the known natural ligand were identified (140). and exhibit a high degree of functionalization. fermentation produces many different protein glycoforms. and then elaborated enzymatically to increase size and complexity of the glycan by using enzyme-catalyzed transglycosylation (132).1. Early on.and disaccharides second. Combinatorial methods for the identification of biologically active carbohydrates are evolving. . and then mono. The resulting mixture. When screening a resin-bound library against a carbohydrate binding protein. Monosaccharides used as scaffolds include D-glucose (141). can be used in subsequent digestion reactions where the sugar chains are cleaved to produce a near homogeneous glycoprotein that.[ T 4 3 Carbohydrate Chemistry I' acids containing high mannose-type oligosac:i charides with as many as 11 primary sugars in a triantennary structure were prepared (131). The degree and heter- ogeneity of glycosyl groups on glycoproteins produced in nature is not only a challenge in synthetic methodology but also is a problem in characterization and in target product specification. however. it was found that the protein bound with high selectivity to only one carbohydrate among more than 1300 closely related compounds in the library (145).and trisaccharides were produced by a split-andpool protocol employing glycosyl sulfoxides as donors for solid phase glycosylations and reacting with C2 azidosugars first. random glycosylation of unprotected disaccharides. A split-and-pool approach was used to generate a small library of chondroitin sulfate disaccharides containing all eight possible sulfated forms (139). Progress in solid phase saccharide chemistry (120). : 3. Although sup- ply continues to be a major challenge to carbohydrate chemists. can be further reacted enzymatically. especially with oligosaccharide libraries that contain identical molecular weight compounds that only differ in the stereochemistry of the anomeric center. cell fermentation is used to glycosylate proteins. complex oligosaccharides and oligosaccharide libraries are becoming accessible to non-chemists. it was observed that the bead-bound carbohydrates exhibited multivalent behavior.1. Prob- lems include hit identification and isolation. Nearly statistical mixtures of all possible trisaccharides were obtained after chromatography. Thus. Screening against a bacterial lectin. a new tool for drug discovery. Monosaccharides are ideal scaffolds for the preparation of libraries because they are enantiomerically pure. Generation of random glycosylationbased libraries as heterogeneous mixtures have given way to focused libraries that employ discrete saccharide cores. D. 138). library synthesis. 2-deoxy-2-aminoglucuronic acid and 3-deoxy-3-aminoglucuronic acid (1431. Unfortunately.135). is now becoming available. in turn. Linear and branched trisaccharide libraries were prepared by the latent-active glycosylation (137. Currently. This concept made use of the convenient access to both glycosyl donors and acceptors from common ally1 glycoside precursors. A 1000compound library of acylated amino di. involving the nonselective coupling of a fucosyl donor.and L-glucose and L-mannose (142). was used to produce three sub-libraries of a-fucosylated disaccharides in one step (136). Efficient methods for the combinatorial synthesis of carbohydrates on the solid support and in solution are known (133).in combination with techniques employed in glycomimetic chemistry. For example. 3.6 Carbohydrate Library. often possess a rigid conformation. will advance carbohydrate library methodology and be a source for promising new therapeutic agents. Reductive conversion of the azides into m i n e s allowed for further differentiation using acyl groups. glycans on N-glycosylated proteins are digested to the core N-acetylglucosamine by endoglycosidases.

Presumably. MALDI-Q-TOF is arguably the most sensitive and useful type instrument for complex carbohydrate analysis (148). electrospray ionization (ES). especially in development. but where more structural detail is needed. monitoring natural ionization-induced fragmentation. For example. NMR still remains a powerful tool. The key to the success bf discovering naturally occurring therapeutic agents rests on bioassay-guided fractionation and purification procedures. With the potential structural diversity of carbohydrates. a new type of disaccharide linkage. it is not clear whether this new linkage has important biological meaning. NMR techniques used in combination with databases is helpful for carbohydrate analysis.2 Analysis To gain an insight into the biological and physicochemical functions of complex carbohydrates at the molecular level. The natural allocation of hydrogen and carbon at- Natural products isolated from higher plants and microorganisms have provided novel. or matrix-assisted laser desorption ionization (MALDI) have advanced carbohydrate analysis (146. Linkage and monosaccharide analysis by a combination of derivitization and LC/MS continues to provide useful carbohydrate characterization.1. clinically active drugs. it is curious that novel linkages. In the study of the structure of the cell-wall polysaccharides of a poorly characterized Proteus microorganism. Soft ionization techniques of fast atom bombardment (FAB).4). and primary sequence of the constituent saccharide residues. in bacterial polysaccharides was discovered (152). namely. Nuclear Overhauser effect spectroscopy (NOESY).3 Natural Products Isolation (NMR) spectroscopy is critical in determining the complete structure of a carbohydrate. knowledge of their primary and secondary structures is a required.2.Carbohydrate-Based Therapeutics 3. and (pseudo) 3D and 4D extensions. Nuclear Magnetic Resonance oms in carbohydrate molecules lends itself to the revelation of their full primary structure through H-1 and C-13 NMR analysis. SUGABASE is a carbohydrateNMR database that combines CarbBank complex carbohydrate structure data (CCSD)with proton and carbon chemical shift values (151).heteronuclear single quantum coherence (HSQC). 3. Although not a major focus for research. Techniques include 2D Total Correlation Spectroscopy (TOCSY). 3. The main emphasis of current carbohydrate structural analysis is the applicability of modern multi-dimensional NMR for solving the two crucial problems in complex carbohydrate structural analysis.1 Mass Spectrometry. Of the three ionization technologies. rotational nuclear Overhauser effect spectroscopy (ROESY). 3. Derivatization can be divided into tagging of reducing ends and protection of most or all of the functional groups.2. and for that matter novel monosacchar- . Highly sensitive mass spectrometric methods for defining the primary structures of glycoprotein glycoforms is advancing structural carbohydrate chemistry (149). 147). number. The complete structural characterization of complex carbohydrates involves determining the type. Devising strategies to optimize fragment ion information include the following: inducing fragmentation by collisional activation. unique discoveries are made in carbohydrate natural products isolation. heteronuclear multiple bond correlation (HMBC). however. This includes the occurrence of branch points and the location of appended non-carbohydrate groups and the three-dimensional conformation(~) and dynamics in solution. the elucidation of the sequence of glycosyl residues and the solution conformation and dynamics of a carbohydrate (150). this is the first example of an open-chain glycosidic linkage. and selecting derivatives that enhance and direct fragmentation. heteronuclear multiple quantum correlation (HMQC).2 NMR. structure (24). A traditional area of carbohydrate isolation is bacterial capsular polysaccharide chemistry for vaccine production (see Section 2. Permethylation is still an important type of full derivatization. Chemical derivatization methods as an aid to MS analysis continue to be important.

Several sialic acid-based polymers have been synthesized that inhibit flu virus receptor-binding activity. when presented on a polymer. Naturally occurring carbohydrate multivalent displays are widespread in nature and include mucins.4 Multivalency Carbohydrate function is often contingent on multiple repeat saccharide presentation and subsequent multidentate binding. A multivalency ap- ides. however. For example. in hemagglutination or the cross-linking of erythrocytes by a virus. The multivalent liposome inhibits not only hemagglutinination but also the viral surface protein neuraminidase activity. and the potency of the lysoganglioside is a million-fold higher than that of the monomeric ligand. The linked dimer also inhibits the bacterial enzyme-catalyzed modification of aminoglycoside drugs that is an antibiotic-resistance mechanism (160). 3. Carbohydrate multivalency is the simultaneous attachment of two or more binding sites on one oligosaccharide. often present multiple repeat carbohydrate binding domains. and mammalian cell surfaces. which in turn impedes flu viruses from sticking to cell surfaces and subsequent viral infection of cells (see also Section 2. Flu viruses attach to cell surfaces through multivalent interactions. employing multiva- proach was used to inhibit influenza virus infectivity. The interaction of multivalent presentations can result in the formation of numerous simultaneous non-covalent bonds that proceed to afford the observed high avidity of carbohydrates (153). polysaccharide. effective inhibition is observed in the picomolar range (157). Indeed. This new antibiotic is active against several bacteria. which is usually a protein.4.3 Carbohydrate Chemistry lent presentations found in nature is a seductive strategy for drug discovery.4) (155. 156). Significant affinity enhancements are observed in these studies. a combination of a sialic acid liposome and a lysoganglioside/poly-L-glutamic acid conjugate act as inhibitors of flu virus hemagglutinin at picomolar concentrations (159).1 Viral Inhibitors. The binding of monosaccharides to target sites on proteins is generally weak. . A bifunctional dimeric aminoglycoside antibiotic that binds to bacterial RNA can be made to bind three orders of magnitude more tightly than the monomeric antibiotic. Multivalent sialic acid-containing liposomes have been shown to inhibit flu virus (158). or glycoconjugate to multiple receptor sites on another molecule.1. The liposome's potency is about 1000-fold higher than that of the corresponding monomeric sialic acid ligand. Indeed. Sialic acid-based polymeric agents can inhibit the interaction of trimers of hemagglutinin and sialic acid.2 RNA Target. Lectins. the monomer shows inhibition in the millimolar range. 3. the use of multivalency is also a new approach to drug design (154). Multivalent inhibitors can be used to target not only proteins but also RNA. are not found more often in nature. pathogen. Because successful drugs usually bind tightly to target receptors with binding affinities often in the nanomolar range and monomeric carbohydrates generally have low receptor binding with binding affinities in the millimolar range at best.4. carbohydrate-binding proteins. including tobramycin-resistant strains. Perhaps a study of the carbohydrate chemistry of microbes that exist in extreme conditions may yield additional novel carbohydrate-based structures and perhaps new therapeutics. yet the affinity and specificity required for effecting a physiological response is high. 3.

In neighboring site participation.The structure of starfish (25) consists of two trisaccharide units at the t i s five tethers that radi. The resulting inhibitor binds to multiple sites on a target bacterial toxin that is closely " related to cholera toxin. starfish (25)is sandwiched between two subunits. The avidity of many multivalent carbo- .~ of ate from a central glucose core yielding a decadent presentation. only five of starfish's trisaccharide arms are engaged.4. 3. a chelate effect. An X-ray crystallographic study of the complex formed between the B-subunit pentamer of SLT-I and starfish(25)revealed the mode of binding. This multivalent carbohydrate ligand exhibits more than six orders of magnitude increase in inhibition over trisaccharide unit. it binds only to one. Mecha- - nisms that contribute to the high avidity observed for multivalent ligands include neighboring site participation. the enterotoxin's five identical binding sites recognize and bind to ganglioside carbohydrate groups in the gut lining(163). and a precipitation process. which is the biological mode of action. leaving the other five free to bind another B-subunit.4. A bacterial enterotoxin related to cholera toxin is also a target for a multivalent inhibitor.4 Mechanism of Multivalency. target protein clustering. is reported to neutralize Shiga-like toxins (162). starfish. The binding prevents the toxin from attaching itself to carbohydrates on cell surfaces. Shiga-like toxins from E. Thus. That is.3 Toxin Inhibitors. Photoaddition of thiols to ally1 ethers in an anti-Markovnikov fashion is used as a key step in the multiple attachment of carbohydrates to cyclodextrin cores(164). One tailored multivalent carbohydrate ligand. many target proteins possess multiple or secondary sites adjacent to the initial binding site (165).Carbohydrate-Based Therapeutics 3. When ingested. Instead of binding to the theoretical two sites per monomer. Multivalent inhibi- tors are being designed to inhibit disease-causing bacterial agents including Shiga-like toxin and enterotoxin. coli can be classified into two subgroups. Another approach is to develop synthetic strategies for attaching carbohydrate residues to different parts of cyclodextrin molecules. This subnanomolar activity is within the range desired for an anti-toxin thera~eutic. SLT-I and SLT-I1 (161).

The total synthesis of a series single glycomimetics of heparin. Ideally. multiple interactions occur after formation of the first contact. Thus. 3. chemical synthesis was employed to optimize the length and the charge of the synthetic oligosaccharides. An additional problem with multivalent presenting molecules is that by their very nature they are high molecular weight molecules. Polymeric multivalent structures will have high molecular weights and as such will not likely to be orally active and would require other routes of administration such as parented delivery. However. it was clear that reducing the size to the minimum that still allowed thrombin inhibition was not sufficient to abolish the undesired interaction with PF4. Heparin is a complex mixture of heterogeneously sulphated polymers consisting of alternating L-iduronic and n-glucosamino sugars and is produced and stored in mast cells present throughout animal tissues.One structurally optimized oligosaccharide heparin mimetic is 10 times more potent in vivo than both standard heparin and low molecular weight heparins and is also devoid of the undesired side effect. the use of multivalent compounds as systemic drugs will be limited. there will be no avidity at biologically relevant concentrations. and if multivalency effects have aggregation events as their origin. Also.was recently announced (169).1. When a chelate effect operates. The coagulation of blood is the result of a biochemical cascade catalyzed by enzymes called activated blood-coagulation factors. In a synthetic series that explored chain length. This accounts for the anticoagulant and antithrombotic activities of heparin. and its action is amplified after binding to heparin.5 Heparin 3. multifunctional proteins that mediate cellular functions by binding to transmembrane FGF receptors (171. The protein antithrombin (AT) is the main physiological inhibitor of these factors. thrombocytopenia.1 Synthesis.5.5). apparent avidities will be strongly dependent on concentration. Divalent and trivalent compounds could have low enough molecular weights (less than a 1000 Da) to be orally active. the optimized oligosaccharide heparin mimetic has 10 times the antithrombotic potency relative to that obtained for heparin but is unable to activate platelets in the presence of plasma from patients with heparininduced thrombocytopenia and is devoid of other side effects attributed to the anionic charges of heparin.2 FGF. which include thrombin. Heparin has numerous biological activities that can be attributed to binding interactions between these anionic groups and the positively charged groups on proteins.hydrate molecules is caused by their ability to cluster target receptors (166). Complexes of the decasaccharides with both native and selenomethionyl aFGF were purified and crystallized. Thus. such as hemorrhage. The next step was to reduce the charge of the molecule for antithrombin. 3. An X-ray structure of a hepa- rin-linked biologically active dimer of fibroblast growth factor was described (170). Another model is that most of the enhancement in avidity observed in multivalent binding arises from an aggregation and precipitation process that follows the initial intermolecular binding step (168) If the avidity increases because of aggregation. heparin mimetics should discriminate between thrombin and other proteins. Synthetic heparin mimetics consisting of 20 monosaccharides (20-mers) exhibited antithrombotic activity similar to that obtained for heparin. that is associated with heparin treatment. providing two crystal forms each containing a total of four inde- . Indeed.5. The fibroblast growth factors (FGFs) form a large family of structurally related. which is related to heparin-induced thrombocytopenia (see also Section 2. and that provided the structurally optimized oligosaccharide heparin mimetic. the cross-reactivity of these compounds with PF4 prompted the exploration of related molecules structurally closer to heparin. which are facilitated because of the high effective concentration of the intramolecular binding groups (167). high molecular weight materials tend to be mixtures that are difficult to characterize and to control in manufacturing. 172). particularly platelet factor-4 (PF4).

heparin-decasaccharide-linked dimer of aFGF exhibits a unique mode of ligand-induced protein dimerization. using property encoded nomenclature/MALDI method. a heparin sequencing technology called property-encoded nomenclature/MALDI sequencing was developed (173). Mass signatures of the fragments are then obtained by MALDI-MS. Carbohydrate chemistry is contributing both to the understanding of biological functions of saccharides and glycoconjugates and to the discovery and development of carbohydrate-based therapeutics. The process involves the use enzymatic and chemical techniques to cut a second. 4. the protomers are positioned to bind sulphate groups of five to six monosaccharide units on opposite sides of the heparin helix axis. Glycobiology is now an established discipline that is expanding rapidly. The maturation of carbohydrate chemistry systems that are easily accessible to both biologists and chemists will have an important impact on our understanding of glycobiology and on the discovery of carbohydrate-based therapeutics. chemical tools have proven indispensable for studies in glycobiology (174). each piece provides structural information at reducing and non-reducing ends.1 Human Genome. the decasaccharide chains are ordered and bind through sulphate groups to residues in the heparin-binding loop of aFGF residues. Advances in carbohydrate chemistry have solved some problems associated with carbohydrate research and " have provided new strategies for solving challenges in glycobiology. Recently. which is in turn determined by MS. The crystal structure of the biologically active. The next step is a compositional analysis in which the sample is fragmented into its individual building blocks and counted as to how many of 32 GAG building blocks it contains. 3.3 Heparin Analytical Tool. With the list of building blocks and the corresponding codes for a given sample in hand. One force is the sequencing of the human genome. and the signatures are used as constraints to search a database of possible sequences for the correct match. a computer analysis generates a list of all possible sequences that satisfy the overall composition. In the structures of all four dimers. a unique code for each of the 32 GAG building blocks was generated.1. What follows is a discussion of a few of the interesting developments and trends that have occurred in this field over the past 10 years. In addition. Technical problems that hinder carbohydrate research still exist. 4. Each code is read by computers and also indicates structural information. The database was constructed as follows. which gives the number and kind of building blocks in each. and an understanding of the major advances in this discipline is essential for the discovery of new carbohydrate-based therapeutics. A crucial milestone in the maturation of carbohydrate chemistry will be the development of robust automated systems for the synthesis of oligosaccharides and glycoconjugates. Instead. the mass of each of these smaller pieces is determined. and immunogen and lectin research.5. each building block is identified by its mass. glycoconjugate structure-function studies. post-translation modifications have become obvious explanation for the source of .1 Driving Forces Several forces drive the surge in interest in the field of glycobiology. First. Then. intact sample into smaller and smaller pieces. This aFGF dimer is bridged by heparin and does not have a protein-protein interface. Because these techniques cut heparin fragments at specific sites. As it becomes clear that the number of genes encoded by the genome cannot alone explain the complexity of life. Repeating this process for smaller and smaller pieces eventually gives the correct sequence. The correct sequence is among the ones listed and is defined by a process of elimination. Collectively. the polysaccharide is first fragmented by enzymatic and/or chemical means.Carbohydrate-Based Therapeutics pendent heparin-linked aFGF dimers. To sequence a heparin-related compound. They include the sequencing of the human genome.

and Arngen regularly rejects batches of EPO because of incorrectly attached sugars (176). For example. erythropoietin (EPO). the glycosphingolipid GM1 was in clinical trials to determine whether it could play a role in repairing damaged nervous tissue caused by stroke or spinal injury (see Section 2. Glycosylation of monoclonal antibodies is important for their recognition by Fc y receptors (1771. Modification of glycosylation sites has also led to second generation drugs. For example. The association of the non-polar function with the cell membranes effectively anchors these molecules to the extracellular surface. and gangliosides. (5) polyisoprenyl-type glycolipids in which the sugar is normally linked through a phosphate bridge to the polyisoprenols. correct glycosylation of the top-selling drug. 4. Third. there have been successes within the last 10 years in making effective drugs that are based on carbohydrate structure. focusing on how the basic research may effect the discovery on new carbohydrate-based therapeutics.2.2). up of oligosaccharides linked covalently to a fatty acid portion by means of inositol or sphingosine moieties. resulting in an enzyme with a slightly higher half-life and lower fibrin-binding affinity than recombinant TPA (179). Furthermore. because it had been es- . gradients of carbohydrate sidechains could play key roles in processes such I as development.2. The complexity of saccharide side-chains could more than sufficiently provide micro-specificity between proteins. ( 4 ) steryl glycosides or glycosylated derivatives of sterol. Glycolipids are subdivided into (1) glycosphingolipids. and removal of glycosylation sites in therapeutic monoclonal antibodies could result in reduced immunogenicity and "first-dose" cytokine release syndrome (178). i 4. and the very complex proteoglycans found in the extracellular matrix.1 Glycolipids. including cell-cell and cell-matrix interactions. The role of glycolipids extends well beyond the essential structural role they play in cell membrane lipid bilayers. For instance.3 Structure-Function. so does the need for understanding the role that sugars play in their function. Second. a form of glycosphingolipids. glycoproteins.2 Therapeutic Glycoproteins. animal studies have suggested this use. the glycans are readily modified after synthesis of the core structure (175). (3) ester-type glycolipids or esters of carbohydrates with fatty acids. including glycophosphatidyl inositol anchors. a tissue plasminogen activator (TPA) molecule in which the glycosylation sites have been removed. which act as sites of attachment for proteins to the cell membrane. The use of neuriminidase inhibitors to prevent the escape of the influenza virus from its host cell and the conversion of heparin that is a complex mixture with poly-pharmacy into a single chemical entity with a well-defined pharmacology are real drug discovery success stories. (2) glycerol-type glycolipids or glycosylated glycerols having fatty acid substituents on the glycerol. It was also expected that glycosphingolipids would have use in the diagnosis and treatment of cancer.2 Glycoconjugates r i $ Glycoconjugates are expressed primarily as three different forms: glycolipids.this complexity. the glycobiology of glycolipids and glycoproteins and their ligands and proteoglycans are reviewed.1. Additionally.1. and ( 6 )glycosylphosphatidyl-inositols. which are thought to be crucial in the development of nervous tissue. is critical to maintaining a reasonable circulating half-life. In the last 10 years. it has been calculated that a hexasaccharide has 1012 isomeric permutations. In this section. Glycolipids are made 4. Whereas its mechanism of action was unclear at that time. there has been a great deal of research and high expectations that glycolipid-based therapeutics would be successful. Future success stories will depend in part on an understanding of glycobiology including structure-function of the principal biomolecular conjugate structures. including glycomimetics and an understanding of glycobiology. 4. as the use of glycoproteins as therapeutic agents increases. such as reteplase. The examples are numerous.

far the most complex glycoconjugates. The knockout mice thus indicate a physiologic function for heparin unrelated to its ability to prevent blood clots. less common family referred to as "hybrid. many of the processing steps are similar. including cartilage. Trimming by a-mannosidases. The components affected are several enzymes called mouse mast-cell proteases. no consensus sequence is known.Carbohydrate-Based Therapeutics tablished that almost all human cancers exhibit aberrant glycosylation patterns. These proteases are used by the immune system to destroy foreign proteins.1. It has been estimated that 90% of all of these sequences are glycosylated (183). In the case of "complextype" carbohydrate side-chains. With GAGs for example. In fact. . The gene encodes glucosaminyl N-deacetylasel N-sulphotransferase-2. In mice. resulting in carbohydrate moieties of differing size and terminal sugars. however. an enzyme that adds sulfur trioxide to heparin. the process begins with the attachment. in general. The biosynthetic processes of elongation and cgpping of oligosaccharides varies.2 Proteoglycans." Members of this family share structural features of the other two families (184). 0-linked side-chains can vary in size from a single GlcNAc residue to structures as large as complex-type N-linked oligosaccharides.2. "trimming" is followed by the stepwise addition of new sugar residues by glycosyltransferases. Attachment is initiated in the Golgi and the initiating sugar is usually " a GlcNAc. the carbohydrate precursor is attached to asparagine residues within the recognition sequence Asn-X-Ser/Thr. Attachment and maturation of the carbohydrate side-chains occurs within the endoplasmic reticulum (ER) and Golgi in a highly ordered manner. GAGs show substantial structural heterogeneity.2. After the transfer from the dolichol lipid. through a dolichol lipid carrier. and unlike N-linked glycans. Loss of this enzyme results in improper packaging of key components of the storage granules in mast cells. Proteoglycans are very 4. occurs through a trisaccharide linker to an O-glycosidic bond at serine residues. Their attachment to the polypeptide backbone. GAGs provide viscosity. and there are two major classes of glycoproteins. there is a third. Unlike N-linked oligosaccharides. The GAGs on aggrecan consist of chondroitin sulfate and keratin sulfate and play an important role in the hydration of the cartilage. which makes proteoglycans ideal molecules for the joint extracellular matrix. Glycoproteins are by large extracellular glyconjugates (up to millions of Daltons) composed of a protein core with polysaccharide chains extending out perpendicularly in brush-like structures. chains are built up of long linear sequences of alternating hexamine (D-glucosamine or D-galactosamine) and hexuronic acid (D-glucuronic acid or L-iduronic acid) units. 0-linked oligosaccharides are attached to the hydroxyl component of serines and threonines. a gene necessary for the production of fully functional heparin was disrupted or knocked out. without subsequent glycosyl addition. 0-linked and N-linked. there are at least seven 0-linked oligosaccharide core structures. Specifically. resultingin similar terminal sugars between the two oligosaccharide types.a progressive disease in which the joint cartilage is gradually eroded. 4.5) (182). and as a result.3 Clycoproteins. the polysaccharides account for 90-95% of the proteoglycan's size by weight. In addition to high mannose and complex-type N-linked oligosaccharides. In the case of N-linked oligosaccharides. results in ManxGlcNAc2 side-chains called "high mannose" structures. The predominant proteoglycan in the cartilaginous matrix is aggrecan (180) and is degraded in osteoarthritis (OA). depending on whether the oligosaccharide chain is linked to the protein through threonine or serine side-chains (0-linked) or aspargine (N-linked). Heparin is now known to play a key role in packaging proteases in mast cells granules (see Section 2. which serve as the first line of defense for the immune system. of Glc3Man9GlcNAc2 structures containing trimannosylchitobiose cores to polypeptides as they enter the ER. and despite the regularity in this alternating sequence. Clinical trials are ongoing to determine whether chondroitin sulfate can be used as a treatment for OA (181). impart low compressibility. the sugar sidechains are "trimmed" by exoglycosidases.

and charge and water-binding capacity. engineering in improved Fc y receptor-binding activity could be beneficial for anticancer 4.suggesting these oligosaccharides play a role in maintaining the proper geometry of that the Fc domain. two antibodies. protease resistance. Understanding the functions of the carbohydrate in these molecules could lead to improved second-generation protein products with improved efficacy and potency and reduced side effects. Over the last several decades.4 Clycobiology The carbohydrates groups on glycoproteins confer important physical properties such as conformational stability. deglycosylation studies demonstrated that carbohydrates are required for this interaction. the interest for understanding the role of carbohydrates in antigen recognition and in mAb function has also increased. Fc yRIIIA. an Asn297-mutated variant of the anti-CD3 mAb. antibody-dependent cellular cytotoxicity (ADCC). were taken into clinical trials (189). there has been sustained interest in developing therapeutic mAbs that bind specific carbohydrate epitopes on diseased cells such as cancer. domain thermal stability and binding to a soluble form Fc yRIIb (1981. domain gradually reduced the CH. and Fc yRIIIB) interactions. A33 and R24. and the solved structure is at the binding interface (196).3. in which the immunogen is the Hib capsular polyribosylribitolphosphate polysaccharide (190). the carbohydrates can act as epitopes for ligand binding and as sinks for carbohydrate-binding growth factors (185). For example. The CH. has been tested in phase I studies with some success. Conversely. Reducing the ability of therapeutic mAbs to bind to Fc y receptor could be an effective way to reduce the side effect of "first dose" cytokine release syndrome often experienced by patients. respectively. A newer strategy has emerged in which people are immunized with carbohydrates immunogens for certain diseases. 4. 0KT3 (muronomonabCD3). In fact. Additionally. There lmmunogens With the ever-increasing number of monoclonal antibody (mAb) therapies entering the clinic. and not protein-protein interactions. which recognize a novel palmitoylated surface glycoprotein (187) and the GD3 ganglioside (1881.Manipulation of this interaction could lead to designer antibodies with varying levels of antibody effector functions.4) (191). phagocytosis of microorganisms.to determine whether vaccinating patients with conjugates of carbohydrate antigens expressed on tumor cells and keyhole limpet hemocyanin (KLH) can be effective therapeutics at . is also much interest in understanding role of glycosylation in monoclonal antibody and Fc y receptor (Fc yRI. none of these efforts has resulted in a therapeutic agent. the binding sites on the Fc portion of IgG for all classes of Fc y receptors has been mapped to high resolution (197). endocytosis of immune complexes. region of the Fc domain is important for both the dimerization of the immunoglobulins and for binding of the IgGs to Fc y receptors. the oligosaccharides attached to Asn297. Fc yRIIA. This may not be surprising given that many of the biologics on the market are immune modulators. Unfortunately to date. Binding of the Fc portion of the IgG to the Fc y receptors leads to release of inflammatory mediators. Crystallographic studies showed that within the CH.1 immunoglobulins. help to drive the dimerization of the IgG monomers. domains. IgG and the Fc yRIIIA have been cocrystallized. One of the areas where the role of carbohydrates and glycoproteins is best understood and is most likely to stimulate new carbohydrate-based therapeutic discovery is in the immune system (186).and regulation of immune cell activation (192-195).3 breaking immune tolerance to cancer cells (see Section 2. Most of the protein therapeutics on the market today are glycoproteins. Additionally. clinical trials are underway . and almost all the key molecules involved in the innate and adaptive immune system are glycoproteins. Furthermore. Fc yRIIB.1. Gradual truncation of the carbohydrates in the CH. an approved therapeutic for the prevention of transplant rejection. While the crystal structure suggests that the carbohydrates are on the periphery of the binding interface between IgGs and the Fc yreceptors. A real success story in this area was the development of the Hib vaccine.

