Smith and Williams' Introduction to the Principles of Drug Design and Action 3rd Edition

Smith and Williams’

Introduction to the Principles of Drug Design and Action Third edition

This page intentionally left blank.

Smith and Williams’

Introduction to the Principles of Drug Design and Action Third edition Edited by

H.John Smith Welsh School of Pharmacy University of Wales Cardiff, UK

harwood academic publishers Australia • Canada • China • France • Germany India • Japan • Luxembourg • Malaysia The Netherlands • Russia • Singapore Switzerland • Thailand

Copyright © 1998 OPA (Overseas Publishers Association) Amsterdam B.V. Published under license under the Harwood Academic Publishers imprint, part of The Gordon and Breach Publishing Group. All rights reserved. First Edition published 1983 This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledges’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” Second Edition published 1988 No part of this book may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying and recording, or by any information storage or retrieval system, without permission in writing from the publisher. Printed in Singapore. Amsteldijk 166 1st Floor 1079 LH Amsterdam The Netherlands British Library Cataloguing in Publication Data Smith and Williams’ introduction to the principles of drug design and action.—3rd ed. 1. Drugs—Design 2. Pharmacology I. Smith, H.J. (Harold John), 1930—II. Williams, Hywel III. Introduction to the principles of drug design and action 615.1 ISBN 0-203-30415-2 Master e-book ISBN

ISBN 0-203-34407-3 (Adobe eReader Format) ISBN 90-5702-037-8 (hard cover) Front cover: A model of the active site of aromatase with the substrate audiosteredione (yellow) as described by C.H.Laughton et al. (1993) Journal of Steroid Biochemistry and Molecular Biology 44, 399–407.

CONTENTS

Chapter 1 Chapter 2 Chapter 3 Chapter 4 Chapter 5

Chapter 6

Chapter 7 Chapter 8 Chapter 9 Chapter 10

Chapter 11 Chapter 12

Preface

vi

List of Contributors

vii

Abbreviations

ix

Processes of Drug Handling by the Body D K Luscombe and P J Nicholls

1

The Design of Drug Delivery Systems I W Kellaway

32

Intermolecular Forces and Molecular Modelling R H Davies and D Timms

60

Drug Chirality and its Pharmacological Consequences A J Hutt

121

Quantitative Structure-Activity Relationships and Drug Design J C Dearden and F K C James

202

From Programme Sanction to Clinical Trials: A Partial View of the Quest for Arimidex™, a Potent, Selective Inhibitor of Aromatase P N Edwards

253

Pro-Drugs A W Lloyd and H J Smith

285

Design of Enzyme Inhibitors as Drugs A Patel, H J Smith and J Stürzebecher

316

The Chemotherapy of Cancer D E Thurston and S G M J Lobo

403

Neurotransmitters, Agonists and Antagonists R D E Sewell, R A Glennon, M Dukat, H Stark, W Schunack and P G Strange

469

Design of Antimicrobial Chemotherapeutic Agents E G M Power and A D Russell

530

Recombinant DNA Technology: Monoclonal Antibodies F J Rowell and J R Furr

599

Chapter 13

Bio-inorganic Chemistry and its Pharmaceutical Applications D M Taylor and D R Williams

620

Index

655

PREFACE The second edition of Introduction to the Principles of Drug Design was published in 1988. In the intervening years considerable strides have been made in the approaches to rational drug design as the result of the flood of knowledge coming from advances made in molecular biology. This has provided a better understanding of biological systems in terms of their structural components, cellular signalling, genomic modulation etc., leading to a more informed approach to chemotherapeutic intervention in disease. In the third edition the aims and objectives, as well as the intended reading audience, remain the same as in previous editions but all the chapters have been revised to take into account of new developments in their subject areas and three new chapters have been included. Chapter 4 dealing with Drug Chirality and its Pharmacological Consequences reviews an ongoing field of considerable importance to pharmacologists and especially industrial concerns in view of the recent requirements imposed by Regulatory Bodies regarding drug registration. Chapter 6 provides a fascinating account of the difficulties inherent in the development of a drug from the bench to the clinic and brings out the trials and tribulations encountered by the multi-disciplinary research teams involved. Chapter 10 on Neurotransmitters, Agonists and Antagonists compensates to some extent for an area neglected in previous editions, that is, the design of drugs for action on the central nervous system, and also provides an account of membrane-bound receptors perhaps overshadowed in previous editions by emphasis on enzyme and DNA related targets. Chapter 3 on Intermolecular Forces and Molecular Modelling has required expansion and revision due to advances in the techniques relating to ligand-receptor interactions and we are indebted to Zeneca, through Dr M.T.Cox, for their generosity in meeting the considerable cost of reproducing the necessary new colour plates in the book. We also wish to thank Dr Charlie Laughton of the School of Pharmacy, Nottingham University for providing the illustration on the front cover of the book.

LIST OF CONTRIBUTORS Chapter 1 Professor David K Luscombe and Professor Paul J Nicholls Welsh School of Pharmacy, University of Wales Cardiff, Cardiff CF1 3XF, UK Chapter 2 Professor Ian W Kellaway Welsh School of Pharmacy, University of Wales Cardiff, Cardiff CF1 3XF, UK Chapter 3 Dr Robin H Davies Welsh School of Pharmacy, University of Wales Cardiff, Cardiff CF1 3XF, UK Dr David Timms Zeneca Pharmaceutical, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK Chapter 4 Dr Andrew J Hutt School of Pharmacy, Kings College, University of London, Manresa Road, London SW3 6LX, UK Chapter 5 Professor John C Dearden School of Pharmacy, John Moores University, Byrom Street, Liverpool, L3 3AF, UK †Dr Kenneth C James Welsh School of Pharmacy, University of Wales Cardiff, Cardiff CF1 3XF, UK Chapter 6 Dr Philip N Edwards Zeneca Pharmaceuticals, CAM Department, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK Chapter 7 Dr Andrew W Lloyd Department of Pharmacy, University of Brighton, Cockcroft Building, Moulescoombe, Brighton BN2 4GJ, UK Dr H John Smith Welsh School of Pharmacy, University of Wales Cardiff, Cardiff CF1 3XF, UK Chapter 8 Dr Anjana Patel The Royal Pharmaceutical Society, Lambeth High Street, London SE1 7JN, UK

Dr H John Smith Welsh School of Pharmacy, University of Wales Cardiff, Cardiff CF1 3XF, UK Dr Jörg Stürzebecher Klinikum der Universität Jena, Zentrum für vasculäre Biologie und Medizin, Institut für Biochemie und Moleckularbiologie, Nordhäuser Strasse 78, D-99089 Erfurt, Germany Chapter 9 Professor David E Thurston and Dr Sylvia G M Lobo School of Pharmacy and Biomedical Sciences, University of Portsmouth, Park Building, King Henry I Street, Portsmouth PO1 2DZ, UK Chapter 10 Dr Robert D E Sewell Welsh School of Pharmacy, University of Wales Cardiff, Cardiff CF1 3XF, UK Professor Richard A Glennon and Dr Malgorzata Dukat Department of Medicinal Chemistry, School of Pharmacy, Medical College of Virginia, Virginia Commonwealth University, Richmond, Virginia 23298–0540 USA Dr Holger Stark and Professor Walter Schunack Freie Universität Berlin, Institut für Pharmazie1, Königin-Luise-Strasse 2+4, D-14195 Berlin, Germany Philip G Strange Research School of Biosciences, The University, Canterbury CT2 7NJ, UK Chapter 11 Dr Edward G M Power Department of Microbiology, United Medical and Dental Schools, Guy’s Hospital, London Bridge, London SE1 9RT, UK Professor A Denver Russell Welsh School of Pharmacy, University of Wales Cardiff, Cardiff CF1 3XF, UK Chapter 12 Professor Frederick J Rowell School of Health Sciences, University of Sunderland, Pasteur Building, Sunderland SR1 3SD, UK Dr James R Furr Welsh School of Pharmacy, University of Wales Cardiff, Cardiff CF1 3XF, UK Chapter 13 Professor David M Taylor and Professor David R Williams Chemistry Department, University of Wales Cardiff, Cardiff CF1 3XF, UK

ABBREVIATIONS mχ f π Ø σ σ* µ µ° AADC 7-ACA ACE ADEPT AG AGP AIDS AMP ANP ATP 6-APA 2-APAs APN Ara-A ARI ATP AUC AZT BiCNU BHT c-AMP 5-CAT CCNU cDNA CFCs ChAT Cmax CMC

molecular connectivities hydrophobic constant hydrophobic substituent constant degree of freedom Hammett substituent constant Taft’s substituent constant dipole moment standard partial free energy aromatic amino acid decarboxylase 7-aminocephalosporanic acid angiotensin I—converting enzyme antibody—directed enzyme prodrug therapy aminoglutethimide α1—acid glycoprotein autoimmune deficiency syndrome adenosine monophosphate atrial natriuretic peptide adenosine triphosphate 6-aminopenicillanic acid 2-arylpropionic acids aminopeptidase N adenosine arabinoside, vidaribine aromatase inhibition adenosine triphosphate area under plasma concentration vs time curve azidothymidine, zidovudine carmustine butylated hydroxytoluene adenosine 3’,5’-cyclic phosphate 5-carboxamidotryptamine Iomustine copy DNA chlorofluorocarbons acetyl-CoA: choline O—transferase maximum plasma concentration critical micelle concentration

CME CMV CNBr CNS CoMFA COMT CSCC CT D4T Da DAG ddC ddI DHFR DHT DICR DMBA DMSO DOX DPI EAQ EBV ED50 EDTA Es EI EIS ER Fab FSH 5-FU GABA GABA-T GC G-CSF GDPT GI GLcNAc GM-CSF GSH GTP

1-cyano-1-methyl-ethyl group cytomegalovirus cyanogen bromide central nervous system comparative molecular field analysis catecholamine methyltransferase side chain cleavage enzyme charge transfer 2’,3’-didehydro-3’-deoxythymidine dalton diacylglycerol 2’,3’-dideoxycytidine 2',3'-dideoxyinosine dihydrofolate reductase dihydrotestosterone dose interval concentration dimethylbenzanthracene dimethyl sulphoxide doxorubicin dry powder inhaler eudismic affinity quotient Epstein—Barr virus effective dose for 50% response ethylenediamine-N,N,N’,N’-tetraacetic acid Taft’s steric substituent constant enzyme inhibitor complex enzyme inhibitor substrate complex oestrogen receptor antibody fragment retaining antigen binding properties follicle stimulating hormone 5-fluorouracil γ-aminobutyric acid GABA transaminase guanidine-cytosine granulocyte colony-stimulating factor gene-directed enzyme pro-drug therapy gastro-intestinal tract N-acetylglucosamine granulocyte-macrophage colony-stimulating factor glutathione guanosine triphosphate

HBV HGH HGPRT HIV HLE HMG HNE HPMA HPV HSA HSAB 5-HT HTLV-1 IC50 IFN Ig IP3 IUD Ki Km LDL LH LHRH MAO MDI MDR MEC MEP MO MPS Mr MR mRNA MSC α-MSH MurNAc NADPH NAPAP NMDA NMR NSAID

hepatitis B virus human growth hormone hypoxanthine-guanine-phosphoribosyl transferase human immunodeficiency virus human leucocyte elastase 3-hydroxy-3-methylglutaric acid human neutrophil elastase N-(2-hydroxypropyl) methacrylamide human papilloma virus human serum albumin hard soft acid base 5-hydroxytryptamine human T-cell leukaemia virus inhibitory concentration for 50% inhibition interferon immunoglobulin 1,4,5-inositol triphosphate idoxuridine equilibrium constant for breakdown of EI Michaelis constant low density lipoprotein luteinising hormone luteinising hormone—releasing hormone monoamine oxidase metered dose inhaler multidrug resistant gene minimum effective plasma concentration membrane metalloendopeptidase molecular orbital mononuclear phagocytic system molecular mass molecular refractivity messenger ribonucleic acid maximum safe concentration melanocyte stimulating hormone N-acetylmuramic acid nicotinamide adenine dinucleotide phosphate Nα-naphthylsulfonylglycyl-4-amidinophenylalanine piperidide N-methyl-D-aspartate nuclear magnetic resonance non-steroidal anti-inflammatory drug

ODC 4-OHA 8-OH DPAT OM OMP P P450 17 P450arom P450scc PAB PAPS PBD PBPs PCMB pD2 pDT Penicillin G α1-PI PIP2 pKa PKC PPE QSAR r Rb rv RMM SAFIR SAH SAM SAR SDAT SLPI S N1 S N2 SRS 3TC TCR TNF tRNA

ornithine decarboxylase 4-hydroxyandrostenedione 8-hydroxy-2(dipropylamino)tetralin outer membrane of bacteria proteins of outer membrane of bacteria partition coefficient 17α-hydroxy: 17,20-lyase aromatase side chain cleavage enzyme p-aminobenzoate 3’-phosphoadenosine-5’-phosphosulphate pyrrolo[2,1] [1,4-c] benzodiazepine penicillin binding proteins p-chloromercuribenzoate −log KA: where KA is the equilibrium constant between drug and receptor photodynamic therapy benzyl penicillin α1-protease inhibitor phosphatidylinositol biphosphate ionisation constant protein kinase C porcine pancreatic elastase quantitative structure—activity relationships correlation coefficient biological activity Van der Waals radius relative molecular mass structure-affinity relationship S-adenosylhomocysteine S-adenosylmethionine structure-activity relationship senile dementia of the Alzheimer’s type secretory leukocyte protease inhibitor unimolecular nucleophilic substitution bimolecular nucleophilic substitution slow releasing substance biological half-life 2’-deoxy-3’-thiathymidine therapeutic concentration ratio tumour necrosis factor transfer ribonucleic acid

TSAR UDP-GA VD VDEPT Vmax Vw

tools for structure-activity relationships uridine diphosphoglucuronic acid volume of distribution virus-directed enzyme pro-drug therapy maximum rate for enzyme reaction van der Waals volume

1. PROCESSES OF DRUG HANDLING BY THE BODY DAVID K.LUSCOMBE and PAUL J.NICHOLLS CONTENTS 1.1 INTRODUCTION

2

1.2 ABSORPTION

2

1.2.1 Transfer of drugs across cell membranes

3

1.2.2 Oral dosing

5

1.2.3 Rectal dosing

7

1.2.4 Topical application

7

1.2.5 Injections

8

1.3 DISTRIBUTION

9

1.3.1 Binding

9

1.3.2 Blood-brain barrier

11

1.3.3 Placental barrier

11

1.3.4 Partition into fat

11

1.4 METABOLISM

12

1.4.1 Phase I metabolism

13

1.4.1.1 Oxidations

13

1.4.1.2 Reductions

15

1.4.1.3 Hydrolyses

15

1.4.2 Phase II metabolism

16

1.4.2.1 Glucuronide formation

16

1.4.2.2 Sulphate formation

16

Introduction to the principles of drug design and action

2

1.4.2.3 Methylation

16

1.4.2.4 Acylation

17

1.4.2.5 Glutathione conjugation

17

1.4.3 Factors affecting metabolism

17

1.4.3.1 Stereoisomerism

18

1.4.3.2 Presystemic metabolism

18

1.4.3.3 Dose-dependent metabolism

18

1.4.3.4 Inter-species variation

18

1.4.3.5 Intra-species variation

18

1.4.3.6 Age

19

1.4.3.7 Inhibition of metabolism

19

1.4.3.8 Induction of metabolism

20

1.5 REMOVAL OF DRUGS FROM THE BODY 1.5.1 Renal elimination

20 21

1.5.1.1 Glomerular filtration

23

1.5.1.2 Active tubular secretion

23

1.5.1.3 Passive reabsorption across the renal tubules

24

1.5.1.4 Renal elimination in disease

24

1.5.2 Biliary elimination

25

1.5.3 Elimination in other secretions

25

1.6 SUMMARY

26

FURTHER READING

26

1.1 INTRODUCTION To be useful as a medicine, a drug must be capable of being delivered to its site of action in a concentration large enough to initiate a pharmacological response. This concentration will depend on the amount of drug administered, the rate and extent of its absorption and its distribution in the blood stream to other parts of the body. The medicine will continue to act until the concentration of drug drops below its threshold for pharmacological activity either due to its removal (excretion) from the body in an unchanged form or after

Process of drug handling by the body

3

its metabolism to a more polar substance. The interrelationship between the absorption, distribution, metabolism and excretion of a drug is referred to as pharmacokinetics and describes how drugs are handled by the body. Such knowledge of a new drug is fundamental to the drug development process, to enable selection of the optimal dose, route and frequency of dosing to produce the desired clinical effect, without producing unwanted side-effects. 1.2 ABSORPTION Whilst most medicines are taken by mouth and swallowed, other routes of administration include sublingual dosing in which the drug is placed under the tongue, rectal, inhalation, application to epithelial surfaces (skin patches), and injection either intravenously, intramuscularly or subcutaneously. With the exception of the intravenous route, in which the drug is administered directly into the bloodstream, a drug must initially be absorbed from its site of administration before it can enter the bloodstream and be distributed to its various sites of action. Clearly, the process of absorption is of fundamental importance in determining the pharmacodynamic and hence the therapeutic activity of a medicine. Delays or losses of drug during absorption may contribute to variability in drug response and may even result in a drug appearing to lack clinical effectiveness in some patients. Different formulations of the same active ingredient may lead to varying rates of absorption resulting in markedly different pharmacokinetic profiles in the same patient. Since the process of absorption involves the passage of a drug across one or more cell membranes, physico-chemical characteristics such as molecular size and shape, as well as solubility of the ionized and non-ionized forms will play an essential role in determining the overall pharmacodynamic activity of a drug. A basic knowledge of the physical and chemical principles governing the active and passive transfer of drugs across biological membranes is therefore necessary. 1.2.1 Transfer of drugs across cell membranes Living cells are surrounded by a semipermeable membrane measuring approximately 8µ in thickness. The ease with which a drug passes across such a membrane will reflect the concentration of drug achieved in the tissues and body fluids and hence at its pharmacological site of action. In general, there are four ways by which substances are able to cross cell membranes; diffusion through the lipid component of the membrane, diffusion through aqueous channels or pores in the membrane, combination with an active carrier molecule, by pinocytosis. The commonest and most important mechanism by which drugs are transferred across biological membranes is by passive diffusion. Transfer takes place along a concentration gradient from a region of higher concentration to one of low concentration following a first-order rate reaction. The greater this concentration gradient, the greater the rate of diffusion of a drug across the cell membrane. However, the ease with which a drug passes across a membrane will depend on the characteristics of both the drug molecule and the cell membrane. The drug’s partition coefficient between the lipid cell membrane and the aqueous environment is a major source of variability. Most drugs are weak acids or weak bases, existing in aqueous

Introduction to the principles of drug design and action

4

solution as an equilibrium mixture of non-ionized and ionized species. The non-ionized form is lipid soluble and therefore diffuses readily across cell membranes. In contrast, ionized compounds partition poorly into lipids and as a result are only slowly transported across biological membranes. In general, the higher the partition coefficient between lipid and water the more rapidly the drug is able to pass across cell membranes. The ratio of non-ionized to ionized drug when in aqueous solution is pH-dependent and can be calculated from the general form of the Henderson-Hasselbach equation:

where pKa is the dissociation constant. For drugs that are weak acids, the acid form is in the non-ionized form whilst for drugs that are weak bases, the base form is non-ionized. Thus, a solution of the weak acid aspirin (pKa 3.5) in the stomach at pH 1 will have greater than 99% of the drug in the non-ionized form and consequently is lipid soluble and will be rapidly absorbed into the bloodstream. Likewise, other weak acidic drugs will be absorbed in the stomach because they exist largely in their non-ionized form at low pH values. In contrast, most basic drugs are so highly ionized in the acid content of the stomach that absorption is negligible whilst in the near neutral fluids of the small intestine the absorption of weak basic drugs such as codeine (pKa8) is rapid. Nevertheless, it should be pointed out that the absorption of all orally administered drugs, weak acids as well as weak bases, probably takes place more rapidly in the small intestine than in the stomach. This is because the gastric mucosa has a relatively small surface area and its covering of protective mucus provides a poor site for absorption compared with the large surface area provided within the small intestine. Consequently, whilst only 0.1% of aspirin is in its non-ionized form at pH 7.0, aspirin is well absorbed from the small intestine following oral dosing. Strong organic acids and bases such as sulphonic acid derivatives and quaternary ammonium bases, are ionized over a wide range of pH values resulting in low lipid solubility and in consequence, such drugs are poorly absorbed from the gastrointestinal tract when administered orally. The pKa values for a number of acidic and basic drugs are illustrated in Figure 1.1.

Process of drug handling by the body

5

Figure 1.1 pKa values of some acidic and basic drugs. Whilst most drugs cross cell membranes by passive diffusion, some drugs such as methotrexate and 5-flourouracil are carried by an active transport mechanism which requires the expenditure of metabolic energy. The carrier is a membrane component capable of forming a complex with the drug to be transported. The complex moves across the membrane releasing the drug on the other side. Not surprisingly, carrier-aided transport systems can be saturated, thus limiting the rate of transport. This is in contrast to the process of passive diffusion across lipid membranes, or passage through pores,

Introduction to the principles of drug design and action

6

where the amount of drug conveyed increases proportionally with an increase in concentration. Active transport processes take place in the gastrointestinal tract (e.g. amino acids), in the renal tubules, and across membranes dividing extracellular from intracellular compartments at the blood-brain and placental barriers. Water-soluble substances such as alcohol are able to readily diffuse through the aqueous channels or pores in cell membranes providing their molecular weights are not greater than 100–200 Da. Since most drugs fall within the molecular weight range 200– 1000 Da, diffusion through these aqueous pores is unimportant for almost all substances with the exception of water, alcohol and other small polar molecules. Drug molecules can also be transported across cell membranes by an active uptake process similar to phagocytosis called pinocytosis. This involves the invagination of part of the cell membrane and the trapping of drops of extracellular fluid containing solute molecules which are thus carried through the membrane in the resulting vacuoles. Whilst this mechanism appears important in the absorption of some large molecules such as insulin which crosses the blood brain barrier by this process, pinocytosis is of little importance in the transport of small molecules across biological membranes except possibly in the case of oral vaccines. However, this process may become important if liposomes are used as a means of targeting a drug at a specific site of action since they may be taken up selectively by cells capable of pinocytosis. 1.2.2 Oral dosing The most common route of drug administration is by swallowing. This provides a convenient, relatively safe and economical method of dosing which, subject to the drug being presented in a palatable and suitable form, is the route preferred by most patients. Normally, about 75% of a drug given orally will be absorbed in 1 to 3 hours after dosing. To be effective a drug must be stable in the acid of the stomach fluids and not cause irritation of the gastrointestinal mucosa which might induce nausea and vomiting. It should not pass too rapidly through the stomach or interact with other drugs being administered concurrently. Whether the drug is formulated as a tablet, capsule or liquid preparation the most important site for drug absorption is the small intestine because it offers a far greater epithelial surface area for drug absorption than other parts of the gastrointestinal tract. Apart from the above, many other factors influence the rate and extent of drug absorption such as the physico-chemical properties of the drug, particle size, its concentration at the absorption site and splanchnic blood flow. In fact, the intestine has an excellent blood supply which ensures that any absorbed drug is rapidly transported into the bloodstream as soon as it passes through the intestinal membrane, maintaining a concentration gradient across the membrane. For highly lipid-soluble drugs, or those that pass freely through the aqueous-filled pores, passage across a membrane may be so rapid that equilibrium is established between the drug in the bloodstream and that at the site of absorption by the time the blood is removed from the membrane. In such cases, the rate-limiting step controlling drug absorption is blood flow and not transportation across the intestinal cell membranes. Drug absorption following oral dosing is generally favoured by an empty stomach. Food will effectively reduce the concentration of drug in the gastrointestinal tract which will limit its rate of absorption although not the total amount of drug absorbed.

Process of drug handling by the body

7

Furthermore, gastric emptying will be delayed slowing the onset of action of drugs such as antibiotics, analgesics and sedatives. In particular, gastric emptying is slowed by fats and fatty acids in the diet, and bulky or viscous foods. Some disorders will also slow gastric emptying, for example, mental depression, migraine, gastric ulcers and hypothyroidism whilst many drugs including propantheline, imipramine and the antacid aluminium hydroxide will all produce the same effect. In contrast, factors which promote gastric emptying will result in an increased rate of absorption of nearly all drugs. Such factors include fasting or hunger, alkaline buffer solutions, diseases such as hyperthyroidism and the anti-emetic agent, metoclopramide. Generally, the gastric emptying of liquids is much faster than that of solid food or solid dosage forms. It is for this reason, that tablets and capsules should be taken orally with at least half a glassful of water. In contrast, drugs known to irritate the gastric mucosa, for example, antiinflammatory agents, should be taken immediately after a meal, even though this may decrease its rate of absorption, as the likelihood of induced nausea will be diminished. The term bioavailability is used to describe the proportion of orally administered drug that passes unchanged into the bloodstream. It is particularly useful because it takes into account absorption and any local metabolic degradation that takes place in the stomach and small intestine. Bioavailability is also influenced by gastrointestinal motility, gastric pH, drug solubility, the presence or absence of food in the gastrointestinal tract and the formulation of the dosage form administered (particularly when a drug is prepared by different manufacturers). Benzylpenicillin, the only naturally occurring penicillin in clinical use, is destroyed by gastric acid and therefore has to be administered by injection. Ampicillin in contrast is acid stable and in consequence can be given orally. However, its bioavailability is variable and absorption is incomplete. In an attempt to improve absorption following oral dosing, lipophilic esters have been prepared with some success. Whilst esters of penicillins are inactive in vivo, once absorbed they are hydrolysed to release the active penicillin (see Section 7.4.1). As a result, a number of these so-called ‘prodrugs’ have been successfully developed. This prodrug approach to increasing bioavailability has also been used with angiotensin-converting enzyme (ACE) inhibitors (see Section 7.4.1). Once absorbed from the gastrointestinal tract, an orally administered drug will enter the portal blood circulation and pass immediately to the liver before entering the systemic circulation and delivery to its site(s) of action. On passing through the liver, the drug may be partially or completely metabolized by hepatic microsomal enzymes to less active metabolites or be excreted in the bile from where it passes into the small intestine. These processes may result in a marked reduction in the amount of unchanged (active) drug that is available to exert a pharmacological effect, a phenomenon known as first-pass metabolism. Oral dosing is clearly inappropriate for drugs such as lignocaine which undergo an extensive first-pass effect. Despite rapid absorption from the gastrointestinal tract, lignocaine is so extensively degraded by the hepatic microsomal enzyme system on its initial passage through the liver that the remaining lignocaine level in the peripheral blood circulation is inadequate to exert its therapeutic effect. For drugs which are absorbed through the mucosa of the buccal cavity when placed under the tongue and allowed to dissolve, first-pass metabolism can be avoided. This sublingual route of administration is not often encountered, but since the drug does not have to enter the stomach or intestines to exert its effect, absorption is generally more

Introduction to the principles of drug design and action

8

rapid than after swallowing and the drug is likely to be effective at a lower dose. Hydrolytic enzymes in the intestinal mucosa inactivate glycerol trinitrate and is the reason for this anti-angina drug being administered sublingually rather than being swallowed. This route offers the patient the opportunity to terminate the therapeutic effect once relief has been achieved simply by spitting out the dosage form from under the tongue. Drugs with an unpleasant taste cannot be administered sublingually and neither can high molecular weight substances which are only poorly absorbed through the buccal cavity mucosa. 1.2.3 Rectal dosing Some drugs cause nausea and vomiting when given orally and these may be formulated as suppositories or enemas and given rectally to be absorbed in the rectum. Drugs administered rectally are not subject to first-pass metabolism in the liver, however, absorption is generally irregular, unpredictable and incomplete. Apart from being used in the treatment of constipation or to evacuate the bowels before surgery, this route is generally avoided. 1.2.4 Topical application The application of drugs to the skin or mucous membranes, such as the conjunctiva, nasopharynx or vagina, is used principally for local effects. However, in the past decade a number of drugs have been successfully formulated as self-adhesive skin patches. When placed on the skin, these patches slowly release drug which passes across the skin and into the systemic blood circulation to produce a generalised effect in the body. For example, glyceryl trinitrate when formulated as a transdermal patch will slowly and continuously release drug into the bloodstream providing prophylactic treatment for angina over a 24 h period. The patch is replaced daily, using a different area of the body on each occasion. The same drug has also been formulated as an ointment which can be applied to the chest, abdomen or thigh without rubbing in, being secured with a dressing. This provides short-term prophylactic cover for angina, being repeated every 3–4 hours as required. Self-adhesive nicotine patches are widely available for smokers who wish to give up the habit. They are applied on waking to dry, non-hairy skin on the hip, chest or upper arm, being removed before retiring. The siting of the replacement patch should be on an unused area, used areas being avoided for several days. Nasal sprays containing nicotine are also available for people wishing to stop cigarette smoking. Nicotine is delivered into each nostril as required up to a maximum of two sprays an hour for 16 hours a day. The strong opioid analgesic, fentanyl, has been introduced recently in a transdermal drug delivery system as a self-adhesive patch which provides pain relief for up to 72 hours before needing to be replaced. Women requiring hormone replacement therapy are now offered oestrogen formulated either as a self-adhesive skin patch or a gel preparation in addition to tablet dosage forms. The anti-motion sickness drug hyoscine hydrobromide is likewise available in a self-adhesive patch dosage form being placed on a hairless area of skin behind an ear some 5–6 hours before travelling as protection against motion sickness.