In the presence of these neuraminidase inhibitors. the larg- 4.3. available for the treatment of influenza A and B. precluding fusion and entry into the cells (see Sections 2.2. which were originally defined as Ca2+ and phosholipid binding proteins. carbo- hydrate-based therapeutics are being explored for anti-viral therapies. but it did greatly impair infectivity. There is increasing evidence that N-butyl deoxynojirimycin (15). A design and discovery of carbohydrate-based therapeutics. prevented gp41 exposure.6) (203).4. which are sialic acid-binding proteins that contain I-type CRDs derived from the immunoglobulin fold. Perhaps . Similarly. thereby reducing infectivity (199). For the most part. mimetics of N-acetylneuraminic acid thioglycosides. not incorporated into the viral coat. and hence production of viable viruses. The C-type lectins are further subdivided as those family members that are membrane proteins (selectins.1. and they are classified by their carbohydrate-specificitiesand defined protein domains. which are metal independent in their binding activity and bind galactose through globular galectin-type CRDs. In the case of HIV. is of special interest to the est subfamily of lectins is the Ca2'-dependent (C-type) lectins.2 Hemagglutinins. but the sugar-binding activity can usually be ascribed to a single protein module within the lectin polypeptide. however. which are neuraminidase inhibitors. (6) the P-type lectins. Increasingly. are being investigated for their ability to inhibit rotavirus infection (see Section 2. lectins do not contain enzymatic activity. and (7) the Rtype lectins. (2)the galectins. is one of the coat protein spikes on these viruses and is reauired for release of the new virions from the host cells.4 Lectins A discussion on proteins that specifically recognize the carbohydrate-containing molecules. which cleaves sialic acid. (5) calnexin and calreticulin. and transmembrane lectins with tandem CRDs) and that are soluble proteins (lecticans and collectins).4) (200). In HBV. (3) the annexins. and as a result. Lectins are defined as those proteins that specifically bind to carbohydrate ligands. - 4. which contain CRDs similar to those found in ricin. Such a module is designated a carbohydrate-recognition domain (CRD). There are already two marketed drugs. which and retain bind to N-linked oli~osaccharides misfolded glycoproteins in the endoplasmic reticulum. the new virions remain aggregated at the host cell surface through their coat hemagglutinin. could be an effective treatment for both hepatitis B virus (HBV) and HIV (202). which seem to block rotavirus adhesion to host cells. treatment with N-butyl deoxynojirimycin (15) did not prevent viral coat assembly. however. zanamivir and oseltamivir. but is important for the maintenance of a structure that permits receptorlligand interactions. This is believed to be because of conformation changes in gp120 coat protein.3 and 2. which prevented its shedding after CD4 binding.2.1 C-Type Lectins. This prevented viral secretion. and as a result. Lectins are found throughout multiple organisms. the lectins. Currently. which inhibits the ER resident a-glycosidases I and I1 and is important for trimming terminal glucoses in the maturation pathway of precursor N-glycans (2011. recent studies have determined they also bind to glycoconjugates. which are sialic acid analogs. They are often complex. (4)siglecs (I-type lectins). The animal lectin family can be loosely divided into seven subfamilies based on their binding specificities and requirements: (1)The Ca2+-dependent (C-type) lectins. multi-domain proteins. which are a diverse group of proteins containing a highly conserved CRD and the requirement of Ca2+ for binding. the discussion will be limited to the animal lectins. In recent years there has been a surge in identification and clarification of the biological role of the lectin family of proteins. N-butyl deoxynojirimycin (15) treatment resulted in hyperglucosylated M coat proteins that were misfolded.Carbohydrate-Based Therapeutics mAbs in which recruitment of immune system to kill target cells could be desired. which bind mannose-6-phosphate moieties. 4. in this review. Other anti-viral strategies involve the use of glycosylation inhibitors to interfere with the folding of viral envelope proteins. The neuraminidase enzyme. The Ca2+ is not directly involved with the CRD binding. type I1 receptors.

Recently efomycine M. has been shown to bind to all the selectins. Early on. Conversely. Mannan- binding lectin (MBL) is a soluble C-type lectin in the collectin family. fucose. E. Conceivably. a macrolide antibiotic produced by Streptomyces with a molecular structure similar to sLex. cutaneous lymphocyte antigen (CLA). synovial fluid.2 Mannan-Binding Lectin. 4. Furthermore. Selectins bind to oligosaccharides with sLex and sialylated Lewis a oligosaccharide determinants. The collectins bind to the terminal non-reducing sugar residues mannose.54. There seems to be a higher frequency of these mutations and low MBL serum levels in patients experiencing high numbers of unexplained infections or those with systemic lupus erythematosus. The selectins are t. MASP-2 (2121. molecule glycomimetics that could inhibit selectin interactions were tested for efficacy in treating of a number of inflammatory diseases. Interestingly. It is hoped that this strategy will be more effective than the previous attempts using drugs that mimicked the sLex and sialylated Lewis epitopes. it recognizes a wide spectrum of oligosaccharides (211) and binds through multiple lectin domains to the repeated arrays of carbohydrate residues present on the cell surface of many microorganism. For example. which acts to trimerize the molecule.2. Another approach for targeting selectin-mediated inhibition is through the glycosylases that modify their ligands. galactosyl transferase levels are decreased. Unfortunately. .ype membrane proteins. they were unsuccessful.4. suggesting that blocking this selectin interaction could be an effective psoriasis treatment. and E-selectin. consisting of I P-. In RA.and P-selectins are expressed on activated endothelial cells and bind ligands on leukocytes. the important T-cell homing receptor to the skin. such as the carbohydrate sulfotransferases (205) and the fucosyltransferases (206). There is also evidence suggesting that HIV-positive men harboring these MBL mutations experience more rapid AIDS progression. MBL is part of the complement system and plays a key role in immune defense. Interestingly. however. or 57 have been reported to occur naturally in humans and disrupt MBL secondary structure. which may allow for inappropriate activation of the innate immune system. resulting in increased concentrations of the agalactosyl glycoforms of IgG (IgGO) in the serum. Mutations in codon 52. a number of immune diseases have been correlated with low MBL activity. rPSGL-Ig was advanced to the clinic for ischemia/reperfusion injury. and synovial tissues of RA patients. The collectins (collagen-like lectins) consist of a carboxyl terminal C-type lectin domain and an amino-terminal collagen domain (Gly-Xaa-Yaa triplet). L-selectin is expressed on leukocytes and binds to ligands on activated endothelial cells. originally identified as a ligand for P-selectin. inhibiting MBL binding activity or increasing galactosyl transferase activity in RA patients could reduce the severity of disease. or binding to phagocyte cell surface receptors.4). MBL has been implicated in contributing to the pathology of rheumatoid arthritis (RA). GlcNAc. Mice null for HEC-GlcNAcGST activity showed markedly reduced L-selectin activity (207). with a single CRD. Evidence strongly suggests that CLA is important for T-cell homing to the skin in psoriatic patients (209). and are involved in the leukocyte-endothelial cell interactions in the initial homing of leukocytes to sites of inflammation. no data has been published on the affinity of PSGL-1 for the selectins. In a similar strategy. In the presence of Ca2+. whose expression is restricted to high endothelial venules. PSGL-1 (P-selectin glycoprotein ligand-l). and glucose and play important roles in innate immunity (210). Perhaps this molecule will be more potent. is slight variant of PSGL-1 containing a carbohydrate element that has likely been modified by fucosyltransferase VII (208). showed selective selectin inhibition (204). is a sulfotransferase responsible for addition of the critical binding determinant of GlcNAc-6-sulfate to L-selection ligands. HEC-GlcNAcGST. This results either in the recruitment of the complement system through an associated serum protease. MBL recognizes these truncated carbohydrate side-chains.the best understood among -type lectins are the selectins. making it hard to predict whether this molecule will be any more clinically effective than its carbohydrate-based prc?decessors(see Section 2.

the complete role of their carbohydrate structures is not fully understood. and the presence of a-1-&linked frucose all influence siglec binding to their ligands (214). Conversely. On cell activation. siglec-2 (CD22) is an inhibitory receptor for Bcell receptor (BCR) signaling.3 Siglecs. With the exception of the sialyl Tn antigen. as well as monoclonal antibody therapies. these sialic acid structures are commonly found on the cell surface and in extracellular spaces. a-2-6-linked sialylactosamine. a membrane mucin that can be bound specifically by siglec-l(213). by an unknown mechanism. and while initially it looked like carbohydrate therapeutics would not be feasible drugs.4. it is clear that carbohydrates have emerged as a significant new approach to drug discovery. - 4. there should be a greater ability to produce "designer" drugs with potentially longer half-lifes and fewer side effects. many siglecs contain immune receptor tyrosine-based inhibitory motifs (ITIMs)in the membrane proximal region of their cytoplasmic domains. and these macrophages bind to carcinomas expressing MUC-1." Modulation of siglec binding can also be controlled by post-translational modifications in their cytoplasmic domains. As scientists dissect the heterogeneity of carbohydrates and understand their role in proteintlipid structure. For example. A sustained launch of two new prod- . and there is a number of examples of successful glycomimetics. underlying sugars. All siglecs. and improving cell lines so that they consistently glycosylate these therapeutics correctly will be an important step for future developments. At least two new carbohydrate-based drugs were registered on an annual basis as therapeutics during the last decade (Table 7.5 Health and Disease The role of carbohydrates in health and disease is increasingly being appreciated. There are at least 1 1 family members. are "masked" bv " cis interactions with sialic acids on the same cell surface. Different sialic acid side-chains.2). like other inhibitory receptors in the immune system. there are currently at least several biotechnical companies taking carbohydrate-based therapies into the clinic. Perhaps the lectin family that has seen the biggest growth in the past couple of years is the siglec family. production of these glycoproteins is still not perfected. and presumably this mechanism could apply to other siglecs. and anti-CD22 (siglec-2) monoclonal antibodies have shown promise as a lymphoma therapy (218). Studies examining binding specificities have determined that siglecs bind with varying preferences to at least four different terminal sugars: a-2-3-linked sialylactosamine. and the sialyl Tn epitope (Neu5Aca2-GGlcNAc). High levels of siglec-1 (sialoadhesinlCD169) are expressed on tumor infiltrating macrophages in breast cancer. which also contain putative serinelthreonine phosphorylation sites (216). For many of these glycoproteins. For example. it is hoped that these molecules will have therapeutic value because a number of them seem to play a role in both the adaptive and innate immune system. Most of the biological therapeutics already on the market are glycoproteins. function. Additionally. Considering the therapeutics now in development (Table 7. a-2-8-linked sialic acid. there is clinical trial data suggesting that humanized antibodies against siglec-3 (CD33) could be used for selective ablation of acute myeloid leukemia (217). which are expressed abundantly on cells within the immune system.1). glycosylation on the siglecs themselves also affect binding (215). thereby preventing cell-cell interactions.Carbohydrate-Based Therapeutics 4. Siglecs are sialic acid-binding Ig-like lectins. 5 SUMMARY-CONTINUE THE PACE Now that we are in the 21st century. protein kinase C phos- phorylation of serinelthreonine residues in the siglec-3 cytoplasmic domain modulated it sialic acid binding. and 6 of these have been identified only within the last 4 years. While the exact mechanisms of action of these molecules are unknown. a tumor marker associated with poor prognosis. While the number of therapeutics based on siglecs is limited. with the exception of siglec-1. the siglecs seem to be "unmasked. Furthermore. this pace should continue into the near future. and ligand interactions. Sialic acid binding also serves to regulate siglec-mediated cell-cell interactions.

G.1684-1690 (1996). Washington. and J. Calfee and F. Hirsh. Vol. The Use ofhtibiotics: A Clinical Review of Antibacterial. Webster. and C. L. Virol. 901. Fisher. S. 478-483 (1992). 5th ed.. Med. A. H. J. Raschke. J. 269. J. P. J. 19. D. Burlington. Clin. 56. 14. Vincent. Sci. D. Kubota. R. E. W. J. J. 39. 604611 (1998). Labeaga.-F. Ishida. Ely. 1997.J. Suzuki. II Congresso Congiunto Italiano-Spagnolo di Chimica Farmaceutica. Am. M. Lancet. LaRosa. E. and S. Grayson. Kinnersley. 11. . Chest. R. 1998. and R. 1998. N. 471476 (1997).673-678 (1998). Garber. Crowe. N. Hoy. 6. J. Ward. Ehrlich. Mukai. Hulihan. and A. Assoc. Timmer. Hogerman. J. H. Osterhaus.. R. Sekino. 19. F. J. Adams. W. Engl. Hayden. K.2589-2753 (1995). G. and M.. W. Ohashi. 699-709 (2001). Fuder. N. 1994. P. (1995). For new carbohydrate-based therapeutics to continue at this pace of product registration in the intermediate and long term. 1-10 (1999). 5th ed.. C. A. Horm. G. R. Liicker. and P. Chang. J. Kageyama. 18771881 (1998). Edwards. Reisinger. Pharmacol. 46. P. F. MIST (Management of Influenza in the Southern Hemisphere Trialists). 7.Jan 06. and Y. T. J. Matsui. J. Likewise. Nakano. G. Madore. Paradiso.ucts annually is an enviable accomplishment even to a large. Vrooman. P. 12. 53. L. Rimmelzwaan. A. a way to make interesting research tools needs to be done. 2. P. M. Ikeda. Ther. Menheere. Balfour and D. or at best. S. examples that illustrate how carbohydrate-based libraries can actually accelerate drug discovery need to be shown. P. Chemother. H. Wilder. DC. Butterworth-Heinemann. E. K. V. and T. S. August 30September 3. 13. G. and A. Wetzelsberger. 33. MA. 15. 1. Pledger. P. S. 3. Antifungal and Antiviral Drugs. Edwards. 15. W. C. Wolff. Seino. Burger's Medicinal Chemistry. 101. Rennels. 46. A. P. G. L. Kajikawa. P. Mizuno. Mosquera. Final Program and Exhibits Addendum of the 38th Interscience Conference on Antimicrobial Agents and Chemotherapy. Antimicrob. J. and A. G. Italy. M. Gen. 720-729 (1997). DC. American Society for Microbiology. Warkentin. Updates Clin. Fujimoto. 21. Hayden. 108. Washington. Lopez-Rodriguez. Bernard. M. W. H. A.. Yamada. Dhainaut. Kimura. Malinoski. F. Kato. June 22-23. Karim. 17. solid and solution phase. Musser. An important factor for the success of carbohydratebased therapeutics in the last decade is the combination of glycobiology and carbohydrate chemistry. Nakamichi. Despite the failures of carbohydrate companies. Orjales. Laver. E. Diabetes Care. Also needed to continue the pace of product introduction is the application of a multi-disciplinary approach to drug discovery. Ward. T.. Anderson in M. Nishimura. D. S. Final Program and Exhibits Addendum of the 38th Interscience Conference on Antimicrobial Agents and Chemotherapy. Kleist. N. D. G. . J.-L. 10. Compans. A. New York. E. Steingrub. 1025-1054 (1993). Work to determine whether carbohydrate multivalency is a successful approach to new drug discovery. Med.537553 (1998). Helterbrand. H.159-163 (1976). Am. J. 18.. Reife. Clin. P. 352. Pediatrics. G. Watanabe. 30. John Wiley and Sons. l . Mills. I. McTavish. 8. Y. and R. Keyserling. J. P Sels. D. Drugs. N. K. S. S. H. B. J. 20. Laterre. and R. Drugs. Kinnersley. L. J. Aoki.. S. P. Palese and R. 344. Metab. Kramer.. Ramsay. A. P. K. et a 221-226 (1993). M. N. Milano. B. 16. E.. This includes any combination of automated. A. K. American Society for Microbiology. Odaka. Y. W. REFERENCES 1. p. R. Faught. the venture community should continue to fund the carbohydrate-based technologies and drug discovery efforts because there will be a significant return on investment. 2001. 5. M. Bischofberger. et al. Kraut. Fugedi. L. Treanor. several objectives must be reached. Neurology. N. J. Ed. Eur. Technologies supplying carbohydrates in quality as well as in quantity must be refined or discovered. A... P. Neurology.. B. G. global pharmaceutical company. 9. and enzymatic synthesis. Res. Nieuwenhuijzen-Kruseman. and H. 4. 149 (1997). et al. W. Kucers. R.-F. Kingma.

Yogo. Jaffee. K. Angewandte Chemie Intern. Ed. Koyama. Oncol. J. Peterson and J. and S. J. 38. 46. H. Acquir. 337. Bornmann. L. 1995. R.-F. M. 38.1149-1154 (1995). S. Akima. 0. USA. Natl. 1999. 31. Stephens. Goodman. I. 137141 (1998). Sunkara. Ragupathi. Goa and P. K. L. Ann. Merhi. J. Bryla. 0. P. Bencsath. CA. S. Pharm. 50. K. J. Geoffroy. Med. Terry. Oncol. Aarestrup. 98. Kumar. M. S. J. 24. K. M. D. 29.527532 (1994). Proc. W. J. pp. 6. M. 27. Benfield. Goy. L. Microb. Diran.Livingston. S. CRC Press. 22. Roeltgen. 345 (2001). K.447 (1997). Kirkwood. 514 (1996). Goldhaber. W. 26. 23. Olkiewicz. 21 (1999). J. and J Rao. and A.10-15 (1993). Theoret. I. Wheeler. R. Sirois. Altman. B. Phillis. 58. G. M..Carbohydrate-Based Therapeutics 22. A.. M. Engl.21 (1998).. Angew Chem. Koga. 30. Haemost. J. Med. S. 0. Davey. Magnani. H. Slovin. J. A. Sakai. Ikuta. N. 47. Boca Raton. 109-111 (1992). N. Langer.694-695 (1999). BioWorld Today. J. S. 20. Weatherley.. 53. T. and S. S. A. abstract 290 (2000).. F. ten Cate. K. I. P. Ed.. G. 57. Dec. Fromell. Therap.. 47. J Immunother. W. Geisler Gangliosides in P. Danishefsky. Bornmann. M. Kim. J. 2203-2012 (1998). S. P. Ragupathi. 139-147 (1994). C. Press Release. XIVth World Congress of Pharmacology. 42. F. Zhang. J. C. T. Musselli. H. and H. D. Coombs. Greten and E. Drug Resist. 54. M. Cohen. J. 34. Infect. T... 20( suppl39). B. I. J. Turpie. Turpie. and A. Dantis. Cancer Detect.3334-3345 (1999). G. 32.90-96 (1990). J. J..Lloyd. R. ten Cate. R. Chapas-Crilly.. T. and B.. Livingston.. M. G. Weitz. 48.. Nurmohamed. J. 52. Morrison. R. Orlando. C. V. R. 19. Dhume.. J. Dis. 53. Danishefsky. Preu. R. Z. A575 (1994). Fischl. Liu. 3270-3275 (2001). E.. A. M. C. F. H. Rothblat. L. J. Exp. M. G. C.11. X. Sci. Spassova. P. J. J. Palmantier. J. and B. Bowlin. Z. J. Drugs. L. Natl.. Haemostasis. Y. Lee Jr. Kim. Eds. C. Med. P. Tyms. H. Slovin. K. Kirkwood. Rodriguez. Thromb. 25. 37. Mitchell. 45. J. Parke.736-751 (1997). Takanashi. A.-F. and T. Scher. K. Rheumatol. Kim. L. J. W. S. Bhuta. G. Int. S. R. 49. D. M. 40. S. 0. 56. Hyalgan. Nova Scotia. (1999). Parish. Sci. FL. G. S. Clin. D. H. July 7-12. 106. Intern. Halifax. Traumatic and Ischemic Injuries to the Brain and Spinal Cord: Prevention and Repair. S. J. and J.5710-5715 (1999). Clin. P. Engl. Tran. H. Podger. K. Sandmaier. Williams. J. 177. M. K. C. 2001. K. B. Hudis. Syndr. Glycoconj. K. S. D. Nagata. NY Acad. A. Creager. M. Bigonzi. F. J. Sang Thrombose Vaisseaux. Holmberg. N. Seraydarian. K. Ozaki. G. 96. D. H. Spassova. Peyron. Neose Technologies.aegis. Lloyd. 36. A. Oparin. P. 158. K. Robbins. M. and S. Acad. Adluri. and S. Chapman. L. Hum. D. 25. A. 554561 (1993). P. G. Engl. G. Sames. Schneider. E. N. F. 35. Morrissey. A. M. 33. A.563-566 (1999). L. P. Suzuki. 17. 688-698 (1997). BioWorld Today. 271. L. 291-310. Gilewski. T. M. Sjoerdsma. Pihl-Carey. Tanguay. Kher. Lassen. Califf. Chin. Clin. Nicolaou. T. Am. Eriksson. Immune Defic. Laporte-Simitsidis. 18. 44. http://www. Rheumatol. 51.2370-2380 (2001). Soc. Houghton. Fox. 616. Pavlakis. Scher. A. M. J . G..1996.. C. 54-66 (1999). N. 45. et al. Noyton.574-579 (1994). L. George. I. G. J . M. Israel. S. A. D. M. Omura. H. M. Raz. 0. I. D. Moskowitz. 39. Canadian Cardiovascular Congress. Bauer. A. Livingston. H. 55. 28. . Reddish. 354. Demers. Oncol. Lancet. Ungem. B. J. M. Med. 7. Shrine. R. and Y. S. H.-F. Ibrahim. P.. November 21. Danishefsky. Schaub. MacLean. FL. G. Richards. Lawson.536-566 (1994). Sci. P. L. Inc.. 76-82 (2000). S. A. and H. J.. San Francisco. 43.com. Satoh. Olden. Klein. N.. F. Resnick.. Cancer Res. Itoh. B. 4. 70. W. Clin. Sano. Premmereur.383 (1998). Atkins. J. Kumar. Fujikura. and H. A. Neurology. K. Gurfinkel. Roberts. Arch. A. Retrovirus. Inohara and A. T.. USA. 1047 (1999). B. 22. ten Cate. Ohshima. 337. G. and J. R. Mismett. Konodu. Acad. 12. R. M. 41. J. Longenecker. K. and K. Fazzari. Szu. 2002. Adluri. J . J. Fylaktakidou. J. and H. and S. Proc. M. J. H. Kang. Drugs. D. M. Gastroenterology. 38. Panageas. Ragupathi. and R.

R. Yednock. C. Am. G. C. R. D. A. Lu. and S. D. Soc. Brandley. Stachel. P.. 3 (2000).. Kannagi and A. Exp. K. Chem. Delpierre. Nature. M. M. 1473. Proc. Schmid. Int. Chang. T. T. and C. Andrews.-Y.-H. V. H. Mobasseri. Acad. J. Science. Bergstrom. Warner. H. Sanchez. Jennings. Weitz-Schmidt. Christopher. L. Vestweber. Rodriguez. Glycobiology. 391415. Sames. and D. H. Soc. Livingston. and R. Kahne. I. J .. Freeze. 84. X. 94. 0. Rev. A. Bertozzi. Thoma. Cell. Wong. 65. L. B.833 (1998). B. and S. M. Sci. J. White. N. P. Engl. Soc. 91. G. Rider. and E.335. J. and S. L. Collard. D. Groth. L. 80. Ley. Qiu. Hartz.. Marth. E. Miller. M. M. Fraser-Reid. K. E. J. 276. 2893 (2000) 85. E. Proc. L. R..rences G. Angew Chem. Roush. J. V. New York. Wang. Schmidt. 3 (1990). Dowbenko. McGarvey. Soc. 96. and S. 88. Leffler. Granovsb.861 (1990). H. Herold. K. Natl. Cummings. Whitesides. Bruns. R. J. D. Thompson. and G. J. Block. 284. D. Horan. Phillips. Vachino. S. 669-675 (1995). H. V. USA. M. Varki. pp. Y.. . Abul Milh. Wong. C. Warren. Cayetano. Beaudet. W.587-591 (1997). P. Acta. C.. R. J. Fukuda.. X. 1337-1348 (1997). 6 2 . J. Danishefsky. 6. S. Harris. Danishefsb. R. Shames. C. D. J. Sako. 610-614 (1998). Eds. Cold Spring Harbor Laboratory Press. Ed. Glycoconj. J . J. Ge. T. Yehiely. Krug. J. L. Chem. J. 49. Chim. W. Karlsson. H. Ulrich.. 181. and H.-A. Fennie. Trends Glycosci. Diabetes. Science. M. Glunz. E. G. Nut. Neuron. 1999. Biophys. Ollmann. 187-191 (1991). F. Kahne and M. and K. Med. and J. 89.. Ge... K. and S. 11014 (1998). H. X. M. Acta. M. A. S. Soc. K.1520-1522 (1996). J. S. Pier. A. Smith. Fata. J. Chem. J. V. L. A.. L. H.-A. J. J. White. R. Simon. Cohen. 63. Chem. Cummings in A. J. 20. S. 83. Pure Appl. 12035 (1995). and K.. C. Wong. 86. K. Angew Chem. 750 (1997). 77.21-34 (1999). Loebach. 1179-1186 (1993). Miller-Podraza. F. G. 98. J. Sekanina. Natl. Jablonowski. 19. M. J. Lloyd.. R. Immun. Dis. Ed. Hintermann. and G. S. S. G. 3264 (2001). DeArmond. L. Van Schaftingen. A. J. Veldman. J. Gustin. B. Cell. C. S. D. Pawling. Science.. E. Abul Milh. Furie. Daggett. USA. 98. 120. Shaw. J. Handlon. Nut. P. J. V. H. Dennis. Infect. 34. I. Arbones. Muller. Tetrahedron Lett. Chem. 1045 (1989). K. Pauley. Tsai. Wandosell. Oncol. et al. M. PBrez. G. Kudryshov. G. Simanek. M. Yan. 41. P. M. 1125-1128 (1997). McKenney. Clin. 37.. Robbins. W. R. M. and W. J. J. 123. Singhal. Acad. Paulsen. 1. Colaqo. R Bertozzi. W. Murthy. J. W. A. W. F. Vanstapel. Chem. 93. J. E. Biswas. J. and J.. and C. Dubois. A. and P. Helin. Lee. 92. Ng.. D. 88. Yang. Paulsen. P. Doring. Goldmann. et al. C.-H.Am. T. X. Kanamori. 453 (1996). Engl. G. K. A. 1990 (1999). Am. M. M. Hiruma. J. Emerg.9265-9270 (1996). 0 . Biol. D. Bullard. Hemmerich. P. Liang. Michaelson. Ninchak-Casey. Kajimoto. Barone. L. Dennis. Yarema. D. J. 83. A. Sci. 11. Med. Rosen. S.23442350 (2001). A. Essentials of Glycobiology. 97. Sears and C. Am. Biochem. R.. S. L. M. Wong. 13.. J. and J. Cummings. Santos. T. and R. D. McEver. and A. Hart. 1627-1634 (2000). Chem. A. Ragupathi. 14.1241 (1998). 6224 (1991). Med. 270. T. Torchia.9499 (2000). A. Khokha. B. M. K. P. Glycoconj. K. J . M. 467 (1997). G. C. Banteli. F. R. Biol. Int. Bosse. Gildersleeve. Allen. Hakomori. Tremblay. Chem. W. J. and S. Nudleman. C. T. Merritt. A. Rademann. S. 65. J. and C. Lasky. and C. and K.779 (1993). Biochemistry. Ahern. M. 275. R. Nguyen. Bergstrom. 155 (1982). G. D. 66. and B. H.-H. 75. Y. Cell. Simon. Danishefsky.369-374 (1998). R. and C. A. Stroolbant. Rosen. 274. Rosen. 389. Mahal. 21. Huang. Kris. Tedder. Leppanen. K. C. and D. Granovsky. S.14271 (1995). Rogers. 79 (1995). L. Avila. 95. Cell. D. 78.329-344 (1999). F.-H. K. Geyer. 118. Singer.. J.1523-1527 (1999). P. R.. M. T. 1. Biochim. J. R. 90. and C. 39569 (2000). M. and L.1890-1897 (2001). 2644 (2001). Scott. K. Osborn. Paminiano-Camerino. Chen. Esko. Uozumi. Y. S. 291. N. Karlsson. Magnani. L. Prusiner. Still. E. W. Patton. 81. F. Glycotechnol. Miller-Podraza. Science.. M. 87. C. Goode. Sweidler. Am. Infect. M. Pouliot. D. Helv. Bean. 82. L. 79. D. Mehta. 121. A.

Res. K. J. and G. 39.1162 (2000). 4262-4271 (1996). Kahne. L. Ge. Angew Chem. 4. Bock. K. 5429-5437 (2000).549 (1998). Am. Koeller and C. R. K. Med. 122. Chem. Winssinger. Nishimura. Chem. 1523-1527 (2001). 1. Soc. 109. and F. N. T. J. Isles. Paulsen. Schultz. J. 107. Swanson. Curr. 274. I. Hilaire and M. and L. Org. Wong. Rev. 108. BonnaffB. C.13113-13114 (1996). J. 118. Wong. W. Soc. Chem. Int. M. Cottaz. K. DeRoose. Schulein. R. 120. X. UK. R. Wang. 120. 5. Smith. G . and S. J. 112... J. A. Wiley. J. 137. 125. Baasov. 133. M. Soc. Y. 291. T. S.8869 (2000). Rev. Warriner. 0.. 135.2845-2847 (1996). Di Bussolo. Huang. Am... J. 33. 25. Am. Horan. Ed. S. Inoue. Zhang. Williams... Wong. 39.. Am. 119. X. Kanie.-C. Chem. P. V.. Q. 113. Ed.. and S. 0. Martin and R. Still. Engl. Inoue. 1. R. St. 204 (2000). Tang. Y. 1520 (1996). Labbe. P. S. 4349-4394 (2000). 0. Walter. Am. Soc. D. Ikegami. 119. P.1979 (1998). 104. 129. Chem. Liu. G. Chen. R. Meldal. 118. P. J. Perkin Trans. 110. Kitajima. Kahne. J. J.. K. K. Chem. S. Soc. Am. 1617 (1997). Dillard.4. C. and W. Ohlmeyer. 120. Kunz. Chem. J. Boons. S. West Sussex.. L. 102. S OC. Nicolaau. Boyer. 2523-2532 (1999). J.11137-11146 (1997). Witte. 36. Soc. H. Chem. R. Sears and C. 100. J. Biol.Carbohydrate-Based Therapeutics 98. N. and C. Wong. Opin. 3193 (1999). P.. Am. T. Am. Whitesides. 114. Christiansen. Chem. Haase. Ding. Y. Machajewski and C. Soc. Science. Chen. Hindsgaul. Wischnat. J. Boons and S. 42. and S. Soc. Reipen. Ed. Chem. 55835584 (1998). and C. C. J. 2001. J . 130 (1997). Rev. Ed.1352-1374 (2000).4465 (2000). M. J. T. D. Am. Paulsen. Liu. Bock. Wong.-H.-I. J. Asouline. Kunz.. Reipen.-Q.-H Wong. Soc. 116.2081-2082 (2001). Schultz. Andreana. T. M... H. Heidlas. J. M. J. Suzuki. Oscarson and F.-H. Reader. Chem. T.J. 123. Yan. Chem. Chem. Yamada and S. L.-H. 100... 51 (1998). Dwek. J. Thompson. J. H. Sekanina. 2359 (1997). G. Z.-H. 128. Ley. Sehgelmeble. Wang. Moroz. 100. Chem. V. Seeberger. J. Engl. Davies. E.. X..23442350 (2001). Witte. G. Soc. C. Seeberger. M. J. Wang. and S. Kobayashi. Perkin Trans. Warren. 61. 124. P. 0. G . Nature. D. Rev. P. K. Chem. J. L.307 (1992). M. Z. et al. Zhang. R. Knorst. 0. Chem. Kimura. 121. Soc. M. 117. Chem. R.. H. L. J. Uozumi.. L. J . 99. Chem. N. Yamada. Wong. Sames. 101. Chem. 105. H. C. J. Science. Chem. D. R. Ollmann. Gildersleeve. H. Kowal. 100. P. Seeberger and W. K. G.-H. J. Schmidt. M. Withers.1979-1989 (1998).587 (1997). Kiyosada. H. Engl. Biswas. N. S. T. Chem. 10555-10556 (1997).. Heskamp.6123 (1992). J.. 685 (1989).. Shoda. 122. Bioorg. Chem. M. J. and H Driguez. Engl..23482350 (1996). and 0. S. 127. 120. Angew Chem... and C. and S. N. S . Commun. M.-C. Am.4495 (2000).. Greffe. J. and Y. Meldal. Angew Chem. W. K. Lubineau and D. Cirillo.. Y. 136.-I. 1. 131. A.-X.. 232-240 (2001).. Chem. and H.683-692 (1996). Chem. J. Eckhardt. 734 (1999). Perkin Trans. J. A. . Liang. T. and C. 449 (1997). K. 35. T. Ye. Chem. Fessner. Chem. T. H. Danishefsky.4465 (2000). 409. H. C. 1. 134.-T. J. Int. A.-H. 291. 106. Soc. Rev. 1882-1890 (1999). Engl. J. Ed. Fort. Danishefsky. Lee. 126. Douglas. Seitz.. Palmacci. Shao. and H. L. Fang. 123. Sofia. S. Meldal.4495 (2000). Pastor. Koeller and C. M. F Mackenzie.-D. Kobayashi. Chem. and D. Chem. 138. S.. 389.. Tetrahedron Lett. Herzner. 115. Eur. M. G. Nishimura and K. U. Herzner. Angew Chem. J. 9493 (1995). M. J. Seeberger. and P. W. 100. Hout. Ed. Tetmhedron Lett. Org. Wong. M. and K. Seitz. W. and K. and P.. Am. 130. Seeberger and W. Hu. F. Honda. Takayama. Zannetti. 132. C. 103. J. Plante. Gin. Int. Fan. L. H. Koeller and C. 119. Acc. Angew Chem. Wong. Med. 100. H. 139. Liicking. Chem. M. 111. E. Eur. 122. Science. B. 121. Soc. S. Soc. Solid Support Oligosaccharide Synthesis and Combinatorial Carbohydrate Libraries. K. Nature. Hashimoto. V. J.4549 (2000). 140. 119. and F. 39. Int. and P. H. 100. Haase. Rev. E.. Chem. 35. Eur. Znt.-H. M..3. Loebach. Mathieux. Meinjohanns. Huffman. J.

Rev. Page and W . 400. Curr. 757 (1996). Vinogradov and K. 68.2002.10922-10926 (1993).. and C.. K. M . Oxford. Pannu. J . E. Proc. Wang. 398. Loebach... Bioorganic Chemistry: Carbohydrates. Chem. R. 276. 1382-1391 (1998). R. Duchaussoy. 65-94 (1992). Ed. 769 (1999). J. K. 393. Am. S. D. Mass Spectrom. Dionne. Minke.. Int. D. O'Brien. Y . G. L. C. R. J. Koeller. J. Am. 14. Wong. Thompson.. S. Cancer Res. Burke. Soc. Whitesides. D. FEBSLett. Nature. Knudson. Immunol. M. J. D. Fan. Hale. G. Ling. Org. Zhao. A..-P. J.-A. Weis and K. J. A. Med. B. M . R. C. Engl. Z. Boehr. 2135 (2000). 669-672 (2000). K. Gildersleeve. Soc. 52305231 (2000). H.nl/ static/sugabase/sugabase. 114. C. K. 1113-1116 (2000). C . B. . Serianni in S.318-324 (1999). Verlinde. S. Angew Chem. Chem. J. G.chem. Hkrault.W .4518 (2000). Int. M. Read. 18. et al.7303-7305 (1992). Fulton and J. M.687-698 (1995). Gavel and H. Jackson. M Bovin. Bock. 1135. R. Arison. L. 291. Chan. Ann.. 841 (1989). Nature. Hirschmann. Ge. 63. Whitesides. G. van Kuik. Soc. Bahr. Dulina. G. M. C. E. J. Radaev and P. J. Toone. 23512356 (2001). and G. J. Proc. L. 37. Harvey. Luzier. 9. M. Meng. D.2663-2664 (2000). D. 291. et al. E. Sun. Chem. B. G. A. J. R. von Heijne. Biol. D. S. 4 . Gange. Digabriele.-A. et al. Semin. Spaltenstein and G.. Oxford University Press. Dell and H. Yan. Bennett. Herbert. J. Kahne. Lormeau. C . 38. Soc. F. Chem.855-858 (2000). Sigal. Cell Dev.95 (2000). 38. Moscatelli. S. P. Nature. B. N . and G. Ed. L. C. G. Top. A. U ...-M. L. Wong. A. Int. Am. Wooster and A.. D. D.Wigler. et al. A. 33.1632-1637 (1999). J. Liang. M ..-K. Natl. Bernat. et al. Angew Chem..185-199 (1992).. Tetrahedron. Dell and H.686-687 (1991).69-78 (2001). C. K. Acad. Bordoli. 3. et al.. D. Microbiol. Wong. Engl. Mammen. A. J. C. Org. J. E. S. 113. Zhang. Ingendoh. Soc. B.41527-41542 (2001). G. G. and W . Sucheck. D. F. C. K. M. Ed. and L. A.209 (1990). P.. Annu. K. I. E. Y . J. Roseman. P. M. Smith. Science. M. Simon. Tyldesley. P. L. Kiessling. and F. Angew Chem. Transplantation.537 (1999). 286. Ed. 28. 37. 11. Science. Horan. Mazieres.1520 (1996). Ed. B. Chatenoud. Biophys. 180.. Armstrong. E. P. Bertozzi and L. J. Drickamer. accessed on September 25. J. Kingery-Wood. M.Whitesides and S.. 90. 128. Pharmacother. R.. Lett. Salvino. J. Sci. A.1678 (1971). 1073-1083 (2002). Wright.Whitehouse. J. Cichy. and W . Nicolaou. W . G. Soc. and D. C. 173-181 (2001). M. J. 433-442 (1990). M. R. M. available online at http://www. Combe. Science.. Choi. Science.. Schuster. Nature.2338-2343 (2001). USA.. and R. C. J. Silva and M. Mass Spectrom. J.. G. Angew Chem. Basilica. W . 41. A. Hunter. Alper. 291. S. Tetrahedron Lett.. Biomed. Still. Schlessinger and J.boc. Chem. Bundle.671-674 (1999).. I. M. Draker. Pietranico. Biochem. et al. 3. Science. Driguez. Sadowska. C. Petitou. Acad. R.2754 (1998). G. J. Williams. Wong. 122. S.. N. Chem. Sofia. Int. V a n Phan. and D. 115. K. R. Jencks. 41. Q. 122. Experimental Sugabase. USA. Sears. Hirschmann. Sekanina.. T . 65. Immunol. Karas.. Uozumi. A. Friend. 1524 (1998).23512356 (2001). Hol.. M. M Whitesides. 274. Biol. R.. W . Fenn.417-422 (1999). Mann. A. 37 (1990). 37. Chem. Rheumatol.-K. Hecht.. Choi. Stoddart. Am. H. D.rences M. J. A. J. Protein Eng. R. E. P. Morris. Spoors. L. Am. 291.12550-12568 (1993). Am. 12.. Biol. 1999.Whitesides. Rapid Commun. Cell Biol. 115-165 (1992).. P. B. and J.. Sofia. S. Humphries. Shakespeare.-H. 122. 272.html. Mass Spectrom. Rev. 2. A.2802-2803 (1998).-H. E. Mol. Chem. Chem. J. B. J. and D.. M. M. Engl.. Z . and C. Still. J. A. UK. Hou. Science. Bateman.. 59. et al. Mulvey. C. Natl. Sci. Biswas. 812-817 (1998). Kiessling. Adv. et al. Kitov. Chem. H. Acta. W .441 (1996). J. Morris. Knudson and W . Hillenkamp. S. Sasisekharan. 403. Environ. Engl. M. J.2754 (1998). I. D.uu. R.Vaughan. Y. Biochem. 68. 2357-2364 (2001). Leahy.

G. Biochern. 208. J. Med. S. 351-354 (1987). Rudd. Crocker and A. Chem. P.. 188. 430. . 191. W. A. Grob and L. FEBS Lett.. 209. P. et al. M. V. Y. Immunobiology. Dordrecht. Clin. G. Kluwer Academic Publishers Group.237-247 (2001). J. 137-145 (2001). Schon. Lattime. Biophys. Pediatrics. Berd. et al. Drug Discov. Schon. Radiopharm. Cohen. 236.. J..1228-1237 (2001). R.. A. et al. A. 195. Varki.T. Cresswell. E. Kieffer. and G. Barendswaard. 0. Bradley. 276. Shields. 210.3292-3301 (2001). Exp. J. J. M. R. J. Fazli. Hemmerich. 275. 33. 54. R. M. M. Singer. C. R. Ritter. Lu. 0. 57. Tschmelitsch. W. J. Hamvas. 12... 6591-6604 (2001). D. Livingston.. et al. The Netherlands. I. Oosthulzen. Sautes.. Immunogenet. 291. L. P. 192. D. M. L. Immunol. J.. H. G. Kornfeld and S. 365-369 (2001). 327-339 (1998). A. et al. Lymphoma..231-248 (1998). W. Karlsson. Sondermann. Mehta. et al. R. Bioessays. Dwek. L. M... Rev.. J. S. S. A. P. 103. 509-518 (1997). Virol. R. Brinkman-Van der Linden and A. Turner. 2. 204. 194. C.. TX. Birrell.. 213. Wilson. Bernstein.. C. C. K. et al. 665-676 (2000).. R. 126. Biol. Weninger. H. Krahn. 17-22 (1998). Biol. 6. 2(suppl I). Kiefel. K. M. D. 207. J. Annu.631-664 (1985). Immunity. 214. et al.978-981 (1997). 13. 80.Carbohydrate-Based Therapeutics 186. Freeman. 8. C. Tamm. I. 99. Blood.. Ghirlando. 215. 11-35. Da'ron. Jonsdottir. 71-80 (2001). Biol. Res. Fischer. 406.366-372 (2002).. The Immunoglobulin Receptors and Their Physiological and Pathological Roles in Immunity.. 268. Biochem. Chem. P.. E. Varki... et al. N. Maguire Jr. Hulett. Today. D'Alessio. J. 0 .. P. 199. 206.2181-2186 (1997). 202. et al..267-273 (2000). 198. Nature. T. Clin. Keiken. Elliott. P. H. Mimura. Cheng. Dillman. 1998. Williams Jr. 199. and P. et al. et al. Austin. et al. 201. Namenuk. et al. Immunology. 127(suppl 2). Immunity. 70. . M. M. Med. M. R. Cancer Biother. Hendley. J.. Ann. Virol. 389. H. K. 217. Bistrup. Sondermann. 44. 200. and R.. D. Hematol. R. 216. Eur.. 193. Gessner. et al. et al. R20-R26 (2001). C. Clin. Gavin. 305-322 (2000). S. Sigrnundsdottir. Rudd.. Immunol. 15. 196. J. Ann. J. Rev. 15. S. P. Cancer Pract. A. Musselli. C. Wenzel. Oncol. Rev. 1997. 203. Nature. W. Science. Cancer Res. E. S 9 S l l (2002). Compans. Gen. Biochem. Zitzmann.203-234 (1997). E. Hong. B. I. S. 211. 189. B. E. 205. 190. Fuhlbrigge.. Chem... pp. Kornfeld. 212.. M. Nut. G. K. 19. Cell-mediated Effects of Immunoglobulins. Landes Co. 682-686 (1997). Nice. Armerding.159-163 (1976). P.. Ashe. C. 187. M. A.8625-8632 (2000).45539-45547 (2001). Huber.. D. H. J. 9. Turner and R. 276. Gudjonsson. Cancer Res. 218. A. 13-23 (1998). 27-35 (2001). M. 7153-7160 (1996). G. 31883196 (2002). Ulfman.2370-2376 (2001). 76. Hemmerich. Commun. et al.. P.. Dwek.. K. C. A. J. Palese and R. 197. J. Hogarth. Powell. Chem. E. Ragupathi.

2 Receptor-Mediated Oral Absorption Systems. 250 2 Strategies to Enhance Drug Permeability.2. 253 2.3 RME of Vitamins and Metal Ions.3 Bacterial and Plant Toxins. 263 4.1 Maternal and Neonatal IgG Transport.2. 264 249 Burger's Medicinal Chemistry and Drug Discovery Sixth Edition. . Ohio Contents 1 Introduction.CHAPTER EIGHT ! Membrane Transport Proteins and Drug Transport i iL PETER W.5 Lectins (Phytohemagglutinins). 263 4.2 Prodrug Design by Increasing Lipophilicity.2.1 Receptor-Mediated Endocytosis. 254 3 Passive Diffusion. Inc. 259 4.1 Receptor-Mediated Transport. 256 4 Facilitated and Active Transport Pathways. 252 2. 259 4. 257 4. 252 2. Abraham ISBN 0-471-37028-2 O 2003 John Wiley & Sons.1 Kinetics of Passive Diffusion.1.2.3.1 Immunoglobulin Transport.2 Structure of Cell Surface Receptors.2.3.1. 255 3.2. 254 2.2 Lipid-Based Oral Drug Delivery. : Division of Pharmaceutics Ohio State Biophysics Program OSU Heart & Lung Institute Core Laboratory for Bioinformatics and Computational Biology The Ohio State University Columbus. 260 4.3.2 Riboflavin.3 Rationale and Considerations for the Use of Prodrugs.1. 257 4. 251 2.3. 261 4. 251 2.1.3 Prodrugs. 262 4. 264 4. 264 4. 261 4. 262 4.1 Folate.4 Absorption Enhancement by Targeting to Membrane Transporters.2 Polymeric IgA and IgM Transport. SWAAN : b i . 261 4.2.1.3 Transcytosis.4 Viral Hemagglutinins.1 Definition of Prodrugs.3. 263 4.4 Potocytosis. 253 2. 258 4.1 Penetration Enhancers. Volume 2: Drug Development Edited by Donald J.2 Bacterial Adhesins and Invasins.1.

a membrane-embedded transport protein transports molecules against a concentration gradient using an energy supply provided by either adenosine 5'-triphosphate (ATP) hydrolysis or cotransport of ions moving down their concentration gradient.5.5. but the general transepithelial transport mechanisms for drug molecules are similar (Fig. The first obstacle can be overcome by changing the route of administration or drug formulation. especially macromolecules. drugs must first reach the systemic circulation by penetrating the epithelial barriers covering the absorptive surfaces of the body. 284 6 Conclusions and Future Directions. Originally thought to play only a minor role in the overall drug absorption process.275 5. With the exception of a few drug classes.3. 274 5. 265 5.5.5. movement of drug molecules is derived by its concentration gradient. 265 4. 280 5. often Naf or Ht. 274 5. 270 5.3. In active carrier-mediated transport.4 Transferrin.4 Techniques for Studying Integral Membrane Protein Structure.5 Case Studies. 271 5. 272 5.1). such as skin.5 Site-Directed Chemical Cleavage.250 Membrane Transport Proteins and Drug Transport 4.3 Structural Models of Transport Proteins and Methods to Design Substrates.3. 267 5.5. 269 5. 271 5..3 Vitamin B. 272 5. Depending on the nature of the driving force. The significance of their role in drug transport is discussed later in this chapter.5.5 Therapeutic Applications of ASBT. for hydrophilic drugs and drugs with high molecular weight. .1 Introduction. 277 5. 8. 265 5 Transport Proteins.2 Physiology of PepT1.1 The ATP-Binding Cassette (ABC) and Solute Carrier (SLC) Genetic Superfamilies. However.4Use of ASBT for Enhancing Intestinal Absorption. such as general anesthetics and osmotic diuretics. whereas molecules cross the epithelium through the cells when they use the transcellular pathway. 277 5.3. epithelial membranes still impose a formidable barrier to drug entry.5. and (2)a physical barrier originating from the lipid bilayer. In most biological epithelia. 274 5. 283 5.2.4..3Structure-Transport Relationship of PepT1.1 Membrane Insertion Scanning.4. for example. 277 5. intestine. In passive diffusion.5. Substrates. by encapsulating drugs in vehicles impenetrable to metabolic enzymes.4.2. 286 1 INTRODUCTION To exert their desired pharmacological activity.5. Based on the route that drug molecules penetrate. drugs must reach their sites of action with certain minimal effective concentration. drug molecules confront two obstacles in reaching the systemic circulation: (1)a biochemical barrier resulting from enzymatic degradation. distribution.2 Therapeutic Implications of Membrane Transporters.3 The Apical Sodium-Dependent Bile Acid Transporter (ASBT). epithelial transport can be classified into two pathways: paracellular and transcellular. To gain access to these cellular targets. 267 5. 271 5. transcellular transport can be further categorized into passive diffusion and either receptor-mediated or active carrier-mediated translocation. and elimination. 271 5.3 N-Glycosylation and Epitope Scanning Mutagenesis.2 Cys-Scanning Mutagenesis.2 The Intestinal Peptide Transporter (PepTl). and Inhibitors. In paracellular transport. most therapeutic agents produce their effects by acting on specific membrane proteins or intracellular enzymes.4 Excimer Fluorescence. molecules move across the epithelium through intercellular junctions between adjacent cells.3.2.3.3 Structure-Activity Relationships for ASBT.4. 285 7 Acknowledgments.4. and lung.1Introduction. transport proteins have been appreciated recently to be involved in all aspects of drug absorption. 270 5. Epithelia may vary in thickness or functions among different tissues.2 Physiology of ASBT.3.1 P-Glycoprotein: Understanding the Defining Features of Regulators.5.5.

Note that receptor and carriler proteins in epithelial cells are expressed on both the apical and basolateral surfaces. P-glycoprotein mediated). Opening of the tight junctions or changing the epithlelial lipid bilayer membrane (penetration cmhancement) 2. compounds move through the cells). Therefore. have been used to improve transport of poorly absorbed drugs. The mechanism behind the increased permeability observed for various compounds after the coadministration of salicylates is thought to be an opening of the tight junctions. The aspecific nature of this opening of the tight junction could possibly lead to severe side effects when applied in vivo. routes 2. (3) passive diffusion. XES TO ENHANCE DRUG 2. Furthermore. In general. including compounds such as surfactants. (5) carrier-mediated route. morphological studies indicate that salicylates cause epithelial cell damage and widen both the intercellular junctional spaces and the pores of the epithelial cell (1). three general stra tegies can be employed to increase the transpoirt of a solute across an epithelial membrane: 1. whereas route 1 is considere!d a paracellular pathway (i.meability by: . Currently. (1)Tight junctional pathway. but it can also become more permeable for toxic xenobiotics or even antigens. chelating agents (e. salicylates. (2) drug efflux pathway (e. ' Generally.. Oral lipidI-based formulations 3. ~bsorption enhancement either change the I~ermeability properties of the epithelial cells Ior alter the physicochemical properties of the compound itself.g. the increased absorption results from either disrupting the tight junctions or altering the membrane fluidity or both. EDTA).e. Design o.1.lipophiilization . or shortchained fatty acids.g.€ prodrugs with increased membrane per. a compound moves between the cells). Although the basement membrane and subepithelial structures are not damaged.e. it is not yet completely understood whether this tissue damage is reversible. Routes and mechanisms of solute transport across epithelial membranes.. the latter leading to severe immunological side reactions. The intestinal barrier becomes permeable not only for the drugs studied. bile salts.. The permeability of drugs that are poorly absorbed beca use of their hydrophilic character can be influ1 enced by manipulating either the drug or the Imembrane. . Permeation-enhancing agents.targetin g to a carrier-mediated transport SJ&em Changing: the permeability of the memenhancers is an aspecific brane by per~etration approach thist has met limited success in the pment process and is discussed drug develo: only briefly Ilere. most strategies for .5 are transcellular pathways (i.2 Strategies t c Enhance Drug Permeability Routes and mechanisms for the transport o f drugs across epithelial barriers 1 2 3 4 Apical Basolateral I I I Figure 81.. Most agents enhance uptake of drugs by compromising the integrity of the cell membrane. (4) receptormediated endocytosis andlor transcytosis pathways.1 Penetration Enhancers TY .

Sandimmun/Neoral (cyclosporin microemulsion). and limited to intestinal tissue. ZOT has been shown to increase the absorption of insulin by 10-fold in rabbit ileum and jejunum without affecting colonic absorption (2). bile salts. interfacial transfer into the membrane. for example. this approach has found rather limited clinical application because of the nonspecific behavior of many tight junctional perturbants. the interaction of lipid-based formulations with the gastrointestinal system and associated digestive processes is not completely understood and appears to be more complex than mere solubility enhancement.Membrane Transport Proteins and Drug Transport Notwithstanding its obvious disadvantages.g. This controlled permeation enhancement would be preferred to the nonspecific disruptions caused by fatty acids. In diabetic animals. and chelators. in that ZOT may be used to enhance the intestinal absorption of proteins by safely modulating the paracellular pathway. which infects the intestinal tract and causes severe diarrhea. however. 2. those treated with ZOT and insulin orally showed comparable survival and decreases in blood glucose to those diabetic animals treated with insulin parenterally. Despite a surge in research in the late 1980s and early 1990s.2 Lipid-Based Oral Drug Delivery In recent years there has been an increased interest in the utility of lipid-based delivery systems to enhance oral bioavailability (4). In an effort to improve this solubility-limited bioavailabiliy. this regulation of the paracellular pathway has been shown to be a safe. distribution coefficient D. formulators have turned to the use of lipid excipients to solubilize the compounds before oral administration. and diffusion across the second stagnant aqueous layer. Vibrio cholera. Furthermore. It is now generally accepted that the overall transfer rate across the membrane increases initially with . the systemic availability of highly lipophilic drugs is impeded by their low aqueous solubility. It is generally known that membrane permeability is directly correlated to a drug's waterlipid partition coefficient. it is well known that lipids are capable of enhancing lymphatic transport of hydrophobic drugs. For example. which takes into account partitioning with respect to aqueous pH).3 Prodrugs The dominant factor governing passive drug transport is the lipophilicity of the compound. reversible. generally described by the oillwater partition coefficient (P) or related parameters (e. 2. Another promising class of permeation enhancers is the biocompatible polymer chitosan and its analogs (31. interfacial transfer out of the membrane. including insulin. From a mechanistic point of view. an increasing body of evidence has shown that certain lipid excipients can inhibit both presystemic drug metabolism and intestinal drug efflux mediated by P-glycoprotein (PGP). thereby reducing d~%g clearance resulting from hepatic first-pass metabolism. which is able to increase the permeability of tight junctions. passage across the membrane.and dose-dependentstrategy. produces a protein known as the zonula occludens toxin (ZOT). ZOT in combination with other targeted delivery techniques such as nanosystems could become a highly versatile approach. Interestingly. and Fortovase (saquinavir). As a result. Transport across a lipid bilayer membrane involves a number of steps: diffusion across a stagnant aqueous boundary layer. ZOT protein seems to be effective only at receptors in the jejunum and ileum but not the colon. Norvir (ritonavir). which were shown to substantially increase the bioavailability of several macromolecules across the intestinal epithelium. As more mechanistic studies emerge we can expect more extensive application of this flexible oral drug delivery approach. Furthermore. the opening of tight junctions between enterocytes by chemical modifiers is still a promising technique to increase epithelial permeability of macromolecules. ZOT specifically targets the actin filaments associated with the tight junction without compromising the overall intestinal integrity or function.. A greater understanding of the factors that govern junction control has led to the discovery and development of more specific and potent permeation enhancers. and time. Several formulations are currently on the market. In vivo. These early findings are very promising.

This low bioavailability can be mainly attributed to poor permeability characteristics. who used the wordprodrug orproagent to describe compounds that undergo biotransformation before exhibiting their pharmacological effects. the parent drugs are then released from their lipophilic derivatives by hydrolysis or specific enzymatic actions.1 Definition of Prodrugs.2) in the treatment of psoriasis and neoplastic diseases was impractical because of its poor oral bioavailability. however. not all poorly absorbed drugs can be subjected to structural modifications. a prodrug is synthesized from the parent compound by converting its hydrophilic residues into less polar moieties.3. the vast majority of prodrugs are designed to increase the intestinal absorption of that is. Although many drugs are known to be inactive until biotransformed. ministration (6). 3 ' .2 Prodrug Design by Increasing Lipophi- licity. Upon hydrolysis. especially for compounds with relatively small molecular weight. there are a large number of highly water soluble compounds whose transfer across the intestinal membrane is limited as a result of extremely low P values. Nowadays. nonli- the term prodrug was first introduced by Albert (5).2 Strategies to Enhance Drug Permeability CH20R H vrNH R=-H 6-azauridine (6-AZA) 2'. 8. 2. Today. and biotransformation associated with certain drug molecules. However. to increase lipophilicity. whose structures are closely related to their biological activities. 5'-tribenzoyl-6-AZA Figure 8.2. but eventually reaches a maximum as diffusion across the aqueous boundary layer becomes a rate-limiting step (see below). The prodrug concept has been most successfully applied to: facilitate absorption and distribution of drugs with poor lipid solubility stabilize against metabolism during oral absorption increase the duration of action of drugs that are rapidly eliminated overcome problems of poor product acceptance by patients eliminate stability and other formulation' problems promote site-specific delivery of a drug increase the aqueous solubility lower the toxicity of a drug In fact. 3 ' . rational prodrug design is widely used to overcome problems of absorption. polar drugs2. their utility as prodrugs was based on serendipity.3. Structural formulas of 6-azauridineand its prodrug analogs. This strategy has successfully improved absorption of numerous drugs. . 5'-triacetyl-6-AZA O R OR R = -COCH3 R = . Consequently. he suggested that this concept could be used for many different purposes. distribution. For example. lipophilicity.COC6H5 2 ' . In general. An example of highly polar. The regular use of the nucleoside analog 6-azauridine (Fig. An approach to increase epithelial permeation of these compounds is to alter the lipophilicity of drugs through chemical modifications. the aggregated parent molecule may no longer be biologically active. During or after absorption. Whereas many nutrients and drugs are readily transported across the intestinal membrane. Historically. macromolecules such as polypeptides and proteins. can potentially aggregate on chemical modification. the term "prodrug" is used to describe a compound that is converted to the pharmacologically active substance after ad- pophilic molecules with resulting poor permeability characteristics and therefore low bioavailability is the structural analogs of the natural purine and pyrimidine nucleosides.

and polyacyl derivatives. the triacetyl ester is converted for 80% into 6-AZA and 17% as its 5'-monoacetyl derivative. was carried out in an effort to obtain orally bioavailable 6-MA. On oral dosing. Nowadays. 2.3 Rationale and Considerations for the Use of Prodrugs. Evidently. Psicofuranine is not absorbed upon oral administration in humans. and temperature-dependenttransport mechanisms. cholera toxin) appears to be highly efficient. short to medium length alkyl esters and many inorganic and organic acid and base salt combinations. In the rational design and palmitoyl.3'. absorption is enhanced by formation of conjugates between .g.72 is well absorbed. As discussed below. Not only the newly created prodrug but also the derivatives formed after bioconversion should be nontoxic. propyl. Prodrugs should be easily synthesized and purified. For pharmaceutical scientists.5'-triacetyl-6-MA caused the same clinical effects as an equivalent intravenously administered dose of 6-AZA (7). Prodrugs must be stable in bulk form and dosage form. Orally administered 2'. lipophilization seems straightforward enough to create orally absorbable prodrugs of many polar substances. Apart from naturally occurring substrates. these membrane transporters provide alternative routes for the delivery of drugs that would normally be impermeable to the biological barriers.3'. Numerous other examples of successful ester prodrug design are presented by Yalkowski and Morozowich (8).. The effective transcellular movement of these molecules is facilitated by specialized transport processes in the epithelia. These approaches can be categorized into methods that manipulate the membrane barrier properties or increase drug solubility. Because of the low permeation at both low and high log P values. the ester has a log P value about 2. growth factors). and ethyl ester groups are frequently used in prodrug design. concentration-.g.Membrane Transport Proteins and Drug Transport The synthesis of various ester prodrugs of 6-azauridine (6-AZA). carrier-mediated transport and receptor-mediated endocytosis/transcytosis (discussed in more detail in subsequent sections). -1 < log P < 2) has to be considered in prodrug design. hydrophilic compounds and pharmaceutical macromolecules encounter difficultiesin permeating the cell membrane. 2. they also differ significantly in ways that transporter proteins are anchored in the membrane and in their ligand internalization mechanisms. Neither the prodrug nor its metabolic derivatives should be toxic.5'-tribenzoyl-6-AZA. Relatively safe moieties include amino acids. pivcephalexin). and toxins (e.. Multistep syntheses increase operator time and decrease yield. the triacetate ester log P = 0.3. 2'. energy-. Elaborate synthetic schemes should be avoided because of increased costs. such as 2'.. the lipophilicity of a drug cannot be increased indefinitely to improve epithelial absorption. that is. sugars.95. whereas ampicillin shows about 33%. in pivampicillin. endogenous proteins (e. On the other hand. having a log P of -1. several factors should also be considered. shows about 90% bioavailability upon oral administration in humans.g. Both processes are operated by specific membrane-associated proteins and share common features of active-that is. and subject to structural analog inhibition. as well as other mono. the systemic absorption of many water-soluble nutrients (e. It was shown that the triacetyl ester can be administered every 8 h and is absorbed completely. insulin.g.7 log units higher than that of the parent compound.. Whereas charged. Several strategies have been developed to enhance the bioavailability of drugs with limited membrane permeability. The pivaloyl ester prodrug of the antibiotic drug ampicillin. it is now well recognized that many drugs can be selectively taken up by active transport processes. an optimal lipophilicity (e.4 Absorption Enhancement by Targeting to Membrane Transporters synthesis of the ideal prodrug derivative. ricin.g. 1. pivaloyl (e.3'. pivampicillin. Using a method similar to the conventional prodrug approach.5'-triacetyl6-AZA. 3. vitamins). 2..