Process of drug handling by the body

9

1.2.5 Injections The most common method of introducing a drug directly into the bloodstream is to inject it intravenously. This route is particularly useful when a rapid therapeutic response is required since absorption is circumvented. It is used for the induction of anaesthesia, relief from some epileptic seizures and for administering antibiotics such as benzylpenicillin which is inactivated by gastric acid if given orally. On intravenous dosing, the drug is rapidly removed from the injection site, being diluted in the venous blood as it is carried initially to the heart and then to other tissues. Since the total circulation time in humans is of the order of 15s, the onset of drug action is almost immediate. Drugs delivered by the intravenous route may be administered either as a single rapid injection lasting only 1–2 minutes, known as a ‘bolus’ injection, or as a slow infusion lasting an hour or longer. This latter choice is preferred when a sustained level of drug is required in the bloodstream over a relatively long period (e.g. antibiotics for lifethreatening infections in hospitalized patients). It is also useful for administering large drug volumes and for diluting otherwise irritant substances. This route is not suitable for water-insoluble drugs and suffers the disadvantage that the dosage form must be sterile. Furthermore, intravenous administration must be by trained personnel and great care is needed to ensure that overdosage is avoided, since the rapidity of drug action may not permit the reversal of any drug-induced toxicity. Due to stability problems most antibiotics such as cloxacillin, flucloxacillin, amoxycillin for intravenous use are provided as a dry sterile powder (i.e. sodium salt) to be reconstituted with water for injection before use. Drugs may also be administered by intramuscular injection enabling the exact quantity of drug to be delivered to a localized site such as the deltoid muscle of the arm, the vastus lateralis of the thigh or the gluteus maximus of the buttocks. From these muscular sites the drug must be absorbed before passing into the general circulation. Factors which influence absorption from these sites include the vascularity of the injection site, the degree of ionization and lipid solubility of the drug, volume of injection, and osmolarity. The intramuscular route is often used in patients who are unable to swallow oral medication, for drugs which are poorly absorbed from the gastrointestinal tract, or for drugs which undergo extensive first-pass metabolism. For example, 4hydroxyandrostenedione is a potent mechanism-based enzyme inactivator of aromatase used to lower oestrogen levels in post-menopausal women with breast cancer. This has to be administered as an intramuscular injection to avoid extensive first-pass metabolism to the inactive glucuronidated conjugate which takes place following oral dosing. Drugs administered by intramuscular injection generally exert their pharmacological effect more rapidly than after oral dosing. However, intramuscular injections tend to be painful and are not generally favoured by patients. Nevertheless, a number of penicillins, such as cloxacillin, flucloxacillin and ampicillin may be given intramuscularly as an alternative to oral dosing. The subcutaneous route of injection is only suitable for small dose volumes and few drugs are currently administered by this route with the notable exception of insulin. Drugs given by this route spread out through the loose connective tissue of the subcutaneous layer. Since the skin is rich in sensory nerves, subcutaneous injections are more painful than intramuscular injections. In general, a subcutaneous injection results in faster absorption than a corresponding intramuscular injection, although the difference is

Introduction to the principles of drug design and action

10

not marked and of little clinical significance. Absorption is influenced by the same factors that determine the rate of drug absorption from intramuscular sites. However, the blood supply to subcutaneous layers may be poorer than in muscle tissues and in consequence absorption may appear slower. In both instances, the local action of an injected drug can be prolonged by decreasing its rate of removal from the site of injection. The action of subcutaneously administered drugs can also be sustained if drugs are injected as solutions in oil, the drug diffusing out only slowly from the vehicle. Oestradiol may be administered in the form of a solid pellet which is implanted under the skin, the active hormone slowly dissolving in the tissue fluid before diffusing through the capillary walls and into the bloodstream. Such implants have the benefit of remaining effective for 4 to 8 months. 1.3 DISTRIBUTION Once absorbed into the bloodstream most drugs are distributed throughout body fluids and tissues with relative ease. The pattern of distribution depends on the drug’s permeability, lipid solubility and capacity to bind to macromolecules (largely proteins). The apparent volume of distribution (VD) is a useful term to describe a drug’s pattern of distribution. It represents the volume in which the drug appears to be dissolved in a body fluid (i.e. compartment) and is a proportionality constant relating drug concentration to the total amount of drug in the body. A drug such as heparin whose distribution is largely restricted to the plasma compartment has a small volume of distribution (i.e. 0.05 litre kg−1) whilst nortriptyline (22–27 litre kg−1) has a large volume of distribution which in fact is in excess of total body water (about 0.6 litre kg−1). This indicates that nortriptyline is not only widely distributed throughout total body water but is being accumulated or stored in extravascular sites. Generally, weak bases have a large VD value owing to their lipid solubility and as a result will be present in low concentrations (i.e. ng ml−1) in plasma (eg. diazepam, morphine, imipramine). The reverse situation applies to weakly acidic drugs which will tend to exhibit high (i.e. µg ml−1) plasma concentrations (eg. aspirin, sulphamethoxazole). For those drugs with a molecular weight of less than 600 Da, and which are being transported as free drug in solution in plasma water, transfer from blood vessels out into interstitial fluid is rapid. This is because capillary walls generally behave like a leaky sieve. The lining endothelial cells of the capillary have junctions with each other that are not continuous (i.e. loose), and allow free passage of such relatively small molecules across the capillary wall. This is important for polar compounds, however, both this route and diffusion through the actual capillary wall are also available pathways for lipidsoluble compounds. 1.3.1 Binding An important factor in the distribution of drugs is their binding to macromolecules such as plasma proteins. Such binding is generally a reversible process. The extent to which a drug is bound to plasma proteins will influence its pharmacological profile. This is because it is only that fraction of a drug which is in solution in the plasma (unbound) that

Process of drug handling by the body

11

is free to cross cell membranes and interact with receptors thus effecting a pharmacological response. Drug-protein binding complexes have such high molecular weights that they will not cross cell membranes and are in effect pharmacologically inactive (i.e. the drug is protein bound). In plasma, the main protein for drug binding is albumin while the binding forces involved may be ionic, van de Waals’, hydrogen and/or hydrophobic bonds. Since binding is mostly reversible, there is an equilibrium in plasma between bound and unbound drug, this interaction following the Law of Mass Action. Plasma drug binding depends on the association constant of the drug, the number of binding sites and the concentration of both drug and plasma protein. As the plasma concentration of drug gradually increases following absorption, the fraction of drug in its free form rises slowly at first, but as the protein binding sites become saturated this fraction rises sharply. In practice, the fraction of free drug in the plasma is essentially constant over the range of therapeutic concentrations for most drugs. Saturation is most likely to occur with drugs which have high association constants and are administered in high doses, such as sulphonamides. The extent to which drugs are bound to plasma proteins, particularly albumin, is variable depending on their physico-chemical characteristics. Examples of drugs which bind to albumin are presented in Table 1.1. Since protein-binding sites are non-specific, one drug can displace another thereby increasing the proportion of free (unbound/active) drug to diffuse from the plasma to its site of action. This is only clinically important if the drug is firstly, extensively bound (greater than 90%) and secondly, is not widely distributed throughout the body (i.e. warfarin VD=0.05 litre kg−1). Hence, the pharmacological activity of warfarin is markedly increased when administered concurrently with

Table 1.1 Some drugs that bind to plasma albumin. Drug Diazepam Diclofenac Warfarin Amitriptyline Chlorpromazine Imipramine Nortriptyline Tolbutamide Valproic acid Phenytoin Hydralazine Sulphadimidine Aspirin Lignocaine Sulphadiazine

Binding 95–99%

90–95%

90–95% 90% 80–90% 60–80% 45–60%

Introduction to the principles of drug design and action

12

sulphonamides due to displacement of the former by the latter drug. This leads to higher free (unbound) warfarin concentrations in the plasma which potentiates its anticoagulant effect which could lead to fatal haemorrhage. It is important to view the degree of binding to plasma proteins in relation to a drug’s apparent volume of distribution. Thus, while nortriptyline is 93% bound under steady state conditions, the drug concentration in plasma is less than 1% of the total amount of drug in the body and any displacement by another drug will be clinically insignificant. 1.3.2 Blood-brain barrier Penetration of drugs from the blood-stream into the brain and cerebrospinal fluid is restricted by a specialised protective lipid membrane, the blood-brain barrier. Whilst highly lipid-soluble compounds reach the brain rapidly following dosing, more polar compounds penetrate at a much slower rate and highly polar drugs will not cross into the brain under normal circumstances. As a general rule, the rate of passage of a drug into the brain is determined by its degree of ionization in the plasma and its lipid solubility. Thus, penicillin which is highly ionized will be excluded from the brain unless very large doses are administered. However, the permeability of the blood-brain barrier can be increased by infections which lead to meningeal or encephalic inflammation. This is the reason why penicillin is used in the treatment of meningococcal meningitis. 1.3.3 Placental barrier Foetal blood is separated from maternal blood by a cellular barrier the thickness of which is greater in early pregnancy (25 µm) than in the later stages (2 µm). Although specific transport systems for endogenous materials are present in the placenta and may provide a method for transporting some drugs such as methyldopa and 5-fluorouracil, it appears that most drugs cross the placenta by passive diffusion. Thus, penetration is rapid with lipid-soluble non-ionized drugs but slow with very polar compounds. However, some degree of foetal exposure is likely to occur with most drugs and so caution is required with drug administration during pregnancy. Some drugs such as the sulphonamides readily cross the placental barrier and may reach concentrations in the foetal blood circulation high enough to be antibacterial and lead to toxicity. 1.3.4 Partition into fat Lipid-soluble drugs may achieve high concentrations in adipose tissue, being stored by physical solution in the neutral fat. Since fat is normally 15 per cent of body weight (in grossly obese subjects it can be as high as 50 per cent), it can serve as an important reservoir for such drugs. It also has a role in terminating the effects of highly lipidsoluble compounds by acting as an acceptor of the drug during a redistribution phase. Thus, after intravenous injection, thiopentone enters the brain rapidly, but also leaves it rapidly because of falling plasma levels and this terminates its pharmacodynamic action. It then slowly redistributes into fatty tissues where as much as 70 per cent of the drug may be found 3 h after administration.

Process of drug handling by the body

13

1.4 METABOLISM Most drugs, prior to removal from the body, are subjected to biotransformation (metabolism). The enzymic reactions leading to such changes are classified as Phase I reactions (asynthetic changes) and Phase II reactions (conjugations). As the original compound is chemically altered by these means, metabolism may be considered as a drug elimination mechanism although the problem of excreting the metabolites remains. In most instances, the metabolites have a markedly different partition character from the parent compound, in that lipophilicity is decreased. Such products tend to be easily excreted, as they are not readily reabsorbed from the renal tubular fluid. Drug metabolites also often have a smaller apparent volume of distribution than their precursors. Metabolism influences the biological activity of a drug in a number of ways (see Table 1.2). In many instances, pharmacological activity is reduced or lost by metabolism and for such drugs this may be an important determinant of duration of action and even intensity of effect. Occasionally a drug may be transformed into a metabolite possessing a pharmacological effect of comparable intensity (see Chapter 7). The benzodiazepines are a good example of this phenomenon. Thus the major metabolite of diazepam, N-desmethyldiazepam, has a similar pharmacological potency and long half life (t0.5). Minor metabolites of diazepam are temazepam and oxazepam which are also active. For a relatively small number of drugs (prodrugs), biologically

Table 1.2 Examples of effect of drug metabolism on pharmacological activity. Effect Drug Metabolic reaction Deactivation Drug metabolite Aminoglutethimide Conjugation (with acetic acid) less active than Amphetamine Oxidation parent molecule Barbiturates Oxidation Chloramphenicol or inactive Conjugation (with Procaine glucuronic acid) Tolbutamide Hydrolysis Oxidation Trans-activation Oxidation (to Drug metabolite Diazepam Phenylbutazone nordiazepam) possessing Propranolol Oxidation (to equivalent oxyphenylbutazone) activity to parent Procainamide Oxidation (to 4molecule hydroxypropranolol) Conjugation (to Nacetyl procainamide) Activation

Introduction to the principles of drug design and action

Metabolite is responsible for (pro-) drug activity

Chloral hydrate Chlorazepate Palmitic ester of chloramphenicol Proguanil Prontosil red

Toxification Drug metabolite Malathion possessing toxic Methanol effects Paracetamol

14

Reduction (to trichloroethanol) Oxidation (to nordiazepam) Hydrolysis (to chloramphenicol) Oxidation (to cycloguanyl) Reduction (to sulphanilamide) Oxidation (to malaoxon) Oxidation (to formaldehyde and formic acid) Oxidation (to an electrophilic imidoquinone)

inactive per se, metabolic activation is a prerequisite for therapeutic utility (see Chapter 7), e.g. the popular angiotensin-converting enzyme inhibitor, enalapril, is hydrolysed in vivo to its active form enalaprilat. An interesting example of the application of this principle to achieve selectivity of pharmacological action is the anti-epileptic drug vigabatrin (γ-vinyl-GABA) which is a substrate for the neuronal GABA-ketoglutarate transaminase responsible for inactivating the inhibitory neurotransmitter GABA. The resultant metabolite of vigabatrin is an irreversible inhibitor of the transaminase and this action leads to increased levels of GABA in the brain. Finally, a growing list of drugs and other xenobiotic compounds is metabolized to intermediates that may subsequently react with tissue macromolecules leading to toxic effects. For example, it is considered that the occurrence of haemorrhagic cystitis in bone marrow transplant patients receiving cyclophosphamide is related to the drug’s metabolism to the toxic compound, acrolein. The main site of drug metabolism is the liver, followed by the gastointestinal tract. However, metabolism also occurs in the kidney, lung, skin and blood but, quantitatively, these sites are less important. 1.4.1 Phase I metabolism The Phase I reactions are oxidation, reduction and hydrolysis.

Process of drug handling by the body

15

1.4.1.1 Oxidations Many of the oxidation reactions, such as aliphatic and aromatic hydroxylation, epoxidation, dealkylation, deamination, N-oxidation and S-oxidation (see Table 1.3), are catalysed by enzymes (mixed function oxidases) bound to the endoplasmic reticulum. This latter is a branching tubular system within cells that is also involved in protein synthesis and lipid metabolism. When a tissue such as liver is homogenized, this reticulum fragments into rounded bodies (microsomes) sedimenting at 10–100 S. Many metabolic oxidations have been studied using this microsomal enzyme fraction. The terminal oxygen transferase of the system is cytochrome P450. This is coupled to the flavoprotein enzyme, cytochrome P450-reductase, and linked to NADPH as a source of electrons. Under the influence of cytochrome P450, an oxygen atom from molecular oxygen is transferred to a drug molecule (DH→DOH). The remaining oxygen atom combines with two protons to yield a molecule of water. Thus the enzyme is characterised as a mixed function oxidase. Cytochrome P450 is so named because its reduced carbon monoxide-ligand spectrum has a maximum absorption at 450 nm. It is now known that cytochrome P450 and its reductase both exist in multiple forms and the cytochrome P450 variants appear to possess overlapping substrate specificities. The differences between the cytochrome P450 isozymes are due to modified sequences of the amino acids in the protein of this haemoprotein. A general nomenclature for the isoforms based on structural homology has been agreed. Thus P450 proteins from all sources with a 40% or greater sequence identity are included in the same family (designated by an Arabic numeral). Those isoforms with greater than 55% identity are then included in the same sub-family (designated by a capital letter). The individual genes (and gene products) are then arbitrarily assigned a number. An example is the major phenobarbitone-inducible cytochrome (see 1.4.3.8) in rabbit liver microsomes. This was originally called form 2 or P450LM2. With the present system, this enzyme has

Introduction to the principles of drug design and action

Table 1.3 Some microsomal oxidations.

16

Process of drug handling by the body

17

been assigned to family 2 and sub-family B; the gene and the enzyme are designated CYP2B4. The enzyme may also be styled P450 2B4. In addition to an ability to bind to cytochrome P450, requirements of a substrate for metabolism by this system include: a molecular size above 150 µ (below this size, compounds are normally capable of ready excretion), a sufficient degree of lipophilicity to enter the endoplasmic reticulum and the appropriate chemical substituents. Chemical reactivity at sites on a molecule influences the site of oxidative enzyme attack. Thus, with nitrobenzene, the main oxidative metabolite is 3-hydroxynitrobenzene, while with aniline, 2- and 4-hydroxyaniline are the major ring-oxidized products. While most interest in the cytochrome P450 system is focused on drug metabolism, it must be recognised that several of the isozymes are responsible for the biotransformation of endogenous compounds such as steroid hormones, leukotrienes, prostaglandins, vitamins and free fatty acids. A non-cytochrome P450-dependent microsomal flavoprotein oxidase has been described in liver that effects sulphoxidation of nucleophilic sulphur compounds (e.g. methimazole), hydroxylamine formation from secondary amines (e.g. desipramine, nortriptyline) and amine oxide formation from tertiary amines (e.g. brompheniramine, guanethidine). Oxidations are also carried out by non-microsomal enzymes such as alcohol and aldehyde dehydrogenases and monoamine and diamine oxidases. Although the oxidations are less varied than those of the microsomal enzymes, they are important pathways for several naturally occurring compounds as well as drugs. 1.4.1.2 Reductions Only a small number of drugs is metabolized by reduction, the reductases being located at both microsomal and non-microsomal sites. Some reductases are also found in the micro-organisms of the gut. Aromatic azo and nitro compounds are reduced by microsomal flavoprotein enzymes. The nitro-reductase converts the substrate (e.g. chloramphenicol, nitrazepam) to the corresponding amine by the following sequential reactions: . Azo-reductase effects a reductive cleavage of its substrate by the following sequence: (e.g. prontosil red→sulphanilamide+1,2,4 triaminobenzene). There is a marked azo-reductase activity in the gut microflora. A hepatic microsomal enzyme, requiring NADPH and oxygen, is responsible for replacing halogen with hydrogen in aliphatic halogenated compounds such as halothane, methoxyflurane and carbon tetrachloride (e.g. CCl4→CHCl3). Examples of reductions carried out by non-microsomal enzymes are the transformation in the blood of disulphiram ((C2H5)2 NCSS-SSCN(C2H5)2) into diethyldithiocarbamate ((C2H5)2NCSSH) and the reduction of chloral hydrate to trichloroethanol by alcohol dehydrogenase. 1.4.1.3 Hydrolyses Drugs containing an ester group may be hydrolysed by esterases which have both microsomal and non-microsomal locations. The former tend to be more concentrated in

Introduction to the principles of drug design and action

18

the liver. Such an enzyme is responsible for the hydrolysis of pethidine. The nonmicrosomal esterases occur in blood and some tissues; procaine is metabolized by a plasma esterase. The esterases also hydrolyse amides (e.g. procainamide), though more slowly than the corresponding esters. Epoxide hydrases, present in the microsomal fraction of many tissues, convert epoxides to the corresponding dihydrodiols. This is an important detoxifying reaction for reactive electrophilic epoxides formed as a result of metabolism. A (minor) epoxide metabolite of phenytoin is possibly associated with a higher than normal incidence of neonatal cleft palate when phenytoin is administered to pregnant women. It is likely that women at risk are those in whom there is a relative deficiency of the epoxide hydrase(s) responsible for the inactivation of the toxic metabolite. 1.4.2 Phase II metabolism Phase II metabolism involves the coupling of a drug or its metabolites with various endogenous components. The reaction, which is carried out by a transferase enzyme, requires that either the endogenous or the exogenous component is activated prior to conjugation. Although generally considered to be detoxication pathways, conjugation reactions may result in “metabolic activation”. An example of this is where the 6glucuronide of morphine acts as a carrier molecule allowing its ready passage across the blood-brain barrier. In the brain, the conjugate is cleaved by a hydrolase releasing the active molecule, morphine. Under the influence of tissue deacetylases, the N-acetyl conjugate of isoniazid may give rise to the hepatotoxin, N-acetylhydrazine. 1.4.2.1 Glucuronide formation Probably the most common conjugation pathway is that of glucuronide formation. The combination with glucuronic acid occurs with compounds possessing a functional group with a reactive proton, usually attached to a hetero-atom (e.g. hydroxyl, carboxyl, amino and sulphydryl). These functional groups may be already present in a drug molecule (e.g. paracetamol) or may be acquired by Phase I metabolism (e.g. phenytoin hydroxylation). Depending on the grouping through which conjugation takes place, these metabolites can be described as O-glucuronides (ether type—combination through a hydroxyl group, e.g. alcohol metabolites of barbiturates; ester type—combination through a carboxyl group, e.g. salicylic acid), N-glucuronides (via amino groups, e.g. meprobamate) and Sglucuronides (via sulphydryl groups, e.g. 2-mercaptobenzothiazole). Glucuronic acid is derived enzymically from glucose and its active form, uridine diphospho-glucuronic acid (UDP-GA), is utlized by the UDP-glucuronyl-transferase to effect the conjugation. Glucuronides are very polar and relatively strong acids . They are thus extensively ionized at the pH of blood and urine; this makes them good candidates for excretion. In mammals, phenolic and carboxylic compounds can be conjugated with glucose, the high energy glucose donor being UDP-glucose. The glucosides are more water-soluble than the free aglycones but less polar than the corresponding glucuronides.

Process of drug handling by the body

19

1.4.2.2 Sulphate formation Sulphase esters are formed by the soluble fraction (i.e. 100 S supernatant), the high energy sulphate being 3′-phosphoadenosine-5′-phosphosulphate (PAPS) and the other component substrate being either a phenol (e.g. paracetamol, salicylamide) or aliphatic and steroid alcohol (e.g. ethanol, androsterone). Sulphamates may also be formed in a similar manner from aromatic amines. The capacity to form sulphate conjugates is somewhat limited, and this appears to be related to the low availability of sulphate. 1.4.2.3 Methylation Methylation is an important physiological process for the conversion of noradrenaline into adrenaline (N-methylation). Both of these catecholamines are also metabolized by Omethylation under the influence of catechol-O-methyl transferase (COMT). The methyl group is derived from methionine, the active methyl donor form of which is Sadenosylmethionine. Drugs or their metabolites containing primary aliphatic amine, phenolic or sulphydryl groups may be N-, O- or S-methylated, respectively, by methyltransferases. Thus the minor catechol metabolite of phenytoin, 5-phenyl-5-(3,4dihydroxyphenyl) hydantoin, is conjugated to give the corresponding 3-methylcatechol. 1.4.2.4 Acylation Several acylation conjugation reactions of importance may occur with some drugs. This pathway involves the reaction between an amine and a carboxylic acid to yield an amide, the high energy molecule required being a coenzyme A derivative of the carboxylic acid. The drug, or its metabolite, can be either of the conjugating molecules. Thus, aromatic primary amines (e.g. sulphonamides, aminoglutethimide) and hydrazine derivatives (e.g. isoniazid) are acetylated, utilizing acetyl coenzyme A. It should be noted that acetylation has little influence on the polarity of a drug, in fact it decreases the basicity of the amino group. The acetylated metabolite of sulphathiazole is some 14 times less soluble in water (37°C) than its parent molecule. Because of this property, and a lowered solubility at acid pH, there is the danger of injury to the kidney resulting from precipitation of the conjugated sulphonamide in the renal tubular fluid as the kidney concentrates urine and lowers its pH. The acetyltransferase appears to be located in the soluble fraction of reticulo-endothelial cells present in the liver and kidney. For some metabolically acetylated drugs, e.g. various sulphonamides, isoniazid and procainamide, enzymic deacetylation may occur. Deconjugation is a phenomenon which is not well recognised and its significance is poorly understood. Examples of this “reversal of metabolism” have been reported for other pathways e.g. hydrolysis of glucuronide conjugates and the reduction of the N-oxide of imipramine. Benzoic acid and its derivatives are activated by combination with coenzyme A and conjugated with glycine to form hippurates (e.g. salicylic acid metabolized to salicyluric acid). This takes place in the mitochondria of the liver and kidney.

Introduction to the principles of drug design and action

20

1.4.2.5 Glutathione conjugation The tripeptide glutathione (cystine-glycine-glutamate) may be coupled via its sulphydryl group to various compounds possessing an electrophilic centre. In the case of paracetamol, such a site is introduced as a result of oxidative metabolism. This conjugation reaction is an important mechanism for the effective disposal of electrophiles (e.g. reactive epoxides) before they are able to react with nucleophilic centres of nucleic acids and enzymes to initiate toxic responses. Myleran (busulphan), azathioprine and urethane are examples of drugs conjugated by this pathway. Glutathione conjugates are polar and of high molecular weight (above 300 Da) and are eliminated as such in the bile. However, the glutathione portion of the conjugate may be further metabolized (via the peptide bonds) to mercapturic acids that are the normal urinary products of this conjugation pathway. 1.4.3 Factors influencing metabolism It will be evident from the foregoing that even the simplest of drugs may be subjected to several types of metabolic transformation. Thus, propranolol is conjugated directly with glucuronic acid, ring-hydroxylated and oxidized in the side chain. A complex molecule like chlorpromazine may give rise to an extremely large number of different metabolites. 1.4.3.1 Stereoisomerism Where a drug exists in stereoisomeric forms, the rate and routes of metabolism may differ between the enantiomers. Thus (−)-hexobarbitone and (−)-warfarin are metabolized faster than the (+)-isomers. While the (+)-isomer of glutethimide is hydroxylated in the 4-position of the glutarimide ring, the (−)-isomer undergoes oxidation of the ethyl substituent on the 2-position of the ring structure. 1.4.3.2 Presystemic metabolism For drugs administered orally, there is the possibility of their metabolism as they pass through the wall of the small intestine and (via the portal circulation) through the liver before they reach the heart for distribution systemically. This first-pass or presystemic metabolism has a profound influence on the bioavailability of drugs such as isoprenaline, terbutaline, propranolol, alprenolol, imipramine, dextropropoxyphene and lignocaine. 1.4.3.3 Dose-dependent metabolism In most cases, the metabolism of a drug is a first order process which means that a constant fraction of the drug is metabolized in unit time. However, the therapeutic doses of some drugs (e.g. phenytoin 300–350 mg daily) result in concentrations able to saturate the metabolizing enzymes and zero order kinetics operate (i.e. a constant

Process of drug handling by the body

21

amount of drug is metabolized per unit of time). In this situation, steady state concentrations of the drug rise very sharply with relatively small increments of daily dose and toxicity may arise. Saturation of one metabolic pathway may allow for a shift in the metabolic pattern of a drug. Thus, after paracetamol overdosage (e.g. 20 g), the glucuronide and sulphate conjugation pathways become saturated, making available a greater fraction of the dose for oxidation to a reactive and potentially toxic metabolite. 1.4.3.4 Inter-species variation Differences in drug metabolism may occur between species and this is of great importance in drug development investigations. The differences may be associated with the rate of drug metabolism, e.g. hexobarbitone is oxidized by the following species in order of decreasing rate: mouse>rat>dog>man. For the mixed function oxidases, there is direct correlation with their activity and the tissue oxygen concentration in a species. The route of metabolism may also be influenced by species. Thus bethanidine is mainly N-demethylated by the dog, ring-hydroxylated by the rat and excreted unchanged by man. Well-documented examples of species differences include the poor acetylation of aromatic amines in the dog, the deficiency of glucuronide formation in the cat and the absence of atropinesterase in man. 1.4.3.5 Intra-species variation Different rates and extents of drug metabolism also occur within a species (including man). After a dose of imipramine to human subjects, the plasma levels of the drug 12 h later show a 12-fold variation between individuals. Similar ranges of variations have been found with desipramine and chlorpromazine. The plasma half-lives of certain drugs oxidized by hepatic microsomal enzymes (e.g. antipyrine, phenylbutazone) show much more marked differences between pairs of fraternal twins than between pairs of identical twins. This indicates that genetic rather than environmental causes give rise to such intersubject variability; pharmacogenetics is now an important specialty within pharmacology. A notable example is the hydrolysis of suxamethonium. In some patients, the normal dose gives rise to prolonged muscle relaxation and apnoea. These individuals possess an atypical pseudocholinesterase with a low affinity for suxamethonium. Drug metabolism defects may sometimes be clearly associated with certain congenital abnormalities, e.g. in Down’s syndrome, glycine conjugation is deficient and in Gilbert’s syndrome, glucuronide formation is impaired. Wide variation in the extent of acetylation of isoniazid, hydrallazine, phenelzine, dapsone and some sulphonamides exists and distinct sub-populations of fast or slow acetylators can be defined. Rapid acetylation is inherited as a dominant character which determines the presence of large amounts of the N-acetyltransferase. The relative proportions of rapid and slow acetylators have been shown to vary between ethnic groups (e.g. proportion of slow acetylators in Canadian Eskimos 10%, Swedes 50%, British 60%, Egyptians 72%). Slow acetylators are more susceptible to adverse effects from isoniazid, hydrallazine and phenelzine. Other well studied genetic polymorphisms are the hydroxylations of debrisoquine (CYP2D6), mephenytoin and sparteine (CYP2C19).

Introduction to the principles of drug design and action

22

Where other factors such as age and pathophysiology do not overly influence the genetic expression of metabolising enzymes, phenotyping of patients into slow or fast metabolisers may be an important means of improving the therapeutic efficiency of a drug, e.g. isoniazid. A clear indication for phenotyping is in healthy volunteers and the first series of patients receiving a new drug in the early stages of clinical trial, where its metabolism is known to be influenced by genetic polymorphism. 1.4.3.6 Age A newborn child is deficient in microsomal enzymes including cytochrome P450 and UDP-glucuronyl-transferase, although this may be modified (induction, see 1.4.3.8) by drugs taken by the mother during the latter part of gestation. As a result, the half-lives of several drugs are prolonged in the neonate compared to the adult (e.g. t0.5 for tolbutamine, 40 h at birth, 8 h in adult). Drugs may therefore have more prolonged or intense effects and adverse reactions may arise. For example, chloramphenicol, requiring conjugation with glucuronic acid, is much more toxic to a newborn infant than to an adult. In general, enzyme activity increases to maximum levels over the first 8 weeks of life. There is some evidence that drug metabolism (of, e.g. theophylline, phenobarbitone, diazoxide) in children, prior to puberty, may be faster than in adults. A decrease in drug metabolism may occur with advancing years, as is shown by the slower oxidation of amylobarbitone in individuals over 65 years of age. This is most likely due to the reduction of liver mass that occurs during the latter part of the age spectrum. Reduced liver blood flow, reduced cardiac output and a degree of hypoxia may also be contributory factors. However, the influence of old age appears to be obscured in many cases by environmental factors (inducers) such as cigarette smoking. 1.4.3.7 Inhibition of metabolism Inhibition of drug-metabolizing enzymes may arise from a competitive interaction of two (alternative) substrates for the enzyme. It can result in the prolongation of the duration of drug effects and/or an enhancement of action including increased toxicity. The overall effect depends on the relative concentrations of the two substrates and their affinities for the active sites. Other types of inhibition may involve specific binding of a drug to the haem iron of cytochrome P450 or the formation of an activated complex with P450. Notable drug-drug interactions involving cytochrome P450 are cimetidine inhibiting the metabolism of phenytoin, theophylline and warfarin, and erythromycin and ketoconazole inhibiting the metabolism of the H1 receptor antagonist, terfenadine. The latter interaction increases the risk of terfenadine-induced cardiac arrhythmias (Torsades de pointes). Dietary components, e.g. citrus juices, may also contribute to inhibition of drug metabolism. This has implications for dietary control during clinical trials. Novobiocin has caused jaundice in the newborn, arising from its inhibition of bilirubin conjugation with glucuronic acid. In the late stages of pregnancy, the high maternal levels of progesterone and pregnanediol inhibit the metabolism of drugs such as pethidine, barbiturates and coumarins.

Process of drug handling by the body

23

1.4.3.8 Induction of metabolism A number of drugs and other compounds, when repeatedly administered, can bring about an increase (induction) in the activity of the hepatic microsomal mixed function oxidases and other enzymes (e.g. UDP-glucuronyl-transferase, epoxide hydrase) as well as at other sites as a result of increased enzyme synthesis. Drugs such as phenobarbitone increase the level of cytochrome P450 and its associated reductase. Other examples of inducing agents in man are dichloralphenazone, phenylbutazone, griseofulvin, phenytoin, glutethimide, aminoglutethimide and rifampicin. Their effect is maximal after 2–3 weeks of repeated dosing. On stopping administration, enzyme levels revert to normal within 3–4 weeks. It is thought that these inducers, because of a high concentration or a slow metabolism, occupy the active sites of the enzyme to be induced for a prolonged period. This leads to derepression of gene function, followed by increased synthesis of the enzyme protein, and hence results in enhanced enzyme activity. The consequences of induction upon drug effects depend on the biological activity of the metabolites that are formed in increased amounts. Induction from phenobarbitone can reduce the hormonal effects of both endogenous and contraceptive steroids. Aminoglutethimide (an aromatase inhibitor) and tamoxifen (an oestrogen receptor antagonist) are each employed in the palliative management of advanced hormone-dependent breast cancer. The observation that addition of tamoxifen to the dose regimen of aminoglutethimide is no more effective than aminoglutethimide alone may be explained by the latter drug inducing the metabolism of tamoxifen to inactive products. Various foods such as Brussels sprouts, cabbage and charcoal broiled beef contain inducers of drug metabolism and this may need to be accounted for in clinical trials. 1.5 REMOVAL OF DRUGS FROM THE BODY Most drugs are removed from the body by the kidneys. For this to take place, the drug must either be water soluble being eliminated largely unchanged in the urine (Table 1.4), or more commonly, will have undergone metabolism in the liver to form more polar (i.e. less lipid-soluble) metabolites capable of being excreted in the urine. An alternative route of elimination is via the biliary system into the small intestine, the drug or its metabolites being available either for reabsorption into the blood stream (enterohepatic recycling) or for elimination in the faeces depending on its lipid solubility. Excretion in expired air

Table 1.4 Drugs that are excreted largely unchanged in the urine. Drug Unchanged Amiloride 75–100% Frusemide Gentamicin Methotrexate Atenolol

Introduction to the principles of drug design and action

Ampicillin Carbenicillin Cimetidine Cephaloridine Oxytetracycline

24

50–75%

occurs with highly volatile or gaseous agents such as the general anaesthetics (i.e. thiopentone sodium) whilst a relatively small amount of alcohol is eliminated in expired air which forms the basis of the breathalyser. Minor pathways of excretion include the saliva, skin and breast milk. This latter route sometimes poses problems for babies being breast-fed, since drugs such as amiodarone, aspirin and meprobamate may be transferred to the baby on feeding in large enough concentrations to produce a pharmacodynamic effect and possible toxicity. 1.5.1 Renal elimination The renal handling of drugs is a complex phenomena involving one or more basic processes: glomerular filtration, active tubular secretion and passive reabsorption across the renal tubule. Removal of drugs or their metabolites from the body by the kidneys is referred to as renal clearance. The efficiency of renal clearance may be expressed in terms of a hypothetical volume of plasma (ml) which is completely cleared of a drug or its metabolites by the kidneys per unit time (min). Most drugs are exponentially cleared from the body, the amount of drug cleared in unit time being proportional to the amount remaining in the body. Thus, clearance is theoretically never complete and so in practice it is convenient to measure the time to clear (eliminate) one-half of the drug from the body (i.e. half-life value). The elimination half-life values of a number of drugs are presented in Table 1.5. Such half-life values are particularly useful when deciding on the frequency of dosing. Substances such as creatinine and insulin are cleared by the kidneys with neither tubular secretion nor reabsorption occurring and consequently they have a renal clearance rate approximately equal to the rate at which plasma water is filtered. This is in the order of 125 ml min−1. Clearance values of a drug which are greater than this indicates that renal tubular secretion is taking place whilst lower clearance values indicate that the drug is undergoing tubular readsorption. For some drugs, secretion and adsorption processes may be taking place at the same time so care is needed when interpreting renal clearance values.

Table 1.5 Approximate elimination half-life values (h) of a number of drugs. Drug Suxamethonium Tubocurarine

Half- Drug life <0.5 Levodopa h Methyldopa

Half- Drug life Thiopentone

Oxprenolol

Tolbutamide

Theophylline

Half- Drug Halflife life Pentobarbitone Pimozide 6–8 Amitriptyline h

30– 40 h

Process of drug handling by the body

Aminosalicylic acid Benzylpenicillin Bupivacaine Cephalexin

Paracetamol

2–3 Vancomycin h

Chlorpropamide Warfarin

Pentazocine Streptomycin 0.5– Triamterene 1h

Cephalothin

Bethanidine Indomethacin Oxytetracycline

Ethacrynic acid

Amidopyrine

Probenecid

Inulin

Aspirin

Propantheline

Azathioprine

Sulphanilamide

Amoxycillin

25

Butobarbitone 8–10 Carbamazepine h Digoxin 40– 50 h Methaqualone Quinalbarbitone

Ampicillin

Chloram-phenicol 3–4 h Ephedrine Amantadine

Ethosuximide

Bacitracin

Pindolol

Chlorpheniramine

Nortriptyline

Chloro-thiazide

Terbutaline

Glibenclamide

Oxyphenbutazone

Cloxacillin

Vinblastine

Imipramine

Cromoglycate

Orphenadrine

Flurazepam

Heparin

1–2 Acetazolamide h Allopurinol

Phenformin

Caffeine

Dicoumarin

Methicillin

Chlor-tetracycline

Carbenoxolone

Orciprenaline

Hydrallazine

Phenacetin

Metoprolol

Vincristine

Pethidine

Hexamethonium Lignocaine

10– 15 h 60– 80 h Phenylbutazone

Quinidine

4–6 Chlorpromazine h Clonidine

80– 100 h Phenobarbitone 15– 20 h Barbitone

Griseofulvin

Chloroquine

Polymyxin Codeine Erythromycin Frusemide Gentamicin Insulin

Propranolol

50– 60 h

Digitoxin Haloperidol

20– Reserpine 24 h

Ouabain 2–3 Diphenhydramine h Sulphacetamide 6–8 Diazoxide h Sulphadimidine Lithium

24– 30 h

100– 120 h

Introduction to the principles of drug design and action

26

1.5.1.1 Glomerular filtration The kidneys receive 1.2 to 1.5 litres of blood per minute with approximately 10% being filtered at the glomerulus (i.e. the glomerular filtration rate). Whilst the pores of the glomerular capillaries are sufficiently large to permit the passage of most drug molecules, substances with a molecular weight of over 66,000 Da, including plasma proteins and drugs bound to these proteins, are held back in the plasma. With the exception of a few macromolecular substances, such as heparin and dextrans, all nonprotein-bound drugs pass freely across the glomerulus so that the drug concentration in the glomerular filtrate is the same as that of free drug in the plasma. After glomerular filtration, water is progressively reabsorbed from the filtrate so that only about 1 per cent of the original filtered volume is eventually voided as urine. 1.5.1.2 Active tubular secretion The proximal convoluted tubules of the kidneys are capable of actively transporting a wide range of substances from plasma into the tubular urine and this process is a major mechanism for the removal of acidic drugs and glucuronide metabolites. Although a transport system exists to handle basic substances, it is unlikely that the secretion of unmetabolised basic compounds contributes substantially to their removal from the body. Both systems are energy-dependent, the energy being derived from tubular cell metabolism. Many drugs and their metabolites are actively secreted in the proximal tubules and some of these are presented in Table 1.6. Penicillin removal from the body takes place mainly in the kidneys. It is a rapid process, usually on a single passage through the kidneys and accounts for the relatively short elimination half-life of 0.5– 1.0 hour. Interestingly, penicillin is 80% bound to plasma proteins and is only slowly cleared by glomerular filtration, whilst proximal tubular secretion is rapid and complete. Since tubular secretion is responsible for transporting drug molecules against an electrochemical gradient it is not only energy-dependent but drugs may compete for this same transportation mechanism, thus delaying the excretion of one or more of the competing drugs. For example, probenecid acts competitively for the same transport system as the penicillins in the proximal tubules and was the rational for administering probenecid with penicillin in the early use of this antibiotic. The result was increased plasma penicillin concentrations

Table 1.6 Some drugs and drug metabolites that are actively secreted into the proximal renal tubules. Acids Bases Chlorpropamide Amiloride Ethacrynic acid Dihydrocodeine Frusemide Dopamine Glucuronic acid Mepacrine conjugates Methotrexate Morphine

Process of drug handling by the body

Penicillins Phenylbutazone Probenecid Salicylic acid Sulphate conjugates Sulphathiazole Thiazide diuretics

27

Pethidine Procaine Quaternary ammonium compounds Quinine Thiamine Tolazoline Triamterene

and hence prolonged antibacterial activity. The development of the newer longeracting penicillins have replaced the need to administer probenecid concurrently with penicillin. 1.5.1.3 Passive reabsorption across the renal tubules Lipid-soluble drugs diffuse readily across the renal tubules, particularly in the distal region, and consequently are reabsorbed from the glomerular filtrate back into the bloodstream until the concentrations in plasma water and urine in the distal tubule are similar. The net result is that drugs with a high lipid-solubility are only slowly cleared from the plasma into the urine until they have been metabolized to more polar substances by the liver. In contrast, if a drug is highly polar it will not pass across the renal tubules and so will not be reabsorbed so that urinary elimination is rapid. Clearly, the degree of passive diffusion (i.e. readsorption) across the renal tubules is dependent on the physicochemical characteristics of a drug or it’s metabolite(s). Readsorption will also depend on physiological variables such as the rate of urine flow and urinary pH. For basic drugs including morphine-like analgesics and tricyclic antidepressants, renal excretion can be increased by acidifying the urine with ammonium chloride or a high protein diet. Since the degree of ionization of basic drugs is increased when the urine becomes more acid, the extent of readsorption is reduced thus increasing clearance. Conversely, when the urine is more acid, the ionization of weak acid drugs is decreased leading to enhanced readsorption and reduced clearance. The reverse situation occurs when the urine is more alkaline. For example, salicylates (being weak organic acids) are ionized in an alkaline urine and their excretion rates can be increased by administering sodium bicarbonate or feeding on a vegetable diet. Since urinary pH can vary over a considerable range, urinary excretion rates may vary widely between individuals. The different physico-chemical characteristics of a series of compounds may result in an active constituent being used for a variety of clinical conditions. For example, sulphonamides are particularly useful for treating both urinary tract and intestinal infections. However, for the former complaint high concentrations of sulphonamide are required in the urine for its antibacterial activity. For this purpose, sulphonamides such as sulphafurazole and sulphamethizole, have been developed for their high aqueous solubility so that rapid elimination takes place with little tubular readsorption.

Introduction to the principles of drug design and action

28

Since these drugs are relatively strong organic acids (pKa values 4.9, 5.5) they are completely ionized at physiological pH values. In contrast N4-succinyl and phthalyl sulphathiazoles have low water solubilities but are highly ionized at alkaline pH values Thus, these sulphonamides are only poorly absorbed after oral dosing, most of the administered dose being retained in the gastrointestinal tract. The carboxyl-amide group becomes hydrolysed by digestive bacterial proteinases, releasing the parent sulphonamide which is then available to treat intestinal infections such as bacillus dysentery. 1.5.1.4 Renal elimination in disease The rate of renal elimination of drugs which are extensively cleared by the kidneys will almost certainly be slowed in patients with renal impairment due to disease or in newborn babies and the very old. This will result in a prolongation of the drug’s elimination half-life value. Differences between normal subjects and anuric patients in eliminating a number of antibacterial drugs are presented in Table 1.7. When the elimination half-life of a drug is prolonged in renal failure, accumulation of the drug will occur, resulting in possible toxicity unless the dosage is reduced to correspond with the decreased rate of elimination.

Table 1.7 Elimination half-life values (h) of some antibacterial drugs in patients with either normal or impaired renal function. Drug Normals Anurics Benzylpenicillin 0.5 23.0 Cephalothin 0.5 12.0 Cephalexin 1.0 23.0 Erythromycin 1.4 5.5 Gentamicin 2.4 35.0 Streptomycin 2.5 70.0 Kanamycin 2.8 70.0 Lincomycin 4.7 12.0 Vancomycin 5.8 230.0 Tetracycline 8.5 87.0 Rifampicin 2.8 2.8 Doxycycline 23.0 23.0 1.5.2 Biliary elimination On passing through the liver, many drugs are actively transported from the bloodstream into bile which subsequently passes into the small intestine. If the drug is lipid-soluble then it may be reabsorbed in the intestine and become involved in enterohepatic cycling. Biliary secretion and intestinal reabsorption may continue until

Process of drug handling by the body

29

metabolism, renal and faecal excretion eventually results in the drug being eliminated from the body. Enterohepatic cycling of drugs such as oestrogens and digoxin results in delayed clearance and prolonged pharmacological activity. In contrast, highly water-soluble drugs will remain in the small intestine and be rapidly excreted in the faeces. Biotransformation of a lipid-soluble drug to a gluconide or other conjugate in the liver results in a metabolite which being polar will not be available for reabsorption in the small intestine. However, in some instances deconjugation may take place in the intestine, with the liberated parent drug being free for readsorption. Anions, cations, and non-ionized molecules containing both polar and lipophilic groups may be candidates for biliary excretion in man, provided that their molecular weights exceed about 400–500 Da. Substances of lower molecular weight may undergo reabsorption during their passage through the small canaliculi of the liver so that they are not available to pass into the bile. Examples of drugs which have a significant biliary component to their excretion pattern include the penicillins. In particular, ampicillin is well excreted into bile and so appreciable quantities of ampicillin are excreted in the faeces. This may cause concern in some patients since it may induce diarrhoea after several days of treatment with this particular antibiotic. The diarrhoea results from the ampicillin on entering the large bowel, killing many of the commensal organisms living in it, resulting in an overgrowth of antibiotic-resistant bacteria which directly leads to diarrhoea. 1.5.3 Elimination in other secretions Some drugs appear in saliva, where they may cause a disagreeable taste or irritate the oral tissues. The transfer of a drug from blood to saliva will depend on its lipid solubility, pKa and the extent to which plasma protein binding occurs. Since the average pH of saliva (6.5) is lower than the pH of plasma (7.4), it is most likely that for weakly acidic drugs the concentration of drug in the saliva will be less than that of free drug in the plasma, whereas for weak bases it will be greater in saliva. Some drugs are actively transported from blood to saliva (e.g. lithium) and in these circumstances the concentration of drug in saliva may be two or three times higher than in plasma. In general, non-ionized drugs of high lipid solubility and low molecular weight are more likely to pass into saliva than drugs of low lipid solubility. Almost any drug present in a lactating mother’s blood will appear in her breast milk, the concentration being dependent on the concentration of drug in maternal blood, its lipid solubility, degree of ionization and extent of plasma protein binding. While the amounts of most drugs in breast milk are relatively small, the immature hepatic and renal function of an infant may result in delayed metabolic inactivation and elimination of a drug, leading to unwanted effects in the infant (e.g. anthraquinone purgatives, diazepam, dichlorolphenazone, phenindione, lithium). In general, drugs should be avoided whenever possible during lactation, since there is only limited information on the possible harmful effects of drugs transferred from mothers to their breast-fed infants. Volatile substances, such as the inhalation anaesthetics, nitrous oxide and halothane, diffuse readily across the lipoidal barriers of alveolar membranes and are eliminated in expired air. Once administration of the anaesthetic is ceased its

Introduction to the principles of drug design and action

30

concentration in the lungs drops below that in the blood. Since these drugs are extremely fat soluble, they readily and rapidly pass back into the lungs form the bloodstream and are then exhaled in expired air, thus eventually terminating anaesthesia. 1.6 SUMMARY To be of potential use as a medicine, a drug must reach its site or sites of action in a concentration sufficient to initiate a therapeutic response. The concentration achieved, whilst being related to the dose of drug administered, will also depend on the rate and extent to which the drug is absorbed from its site of administration and its distribution in the bloodstream to other parts of the body. The characteristic effect of a drug will disappear when the drug is removed from the body and consequently from its site of action, either in an unchanged form or more generally after metabolism has taken place giving metabolites which can be more readily excreted. The processes of absorption, distribution, metabolism and excretion are thus important determinants that must be considered in the strategic planning of a new drug substance if the correct dose, route and form of drug administration is to be selected to produce the intended clinical effect with minimal unwanted side-effects. FURTHER READING Belaune, Ph. (1993) Les cytochromes P450 humains. Applications en pharmacologie. Therapie 48, 521–526. Benet, L.Z., Kroetz, D.L. and Sheiner, L.B. (1996) Pharmacokinetics: The Dynamics of Drug Absorption, Distribution and Elimination. In Goodman and Gilman’s The Pharmacological Basis of Therapeutics, 9th edn. pp. 3–27. New York: McGrawHill. Benet, L.Z., Massoud, N. and Gambertoglio, J.G. (1984) Pharmacokinetic basis for drug treatment. New York: Raven Press. Bradbury, M.W.B. (1984) The structure and function of the blood-brain barrier. Federation Proceedings 43, 186–190. Commandeur, J.N.M., Stijntjes, G.J. and Vermeulen, N.P.E. (1995) Enzymes and transport systems involved in the formation and disposition of the glutathione sconjugates: role in bioactivation and detoxification mechanisms of xenobiotics. Pharmacological Reviews 47, 271–330. Curry, S.H. (1980) Drug Disposition and Pharmacokinetics, 3rd edn. Oxford: Blackwell Scientific. Davies, D.A. (1991) Textbook of Adverse Drug Reactions, 4th edn. Oxford: Oxford University Press. Evans, W.E., Schentag, J.J. and Jusko, W.J. (1992) Applied Pharmacokinetics: Principles of Therapeutic Drug Monitoring, 3rd edn. Vancouver: W.A., Applied Therapeutics.

Process of drug handling by the body

31

Gibaldi, M. (1992) Biopharmaceutics and Clinical Pharmacokinetics, 4th edn. Philadelphia: P.A., Lea & Febiger. Gibaldi, M. and Perrier, D. (1982) Pharmacokinetics, 2nd edn. New York: Marcel Dekker Inc. Gibson, G.G. and Skett, P. (1994) Introduction to Drug Metabolism. 2nd edn. Glasgow: Blackie Academic and Professional. Guengerich, P.P. (1989) Characterization of human microsomal cytochrome P450 enzymes. Annual Reviews of Pharmacology and Toxicology 14, 259–307. Jeffery, E.H. (ed.) (1993) Human Drug Metabolism: from Molecular Biology to Man. Boca Raton: CRC Press. La Du, B.N., Mandel, H.G. and Way, E.L. (1971) Fundamentals of Drug Metabolism and Drug Disposition. Baltimore: Williams & Wilkins. Lamble, J.W. (1983) Drug Metabolism and Distribution. Amsterdam: Elsevier. Lewis, D.F.V. (1996) Cytochromes P450; Structure, Function and Mechanism. London: Taylor and Francis. Mitchell, J.R. and Horning, M.G. (1984) Drug Metabolism and Drug Toxicity. New York: Raven Press. Nebert, D.W., Nelson, D.R., Coon, J.J., Estabrook, R.W., Feyereisen, R., FujiiKuriyama, Y. et al. (1991) The P450 superfamily: update on new sequences, gene mapping and recommended nomenclature. DNA. Cell Biology 10, 1–14. Pacific, G.M. and Fracchia, G.N. (eds.) (1995) Advances in Drug Metabolism in Man, pp. 834. Brussels: ECSCEC-EAEC. Pardridge, W.M. (1988) Recent advances in blood brain barrier transport. Annual Reviews of Pharmacology and Toxicology 28, 25–39. Parke, D.V. (1968) The Biochemistry of Foreign Compounds. Oxford: Pergamon. Pratt, W.B. (1990) The entry, distribution and elimination of drugs. In Principles of Drug Actions, edited by W.B.Pratt and P.Taylor, 3rd edn., pp. 201–296. New York: Churchill Livingstone. Prescott, L.F. and Nimmo, W.S. (1981) Drug Absorption. New York: ADIS. Prescott, L.F. and Nimmo, W.S. (1989) Novel Drug Delivery. Chichester: John Wiley. Renwick, A.G. (1996) Pharmacokinetics and Toxicokinetics; Understanding Biodisposition Data. London: Taylor and Francis. Rowland, M. and Tozer, T.N. (1995) Clinical Pharmacokinetics, 3rd edn. Philadelphia: Lea & Febiger. Timbrell, J.A. (1995) Introduction to Toxicology, 2nd edn. London: Taylor and Francis. Williams, R.T. (1959) Detoxication Mechanisms, 2nd edn. London: Chapman and Hall.