Knipp and coworkers (12) used a group of structurally unrelated compounds of known hydrodynamic radii and estimated the pore radius of human Caco-2 cells to be 5. Furthermore. and human (14) show a significant molecular mass cutoff at about 600 Da. scientists are now starting to design prodrugs based on the structural requirements of the transporter systems. Using a series of polyethyleneglycol (PEG) beads of known radii that span that of the restrictive paracellular pore. drug candidates can be shuttled across or into the cells and eventually be released from the ligands.2 A.g. In absorptive epithelia such as enterocytes. formulating pharmaceuticals in drug vehicle systems compensates the limited capacity of RME systems. recent success in transport of RME ligand-drug vehicle conjugates (e. Yet. The overall surface area available for transcellular transport is significantly greater than that of the paracellular pathway. 3 PASSIVE D I F F U S I O N In general. Ma and colleagues (11)proposed that the paracellular pore in rat colon was accessible to molecules with a radius >11 A on the basis of significant permeation of inulin.. Taking advantages of recent advances in molecular biology and computer modeling. In general. this type of conjugation avoids direct chemical reaction between drug molecules and ligands. Consequently. Madara and Dharmsathaphorn (10) suggested a pore size for human colonic T84 cells in the range 3. In vivo PEG absorption profiles in rat. probably because larger conjugates fail to be shuttled through the restricted space within the carrier protein.3 A for Caco-2 and T84 cell monolayers. drug vehicle systems also protect drug molecules from possible enzymatic degradation in the biological membrane. Lipophilic and small amphiphilic compounds can traverse the epithelium efficiently by partitioning into and out of the lipid bilay- . prodrug strategies involving carrier-mediated pathways have the advantage of high uptake capacity. they not only maintain cell polarity by confining surface proteins to their appropriate membrane domains but also prevent diffusion of water-soluble molecules and backflow of the absorbed nutrients. corresponding to a hydrodynamic radius of about 5. First of all. although the fact that all of the probes used were smaller than the estimated pore size was an acknowledged limitation. mannitol and inulin. Studies in rat small intestine have shown that only water and small hydrophilic solutes with molecular radius smaller than 25 A can move across the paracellular pathway (9). nanoparticles. through specific interaction of ligands between the moiety and its transporter. resulting in lo3-to lo6-foldincrease in uptake.Pas! drugs and the endogenous ligands of the membrane transporters. its functional dimensions have remained poorly defined. the intercellular space is sealed by tight junctions or zonula occludens (ZO). allowing incorporation of drugs with more diverse structural properties. which in some cases is insufficient to elicit pharmacological activities. Compared to active carriers. the transcellular pathway is n aturally the preferred route of transport for most molecules. Similarly. However. More important. RME pathways are perfectly suited to accommodate large molecular weight peptide and/or protein conjugates. Second.5 and 4. the size of drug conjugates is relatively limited (-1000 Da). For peptide and protein delivery.6-15 A based on the ability of two probes of widely different size. Despite intense interest in the structure and regulation of the human intestinal tight junction.3 4 which compares well with data obtained in cell culture. respectively. receptor-mediated endocytosis (RME) systems have a rather limited uptake capacity. Watson and colleagues (13) calculated a pore radius of 4. carrier-mediated pathways could facilitate peptides only up to four amino acids. to cross T84 monolayers. because of the endocytic pit formation (up to several hundred nanometers) and the vesicular internalization mechanism. liposomes) by way of RME pathways opens new possibilities for macromolecular delivery across biological barriers. dog. The junctional proteins ZO-1 and 20-2 play essential roles in epithelial barrier function. Therefore. drug transport across any epithelium is dictated by the characteristics of the cell membrane and the physicochemical prop- erties of drugs.

Analyzing the physicochemical properties and permeability characteristics of several thousand drug molecules. Important boundary conditions to the first law are (1)steady state (i. homogeneously mixed compartments on both sides of the barrier). is proportional to the concentration gradient dC/dx: where D is the diffusion coefficient of the solute in cm2/s. 3. Fick's first law may be written as The flux.e. In constrast. The distribution or partition coefficient K is expressed by . respectively. In experimental situations. Cis its concentration in mol/cm3. dC/dt = 0). a barrier (e. z ) in the general form Passive diffusion refers to movement of a solute along its concentration gradient. Factors influencing the transcellular passive diffusion of drugs have been thoroughly characterized. on the donor side or C. The diffusion constant D or diffusivity does not ordinarily remain constant and may change at where (C. concentration gradient) favors such action..1 Kinetics of Passive Diffusion higher concentrations.. epithelial tissue) usually separates two . and is affected by temperature. even when thermodynamic conditions (e..e. and C.2 is known commonly as Fick's first law. i.256 Membrane Transport Proteins and Drug Transport ers. a constant rate of diffusion) and (2)sink conditions (i. If the concentrations in the membrane on the donor and receptor sides are C.ompartments of a diffusion cell of cross-sectional area S and thickness h. As long as the diffusing molecule does not interact with elements of the membrane. Lipinski (15) deduced that only compounds with a molecular weight lower than 500. pressure. on the receiver side. . Fick's second law applies: which is usually differentiated to express changes in concentration in three dimensions (x. y. large hydrophilic molecules cannot diffuse freely through the cells. The change in 'mass (M) of a solute as a function of time (t) during its diffusion through a membrane barrier with area S is known as the flux J: The reader should appreciate that this form of Fick's law does not have to be taken into consideration if the aforementioned boundary conditions are met (steady state..e. Equation 8. solvent properties.. the driving force behind the diffusion of a molecule through the lipid bilayer is the electrochemical potential difference of the compound on both sides of the membrane.C. This set of characteristics for well-permeating molecules is now popularly known as Lipinski's "rule of five" and has served in the drug industry as an extremely helpful screening mechanism for recognizing drug permeability issues early on in the drug discovery process (15)..g. and less than 5 hydrogen bond donors and 10 hydrogen bond acceptors are likely to permeate efficiently across the cell membrane by passive diffusion. and C2 within the membrane generally are not known but can be replaced by the partition coefficient K multiplied by the concentration C. in turn. Without these boundary conditions.)/h is an approximation of dC/ dx. a logP less than 5. and x is the distance in cm of movement perpendicular to the surface of the barrier.g. The concentrations C. and the chemical nature of the solute.

at high P values (log P > 3). 4 FACILITATED AND ACTIVE TRANSPORT PATHWAYS 4.02 0 I I I I Figure 8. 8.and receptor-mediated systems should be pointed out. The curve is described by log .08 - C .1 0. k = 0. Because of the low intestinal permeation at both low and high P values. The rate of absorption.09(log PI2 . Empirical relationship between i n situ intestinal absorption rate constants (k. this.0.12 r.833 (16).g. an optimal P value (e. a decrease in water solubility takes place that would make the compound highly insoluble.0.P can be assessed in experimental systems and provides a relative parameter to classify solute penetration through a lipid bilayer. in turn. F 0. 0. and h cannot always be determined independently and commonly are lumped together to provide the permeability coefficient P with units of linear velocity (cmls). Conversely.U) n I 0. It is common to express the partition coefficient in terms of log P.8 demonstrates that the absorption rate of solutes through biological membranes is directly proportional to the value of the oillwater partition coefficient. under physiological conditions. The reader should be aware that P in this case is used to denote the partition coefficient and not permeability.Cr ) (8.16 0. At low P values (log P < -2).3). - J= DKCd = P. K.) -2 -1 0 1 -3 2 3 log P (octanol/buffer) and apparent partition coefficients.04 0.1031og P . would prevent the compound from reaching the membrane surface.4 Facilitated and Active Transport Pathways 0.8) The variables D.7) In general. Carrier-mediated systems involve trans- . C. (16) studied the absorption of a series of carbarnate esters through rat everted intes- tine and observed a bell-shaped relationship between the absorption rate and the partition coefficient (Fig. Houston et al. Equation 8.14 0. sink conditions will apply in the receptor compartment. and the pH at the absorption site are interrelated. J=DK( Cd . the compound becomes so lipid soluble that the diffusion through the unstirred water layers flanking both sides of the membrane becomes the rate-limiting step in the overall absorption process.. Thus. the pK. the compound cannot penetrate the lipid membrane because of an excessive thermodynamic barrier. of the diffusing solute.Cd h (8..1 Receptor-Mediated Transport A distinct difference between carrier. Moreover. -1 < logP < 2) has to be considered in drug design.06 0.3. Numerous studies support the idea that drug molecules are absorbed through lipid bilayer membranes in the un-ionized state by the process of passive diffusion.

RME is a highly specific cellular biologic process by which. The endocytotic recycling pathways of polarized epithelial cells (Fig. but the interested student may consult alternative reviews covering the complete spectrum of endocytotic processes in other cell types (17. Coated pits are areas of the membrane with high concentrations of endocellular clathrin subunits.4 presents an overview of the existing possibilities in the fate of ligands and receptors. This process ensures that the correct receptors are concentrated in the coated pit areas and minimizes the amount of extracellular fluid that is taken up in the cell. Alternatively.5 h (19). the ligand is directed to lysosomes for destruction. and potocytosis [nonclathrin (caveo1in)-mediated RME]. Specialized coat proteins. these processes of membrane deformations are termed "endocytosis. Endosomes may then move randomly or by saltatory motion along the microtubules (24) until they reach the trans-Golgi reticulum. trap specific membrane receptors (which move laterally through the membrane) in the coated pit by binding to a signal sequence (Tyr-XArg-Phe. Collectively. where the ligand is released into the extracellular milieu and the receptor is recycled. exogenous ligands bind to specific externally oriented membrane receptors. The fate of the receptor and ligand are determined in these sorting vesicles.. transcobalamin) (23). receptor-mediated endocytosis (clathrin-mediated). as its name implies. which causes dissociation of the receptor-ligand complex (22).g." or CURL (compartment of uncoupling receptor and ligand) (20)l. Localized membrane proteins. are generally more complex than those in nonpolarized cells. where X = any amino acid) at the endocellular cai-boxy terminus of the receptor.51. In many cells the process of endocytosis is so active that the entire membrane surface is internalized and replaced in less than 0. Figure 8. as explained in more detail below. Carriers operate by shuttling their substrates across the membrane by means of an energy-dependent (ATP or cotransport) flip-flop mechanism and receptors are internalized in vesicles after binding to their substrate.. which are actually a multisubunit complex.Membrane Transport Proteins and Drug Transport port proteins that are anchored to the membrane by multiple membrane-spanning fragments or protein loops. Initially. After the internalization process. "receptosome. 4. Some ligands and receptors are returned to the cell surface. the ligand-receptor complex accumulates in coated pits. where they are believed to fuse with Golgi components or other membranous compartments and convert into tubulovesicular complexes and late endosomes or multivesicular bodies. 18).5. whereas the receptor is recycled to the cell membrane. but are presumably similar to . various ligands bind to cell surface receptors and are subsequently internalized and trafficked within the cell. The signals required for this sorting step have not been defined as of yet. Mam- malian cells have developed an assortment of mechanisms to facilitate the internalization of specific substrates and target these to defined locations inside the cytoplasm. lipids. the clathrin coat is lost through the help of chaperone proteins and proton pumps lower the endosomal pH to approximately 5. whereas receptormediated systems use receptor proteins that span the membrane only once." consisting of phagocytosis. RME appears to require the GTP-binding protein dynamin. and extracellular solutes are also internalized during this process. The assembly of clathrin molecules on the coated pit is believed to aid the invagination process. pinocytosis. The emphasis of this section is receptor-mediated endocytosis in the intestinal tract. RME can be dissected into several distinct events. but the process by which dynamin is recruited to clathrin-coated pits remains unclear (21). When the ligand binds to its specific receptor. transferrin) from the receptor involved in transcytosis (e. In these enterocytes there is a common recycling compartment (CRC)that receives molecules from both apical and basolateral membranes and is able to correctly return them to the appropriate membrane or membrane recycling compartment (ARC or BLRC) (25).g. such as enterocytes. called adaptins. 8. Binding occurs within 2 min and is followed by membrane invagination until an internal vesicle forms within the cell [the early endosome.CURL serves as a compartment to segregate the recycling receptor (e.1 Receptor-Mediated Endocytosis.1.

ligand degraded L = low-density lipoprotein Receptor recycles. ligand. such as an extracellular ligand binding site. a fungal metabolite that has profound effects on the structure and function of the Golgi apparatus. 28). recycling or transcytosis of either molecule (see text). ligand recycles ligand transported ligand released L = transferrin L = IgA L = viruses. it strongly influences the eventual endocytotic mechanism (26). Receptor transported. One of the least under- Our general understanding of RME receptor structure and related structure-function relationships has been significantly enhanced by ongoing efforts to clone mRNA sequences coding for endocytotic receptors. that are particularly found on transcytotic vesicles and are believed to be required for fusion with target membrane (27. The initial binding and uptake steps [including receptor clustering in coated or noncoated pits. ligand and receptors are sorted and trafficked independently. BFA enhanced the TfR-mediated transcytosis . and possibly one of the most important aspects for successful oral drug delivery by RME.1. lysosomes are depicted as shaded circles. a single hydrophobic transmembrane domain (unless the receptor is expressed as a dimer). It appears that most RME receptors share several structural features. Shah and coworkers (29) showed in Caco-2 cell monolayers that BFA causes a marked decrease in the number of basolateral Tf receptors (TfR) along with a slight increase in the number of apical TfR. 10. 4. receptor. R. Intracellular sorting pathways of RME.4.1. After entry into acidic endosomes.) the peptide sequences required for proper sorting in the trans-Golgi network. Although protein orientation may appear trivial. (Adapted from Ref.3 Transcytosis. stood aspects in vesicular trafficking and sorting. toxins Figure 8. more commonly referred to as transcytosis. Other methods to stimulate transcytosis have been recently explored. internalization of the receptor-ligand complex into coated vesicles (noncoated in the case of potocytosis). and fusion of vesicles to form endosomes] are common to all pathways. L. 4. Recent studies in the area of cellular biology have reported specific proteins.4 Facilitated and Active Transport Pathways 1 Receptor recycles. Two classes of receptors are proposed based on their orientation in the cell membrane: the amino terminus of Type I receptors is located on the extracellular side of the membrane.2 Structure of Cell Surface Receptors. Receptor recycles. Transcytosis of transferrin (Tf) was found to be stimulated in the presence of Brefeldin A (BFA). whereas Type I1 receptors have this same protein tail in the intracellular milieu. named TAPS (transcytosisassociated protein). which may result in degradation. and a cytoplasmic tail encoding endocytosis and other functional signals (26). is the transport of endocytotic vesicles to the opposite membrane surface.

or direct exocytosis of the M E to apical membrane.Membrane Transport Proteins and Drug Transport Apical CP Basolateral Figure 8. BLSE. Question marks along an arrow indicate that only circumstantial evidence exists for those pathways a t the present time. Potocytosis. Double-headed arrows indicate that similar magnitudes of transfer occur in either direction: Ligand and receptor complex is initially taken up from coated pits (CP) into coated vesicles (CV). . a recycling pathway to the central recycling compartment (CRC) andlor the apical recycling compartment (ARC). BLRC. of both 1251-Tf and the horseradish peroxidase-Tf conjugate across Caco-2 cells in both apical-to-basolateral and basolateral-to-apical directions.5. basolateral recycling compartment. similar pathways occur a t the basolateral membrane. cell nucleus.1. N.4 Potocytosis. Golgi apparatus. which eventually exocytose the contents of the vesicle. whereas basolateral endocytosis was unaffected (30). tight junction. which are uniform omegaor flask-shaped membrane invaginations (5080 nm diameter) (17) and was first described as the internalization mechanism of the vitamin folic acid in a mouse keratinocyte cell line (34).Years before the term "potocytosis" was coined.The CV loses its clathrin coat as described in the text and transforms into an apical sorting endosome (ME). TJ. basolateral sorting endosome. which fuse with lysosomes (LY). including cholera and tetanus toxins (35). In polarized cells. in combination with the trans-Golgi network (TGN). or non-clathrin-coated endocytosis takes place through caveolae. interconnects the apical and basolateral sorting pathways and enables the potential for transcytosis. The CRC. Prydz and colleagues found that BFA treatment rapidly increased apical endocytosis of both ricin and HRP in MDCK cells. G. 4. Schematic representation of endocytotic pathways in polarized cells. The fate of receptor and ligand can be: a degradative pathway where the material is passaged to multivesicular bodies (MVB)and eventually late endosomes (LEI. various ligands had been reported to localize in non-clathrin-coated membrane regions. It was not until recently that potocytosis has been accepted as a dis- tinct RME pathway (31-33).

The Fc receptor resembles the major histocompatibility complex (MHC)class I antigens. Hussain and colleagues (40) deduced from their own work and the work of others that "the importance of Peyer's patches as the principal site of particulate absorption may have been overemphasized. Interestingly." In this light. Soluble macromolecules. the advantage of potocytosis pathways over clathrincoated RME pathways lies in the absence of the pH-lowering step.2. From a drug delivery standpoint. suggesting that the ligand causes redistribution of receptors at this site. (3) interactions with the actin-based cytoskeleton. in addition to the regulation of IgG catabolism (43.371.This may be of invaluable importance to the effective delivery of pH-sensitive macromolecules. to absorb particulate matter. and that normal epithelial cells can also be induced. In mature absorptive cells both subunits are colocalized in each of the membrane compartments that mediate transcytosis of IgG. it should be stressed that the surface area of M-cells is only 10% compared to the surface area of normal epithelial cells. (2)the uptake of small molecules by potocytosis involving GPIlinked receptor molecules and an unknown anion transport protein.2. Mcells are specialized epithelial cells of the gutassociated lymphoid tissues (GALT) that transport antigens from the lumen to cells of the immune system. in that it consists of two subunits: a transmembrane glycoprotein (gp50) in association with p2-microglobulin (41). Site-directed mutagenesis of a recombinant Fc hinge fragment has been used to localize the site of the mouse IgGl (mIgG1) molecule that is involved in the intestinal transfer of recombinant Fc hinge fragments in neonatal mice. whereas a degradative pathway dominates at higher mucosal IgG concentrations. These results support a model for transport in which IgG is transferred across the cell as a complex with both subunits. with appropriate ligands such as plant lectins and bacterial adhesins. The importance of M-cells in the uptake of particles is still a point of discussion. given that unbound proteins are removed from the transport pathway before exocytosis. and (4) the compartmentalization of certain signaling molecules involved in signal transduction.1. 4. IgG in milk selectively binds to neonatal Fc receptors (FcRn) expressed on the surface of the proximal small intestinal enterocytes during the first 3 weeks after birth.5) of the jejunum and release at the pH of plasma (7. including G-protein-coupled receptors. In rats. Accord- 4.2 Receptor-Mediated Oral Absorption Systems RME in Enterocytes versus M-Cells. which coats the cytoplasmic surface of the membrane (36. FcRn. with binding occurring at the luminal pH (6-6. Caveolae are characterized by the presence of an integral 22-kDa membrane protein termed VIP21caveolin. Receptor-mediated transcytosis of im- ing to Walker and Sanderson (38) the preferred route of intestinal uptake of low concentrations of antigens is through the M-cells (microfold or membranous). Recently.1 Maternal and Neonatal IgG Transport.1 immunoglobulin Transport 4. Benlounes and coworkers (42) recently showed that IgG is effectively transcytosed at lower concentrations (<300 pg/mL). although at higher concentrations the regular enterocytes are also involved. munoglobulin G (IgG) across the neonatal small intestine serves to convey passive immunity to many newborn mammals (41). is involved in transcytosis across both yolk sac and neonatal intestine. FcRn binds IgG in a pH-dependent manner. These results favor the proposal that IgG is transferred across cells as an IgG-receptor complex (45). IgG administered in situ apparently causes both subunits to concentrate within endocytic pits of the apical plasma membrane.4).and also entire microorganisms are transported by M-cells. thereby circumventing the classical endosomal~lysosomal pathway (32). Drug carriers such as liposomes are not readily transported intact across epithelial barriers. These studies definitively indicate that the neonatal Fc receptor. Pate1and Wild (46) showed that coat- . Continued binding to vesicle membranes appears to be required for successful transfer. thereby initiating an immune response or tolerance.4 Facilitated and Active Transport Pathways Morphological studies have implicated caveolae in (1)the transcytosis of macromolecules across endothelial cells. small particles (39). 44). located in Peyer's patches.

IgG is transcytosed from the apical to basolateral surface in several epithelial tissues such as the placenta and the small intestine of newborn rats. but not F41 and 987P adhesins (55). For endocytosis there are two signals. 1993. These results suggested that transported liposomes followed a pathway of transcytosis in clathrin-coated vesicles. such as a mannose-specific adhesin in Lactobacillus plantarum (52) and V. protein kinase C.2. and calmodulin. The internalization process of many microorganisms is impres- . This single factor is sufficient to promote entry of inert particles by binding multiple integrin receptors during cellular uptake (57).1. Livrelli and coworkers (50) found no relationship between the adhesive pattern and the production of specific fimbriae.481. depending on the environmental conditions and nature of the adhesins as well as receptors. No. This attachment is usually mediated by bacterial fimbriae or pilus structures. and by the hypoglycemic effect of entrapped insulin. which is located on the basolateral surface of the cell.Membrane Transport Proteins and Drug Transport ing liposomes with appropriate IgG enhances their transport across rabbit yolk sac endoderm and enterocytes of suckling rat gut proximal small intestine. For many bacterial species. Kg9 and K88 adhesin were detected. The receptor for intestinal transport of IgG is structurally similar to class I MHC molecules (Mostov. In contrast to other infective agents. cholerae (53).to 42-kDa glycoprotein K88-specific receptor. Mostov. both of which contain tyrosines. Using monoclonal antibodies against fimbrial i d hesins of porcine enterotoxigenic E. coli. although other cell surface components may also take part in the process. pIgR expression can be upregulated by cytokines (49). who identified E-cadherin as the ligand for internalin. In a study on the colonization mechanism of Klebsiella. The colonization mechanism of the enteropathogenic bacterium Yersinia pseudotuberculosis has been studied in great detail. No. The pIgR-IgA complex is internalized into endosomes.2. 4. Targeting from the trans-Golgi network to the basolateral surface is determined by the membrane-proximal 17 residues of this domain. adherence to host cells is the initial key step toward colonization and establishing an infectious disease. Bacteria usually ex- press various cell adherence mechanisms. Adherent bacteria colonize intestinal epithelium by multiplication and initiation of a series of biochemical reactions inside the target cell through signal transduction mechanisms (with or without the help of toxins) (51). The pIgR contains a cytoplasmic domain of 103 amino acids that contains several sorting signals. including bile and saliva (47. a Listeria monocytogenes protein essential for entry into epithelial cells.2 Polymeric IgA and IgM Transport. polymeric IgA and IgM bind to the polymeric immunoglobulin receptor (pIgR). invasin. where it is sorted into vesicles that transcytose it to the apical surface. and the large extracellular fragment (known as secretory component) is released together with the ligand. Enterobacter. At the apical surface the pIgR is proteolytically cleaved. Several adhesin and invasin molecules have been identified. 2734). Metcalfe and coworkers (54) found that adherence of Escherichia coli K-12(K88ab) to immobilized porcine small intestine mucus was caused by a 40. suggesting that several unrecognized adhesive factors are involved that remain to be identified. 4. pseudotuberculosis is mediated by a single 986 amino acid protein. 1994. This phenomenon has also recently been found by Mengaud and coworkers. coli (EIEC). invasion and transcytosis of Y. They measured the effect of liposomal transcytosis both by radiolabel assay of entrapped [1251]PVP and [3H]inulin. Polymeric IgA is produced by plasma cells and found in all external excretions. Transcytosis of the pIgR is signaled by serine phosphorylation and may be regulated by the heterotrimeric Gs protein. such as Salmonella strains or enteroinvasive E. thus escaping lysosomal degradation. Bacteria causing gastrointestinal infection need to penetrate the mucus layer before attaching themselves to the epithelial surface.2 Bacterial Adhesins and invasins. 2735. Two components are necessary for the adherence process: a bacterial "adhesin" (adherence or colonization factor) and a "receptor" on the host (eukaryotic) cell surface. In the small intestine. on the bacterial surface that binds to a501 integrin (56). and Serratia strains.