2. THE DESIGN OF DRUG DELIVERY SYSTEMS IAN W.KELLAWAY CONTENTS 2.1 INTRODUCTION

29

2.2 FORMULATION AIMS

31

2.3 PHYSICO-CHEMICAL FACTORS INFLUENCING DRUG BIOAVAILABILITY

31

2.3.1 Rate of solution

32

2.3.2 Complexation

34

2.3.3 Drug stability

34

2.4 INFLUENCE OF ROUTE OF ADMINISTRATION AND TYPE OF DOSAGE FORM

35

2.5 FORMULATION FACTORS

35

2.5.1 Solutions

36

2.5.2 Emulsions

36

2.5.3 Soft gelatin capsules

37

2.5.4 Suspensions

37

2.5.5 Hard gelatin capsules

37

2.5.6 Tablets

38

2.5.7 Coated tablets

38

2.6 DRUG DELIVERY TO THE LUNG

38

2.6.1 Therapeutic aerosol generation and particle fate

39

2.6.2 Metered dose inhalers (MDIs)

40

The design of drug delivery systems

33

2.6.3 Nebulizers

41

2.6.4 Dry powder inhalers (DPIs)

41

2.6.5 Pulmonary drug selectivity and prolongation of therapeutic effects

42

2.6.5.1 Prodrugs

42

2.6.5.2 Polyamine active transport system

42

2.6.5.3 Rate control achievable by employing colloidal drug carriers

42

2.6.6 Delivery of drugs to the systemic circulation by the pulmonary route

43

2.7 SUSTAINED AND CONTROLLED RELEASE DOSAGE FORMS

44

2.7.1 Potential advantages of sustained controlled release products

45

2.7.2 Therapeutic concentration ranges and ratios

45

2.7.3 Dosage interval concentration ratio and rate of elimination

46

2.7.4 Mechanisms of achieving sustained release by the oral route

46

2.7.4.1 Hydrophilic gel tablets or capsules

46

2.7.4.2 Matrix tablets

46

2.7.4.3 Capsules containing pellets with disintegrating coatings

46

2.7.4.4 Pellets or tablets coated with diffusion controlling membranes

47

2.7.5 Positional controlled release

47

2.7.6 Gastrointestinal transit of sustained release dosage formulations

47

2.8 SITE-SPECIFIC DRUG DELIVERY

48

2.8.1 Carrier systems

48

2.8.2 Fate of site-specific delivery systems

49

2.9 BIOEQUIVALENCE

50

Introduction to the principles of drug design and action

FURTHER READING

34

50

2.1 INTRODUCTION Drugs are rarely, if ever, administered to patients in an unformulated state. The vast majority of the available medicinal compounds which are potent at the milligram or microgram levels could not be presented in a form providing an accurate and reproducible dosage unless mixed with a variety of excipients and converted by controlled technological processes into medicines. Indeed, the primary skills of the pharmacist lie in the design, production and evaluation of a wide range of dosage forms, each providing an optimized delivery of drug by the selected route of administration. The aims of this chapter, therefore, are to outline mechanisms by which the onset, duration and magnitude of the therapeutic responses can be controlled bv the designer of the drug delivery system. It has been appreciated for a considerable time that dosage forms possessing the same amount of an active compound (chemically equivalent) do not necessarily elicit the same therapeutic response. The rate at which the drug is liberated from the dosage form and the subsequent absorption, distribution, metabolism and excretion kinetics will determine the availability of the active species at the receptor site. The majority of systemically acting drugs are administered by the oral route and therefore must traverse certain physiological barriers including one or more cell membranes. Pro-drugs may alter this part of the overall rate process (see Chapter 7) although generally, control of plasma levels is achieved by modulation of the drug liberation process from the dosage form. The critical drug activity at the receptor site is usually related to blood and other distribution fluid levels, as well as elimination rates. Other factors affecting activity include deposition sites, biotransformation processes, protein binding and the rate of appearance in the blood. Hence in order to obtain the desired response, the drug must be absorbed both in sufficient quantity and at a sufficient rate. The term bioavailability is used to express the rate and extent of absorption from a drug delivery system into the systemic circulation. The crucial influence of rate as well as extent of absorption in considerations of bioavailability can be seen in Figure 2.1. The plasma levels are illustrated following a single oral administration of three chemically equivalent delivery systems (A, B and C) but with different drug liberation rates (A>B>C). Formulation A has a shorter duration of activity but results in a more rapid onset of activity compared with formulation B. The magnitude of the therapeutic response is also greater for A than B. Formulation C is therapeutically inactive, as the minimum effective plasma concentration (MEC) is not achieved. Therefore, unless a multiple dosing regimen is to be considered, C has no clinical value. It should also be noted that the plasma concentrations from A exceed the maximum safe concentration (MSC) and some toxic side-effects will be observed. Unless rapidity of action is of paramount importance and the toxic effects can be tolerated, B therefore becomes the formulation of choice. Generally, however, a rapid and complete absorption profile is required to eliminate variation in response due to

The design of drug delivery systems

35

physiological variables which include gastric emptying rate and gut motility. Bioavailability can also therefore be influenced by physiological and pathological factors, although in this chapter only the pharmaceutical or formulation aspects will be considered. Bioavailability may be assessed by the determination of the induced clinical response which, because it often involves an element of subjective assessment, makes quantitation

Figure 2.1 The influence of drug release rate on the blood level—time profile following the oral administration of three chemically equivalent formulations. MSC=maximum safe concentration; MEC= minimum effective concentration. difficult. Measurement of drug concentrations at the receptor site is not feasible; therefore, the usual approach is the determination of plasma or blood levels as a function of time making the implicit assumption that these concentrations correlate directly with the clinical response. Areas under the concentration—time profiles give the amount of drug absorbed and hence (if related to those of an intravenous solution of the same drug) permit an absolute bioavailability to be determined, whilst if related

Introduction to the principles of drug design and action

36

to a ‘standard’ formulation (often the original or formula of the patent holder) then the term ‘relative bioavailability’ is employed. The constraints of space dictate the limitation of both discussion and examples to the oral route, which is the most widely used route for systemically active agents, and the pulmonary route for which there is an interdependence between the device and the formulation in order to optimise drug efficiency. However, there are alternative routes to be considered including nasal, ocular, transdermal, buccal, vaginal and rectal— details are available from specialist textbooks (see Further Reading). 2.2 FORMULATION AIMS Formulation aims, in the light of bioavailability considerations, are to produce a drug delivery system such that: (a) A unit dose contains the intended quantity of drug. This is achieved by homogeneity during the manufacturing process and a suitable choice of excipients, stabilizers and manufacturing conditions to ensure both drug and product stability over the expected shelf-life. (b) The drug is usually totally released but always in a controlled manner, in order to achieve the required onset, intensity and duration of clinical response as previously outlined. Most dosage forms can be designed to give a rapid response: if, however, a long duration of response is required then it is easier to achieve sustained release using solid rather than liquid formulations.

2.3 PHYSICO-CHEMICAL FACTORS INFLUENCING DRUG BIOAVAILABILITY Drug concentrations in the blood are controlled either by the rate of drug liberation from the dosage form or by the rate of absorption. In many cases it is the drug dissolution rate that is the rate-determining step in the process. Dissolution is encountered in all solid dosage forms, i.e. tablets and hard gelatin capsules as well as suspensions, whether intended for oral use or administration via the intramuscular or subcutaneous routes. If absorption is rapid, then it is almost inevitable that drug dissolution will be the rate-determining step in the overall process and hence any factor which affects the solution process will result in changes in the plasma—time profile. Hence the pharmacist has the opportunity of controlling the onset, duration and intensity of the clinical response by controlling the dissolution process.

The design of drug delivery systems

37

2.3.1 Rate of solution Dissolution of a drug from a primary particle in a non-reacting solvent can be described by the Noyes-Whitney equation

(2.1) where dw/dt is the rate of increase of the amount of drug dissolved; k is the rate constant of dissolution; cs the saturation solubility of the drug in the dissolution media; c the concentration of drug at time t; A is the surface area of drug undergoing dissolution; D the diffusion coefficient of the dissolved drug molecules and h, the thickness of the diffusion layer. Hence it can be readily appreciated that the dissolution rate is dependent on the diffusion of molecules through the diffusion layer of thickness h. Closer examination of this equation will demonstrate some of the mechanisms for controlling solution rate. (1) dw/dt A. Reduction in the particle size of the primary particle will result in an increase in surface area and hence more rapid dissolution will be achieved. A change in the shape of the plasma—time profile will result and it is possible also to increase the area under this curve, which of course means an increase in bioavailability. It is therefore possible to achieve a reduction in the time necessary for the attainment of maximum plasma levels, an increase in the intensity of the response and an increase in the percentage of the dose absorbed. Griseofulvin is one of the most widely studied drugs in relation to bioavailability, as this poorly water-soluble, antifungal drug exhibits a striking example of dissolution rate-limited absorption. Plasma levels have been shown to increase linearly with an increase in specific surface area and thus, despite the cost of micronization, griseofulvin is marketed as a preparation in this form because identical blood levels can be achieved by using half the amount of drug present in the unmicronized formulation. Micronization, however, is not the only solution to the griseofulvin bioavailability problem. For example, microcrystalline dispersions have been formed in a water-soluble solid matrix in which the dispersion state is determined by the preparative procedures, some of which result in true solid solutions. The two most widely accepted approaches are (a) crystallization of a melt, resulting from fusing of drug and carrier and (b) co-precipitation of drug and carrier from a common organic solvent. In the latter case a griseofulvin— polyvinylpyrrolidone dispersion resulted in a ten-fold increase in solution rate, compared with a micronized preparation. It should be emphasized, however, that griseofulvin is at the extreme end of the bioavailability spectrum! For drugs exhibiting good aqueous solubility, little is to be gained by reducing the particle size of the drug, as plasma levels are unlikely to be dissolution rate-limited. Indeed, if enzymatic or acid degradation of the drug occurs in the stomach, then increasing dissolution rates by reducing particle size can result in reduced bioavailability. (2) dw/dt cs. Many drugs are weak acids or bases and hence exhibit pH-dependent solubility. It is therefore possible to increase cs in the diffusion layer by adjustment of

Introduction to the principles of drug design and action

38

pH in either (a) the whole dissolution medium or (b) the microenvironment of the dissolving particle. The pH of the whole medium can be changed by the coadministration of an antacid. This raises the pH of the gastric juices and hence enhances the dissolution rate of a weak acid. However, this is rarely a practical proposition and therefore most pH adjustments are made within the very localized environment of the dissolving drug particles. Solid basic substances may be added to a weakly acidic drug, which raises the pH of the microenvironment. Probably the best known example is that of buffered aspirin products which use the basic substances sodium bicarbonate, sodium citrate or magnesium carbonate. Rather than employ another agent to alter the pH, a highly water-soluble salt of the drug can be equally, if not more, effective. The dissolving salt raises the pH of the gastric fluids immediately surrounding the dissolving particle. On mixing with the bulk of the gastric fluids the free acid form of the drug will be precipitated, but in a microdispersed state with a large surface area to volume ratio which will rapidly redissolve. The process is represented diagrammatically in Figure 2.2. Many examples exist to illustrate the importance of salt formation on bioavailability. One such example is provided by the antibiotic novobiocin, where the bioavailability was found to decrease in the order, sodium salt>calcium salt>free acid. The dissolution rate of weak bases can be similarly changed by salt formation: however, dissolution rate-limited absorption is less important for bases than acids. This is because little absorption occurs in the stomach where the bases are ionized: most of the drug being absorbed post gastric emptying and this delay compensates any benefits accruing from more rapid solution rates. However, basic drugs are often administered as salts, e.g. phenothiazines and tetracyclines, to ensure that gastric emptying, and not dissolution, will be the rate-limiting factor in the absorption process. A large number of drugs exhibit polymorphism, that is, they exist in more than one crystalline form. Polymorphs exhibit different physical properties including solubility, although only one polymorph will be stable at any given temperature and pressure. Others may exist in a metastable condition, reverting to the stable form at rates which may permit their use in drug delivery systems. The most desirable property of the metastable forms is their inherently higher solubility rates which arise from the lower crystal lattice energies. Amorphous or non-crystalline drugs are always more soluble than the corresponding crystalline form because of the lower energy requirements in transference of a molecule from the solid to the solution phase. Crystalline novobiocin dissolves slowly in vitro

The design of drug delivery systems

39

Figure 2.2 The dissolution of a highly watersoluble salt of a weak acid in the stomach. compared with the amorphous form, kinetics which correlate well with bioavailability data. Amorphous chloramphenicol stearate is hydrolysed in the gastrointestinal tract to yield the absorbable acid, whilst the crystalline form is of such low solubility that insufficient is hydrolysed to give effective plasma levels. Solvates are formed by some drugs: when the solvent is water, the hydrates dissolve more slowly in aqueous solutions than the anhydrous forms, e.g. caffeine and glutethimide. For ampicillin, greater bioavailability has been shown for the higher energy form anhydrate than the trihydrate, which illustrates the dependence of solubility and dissolution rates on the free energy of the molecules within the crystal lattice. Conversely, organic solvates such as alkanoates dissolve more rapidly in aqueous solvents than the desolvated forms. 2.3.2 Complexation Increased solubility or protection against degradation may be achieved by complex formation between the drug and a suitable agent. Complexes may also arise unintentionally as a result of drug interaction with an excipient or with substances occurring in the body. Complex formation is a reversible process and the effect on bioavailability is often dependent on the magnitude of the association constant. As most complexes are non-absorbable, dissociation must therefore precede absorption. The formation of lipid-soluble ion-pairs between a drug ion and an organic ion of opposite charge would result in greater drug bioavailability. Rarely have such results

Introduction to the principles of drug design and action

40

been achieved, presumably due to the dissociating influence of the mucosa and the poor membrane partitioning of the bulky ion-pair. Surfactants are used in a wide range of dosage forms often to increase particle wetting, control the stability of dispersed particles, and to increase both solution rates and the equilibrium solubility by the process of solubilization. Bioavailability may, however, be enhanced or retarded and often exhibits surfactant concentration dependent effects. Below the critical micelle concentration (CMC), enhanced absorption may be encountered due to partition of the surfactant into the membrane, which results in increased membrane permeability. At post-CMC levels, the dominant effect is the ‘partitioning’ of the drug into the micelle, a lower drug thermodynamic activity results and absorption is reduced. Micellar solubilization of membrane components with a loss of membrane integrity can also occur. Thus it is not easy to predict the effect of surfactants on bioavailability for, although dissolution rates will be increased by high concentrations of surfactant, the effect on the absorption phase may be complex. 2.3.3 Drug stability Drug stability, in addition to being of paramount importance to product shelf-life, can also affect bioavailability. Some therapeutic substances are degraded by the acid conditions of the stomach or by enzymes encountered in the gastrointestinal tract. Reduced or zero therapeutic effectiveness will result. Penicillin G is an example of a drug rapidly degraded in the stomach and for which enteric coating is not a solution to the problem, as the drug is poorly absorbed from the small intestine. The semisynthetic penicillins such as ampicillin and amoxacillin show much greater acid stability. Improved bioavailability of acid-labile drugs can sometimes be achieved by reducing the rate of drug release from the dosage form. 2.4 INFLUENCE OF ROUTE OF ADMINISTRATION AND TYPE OF DOSAGE FORM Although many routes exist for the administration of a systemically acting drug (including parenteral, rectal, vaginal, pulmonal, nasal, transdermal, etc.) by far the most popular is the oral route. Bioavailability, in addition to being dependent on the route of administration, will also be influenced by the dosage form selected. Although it is not possible to generalize completely regarding the relative drug release rates and hence bioavailabilities from different dosage forms. Table 2.1 attempts to provide guidelines. It is however possible, for example, to produce a tablet with bioavailability equivalent to an aqueous solution! Aqueous solutions are rarely used due to solubility, stability, taste and non-unit dosing problems. The use of oils as drug carriers either as an emulsion, in which homogeneity and flavour masking are important, or in a soft gelatin capsule, provides efficient oral dosage forms. The release of the oil from the soft gelatin capsule shell is rapid but the surface area of the oil/water interface is lower than in an emulsion and hence partitioning of the drug is slower. Suspensions are suited to drugs of low

The design of drug delivery systems

41

solubility and high stability. Although a large surface area is provided, a dissolution stage nevertheless exists. On proceeding along the sequence from powders to hard gelatin capsules to tablets (see Table 2.1), the particles become more compacted and hence the deaggregation/dissolution phase becomes longer (see Figure 2.3). 2.5 FORMULATION FACTORS It should by now be appreciated that, by design, it is possible to formulate a potent, well-absorbed drug in such a manner that it is essentially non-absorbable. Hence the pharmacist with his unique skills in designing drug delivery systems can significantly influence the therapeutic efficacy of a drug. In most cases, the formulator can only influence bioavailability if the drug release phase is the rate-controlling step in the overall process.

Table 2.1 The ranking of dosage forms for oral administration with respect to the rate of drug release. Aqueous solutions Emulsions Soft gelatin capsules Suspensions Powders Increasing release rates and bioavailability Granules Hard gelatin capsules Tablets Coated tablets

Introduction to the principles of drug design and action

42

Figure 2.3 Summary of the processes following oral administration of dosage forms. Processes (a) dissolution; (b) deaggregation; (c) disintegration; (d) partitioning; (e) dispersion; (f) precipitation; (g) absorption. 2.5.1 Solutions As the drug is in a form readily available for absorption, few problems should exist. However, if the drug is a weak acid or a cosolvent is employed, then precipitation of the drug in the stomach may take place. Rapid redissolution of these ‘microprecipitates’ normally occurs. Aqueous solutions will require the addition of a suitable selection of colours and flavours to minimize patient non-compliance, and preservatives and perhaps buffers to optimize stability. Such factors would be elucidated in preformulation studies. 2.5.2 Emulsions The use of oral emulsions is on the decline. Most oils are unpalatable and an emulsion is an inherently unstable system. The choice of carrier oil dictates the extent and rate of drug partitioning between the oil and water. Emulsifying agents are either a mixture

The design of drug delivery systems

43

of surfactants or a polymer. Polymers may also be present to control the rheological properties of the emulsion and achieve an acceptable rate of creaming. The effect of surfactants on bioavailability has been previously discussed. Polymers can form nonabsorbable complexes with drugs and an increase in viscosity brought about by ‘thickening agents’ can delay gastric emptying which in turn may affect absorption. Viscosity effects, however, are not likely to be encountered with small dose volumes (5–10 ml). 2.5.3 Soft gelatin capsules After rupture of the glycero-gelatin shell, a crude emulsion is formed when the oil containing the drug is dispersed in the aqueous contents of the gastrointestinal tract. Oils are not always used to fill soft gelatin capsules; indeed occasionally watermiscible compounds such as polyethylene glycol 400 are used as vehicles. Soft gelatin capsules are a convenient unit dosage form generally exhibiting good bioavailability. 2.5.4 Suspensions A high surface area of the dispersed particles ensures that the dissolution process begins immediately the administered dose is diluted with the fluids of the gastrointestinal tract. Most pharmaceutical suspensions mav be described as coarse, that is they have particles in the size range 1–50 µm. Colloidal dispersions are expensive to produce and the theoretically faster solution rates arising from increased surface area are often offset by the spontaneous aggregation of the particles due to the possession of high surface energy. Particles>50 µm result in poor suspensions with rapid sedimentation, slower solution rates and poor reproducibility of the unit dose. In order to achieve desirable settling rates and ease of redispersion of the resulting sediments, controlled flocculation of the suspension is necessary. This is normally achieved by the use of surfactant or polymers, both of which may significantly influence drug bioavailability for reasons previously discussed. Polymers are also used, as with emulsions, as thickening agents to achieve the desired bulk rheological properties. On storage, the particle size distribution of suspensions may change with the growth of large particles at the expense of small. Hence solution properties and bioavailability may well be altered on storage. 2.5.5 Hard gelatin capsules It might be assumed that powders distributed into loosely packed beds within a rapidly dissolving hard gelatin capsule would not provide bioavailability problems. However, in practice, this is not true. One of the classic bioavailability cases in the pharmaceutical literature arose when the primary excipient in phenytoin capsules, calcium sulphate dihydrate, was substituted (by the manufacturing company in Australia) by lactose. Minor adjustments were also made to the magnesium silicate and magnesium stearate levels. The overall effect was that previously stabilized epileptic patients suddenly developed the symptoms associated with phenytoin

Introduction to the principles of drug design and action

44

overdose. It is now generally accepted that the calcium ions form a poorly absorbable complex with phenytoin. Another study demonstrated the reduced bioavailability of tetracycline from capsules in which calcium sulphate and dicalcium phosphate were used as fillers. The calcium—tetracycline complex formed in such formulations is poorly absorbed from the gastrointestinal tract. The choice and quantity of lubricant employed can greatly influence bioavailability. Even with a water-soluble drug it is possible to vary the drug release patterns from rapid and complete to slow and incomplete. With hydrophobic drugs, the problems can be even more acute. Hence, hydrophilic diluents should be employed to aid the permeation of aqueous fluids throughout the powder mass, reduce particle clumping and hence increase solution rates. 2.5.6 Tablets For economic reasons as well as for the convenience of the patient, the compressed tablet is the most widely used dosage form. However, by virtue of the relatively high compression forces used in tablet manufacture, together with the inevitable need of a range of excipients (including fillers, disintegrants, lubricants, glidants and binders), tabletting of drugs can give rise to serious bioavailability problems. As was seen in Figure 2.3, the active ingredient is released from the tablet by the processes of disintegration, deaggregation and dissolution: the latter occurring, however, at all stages in the overall liberation process. The rate-limiting step is normally dissolution, although by the use of insufficient or an inappropriate type of disintegrant, disintegration may become the all-important rate-limiting step. Division of the disintegrant between the granule interior and the intragranular void spaces can accelerate the disintegration process. Several interdependent factors determine disintegration rates, including concentration and type of drug, the nature of diluent, binder and disintegrant as well as the compaction force. High compression forces will often result in the retardation of disintegration due to reduced fluid penetration and extensive interparticulate bonding. Soluble drugs and excipients may lead to a decrease in disintegration due to the local formation of viscous solutions. The effect of hydrophobic lubricants is similar to that observed for capsules. The method by which the lubricant is incorporated, as well as the efficiency of mixing, have also been shown to influence drug dissolution rate from tablets. When the excipient-drug ratio is increased, thus increasing tablet size, solution rates of poorly water-soluble drugs also increase. 2.5.7 Coated tablets The application of an outer coat to a tablet presents a further barrier between the fluids of the gastrointestinal tract and the drug particles and one which is the first to be dissolved or ruptured prior to the fluid penetration of the tablet mass. Film coats are usually thin and readily soluble and hence would be expected to have but a negligible effect on bioavailability. The more traditional sugar coat is similarly water soluble.

The design of drug delivery systems

45

Enteric coatings can give rise to considerable variations in drug plasma levels due primarily to variation in stomach residence times, which, for non-disintegrating tablets, can vary between 1·5 and 6 hours. As a single enteric coated tablet or capsule empties from the stomach in an all or none manner, better control of the plasma concentration—time profile is obtained by the use of individually enteric-coated granules either packed into a capsule or compressed into a rapidly disintegrating tablet. 2.6 DRUG DELIVERY TO THE LUNG Drug delivery to or via the respiratory tree has been a long-standing pharmaceutical objective. For locally acting agents it is desirable to confine the action of the drug to the lung in order to eliminate unintended side effects which might result following absorption and distribution to other extravascular sites. Oral inhalation is often the preferred route in order that such effects be minimised. The large surface area for absorption provided by the alveolar region, together with reduced extracellular enzyme levels compared with the gastrointestinal tract, ensures that pulmonary administration is a potentially attractive route for the delivery of systemically active agents including the new generation of biotechnology molecules. 2.6.1 Therapeutic aerosol generation and particle fate There are three principal types of aerosol generators currently used in inhalation therapy, viz. the pressurized pack (metered-dose) inhaler (MDI), the nebulizer for continuous administration and the unit-dose dry powder inhaler (DPI). The pharmaceutical formulator is not only concerned with the drug formulation but also the selection of the appropriate device as it is the intimate relationship between device and formulation that leads to optimal drug deposition within the lower respiratory tract. The latter consists of the bronchial and pulmonary regions and in order to deliver drug to these regions, the polydispersed therapeutic aerosol containing particles/droplets of the drug should ideally be in the size range of 2–5 µm in diameter. The influence of particle diameter in determining deposition site is illustrated in Figure 2.4, where the fraction deposited in the alveolar and tracheo-bronchial regions of the lung is shown as a function of aerodynamic particle diameter. Tracheobrochial deposition may occur by various mechanisms but inertial impaction, sedimentation and Brownian diffusion predominate. Mouth breathing—the normal route of pulmonary delivery of medicinal agents—bypasses the nasal removal of large particles, which are

Introduction to the principles of drug design and action

46

Figure 2.4 Particle diameter dependence of alveolar and tracheobronchial deposition for mouth breathing. Tidal volume 1 1, breathing frequency 7.5/min, mean flow rate 250 cm3/s, inspiration/expiration times 4 s each. (Reproduced from Routes of Drug Administration. Eds. A.T.Florence and E.G.Salole (1990), p. 53. London: Wright. therefore deposited in the throat and part of the tracheobronchial region. In the bronchioles, ciliated cells are dominant and in conjunction with mucus secreted by goblet cells and submucosal glands, constitutes the ‘mucociliary escalator’ which ensures rapid (within hours) removal of insoluble or slowly soluble deposited particles by transport to the mouth for subsequent swallowing. Soluble particles, in contrast, dissolve and may enter the bloodstream. Particles penetrating to the pulmonary compartment may be retained on the pulmonary surfaces as a result of settling, diffusion and interception processes. Several mechanisms ensure clearance, including dissolution with absorption, phagocytosis of particles by macrophages with translocation to the ciliated airways, and lymphatic uptake. Aerosol characteristics will therefore determine the depth of penetration within the airways and hence particle fate.

The design of drug delivery systems

47

2.6.2 Metered Dose Inhalers (MDls) This is a sprayable product in which the propellant force is a liquified or compressed gas (Figure 2.5). They are currently the major device used by domiciliary patients and consist of a container hermetically sealed by a metering valve and composed of aluminium or glass protected with a plastic outer casing. As most drugs are of low propellant solubility, they are frequently formulated as micronised suspensions. Stability is achieved by the addition of surfactants such as lecithin, oleic acid or sorbitan esters which also serve as a lubricant of the metering valve assembly. Solution formulations may be achieved by addition of a cosolvent such as ethanol or by solubilization in the added surfactant. Chlorofluorocarbons (CFC’s) are currently the propellant of choice, blended to achieve a vapour pressure of 350–450 kPa although the Montreal Protocol on Substances that Deplete the Ozone Layer calls for the phasing out of CFC’s by the year 2000. Replacement propellants are being investigated with the hydrofluorocarbons HFA-134A and HFA-227 being the most likely substitutes. In 1995 a number of products were marketed using hydrofluorocarbon propellants.

Figure 2.5 Diagram of a metered dose inhaler. Within the container is the drug formulation which typically comprises micronised drug suspended in the propellant and stabilized by a surfactant. (Reproduced from Morén, F. (1981). Pressurized aerosols for oral inhalation. Int. J. Pharm. 8, 1–10).

Introduction to the principles of drug design and action

48

2.6.3 Nebulizers Nebulizers are devices for converting aqueous solutions or micronised suspensions of drug into an aerosol. This is effected by two principal mechanisms, either high velocity airstream dispersion (the air-jet nebulizers) or by ultrasonic energy dispersion (the ultrasonic nebulizers). The former require a source of compressed gas (cylinder or air compressors) and hence tend to be more frequently encountered in hospitals than the domiciliary environment. Ultrasonic nebulizers are, in contrast, easily portable but, although producing a dense aerosol plume, often the population of droplets have a higher mass median aerodynamic diameter compared with the air jet nebulizers. Drug formulations for use in nebulizers are, wherever possible, aqueous solutions. Selection of appropriate salts and pH adjustment will usually permit the desired concentration to be achieved. If this is not feasible, then the use of cosolvents such as ethanol and/or propylene glycol can be considered. However, such solvents change both the surface tension and viscosity of the solvent system which, in turn, influence aerosol output and droplet size. Water insoluble drugs can be formulated by either micellar solubilization or by forming a micronised suspension. Nebulizer solutions are often presented as concentrated solutions from which aliquots are withdrawn for dilution before administration. Such solutions require the addition of preservatives, e.g. benzalkonium chloride and antioxidants (e.g. sulphites). Both excipient types have been implicated with paradoxical bronchospasm and hence the current tendency to use small unit-dose solutions that are isotonic and free from preservatives and antioxidants. Nebulizers of different design produce aerosols of different output and particle size of droplets. For maximum efficacy, the drug-loaded droplets need to be less than 5 µm. In the treatment of prophylaxis of Pneumocystis carinii pneumonia with nebulized pentamidine and where the target is the alveolar space, it is desirable to use nebulizers capable of generating droplets of less than 2 µm. During the nebulization from air jet nebulizers, cooling of the reservoir solution occurs which, together with vapour loss, results in concentration of the drug solution. This can lead to drug recrystallisation with subsequent blockage within the device or variation in aerosol droplet size. In contrast, ultrasonic nebulization results in a rise in solution temperature and a decrease in aerosol size. Although aerosol size distributions are a critical determinant of effective pulmonary drug delivery, it is also desirable to consider output in selection of a nebulizer. For most applications drug administration should occur over a maximum of 10–15 minutes to optimise patient compliance. 2.6.4 Dry Powder Inhalers (DPls) These breath activated devices aerosolise a set dose of micronised drug on an airstream. The earliest devices consisted of the micronised drug contained within a single-dose capsule which often contained lactose as an inert drug carrier and diluent. On rapid inhalation, mechanical deaggregation of the powder occurs but the high inertia ensures a significant deposition of the powder on the back of the throat. DPls

The design of drug delivery systems

49

tend to be even less efficient than MDIs but because of the higher doses employed, an equivalent therapeutic effect can be achieved. Multidose systems are now available, e.g. Diskhaler® and Turbuhaler®, the latter functioning at low inspiratory flow rates with the capability of delivering, for example, 200×1 mg doses of terbutaline sulphate. 2.6.5 Pulmonary drug selectivity and prolongation of therapeutic effects 2.6.5.1 Prodrugs In addition to improved selectivity of action in the lung relative to other organs, it is possible to obtain prolongation of therapeutic effects and enhancement of pulmonary activity by the design of appropriate prodrugs. Lung accumulation from the blood pool is achieved by many drugs which are both highly lipophilic and strongly basic amines. Such drugs exhibit very slowly effluxable lung pools. Lung tissue exhibits high nonspecific esterase activity which is species dependent and capable of cleaving carboxylate or carbonate ester linkages. In vivo prodrug conversion to active drug moiety can be controlled by use of different aliphatic or aromatic coupling agents, together with stereochemical modifications. Terbutaline (2.1) is an example of a bronchodilator drug for which a number of prodrugs exist. Terbutaline exhibits little affinity for lung tissue being rapidly absorbed following inhalation with peak plasma concentrations occurring within 0.5 h. The diisobutyryl ester (ibuterol) (2.2) results in an increased bioavailability of 1.6 fold over terbutaline following oral administration. However, it is 3 times as effective as terbutaline post-inhalation in inhibiting bronchospasm. Enhanced effects are attributable to more rapid absorption and better tissue penetration. Bambuterol (2.3) is the bis-N,N-dimethylcarbonate of terbutaline and as such is well absorbed from the gastrointestinal tract and is relatively resistant to hydrolysis leading to a sustained release oral product. However, it is not readily metabolised in the lung which precludes its administration by the pulmonary route. 2.6.5.2 Polyamine active transport system The cell types which accumulate polyamines such as endogenous putrescine, spermidine and spermine, together with compounds such as paraquat, are the Clara cells and the alveolar Type I and Type II cells. 2.6.5.3 Rate control achievable by employing colloidal drug carriers Control of the duration of local drug activity and of the plasma levels of systemically active agents may be achievable by employing a colloidal carrier possessing appropriate drug release characteristics. Tracheobronchial deposition of such carriers may not be

Introduction to the principles of drug design and action

50

desirable as their clearance will occur in a relatively short time period on the mucociliary escalator. Pulmonary deposition will, in contrast, result in extended clearance times which may be dependent upon the composition of the colloid. The mechanism by which clearance is effected will also vary, but will involve alveolar macrophage uptake, with subsequent metabolism or deposition on to the mucus blanket in the ciliated regions or lymphatic uptake. Colloidal carriers, of which liposomes are an example, can therefore control both drug delivery rates and availability. Technological problems, however, exist such as the design of delivery

The design of drug delivery systems

51

devices to ensure deposition in the appropriate regions of the lung without degradation or loss of entrapped drug. Toxicological considerations, foremost amongst which is the processing of the colloid, also require to be addressed. 2.6.6 Delivery of drugs to the systemic circulation by the pulmonary route The large surface area, thin epithelial membrane provided by Type I cells and a rich blood supply, ensures that many compounds are readily transported from the airways into the systemic circulation. Gaseous anaesthesia and oxygen therapy are examples of efficient clinical utilisation of the pulmonary absorption process. Compounds are absorbed by different processes including active transport and passive diffusion through both aqueous pores and lipophilic regions of the epithelial membranes. Absorption can be both rapid and efficient; for example, sodium cromoglycate is well absorbed from the lung whereas less than 5% is absorbed from the gastrointestinal tract. Small lipophilic molecules, such as the gaseous anaesthetics, are absorbed by a nonsaturable passive diffusion process. Hydrophilic compounds are absorbed more slowly and generally by a paracellular route. Aqueous pores are, by virtue of their size, capable of controlling the rate and extent of hydrophilic compound absorption. Sodium cromoglycate is absorbed by both active and passive (paracellular) mechanisms. The rates of absorption by the paracellular route decreases as the molecular weight of the compound increases. The efficiency of absorption from the lung is species dependent. For example, insulin is absorbed from the human lung but less efficiently than in the rat or rabbit. Human growth hormone (molecular weight 22 kDa) is absorbed from the lungs of hypophysectomised rats with an estimated bio-equivalence of 40% relative to the subcutaneous route and an absolute bioavailability of 10%, sufficient to induce growth. A nonapeptide (leuoprolide acetate) has been shown to have an absolute bioavailability following aerosolization to healthy male volunteers of between 4 and 18% which, when corrected for respirable fraction, corresponds to 35–55%. Protein absorption, however, is postulated to occur through the extremely thin Type I cells by the vesicular process of transcytosis. The passage from lung to blood of proteins in the rat has recently been shown to increase during inflammatory conditions with the observed transport correlating to the severity of the lung injury. The pulmonary route therefore warrants further investigation for the systemic delivery of peptides and proteins. 2.7 SUSTAINED AND CONTROLLED RELEASE DOSAGE FORMS Figure 2.6 illustrates the differences between three distinct drug release profiles achieved by the use of (A) the usual single dose preparation, (B) a sustained release preparation and (C) a prolonged release preparation. Sustained release products are rarely achieved in practice although in many respects they represent an ideal delivery

Introduction to the principles of drug design and action

52

system. Initially a loading dose is rapidly released from the sustained action delivery system to provide the necessary blood levels to elicit the desired pharmacological response. The remaining fraction of the dose (maintenance dose) is then released from the preparation at rates which ensure the maintenance of a constant blood level. Prolonged action delivery systems merely extend the duration of the pharmacological response compared with the usual single dose preparation. Not all drugs are suitable candidates for prolonged action medication as (a) the drug must be absorbed efficiently over a substantial portion of the gastrointestinal tract, (b) the drug must possess a reasonably short biological half-life (<12 hours), (c) the size of the prolonged dosage form must not be too large for ease of swallowing, i.e. the drug must be effective at a ‘reasonable’ dose level and (d) the pharmacological activity of the drug should be clinically desirable. In some instances, the latter has been questioned if tolerance to the drug may result. It should be noted that various terms have been employed to describe oral dosage forms which provide long-term therapeutic action. These include ‘sustained’, ‘prolonged’, ‘slow’, ‘gradual’, ‘timed’, ‘extended’, and ‘controlled’. Often such terms are used interchangeably, although ‘controlled’ should be reserved for drug delivery systems

Figure 2.6 The difference between sustained and prolonged release dosage forms as illustrated by the blood concentration— time profiles. where the rate of drug release is determined solely by the device and is therefore independent of any anatomical or physiological constraints.

The design of drug delivery systems

53

2.7.1 Potential advantages of sustained controlled release products Prolonged drug absorption and reduced peak blood concentrations are two obvious advantages of effectively designed sustained release products. As a consequence of the prolonged absorption phase, therapeutic effects should also be extended and a more regular and even pattern established. It has been claimed that a reduction in dosing frequency to once daily will lead to improved compliance and a concomitant reduction in unwanted side-effects from high peak blood levels. Irritant drugs such as nonsteroidal anti-inflammatory agents, which are slowly released within the gut, should result in reduced inflammatory responses in the gastric mucosa. 2.7.2 Therapeutic concentration ranges and ratios Rapid diffusion of the drug across capillary walls will result in equilibrium drug concentrations at the target site equivalent to the free serum concentration. Under these conditions, concentration-effect relationships can be established. Often it is difficult to define both MSCs and MECs due to variability in reported values (Figure 2.4). Therapeutic ranges are often related to patient age, disease state and concomitant therapy. The presence of active metabolites and intersubject variation in plasmaprotein binding often complicates both the highest tolerable and the minimum therapeutic concentrations; the ratio of which is the therapeutic concentration ratio (TCR). For drugs exhibiting a low TCR, it is critical to minimize variations in peak and trough plasma concentrations. Hence, a controlled release dosage form becomes highly desirable because, in addition, minimization of variations in serum concentration between doses will achieve both increased therapeutic effectiveness and safety. 2.7.3 Dosage interval concentration ratio and rate of elimination The dosage interval concentration (DICR) is the ratio of the peak to the minimum plasma concentration achieved during a single dosing interval. It is dependent on and will increase with dosing frequency and absorption and elimination rates (see Chapter 1). Elimination rate is an intrinsic property of the drug molecule and therefore, unlike absorption rate, cannot be controlled by formulation factors. Minimization of the DICR for rapidly cleared drugs can be achieved by frequent administration, resulting in patient inconvenience and hence poor compliance, or by the more pragmatic approach, the design of sustained release formulations in order to prolong the absorption phase. 2.7.4 Mechanisms of achieving sustained release by the oral route Oral sustained release products have been produced employing various drug release mechanisms. Unfortunately, no single approach is universally acceptable and selection is inevitably related to drug properties. For bona fide controlled release, in vitro release profiles are superimposable on those achieved in vivo. If drug release is pH

Introduction to the principles of drug design and action

54

dependent, then greater variability in in vivo performance is to be expected. Excluding molecular modification in order to change drug solubility (prodrug, salt or complex formation) and the use of the limited applicability of ion exchange resins, the design of sustained release products is often accompanied by employing one or more of the following approaches. 2.7.4.1 Hydrophilic gel tablets or capsules Hydrophilic gums are used to form a gel layer surrounding the tablet upon introduction to an aqueous medium. Diffusion across this layer constitutes the ratedetermining release step. Nitroglycerin (glyceryl trinitrate) dispersed in hydroxypropylmethylcellulose for buccal administration permits sustained release for over 4 hours. Capsules, the contents of which swell within the stomach to produce a plug which is buoyant on the gastric contents, provide a further example of this type of technology. 2.7.4.2 Matrix tablets The eroding variety of matrix tablets are generally slowly disintegrating tablets, although similar release systems can be achieved by using semi-solid lipophilic materials in hard gelatin capsules. This approach generally leads to poorer control of in vivo performance. Drugs dispersed in inert matrices can also be employed for sustained release tablets, and better in vivo reproducibility generally results, as drug release rates are not dependent on enzyme levels or gastric intestinal transit rates. Zero-order release kinetics are not obtained from these tablets; cumulative drug release is often proportional to the square root of time. 2.7.4.3 Capsules containing pellets with disintegrating coatings By employing differentially coated pellets, i.e. pellets with varying thickness of a slowly dissolving or eroding polymer, it is possible to provide an extended dissolution profile of a drug. Such a coating may be pH sensitive or insensitive, the former to provide positional release (e.g. classic enteric coatings), the latter to achieve drug release rates independent of transit profiles within the gastrointestinal tract. 2.7.4.4 Pellets or tablets coated with diffusion controlling membranes Pellets or a compressed tablet may be coated with a rate-controlling membrane (nondisintegrating) across which the drug may diffuse. Pellets are normally presented in hard gelatin capsules. By encapsulating drugs in excess of their solubility, a constant concentration gradient will be maintained as long as the saturation state exists and zero-order kinetics will prevail. The OROS® elementary osmotic pump comprises a central core of a salt to provide an osmotic gradient together with drug particles. Surrounding this core is a semipermeable membrane. Fluid is drawn into the core at a rate controlled by both the membrane characteristics and the osmotic gradient. Saturated drug solution is then

The design of drug delivery systems

55

forced from the device through a small orifice into the surrounding membrane at a constant rate. 2.7.5 Positional controlled release Considerable benefits may ensue from drug delivery to a specific region of the gastrointestinal tract. An example of buccal absorption to eliminate first-pass metabolism has previously been described (2.7.4.1). Gastric retention of the drug delivery system may be required to achieve localized drug concentrations or to delay passage of the dosage form past the absorbing membranes of the small intestine, which may lead to improved bioavailability. Prolonged gastric retention may be achieved by the use of gel rafts which ‘float’ on the gastric contents (as with the hydrodynamically balanced capsule™) or by employing mucoadhesive delivery systems, although the latter are in early stages of development. It is often necessary to prevent drug release in the stomach to avoid gastric degradation, reduce gastric side-effects such as inflammatory responses, or provide localized drug concentrations in the small intestine or colon. Enteric coating of both tablets and capsules provides the most widely used approach to avoid drug release in the stomach and is achieved by the use of film-forming ionizing polymers of suitable pKa and degree of substitution. Alternative, more sophisticated, technologies exist to achieve colon specific drug delivery. For example, the Pulsincap® utilizes a novel hydrogel plug fitted into the neck of a water insoluble capsule. A water soluble cap fits over the capsule body and dissolves within the gastric juice to expose the underlying plug. The hydrogel swells at a controlled rate and ejects when it can no longer be contained, thus giving rise to pulsed delivery of the capsule contents within the colon approximately 5 hours after administration. Applications include treatment of local disorders, e.g. irritable bowel disease, and peptide delivery for absorption at what is the preferred gastro intestinal site. 2.7.6 Gastrointestinal transit of sustained release dosage formulations Transit profiles are greatly influenced by diet. Comparison of gastric emptying of a single unit with a multiple or pelleted dosage form resulted in similar residence times when dosing followed a light breakfast. However, when a heavy breakfast was substituted for the light meal, an increase in residence time was noted for the pellets but the delay was considerably shorter than for the single dose unit (residence time approximately 10 hours). Less intersubject variability in plasma concentrations would therefore be expected from pelleted formulations compared with single unit dosage forms, especially where there is no attempt to regulate diet and where the dosage forms show pH-dependent drug release profiles.

Introduction to the principles of drug design and action

56

2.8 SITE-SPECIFIC DRUG DELIVERY Having briefly reviewed the principal approaches to the temporal control of drug release from the dosage form, it is pertinent now to examine the subject of site-specific drug delivery or drug targeting. Originally driven out of a desire to therapeutically optimise the delivery of cytotoxic agents, the interest in site specific delivery has expanded rapidly as molecular biology advances have led to a better understanding of a number of disease states and the identification of potential target sites (receptors) for drugs. We are currently witnessing a revolution in availability of biotechnic drug products (e.g. genes, peptidergic mediators, antisense oligonucleotides), the successful clinical utilization of which requires significant application of drug delivery science. In essence, site specific drug delivery attempts to optimise drug activity by insuring exclusive availability to the specific receptors, i.e. differential accessibility and in so doing provide protection to both drug and body. 2.8.1 Carrier systems Target selectivity by differential sensitivity (drug distributed throughout the body but acting exclusively on target) is virtually impossible to achieve. Therefore, in order to selectively deliver drugs it is necessary to utilize a drug carrier system (Table 2.2). Such a site-specific carrier is required, in addition, to being a guiding device, to protect against drug excretion or inactivation, prevent drugs from eliciting adverse immune reactions, provide for site recognition and being retained at that site to achieve drug release over an appropriate timescale, and finally to be degraded/excreted.

Table 2.2 Classification of drug carriers. Macromolecular carriers Microparticulate carriers Cellular carriers Proteinaceous carriers (e.g. erythrocytes, leucocytes, lymphoid (e.g. antibodies, albumin, glycoproteins, lipoproteins, cells, fibroblasts) gelatin, polypeptides) Lectins Vesicular carriers (e.g. liposomes, niosomes) Hormones Polysaccharides Lipid carriers (e.g. dextran) (e.g. emulsions, waxes, lipoproteins, chylomicrons) Deoxyribonucleic acid Microspheres/Nanoparticles (e.g. albumin, starch, dextran, polyalkylcyano-acrylate, polyamide, polyanhydrides, poly(lactic glycolic) acid)

The design of drug delivery systems

57

Soluble synthetic polymers Viruses and viral envelope products Parenteral administration is the simplest and most efficient approach for any of the carrier systems, although considerable efforts have been expended to develop systems that will be capable of reaching the blood pool via mucosal epithelia (nasal, pulmonary, buccal intestinal, vaginal, rectal). Transport mechanisms include fluidphase pinocytosis, paracellular and transcellular diffusion, receptor-mediated transand endo-cytosis, nutrient carrier processes and lymphatic translocation. 2.8.2 Fate of site-specific delivery systems The target site may be reached by either active or passive events. Microparticulate carriers (Table 2.2) injected into the circulation are usually recognised as foreign and undergo opsonisation. This process consists of adsorption of serum components or opsonins (mainly proteins) which trigger the microparticle uptake by the cells of the mononuclear phagocytic system (MPS). Depending on surface properties of the carrier particles, some 50–90% of the injected dose will distribute to the liver where the main target cells are the Kupffer cells with a smaller portion taken up by hepatocytes. The spleen, lungs and bone marrow will also accumulate the opsonised particles but to a lesser degree. This passive targeting process can be controlled by selecting particles with different surface properties leading to a different spectrum and mass of adsorbed opsonins. Attempts to produce microparticles with long circulation times have focused on preventing/reducing opsonisation by pre-adsorption or surface grafting of surfactants or polymers on the microparticle surface. By utilising surfactants or polymers which contain hydrophilic moieties such as polyoxyethylene chains, a steric barrier is created preventing serum protein adsorption, which is predominantly a hydrophobic interaction. ‘Stealth’ liposomes are an example of sterically-stabilised microparticulate drug carriers. Active targeting, in contrast, involves some cell-specific recognition event achieved by incorporating a target recognition moiety at the surface of the carrier. Immunoliposomes, for example, may be prepared by covalent coupling of Fab/ fragments to the preformed liposome surface. The Fab/ is prepared from IgG raised against a tumour-specific cell surface antigen. Although microparticulate carriers have the advantage of providing a high, wellprotected drug payload, there are two principal disadvantages to their usage. The first already discussed is their potential capture by cells of the MPS before reaching their targets. The second is the barrier afforded by the endothelia to extravasation. There are three main types of endothelia: (1) Continuous—where the cells are close together and no gaps exist. The cells overlie a continuous basement membrane. This is the most frequently occurring form and its morphology prevents colloid extravasation. Molecules of 5–10 kDa can pass into the tissue space and subsequently into the lymphatic system. (2) Fenestrated endothelia occur in the exocrine glands, possess a continuous basement membrane, but differ from the continuous type in that the endothelial surface contains fenestrae (50–60 nm holes) which are covered

Introduction to the principles of drug design and action

58

by a diaphragm. Although colloids might penetrate the fenestrae, transport would not occur across the basement membrane. (3) Sinusoidal endothelia are a discontinuous type which is found in the liver, spleen and bone marrow. Gaps are present in the wall and the basement membrane is absent. In the liver, sieve plates exist through which colloids of less than 100 nm can pass into the space of Disse and hence into the parenchymal cells. Sinusoidal endothelia therefore provides the only opportunity for the extravasation of microparticulate drug carriers. The opportunity therefore exists for treating abnormal cells, such as cancerous or virally-infected through macrophage activation, by incorporating an immunomodulatory drug such as muramyl dipeptide in a colloidal carrier. In rheumatoid arthritis, damaged endothelia may provide the opportunity for colloidal carrier accumulation. Intravascular targets, for example, diseased macrophages (fungal, parasitic and viral, autoimmune diseases and enzyme storage diseases) and other bood cells (cancer, gene and antiviral therapy) are readily accessible to sub 1.0 µm colloids. Also, the opportunity exists of providing long-circulating microparticulates for the slow release of a range of pharmacological agents. 2.9 BIOEQUIVALENCE There are many instances reported in the pharmaceutical literature of chemically equivalent products which have been shown to be bioinequivalent. Any change in a formulation can potentially change the bioavailability. Even with identical formulae, changes in the many process variables can occur in the production of a generic product undertaken by different manufacturers, giving rise to bioinequivalent products. In a granulation process, for example, the nature, quantity and method of addition of the granulating fluid; the drying process; the ageing and storage conditions of the granules prior to tabletting; can all influence the drug release characteristics of the final product. The choice of granulating agent can lead to differences in tablet hardness upon storage, which of course in turn will lead to prolonged disintegration and dissolution times. However, some drugs are known to provide greater bioinequivalence problems than others. The Food and Drug Administration (U.S.A.), in response to this situation, published a list of 115 drug substances for which in vitro or in vivo bioequivalence data are required. When a patient is being successfully treated or stabilized on a branded product, it is undesirable to change to a chemically equivalent product from an alternative manufacturer unless bioequivalence has been proven. Economic pressures advocating change of product should be resisted—at least until bioequivalence data are presented.

The design of drug delivery systems

59

FURTHER READING Chien, Y.W. (1992) Novel Drug Delivery Systems, 2nd edn. New York: Marcel Dekker. Florence, A.T. and Salole, E.G. (eds.) (1990) Routes of Drug Administration. London: Wright. Junginger, H.E. (ed.) (1992) Drug Targeting and Delivery: Concepts in Dosage Form Design. Chichester, England: Ellis Horwood. Kreuter, J. (ed.) (1994) Colloidal Drug Delivery Systems. New York: Marcel Dekker. Lee, V.H.L. (ed.) (1991) Peptide and Protein Drug Delivery. New York: Marcel Dekker. Robinson, J.R. and Lee, V.H.L. (eds.) (1987) Controlled Drug Delivery: Fundamentals and Applications, 2nd edn. New York: Marcel Dekker. Wilson, C.G. and Washington, N. (1989) Physiological Pharmaceutics: Biological Barriers to Drug Absorption. Chichester, England: Ellis Horwood.

3. INTERMOLECULAR FORCES AND MOLECULAR MODELLING ROBIN H.DAVIES and DAVID TIMMS CONTENTS 3.1 INTRODUCTION

52

3.2 MOLECULAR INTERACTIONS

53

3.2.1 Electrostatic interactions

53

3.2.2 The hydrogen bond

54

3.2.3 van der Waals interactions

56

3.2.4 Exchange repulsion

56

3.3 FREE ENERGIES OF INTERACTION—GAS PHASE AND SOLUTION

56

3.3.1 Entropy and free energy contributions in the gas phase

56

3.3.2 Entropy and free energy contributions in solution

57

3.3.3 Electrostatic interactions in solution

58

3.3.4 van der Waals interactions in solution and the hydrophobic effect

58

3.3.5 Some experimental observations— Thermodynamics of ligand binding to receptor proteins

59

3.4 INTRAMOLECULAR FORCES AND CONFORMATION

61

3.4.1 Conformation in the gas phase. Intrinsic conformation

61

3.4.2 General rules for the interaction between orbitals of different energy

61

Intermolecular forces and molecular modelling

61

3.4.3 Examples of orbital interaction e.g. C-C σ bonds, C-Cπ bonds

62

3.4.4 Electron donor-acceptor interaction

63

3.4.5 Hyperconjugation

63

3.4.6 General remarks

66

3.5 MOLECULAR MODELLING

68

3.5.1 Introduction

68

3.5.2 Thermodynamics of ligand binding and conformer identification

69

3.5.3 Ligand design—macromolecular structure known

74

3.5.3.1 Multiple fragment probes—locate and link methods 3.5.3.2 Linking the probes 3.5.3.3 Single fragment probes and ligand evolution 3.5.3.4 Filling the target site 3.5.4 Accommodation of the protein to ligand binding. Estimating interaction free energies 3.6 PROTEINS 3.7 ACCURATE CALCULATION OF INTERMOLECULAR INTERACTIONS REFERENCES

74 75 76 76 77 78 79 80

PLATES 3.1 Serine proteases. Proton movement and enzymatic cleavage of the peptide bond

82

3.2 Inhibition of a serine protease and protein-protein recognition

84

Introduction to the principles of drug design and action

62

3.3 Aspartate proteases. Proton movement and enzymatic cleavage of the peptide bond

85

3.4 Protein-protein recognition. The influence of a hormone on protein dimerisation

87

3.5 A potential ligand-activated proton pathway for signalling in a guanine-nucleotide-coupled receptor ternary complex acting as a guanosine triphosphate synthase

88

3.6 Adenosine triphosphate synthase 3.7 The influence of strong charge on conformation. The structure of calmodulin with and without the interaction of 4 calcium ions 3.8 Protein—single strand DNA recognition 3.9 Protein—double strand DNA recognition. The selectivity of protein binding in the major and minor grooves of the DNA

90

92 94

95

3.1 INTRODUCTION Structural specificity in biological interactions involving macromolecules is generally dominated by non-covalent intermolecular forces. While localised biological interactions may be dependent on proton or electron transfer between groups or, in the case of an enzyme action, on the transfer of a functional group between molecules, the orientation and complementarity of the interacting groups usually dictate a much larger structure. The resulting interacting surface is dominated by the non-covalent intermolecular binding forces of the two molecules. Often a biological reaction may require a signal to be transmitted over a distance to a site of action, messages, for example, from periplasmic to cytoplasmic molecules having to traverse the cell membrane through receptor proteins. Apart from the minimal structures essential to allow a given mechanism to occur, the evolutionary selectivity of biological interactions has produced additional structure to ensure relatively unique complementarity between the non-covalently interacting molecules. However, even the localised interaction may be dominated by non-covalent forces. Attractive non-covalent hydrogen bond forces are able to dominate the activation of proton transfer from a tyrosine residue by excitation though specific hydrogen bond proton donor interactions on the residue’s phenolic oxygen atom. If energy is not to be wasted in such excitation, accurate directionality in hydrogen bonding is required. In the case of an enzyme interaction, non-covalent forces can dominate the formation of the Michaelis complex prior to activation of the reaction.

Intermolecular forces and molecular modelling

63

Again for biological efficiency, the favourable binding must be such as to eliminate further energetic input, the binding being such as to prepare for the reaction with the substrate close to the transition state. Further structure may have evolved on interacting macromolecules as the efficient deployment of excess energy involved in a specific mechanism may be utilised in an associated reaction to absorb the surplus energy. Signals may have evolved between the two mechanisms for efficient deployment of the energy resource. The wider aspect of signal control and stability may again have utilised further additional structure to allow interactions for feed-back mechanisms to switch off the action or reaction concerned. In modelling such reactions, it is useful to bear in mind such problems and, in view of their complexity, to remember that we are usually experimentally led. The theoretical determination of protein structure from first principles based on the intramolecular interactions of the individual amino acids would have high significance in the design of inhibitory or stimulatory ligands in many areas of drug therapy but the search for a full three dimensional structure by seeking a global energy minimum in any sizeable protein is of very high dimensionality. This so-called multiple minima problem has meant, in practice, a recourse to a variety of empiric assumptions to reduce the scale of the problem although theoretical effort in this field is intense. As the size of the macromolecule increases and attempts to compute free energies of small molecules interacting with macromolecules gain ground, it becomes essential to base the intermolecular forces on fast empiric methods using interatomic potentials. These, themselves may be conveniently developed from more rigorous theoretical methods and we review later the current scale of attack on these problems in the light of existing computational hardware. We define first, the characteristic interactions in molecular recognition. 3.2 MOLECULAR INTERACTIONS 3.2.1 Electrostatic interactions The electrostatic potential energy between two isolated charge distributions A and R is given by

(3.1) where qi, qj are the charge on atoms i of A and j of R respectively and r is the interatomic distance. Scaling electronic units of charge and distance to chemical energies in kcal/mol, the interaction energy associated with two units of charge sited 1 Å apart is 332.0. Thus, for an ion pair interacting over 3.0 Å in the gas phase, the interaction energy is over 100 kcal/mol. and electrostatic interactions are large. On the other hand, in aqueous solution, the energies of ion hydration must be overcome for the ions to interact. The

Introduction to the principles of drug design and action

64

addition of a proton to an ammonia molecule in the gas phase is calculated to be over 200 kcal/mol while in aqueous solution the net interaction is 4.0 kcal/mol. Thus the binding of a strong polar group involves a relatively small competitive difference between hydration and the specific interaction making the likelihood of entropic effects relevant to the overall equilibrium. An equivalent description of the potential energy of two charge distributions as given in 3.1 can be made by expanding the 1/rij term to give an expression of the energy in terms of all the electric multipole moments. For polar molecules with no net charges, the first term is a dipole-dipole interaction where the individual polar charges (the interactions of which, as stated, can be summed separately) are conveniently replaced by the dipoles. Non-vanishing higher terms such as the dipole-quadrupole and quadrupole-quadrupole etc. may be necessary for the equivalent accurate description. To exemplify the size of dipole-dipole interactions in simple systems, for charges of 0.15 electrostatic units taken over a length of 2.0 Å, the effective dipole moment µ is 1.4 Debyes. For two such dipoles aligned 4.0 Å apart, the energy of interaction is ~1 kcal/mol, the energy contribution being proportional to µ/r3 where r is the distance between the dipole centres. An isolated charge interacting with this dipole, on the other hand, would produce an equivalent interaction of over 6 kcal/mol. Here, the interaction is proportional to µ/r2. The electrostatic interactions are, for modelling purposes, scaled by a dielectric factor. The dielectric constant is 80 in an aqueous medium, but this factor reduces to 3.0 for interactions in a hydrocarbon liquid. An enzyme reaction may have a heterogeneous environment but for membrane and many other proteins, there is evidence that the environment is close to a hydrocarbon-like medium. A favoured empiric factor used in modelling interactions within proteins is a distance dependent 1/4r. For interactions in close proximity, the relevance of a general dielectric effect may be questioned. The presence of strong charge creates two further effects. Strong charge may induce a further moment in the second molecule producing an induced moment—a polarising effect, while an actual transfer of charge within the interacting molecules may take place. Polarisation may be regarded as the the influence of one set of charges upon the second molecule. Charge transfer is a quantum mechanical effect and as it now emerges over the last decade is a dominant feature of the hydrogen bond. When occupied orbitals overlap, there is repulsion between the orbitals. If, on the other hand an occupied orbital overlaps with an unoccupied antibonding orbital, there is some charge transfer to the unoccupied orbital and increased stabilisation. The closer the two orbitals are together in energy, the greater the stabilisation. The largest effect is found from the interaction of the highest energy occupied molecular orbital (HOMO), usually a lone electron pair (n) with that of the lowest energy unoccupied orbital (LUMO), often a σ* orbital, spectroscopically observed, for example, as an n−σ* transition. The contribution of these terms to the overall energy may be determined theoretically.