Oral immunizations with 10' pfu (plaque-forming units) of W-HA generated low numbers of HA-specific IgA-producing cells in the lamina propria of the gut. therefore. Staphylococcus aureus produces a set of proteins [e. The study of bacterial adhesins and invasins for the application in drug delivery strategies has recently become the focus of much attention. coli heat-labile toxin binds to the brush border of intestinal epithelial cells in a highly specific. beans. pseudotuberculosis into the lumen of BALBIC mice. that modified and. Fisher and coworkers expressed the transmembrane domain of di~htheria toxin in E. ricin and saporin (70). . The latter group found that f latex microspheres up to 1 pm coupled to mal: tose-binding protein.5 Lectins (Phytohemagglutinins). After reaching early endosomes by RME.3 Bacterial and Plant Toxins. which act both as superantigens and toxins. Etchart and colleagues (79) compared the immune response to a vaccinia virus recombinant..Semliki Forest virus (75). have been identified that bind to any mammalian cell surface expressing galactose units and are subsequently internalized by RME (67). Uptake of this toxin and transcytosis to the basolateral side of the enterocytes was observed both in vivo (63) and in vitro (64). Toxins such as nigrin b (681. wildtype bacteria can be found in the Peyer's patch (58).A can induce measles-specific immunity in the intestine.2. coli as a maltose-binding fusion protein and coupled it chemically to high molecular weight poly-L-lysine. they are abundant in the human diet (e. adenoviruses (76). which was fused with invasin. expressing the measles virus hemagglutinin (W-HA). are rapidly internalized). 61) t used a similar approach for intestinal delivery j of nanoparticles. tins are plant proteins that bind to specific sugars. A large pool enters the degradative pathway. They extended their studies in mice in vivo by showing that ingested SEB appears in the blood more efficiently than does SEA. whereas a few molecules become cytotoxic by translocating their catalytic fragment A (DTA) into the cytosol(62). 61).. The possibility exists. The resulting complex was successfully used to mediate the internalization of a reporter gene in vitro (65).2. and reovirus (78). lectinlike manner. after p a r e n t e d or mucosal immunizations in mice.and modeccin (72) are highly toxic upon oral administration (i. These binding proteins have been identified for most viruses. Such binding may result in specific hemagglutinating activity. SEB. The 4. including rotaviruses ( 7 3 . initial step in many viral infections is the binding of surface proteins (hemagglutinins) to mucosal cells. whereas oral coimmunization with W-HA and cholera toxin greatly enhanced the level of HA-specific spot-forming cells (IgA > IgG).Mutants expressing defective invasin derivatives were unable to promote efficient translocation into the Peyer's patch and instead colonized on the luminal surface of the intestinal epithelium.4 Viral 4. 4.e. a-sarcin (691.viscumin (71). intrajejunal immunizations with 10' pfu W-HA alone induced high levels of anti-HA IgG-producing cells in the spleen and anti-HA IgA-secreting cells in the lamina propria of the gut. varicella zoster virus (74). can be internalized by MDCK cell monolayers (60. Recently. staphylococcal enterotoxin A (SEA).. which are found on the surface of glycoproteins and glycolipids of eukaryotic cells. The B subunit of the E. diphtheria toxin (DT) molecules have two possible fates. potato leafroll virus (77). Hamad and coworkers (66) found dose-dependent. most important. and other seeds). cereals. This study shows that W . provided that it is protected from degradation in the gastrointestinal tract. Paul and colleagues used an invasin fusion protein system for gene delivery strategies (59) and Easson and coworkers (60. in Caco-2 cells. but not SEA. toxic shock syndrome toxin 1 (TSST-111.g. mostly ribosome-inactivating proteins (RIPS). or that cholera toxin is used as an adjuvant. facilitated transcytosis of SEB and TSST-1. ' Various plant toxins. Because lectins are relatively heat stable. Interestingly.2. LecHemagglutinins.g. less toxic subunits of these compound can be used to facilitate the uptake of macromolecular compounds or microparticulates.4 Facilitated and Active Transport Pathways ' sively fast: it was recently shown that within 45 min after introduction of Y.

to our knowledge there are currently no reports in the literature describing the use. . whereas the p-isoform is often found in nonepithelial lineage tumors (86). Again. However. the folate receptor is principally overexpressed in human tumors.3. Two homologous isoforms (a and p) of the receptor have been identified in humans. of this system for intestinal drug delivery purposes. Thus.81). be attributable to the low expression level of the receptor in (healthy) enterocytes. Sharma and colleagues (85) studied the nature of cell-associated carbohydrates in the human intestine that may mediate transepithelial transport of bacterial and dietary lectins and their processing by the lymphoid cells of Peyer's patches. They covalently coupled polystyrene nanoparticles (500 nm) to tomato lectin and observed 23% systemic uptake after oral administration to rats.2 Riboflavin.5%. local targeting to intestinal cancer cells (e. Hussain and coworkers showed that the uptake mechanism for lectins can be used for intestinal drug targeting in vivo (40). Upon comparison of human and mouse glycoconjugates of follicle-associated epithelium and GALT. which explains why so-called high fiber foods (rich in ledins) are thought to be responsible for stimulating bowel motility (80.. or attempt. and targeting of antisense oligonucleotides (90) to a variety of cell types. Although considerable success has been met in other areas of drug targeting. it should be pointed out that the overall contribution of M-cells to the absorptive surface area of the gastrointestinal tract is minimal. they found a distinct difference in glycosylation between mouse and human Peyer's patches and their associated lymphoid cells. colon carcinoma) appears to be a fertile approach. choosing the appropriate lectin is apparently important when considering cell surface glycoconjugates as target molecules for intestinal drug delivery strategies. which could jeopardize the widespread application of drug targeting to these specialized cells. 4. 4. 84). The cellular uptake of free fo- lic acid is mediated by the folate receptor and/or the reduced folate carrier. Weaver and colleagues (82) directly infused concanavalin A. The a-isoform is found to be frequently overexpressed in epithelial tumors. indicating a 50-fold increase in oral absorption.3 RME of Vitamins and Metal Ions 4. Although lectins are generally believed to be transported by means of an RME mechanism. it was recently shown by Low and coworkers (91) that serum albumin coupled to riboflavin showed RME-mediated uptake in distal lung epithelium.3. both villous and crypt epithelial cells contained gold particles. In another study demonstrating the rapid RME uptake of lectins. gene delivery (89). In the future. knowledge of the site and species-related variations in Mcell surface glycoconjugate expression may allow lectins to be used to selectively target an- tigenic material and oral vaccines to the mucosal immune system at specific locations (84). in part. Consequently. they showed the intestinal uptake of tomato lectin-conjugated nanoparticles through the villous tissue to be 15 times higher than uptake by gut-associated lymphoid tissue (GALT) (40). but also in protein delivery (881.g. Control animals exerted a systemic uptake of <0. In this paperstudy the authors speculate that a similar uptake process exists in the small intestine.Whereas the expression of the reduced folate carrier is ubiquitously distributed in eukaryotic cells.1 Folate. conjugated with 10 nm colloidal gold particles. into the lumen of the jejunum in neonatal guinea pigs. this receptor system has been used in drug-targeting approaches to cancer cells (871. given that the a-isoform of the folate receptor is overexpressed in epithelial cell lines. Binding studies have revealed that M-cells exhibit pronounced regional and species variation in glycoconjugate expression. there is substantial evidence that these compounds have significant affinity to intestinal M-cells (83. demonstrating the rapid accessibility of crypt cells to the lectin. Within 60 min. Interestingly. Although not as exten- sively investigated as the transferrin and folate pathways. This may. The folate receptor is a glycosylphosphatidylinositol (GP1)anchored 38-kDa glycoprotein clustered in caveolae mediating cell transport by photocytosis (32).Membrane Transport Proteins and Drug Transport Concentrated solutions of lectins have a "mucotractive" effect as a result of irritation of the gut wall.

3. The Cbl-IF complex binds to an IF-receptor located on the surface of the ileum. RME system is probably the most extensively studied system for the oral delivery of peptides and proteins. In humans. According to current knowledge of intestinal iron absorption. and the LHRH antagonist ANTIDE (96). More recently. Recently. a dimeric transmembrane glycoprotein of 180 kDa (102). was transported across Caco-2 cell monolayers by RME.5 kDa).to threefold higher compared to that of control (nonspecific uptake of nanoparticles) (98. which was further potentiated by BFA pretreatment (-41%) (106). and the ligand-receptor complex is endocytosed within clathrin-coated vesicles. it is clearly not sufficient for the delivery of insulin or G-CSF..104). granulocyte colony-stimulating factor (G-CSF. and on many kinds of tumors. 991. It is clear. 100). The vitamin B. TfR. Huang and Swaan (92. Thus far. 18. the colloquial name for cobalamin (Cbl).. The role that transferrin.101) are promising and eliminate the requirement of covalently coupling cobalamin and the substrate to be delivered. they showed (both in vitro and in vivo) the intestinal uptake of biodegradable polymeric nanoparticles coupled to cobalamin to be two.3 Vitamin B. in elevated numbers on erythroblasts. Shah and Shen (105) showed that insulin covalently coupled to transferrin. a-interferon (98.Indeed. a protein released from the parietal cells in the stomach and proximal cells in the duodenum. which triggers a yet undefined endocytotic process. It was reported that IF-Cbl complex dissociates at acidic pH. the recent successes with cobalamin-conjugated nanoparticles in vivo (100. More recently. 4. they showed that oral administration of this complex to streptozotocin-induced diabetic mice significantly reduced plasma glucose levels (-28%). that Cbl is transported into all other cells only when bound to transcobalarnin 11. the universal application of this transport system for the oral delivery of peptides and proteins seems to be hampered only by its limited uptake capacity: 1 nmol per dose.. and both proteins have been isolated from intestinal enterocytes (103. the fate of the IF-Cbl complex has yet to be clarified.4 Transferrin.8 kDa). in turn. After acidification of these vesicles. 93) recently showed that riboflavin is taken up in human small intestinal and placental epithelial cells by a riboflavin-specific RME process. and Cbl is transferred to transcobalarnin I1 by Ramasamy and coworkers (941. Even though this amount of uptake may be adequate for molecules such as LHRH or erythropoietin. Russell-Jones and coworkers have shown that this particular system can be employed for the intestinal uptake of luteinizing hormone releasing factor (LHRH) analogs (97). After oral administration it is bound to intrinsic factor (IF). diferric transferrin binds to transferrin receptor (TfR).. erythropoietin (29. In most cells. uptake mechanism. the uptake of cobalamin is approximately 1 nmol per intestinal passage. an 80-kDa iron-transporting glycoprotein.is a large polar molecule that must be bound to specialized transport proteins to gain entry into cells. Transferrin. transferrin is excreted into the intestinal lumen in the form of apotransferrin and is highly stable to attacks from intestinal peptidases.. This. is efficiently taken up into cells by the process of carriermediated endocytosis. iron dissociates from the transferrin1TfFi com~lex and enters the * cytoplasm. Transferrin receptors are found on the surface of most proliferating cells. where it is bound by ferritin (Fn). with a potential for multiple dosing (2-3 times per hour) (96). After internalization. Vitamin B. and Fn have in regulating dietary iron uptake and maintaining total body iron stores is unknown. whereas Dan and Cutler (95) found evidence of free Cbl in endosomes and the basolateral side of the membrane after administration to the apical surface of Caco-2 cell monolayers. The uptake of iron in the intestinal tract amounts up to 20 mglday and is primarily mediated by transferrin. 5 TRANSPORT PROTEINS Membrane transporters in general are a large group of membrane proteins that have one (bi- . would permit the delivery of any macromolecule through the vitamin B. However..3. 4. however. Both the TfR and Fn genes can be detected on small intestinal mRNA (102).

secreting. and as a therapeutic target for hypocholesterolemic agents (108-110). Mediate uptake and efflux of ion species that must be maintained at concentrations dramatically different from those in the external milieu. most are poorly characterized at the molecular level. conferring an important field of study for pharmaceutical scientists involved in these areas. and lipids into and beyond the cytoplasmic membrane. cholesterol homeostasis. nitrogen. Transporters allow living organisms to conduct biological warfare. for example. and phosphorus. which transports di.and tri-peptide as well as numerous therapeutic compounds. 3.. Provide a means for regulation of metabolite concentrations by catalyzing the efflux of end products of metabolic pathways from organelles and cells. and toxicology). and toxicity of drug molecules. 8. which exports metabolites as well as drugs from various cell types. ABC-B1) gene. allowing genetic exchange between organisms and thereby promoting species diversification. Mediate the active extrusion of drugs and other toxic substances from either the cytoplasm or the plasma membrane. After all. antifungal agents. Mediate entry of all essential nutrients into the cytoplasmic compartment and subsequently into organelles. and a variety of other signaling molecules that allow a cell to participate in the biological experience of multicellularity. it is now well appreciated that the most critical parameter for a new drug to survive the drug development pipeline on its way to the market is its ADMET profile. sulfur. and detection of drug-transporter interactions remains unacceptably serendipitous. antibiotics. 7. 2. As a result. in-depth knowledge of membrane transport systems may be extremely useful in the design of new chemical entities (NCE). During the past decade it has become clear that a vast number of drugs share transport pathways with nutrients. Transfer of nucleic acids across cell membranes. a prototypical ABC-transporter and a product of the multidrug resistance (MDR-1. Thus. P-glycoprotein and cytochrome P450 regulation by the pregnane X receptor). distribution.266 Membrane Transport Proteins and Drug Transport topic) or more (polytopic) hydrophobic transmembrane segments.g. SLClOA2). we are unable to predictsthe interaction of drugs with this important class of membrane proteins a priori. 4. Their involvement in cellular function can be classified as follows [after Saier (10711: 1 . This section aims to give an overview of current strategies for modeling transporter systems illustrated by three well-characterized transport systems: (1)the P-glycoprotein efflux pump. . complex carbohydrates. Despite the involvement of solute transporters in fundamental cellular processes. 6. Polytopic membrane proteins are indispensable to the cellular uptake and homeostasis of many essential nutrients. SLClSAl). metabolism. as well as in their pharmacokinetic and pharmacodynamic ( P W D ) profiles. excretion. and (3) the apical sodium-dependent bile acid transporter (ASBT. Participate in the secretion of proteins. transport proteins are involved in all facets of drug ADME and ADMET (absorption. 5. (2) the small peptide transporter (PepT1. excretion. antiviral agents. Because cellular transporter expression is often regulated by nuclear orphan receptors that simultaneously regulate the translation and expression of metabolic enzymes in the cell (e. thus facilitating the metabolism of exogenous sources of carbon. Moreover. hormones. they indirectly control drug metabolism.which plays a key role in intestinal reabsorption and enterohepatic recycling of bile salts. As a result. Facilitate the uptake and release of pheromones. These transporters are involved in almost all facets of biological processes in the cell. and toxins of humans and other animals that may confer upon the organisms producing such an agent a selective advantage for survival purposes. a critical role has been recognized for transport proteins in the absorption. neurotransmitters. Many of these toxins are themselves channelforming proteins or peptides that serve a cell-disruptive transport function.

Paulsen predicts that 15% of all genes in the human genome will code for transport proteins.k B 5.e. Eventually. and common evolutionary origins (107). The ABC superfamily contains 7 families with 48 members (113). Several classes of pharmacologically active compounds share transport pathways with nutrients (117). excretion of drugs and their metabolites. the Human Gene Nomenclature Committee (HUGO) has taken on the task of directing gene classification and defining distinct subclasses in the ABC and SLC families. clearly defined phenotypes shown to be inherited as monogenic Mendelian traits) has been attributed to mutations in transport proteins. and drug pharmacokinetics and pharmacodynamics (122-124). With a current number of estimated sequence-tagged sites (STS) of 30. mediated by transporters in the kidney and liver. we can expect an additional 4500 membrane transporters to emerge.g. 112). the SLC superfamily is anticipated to consist of at least 2300 members. only a fraction (10%) has been characterized in certain detail (i. Initial studies in Xenopus leavis oocytes by Hediger and colleagues resulted in the isolation of the intestinal sodium-dependent glucose transporter SGLTl (111.2). The current status of transporter nomenclature and classification is in a similar state of disarray to that which the cytochrome P450 enzyme field found itself in during the early 1980s. H+. at this moment.drug resistance [e. Table 8... Na+.e. Furthermore. and bile acids). Paulsen and colleagues (114.ac. similar energy-coupling mechanisms. With the completion of the Human Genome Project. it can be anticipated that a vast number of membrane transport proteins will be identified without known physiological function. The other half of these genes will be members of the SLC superfamily. the ATPbinding cassette (ABC) and major facilitator (MFS)superfamilies account for nearly 50% of all transporters in each organism. A substantial role has been recognized for transport proteins in oral absorption and drug bioavailability (118). the pathophysiology of several hereditary diseases (i. vitamins.gene. class B contains the well-known multidrug-resistance gene (MDR1) derived Pglycoprotein (ABCBl). Thus. 120)l. 5.1 The ATP-Binding Cassette (ABC) and Solute Carrier (SLC) Genetic Superfamilies Organic solutes such as nutrients (amino acids.uk/nomenclature/). organ expression pattern). neurotransmitters.whereas class C is composed of members of the multidrug-resistance protein (MRP) subfamily. efflux of antineoplastic compounds from tumor cells mediated by multidrug resistance (MDR) gene products (119. and drugs are transferred across cellular membranes by specialized transport systems.2 Therapeutic Implications of Membrane Transporters It has been generally acknowledged that transporters play an important role in clinical pharmacology (Table 8. Most of these mutations in human genes and genetic disorders have been recorded and can . substrate specificity. sugars.g. Furthermore. However. sequence information and functional data derived from numerous transporters have revealed unifying designs. the SLC class contains 37 families with 205 members and is rapidly expanding (http:/J www. The plethora of isolated transporters motivated the Human Gene Nomenclature Committee to classify these proteins into a distinct genetic superfamily named SLC (for SoLute Carrier). membrane topology. the latter energized directly by the hydrolysis of ATP or indirectly by coupling to the cotransport of a counterion down its electrochemical gradient (e.1 present a concise overview of ABC and SLC members that have been classified and characterized. Our understanding of the biochemistry and molecular biology of mammalian transport proteins has significantly advanced since the development of expression cloning techniques. Two superfamilies. To date..ucl. facilitated transporters) or an active process (carriers).. drug toxicity (121). These systems encompass integral membrane proteins that shuttle substrates across the membrane by either a passive process (channels. Currently. c1-1. 115) determined the distribution of membrane transport proteins for all organisms with completely sequenced genomes and identified 81 distinct families. this may eradicate the rampant use of trivial names for these classes of proteins.000 (116).

ALDL OABP. vitamin transporters (A6). y + Sodium/calcium exchangers Sodium/hydrogen exchangers Sodium/bile acid cotransporters Proton-coupled divalent metal ion transporters Sodium/potassium/chloridetransporters Sodium/sulfate and dicarboxylate symporters Urea transporters Oligopeptide transporters Monocarboxylic acid transporters Sodium phosphate transporters Vesicular monoamine transporters Folatelthiamine transporters Phosphate transporters Organic anion transporters EAAT GLUT ATR. retina skeletal and smooth muscles Ubiquitous (intestine. GI tract. GABA Cationic amino acid transporter. NADC UT PepT MCT NPT VAT FOLT. photoreceptor proteins Efflux transporters. intestine. and kidney Kidney. OATP Neurons. dibasic. ASBT NRAMP. brain. kidney. DCT NKCC NaSi. kidney. THTR GLVR OAT. TAP. Ileum Ubiquitous Ubiquitous (kidney) Kidney and red cells Intestine. intestine Brain Placenta. intestine. muscle. and neutral amino acid transporters Anion exchangers (1-31. NIS GAT. thyroid Brain Ubiquitous (intestine. GLYT LAT. brain. GCN20 WHITE Brain Ubiquitous Synoviocytes Eye Solute Carrier Family SLCl (7) SLC2 (11) SLC3 (2) SLC4 (10) SLC5 (7) SLC6 (16) SLC7 (11) SLC8 (3) SLC9 (6) SLClO (2) SLCll (3) SLC12 (1-7) SLC13 (2) SLC14 (2) SLC15 (2) SLCl6 (7) SLC17 (5) SLC18 (3) SLC19 (2) SLCZO (2) SLC21(14) High affinity glutamate transporter Facilitated glucose transporters 1 Cystine. CAT NCX NHE Ntcp. RBAT EPB. retina Most cells Kidney and intestine Erythrocytes. kidney) Heart. intestine . fatty acid transporters. AE1. kidney Erythrocytes. brain. choline (A71 Neurotransmitter transporters. small intestine and other tissues Liver. tumor cells Lung (CFTR). p-a CFTR. kidney) Liver. SMVT. brain. NBC SGLT. RNS4I ABC50. multiple drug resistance associated proteins Cholesterol. iodide transporter (A5). MRP Epithelial cells. liver (MRP) Organ Expression ATP-Binding Cassette Family ABC-A (12) ABC-B (11) ABC-C (12) Cholesterol efflux regulatory proteins. multiple drug resistance Cytic fibrosis transconductance regulators.1 Overview of the ABC and SLC Genetic Superfamilies: Nomenclature and Expression Gene Symbol NameISubstrate Trivial Names CERP.268 Membrane Transport Proteins and Drug Transport Table 8. kidney. adrenoleukodystrophy-associated Rnase L inhibitor TNF-a-stimulated ABC-member. reproductive organs Intestine. sodium bicarbonate cotransporter (4-10) Sodium/glucose cotransporter. RMP MDR. nonmembrane-bound Eye pigment transmembrane permeases ABC-D (4) ABC-E (1) ABC-F (3) ABC-G (6) ALD. kidney.

Recently. kidney.2 Top-Selling 100 FDA-Approved Drugs Molecular Target Transporterslchannels Membrane receptors Enzymes Nuclear receptors Foreign molecules (pathogens) Market (%) 30 25 20 15 5 might change in the future if more research and technologies are applied to this area. resulting in a P290S mutation that abolishes bile acid reuptake (126). mostly undefined) Fatty acid transporters Sodium-coupled nucleoside transporters Nucleoside transporters Zinc transporters Copper transporters GABA vesicular transporter Sodium phosphate CMP-sialic acid and UDP-galactose transporter Glycerol-3-phosphate transporter be found online at the National Institute for Biotechnology Information (http://www. and fat Ubiquitous Ubiquitous Most organs Kidney.5 Transport Proteins Table 8. resulting in high resolution three-dimensional structural information. D28G) in SGLTl(125) and bile acid malabsorption syndrome can be attributed to a single C to T transition in the ASBT gene. congenital glucose-galactose malabsorption syndrome is caused by a defect (D28N. lung.3 Structural Models of Transport Proteins and Methods to Design Substrates The primary and.ncbi.5 A (131). only a few polytopic membrane proteins have yielded to X-ray crystallographic analyses (127-1301. intestine Brain. liver Ubiquitous Liver. Unfortunately. intestine. In this study. liver. pancreas Ubiquitous Sodium/potassium/calcium exchangers Mitochondria1 carriers (citrate. NPT CST. adenine. In summary.000 crystallization conditions using about 20 detergents were tested to yield crystals with good quality for X-ray structure determination. approximately 96. skeletal muscle. secondary structures of many transport systems are known. eye. 5.1 (Continued) Gene Symbol SLC22 (8) SLC23 (2) SLC24 (4) SLC25 (21) SLC26 (11) SLC27 (6) SLC28 (2) SLC29 (2) SLC30 (4) SLC31 (2) SLC32 (1) SLC34 (2) SLC35 (3) SLC37 (1) NameISubstrate Organic cation transporters Nucleobase transporters Trivial Names OCT SVCT NCKX CTP. and function as important mediators governing all aspects of drug therapy. For example. however. the knowledge of their structure and mechanism of action has lagged far behind the knowledge of these properties of proteins in general. this Table 8. DTD FATP CNT ENT ZNT COPT VGAT NaPi3. heart Cartilage and intestine Adipocytes. heart. resulting in the structure of a homolog of the multidrug-resistant ATP-binding cassette transporters to a resolution of 4. gut. to a lesser extent. UGALT Organ Expression Kidney. The combinatorial crystallization approach used by Chang and Roth (131) will likely stimulate . solute transporters play an invaluable role in fundamental cellular processes in health and disease. but the structural information corroborated several previously obtained experimental results. however.gov) OMIM database (Online Mendelian Inheritance in Man). the resulting protein was not functionally active. combinatorial approaches have been applied to successfully crystallize membrane proteins in a high throughput approach. ANT SAT. Despite the apparent clinical importance of SLC proteins. nlm. Because of the absence of suitable crystallization methods. carnitine) Solute carrier family 26 (sulfate.nih.

the PHD Topography neural network system developed by Rost and coworkers (138) (http://www. Each has advantages and disadvantages. often producing a "sliding window" of groups of hydrophobic amino acid residues that are suitable candidates for the formation of a membrane-spanning segment. The membrane insertion scanning technique has been used to determine the membrane topology of various transporters.3). Stop transfer properties of hydrophobic sequences can be examined in a similar manner using the first 139 N-terminal amino acids of the rabbit Hf . Meanwhile.3 lists the techniques that this and other groups have used to solve the structure and topology of membrane transporters in the absence of a crystal structure. Examples of these software programs are TopPred I1 (137) (http://www. 5. 5.e. biokemi. the degree of tertiary structural similarity correlates well with the degree of primary structural similarity.K+-ATPasesubunit (HK MI) containing the first membrane sequence of the Ht. who have taken numerous biophysical approaches toward studying the structure of the bacterial transporter lactose permease (136).4. 4. de/predictprotein). 2.1 Membrane Insertion Scanning. They tested the ability of individual protein segments to function as signal anchors for membrane insertion by placement between a cytoplasmic anchor encompassing the first 101 amino acids of the rabbit Ht. these data can be applied to all other members within limits dictated by their degrees of sequence similarity. share sufficient primary structural similarity to have evolved from a common ancestor) will prove to exhibit strikingly similar three-dimensional structures (135). and only rarely does one method enable conclusive topographic analysis of the membrane-embedded segments of a polytopic integral membrane protein. Considerable progress has been made over the past 10 years by the group of Kaback at UCLA. 9. once threedimesnional structural data are available for any one family member. Table 8. reaction with sided reagents. Mem- brane topology of transporter proteins can be predicted by various computer algorithms.. Sachs and colleagues at UCLA pioneered a technique aptly named "membrane insertion scanning" to determine whether a sequence of amino acids is capable of spanning a membrane. Phylogenetic analyses allow application of modeling techniques to a large number of related proteins and additionally allow reliable extrapolation from one protein family member of known structure to others of unknown structure.3 Techniques for Studying Polytopic Membrane Protein Structure 1. Thus. 3. however. N-glycosylation site engineering There are various approaches to study the topography of integral membrane proteins in the absence of a resolved crystal structure.K+-ATPase as a signal anchor upstream of individual predicted transport transmembrane sequences linked to the N-glycosylation flag.se).enzim. 6. Furthermore.su.Membrane Transport Proteins and Drug Transport other researchers to attempt crystallization of other important membrane proteins.embl-heidelberg.4 Techniques for Studying Integral Membrane Protein Structure Table 8. 5.hu/hmmtop). It is well recognized that any two proteins that show sequence homology (i. Site-directed Ala scanning Membrane insertion scanning Site-directed thiol crosslinkers Excimer fluorescence Engineered divalent metal-binding sites EPR Metal-spin-labeling interactions Site-directed chemical cleavage Identification of discontinuous monoclonal antibody probes 10. (Table 8. we base our views of solute transport on molecular models that provide working models of transport systems (132-134).K+ATPase (HK MO) subunit and a glycosylation flag sequence consisting of five N-linked glycosylation sites located in the C-terminal 177 amino acids of the rabbit HK M O subunit (140). 8. 7. or the hidden Markov model recently described by Tusnady and Simon (139) (http://www. chan- . and proteolysis of the purified membrane-inserted protein have all been used individually or in combination. none of the existing methods can definitively define these regions. I n vitro and in vivo translation of constructs containing one or more transmembrane sequences. epitope localization.