Intermolecular forces and molecular modelling

65

3.2.2 The hydrogen bond It was first noted by Coulson that although hydrogen bonding may appear to be dominantly electrostatic, the distance of approach of the proton donor and acceptor produces an almost equal repulsive effect. Decomposition of the terms contributing to the hydrogen bond, at first suggested that the main component contribution to the hydrogen bond was electrostatic contrasting with early spectroscopically-based theories of charge exchange. It has been subsequently shown that the dominant feature of the hydrogen bond does indeed involve charge exchange and that the form of the earlier decomposition underestimated this charge transfer effect. The size of these charge transfer terms can be used to distinguish hydrogen bond and non-hydrogen bond interactions and are exemplified in Table 3.1. Even where the net change in charge transfer is less than 0.01e, the net attractive energy is still around 6 kcal/mol. The more favourable are the charge transfer interactions, the more repulsion can be overcome and the greater penetration of the van der Waals radius of the atom. The degree of charge transfer, in general, follows simple electronegativity rules. Thus, for example, for complexes involving NH3 as the electron donor (Lewis base), one can rank Lewis acid strength

Electron donor (Lewis base) strength decreases in the order

There is an exception to the ranking of F2 and O2 which can be understood in terms of the hybrid p character of the donor lone pairs where the mixing of π-type lone pairs with the σ-type lone pair increases the charge transfer. On the other hand in the nonhydrogen

Table 3.1 Natural bond orbital charge transfer analysis of some H-bonded and non H-bonded complexes (calculations at HF/6–31G* level) involving HF, H2O, and NH3. Energies in kcal mol−1. Complex (A– ΔE ΔECT ΔEA→B ΔEB→A qAa db B) H3N…HF −12.19 −21.96 −21.20 −0.82 +0.0339 +0.85 (+0.0340) H3N…HOH −6.47 −11.42 −11.00 −0.44 +0.0176 +0.61 (+0.0217) H3N…HNH2 −2.94 −4.96 −4.75 −0.21 +0.0078 +0.27 (+0.0152) −9.21 −15.17 −14.58 −0.66 +0.0206 +0.78 H2O…HF (+0.0305)

Introduction to the principles of drug design and action

H2O…HOH

−5.64

−9.17

−8.88 −0.31

H2O…HNH2

−2.86

−4.10

−3.99 −0.11

HF…HF

−5.85 −10.47 −10.18 −0.34

HF…HOH

−3.99

−5.15

−5.03 −0.13

HF…HNH2

−2.65

−3.45

−3.40 −0.06

CO2…FH (T)c

−1.58

−1.42

−0.13 −1.30

CO2…OH2 (T)

−3.14

−2.67

−0.33 −2.34

N2…FH (T)

−0.35

−0.24

−0.07 −0.17

N2…OH2 (T)

−0.51

−0.21

−0.09 −0.12

N2…NH3 (T)

−0.52

−0.23

−0.08 −0.16

O2…FH (T)

−0.29

−0.30

−0.16 −0.14

O2…OH2 (T)

−0.40

−0.26

−0.15 −0.11

O2…NH3 (T)

−0.36

−0.25

−0.11 −0.14

F2…FH (T)

−0.12

−0.34

−0.21 −0.13

F2…NH3 (T)

−0.02

−0.17

−0.06 −0.11

66

+0.0130 (+0.0253) +0.0065 (+0.0184) +0.0146 (+0.0318) +0.0078 (+0.0238) +0.0056 (+0.0184) −0.0019 (−0.0069) −0.0037 (−0.0053) −0.0001 (−0.0007) +0.0000 (−0.0007) −0.0001 (−0.0009) +0.0000 (−0.0003) +0.0001 (−0.0002) −0.0000 (−0.0005) +0.0001 (+0.0001) −0.0000 (−0.0002)

+0.57 +0.21 +0.69 +0.40 +0.18 +0.17 +0.31 −0.41 −0.54 −0.68 −0.37 −0.51 −0.67 −0.28 −0.72

a

Charge on monomer A by natural bond orbital analysis. Values in parentheses are the corresponding Mulliken charges, shown for comparison purposes only. b Distance of penetration of van der Waals radius in Å. c T shaped complex. (From Reed, Weinhold, Curtiss et al. (1986), by permission) bond complexes, the electrostatic interaction is more dominant, the overall energy being greater than the charge transfer. For the T-shaped complexes the dominant charge transfer interaction is of the n-π* type.

Intermolecular forces and molecular modelling

67

3.2.3 van der Waals interactions Electrons at any moment in time are in arbitrary positions round an atom and may give rise to instantaneous local induced moments, although the net average will be zero. The main effect will be an induced dipole which, in turn, will polarize a second atom or molecule to induce a further moment. The dominant induced dipole-induced dipole interaction will produce a net attraction even though the atoms interacting have no net charge. Such interactions are termed non-polar or van der Waals interactions and their energies vary inversely as the 6th power of the interatomic distance. The induced interaction is proportional to the polarizabilities of the atoms and hence the van der Waals interaction increases with the extent of the outer electron shell of the atom. Although individual van der Waals interactions are quite small, since they are common to all atoms, the net interaction effect may become dominant in the overall drug-receptor interaction. 3.2.4 Exchange repulsion The overlap between closed electron shells on too close an approach between atoms is strongly repulsive. This ‘exchange repulsion’ is exponential in character i.e. of the form Ae−br where A and b are constants. For speed of computation of summed energies over all interacting atoms, these terms are usually replaced by an inverse 12th power of the interatomic distance or by an effective cut-off at the van der Waals contact distance. 3.3 FREE ENERGIES OF INTERACTION—GAS PHASE AND SOLUTION 3.3.1 Entropy and free energy contributions in the gas phase Typical entropy and free energy contributions from translations, rotations and vibrations for simple molecules in the gas phase at 298°K are given in Table 3.2. Vibrational entropies due to the ‘hard’ modes of vibration (from bond stretching and large bond angle opening) are relatively unimportant for normal small rigid molecules lacking low frequency vibrational modes but on binding and coupling there is spreading of the perturbed frequency vibrational modes to both higher and lower frequency modes. The latter will dominate the entropy contributions and may become of significant importance. For the isolated molecule, however, the internal rotations of the molecule and the ‘soft’ modes of vibration due to dihedral angle variation are generally regarded as more entropy rich than all but the lowest frequency ‘hard’ vibrations, and for this reason, the latter contributions are usually neglected. These entropic and enthalpic vibrational contributions may be readily calculated using a parabolic approximation for a vibration where, at the lowest frequencies most of the contributions to the thermodynamic functions occur. The fact remains that most of the significant effects in vibration may arise from the coupling contributions between the molecules whether in the gas phase or in solution. Thus, for example in a distorted ligand-receptor three point interaction where most of the energetic contributions arise,

Introduction to the principles of drug design and action

68

on average, from predominantly two out of the three points of primary interaction and some twelve degrees of freedom may contribute to balancing the unfavourable distortion against slightly more favourable attraction, a dominant entropic TΔS contribution of over 5 kcal mol−1 can contribute to the binding. Such contributions only arise from the direct interaction of the two molecules and should be treated explicitly. Experimental data on receptor ligand binding to a guanine nucleotidecoupled receptor are given in Section 3.3.5. 3.3.2 Entropy and free energy contributions in solution In the case of ligand-receptor binding the net effect is that the free energy of the binding is greater than that of the sum of solvated ligand and solvated receptor free energies. Using a simple liquid/gas thermodynamic cycle to relate gas and liquid phase interactions

(3.2) it is seen that liquid/gas partitioning of the appropriate species of drug (A), receptor (R) and complex (AR) may be incorporated to understand the predicted solvational behaviour. The effects of solvation on the electrostatic and van der Waals interactions, as discussed earlier give rise to competitive effects of hydration both on electrostatic and on hydrophobic

Table 3.2 Typical entropy and free energy contributions from translations, rotations and vibrations at 298°K. S0 H0−H00 G0−H00 (cal (kcal (kcal deg−1 mol−1) mol−1) mol−1) Motion Three degrees of translational freedom for 29–36 1.48 −7.2 to molecular weights 20–200, standard state 1 −9.1 Ma Three degrees of rotational freedoma Moments of inertiab Water 5.8×10–120 10.5 0.89 −2.24 21.5 0.89 −5.53 n-Propane 5.0×10−116 endo-Dicyclopentadiene 3.8×10−113 27.2 0.89 −7.21 c Internal Rotation 3–5 0.3 −0.6 to

Intermolecular forces and molecular modelling

69

−1.2 Vibration v, cm−1 1000 800 400 200 100

0.1 0.2 1.0 2.2 3.4

0.03 0.05 0.20 0.35 0.46

0.0 −0.01 −0.10 −0.31 −0.56

a

Calculated. Product of three principal moments of inertia g3 cm6. c Typical value; this quantity is a function of the barrier to rotation and the partition function. (From Page and Jencks (1971), by permission). b

interactions. We consider first the effect of solvation on translational and rotational entropies. There will be some loss in translational and rotational entropy on solution. For a non-polar molecule, the magnitude of this loss is ~10 e.u. or a TΔS contribution of ~ 3 kcal/mol at blood temperature when the same reference concentration is taken in the gas and liquid phases. The translational and rotational entropy loss on binding to a receptor site is thus expected to be not very different between binding in solution and in the gas phase. 3.3.3 Electrostatic interactions in solution The dramatic efffect of hydration on electrostatic interactions has been mentioned in Section 3.2.1. Thus net ion-ion interactions are greatly reduced on hydration or may become unfavourable and only the consequent liberation of solvent molecules may create a significant entropy effect and favourable interaction. Except in close comparison of related molecules, therefore, gas phase comparisons are of limited utility. Weaker interactions such as ion-dipole interactions may similarly be largely suppressed by competitive solvent dipoles and contribute little to the overall free energy of interaction in a polar phase. While electrostatic effects will be largely suppressed by hydration, the final binding site of the ligand molecule will usually be within a macromolecule surrounded by mobile regions of polar and non-polar phases. Ligand concentration is usually referenced to aqueous solution. To estimate the energetics of hydration that must be lost for ligands to interact with a membrane or related protein, it can, therefore, be informative to reference the concentration to a non-aqueous hydrocarbon environment, when van der Waals interactions are automatically taken into account. The best reference model is obviously some hydrocarbon phase, but the partitioning model which has received the most attention has been the solvent octanol with its attendant problem of interaction with strong hydrogen bond acceptor solutes. This latter model is discussed in Chapter 5. An

Introduction to the principles of drug design and action

70

example of the insight gained on applying partitioning data to the thermodynamics of ligand-receptor protein binding is given in the next section. 3.3.4 van der Waals interactions in solution and the hydrophobic effect There is a driving force for non-polar molecules to interact in aqueous solution which is termed the hydrophobic effect. This force at the macroscopic level causes aggregation of lipids in solution and the folding of proteins in self assembly. The source of this important effect has been of dispute. Earlier theories had interpreted this effect as being entropically driven due to the ordering of water molecules around nonpolar solutes with their resultant liberation on non-polar interatomic contact. More recent evidence has shown that this effect is predominantly or equally enthalpic in character. Table 3.3 shows the incremental thermodynamics of partitioning of a methylene group in a homologous series between an aqueous and a hydrocarbon phase. There are relatively weak favourable enthalpic and large unfavourable entropic components in aqueous solution while in the hydrocarbon environment there is marked favourable enthalpy and weaker unfavourable incremental entropy. The thermodynamic contributions for transfer from the aqueous to the hydrocarbon phase then show the entropic and enthalpic components to produce similar difference contributions to the free energy transfer.

Table 3.3 Incremental thermodynamics of partitioning of the −CH2 group. 310°K kcal mol−1 0 0 Partitioning phases ΔG ΔH −TΔS0 1. Cyclohexane/gas −0.76 −1.12 +0.36 2. H2O/gas +0.18 −0.67 +0.85 Cyclohexane/H2O −0.94 −0.45 −0.49 (From Abraham (1982), by permission). Further insight into the cause of the hydrophobia effect comes from cavity models of solution. The unfavourable entropic effect on aqueous solvation appears to arise from the high number of states available to water and the resultant loss in entropy on forming a cavity to adapt the solvent. The free energy of solvation may be written as a sum of the free energies of formation of the cavity formation and of the solute-solvent interaction. Although the free energies of formation of the cavity are relatively similar in the aqueous and non-aqueous phases, the enthalpic and entropic components are quite different. In the non-aqueous phase, most of the work of cavity formation goes to the enthalpic maintenance of the excluded volume and only a small contribution to the entropy or configurational exclusion of volume; for water, the reverse is the case. As molecules become progressively larger and structures more ordered, it is not possible to be categoric in the relation of forces to their resultant effects in solution, particularly when structural reorganisation becomes critical. Thus as more states become available, there is usually a weakening of enthalpy changes but a compensation in entropy effects. These compensatory effects can be shown to be

Intermolecular forces and molecular modelling

71

large. The concept of molar concentration applied to thermodynamic changes in ordered structures such as liposomes is also less certain. 3.3.5 Some experimental observations—Thermodynamics of ligand binding to receptor proteins The thermodynamics of binding of small ligand molecules within known protein sites should be computable to a good degree of accuracy. The difficulty lies not with the flexibility of the small ligand but with the uncertainty in accommodating the potential flexibility of the macromolecular structure when based on a crystal structure. The position may be exemplified by data on ligand binding to guanine nucleotide-coupled receptor proteins (GCPRs, refer to Plate 3.5) where the structures are not yet experimentally determined. The GCPRs are a dominant class of hepta-helical membrane-spanning proteins linking cytoplasmic events through a heterotrimeric Gαβγ-protein on the cytoplasmic side of the cell to a signalling hormone binding to the receptor. Typical data for binding of related phenoxypropanolamine ligands to a turkey erythrocyte β-adrenergic receptor are given in Table 3.4, a receptor closely related to the mammalian β1-adrenergic receptor. The prediction for the binding of the antagonist, propranolol from the weak partial agonist, practolol may be made with good accuracy. Practolol possesses a p-NHCOCH3 group, and data at the free energy level on the mammalian receptor are concordant with an -NH hydrogen bond proton donor interaction with the

Introduction to the principles of drug design and action

72

Table 3.4 Flexibility of the protein and ligand hydration effects in the thermodynamics of binding of phenoxypropanolamine (a) and phenethanolamine (b) ligands to the turkey erythrocyte βadrenoceptor. (a) Prediction of the binding of propanolol from practolol(†) and (b) adrenaline from isoprenaline(‡).

Intermolecular forces and molecular modelling

73

receptor, not of particular strength and which can be modelled empirically using data from a long chain ester solvent. The main enthalpic difference between the binding of the two compounds is due to loss of hydration on the amidic C=O moiety of practolol. The structure-activity relations have indicated high flexibility in the hydrophobic residues surrounding given regions of the ligand and those residues surrounding the bound 2-substituents in phenoxy ring compounds show a receptor environment akin to a hydrophobic liquid accommodating even large substituents. Agonists on the other hand such as isoprenaline show a strong enthalpic binding with a marked negative entropy. Despite ~12 kcal/mol enthalpic difference in the binding, the free energies of binding of the agonist and antagonists are quite similar. Without some indication of the regions of flexibility of the protein residues and an understanding of the energetics controlling activation of the signal, small empiric perturbations about the structure of the known ligand might still offer the best way of achieving partial agonism even if a detailed crystal adrenoceptor structure were available. Given the crystal structure on the other hand, a theoretical mobilisation of the structure using molecular dynamics or molecular mechanics with developed potentials for the interactions of interest should provide a good thermodynamic prediction. 3.4 INTRAMOLECULAR FORCES AND CONFORMATION 3.4.1 Conformation in the gas phase. Intrinsic conformation Intrinsic conformational preference in small molecules is a guide to interpretation in larger systems. Conformational preference in the gas phase is very largely dictated by the net outcome of electrostatic and bond orbital interactions. We have already commented on the repulsive or destabilising (‘4 electron’) interaction, the ‘exchange repulsion’, of occupied bond orbitals in the overlap of closed shells of electrons and the stabilising (‘2 electron’) interaction of an occupied bond orbital with an unoccupied antibonding orbital giving rise to some charge transfer. All chemists are familiar with the concept of delocalised (molecular orbitals arising from overlap of the atomic π orbitals allowing reactivity at a site remote from the site of substitution. A set of molecular orbitals can be given an equivalent representation in terms of local bond or group orbitals of the molecule. In the case of π orbitals, the resultant interaction may extend over several atomic centres. For singly bonded flexible systems, there are more localised bond orbital interactions from vicinal orbitals about the bond which can dictate or contribute to structural preference. A knowledge of bond or group orbital interactions can thus give insight into the resultant preferred conformation of flexible systems. The term hyperconjugation has been used to define the favourable interaction of these orbitals and in view of their importance in conformational studies, a wider simple introduction to bond orbitals and their interaction is given, following Jorgensen and Salem (1973). 3.4.2 General rules for the interaction between orbitals of different energy 1. When two orbitals interact, they yield a lower energy bonding combination and a

Introduction to the principles of drug design and action

74

higher energy antibonding combination

2. The destabilisation of orbital φA (energy EA) is always slightly larger than the stabilisation of orbital φB (energy EB and EB<EA). 3. Only energy levels which are close together interact strongly, the closer the better. 4. Only orbitals which overlap significantly interact. 5. If a given energy level interacts with several others of significantly different energy, the interactions are pairwise additive. 3.4.3 Examples of orbital interaction e.g. C−C σ bonds, C−C π bonds Two carbon p atomic orbitals interacting ‘end on’ (in this interaction there is zero angular momentum about the bond which is defined as a σ interaction) are shown diagrammatically (Figure 3.1) to produce a σ C−C orbital of lower energy and a σ* C−C antibonding orbital. The atomic contributions are out of phase in the antibonding orbital and characterised by a node (where there is zero charge density). The two electrons occupy the lower bonding orbital and there is a net energy stabilisation on interaction.

Intermolecular forces and molecular modelling

75

Figure 3.1 Two carbon p orbitals interacting to produce a bonding σ C−C orbital of lower energy and a higher σ* antibonding orbital. (From Jorgensen and Salem (1973), by permission).

Figure 3.2 Interaction of 2p carbon atomic orbitals to produce π-orbital overlap for a double bond. (From Jorgensen and Salem (1973), by permission). Figure 3.2 shows the interaction of 2p orbitals to produce a π orbital overlap, in the case of forming the second bond in a double bond. Three carbon orbitals lie in the plane at right angles to the paper, and the 2p carbon orbitals are perpendicular to them. On interaction, the higher energy π* antibonding orbital has the atomic contributions out of phase and there is now a nodal plane perpendicular to the C−C bond. The two electrons occupy the the lower energy π bonding orbital and there is net energy

Introduction to the principles of drug design and action

76

stabilisation. The carbon double bond is thus seen to consist of a σ C−C bond and a π C−C bond. 3.4.4 Electron donor-acceptor interaction In the case of azoborane the relevant atomic orbitals are the 2p nitrogen orbital containing the electron lone pair, and the vacant 2p boron orbital. On interaction there is net stabilisation in energy, the two electrons of the nitrogen atom occupying the N−B bonding orbital and a planar structure is formed (Figure 3.3). When electron lone pairs are present in both orbitals as in hydrazine (3.1), the additional electrons would have to enter the π* antibonding orbital and from rule 2 (Section 3.4.2) the net energy would be destabilising. The hydrazine structure is thus staggered, with the electron pairs lying in a gauche position. For the same reason the azadipeptide (3.2) would be expected to show no tendency to delocalize across the N−N bond and calculation shows the amidic groups to lie preferentially at 90° to one another. In the N,N/ dialkyl hydrazino group in the ring system (3.3) the X-ray structure shows the amidic groups to lie in a similar orientation (Figure 3.4) although other forces in this cyclic system may be acting. 3.4.5 Hyperconjugation For single bonds involving a heteroatom, the possibility exists that a vicinal atom may have a vacant antibonding orbital to produce 2-electron stabilisation. The strength of this interaction will be dependent on the energy difference between these orbitals and the

Figure 3.3 Overlap of 2p nitrogen orbital containing the electron lone pair and the vacant 2p boron orbital in azaborane (From Radom (1982), by permission).

Intermolecular forces and molecular modelling

77

extent of their overlap. In principle, any bonding-antibonding interaction will produce some effect but the highest occupied molecular orbital is usually that occupied by a heteroatom lone pair and the high energy of the localised orbital will have a dominant effect on structural preference. In the case of the vacant antibonding orbital, the more electronegative the neighbouring atom or its substituent, the lower will be its energy. The typical shapes of bonding and antibonding hybrid C−X orbitals are shown in Figure 3.5. Maximum overlap of the bonding and antibonding orbitals tends to occur when the bonds of the neighbouring groups are antiperiplanar (trans), or in the case of the lone pair when it is similarly antiperiplanar to the C−X bond. The strongest hyperconjugative interaction will thus tend to occur when a heteroatom lone pair is antiperiplanar to an electronegative substituent. The anomeric effect in sugars or in substituted pyranose or

Introduction to the principles of drug design and action

78

Figure 3.4 Structure of a fraction of the phthalazino (2,3-b)phthalazine-5,12dione molecule. (From Cariati, Cauletti, Ganadu, Piancastelli and Sgamelloti (1980), by permission).

Figure 3.5 Bonding and anti-bonding C−H hybrid sp3 orbitals. The solid (dashed) lines represent orbital amplitude contours of positive (negative) phase. The position of the C−C bond in the fragments is indicated. The corresponding overlap of the orbitals (with the appropriate phase) may be judged by superposition of the

Intermolecular forces and molecular modelling

79

two C−C bonds. Each contour corresponds to half the amplitude of the preceding one. (From Brunck and Weinhold (1979), by permission). dioxan rings where an electronegative substituent lies preferentially axial is an example of an effect where bond orbital interaction dominates the conformer preference. The effect even so is not large being of the order 1–1.5 kcal mol−1 at blood temperature giving an axial to equatorial preference of 5–15:1. Possible acetal conformations (3.4 a–f) are shown where R, R/ are alkyl substituents (Deslongchamps, (1983)). The antibonding orbitals of interest will lie on the bond with the electronegative oxygen heteroatom and be preferentially antiperiplanar to an oxygen atom electron lone pair. Conformers d, e and f have two anomeric effects, a and b have only one, while conformer c has no suitable bond orbital interaction. However, conformers e and f have steric repulsion from alkyl R/ substituents and the order of stability is found to be d, a, b, c with estimated energies of 0, +1.0, +1.9, +2.9 kcal mol−1 respectively.

An example involving the nitrogen lone pair is shown in the conformer preference of (3.5a) and (3.5b) where conformer (3.5b) has a 500-fold population preference. 3.4.6 General remarks Conformer preference in flexible σ bonded systems is more usually a balance between electrostatic, exchange repulsion and bond orbital effects. The favourable ‘two electron’ interaction has been emphasized here to give some insight into the structure of conformer preference. More detailed reading may be cited (Csizmadia (1982), Deslongchamps (1983)).

Introduction to the principles of drug design and action

80

It is possible to estimate the relative components contributing to the conformer preference in saturated systems by the following considerations. Figure 3.6a shows eclipsed and staggered forms of an aliphatic system using Newman projections. On rotating about the bond there is an energy well or barrier every 60° due to the exchange repulsion and the rotation is three-fold symmetric. In the case of hyperconjugation, the rotation of the antibonding orbital through 90° minimises the interaction and there is a 2-fold interaction on rotation about the bond through 360° (Figure 3.6b). For an electrostatic interaction, on the other hand, there is a 1-fold interaction on bond rotation though 360° (Figure 3.6c). The components and their resultant interaction may thus be separated and are shown experimentally in Figure 3.7.

Figure 3.6 Relative components contributing to the conformer preference in a saturated aliphatic system (a) steric (b) bond orbital (c) electrostatic.

Intermolecular forces and molecular modelling

81

Figure 3.7 Resultant interaction of components contributing to conformer preference illustrated in Figure 3.6 (from Radom (1982) by permission). 3.5 MOLECULAR MODELLING 3.5.1 Introduction The term ‘molecular modelling’ embraces a wide definition and it is convenient to categorise approaches to modelling dependent on the degree of information known. Even where the target structure is totally unknown it should be possible to determine structural information from the target site ligand either by selective synthetic ligand constraints or by analysis of the available pharmacological data. Early development in the pharmaceutical industry relied on such methods utilising small perturbations about the structure of a target hormone and such methods continue to have strong utility. While the structural information obtained from such approaches is not independent of the mode of binding of the particular set of ligands, even here, there are indications of efficiency from the overall gross ligand potency. Synthetically-based identification of the bioactive conformers using constrained molecules aided by temperature studies on receptors using isolated membranes or intact cells ‘in vitro’ yield thermodynamic conformer binding data and quantitative information on the mode of binding which should allow determination of the geometry around localised bonds of the ligand in many instances and, importantly, allow for some decomposition of the energetics of the binding in closely related ligands. While considerable localised information on the target site can be obtained from such methods, selective binding to sites remote from the biological action remain elusive without wide-scale random screening.

Introduction to the principles of drug design and action

82

There is an increasing data base on target macromolecules. The three dimensional structures of 4,000 now exist in the Brookhaven Protein Data bank and amino acid sequences of a further 150,000 are available. Structure-based ligand design is, therefore, a reality in a number of therapeutic areas. Plate 3.1 shows the localised structure of a typical serine protease containing the characteristic Asp102—His57— Ser195 catalytic triad involved in the peptide bond rupture. The catalytic site of trypsin in the presence of the bovine pancreatic trypsin inhibitor shows the presence of an anionic site for preferential binding to basic residues involved in the peptide bond rupture. Plate 3.2 shows the inhibition of this site in the enzyme a-thrombin by the natural ligand inhibitor Hirudin where selective binding to remote ‘exo-sites’ is exemplified. A variety of simple logical approaches can be deployed to examine the possibilities of occupying the binding site efficiently. Their disadvantage in accurate prediction, as mentioned earlier, is the inability of the crystal structure to convey the varying degrees of flexibility within the site. The selective binding to remote sites, however, should be efficient and of great therapeutic advantage. It will be convenient to classify possible predictions in terms of ligand, protein and DNA targets based on the scale and predictability of the problem. As we are concerned, here with ligand design, the main emphasis in this chapter will be to concentrate on the strategies available to rational ligand design both when the macromolecular structure is known and unknown. Examples of the scale of some interactions involving protein-protein, protein-single strand DNA and protein-double stranded DNA are given in Plates 3.4, 3.8 and 3.9. Plate 3.8 shows the binding of a zinc finger domain to a single strand of DNA while protein ocupancy of the major and minor grooves of a piece of double stranded DNA is shown in Plate 3.9. 3.5.2 Thermodynamics of ligand binding and conformer identification When a ligand binds to a receptor or its target enzyme, often the energy of hydration is lost from most regions of the molecule and the interaction becomes essentially nonaqueous in character. It can, therefore, be useful to change the reference phase for binding to that of a model hydrocarbon liquid when simple correlations of potency and change in reference phase indicate the inherent flexibility of the target macromolecule in given regions of the molecule. Some consequences are exemplified in Section 3.3.5. Such correlations at the free energy level automatically introduce a good approximation to the van der Waals forces operating in the binding in closely related molecules. Often a 2–3 kcal/mol variation in observed binding is reduced to little more than 0.15 kcal/mol when introducing this reference change providing a useful base line for exploring other effects within a given mode of binding. As a major target is to maximize efficient binding it is important to identify whether the binding conformation is dominant or whether only a small fraction of the ligand productively binds to the macromolecule. It is useful, therefore, to represent the gross ligand binding constant in a conformer representation and to examine possible relations between the phase environment and the conformer representation (Davies (1987)). In terms of standard partial free energies, µ° the gross binding constant may be written as

Intermolecular forces and molecular modelling

83

(3.3) where the subscripts AR, A and R refer to the complex, drug and receptor respectively, and k is the Boltzmann constant, T the absolute temperature. Using second indices to identify the conformer i of the drug A engaged in binding, with j* its receptor counterpart, then, for the ij* conformer interaction, Equation [3.3] may be written

(3.4) Using conformer populations fi of A, and fj* of R, and the relations

(3.5) and summing over the bound states

(3.6) where Kij*, the conformer binding constant is given by

(3.7) The binding constant is a sum of the conformer binding constants weighted by their appropriate conformer fractions. It is more convenient to define the conformer binding constant referenced to the average states of A and R.

(3.8)

Introduction to the principles of drug design and action

84

It is often helpful to consider comparative drug binding with a change of reference to a hydrocarbon lipid phase L. The standard free energy change of A can then be written

(3.9) and since

(3.10) where fiL is the conformer fraction of i in a nonaqueous medium and Pi is the conformer-or micro- partition coefficient of the species i which, often, is easily estimated (Davies, Sheard and Taylor (1981)). It follows that

(3.11) These relations may be observed from the free energy diagram in Figure 3.8. The appropriate thermodynamic relations may be similarly expressed. The two equations show the relations between conformer populations in aqueous and nonaqueous phases. For a set of close analogues which bind to the receptor in the same way, Kij* fj* is often invariant and the binding constant will vary directly with the relevant conformer fraction of the ligand. For a rotation about a single bond, the conformer population can be readily calculated from the rotamer energetics by use of the Boltzmann distribution. The reason that classical statistics can be applied to rotamer energetics is that, unless the barrier to rotation is very high, conformer interchange is very fast. The number of molecules with energy Ei is given by

(3.12)

Intermolecular forces and molecular modelling

85

The relative population between two states 1 and 2 is given by

(3.13) More strictly, writing ΔG2−1 for the free energy difference between the two rotamers and taking logarithms

(3.14) where N is Avogadro’s number. Using the log10 scale, NkT=1.418 kcal mol−1 at blood temperature (37°C). Thus for two conformers differing in energy by l kcal mol−1, log10 n2 /n1~0.7 (where conformer 2 is the more favourable) giving a conformer population ratio n2/n1 of 5:1 at this temperature. (From Davies (1987), by permission) The utility of Equation [3.10] is shown by a simple example in Figure 3.9 where the bioactive conformer with the basic side chain perpendicular to the aromatic ring (calculated on the intrinsic conformer preference in Table 3.5) of the CNS agent viloxazine (3.6, R=2-OCH3), an inhibitor of biogenic amine release, is plotted as a function of the potency component which has been referenced to a non-polar phase environment. The relation is of unit slope. (In this early example, octanol has been used as the reference non-polar solvent. As the relatively weak hydrogen bond proton acceptor properties of the solute phenoxy oxygen atom are weaker than those of the polar reference solvent octanol, little error is introduced in this set of data by the use of this solvent compared with that of a hydrocarbon). While most data of this type are very much related to details of ligand conformation and of localised energetics in the target site, an advantage of such information is an understanding of the detailed thermodynamics of binding of closely related molecules as exemplified in Section 3.3.5. Plate 3.5 shows the development of a potential guanine nucleotide receptor α-helical model based on geometric constraints of the bound ligand hormone and the resultant constraints on the receptor α-helices. The comparative

Introduction to the principles of drug design and action

86

Figure 3.8 Schematic representation of the free energy relations for the conformer i of the drug (A) interacting with the relevant receptor conformer jx of the receptor protein complex and possible pathways for relating the bound conformer free energy to the reference free energy GA. The standard free energy of the conformer i of the drug is related to the average free energy GA by the conformer fraction or population fi. A change of reference phase from aqueous to hydrocarbon is shown by the subscript L. The partition coefficient P defines the average free energy difference of A between the two phases and individual conformers in the different phase environments may be related similarly by conformer partition coefficients Pi (From Davies (1987), by permission).

Intermolecular forces and molecular modelling

87

Figure 3.9 (a) Potency ‘in vivo’ of viloxazine analogues plotted against a partitioning effect using the octanol/water model on

Introduction to the principles of drug design and action

88

the log10 scale, (b) Residual variation in potency of viloxazine analogues after allowance for a partitioning effect plotted on the log10 scale against the fraction of the conformers having the side chain perpendicular to the aromatic ring. (From Davies (1987), by permission).

Intermolecular forces and molecular modelling

89

Table 3.5 Intrinsic conformer preference of substituted anisoles and related molecules at 37°C. Ab initio estimates and NMR data.

Introduction to the principles of drug design and action

90

energetics between agonist and antagonist and the thermodynamics of binding and response of a partial agonist have led to ideas on proton shuttle mechanisms between Tyr-Arg-Tyr residue triads which show promise in identifying general proton transfer signalling mechanisms (Nederkoorn, Timmerman, Timms et al. (1997)). 3.5.3 Ligand design—macromolecular structure known There are a number of simple ligand modelling strategies that have evolved to take advantage of the structural information on target proteins derived from X-ray crystallography or NMR spectroscopy. Given that the structure of the site is known, the strategies resolve to devising efficient schemes for the logical exploration of the space of the target site and the housing of the ligand’s appropriate interacting groups. Whether to build upon interacting groups to probe obvious target sites and link these probes back to some representative molecule or whether to fill the volume of the site with nominal atoms and then to choose viable sub-sets for efficient interaction, the choice is perhaps dependent on the degree of understanding of the mechanism involved. Binding it should be remembered is a free energy process and those methods which incorporate the statistics of the binding both in macromolecule and in ligand should prove the most powerful. The limiting problem is likely to be computational effort but there is no substitute for knowledge of molecular structure. 3.5.3.1 Multiple fragment probes—locate and link methods In the probe approaches, the so-called locate and link methods, a site specific small probe can be geometrically constructed or better its interaction calculated and the orientation for the best localised orientation of the small probe molecule determined. Variants on optimising the location of the probe can be generalised. One may place a set of small groups randomly on a coarse grid (0.5 Å) and optimise the translational and orientational variables using search methods based on the rate of change in energy as a function of the variables or by stochastic methods such as Monte Carlo. Similar approaches using the protein-fragment interaction forces and employing molecular dynamics for locating probes on a large number of polar fragments (e.g. 1000) randomly distributed within the binding site are used to calculate, via Newton’s laws, the independent motion of each fragment. By slowly cooling the system to absolute zero, optimal binding positions for the probe groups can be determined. Steric features of a site can be exploited by constructing spheres in contact with the protein surface such that the centroids represent positions for locating interacting atoms. In place of calculation, optimal positions for groups to partner hydrogen bonding moieties in the protein can be derived from data surveys of small-molecule Xray structures and via microwave spectroscopy and quantum mechanical calculations. The resulting positions for donor hydrogen or acceptor atoms and their connected atoms form a set of vectors on which candidate probe hydrogen bond groups can be overlayed. A special case in refined X-ray structures is given by bound solvent water molecules which represent experimentally located probe fragments. Such water molecules can indicate opportunities for the location of hydrogen bond groups,

Intermolecular forces and molecular modelling

91

although, dependent on their degree of interaction, not all can be replaced in an energetically favourable manner. Water is potentially tetracoordinate in hydrogen bonding through its two hydrogen bond proton donors and its two electron lone pairs as proton acceptors. Whether to treat a located water molecule as a candidate for replacement or as strongly held by the protein depends on the number of potential interactions made with the macromolecule. The following simple table on the categories of bound water and their implications for substitution may be constructed.

H1

Category of water protein protein protein protein sequestered protein protein protein available structural available protein protein protein structural LP1

H2

LP2

available protein available protein

ligand-like

protein available protein available

ligand-like

protein

ligand-like

protein available available

Implication not available locate a donor locate an acceptor replace by e.g. −CO? replace by e.g. −NH2 replace by e.g. −OH

Where only one interaction with the protein occurs, then replacement should be possible unless the water molecule under consideration forms a link in a chain of interacting hydrogen bonded groups from the protein or other water molecules. In this case preservation of the chain may be an important consideration. For two or less interactions with the protein or surrounding system, the bound water may be described as ligand-like and it should be possible to displace it with a favourable energetic outcome provided that there is no degradation in the quality of the replacement interactions. 3.5.3.2 Linking the probes The construction of potential intramolecular links between two probe groups is a straightforward if tedious problem of determining the possible spans by constructing a series of bonds with standard lengths, angles and torsions and elucidating those links which do not clash with the protein. A number of methods have been developed to address this problem. For up to six bonds, given the bond lengths and angles, the required torsion angles can be solved analytically. Beyond this, connection can be achieved by constrained optimisation of the torsion angles introducing constraints using the method of Lagrange multipliers. If torsion angles are sampled at appropriate minima, combinations of bond geometries (tetrahedral and trigonal) can be assembled into a growing network which terminate when a connection between fragments is made. Generally, several linker chains of varying length and composition will connect the probes and often these can be combined to give cyclic structures, eliminating unwanted conformational freedom and associated entropic effects. Similarly when the

Introduction to the principles of drug design and action

92

linkers show certain patterns of torsion angles, for, example, a series of planar torsions, they may be reinforced by constructing rings incorporating those torsions. Alternatively, the spatial arrangement of functional groups within the binding site allows this geometric structure to interrogate a data base of small-molecule threedimensional structures such as the Cambridge structure data base. In pharmaceutical companies, such data bases contain up to 1 million molecules. Molecules matching the required criteria can be tested for their ability to bind to the protein. The fastest searches assume a single conformation for each small molecule, but multiple conformations can be sampled if pre-stored in the data base. A method requiring less data storage at the expense of description of strain within the molecule and of computer time, attempts to fit the spatial constraints of the search query using distance geometry methods (Blumenthal (1970)). Further data base methods utilise the vector nature of the probe to its potential link with the putatative ligand. The geometric relationship between the vectors can be defined in terms of distances, angles and torsions. A searchable vector data base can be generated from any set of molecules by identifying templates with a number of connector bond vectors and then tabulating the geometric relationships between them. These templates are often rigid ring systems and the connectors, C-H bonds. Starting from commercially available compounds, a data base of more than 30,000 templates can be derived. Searching a vector data base for templates capable of connecting the localised functional groups simply correlates matching the appropriate distances, angles and torsions within given tolerances. An ability to synthesise the appropriate template is usually an overriding choice amongst the matching templates. Finally, it may be noted, that the whole process of matching a ligand to its site can be machine based without recourse to an experimental data base. If a decision is taken on the basis of the synthetic chemistry to be exploited, for example that of substituted benzdiazepines, then the most promising substitution patterns can be identified. Since the chemical reactions are specified, the reagents that are available commercially can be used as input to the computation and the output can be exploited using robotic methods in multiple parallel syntheses to generate libraries of candidate compounds. 3.5.3.3 Single fragment probes and ligand evolution As the name implies, an initial target binding site is selected and an initial fragment probe developed from which the ligand is allowed to grow. This growth can be done by successive addition of atoms using a correlated acceptance or rejection procedure on each addition, the choice being dependent on their fitness to the protein environment. At a geometric level, the quality of the ground rules and the range of atom types considered are critical to the validity of the method. A variant of the method allows atoms to be ‘mutated’ and segments of one molecule to be exchanged for a second. These evolutionary steps of addition, mutation, and crossover form the basis of a ‘genetic’ algorithm. The number of structures evolved is controlled by assessing the fitness of the protein environment. Further criteria are required for realistic segments to be be identified and to ensure that the consequences of mutating atoms on surrounding atoms are transmitted into the next generation.

Intermolecular forces and molecular modelling

93

Again a fragment database with connector bonds can be attached to candidate ‘hooks’ within the seed. Selection procedures based on some ‘protein binding’ score catering for the interaction and the degree of distortion involved with the new link are used in such procedures. The quality of the criteria and the potentially regressive effect of the enlargement on previous substitutions where a net favourable gross interaction may occur, highlight difficulties with these automated procedures. As further steps become involved, the enlargement may lead to combinatorial explosion in the number of candidates. 3.5.3.4 Filling the target site The final approach is to fill the target site with nominal atoms and then choose viable subsets and determine the chemical nature of the atoms constituting the candidate ligand. Again much of these automated procedures are based on simple logical procedures. A regular lattice such as the diamond, tetrahedral or planar hexagonal is positioned in the binding site using interactive graphics or by calculating minimal steric clashes with the protein surface. Complementary ligand-receptor interatomic interactions may be assigned and viable subsets of atoms selected. It is, however, difficult to mix different regular lattices to form realistic molecules. Alternatively one may place a set of small acyclic and cyclic fragments to fill the site with all possible combination frameworks. It is then necessary to select candidate subgraphs and assign atom types via the protein environment. This method allows different geometries to be used together and overcomes the combinatorial explosion by using a small set of fragments involving only carbon atoms. Here, the main difficulties lie in the selection of viable sub-graphs and the assignment of atom types. A much simpler approach is to characterise the shape of the protein binding site as a defined ellipsoid and search a data base of small molecules for identifying suitable ligands which fill the site approximately. The structure may then be substituted to adapt and complement the target site. The last approach in this category is to fill the site with atoms whose individual nature is randomly assigned. The system is equilibrated using molecular dynamics with a force field that allows for ‘soft’ repulsion between the atoms. A ‘mother’ atom is randomly selected and attempts are made to form bonds with neighbouring atoms using probabilistic rules. If accepted, the system is then relaxed using molecular dynamics and a new ‘mother’ atom selected. The process is repeated for a specified number of selections, resulting in the emergence of a candidate ligand from the initial aggregate of atoms. The process is thus stochastic and may take many repeats to arrive at a synthetically useful ligand, The rules for bond formation and the associated acceptance criteria are crucial to this approach. 3.5.4 Accommodation of the protein to ligand binding. Estimating interaction free energies In the previous Section, the structure of the protein was taken to be fixed at the average determined by X-ray crystallography or NMR spectroscopy. By comparing native protein structure with those of complexes, it is apparent that some degree of

Introduction to the principles of drug design and action

94

accommodation to the ligand always occurs on binding. Indeed, in many cases, significant conformational changes accompany the ligand binding. Given the choice of a native protein structure or the structure of the protein partner from a ligand complex, experience has indicated that the latter is the better starting point for ligand design. The problem here, as with basing new design on an active ligand conformer when the structure of the protein binding site is unknown, is the inherent bias of the bound ligand conformation. Clearly one should design much better, if the accommodation of the protein to novel structure were taken into account. The local fluctuations in protein and ligand structures can be introduced in a given mode of binding to yield a free energy. Using Monte Carlo or molecular dynamics methods (see for example, Beveridge and DiCapua (1989), Allen and Tildesley (1987), Valleau and Whittington (1977)), an ensemble of local fluctuations within the ligand and the protein are calculated to yield the thermodynamic functions of the binding. In the former method, the sample space is efficiently explored using an algorithm based on Boltzmann weighting while in the latter, the dynamics of the interactions are explored over a period of nanoseconds. Both methods thus allow for an ensemble of protein structures to be explored and replace the single rigid structure used hitherto. As indicated earlier in the context of building fragments, one problem is the expansion in the number of protein ‘structures’ which are associated with individual designed ligands. and the limits on computing time. Again restriction on the variables undergoing change in the fluctuations to torsional angle subsets may alleviate the problems to some extent. If large conformational changes occur on binding, then the changes are difficult to simulate in any predictive way. Some experimental information on restricting the scale of the structure to be relaxed can be given by the X-ray or NMR structure. NMR determined structures are defined by an ensemble of structures that meet the NMR structural criteria. This ensemble can be used instead of a single structure. The mobility of atoms in structure determined by X-ray crystallography is often represented by an associated temperature factor, and these data could be incorporated into the design process. There is a structural hierarchy in relation to the protein’s accommodation to the ligand, from side-chain reorientation, then local main-chain adjustments and finally large hingebending movements of whole regions of the protein structure. Although many of these decision taking processes may be introduced into automated regimes, the introduction of specific constraints removes some of the objective character of the procedures involved, and all methods are limited by the adequacy of the physical descriptions of the interactions defined in Sections 3.2–3.4 Specific polarizing effects of strong charge interactions inducing changes in the charge distribution both in ligand and in protein are not introduced into standard fast potential routines unless potentials are specifically developed over the sets of ligand and protein atoms for the particular interaction concerned using more fundamental quantum mechanical calculations. There is a case for doing this in any area of detailed study. The difficulties of estimating accurate free energies of binding should not be underestimated. It would, of course, be desirable to calculate all interactions by fundamental quantum mechanical methods but the physical constraint on machine time becomes quickly rate limiting. The scale of the problem with current machine capabilities is summarised in Section 3.7.

Intermolecular forces and molecular modelling

95

3.6 PROTEINS The theoretical determination of protein structure from first principles based on the intramolecular interactions of the individual amino acids, as we remarked earlier, would have high significance in the design of inhibitory or stimulatory ligands in many areas of drug therapy. This is a large subject and we refer to more specialised treatments. The possible number of sequences in an average sized protein of some 400 amino acids is 20400 based on the 20 amino acids and the question as to why only a very small fraction occurs in nature may resolve to structures that have unique and stable native states. A recent paper (Li, Helling, Tang and Wingreen (1996)) which avoids most details of the chemistry of the amino acid interactions examines a polymer of 27 amino acids occupying all sites of a 3×3×3 cube employing simple interactions on a lattice (hydrogen bonding or otherwise). The great majority of sequences have multiple ground states and hence may fold into different structures assuming no inherent large kinetic barrier. Thus ‘foldability’ focusses on the sequence selecting potentially functional ones while ‘designability’ is based on the structure of the resulting protein, which is quantified by measuring the number of sequences that uniquely fold into a particular structure. In evaluating the 227 structures in the simple amino acid scheme, the distribution gives a number of patterns. At the tail of the distribution, there are structures that are highly desirable, they are also more stable. The number of sequences (NS) associated with a given structure (S) differs from structure to structure but preferred structures emerge with NS values much larger than the average. Analysis of the mutation patterns of the homologous sequences for highly designable structures revealed phenomena similar to those observed in real proteins, some sites being highly mutable while others are highly conserved. Although the initial categorisation is elementary, such an approach may offer a pathway to introducing constraints on the multiple minima problem in addition to the already established methods. 3.7 ACCURATE CALCULATION OF INTERMOLECULAR INTERACTIONS In view of the problems of determining molecular interactions accurately, particularly for stronger interactions where the involvement of charge transfer and polarisation are significant, the question may be asked as to why one does not work at a more significant level of accuracy. Here, to provide some perspective to this problem, we simply outline the scale of calculations currently feasible with existing parallel computers. Undoubtedly, the best compromise is to achieve the greatest accuracy possible dependent on the scale of the problem. Given time and effort, if the scale demands the use of fast methods involving empiric potentials, it would undoubtedly be best to develop the best empiric potentials for each particular interaction, dependent on its environment and degree of local interaction. Considerable effort has been made to provide more flexibility in this direction by the categorisation of potentials for particular interactions.

Introduction to the principles of drug design and action

96

In fundamental or ‘ab initio’ quantum mechanical calculations, each electron’s interaction with the nuclei is not strictly independent of the position of other electrons in the system. To simplify the problem, the initial approximation is made that each electron interacts with the average field of the other electrons, i.e. the motions of the electrons are uncorrelated (Hartree-Fock approximation). An electron will thus have kinetic energy while its potential energy will consist of its interaction with other nuclei and with the average field of the other electrons, so that the problem is reduced to a set of one electron equations. Molecular geometries, dipole moments and electrostatic effects may be calculated to good accuracy with this approximation. The neglect of electron correlation, however, means that dispersive or van der Waals interactions are not present, while in situations where electron correlation is important, for example in transition states with molecules near dissociation limits, the approximation is completely invalid. The electrons on each atom are characterised by molecular orbitals and a molecular orbital is constructed from a linear combination of the atomic orbitals. The set of vector functions defining the atomic orbitals is known as a basis set. If one function is used to characterise the atomic orbitals, the set is known as a minimal basis, and broadly viewed, a minimal basis has insufficient flexibility to enable the valence electrons to spread themselves out satisfactorily and such conditions can have different consequences dependent on the occupied and unoccupied orbitals. Providing two functions to characterise each orbital (doubling the basis set) allows much more flexibility in the wave functions but can produce exaggerated properties. As interactions become stronger, the introduction of d orbitals in atoms in the first row of the periodic table becomes significant leading to some 15 basis functions for a first row atom and the size of the basis set rapidly expands even with relatively small molecules. In Hartree Fock-theory, the number of two-eletron integrals rises as n4 where n is the number of basis functions. An alternative approach which has gained ground in recent years is based on the Kohn-Sham theory that an exact solution to the Schrodinger equation exists which leads to self-consistent equations as in Hartree Fock theory. The many-electron problem can be replaced by an exactly equivalent set of one-electron equations with an effective one-particle potential. This effective potential will reproduce the exact density and the exact total energy if the definition of this potential can be defined. The advantage of this density functional approach which has required considerable development is that the scale of this approach rises as n2. Thus for larger problems of chemical interest, the potential becomes high. A decade ago, the cutting edge of computation was a machine with a speed of some 100 mflops/sec but now speeds approach 100–1000 gigaflops/sec. Utilising the benefits of parallelisation of machines applies only to certain calculations where the problem can be dismembered satisfactorily to run time-limiting sections in parallel as with the calculation of two-electron integrals. Using 64 node machines, the practical limit on basis functions using ab initio methods is approximately 4000. This allows dependent on accuracy, an interaction of some 250–1000 atoms. The basis set limit using density functional theory is perhaps 5000. Semi-empirical quantum mechanical methods cannot utilise the benefit of parallelisation beyond about 8 nodes and the practical limit of scale is again of this order. For free energy calculations using empiric potentials, molecular dynamics methods can handle up to 500,000 atoms for 1 nanosec

Intermolecular forces and molecular modelling

97

time scale. Vibrations and rotations with time scales of 10−15 and 10−12 sec. respectively can be handled by such calculations. However docking in molecular recognition (of the order of 10−6 sec.) and translational motion in liquids are on longer time scales. REFERENCES Allen, M.P. and Tildesley, D.J. (1987) Computer Simulations of Liquids. Oxford: Clarendon. Abraham, M.H. (1982) Free energies, enthalpies and entropies of solution of gaseous nonpolar nonelectrolytes in water and nonaqueous solvents. The hydrophobia effect. Journal of the American Chemical Society 104, 2085–94. Abrahams, J.P., Buchanan, S.K., van Raaij, M.J., Fearnley, I.M., Leslie, A.G.W. and Walker, J.E. (1996) The structure of bovine F1-ATPase complexed with the peptide antibiotic efrapeptin. Proceedings of the National Academy of Sciences 93, 9420–4. Beveridge, D.L. and DiCapua, F.M. (1989) Free Energy via molecular simulation: A primer. In Computer Simulations of Biomolecular Systems, edited by W.F.van Gunsteren and P.K.Weiner, pp. 1–26. Leiden: ESCOM. Blumenthal, L.M. (1970) Theory and Applications of Distance Geometry, 2nd edn., Bronx, New York: Chelsea. Brunck, T.K. and Weinhold, F. (1979) Quantum Mechanical Studies on the origin of barriers to internal rotation about single bonds. Journal of the American Chemical Society 101, 1700–9. Cariati, F., Cauletti, C., Ganadu, M.L., Piancastelli, M.N. and Sgamellotti, A. (1980) Spectroscopic investigations on phthalazino(2,3-b)phthalazine-5,12-dione and some of its mono and di-substituted derivatives. Spectrochimica Acta 36A, 1037– 43. Csizmadia, I.G. (ed.) (1982) Molecular Structure and Conformation. Amsterdam: Elsevier. Davies, R.H. (1987) Drug and Receptors in Molecular Biology. International Journal of Quantum Chemistry and Quantum Biological Symposica 14, 221–43. Davies, R.H., Sheard, B. and Taylor, P.J. (1981) Conformation, partition and drug design. Journal of Phamaceutical Sciences 68, 396–97. Deslongchamps, P. (1983) Stereoelectronic Effects in Organic Chemistry. Oxford: Pergamon. Feng, J.-A., Johnson, R.C. and Dickerson, R.E. (1994) Hin recombinase bound to DNA: The Origin of specificity in major and minor groove interactions. Science 263, 348–55. Jorgensen, W.L. and Salem, L. (1973) The Organic Chemist’s Book of Orbitals. New York and London: Academic Press. Kuboniwa, H., Tjandra, N., Grzesiek, S., Ren, H., Klee, C.B. and Bax, A. (1995) Solution structure of calcium-free calmodulin. Natural Structural Biology 2, 768– 76. Li, H., Helling, R., Tang, C. and Wingreen, N. (1996) Emergence of preferred structures in a simple model of protein folding. Science 273, 666–9.