5. 5. EYMPME-tagged hPEPTl mutants with an anti-EYMPME monoclonal antibody in nonpermeabilized and permeabilized cells. Covitz and colleagues (144) used this technique to determine the topology of the peptide transporter. a polytopic membrane transport protein that catalyzes p-galactoside/H+ symport. was inserted into different extramembranous locations of hPEPTl by site-directed mutagenesis. Transmembrane regions in polytopic membrane proteins may require flanking topogenic information to become integrated into the lipid bilayer and. This method has been successfully used to determine which domains of the receptorslchannels are located extracellularly..4. Based on the positioning of the glycosylation groups within the extracellular membrane. The insertion of short reporter sequences (e. A common criticism of membrane insertion studies is the placement of a hydrophobic amino acid sequence out of its physiological or environmental context. 5. this approach requires isolation of homogeneous preparations of intact membranes and many tedious control experiments.3 N-Glycosylation and Epitope Scanning Mutagenesis. A drawback of this technique is the dependency of glycosylation on efficiency and accessibility of the concensus sequence to the glycosylation machinery. the proximity between paired labeled residues can be determined.g. Regardless. A technique that can be used in thus. In general. An epitope tag. In general. the topology of various proteins has been successfully solved using this technique. Moreover. Site-directed addition to membrane insertion scanning is based on the fact that N-glycosylation occurs only on the luminal side of the endoplasmic reticulum. Alternatively. Frillingos and colleagues (141) used Cys-scanning mutagenesis to determine which residues play an obligatory role in the mechanism and to create a library of mutants with a single-Cys residue at each position of the molecule for structure/function studies.4. thus "forcing" an abbreviated sequence through the membrane. .4. a molecular weight shift may indicate successful glycosylation and reveals definitive information on protein topology. Because pyrene exhibits excimer fluorescence if two molecules are within about 3.146). helix tilt. topology determination by this technique alone may not be definitive. excimer fluorescence (SDEF)and site-directed spin labeling (SDSL) are two particularly useful techniques to study proximity relationships in membrane helices. helix packing.2 Cys-Scanning Mutagenesis. Using this technique.5 Transport Proteins nels. PepT1. Kaback and coworkers showed that ligands of the lac permease cause a dramatic increase in reactivity that is consistent with the notion that the mutated amino acid positions are transferred into a more hydrophobic environment (145. Studying the lactose permease of Escherichia coli.4 Excimer Fluorescence. The membrane topology was solved by labeling reconstituted. 5.5 Site-Directed Chemical Cleavage. small peptide epitopes can be inserted into the extramembranous parts of an SLC prote in that are recognized by a wellcharacterized monoclonal antibody. factor Xa protease cleavage sites) into hydrophilic loops has proved to be a useful alternative to N-glycosylation scanning mutagenesis. and ligand-induced conformational changes can be determined by using the library of mutants in conjunction with a battery of sitedirected techniques.5 A. including the sodium-dependent glucose transporter (SGLTI) (142) and the y-aminobutyric acid transporter GAT-1 (143). The experiments are based on site-directed pyrene labeling of combinations of paired Cys replacements in a mutant devoid of Cys residues. N-glycosylation consensus sequences (NXSIT)are engineered into hydrophilic regions of an aglycomutant. and receptors. However. thus. EYMPME. this type of study will define amino acid side chains that play an irreplaceable role in the transport mechanism and positions where the reactivity of the Cys replacement is altered upon ligand binding. interspin distances in the range of 8-25 A between two spin-labeled Cys residues can be measured in the frozen state. nonglycosylated mutants do not provide conclusive information about the topology and these data should be interpreted carefully.4. functionally active. a glycosylation-free mutant is first designed and subsequently. Furthermore.

a drug-metabolizing enzyme with broad substrate specificity. and capillary endothelial cells in the central nervous system (CNS) (150).1 P-Clycoprotein: Understanding the Defining Features of Regulators. fragments are isolated on SDSPAGE and can be analyzed to further determine membrane topology.Moreover. competitive. The factor Xa protease recognizes the tetrapeptide motif IEGR and specifically cleaves the protein sequence COOH-terminal of the arginine residue (148). Interestingly. 158). appears to coexist with P-gp in organs such as the intestine and liver. Substrates.and COOH-terminal domains. P-gp expression at the bloodbrain barrier is a key factor in the limited CNS entry of many drugs. The key to this linkage between P-gp and CYP3A4 appears to be their coregulation by the pregnane-X-receptor at the level of transcription (157. The ABC efflux transporter P-gly- coprotein (P-gp)is a large 12 transmembranedomain bound protein initially noted to be present in certain malignant cells associated with the multidrug resistance (MDR) phenomenon that results from the P-gp-mediated active transport of anticancer drugs from the intracellular to the extracellular compartment (150). After digestion of purified protein vesicles with the factor Xa enzyme..5 Case Studies 5. More recent data have suggested that there may be a dissociation of inhibitory potencies for molecules against these proteins. However. in-frame factor Xa protease sites are inserted into a target sequence at positions within the NH. we may be closer to understanding the structural features necessary for PXR ligands. cytochrome P450 3A4 (CYP3A4). brush border of proximal tubule cells. 5. In addition. Recently. These observations led to the hypothesis that there may be a relationship between these two proteins in the drug disposition process.Generally. and noncompetitive interactions between modulators were found to interact with a t least two binding sites in P-gp (164). Wacher et al. Siniilarly. the presence of multiple drug binding sites has been proposed (160-163).Steady-state kinetic analyses of P-gp-mediated ATPase activity using different substrates indicate that these sites can show mixed-type or noncompetitive inhibition indicative of overlapping substrate specificities (169). found that modulators and substrates coordinately upregulate both proteins in human cell lines (154). for the most part the potency of inhibition for CYP3A4 did not predict the potency of inhibition for P-gp. P-gp-mediated transport was found to be important in influencing the extent of CYP3A induction in the same cell lines and also in mice (155). Other researchers have determined that immobilized P-gp demonstrates competitive behavior between vinblastine and doxorubicin. and into hydrophilic loops. canalicular domain of hepatocytes. Although some molecules can interact with CYP3A4 and P-gp to a similar extent. Subsequent results have indicated there may be three or more binding sites (168). the recognition motif is tandomly (IEGRIEGR) inserted to increase the probability of cleavage (149). and Inhibitors. a pharmacophore for PXR ligands may also enable us to select new drugs that are less likely to induce P-gp and CYP3A4 (159).5.272 Membrane Transport Proteins and Drug Transport and experimental difficulties associated with protease accessibility are well documented (147). The expressed level of P-gp as well as altered functional activity of the protein attributed to genetic variability in the MDRl gene also appears to impact the ability of this transporter to influence the disposition of drug substrates (152). have described the overlapping substrate specificity and tissue distribution of CYP3A4 (153) and P-gp. Schuetz et al. and vice versa (58). In general. cooperative allosteric interactions between cyclosporin and .Expression of P-gp in such locations results in reduced oral drug absorption and enhanced renal and biliary excretion of substrate drugs (151).Moreover. The first elegant experimentally determined signs of a complex behavior for P-gp appeared in 1996 when cooperative. To account for the observed broad substrate specificities for both CYP3A4 and P-gp. P-gp is normally expressed at many physiological barriers including the intestinal epithelium. The multiple site hypothesis was confirmed by other groups (165-167). not all CYP3A substrates such as midazolam and nifedipine are P-gp substrates (156). with the description of the X-ray crystal structure of this protein with SR12813 bound.

This is not surprising. The P-gp modulator LY 335979 has been shown to competitively block vinblastine binding (1731. This hypothesis derives from experimental results describing a complex behavior for P-gp. there has been over a decade's worth of studies that have identified pharmacophores for this efflux pump that could have been used to help predict P-gprelated bioavailability issues. Clearly allosteric behavior by multiple substrates.5-4. vinblastine and calcein accumulation in P-gp expressing LLC-PK1 (L-MDR1)cells. and simple regression models of propafenone analogs (184. In 1997 the first 3D-QSAR (quantitative structure-activity relationships) analysis of phenothiazines and related drugs known to be P-gp inhibitors was described (181). Early computational studies using P-gp modulators such as verapamil. suggested to form a hydrogen bond with P-gp (178). photoaffinity experiments have been valuable in defining the cyclosporin binding site in hamster P-gp (171) and indicating that trimethoxybenzoylyohimbine (TMBY)and verapamil bind to a single or overlapping sites in a human leukemic cell line (172). it would be valuable to use structural information to define whether unrelated molecules are likely to interact with P-gp. (179). 175). This was followed by Hansch-type QSAR studies with propafenone analogs (182). vinblastine binding in vesicles derived from CEMI VLBlOO cells. However.5 Transport Proteins vinblastine or ATP.6 A apart (186). more extensive study with 232 phenothiazines and structurally related compounds indicated that molecules with a carbonyl group that is part of an amide bond plus a tertiary arnine were active P-gp inhibitors (176). 185). and verapamil (170). such that cooperative. A subsequent. and linear discriminant analysis with topological descriptors (180). Interestingly. The inhibitor pharmacophores ' . With the growth in knowledge derived from these and other studies. polarizability. and noncompetitive interactions between modulators may occur (164). vinblastine. Others have used MULTICASE to determine important substructural features like CH2-CH2-N-CH2-CH. CoMFA studies of phenothiazines and related drugs (1831. 19 propafenone-type P-gp inhibitors were then used to confirm the requirement for a carbonyl oxygen. A number of computational approaches and models of P-gp have yielded useful information that is usually derived from a series of structurally related molecules. One study using a diverse array of inhibitors with P-gp-ATPase activity noted that size of the molecular surface. indicative of multiple binding sites within P-gp (165-167). given that in vitro studies with P-gp inhibitors frequently take this approach. 18-epireserpine. Additional studies have shown that TMBY is a competitive inhibitor of vinblastine binding to P-gp (173). These latter models confirmed the relevance of hydrogen bond acceptors and the basic nitrogen for inhibitors (184. and hydrogen bonding had the largest impact on the ATPase activity (187). inhibitors. some complexity arises if one considers more structurally diverse molecules because they may bind to different sites within P-gp. A recent example suggested P-gp inhibitors with high lipophilicity and polarizability were found to be more likely to be high affinity ligands for the verapamil-binding site (188). 185) and multiple hydrogen bond donors in substrates 2. Similarly. and verapamil binding in Caco-2 cells (159) suggested there may be some overlap in the binding sites for these four substrates. and anticooperative allosteric interactions between ATP. competitive. or modulators of CYP3A4 or P-gp complicates predicting the behavior and drug-drug interactions of new molecules in vivo and has important implications for drug discovery. Recently.whereas vinblastine itself can competitively inhibit verapamil stimulation of P-gp-ATPase (172). In terms of understanding P-gp structureactivity relationships. reserpine. A model built with 21 molecules of various structural classes that modulate P-gpATPase activity suggested these molecules competed for a single binding site (177). specific models addressing the individual P-gp binding sites using a diverse array of inhibitors have been described. and TMBY showed that they could be aligned. The use of a computational approach to model in vitro data derived from structurally diverse inhibitors of digoxin transport in Caco-2 cells. suggesting the importance of aromatic rings and a basic nitrogen atom in P-gp modulation (174.

di. The latter may be of particular importance because we are now in an era in which membrane proteins are being crystallized. 196) and angiotensin-converting enzyme (ACE) inhibitors (197-207).and tripeptides occurring in food products (190-194). the transporter has been recognized as an important intermediate in the oral bioavailability of peptidomimetic compounds (199). coli. targeting delivery to nutrient transport systems in the gastrointestinal tract has emerged as a strategy to increase oral availability of poorly absorbed drugs and it has met with several successes. POT family members have two characteristic protein signatures assigned by Steiner and Becker. in that it appears to possess a pore region ringed by protein (189). In this approach. such as p-lactam antibiotics (195. 5 . The question of whether P-gp behaves like this flippase or like more traditional pores still remains.6). given that strong correlations between these models were observed. 211). the proton/oligopeptide transporter (POT) family. but computational techniques will be at the forefront of continuing to provide some enlightenment and aid in drug discovery. the intracellular loop. Over the past two de- cades. Currently there is a keen interest in understanding the structural determinants for substrates and inhibitors of this transport protein. thereby facilitating transport across the intestinal wall. The first of these signature sequences includes the end of transmembrane span 2 (TMDZ). onto which vinblastine was aligned (159). In general. where currently only two members have been identified in humans. 5 . It is suggested that P-gp pharmacophores could be docked into these structures and the amino acids likely to correspond on both monomers defined. hPepTl and hPepT2.5-A resolution X-ray structure of the lipid flippase MsbA from E. drug penetration is enhanced by coupling a drug molecule to natural ligands for nutrient transporters. 5 . and verapamil are likely to bind a single site. low affinity. active. this approach aims to create novel drug entities that mimic the three-dimensional features of natural ligands. In fact. As models are generated for the other binding sites and substrates. 209). and TMD3. A third consensus sequence (GTGGIKPXV). known as the PTR2 family signatures (210. 8. provides a further structure on which to build homology models for P-gp because its amino acid sequence is similar (73%) (131). 2 .Membrane Transport Proteins and Drug Transport were suggested to be mainly large with hydrophobic and hydrogen bonding features. 2 Physiology of PepT1.2 The Intestinal Peptide Transporter (PepT1) 5 . For this reason. Such models will be of value for modulating the bioavailability of drugs and understanding the locations of the P-gp binding site(s). However. The 22-A resolution X-ray structure of the monomer for the related ABC family member (MRPlIABCCl) is suggested to be structurally similar to P-gp. they may help in elucidating important structural differences between ligands for each site. PepTl belongs to a larger family of oligopeptide transporters. the intestinal small peptide carrier (PepT1) is a proton-coupled. In addition. it shows affinity toward a broad range of peptidelike pharmaceutically relevant compounds. the lack of knowledge regarding structural specificity toward its substrates has prevented the use of this transporter on a more rational basis. 1 introduction. oligopeptide transport system with a broad substrate specificity. Alternatively. The second PTR2 signature corresponds to the core of TMD5. The recent determination of the 4. The toxicity and immunogenicity of drug-ligand conjugates is expected to be inherently low because of the use of endogenous substrates. an approach named "substrate 'mimicry" can be used. A simple P-gp substrate pharmacophore was generated using the alignment of verapamil and digoxin. these molecules can oftentimes be viewed as "peptidelike" in their molecular composition (Fig. It is hypothesized that vinblastine. In addition to transporting its natural substrates. 5.proposed by Fei and cowork- . another ABC family member. its cellular localization to the apical membrane does not provide a mechanism for cellular exit into the basal compartment (208. Both approaches should result in compounds that are recognized by a specific transport protein embedded in the enterocyte brush-border membrane. digoxin. 2 .5. This model also contained many hydrophobic and hydrogen bond acceptor features.

respectively. originally developed by Marshall and coworkers (217). Early studies were directed to define a pharmacophoric pattern for this transporter (214-216). Hydropathic analysis also predicts that the N. The three amino acids phenylalanine. and cysteine are shown to merge through (pseudo)peptide bonds. and 9TM and 10TM. Because no three-dimensional structure of the transporter is available. Structural information on PepTl has been limited to its primary sequence and predicted structural membrane topology. 5.3 Structure-Transport Relationship of PepT1. with an estimated pI of 8. is well conserved between mammalian and C. elegans peptide transporters CPTA and CPTB.6. Molecular building blocks of the p-lactam antibiotic cephalexin.213).5. The results using EE epitope insertions in the amino terminal region were inconclusive. rabbit. has proved to be useful in rationalizing and predicting pharmacological data of active and inactive substrates. In the AAA. The human PepTl sequence has been predicted to contain a 2127 base-pair (bp) open reading frame that encodes a 708 amino acid. Hydropathy analysis of the human.2. were extracellularly localized (144). whereas epitope tags at positions 106 and 412 showed that the predicted loops between 3TM and 4TM. This model has been partially proved by other investigators (144. leaving ambiguity to the predicted structure from the Nterminus to TMD3.CH3 CH2-CH-COOH I A I H2N-CH-CH2 COOH HZN 'CH I CH3 s Phenylalanine t Cysteine Valine COOH Cephalexin Figure 8. the structural requirements common to a set of compounds showing affinity for the same transporter may be used to define a pharmacophoric pattern of atoms or groups of atoms mutually oriented in space. Two factors have to be considered . A pharmacophore or "recognition site" is defined based on a common arrangement of essential atoms or groups of atoms appearing in each active molecule.5 Transport Proteins COOH I . design of new substrates for PepTl has relied on indirect structure-affinity relationships.58 (212). Site-directed mutagenesis. possibly because of EE effects on function.CH N H2 . and rat PepTl isoforms have predicted the presence of 12 transmembrane domains (TMD) in each isoform (212). using EYMPME (EEbepitope tag insertion at different locations of the human PepTl transmembrane regions have identified the COOH-terminal to be intracellular.These results support the predicted loop and TMD numbers and orientations from TMD4 through the C-terminus (144). In many published studies the active analog approach (AAA). but is not specific to or well conserved in other members of the POT superfamily. valine. 79-kDa protein.and C-termini are located on the cytoplasmic side of the membrane. ers (212). to form a peptidomimetic compound with the molecular features of a tripeptide. which are necessary for binding to this transporter.

Enalapril and lisinopril are substrates for PepTl. Another study examined the three-dimensional structural features of three structurally closely related PepTl substrates: enalapril. values) appearing not to change substantially with chain lengths from 5-11 CH. The conclusions of this study indicate that a peptide bond is not essential for substrates of PepTl. it allowed explanation of the variation in the permeability with respect to steric and electrostatic interaction energies of the molecules (222). the active compound (enalaprilat) is formed by bioconversion of enalapril.. After oral administration of enalapril. enalaprilat. and finally an amine nitrogen atom (hydrogen bonding region) were important features of substrates (218). The CoMFA approach required manual alignment of relatively rigid molecules. Binding of o-AFAs was demonstrated to inhibit D-Phe-Ala uptake competitively in transiently transfected hPepTl cells. a hydrophobic site. we showed that intramolecular hydrogen bond formation between the lysyl side chain in enalaprilat and its carboxylic acid groups may explain decreased affinity for PepTl (220). These studies clearly demonstrated that 4-10 methylene carbon groups (CH.7) (220).--CH. Molecular structure of the ACE-inhibitors enalapril. but is poorly absorbed from the gastrointestinal tract (3-12% bioavailability) (221). By use of a combined in vitro and molecular modeling approach. The distance between the charged functionalities for PepTl recognition has been demonstrated to be critical (223). with the competitive inhibition (EC. One study used p-lactams and showed that a carboxylic carbon (likely to position in a positively charged pocket). The prodrug approach of esterifying enalaprilat to enalapril is required to enhance the oral bioavailability to 60-70%. 219). Enalapril is an ester prodrug of the pharmacologically active enalaprilat. Insights into the affinity of p-amino groups were also elucidated in these studies. units between the terminal functional groups (223). in that it requires structurally similar rigid molecules. 215.7.CH3 -CH3 -H Figure 8. binds slowly and tightly to ACE. such as a p-lactam nucleus (214. units. di. the AAA is limited. however. Structural studies of PepTl substrates were extended using comparative molecular field analysis (CoMFA)of 10 known substrates for the peptide transporter with data derived from an in situ rat model (222). Studies by other research groups have sought to address the requirements for binding and transport by PepTl.and tripeptides. This 3DQSAR produced by our laboratory provided a valuable starting point for future prediction of Enalapril Enalaprilat -C2H5 . contain many freely rotatable bonds. Enalaprilat. and that two ionized amino or carboxyl groups with at least four CH. are required for transport by PepTl (223). and lisinopril (Fig. two carbonyl oxygen atoms (hydrogen bond acceptors). most studies have concentrated on substrates with more rigid backbones. where 8-amino-o~- . 218. the affinity of substrates for this transporter and at the same time confirmed the earlier findings from AAA studies.Membrane Transport Proteins and Drug Transport for the successful application of this technique: the included compounds should be structurally homogeneous and the individual compounds should have a low level of conformational flexibility. Therefore. 8. This study also raised considerable concerns about conformation during presentation because the CH. enalaprilat. a diacid. Because the natural substrates. and lisinopril. units) could be accommodated between the termini and allow direct binding with PepTl. providing a distance of 500 to 635 pm between them. producing well-defined clinical effects. The allowable backbone distance between the amine and the carboxylic acid terminals for binding to and transport by PepTl have been demonstrated in part with a series of o-amino fatty acids (w-AFA) (223). midsection has considerable flexibility in contrast to that of a peptide bond.

whereas 2-amino-octanoic acid had no inhibitory effect (223).5. A detailed description of the enterohepatic circulation of bile acids may be found in several reviews (234-239).5 Transport Proteins tanoic acid effectively inhibited D-Phe-Ala uptake. medium. and serve as the key determinant in PepT2 affmity preference over PepTl. anionic p-lactam antibiotics. whereas the residue at position 260 remains to be completely defined (228). Recent studies have used Caco-2 cells and assessed the inhibition (Ki) of Gly-[3Hl-~-Pro transport by ACE inhibitors as a means to understand dipeptide transport. The primary bile acids cholate and chenodeoxycholate are synthesized from cholesterol in the liver mediated by the enzyme cholesterol 7a-hydroxylase (CYP7A) (Fig.2-0. circulates 6-10 times a day. The "apical sodium- dependent bile acid transporter" (ASBT) has been targeted for oral drug delivery as well. Another group has generated Ki values for 23 p-lactams and suggested their data supported the need for all of the structural features previously described for recognition by the transporter (226). The molecular mechanismb) by which these residues may interact with H+ remains to be elucidated. Bile acids mediate the digestion and absorption of fat and fat-soluble vitamins (232). and thus being fundamental for transport function. 228).5. However. The total bile acid pool in humans. The activity of this enzyme is under inhibitory feedback control by bile ' . 3-5 g. and a carboxylate binding site (225). Inhibition of intestinal bile acid reabsorption elicits increased hepatic bile acid synthesis from its precursor cholesterol. whereas Tyr at position 167 (TMD5) inactivated the transporter completely (227). A two-dimensional model was constructed that indicated the types of functional groups favored for inhibition (197). 5.5. This group is speculated to be involved in the Hf binding. ASBT has recently received much attention because of its pivotal role in cholesterol metabolism. lacking an a-amino group.3. The high efficacy of ASBT combined with its high capacity makes this system an interesting target for drug delivery purposes. the presence of a p-amino group resulted in the substrate having a higher affinity for transport by PepT2 than by PepTl (224). further studies have demonstrated that the a. such as cholestyramine and colestipol (230). a hydrogen bond to the carbony1 group of the first peptide bond. Interestingly. Site-directed mutations of single amino acids located within the PepTl transmembrane domains (TMDs) showed that Trp at position 294 and Glu at position 595 reduced significantly the glycylsarcosine uptake by human embryonic kidney cells (HEK293). 5. several novel ASBT inhibitors are in clinical trials for the treatment of hypercholesterolemia (231).5 g of bile acids is lost in feces per day and this amount is repleted by the de novo synthesis of bile acids (233). Only 0. demonstrated a preference for PepTl. site. and H260 residues (corresponding positions in the PepTl sequence) in PepTl and PepT2. A recent study has taken this understanding of PepT1 further by using a meta-analysis of Ki data for 42 substrates across eight classes of molecules (225).2 Physiology of ASBT. including local drug targeting to the intestine and improving the intestinal absorption of poorly absorbable drugs.3 The Apical Sodium-Dependent Bile Acid Transporter (ASBT) 5. Sequence alignment demon- strated the presence of H57.This model was also used to classify the compounds as possessing high. a hydrophobic pocket. Histidyl residues at positions 57 and 121 have been demonstrated to be essential to maintain PepTl function.8). where coexpressed (224). thereby lowering plasma cholesterol levels (229). or low affinity for PepTl.or p-amino carbonyl functionality may affect PepTl substrate affinity.3. PepTl has shown to present another important structural characteristic: the presence of conserved histidyl residues (H) in several transmembrane regions. giving rise to a daily bile acid turnover of 20-30 gin humans. For example. H121. Application of this approach has met considerable clinical success using unspecific bile acid sequestrants. The findings by these authors were in agreement with previous studies and provided a template consisting of an Nterminal NH.1 Introduction. Currently.8. Other studies have shown the importance of certain amino acid residues in determining PepTl peptide transport activity (227.

although it was originally named gastrotropin (241. In light of the importance of this pathway for the removal of cholesterol. This protein is generally known as ileal bile acid binding protein (IBABP) (240). acids. but active reabsorption is confined to the distal ileum to minimize loss of bile salts in the feces. mtest~ne ?/ Loss in urine circulation Intestinal absorption Loss in feces Figure 8. Within the cytoplasm of the enterocyte.to 1BkDa soluble protein binds the absorbed bile acid and mediates its transfer across the cell. Intestinal uptake of bile acids takes place along the entire length of the small intestine.8. They are stored in the gallbladder and released through the bile duct into the duodenum. Biosynthesis of bile acids and the enterohepatic circulation. The portal circulation carries bile acids from the intestine to the liver. The exact in- . In the ileum. where they aid in the digestion of dietary fats. Bile acids are synthesized from cholesterol in the liver under feedback regulation of the nuclear orphan receptors farnesoid X receptor (FXR) and lignane X receptor (LXR). 242). conjugated bile acids are reabsorbed by ASBT (also named IBAT) gene SLClOA2. a second 14. where they are actively absorbed by hepatocytes and secreted into bile. the mechanisms for transcription regulation of the CYP7A gene have been extensively studied.278 Membrane Transport Proteins and Drug Transport # lnto Secretion .

Based on membrane insertion scanning techniques. whereas cholic acid was much less effective and the secondary bile acids deoxycholic and lithocholic acid had variable responses. 258) were the f i s t to clone and characterize a 43-kDa protein as the rat ileal bile acid transporter protein. which plays a role in many transport processes in the cell (245). Timed photoaffinity labeling techniques revealed the brush- border (254). Only four of these 12 cysteines are conserved in the human hepatic bile acid transporter (Ntcp).261). Microinjection of guinea pig and rabbit ileal mucosal Poly (A+) mRNA into Xenopus laevis oocytes resulted in translation of a functional Naf/bile acid carrier that is expressed in the surface membrane (253). active cotransporter driven by a Na+-gradient across the apical membrane of ileocytes with a 2:l Nat:bile acid coupling stoichiometry (259).JCcfl) (126. 247). appears to be regulated by oxysterols (249). Human ASBT contains 13 cysteine residues. rat. but there has been speculation that IBABP may bind to ASBT at the cytosolic surface and may play a role in mediating bile acid affinity for ASBT (243). 256). The response was greatest in the presence of chenodeoxycholic acid (CDCA). In addition to IBABP. named bile acid response element (BARE) (248). incubation with rabbit jejunal-mucosa Poly (A+) mRNA did not result in expression of the bile acid transporter.9). structural information on ASBT has been limited to its primary (sequence) and secondary (membrane topology) structures. Expression of IBABP and ASBT is regulated by the farnesoid X receptor (FXR). the 7TM model appears to be the most feasible topology for ASBT to date (Fig. this model suggests the presence of two very short TM domains. a related orphan transporter. a and P. Furthermore. the a isoform is responsible for regulating CYP7A and some SAR and pharmacophore studies suggest the 24-0x0 ligands acting as hydrogen bond acceptors bind a and may be candidates for the tightly to m natural ligand (250. 8. although the specific affinity of these proteins for bile acids is probably low. ASBT is primarily expressed on the apical membranes of ileal enterocytes in mammals. For example. the putative 43-kDa protein was shown to share high homology with actin. A specific binding site for FXR was found on the promotor region of the IBABP gene. The 99-kDa protein that was detected previously can be explained as a dimeric form of the glycosylated transport protein (Fig. Topography models of ASBT predict seven transmembrane (7TM) regions (258).10). a nuclear orphan receptor (246. As expected.5 Transport Proteins teraction between ASBT and IBABP is currently unknown. To date. these observations remain controversial and may have resulted from (patho)physiologic induction. Dawson and coworkers (257. Because bile acids are biosynthesized from - . of which 12 are conserved in the hamster. Two isoforms of LXR. and basolateral proteins (255) involved in the transfer of bile acids across the rat ileal enterocyte. a member of the steroid receptor superfamily that. Overall. in turn. and 43-kDa proteins that have remained unidentified appear to bind bile acids in the cytosol (244). but eight or nine membrane-spanning regions have been suggested by topology/hydropathy analysis (260). 35-. however. have been identified. Studies with antibodies raised against the terminal amino acid sequences verify that the amino terminal is located extracellularly. although weak translation has been observed in the jejunum. A putative transport protein involved in coupled Naf bile acid transfer across the ileal brush-border membrane had a molecular weight of 99 kDa (254. whereas the carboxy terminal is located at the cytoplasmic side of the membrane (262). 258. 251). which effectively eliminates the putative 8TM model. 8. a 9TM model was recently suggested by Hall6n and coworkers (260). cytosolic (244). presumably in the p-helix conformation. 20-. which has 37% homology and 48% similarity with ASBT. Experimental evidence indicates a trans position of the N. Molecular regulation of CYP7A appears to be regulated by the liver X receptor (LXR). however. This unique topology makes ASBT different from other members of the SLC family in that it does not adhere to the "positive-inside" rule (263).where it probably functions as part of a bile acid secretion feedback mechanism. ASBT has been detected in rat bile duct epithelial cells (cholangiocytes) (252). rabbit.and C-terminal protein domains (Ne. Subsequent electrophsyiological characterization revealed that ASBT is an electrogenic. and mouse. and only two cysteines are conserved in the P3protein.

I and Weiner were the first to establish a 1: structure-activity relationship for intes bile acid transport using the rat evertec model (268). located at the basolateral membrane. The V . 8. 2. ASBT. 5. 272). The studies with endogenous bile acids addition to or alterations at the C-17 chain (Fig. such as taurodehy cholic acid.These findings substantiate the feasibility of using ASBT as the new pharmacological target for cholesterol-lowering therapy. : have led to physiological understanding of the function of bile acids and the enterohepatic circulation. Krag and P (275) found similar structure-V. These modifications typically entail either the substitution of the hydroxyl groups at the 3. In the cytosol bile acids are shuttled through the cell by the aid of various proteins. The search for a deeper understanding of the molecular mechanism behind the affinity and recognition of molecules by the bile acids carriers in both ileum and liver has led researchers to modify bile acids and study the carrier affinity of these compounds. In a series of seminal experiments. Recently.7. Physiology and molecular biology of intestinal bile acid transport. most importantly the ileal bill acid binding protein. The sodium gradient i maintained by the sodium-potassium ATPase. following general observations were repol 1 .11).3 Structure-Activity Relationships for ASBT. several compounds have been shown to be able to lower serum cholesterol in animal studies by specifically inhibiting ASBT (264-267). no single droxy group is essential for trans] Triketo bile acids. are devoid of all hydr groups and show considerably less tr k port capacity (273. ASBT Tu> K+ Na+ ATP ' ADP + Pi %GAnionexchanger Figure 8. or 12 positions by other functionalities or the . The K.. a specific nonabsorbable inhibitor of the ileal bile acid transporter system would lower plasma cholesterol level by blocking the intestinal reabsorption of bile acids and consequently raising the conversion of cholesterol to bile acids. is related to whether the bile a( conjugated or unconjugated (269. iBABP. cholesterol and are the only forms for cholesterol to be excreted in uiuo. An anion exchanger transports bile acids across the basolateral mem brane into the portal circulation.Membrane Transport Proteins and Drug Tran Brush-border membrane A Na+/K+ATPase Basolateral membrane > . Bile acids are activel: absorbed in enterocytes through a sodium-dependent cotransporter.9. is independent of conjugatior appears to be related to the number o: droxyl groups attached to the sterol dihydroxy > mon cleus: trihydroxy 8droxy (271. However. 270 general. Important additional find were recorded by the group of Kramer. value than that of t h e i ~ conjugated parent compounds (271). conjugated bile acids have a fold higher K .5. 274).3.

7. no correlation was observed between KT values and C01njugation. the addition of an extra: negative charge in the form of sulfonati. Lack and coworkers have proposed that the recognition site for carrier-mediated bile acid transport is a hydrophobic pocket on the membrane surface that consists of three components: a recognition site for interaction with the steroid nucleus. Studies with antibodies to terminal epitopes have confirmed the inside-out ientation of this membrane protein. an . and an anionic site for interaction with Na+. For efficient transport. All. however. 5. Replacing the anion~ i c moiety from the C-17 side-chain to pot&ion 3 results in a complete loss of affin: ity (277).hough bile acids with two negative chiarges around the C-17 position give minimal active transport. A single negative charge in the general vicinity of the C-17 side-chain. on at the C-3 position does not preclude act. a cationic site for coulombic interaction with the negatively charged side chain. or 12. Hydropathy analysis proposes a seven transmembrane pology for ASBT. 3. 2. Membrane topology of ASBT. St€ !reospecificity of the hydroxyl groups on ! sterol nucleus was shown by substitut h€ tio:n of this group at the 3-position with a lroxyethoxy moiety (279). Supposedly.ive transport (278).10. 4. Generalizations on the structural requirements for ASBT affinity include: 1 . however. The presence of at least one hydroxyl group on the steroid nucleus at position 3. a bile acid molecule ml1st possess a single negative charge (276) thi~t should be located on the C-17 sidechisin of the sterol nucleus (277).5 Tran sport Proteins igure 8. this anionic site could be responsible for the reduced affinity of bile acid derivatives with a dianionic side chain (280). The 3a-iso~ Y C mer was able to inhibit transport of L3FI]taurocholate in rabbit ileal brush-borvesicles. tioinships in humans. whereas the 3P-isod emembrane ~ mer showed very weak affinity and was unable to inhibit taurocholic acid transport.

and (4) the 3a-OH group does not necessarily map a hydrogen bond functionality.Membrane Transport Proteins and Drug Transport RI -OH -OH -OH -OH -OH -OH -OH -OH -OH -NHCH2COOH -NH(CH2)2S03H R2 R3 R4 R5 Prefix -OH -OH -OH -OH -OH -OH -OH (p) -OH -OH -H -H -H -H -H -H -H -H -OH -OH -OH -H -H -OH (B) -OH (.@ -OH (B) -H -OH -OH -H -OH -H -OH -H -H -OH (B) -H - ChenodeoxyDeoxyLithoUrsoUrsodeoxyIsoursodeoxyLagodeoxyHyo- - - GIYCO-' ~aurol 'The prefixes glyco. our group and the joint laboratories of Baringhaus and Kramer have reported on the three-dimensional structural requirements of ASBT that will be of specific use in the development of novel substrates for this transport protein.e. 3.11. A cis configuration of rings A and B within the sterol nucleus. al- . additional negative charge at the C-3 position does not prevent active transport.prevail all others. glyco-hyocholic acid. Substitutions at the C-17 position do not interfere with transport as long as a negative charge around the C-24 position remains present. The ASBT pharmacophore model is in good agreement with the 3D-QSAR model we previously developed using a series of 30 ASBT inhibitors and substrates (282). By use of a training set of 17 chemically diverse inhibitors of ASBT. More recently.and tauro. one hydrogen bond acceptor. and three hydropho- bic features. In this study. This model enabled the in silico design of substrates for ASBT. however. (2) an a-OH group at position 7 or 12 constituted a hydrogen donor. the electrostatic and steric fields around bile acids were mapped using comparative molecular field analysis (CoMFA) to identify regions of putative interaction with ASBT. Figure 8. It should be pointed out that our model was silent on biological diversity around the C-3 position because of limited molecular variability in this region. i. 5. 4. especially for conjugation at the C-17 position. For natural bile acids they found that (1)ring D in combination with methyl-18 mapped one hydrophobic site and methyl-21 mapped a second. (3) the negatively charged side chain constituted the hydrogen bond acceptor. Structure and nomenclature of bile acids. Baringhaus and colleagues (281) developed an enantiospecific catalyst pharmacophore that mapped the molecular features essential for ASBT affinity: one hydrogen bond donor. Substitutions at the C-3 position do not interfere with active transport.

g knowledge-based homology modeling (284). describing acid transport systems for: (1) the improvethe sy: nthesis of sulphonylcholylamide. 8. was clisimed to be particularly suitable for the biologically active. BY PIrobing the protein surface with cholic acid w ec identified five binding domains. Bile acids display a physiologic c~rganotropism for the liver and the small ine. shuttle compounds across the intestinal epim or to target them to the liver and the theliu: hepatc~biliary system. we haveI recently constructed a model for the ligandl binding of the ASBT transport protein usin. which suggests that bile acid transintest: Figure8. Currently. In addition to the previously described pharmaco] phore and 3D-QSAR models. (2)liver site-directed delivery of a attack the liver.msport Proteins terations at that position have been described letail by Kramer and coworkers (283). there is no crystal structure for ASBT and it is antic:ipated that suitable methods for crystallizing intrinsic membrane proteins will not be avai: lable for several years. athways can be exploited in two ways: to of ASBT and its binding sites. As an alternative to high resolution molecular information. tial therapeutic applications of using the bile patent. the molecular Ipepresentation of ASBT may provide a powel*fultool in the design of novel substrates or inh~ibitors for this transporter.5 L3. three of which are located on the outer surface of the protetin (Fig. the extracellular' loop regions were superposed and optimizem d with molecular dynamics simulations.4 Use of ASBT for Enhancing Intestinal A. which ment of the oral absorption of an intrinsically. Another binding domain was 1' ocated between two extracellular loops in close enough proximity to the first binding region 1to accommodate dicholic acid conjugates.12. There may also be the possibility of combining models from different grotIPS. This approach was first 1980s when Ho (285) suggested several potenenvisicmed in 1948 by Berczeller in a U.application (No. 2.441.1291. Putative three-dimensional structure port p.bsorption. 5.12). Ch e serious limitation of indirect structure-a ctivity models is the inherent lack of inforniation on substrate-transport protein interalctions at the molecular level." It was not until the middrug to bring about high therapeutic concen- .although this may present its own prokdems.S. Using the transmembrane domains of bactc~riorhodopsin as a scaffold. but poorly absorbed hydrotreatrrlent of "germ and virus diseases which philic drug. this c~bservation explains the extreme inhibitory c:apacity of this class of compounds (110). The in d two indirect models outlined above should facilitate the rational design of (pro)drugs for targeting to ASBT.

Thus far.e.. liver targeting is easily accomplished. within either the enterocyte or the portal vein. the following drug-targeting strategies are feasible with bile acids: 1.5. In clinical practice.5. 3-tosyl. The 24-position appears to be an attractive site in the bile acid molecule for coupling purposes. will systemic delivery using a prodrug approach be successful. the drug moiety must be released from the bile acid it is coupled to. and (3)gallbladder-site delivery systems of cholecystographic agents and cholesterol gallstone dissolution accelerators. It should be stressed. Hypercholesterolemia is well known as a major risk factor for coronary heart disease. however. From a structural perspective.3 Cholesterol-Reducing Agents. In that case. One is the 3-hydroxy-3-methylgrutaryl coenzyme A (HMGCoA) reductase inhibitors (such as Lipitor). The uses of carrier-mediated bile acid for drug delivery purposes can be divided in three groups: liver. whereas systemic drug delivery needs to address the problem of rapid biliary excretion. another is the bile acid sequestrants.5. most studies have focused on using the bile acid transport system for liver targeting. By use of a prodrug approach. who primarily used the 3-position for drug conjugation.1 Liver and Gallbladder Delivery. 5. and lowering serum cholesterol. So far. 5.3. 5. The carboxylic acid moiety is easily linked to an amine using conventional peptide-synthesizing techniques (292-294). Although promising.3. When one compares the maximal transport f l u x (J-) values of taurine conjugated bile acids measured in rat liver and distal ileum. 5.Membrane Transport Proteins and Drug Transport trations in the diseased liver with the minimization of general toxic reactions elsewhere in the body. 2. Our recent studies with small peptides attached to the C-24 position further demonstrate the general applicability of this transport system for drug targeting and delivery purposes (286). The feasibility and viability of using the bile acid transport pathway for targeting of drug-bile acid derivatives was more definitively demonstrated by the laboratory of Kramer. 7. Apart from the necessity of a negatively charged group around the C-24 position. The main drawback of .and gallbladder-directed delivery. respectively. however. and 3-benzoylcholate) that were handled like natural bile acids by the liver and intestine during in situ perfusion experiments.2 Systemic Delivery. making the synthesis of these compounds relatively easy. Oral absorption enhancement can be directed to either the liver or gallbladder or systemic delivery. Both the 3-position and the 24-position appear to be usable coupling points for a prodrug strategem. 3-iodo. His experimental evidence was based on bile acid analogs with minor structural modifications (i. with a bile acid molecule as a shuttle. and L-T3. a trend for relatively higher hepatic maximal transport rates can be observed. HMGCoA reductase inhibitors. oral absorption enhancement.5. These studies prove that the coupling of drug entities to bile acids does not cause a loss of affinity for the hepatic bile acid carrier.5.3. that the need for a negatively charged group around the C-24 position is required. the suitability of this transport system for systemic drug delivery remains to be demonstrated.3. or 12 with conservation of the C-17 side-chain. that no studies so far have attempted to develop a prodrug approach in which the drug will be released before arrival in the liver.5 Therapeutic Applications of ASBT. Only if these conditions are met. largely based on the above-outlined structure-activity relationships. such as cholestyramine and colestipol(97). This observation can have important consequences when using the bile acid transporter for oral drug delivery. Attachment of drugs to the steroid nucleus at hydroxyl positions 3. The research groups of Kramer (287-289) and Stephan (290) have successfully shown specific hepatic delivery of chlorambucil. It has to be mentioned. which bind bile acid in the intestinal lumen and thus increase their excretion.5. no single study has unequivocally shown the release of the parent compound from the conjugate after passage across the intestinal wall. Attachment of drugs to the bile acid sidechain ((2-17) with positional retention of the negatively charged group. Kim and colleagues (291) have shown that some size restrictions apply when compounds are coupled to the C-3 position. two main hypocholestrolemic agents are commonly used.5.

The inhibition of the intestinal bile acid transport system is thought to be the underlying mechanism for an increased fecal bile acid excretion and lower plasma LDL cholesterol levels after oral administration of these drugs. was able to selectively inhibit active ileal bile acid absorption in rats. mice. several molecules have been found to possess this effect in animal studies (108-110. . such as high dosages of 10-30 g per day. any reagent that can inhibit the bile acid active transport system could block the reabsorption of bile acids and consequently reduce the serum cholesterol level. mice. resulting in an efficient inhibition of bile acid reabsorption without or with only low absorption of the inhibitor itself (110) (Fig. So far. was able to reduce serum cholesterol in hamsters. and rabbits. and malabsorption syndromes.131.13). Structure of specific bile acid transporter inhibitors. maldigestion. 8. See text for additional details. these agents is poor compliance of patients stemming from adverse side effects. it was shown that a benzothiazepine derivative. another compound. mon- keys. 6 CONCLUSIONS AND FUTURE DIRECTIONS It is clear that over the past decade we have gained a great deal of knowledge around the functioning of SLC proteins and how they can be manipulated as drug targets for maximizing absorption and bioactivity through prodrug approaches. S-8921 (Fig. a lignan derivative. 8. 2164U90. Recently. As an alternative method to bile acid sequestrants. 297).6 Conclusions and Future Directions OMe@ OMe S-8921 OMe Figure 8. and humans (296. Similarly. The first such inhibitors consisted of the coupling of two bile acid molecules by means of a spacer to allow simultaneous interaction with more than one transporter site.295).13. constipation. Although we are still without a crystal structure for the SLC proteins.

Sci. Yamashiro and F.. J. It is likely that the integration of bioinformatics and both computational and in vitro models will drive our understanding of SLC proteins to new levels and afford an opportunity for further possible therapeutic targets and drug design opportunities. As with all models. Welch. Uzzau. 1980.. Lehr i n A. J. P. A. D. Pharmacol. Pharm. 1105 (1997). Hopkins. Wigge. 52. Peppard. R. M.626 (1998). C.Membrane Transport Proteins and Drug Transport there has been an explosion in interest in structure-activity relationships. Cell. 142. Bennett. Lipinski. Strous. G. Drug Design. P. Enrich. Kingham. Maxfield. 182. Pappenheimer. 15. Maxfield.235 (2000). J. p.. higher resolution screening systems.. Loehry. S127 (2001). Possibilities. Rowland. 37. 2124 (1985). J. J. Gastroenterology. DrugAbsorption Enhancement: Concepts. C. H. Biochem. M. Cell Physiol. L. Borchard. S. Erickson. Biol. Sean Ekins (Concurrent Pharmaceuticals. and ADME properties (298). Junginger. McMahon. Sci. Warhurst. 99. L.527 (1995). L. Madara and K. and H. ACS Symposium Series 14. 323. A. 13. 121. and E. F. 1. Rowland. 195.. New York. N. D. J. Washington. T . 759 (1997). however. and S. Truong. N . 439 (1980). h a d . Pharmacol. J. Ther. 14. 26. 108.. Harwood. J. Toxicol. Switzerland. Clin.. Borchardt. A. Proc. 23.. Walker. T . J. C388 (2001). S. K. USA. R. Fasano and S. 101. 92. . Physiol. and R. I. H. 7. H. Pediatr. J. Houston. Ghosh. van der Lubben. Y. Schwartz. Res. L. J. Cell Biol. Shah and W . J. 38. J. 1158 (1997). 2. Biol. 281. G. G. Whorwell. Albert. A. He. Exp. Adv. 27. Gut. Biol. 554 (1997). d..The resolution of the three-dimensional structures of solute transporter protein models is presumably low. and M.. P. Curr. Lemons. 17. Yalkowski and W . L. 9. J. 44. Invest. J.. Murby. S. Y . von Figura. J.93 (1994).. This has been seen largely as part of a wider growth in using computational approaches for understanding in vitro data. 738 (1983).. A. A. H. W . J. 18. E. Vallis. DC. particularly in the fields of drug design. R. p. R. Barsuhn. Ho. Chen. 17. 11. Stella in V . Ed. P. ACKNOWLEDGMENTS I thank Se-Ne Huang and Yongheng (Eric) Zhang for their contributions. 100. Verhoef. Porter and W .Physiol. Vol. J. D. 1475 (1961). Cell. Ed. Upshall. Whiteheart. B. 325. N. C. 12. D. 12 (1995). R. Mol.. C. Membr. Nelson. Pol. Ayrton. G. Sci. C.. 110. J. Nature. 10. Ariens. 29. Stella. J. Morozowich i n E. Geuze. G. Charman.435 (1997). Ortega. Methods.Y . 9. J. Watson. 195 (1984). 139 (2001). 1490 (1997). 24. A. J. and P. Slot. Drug Target. they can be justified by their ability to confirm biologically relevant phenomena.. J. Marsh and A. we can expect that a combination of indirect (3DQSAR) and direct (homology models) techniques may lead to newer. G. and J. A. Dharmsathaphorn. M. Eastmond. Sztul. S. Knipp. 22.354 (1976). Nature. Academic Press. REFERENCES J. W . Schwartz. M. Pharm. 19. and F. R.American Chemical Society. Shen. 25. J. Nguyen. Krugliak. Mukherjee. 16. A.. A. and C.. Adv. Bernstein. Rev. 26. Rev. D. p. and H. 90.421 (1958). Natl. L. 1994. Hollander. M. 86. F. K. the continued input of novel experimental data and revalidation of the model may eventually lead to highly predictive systems capable of in silico detection and design of novel solute transporter substrates. Ed. T. Ma. Am. Furthermore. T. Barroso. M. Prodrugs as Novel Drug Delivery Systems. D. Biochem. Bridges. Warhurst. C. Biol. D. Albertini. J. J. Herman and D. 77. Drug Deliv. 21. W . Blood. G. Dr. Cambridge.D. 67 (1975). Boer.. E. 50. optimization. 28. V . and C. B. 21. Madara and J. Hasilik. R. Helenius. Gifford. and A.835 (1995). Drug Deliv. 149 (1987). D.231 (1984). C. MA) contributed to the P-glycoprotein section in this chapter. L.. Cell. 7 8. 87. Rev. Cancer Res. 20. 1975. J. J.. Hughson and C.337 (1990). M. Limitations and Trends. and G. 304.

Virchows Arch. A. E. M. Hansen. Lehr. and R. Crennell. Weisser. 673 (1992). Zmmunol. 37. V. 13. J. Forestier. 613 (1997).1838 (1996). G. 27.. A.-M. Acad.. 89. Lazorova. U. Youngman. K. 59. 65. A. W. C. 67. Girbes. FEBS Lett.. Hamad. 296. Citores.. 68. Metcalfe.. Frutiger.651 (1982). Med. Lacey. V. 104. 62. Neutra. Wild. Y.410 (1992).23. M. Hussain. Di Martino. D. de Benito.. Indian J. Cell. J. J. 1 (1986). Sci. L. Desai. Ludert.. 1447 (1997). A. 68. 119. Marrack. B.1253 (1997). 49. 20. and R. Isberg. Microbiol. Prydz. R. L. and E. Zmmunol. M.53 (1997). P. J. Med. Geczy. 1358. Glenney. 64. and A.535 (1994). S. G. 40.360 (1997). M. L. N. R. Livrelli. T. R. G. Cohen. R. Bridges. 74. E. M. Heuser. J. J. Microbiol. F. Monier. Sloane. R. Ghetie. G. Gil. Lacey. Hum. Polito.675 (1994). Rousselet. 38 (1996). 58. and F. Ying. Thuillier. and T. G. Microbiol. S. Bioact. Res. Zmmun. Rodewald. R. J. J. P. W. Abrahamson and R.. Laver. Orci. Chem. C. 65. D. Clin. Biochem. V. Boquet. W.. S21 (1997). K. Y. Res. G. Michelangeli.J. A. Molin.. G. Higgins. Vimr. Heyman. R. 34. and B. and J. Levy. Proc. C. Sanderson. V. Poulain-Godefroy. 10 (1992). K. A. L. and S. Kaetzel. Garman. K.2347 (1995). Michetti. Bolognesi. A. 321 (1988). 31. Adlerberth. K. K. Cell Biol. K. Fiocchi. De Champs. E.495 (1979). P. Takashima. Znt. 60. K. Donzell. F. G. 91. Nature. 270 (1981). J. L. J.245 (1993). Paul. D. 36. J. Pharm. R. A. K. B. 664. 254 (1995). 108. vanderspek. Kurzchalia. Gasperi-Campani. Acad. and P. M.. J. Stirpe and L. Vet. Montesano. Wilson. 54. F. J. 7. S. C. Biochim. J. and B. 861 (1990). M. A. Mol. Desjeux. J . Isberg. G. H. Loomis.. Rodewald. Sci. P. G.-M. L. Stirpe. Natl. H. R. 185. Roberts. Environ. Biol. Zmmun. Fisher and J. S. Ghetie. Krogfelt. M. R. Eur. C. Labhasetwar. R. P. Roth. Pharm. 73 (1990). J. L. J. 62. 42. Zntervirology. G.259 (1992). Znfect. 72. Kappler. Buonamici. Sandvig. Acta. Matsue. LaFoe. 49 (1997). G. Cell Biol. 257. M. P. Robert. J. 39. G. Sandvig. Hughes. 331 (1993). D. Wold. F.6006 (1992). E. C. D. A. Ferreras. 1963 (1996). C. J. Biochim. L. 46. Chedid.265 (1987). C. H. A. FEBS Lett. S. M. Ghose. Pharm Res. S. C. N. Overell. E. A. Ward. J. 562. V. and L. and J. Olsnes. Hirt. A. J. Lehr. Severs. Buonamici. Rothberg. Ann.. 67. A. 1997. E. 427.. D. Radu. 70.. Lord. Lemichez. USA. G.. 45. P.321. M. 52. Leong. S. Kim. Stirpe. 56.. Barbieri. Florence. 8. Grose. Thorns. D. Isberg and J. J . Sperti. F.2533 (1996). H. J.10909 (1993). F. N. Structure. L. 73. and F. Anderson. P.. 346. Pihl. 234. 564 (1994). Zmmunol. FEBS Lett. V. 32. S. M. R. G . Hanson. 69. J. Rothberg. R. . Mestecky. E. J. Controlled Rel. Ahrne. S. 377 (1989). Science. K. J. and S. M. 7.References 30. 47. 2. 158.. and P. Biol. Amidon. Cell Sci. Kamen. F. Acad. Perregaux.. J. and M. Liprandi. 57. Acta. Anderson. and G. Morris. Drug Target. Devilliers. Sylvester. I. Scand. 35. and R. Pate1 and A. 255. R. 13271 (1982). Muter. Appl.. 71. Anderson. T. W. and J. 63. 2244 (1996). Biophys. M. Cell.. Proc. and E. Stirpe. J. Lukianov. J. Gene Ther.26.. Rothberg. Derenzini. Radu. A. F. S. Virol. R. 51. W. Artursson. and J. Wells. Laux. 529 (1996). H. S. 341 (1988). 41. 0. F. and S. and G. Olson and C. 43. Biophys. R. G. R. Walker and I. Medesan. J. N. Clin. 91 (1991). J. 66. M. Montanaro. J.445 (1997). J. Arch. 153.2211 (1997). I.. Symp. Stirpe. Easson and C. 33. Jeanguenat. Medesan. Darfeuille-Michaud. A. 50. Kamen. W. Matesoi. Ward. Olsnes. J. Berryman and R. J. Zmmunol. Toxicol. Cell Sci. A. 59. 55. E. 90. Znfect.. Exp.. Marra and R. 90. 14. Res. 44. 1. C. F. N.. Anderson. 95 (1987).. 71. Esparza. Neonate. R. Microbiol. C. Jani. 60. R.. Battelli. and D. Zmmunol. Proc. Atkinson. H. Sjolander. Med. 195. Sci. Krivan. Lindner. Russell-Jones. 53. 61. and C. 3412 (1997). J. 71. USA.. Easson and C. K. Joly. M. 38. K. and A. J. W. and R. Benlounes. E. Murphy. F. J. Linder. J. P. Johansson. A. L. C. and A.. Bomsel. 4042 (1997). 34. Dupree. J. Mol. Biol. Kraehenbuhl. C. van Deurs. Cell. 48. W.. RussellJones. L. 40. A. Battelli. S.88 (1994). Taylor. 74. E. Natl. 14..

Membrane Transport Proteins and Drug Transport

75. A. Irurzun, J. L. Nieva, and L. Carrasco,Virology, 227,488 (1997). 76. B. S. Worthington and J. Syrotuck, J. Nutr., 106,20 (1976). 77. A. Garret, C. Kerlan, and D. Thomas, Arch. Virol., 131,377 (1993). 78. J. L.Wolf, R. S. Kauffman,R. Finberg, R. Dambrauskas, B. N. Fields, and J . S. Trier, Gastroenterology, 85,291 (1983). 79. N. Etchart, F. Wild, and D. Kaiserlian, J. Gen. Virol., 77, 2471 (1996). 80. S. E. Bardocz, S. W . Grant, and G. Pasztai, in S. P. Bardocz, Ed., Lectins: Biomedical Perspectives, Taylor & Francis, London, 1995, p. 103. 81. C.-M. P. Lehr i n S. P. Bardocz, Ed., Lectins: Biomedical Perspectives, Taylor & Francis, London, 1995, p. 117. 82. L. T . Weaver and D. S. Bailey, J. Pediatr. Gastroenterol. Nutr., 6,445 (1987). 83. A. Gebert, H. J. Rothkotter, and R. Pabst, Znt. Rev. Cytol., 167,91 (1996). 84. M. A. Jepson, M. A. Clark, N. Foster, C. M . Mason, M . K. Bennett, N. L. Simmons, and B. H. Hirst, J. Anat., 189,507 (1996). 85. R. Sharma, E. J . van Damme, W . J. Peumans, P. Sarsfield, and U . Schumacher, Histochem. Cell. Biol., 105,459 (1996). 86. J. F. Ross, P. K. Chaudhuri, and M . Ratnam, Cancer, 73,2432 (1994). 87. J. Fan, N. Kureshy, K. S. Vitols, and F. M. Huennekens, Oncol. Res., 7,511 (1995). 88. C. P. Leamon and P. S. Low, Biochem. J., 291, 855 (1993). 89. K. A. Mislick, J . D. Baldeschwieler, J. F. Kayyem, and T . J. Meade, Bioconjug. Chem., 6, 512 (1995). 90. S. Wang, R. J. Lee, G. Cauchon, D. G. Gorenstein, and P. S. Low, Proc. Natl. Acad. Sci. USA, 92,3318 (1995). 91. 0. D. Wangensteen, M . M . Bartlett, J. K. James, Z. F. Yang, and P. S. Low, Pharm. Res., 13, 1861 (1996). 92. S. N. Huang and P. W . Swaan, J. Pharmacol. Exp. Ther., 298, 264 (2001). 93. S. N. Huang and P. W . Swaan, J. Pharmacol. Exp. Ther., 294, 117 (2000). 94. M. Ramasamy, D. H. Alpers, C. Tiruppathi, and B. Seetharam, Am. J. Physiol. Gastrointest. Liver Physiol., 257, G791 (1989). 95. N. Dan and D. F. Cutler, J. Biol. Chem., 269, 18849 (1994).

96. G. J. Russell-Jones i n M . D. A. Taylor, Ed., Peptide-Based Drug Design: Controlling Transport and Metabolism, American Chemical Society, Washington, DC, 1995, p. 181. 97. G. J. Russell-Jones, S. W . Westwood, P. G. Farnworth, J. K. Findlay, and H . G. Burger, Bioconjug. Chem., 6, 34 (1995). 98. A. D. Habberfield, K. Ralph, S. W . Westwood, and G. J. Russell-Jones, Znt. J. Pharm., 145,l (1996). 99. G. J. Russell-Jones, S. W . Westwood, and A. D. Habberfield, Bioconjug Chem., 6,459 (1995). 100. G. J. Russell-Jones, L. Killinger, and S. W . Westwood, Proc. Zntl. Symp. Controlled Rel. Bioact. Mater., 1997; 101. G. J . Russell-Jones, S. W . Westwood, and A. D. Habberfield, Proc. Znt. Symp. Controlled Rel. Bioact. Mater., 1996. 102. A. Pietrangelo, E. Rocchi, G. Casalgrandi, G. Rigo, A. Ferrari, M . Perini, E. Ventura, and G. Cairo, Gastroenterology, 102,802 (1992). 103. G. J. Anderson, L. W . Powell, and J. W . Halliday, Gastroenterology, 98,576 (1990). 104. G. P. Jeffrey, K. A. Basclain, and T . L. Allen, Gastroenterology, 110, 790 (1996). 105. D. Shah and W.-C. Shen, J. Pharm. Sci., 85, 1306 (1996). 106. J. S. Wang, D. Shah, and W.-C. Shen, Pharm. Res., 14, S469 (1997). 107. M. H. Saier Jr., Znt. Rev. Cytol., 190,61(1999). 108. W . Kramer and G. Wess, Eur. J. Clin. Invest., 26, 715 (1996). 109. M. B. Tollefson,W . F. Vernier, H . C. Huang, F. P. Chen, E. J. Reinhard, J. Beaudry, B. T . Keller, and D. B. Reitz, Bioorg. Med. Chem. Lett., 10, 277 (2000). 110. G. Wess, W . Kramer, A. Enhsen, H. Glombik, K. H. Baringhaus, G. Boger, M . Urmann, K. Bock, H. Kleine, G. Neckermann, A. Hoffmann, C. Pittius, E. Falk, H. W . Fehlhaber, H . Kogler, and M. Friedrich, J. Med. Chem., 37, 873 (1994). 111. M. A. Hediger, M. J. Coady, T . S. Ikeda, and E. M . Wright, Nature, 330,379 (1987). 112. M. A. Hediger, T . Ikeda, M. Coady, C. B. Gundersen, and E. M. Wright, Proc. Natl. Acad. Sci. USA, 84,2634 (1987). 113. M . Dean, A. Rzhetsky, and R. Allikmets, Genome Res., 11, 1156 (2001). 114. I. T . Paulsen, M . K. Sliwinski, B. Nelissen, A. Goffeau,and M . H. Saier Jr., FEBS Lett., 430, 116 (1998).

1

References

115. I. T . Paulsen, M. K. Sliwinski, and M . H. Saier Jr., J. Mol. Biol., 277,573 (1998). 116. G. D. Schuler, M. S. Boguski, E. A. Stewart, L. D. Stein, G. Gyapay, K. Rice, R. E.White, P. Rodriguez-Tome, A. Aggarwal, E. Bajorek, S. Bentolila, B. B. Birren, A. Butler, A. B. Castle, N. Chiannilkulchai, A. Chu, C. Clee, S. Cowles, P. J. Day, T . Dibling, N. Drouot, I. Dunham, S. Duprat, C. East, and T . J. Hudson, Science, 274,540 (1996). 117. P.W . Swam, S. Oie, and F. C. Szoka Jr., Adv. Drug Del. Rev., 20, 1 (1996). 118. S. A. Adibi, Gastroenterology, 113,332 (1997). 119. D. M. Bradshaw and R. J. Arceci, J. Clin. Oncol., 16,3674 (1998). 120. A. Krishan, C. M. Fitz, and I. Andritsch, Cytometry, 29,279 (1997). 121. W . 0. Berndt, Toxicol. Pathol., 26, 52 (1998). 122. M. J. Owens and C. B. Nemeroff,Depress. Anxiety, 8, 5 (1998). 123. M. E. Reith, C. Xu, and N . H. Chen, Eur. J. Pharmacol., 324, 1 (1997). 124. D. L. Baly and R. Horuk, Biochim. Biophys. Acta, 947, 571 (1988). 125. E. M. Wright, E. Turk, B. Zabel, S. Mundlos, and J . Dyer, J. Clin. Invest., 88,1435 (1991). 126. M . H. Wong, P. Oelkers, and P. A. Dawson, J. Biol. Chem., 270,27228 (1995). 127. J. Deisenhofer and H. Michel, Annu. Rev. Biophys. Biophys. Chem., 20,247 (1991). 128. J. Deisenhofer and H. Michel, Annu. Rev. Cell Biol., 7,1(1991). 129. N. Unwin, Nature, 373,37 (1995). 130. T . Tsukihara, H. Aoyama, E. Yarnashita, T . Tomizaki, H . Yamaguchi, K. Shinzawa-Itoh, R. Nakashima, R. Yaono, and S. Yoshikawa, Science, 272, 1136 (1996). 131. G. Chang and C. B. Roth, Science, 293, 1793 (2001). 132. H. R. Kaback, J. Voss, and J. W u , Curr. Opin. Struct. Biol., 7, 537 (1997). 133. V . C. Goswitz and R. J. Brooker, Protein Sci., 4, 534 (1995). 134. M. F. Varela and T . H. Wilson, Biochim. Biophys. Acta, 1276,21 (1996). 135. R. F. Doolittle, M. S. Johnson, I. Husain, B. V a n Houten, D. C. Thomas, and A. Sancar, Nature, 323,451 (1986). 136. J. Voss, J. W u ,W . L. Hubbell, V . Jacques, C. F. Meares, and H. R. Kaback, Biochemistry, 40, 3184 (2001).

137. M. G. Claros and G . von Heijne, Comput. Appl. Biosci., 10,685 (1994). 138. B. Rost, P. Fariselli, and R. Casadio, Protein Sci., 5,1704 (1996). 139. G. E. Tusnady and I. Simon, J. Mol. Biol., 283, 489 (1998). 140. K. Bamberg and G. Sachs, J. Biol. Chem., 269, 16909 (1994). 141. S. Frillingos, M. Sahin-Toth, J. W u , and H. R. Kaback, FASEB J., 12,1281 (1998). 142. E. Turk, C. J. Kerner, M. P. Lostao, and E. M. Wright, J. Biol. Chem., 271, 1925 (1996). 143. E. R. Bennett and B. I. Kanner, J. Biol. Chem., 272, 1203 (1997). 144. K. M. Covitz, G. L. Amidon, and W . Sadee, Biochemistry, 37, 15214 (1998). 145. Q. Wang, J. Voss, W . L. Hubbell, and H. R. Kaback, Biochemistry, 37,4910 (1998). 146. M. Zhao, K. C. Zen, W . L. Hubbell, and H. R. Kaback, Biochemistry, 38, 7407 (1999). 147. R. C. Hresko, H. Murata, B. A. Marshall, and M. Mueckler, J. Biol. Chem., 269, 32110 (1994). 148. K. Nagai and H. C. Thogersen, Nature, 309, 810 (1984). 149. M. Sahin-Toth, R. L. Dunten, and H. R. Kaback, Biochemistry, 34,1107 (1995). 150. C. Wandel, R. B. Kim, S. Kajiji, P. Guengerich, G. R. Wilkinson, and A. J . Wood, Cancer Res., 59, 3944 (1999). 151. A. George, Curr. Opin. Drug Discou. Dev., 2, 286 (1999). 152. S. Hoffmeyer, 0. Burk, 0. von Richter, H. P. Arnold, J. Brockrnoller, A. Johne, I. Cascorbi, T . Gerloff, I. Roots, M. Eichelbaum, and U. Brinkmann, Proc. Natl. Acad. Sci. USA, 97, 3473 (2000). 153. V. J. Wacher, C. Y . W u , and L. Z. Benet, Mol. Carcinog., 13, 129 (1995). 154. E. G. Schuetz, W . T . Beck, and J. D. Schuetz, Mol. Pharmacol., 49,311 (1996). 155. E. G. Schuetz, A. H. Schinkel, M. V . Relling, and J. D. Schuetz, Proc. Natl. Acad. Sci. USA, 93,4001 (1996). 156. R. B. Kim, C. Wandel, B. Leake, M. Cvetkovic, M. F. Fromm, P. J. Dempsey, M. M. Roden, F. Belas, A. K. Chaudhary, D. M. Roden, A. J. Wood, and G. R. Wilkinson, Pharm. Res., 16, 408 (1999). 157. E. G. Schuetz, S. Strom, K. Yasuda, V . Lecureur, M . Assem, C. Brimer, J. Lamba, R. B. Kim, V . Ramachandran, B. J. Komoroski, R.

Membrane Transport Proteins and Drug Transport

Venktaramanan, H. Cai, C. J. Sinal, F. J. Gonzalez, and J. D. Schuetz, J. Biol. Chem., 276, 39411 (2001). T . W . Synold, I. Dussault, and B. M. Forman, Nut. Med., 7, 584 (2001). S. Ekins and J. A. Erickson, Drug Metab. DisS. Ekins, B. J. Ring, S. N. Binkley, S. D. Hall, and S. A. Wrighton, Znt. J. Clin. Pharmacol. Ther., 36, 642 (1998). K. R. Korzekwa, N . Krishnamachary, M. Shou, A. Ogai, R. A. Parise, A. E. Rettie, F . J. Gonzalez, and T . S. Tracy, Biochemistry, 37, 4137 (1998). J . B. Houston and K. E. Kenworthy, Drug Metab. Dispos., 28,246 (2000). L. M. Greenberger, C. P. Yang, E. Gindin, and S. B. Horwitz, J. Biol. Chem., 265,4394 (1990). S. Ayesh, Y . M. Shao, and W . D. Stein, Biochim. Biophys. Acta, 1316,8 (1996). A. B. Shapiro and V . Ling, Eur. J. Biochem., 250, 130 (1997). S. Scala, N. Akhmed, U . S. Rao, K. Paull, L. B. Lan, B. Dickstein, J . S. Lee, G. H. Elgemeie, W . D. Stein, and S. E. Bates, Mol. Pharmacol., 51, 1024 (1997). S. Dey, M. Ramachandra, I. Pastan, M. M. Gottesman, and S. V. Ambudkar, Proc. Natl. Acad. Sci. USA, 94, 10594 (1997). A. B. Shapiro, K. Fox, P. Lam, and V. Ling, Eur. J. Biochem., 259,841 (1999). E. J. Wang, C. N. Casciano, R. P. Clement, and W . W . Johnson, Biochim. Biophys. Acta, 1481, 63 (2000). L. Lu, F. Leonessa, R. Clarke, and I . W . Wainer, Mol. Pharmacol., 59, 62 (2001). M. Demeule, A. Laplante, G. F. Murphy, R. M. Wenger, and R. Beliveau, Biochemistry, 37, 18110 (1998). R. L. Shepard, M. A. Winter, S. C. Hsaio, H. L. Pearce, W . T . Beck, and A. H. Dantzig, Biochem. Pharmacol., 56, 719 (1998). A. H. Dantzig, R. L. Shepard, J. Cao, K. L. Law, W . J. Ehlhardt, T . M. Baughman, T . F. Bumol, and J. J. Starling, CancerRes., 56,4171 (1996). H. L. Pearce, M. A. Winter, and W . T . Beck, Adu. Enzyme Regul., 30, 357 (1990). H. L. Pearce, A. R. Safa, N. J. Bach, M. A. Winter, M. C. Cirtain, and W . T . Beck, Proc. Natl. Acad. Sci. USA, 86,5128 (1989). A. Ramu and N. h u , Cancer Chemother. Pharmacol., 30, 165 (1992).

M. J. Borgnia, G. D. Eytan, and Y . G. Assaraf, J. Biol. Chem., 271,3163 (1996). P. Chiba, G. Ecker, D. Schmid, J. Drach, B. Tell, S. Goldenberg, and V . Gekeler, Mol. Pharmacol., 49, 1122 (1996). G. Klopman, L. M. Shi, and A. h u , Mol. Pharmacol., 52, 323 (1997). G. A. Bakken and P. C. Jurs, J. Med. Chem., 43,4534 (2000). M. Wiese and I. K. Pajeva, Pharmazie, 52,679 (1997). C. Tmej, P. Chiba, M. Huber, E. Richter, M . Hitzler, K. J. Schaper, and G. Ecker, Arch. Pharm. (Weinheim), 331,233 (1998). I. Pajeva and M. Wiese, J. Med. Chem., 41, 1815 (1998). D. Schmid, G. Ecker, S. Kopp, M. Hitzler, and P. Chiba, Biochem. Pharmacol., 58, 1447 (1999). G. Ecker, M. Huber, D. Schmid, and P. Chiba, Mol. Pharmacol., 56, 791 (1999). A. Seelig, Znt. J. Clin. Pharmacol. Ther., 36,50 (1998). T . Osterberg and U . Norinder, Eur. J. Pharm. Sci., 10, 295 (2000). S. Neuhoff, P. Langguth, C. Dressler, T . B. Anderson, C. G. Regardh, and H. SpahnLangguth, Znt. J. Clin. Pharmacol. Ther., 38, 168 (2000). M. F. Rosenberg, Q. Mao, A. Holzenburg, R: C. Ford, R. G. Deeley, and S. P. Cole, J. Biol. Chem., 276,16076 (2001). G. K. Grimble a n d D. B. A. Silk, Nutr. Res. Rev., 2,87 (1989). D. M. Mathews a n d S. A. Adibi, Gastroenterology, 71, 151 (1976). S. A. Adibi, Metabolism, 36, 1001 (1987). P. Furst, S. Albers, and P. Stehle, Proc. Nutr. SOC., 49,343 (1990). G. K. Grimble, Annu. Reu. Nutr., 14, 419 (1994). N. J. Snyder, L. B. Tabas, D. M. Berry, D. C. Duckworth, D. 0.Spry, and A. H. Dantzig,Antimicrob. Agents Chemother., 41, 1649 (1997). H. Jiang, L. B. Tabas, and A. H. Dantzig, submitted. V . A. Moore, W . J. Irwin, P. Timmins, P. A. Lambert, S. Chong, S. A. Dando, and R. A. Morrison, Znt. J. Pharm., 210,29 (2000). F. H. Leibach and V. Ganapathy, Annu. Rev. Nutr., 16,99 (1996).

1 . b

!.

1

References

199. H.K. Han, J. K. Rhie, D. M. Oh, G. Saito, C. P. Hsu, B. H. Stewart, and G. L. Amidon, J.Pharm. Sci., 88,347(1999). 2 0 0 . K. M. Covitz, G. L. Amidon, and W. Sadee, Pharm. Res., 13,1631(1996). 201. J. S. Kim, R. L. Oberle, D. A. Krummel, J. B. Dressman, and D. Fleisher, J. Pharm. Sci., 83, 1350 (1994). 202. J. P. Bai and G. L. Amidon, Pharm. Res., 9,969 (1992). 203. M. Hu, P. Subramanian, H. I. Mosberg, and G. L. Amidon, Pharm. Res., 6,66(1989). 204. D. M. Oh, H. K. Han, and G. L. Amidon, Pharm. Biotechnol., 12,59(1999). 205. M. Boll, D.Markovich, W. M. Weber, H. Korte, H. Daniel, and H. Murer, Pfliigers Arch., 429, 146(1994). 206. M.E. Ganapathy, M. Brandsch, P. D. Prasad, V. Ganapathy, and F. H. Leibach, J. Biol. Chem., 270,25672 (1995). 207. M.E.Ganapathy, W. Huang, H. Wang, V. Ganapathy, and F. H. Leibach, Biochem. Biophys. Res. Commun., 246,470(1998). 2 0 8 . H. Shen, D. E. Smith, T. Yang, Y. G. Huang, J. B. Schnermann, and F. C. Brosius 3rd, Am. J. Physiol. Renal Fluid Electrolyte Physiol., 276,F658 (1999). 209. H. Ogihara, T. Suzuki, Y. Nagamachi, K. Inui, and K. Takata, Histochem. J.,31,169(1999). 210. H. Y. Steiner, F. Naider, and J. M. Becker, Mol. Microbiol., 16,825 (1995). 211. D. Meredith and C. A. Boyd, Cell. Mol. Life Sci., 57,754(2000). 212. Y. J . Fei, V. Ganapathy, and F. H. Leibach, Prog. Nucleic Acid Res. Mol. Biol., 58, 239 (1998). 213. S. Nussberger, A. Steel, D. Trotti, M. F. Romero, W. F. Boron, and M. A. Hediger, J. Biol. Chem., 272,7777(1997). 214. P. W. Swaan and J. J. Tukker, Pharm. Weekbl. Sci. Ed., 14F,62(1992). 215. P.W. Swaan and J. J. Tukker, Pharm. Weekbl. Sci. Ed., 14M,4 (1992). 216. J. J. Tukker and P. W. Swaan, Pharm. Res., 9, S180 (1992). 217. C. Humblet and G. R. Marshall, Ann. Rep. Med. Chem., 15,267(1980). 218. J. Li and I. J. Hidalgo, J. Drug Target., 4,9 (1996). 219. P. W. Swam and J. J. Tukker, J. Pharm. Sci., 86,596(1997). 220. P. W. Swaan, M. C. Stehouwer, and J. J. Tukker, Biochim. Biophys. Acta, 1236,31(1995).

221. S. H. Kubo and R. J. Cody, Clin. Pharmacokinet., 10,377(1985). 222. P. W. Swaan, B. C. Koops, E. E. Moret, and J . J. Tukker, Receptors Channels, 6,189(1998). 223. F. Doring, J. Will, S. Amasheh, W. Clauss, H. Ahlbrecht, and H. Daniel, J. Biol. Chem., 273, 23211(1998). 224. T. Terada, K. Sawada, M. Irie, H. Saito, Y. Hashimoto, and K. Inui, Pfliigers Arch., 440, 679(2000). 225. P. D. Bailey, C. A. R. Boyd, J. R. Bronk, I. D. Collier, D. Meredith, K. M. Morgan, and C. S. Temple, Angew. Chem. Znt. Ed. Engl., 39,506 (2000). 226. B. Bretschneider, M. Brandsch, and R. Neubert, Pharm. Res., 16,55(1999). 227. M. B. Bolger, I. S. Haworth, A. K. Yeung, D. Ann, H. von Grafenstein, S. Hamm-Alvarez, C. T. Okamoto, K. J . Kim, S. K. Basu, S. Wu, and V. H. Lee, J. Pharm. Sci., 87,1286 (1998). 228. X.Z.Chen, A. Steel, and M. A. Hediger, Biochem. Biophys. Res. Commun., 272, 726 (2000). 229. H. Buchwald, D. K. Stoller, C. T. Campos, J. P. Matts, and R. L. Varco, Ann. Surg., 212, 318 (1990). 230. M. Ast and W. H. Frishman, J. Clin. Pharmacol., 30,99 (1990). 231. K. Takashima, T. Kohno, T. Mori, A. Ohtani, K. Hirakoso, and S. Takeyama, Atherosclerosis, 107,247(1994). 232. M. D. Carey, D. M. Small, and C. M. Bliss, Annu. Rev. Physiol., 45,651(1983). 233. F. A.Wilson in R. A. Frizzell, Ed., The Gastrointestinal System, Vol. IV,American Physiological Society, Bethesda, MD, 1991,p. 389. 234. R. H. Dowling, Gastroenterology, 62, 122 (1972). 235. A. F. Hofmann, Adu. Intern. Med., 21, 501 (1976). 236. S. Ewerth, Acta Chir. Scand.Suppl., 513, 1 (1982). 237. R.H. Erlinger in L. R. Johnson, Ed., Physiology of the Gastrointestinal Tract, 2nd ed., Raven Press, New York, 1987,p. 1557. 238. C. D. Klaassen, Toxicol. Pathol., 16, 130 (1988). 239. A. F. Hofmann in J. Forte, Ed., The Gastrointestinal System, Vol. 111,American Physiological Society, Bethesda, MD, 1989,p. 567. 240. S. Stengelin, S. Apel, W. Becker, M. Maier, J. Rosenberger, U. Bewersdorf, F. Girbig, C.

Membrane Transport Proteins and Drug Transport

Weyland, G. Wess, and W. Kramer, Eur. J.Biochem., 239,887 (1996). A. D. Vodenlich, Y.-Z. Gong, K. F. Geoghegan, M. C. Lin, A. J. Lanzetti, and F. A. Wilson, Biochem. Biophys. Res. Commun., 177, 1147 (1991). S. Iseki, 0. Amano, T. Kanda, H. Fujii, and T. Ono, Mol. Cell. Biochem., 123, 113 (1993). W. Kramer, D. Corsiero, M. Friedrich, F. Girbig, S. Stengelin, and C. Weyland, Biochem. J., 333, 335 (1998). M. C. Lin, W. Kramer, and F. A. Wilson, J. Biol. Chem., 265, 14986 (1990). M. C. Lin, E. Mullady, and F. A. Wilson, Am. J. Physiol. Gastrointest. Liver Physiol., 265, G56 (1993). J. Grober, I. Zaghini, H. Fujii, S. A. Jones, S. A. Kliewer, T. M. Willson, T. Ono, and P. Besnard, J. Biol. Chem., 274,29749 (1999). J. Y. Chiang, R. Kimmel, C. Weinberger, and D. Stroup, J. Biol. Chem., 275, 10918 (2000). M. Crestani, A. Sadeghpour, D. Stroup, G. Galli, and J. Y. Chiang, J. Lipid Res., 39,2192 (1998). J . M. Lehmann, S. A. Kliewer, L. B. Moore, T. A. Smith-Oliver, B. B. Oliver, J. L. Su, S. S. Sundseth, D. A. Winegar, D. E. Blanchard, T. A. Spencer, and T. M. Willson, J. Biol. Chem., 272,3137 (1997). 250. B. A. Janowski, M. J. Grogan, S. A. Jones, G. B. Wisely, S. A. Kliewer, E. J. Corey, and D. J. Mangelsdorf, Proc. Natl. Acad. Sci. USA, 96, 266 (1999). 251. T. A. Spencer, D. Li, J. S. Russel, J. L. Collins, R. K. Bledsoe, T. G. Consler, L. B. Moore, C. M. Galardi, D. D. McKee, J. T. Moore, M. A. Watson, D. J. Parks, M. H. Lambert, and T. M. Willson, J. Med. Chem., 44,886 (2001). 252. K. N. Lazaridis, L. Pham, P. Tietz, R. A. Marinelli, P. C. deGroen, S. Levine, P. A. Dawson, and N. F. LaRusso, J. Clin. Invest., 100, 2714 (1997). 253. S. Sorscher, J. Lillienau, J. L. Meinkoth, J. H. Steinbach, C. D. Schteingart, J. Feramisco, and A. F. Hofmann, Biochem. Biophys. Res. Commun., 186,1455 (1992). 254. W. Kramer, G. Burckhardt, F. A. Wilson, and G. Kurz, J. Biol. Chem., 258,3623 (1983). 255. M. C. Lin, S. L. Weinberg, W. Kramer, G. Burckhardt, and F. A. Wilson, J.Membr. Biol., 106, l(1988). 256. G. Burckhardt, W. Kramer, G. Kurz, and F. A. Wilson, J. Biol. Chem., 258,3618 (1983).

257. B. L. Shneider and M. S. Moyer, J. Biol. Chem., 268, 6985 (1993). 258. B. L. Shneider, P. A. Dawson, D.-M. Christie, W. Hardikar, M. H. Wong, and F. J. Suchy, J. Clin. Invest., 95, 745 (1995). 259. S. A. Weinman, M. W. Carruth, and P. A. Dawson, J. Biol. Chem., 273,34691 (1998). 260. S. Hallen, M. Branden, P. A. Dawson, and G. Sachs, Biochemistry, 38, 11379 (1999).

261. C. J. Sippel, P. A. Dawson, T. Shen, and D. H. Perlmutter, J. Biol. Chem., 272,18290 (1997).

262. M. H. Wong, P. Oelkers, A. L. Craddock, and P. A. Dawson, J.Biol. Chem., 269,1340 (1994). 263. G. von Heijne and Y. Gavel, Eur. J. Biochem., 174,671 (1988). 264. J. Higaki, S. Hara, N. Takasu, K. Tonda, K. Miyata, T. Shike, K. Nagata, and T. Mizui, Arterioscler. Thromb. Vasc. Biol., 18, 1304 (1998). 265. T. Ichihashi, M. Izawa, K. Miyata, T. Mizui, K. Hirano, and Y. Takagishi, J. Pharmacol. Exp. Ther., 284,43 (1998). 266. M. W. Love and P. A. Dawson, Curr. Opin. Lipidol., 9,225 (1998). 267. A. F. Hofmann,Arch. Intern. Med., 159,2647 (1999). 268. L. Lack, Environ. Health Perspect., 33, 79 (1979).

269. M. P. Tyor, J. T. Garbutt, and L. Lack,Am. J. Med., 51, 614 (1971). 270. S. Walker, A. Stiehl, R. Raedsch, P. Kloters, and B. Kommerell, Digestion, 32,47 (1985). 271. R. Aldini, A. Roda, P. L. Lenzi, G. Ussia, M. C. Vaccari, G. Mazzella, D. Festi, F. Bazzoli, G. l . ,Eur. J. Clin. InGalletti, S. Casanova, et a vest., 22, 744 (1992). 272. E. R. Schiff, N. C. Small, and J. M. Dietschy, J. Clin. Invest., 51, 1351 (1972). 273. K. W. Heaton and L. Lack, Am. J. Physiol., 214,585 (1968). 274. L. Lack and I. M. Weiner, Am. J.Physiol., 210, 1142 (1966). 275. E. Krag and S. F. Phillips, J. Clin. Invest., 53, 1686 (1974). 276. K. Gallagher, J. Mauskopf, J. T. Walker, and L. Lack, J. Lipid Res., 17, 572 (1976).

277. L. Lack, A. Tantawi, C. Halevy, and D. Rockett, Am J Physiol. Gastrointest. Liver Physiol., 246, G745 (1984). 278. T. S. Low-Beer, M. P. Tyor, and L. Lack, Gastroenterology, 56, 721 (1969).

References

279. G. Wess, W. Kramer, W. Bartmann, A. Enhsen, H. Glombik, S. Miillner, K. Bock, A. Dries, H. Kleine, and W. Schmitt, Tetrahedron Lett., 33, 195 (1992). 280. L. Lack, J. T. Walker, and G. D. Singletary, Am. J. Physiol., 219,487 (1970). 281. K. H. Baringhaus, H. Matter, S. Stengelin, and W. Kramer, J. Lipid Res., 40,2158 (1999). 282. P. W. Swaan, F. C. Szoka Jr., and S. Oie, J. Cornput.-Aided MoL Des., 11, 581 (1997). 283. W. Kramer, S. Stengelin, K. H. Baringhaus, A. Enhsen, H. Heuer, W. Becker, D. Corsiero, F. Girbig, R. Noll, and C. Weyland, J. Lipid Res., 40, 1604 (1999). 284. F. Helsper and P. W. Swaan, submitted. 285. N. F. H. Ho, Ann. N. Y. Acad. Sci., 507, 315 (1987). 286. P. W. Swaan, K. M. Hillgren, F. C. Szoka Jr., and S. Oie, Bioconjug. Chem., 8, 520 (1997). 287. W. Kramer, G. Wess, G. Schubert, M. Bickel, F. Girbig, U. Gutjahr, S. Kowalewski, K.-H. Baringhaus, A. Enhsen, H. Glombik, S. Miillner, G. Neckermann, S. Schulz, and E. Petzinger, J. Biol. Chem., 267, 18598 (1992). 288. W. Kramer, G. Wess, A. Enhsen, K. Bock, E. Falk,A. Hoffmann, G. Neckermann, D. Gantz, S. Schulz, L. Nickau, et al., Biochim. Biophys. Acta, 1227, 137 (1994).

289. G. Wess, W. Kramer, X. B. Han, K. Bock, A. Enhsen, H. Glombik, K. H. Baringhaus, G. Boger, M. Urmann, A. Hoffmann, and E. Falk, J. Med. Chem., 37, 3240 (1994). 290. Z. F. Stephan, E. C. Yurachek, R. Sharif, J. M. Wasvary, R. Steele, and C. Howes, Biochem. Pharmacol., 43, 1969 (1992). 291. D.-C. Kim, A. W. Harrison, M. J. Ruwart, K. F. Wilkinson, J. F. Fisher, I. J. Hidalgo, and R. T. Borchardt, J. Drug Target, 1,347 (1993). 292. C. 0. Mills, G. H. Martin, and E. Elias, Biochim. Biophys. Acta, 876,677 (1986). 293. M. S. Anwer, E. R. 07Maille,A. F. Hofmann, R. A. DiPietro, and E. Michelotti, Am. J. Physiol. Gastrointest. Liver Physiol., 249, G479 (1985). 294. C. 0. Mills, S. Iqbal, and E. Elias, J. Hepatol., 1,199 (1985). 295. W. F. Caspary and W. Creutzfeldt, Diabetologia, 11, 113 (1975). 296. C. Root, C. D. Smitth, D. A. Winegar, L. E. Brieaddy, and M. C. Lewis, J. Lipid Res., 36, 1106 (1995). 297. M. C. Lewis, L. E. Brieaddy, and C. Root, J. Lipid Res., 36, 1098 (1995). 298. S. Ekins, C. L. Waller, P. W. Swaan, G. Cruciani, S. A. Wrighton, and J. H. Wikel, J. Pharmacol. Toxicol. Methods, 44,251 (2000).

CHAPTER NINE
b

Allosteric Proteins and Drug Discovery
J. ELLIS BELL
Department of Chemistry Gottwald Science Center University of Richmond Richmond, Virginia

!

JAMES C. BURNETT
School of Pharmacy and Department of Medicinal Chemistry Institute for Structural Biology and Drug Discovery Virginia Commonwealth University Richmond, Virginia

JESSICA K. BELL
Department of Pharmaceutical Chemistry University of California a t San Francisco San Francisco, California

PETER S. GALATIN DONALD J. ABRAHAM
School of Pharmacy and Department of Medicinal Chemistry Institute for Structural Biology and Drug Discovery Virginia Commonwealth University Richmond, Virginia

Contents
1 Introduction, 296 2 Allosteric Theory and Models, 296 2.1 The Monod-Wyman-Changeux and Koshland-Nemethy-Fimerl Dalziel-Engel Models, 300 2.1.1 Monod-Wyman-Changeaux: The Concerted Model, 300 2.1.2 Koshland-Nemethy-Filmer/Dalziel-Engel: The Sequential Model, 301 2.2 Hemoglobin: Classic Example of Allostery, 301 2.3 Common Features of Allosteric Proteins, 304 2.4 Understanding Allosteric Transitions at the Molecular Level, 305 2.5 The Molecular Mechanisms of Allosteric Changes, 306
295

Burger's Medicinal Chemistry and Drug Discovery Sixth Edition, Volume 2: Drug Development Edited by Donald J. Abraham ISBN 0-471-37028-2 O 2003 John Wiley & Sons, Inc.

Allosteric Proteins and Drug Discovery 2.6 The Basis for Allosteric Effectors as Potential Therapeutic Agents, 312
3 Allosteric Targets for Drug Discovery and Future Trends, 313

1

INTRODUCTION

2

ALLOSTERIC THEORY A N D M O D E L S

The concept of allostery for the regulation of protein function has been known for approximately half a century. However, medicinal chemists have only recently been successful in developing allosteric effectors as therapeutic agents. Allosteric modifiers do not act as agonists or antagonists; instead they regulate substrate binding or release, or catalysis. Allosteric proteins regulate some of the most important biochemical pathways in living systems. Allosteric effectors represent a new mechanistic approach in drug therapy. Pharmacologists have explored similar and closely related areas using different terminology such as inverse agonists. The purpose of this chapter is to integrate the underpinnings grounded in structural biology and mechanistic biochemistry with drug discovery. Because an increasing number of human illnesses are now know to result from mutations affecting the it allosteric regulation of various enzymes (I), is safe to assume that a number of new drug entities will emerge in the near future that modulate allosteric interactions. The major barrier to more rapid discovery of allosteric effectors has been the lack of detailed structural information about the different allosteric states that an allosteric protein assumes. Until recently, hemoglobin represented the only well-defined allosteric protein with a plethora of structural and mechanistic studies. However, those involved in unraveling the mechanistic details of the allosteric transition still argue over what triggers it, and what is the order of tertiary and quaternary conformational changes. Despite the difficulty in obtaining structural information for each allosteric state in any given protein or receptor, drug discovery with allosteric systems is well under way and bearing fruit. The purpose of this chapter is to review the underlying principles of allosteric proteins and how these may be used to design "allosteric" drugs.

Enzymes function as monomers, multidomained protomers, or complex oligomeric species. They bind substrates, cofactors, and/or effectors, and stabilize transition states to increase the rate of catalysis and release product. Aside from regulation arising at the transcriptional, translational, or posttranslational levels, the control of an enzymatic reaction relies on the enzyme's ability to regulate substrate binding or product release (so called Ktype allosteric effects), or the rate of catalysis (so-called V-type allosteric effects) in response to cellular and whole organism changes. Most proteins exhibit hyperbolic saturation curves; however, a number of important biological processes involve multi-subunit proteins whose overall activities are regulated by allosteric effects. These proteins exhibit saturation curves that deviate from the normal hyperbolic curve (Fig. 9.1), and may involve effects where the activity increases disproportionately with saturation [a so-called positive cooperativity, giving a sigmoidal saturation curve (Fig. 9.1)] or decreases relatively as saturation is increased (a so-called negative cooperativity), giving rise to a saturation curve (Fig. 9.1) that is proportionally too steep below half saturation, and less steep relative to a normal saturation curve. Above half saturation a further deviation from a normal (hyperbolic) saturation curve (Fig. 9.2 a) that is often seen is the phenomenon of substrate inhibition, which also (among other alternatives) can have its root in allosteric interactions. In a Lineweaver-Burk plot, substrate inhibition results in