Introduction to the principles of drug design and action

98

Marquart, M., Walter, J., Deisenhofer, J., Bode, W. and Huber, R. (1983) The geometry of the active site and of the peptide groups in trypsin, trypsinogen and its complexes with inhibitors. Acta Crystallographica B39, 480–90. Nederkoorn, P.H.J., Timmerman, H., Timms, D., Wilkinson, A.J., Kelly, D.R., Broadley, K.J. and Davies, R.H. (1997) Stepwise phosphorylation mechanisms and signal transmission within a ligand-receptor-Gαβγ-protein complex recent submission. Page, M.I. and Jencks, W.P. (1971) Entropic contributions to rate accelerations in enzymic and intramolecular reactions and the chelate effect. Proceedings of the National Academy of Sciences 68, 1678–83. Radom, L. (1982) Structural consequences of hyperconjugation. In Molecular Structure and Conformation: Recent Advances, edited by I.G.Csizmadia, pp. 1–64. Amsterdam: Oxford, New York, Elsevier. Reed, A.E., Weinhold, F., Curtiss, L.A. and Potachko, D.J. (1986) Natural bond orbital analysis of molecular interactions: The theoretical studies of binary complexes of HF, H2O, NH3, N2, O2, F2, CO and CO2 with HF, H2O and NH3 . Journal of Chemical Physics 84, 5687–705. Sielecki, A.R., Fedorov, A.A., Boodhoo, A., Andreeva, N.S. and James, M.N.G. (1990) Molecular and crystal structures of monoclinic porcine pepsin refined at 1.8 Å resolution. Journal of Molecular Biology 214, 143–70. South, T.L., Blake, P.R., Hare, D.R. and Summers, M.F. (1991) C-Terminal retroviraltype zinc finger domain from the HIV-1 nucleocapsid protein is structurally similar to the N-terminal zinc finger domain. Biochemistry 30, 6342–9. Taylor, D.A., Sack, J.S., Maune, J.F., Beckingham, K. and Quiocho, F.A. (1991) Structure of a recombinant calmodulin from drosophila melanogaster refined at 2.2 Å resolution. Journal of Biological Chemistry 266, 21375–80. Valleau, J.P. and Whittington, S.G. (1977) A Guide to Monte Carlo for Statistical Mechanics: 1. Highways. In Statistical Mechanics Part A: Equilibrium Techniques edited by B.J.Berne, pp. 137–168. New York and London: Plenum. Vitali, J., Martin, P.D., Malkowski, M.G., Robertson, W.D., Lazar, J.B., Winant, R.C., Johnson, P.H. and Edwards, B.F.P. (1992) The structure of a complex of bovine αthrombin and recombinant hirudin at 2.8 Å resolution. Journal of Biological Chemistry 267, 17670–8. Vos, A.M.de, Ultsch, M. and Kossiakoff, A.A. (1992) Human growth hormone and extracellular domain of its receptor: crystal structure of the complex. Science 255, 306–12.

Intermolecular forces and molecular modelling

99

Plate 3.1 Serine proteases. Proton movement and enzymatic cleavage of the peptide bond. The serine proteases are characterised by an Asp102 – His57 – Ser195 catalytic triad. Experimental (NMR) and theoretical results have indicated that

Introduction to the principles of drug design and action

100

the histidine residue remains neutral thoughout the course of the reaction. The initiating attack of Ser195 on the peptide carbonyl carbon atom is facilitated by the abstraction of the hydroxyl proton by His57. The proton originally residing on His57 is transferred to Asp102 and the incipient negative charge developing on the peptide carbonyl oxygen is stabilised by hydrogen bonding from the main chain -NH groups of residues 193 and 195. The tetrahedral intermediate collapses to an acylated enzyme with the delivery of a proton to the leaving amino group. This proton originates from His57 but delivery may be mediated by a water molecule. Concomitantly, the histidine accepts the proton from Asp102 to regenerate the initial protonation state. Deacylation follows an analogous cycle of proton transfers with a water molecule replacing Ser195 as the nucleophile and with the serine becoming the leaving group.

The figure (Marquart, Walter, Deisenhofer et al. (1983)) shows the catalytic site of trypsin in the presence of the bovine pancreatic trypsin inhibitor (BPTI). The Cα trace of the enzyme (pink) shows the catalytic triad to the right. The scissile carbonyl carbon atom is shown in green. The primary recognition of the peptide bond to be cleaved results from a binding pocket for the substrate side chain in the vicinity of residue 189. The nature of the residues in this pocket predicate the particular specificity of the protease. In the case of trypsin, this residue is an aspartate and specificity is for basic side chains. In the left of the

Intermolecular forces and molecular modelling

101

figure, a lysine side chain is shown interacting with Asp189 and Thr190 via two water molecules.

Plate 3.2 Inhibition of a serine protease and protein-protein recognition. The natural ligand inhibitor, Hirudin binding to the catalytic Asp-His-Ser triad within the serine protease α-Thrombin (Vitali, Martin, Malkowski et al. (1992)). αThrombin has a high specificity for peptide bonds associated with arginine residues and plays a central role in thrombosis and haemostasis. It is the product of prothrombin cleavage by factor Xα in the final step of the blood clotting cascade, and consists of two polypeptide chains, A and B, connected through a single disulphide bond.

Introduction to the principles of drug design and action

102

During clotting, α-thrombin converts fibrinogen into fibrin by removing fibrinopeptide A from the Aα-chain and fibrinopeptide B from the Bβ-chains of fibrinogen. Hirudin is a small protein of 65 residues and 3 disulphide bonds that is isolated from the glandular secretions of the leech Hirudo medicinalis and is a potent natural inhibitor of thrombin. The figure shows the large surface area of contact of the Hirudin inhibitor (blue) with the serine-protease, bovine α-thrombin (brown). The Asp102 – His57 – Ser195 catalytic triad of the enzyme (elemental colouring) is blocked by the first three residues of the N-terminal chain of Hirudin (refer to the Hirudin N-terminus in green). In human thrombin, two hydrogen bonds from the amino terminal group exist. In the crystal developed at pH 4.7, one is to the carbonyl group of Ser214 and the second is to the catalytic serine residue 195. For the crystal developed at pH 7, this second bond is to His57. Neither bond is formed in this bovine complex at pH 4.7 indicating that a second bond may not be essential for Hirudin binding. Specific binding to the associated binding site for arginine residues does not occur (compare the bovine pancreatic trypsin inhibitor in Plate 3.1) but a number of exo-sites on the surface of the thrombin can interact with the inhibitor. The last sixteen residues of hirudin are in an open conformation and bind between the two loops of the enzyme surface formed by Phe34 to Leu41 and by Lys70 to Glu80. This region of the enzyme is marked by positively charged side chains and interaction with Hirudin’s anionic residues Asp53, Asp55, Glu56, Glu57. The latter three residues are shown in red. Salt bridges are formed by Asp55 and Glu57 interacting with the enzyme residues Arg73 and Arg75 respectively.

Intermolecular forces and molecular modelling

103

Plate 3.3 Aspartate proteases. As for the serine proteases, electron reorganisation coupled to proton movement is critical

Introduction to the principles of drug design and action

104

to the cleavage of the peptide bond. In this case, the catalytic site consists of two adjacent aspartate residues, Asp32 and Asp215 (pepsin numbering) which localise a solvent water molecule between their carboxyl groups. This highly polarised water molecule is the initiator of the peptide bond hydrolysis. Studies of the pH dependence of catalysis by porcine pepsin leads to estimates of two pKa values of 1.2 and 4.7 and, hence, one of the apartate residues is thought to be protonated in the resting state. High refinement (1.8 Å (Sielecki, Fedorov, Boodhoo et al. (1990)) of the pepsin structure shows that the oxygen atom of the catalytic water molecule lies in the plane of Asp215 whereas the carboxylate group of Asp32 is twisted by some 22° with repect to this common plane. In the resting state, the carboxylate oxygen atoms are arranged such that one from each residue is within hydrogen bonding distance of the water molecule and the two adjacent ‘inner’ carboxylate oxygen atoms from each residue are also hydrogen bonded together. From the interatomic distances, the ‘inner oxygen atom of Asp32 and the ‘outer’ oxygen of Asp215 are hydrogen bonded to the water. Hence the probable location of the proton that forms the hydrogen bond between the carboxylate groups is on the ‘inner’ oxygen atom of Asp215. The shorter contact distance from the water molecule is also to Asp32 (2.6 Å) rather than to Asp215 (2.9 Å.) suggesting that in the resting state, it is Asp215 that is protonated. The proposed hydrogen bond length between Asp32 and Asp215 is 2.8 Å. The precise

Intermolecular forces and molecular modelling

105

mechanism of catalysis is not known and the following description represents one possible hypothesis. Nuceophilic attack on the peptide carbonyl carbon atom is probably facilitated by movement of a proton from the water molecule to Asp215 along with proton transfer from Asp215 to Asp32. The ‘inner’ oxygen atoms are located via hydrogen bonds from the main chain NH groups of Gly34 and Gly217 and two residues, Ser35 and Thr218 may hydrogen bond to the ‘outer’ carboxylate oxygen atoms. Ser35 may help to stabilise the incipient oxyanion of the tetrahedral intermediate and Thr218 may position a second water molecule in order to mediate the transfer of the proton from Asp215 to the leaving amino group on breakdown of the tetrahedral intermediate. As the proton is transferred from the ‘outer’ oxygen atom of Asp215, the proton on Asp32 is transferred to the ‘inner’ carboxylate of Asp215 so restoring the initial state. The hypothetical proton and electron reorganisations are shown in the scheme below

Inhibitors of aspartate proteases such as pepstatin, displace the catalytic water molecule by an appropriately orientated hydroxyl group. The figure shows the pepsin catalytic site with the resident water molecule superimposed on a second structure determined in the presence of pepstatin (green). In the centre of the figure are the hydroxyl group of pepstatin and the catalytic water molecule with Asp32 located to the left.

Introduction to the principles of drug design and action

106

This page intentionally left blank.

Plate 3.4 Protein-Protein recognition. The influence of a hormone on protein dimerisation. Human growth hormone (hGH) binding to the extracellular domain of its receptor (de Vos, Ultsch and Kossiakoff (1992)). The binding of hGH to its receptor is required for regulation of normal human growth and

Intermolecular forces and molecular modelling

107

development. The extracellular domain of the receptor (hGHbp) complex, here shown as a ribbon structure, consists of one molecule of growth hormone per two molecules of receptor (orange and blue respectively). The hormone (lilac) is a four helix bundle. The binding protein consists of two distinct domains which have some similarity to immunoglobulin domains. In the complex, both receptors donate essentially the same residues to interact with the hormone even though the two binding sites on hGH have no structural similarity. In addition to the hormonereceptor interfaces, there is also substantial contact between the carboxyl-terminal domains of the receptors. The core of the helix bundle is made up of primarily hydrophobic residues. The extracellular part of the receptor consists of the two domains linked by a four residue segment of polypeptide chain. Each domain contains seven β-strands that together form a sandwich of two antiparallel β-sheets, one with four strands and one with three with the same topology in each domain. The thirty residues of the receptor’s amino terminal domain show conformational flexibility and are not given in the crystal structure. The carboxy-terminal domains are closely parallel, the termini pointing away from the hormone in the expected direction of the membrane. Intact receptors would have an additional eight residues at the end of the seventh strand (bottom right) which form the putative membrane spanning helix.

Introduction to the principles of drug design and action

108

Plate 3.5 A potential ligand-activated proton pathway for signalling in a guaninenucleotide-coupled receptor ternary complex acting as a guanosine triphosphate synthase (Nederkoorn, Timmerman, Timms et al. (1997)). The ternary complex consists of a seven (helical trans-membrane receptor (yellow), a heterotrimeric Gαβγ-protein

Intermolecular forces and molecular modelling

109

(α-pink, β-brown, γ-white) and an activating ligand. The nucleotide guanosine di-phosphate (GDP) resides within the Gα-subunit (elemental colouring to the left of the figure). On ligand activation, a series of events are set in train and the existing interpretation is that on signal stimulation GDP is exchanged for the triphosphate, GTP, causing separation of the Gα- and Gβγ-subunits from the receptor on the cytoplasmic site of the cell. The high energy Gα..GTP and the Gβγ-subunits can then both activate second messengers within the cell. An alternative interpretation of this mechanism is that a direct phosphorylation of the GDP initially occurs. A proton signalling pathway can exist to a histidine residue holding a metaphosphate group as an acid-labile phosphoramidate on the Gβ-subunit. Transfer of the metaphosphate group to an arginine residue (blue) at the base of the Gα- α2-helix (green) at the interface between the Gα- and Gβγ-proteins to form a high energy phosphonoarginine intermediate, allows transport of the phosphate group to the phosphorylation site. The histidine phosphoramidate is shown at the base of the figure in elemental colouring. The primary mechanism for delivering a proton over the 43 Å through a set of local TyrArg/Lys-Tyr proton shuttles is seen to reside in the balance between TyrArg/Lys ion pair and neutral complexes under multiple hydrogen bonded conditions within a hydrophobic environment. Under such conditions the isolated neutral Tyr-Arg complex is some 12–14 kcal/mol more stable than

Introduction to the principles of drug design and action

110

the ion pair form but transfer of a proton can occur under the influence of two hydrogen bond proton donor interactions on the phenolic oxygen atom. The figure shows six tyrosine residues (yellow) within the proposed signalling pathway together with their associated bases. A comparable mechanism is likely also to occur with cysteine/base residues under appropriate conditions. The partially stimulating analogue, prenalterol containing the 4-hydroxy-phenoxy moiety is shown at the top of the receptor figure interacting with Asp138 and initiating proton transfer from Tyr377 in the β1-adrenoceptor. At the top of the α2-helix two retaining bonds are broken within the full ternary complex allowing movement of the α2helix and carrying the metaphosphate group over the last 30 Å to the phosphorylation site.

Intermolecular forces and molecular modelling

111

Plate 3.6(a) Adenosine triphosphate synthase (ATP synthase, F1 F0 synthase) is the central enzyme in energy conversion in mitochondria, chloroplasts and bacteria. and uses a proton gradient across the mebrane to synthesis ATP from the diphosphate, ADP and inorganic phosphate. The multi -subunit assembly consists of a globular domain, F1, and an intrinsic membrane domain, F0, linked by a slender stalk about 45 Å long. The F1 domain is an approximate sphere 90–100 Å in diameter and contains the catalytic binding sites for the substrates ADP and inorganic

Introduction to the principles of drug design and action

112

phosphate. About three protons flow through the membrane per ATP synthesised but the mechanism of synthesis is not known. The F1 structure is a complex of five different proteins with the stoichiometry 3α:3β:1γ:1δ:1ε. The sequences of the α- and β-subunits are homologous (~20% identical), including the P-loop nucleotide-binding motif. The catalytic sites are in the βsubunits while the function of the asubunits are obscure. It has been suggested that the structures of the three catalytic sites are always different, but each passes through a cycle of ‘open’, ‘loose’ and ‘tight’ states. In this respect crystals developed with AMP-PNP (where the nitrogen atom defines the analogue of ATP) show occupancy of the nucleotide sites in different states of phosphorylation. The α- and β-subunits are arranged alternatively like the segments of an orange around a central α-helical domain containing both the Nand C- terminals of the γ-subunit. As the three β-subunits vary in nucleotide occupancy (ADP, AMP-PNP, and empty) and have different conformations, the structure as found in the crystal (2.8 Å resolution) is compatible with one of the states to be expected in the cyclical binding change mechanism (Abrahams, Buchanan, van Raaij et al. (1996)). The figure shows the arrangement of the three α-(A, pink; B, blue; C, green) and β-(D, purple; E, yellow; F, white), around the central F0 stalk (orange). The positions of nucleotides are given in elemental colouring. Plates 3.6(b) and (c) show the similarity of the binding sites of the nucleotides in the α- and β-subunits. (b) the

Intermolecular forces and molecular modelling

113

nucleotide AMP-PNP is between the A, α- and D, βsubunits. All the nucleotide binding sites are in the α(A, pink) except for those indicated (β-, D (purple)). The magnesium ion assisting the phosphorylation is shown in red between the two terminal phosphate groups, (c) The ADP is bound very predominantly to the, β-(D, purple) subunit. The relations to the α-, (C (green)) subunit are indicated.

Introduction to the principles of drug design and action

114

Intermolecular forces and molecular modelling

115

Plate 3.7 The influence of strong charge on conformation. The structure of calmodulin with and without the interaction of 4 calcium ions. Calmodulin is the principal calciumdependent regulator of a variety of intracellular processes. The 148 residue protein has four Ca++ sites and a number of acidic residues. It is a ubiquitous protein in eukaryotes and plays a critical role in coupling transient Ca++ influx, caused by a stimulation at the cell surface, to events in the cytosol. The Ca++ binding sites have the ‘EF hand’ configuration also identified in other Ca++ binding proteins such as intestinal calcium binding protein and troponin C. The ‘EF hand’ comprises a helix-loop-helix structure which can be identified from the sequence homology alone. The basic structural unit of the globular domain consists of a pair of EF-hands rather than a single binding site. Plate 3.7(a) Left. Calcium-bound calmodulin from Drosophila melanogaster (2.2 Å resolution—Taylor, Sack, Maune et al. (1991)) has a seven turn a-helix connecting the two calcium-binding domains. The dumbbell shaped molecule contains seven α-helices and four ‘EF’ calcium-binding sites and closely resembles the mammalian structure. The six-coordination octahedral form of a binding site is shown in Plate 3.7(b) where the Ca++ ion is held by four acidic residues. In each site, the coordination (one shared) comes from five side-chain oxygen atoms, a carboxyl oxygen (not shown) and one water molecule. Plate 3.7(a) Right. The NMR determined calcium-free structure of calmodulin (Kuboniwa, Tjandra, Grzesiek et al. (1995)). Each calmodulin domain consists of a strongly twisted but tightly packed bundle of four helices. Upon binding of Ca++ most of the change occurs within each of the ‘EF hands’ with inter-helix angle changes. The structural rearrangement on binding Ca++ ion results in a pronounced hydrophobic pocket on

Introduction to the principles of drug design and action

116

the surface of each domain. These pockets appear to be of importance from structure studies on Ca++ bound complexes with different synthetic target peptides. The accuracy of NMR determined structures is highest at the centre of the protein and decreases as one moves towards the surface. The accuracy in the determination of the Ca++ binding loops requires, in principle, further refinement. The conformation of the long central helix in the crystal structure was not previuosly consistent with extensive biochemical data on these proteins. The Ca++ free structure shows increased flexibility and this ‘connecting spacer’ can be viewed as a flexible tether between the two domains. This is confirmed by by Xray structures on calmodulin complexed with peptide fragments of its intracellular receptors, e.g. myosin light-chain kinase where the two domains of cadmodulin swing round and envelope the target peptide.

Intermolecular forces and molecular modelling

117

Plate 3.8 Protein- Single strand DNA recognition. A zinc finger domain binding to a single stranded DNA sequence. Interaction of an NMR-determined zinc finger domain in the HIV-1 nucleocapsid protein (South, Blake,

Introduction to the principles of drug design and action

118

Hare and Summers (1991)). A common feature of proteins containing the ‘retroviral-type’ (r.t.) zinc finger domain (Cys – X2 – Cys – X4 – His – X4 – Cys) is that they appear to be involved at some stage in sequencespecific single-stranded nucleic-acid binding analogous to the zinc finger motif found widely in duplex-DNAbinding proteins. Zinc finger r.t. domains are found both in the Nterminal and C-terminal chains of the intact HIV-1 nucleocapsid protein isolated from virus particles. The sequences have been shown to bind zinc stoichiometrically and with high affinity. The figure shows an eighteen amino acid HIV1-F1 peptide Cα sequence (Val-Lys-Cys-Phe-Asn-CysGly-Lys-Glu-Gly-His-Ile-Ala-Arg-AsnCys-Arg-Ala in pink) bound to a single strand DNA sequence A-C-G-C-C). The tetrahedral coordination of the Zn ion with the three cysteine residues and His11 is shown bonded schematically on the right of the figure. The hydrophobic interactions of the peptide residues (Phe4, Ile12, Ala13,) are shown in green while the strong polar interaction of Arg14 with DNA backbone phosphate groups is seen at the end of the finger.

Intermolecular forces and molecular modelling

119

Plate 3.9 Protein-Double strand DNA recognition. The selectivity of protein binding in the

Introduction to the principles of drug design and action

120

major and minor grooves of the DNA. The binding of the prokaryotic enzyme Hin recombinase to DNA in the Salmonella chromosome (Feng, Johnson and Dickerson (1994)). This site-specific recombination reaction controls the alternate expression of two flagellin genes by reversibly switching the action of a promoter. During the process of inverting the extended segment of DNA, two Hin proteins in the form of a dimer bind to the the left and right recombination sites located at the boundaries of the invertible DNA segment. Through interaction with a third interacting site (held by an additional protein) the overall complex aligns the two recombination sites correctly and the Hin protein is activated to initiate the exchange of DNA strands leading to inversion of the intervening DNA. The recombination half-site of the double helical sugarphosphate backbone of the DNA (elemental colouring) linked by the heterocyclic base pairs (blue) is shown occupied by the helix-loop-helix-loop-helix of the Hin protein. The third Hin helix (green) sits in the major groove of the DNA where the residues Arg 178, Thr 175 and Tyr 179 are shown on the lower side of this helix. Helices 1 and 2 (purple) are approximately orthogonal to helix 3. The amino terminal loop (white) at the bottom right of the picture attached to Helix 1 lies in the minor groove with two arginine residues (140 and 142) interacting with the helical backbone of the DNA. The carboxyl terminal chain extending from helix 3 (white) leads again into the minor groove at the upper left of the figure where the portion of the chain interacting with the DNA is shown in pink. The short loops joining helices 1 and 2 (top right) and helices 2 and 3 (middle right) are also indicated in white. Water molecules within the X-ray crystal structure (determination at 2.3 Å resolution) are shown with a white cross.

4. DRUG CHIRALITY AND ITS PHARMACOLOGICAL CONSEQUENCES ANDREW J.HUTT CONTENTS 4.1 INTRODUCTION

98

4.2 DEFINITIONS AND NOMENCLATURE

99

4.2.1 Nomenclature and designation of stereoisomers

103

4.2.2 The nomenclature problem in generic names

111

4.2.3 Prochirality

112

4.3 BIOLOGICAL ACTIVITY

112

4.3.1 Terminology used in the pharmacological evaluation of stereoisomers

116

4.3.2 “Purity” of enantiomerically pure drugs

117

4.3.3 Receptor selectivity

119

4.3.4 Quantitative structure—activity relationships

120

4.4 PHARMACOKINETIC CONSIDERATIONS

122

4.4.1 Absorption

122

4.4.2 Distribution

124

4.4.3 Metabolism

126

4.4.4 Excretion

133

4.4.5 Pharmacokinetic parameters

134

4.5 PHARMACODYNAMIC CONSIDERATIONS

135

4.5.1 The pharmacological activity resides in one enantiomer the other being biologically inert

135

Introduction to the principles of drug design and action

122

4.5.2 Both enantiomers have similar activities

135

4.5.3 Both enantiomers are marketed with different indications

136

4.5.4 The enantiomers have opposite effects

136

4.5.5 One enantiomer may antagonise the side effects of the other

139

4.5.6 The required activity resides in one or both enantiomers but the adverse effects are predominantly associated with one enantiomer

139

4.6 SELECTED THERAPEUTIC GROUPS 140 4.6.1 Antiarrhythmic agents 140 4.6.2 β—Blockers 143 4.6.3 Anticoagulants 147 4.6.4 Antihistamines 148 4.6.5 Non-steroidal anti-inflammatory drugs 150 4.6.6 Antimicrobial agents 155 4.7 TOXICOLOGY 160 4.8 RACEMATES VERSUS ENANTIOMERS: THE FUTURE

163

4.9 CONCLUDING COMMENT 165

Drug chirality and its pharmacological consequences

FURTHER READING

123

165

4.1 INTRODUCTION One in four of all therapeutic agents are marketed and administered to man as mixtures. These agents are not drug combinations in the accepted meaning of the term, i.e. two or more coformulated therapeutic agents, but combinations of isomeric substances the biological activity of which may well reside predominantly in one isomer. The majority of these mixed formulations arise due to the use of racemic mixtures of synthetic chiral drugs and less frequently to mixtures of diastereoisomers. Over the last ten years there has been considerable interest in the area of drug chirality as a result of recent advances in the stereoselective synthesis and stereospecific analysis of chiral molecules. As a result of these advances and the realization of the significance of the pharmacodynamic and pharmacokinetic differences between the enantiomers of chiral drugs there has been increasing concern over the use of racemates, and other stereoisomeric mixtures, in therapeutics. The use of such mixtures may present problems particularly if the adverse effects, or toxicity of the drug is associated with the less active, or inactive isomer, or does not show stereoselectivity. Many authors regard racemates as “compounds which contain 50% impurity” and that their use is “polypharmacy” with the proportions of the materials being dictated by chemical rather than pharmacological or therapeutic criteria. As a result of these concerns drug stereochemistry has become an important consideration for both the pharmaceutical industry and the major drug regulatory authorities. The extent of the problem can be appreciated from the results of a survey of 1675 drugs carried out in the early 1980s. Of these agents 1200 (72%) were classified as synthetic and 475 (28%) as natural products or semisynthetic agents. Of the compounds classified as natural products or semisynthetics 469 were chiral and of these 461 (98%) were marketed as single isomers. In contrast 29% (480) of the synthetic compounds were chiral with only 3.5% (58) being available as single isomers the remainder (25%) being marketed as racemates. More recent investigations have indicated that the position regarding natural and semisynthetic agents has not changed greatly but that the proportion of synthetic agents available as single isomers had increased. From the above figures it is obvious that drug chirality is an “acrossthe-board” problem, mixtures of stereoisomers being found in the majority of therapeutic groups. Biological environments at a molecular level are highly chiral being composed of chiral biopolymers, e.g. proteins, glycolipids and polynucleotides, from the chiral building blocks of L-amino acids and the D-carbohydrates. As nature has made a preference in terms of its stereochemistry it is not surprising that enzymes and receptor systems show a stereochemical preference for one of a pair of stereoisomers. The interaction of a drug with a receptor, or enzyme active site, involves interaction between the functionalities of the drug molecule and complementary sites or groups on the receptor. Such interactions may have considerable steric constraints in terms of interatomic distance and steric bulk between such functionalities. In the case of

Introduction to the principles of drug design and action

124

stereoisomers the three dimensional spatial arrangement of the functionalities is of considerable significance. Enantioselectivity in drug action should not be surprising as many of the natural ligands are themselves chiral, e.g. neurotransmitters, autocoids, hormones, endogenous opioids etc. Indeed Lehmann has stated that “the stereoselectivity displayed by pharmacological systems constitutes the best evidence that receptors exist and that they incorporate concrete molecular entities as integral components of their active sites”. 4.2 DEFINITIONS AND NOMENCLATURE Stereochemistry is concerned with the three dimensional spatial arrangement of the atoms within a molecule. The prefix stereo originating from the Greek stereos meaning solid or volume. Stereoisomers are compounds which differ only in the threedimensional arrangement of their constituent atoms in space and such isomers may be divided into two groups namely enantiomers and diastereoisomers. Enantiomers are pairs of compounds which are non-superimposable mirror images of one another and in terms of physicochemical properties, differ only in their ability to rotate the plane of plane polarised light which is equal in magnitude but opposite in direction. Such isomers are said to be chiral (from the Greek chiros meaning handed) and are variously referred to as optical isomers or enantiomorphs (Greek enantios opposite, morph form). The term diastereoisomers refers to all other stereoisomeric compounds regardless of their ability to rotate the plane of plane polarised light and the definition therefore includes both geometrical, i.e. cis/trans isomers and optical isomers. A fundamental distinction between enantiomerism and diastereoisomerism is that in a pair of enantiomers the distances between nonbonded atoms are identical, whereas in diastereoisomers they are not. Thus, the energy content of a pair of enantiomers is essentially identical, whereas a pair of diastereoisomers differ in energy and hence in their physico-chemical properties. This fundamental difference in the properties of the two types of stereoisomer has considerable significance as mixtures of enantiomers cannot be readily separated by standard chemical techniques, whereas diastereoisomers may be separated, in principle at least by distillation, recrystallisation and chromatography. In terms of the compounds of interest in medicinal chemistry and pharmacology the most frequent cause of chirality arises from the presence of a tetracoordinate carbon atom in a molecule to which four different atoms or groups are attached (4.1), i.e. a chiral or asymmetric centre. The presence of one such centre in a molecule gives rise to a pair of enantiomers, the presence of n such different centres yields 2n stereoisomers and half that number of pairs of enantiomers. Those stereoisomers which are not enantiomeric being diastereoisomeric. Diastereoisomers which differ in configuration about one chiral centre only are termed epimers. In additional to carbon other atoms frequently found in organic molecules have a coordination number of four and a tetrahedral arrangement of the attached ligands,

Drug chirality and its pharmacological consequences

125

e.g. silicon, nitrogen, phosphorus and sulphur, it is not surprising that optically active compounds of these elements are also known. In the case of trivalent derivatives of nitrogen the lone pair of electrons may be considered to be the fourth ligand. However, inversion of the pyramidal forms occurs very rapidly via a planar transition state (4.2→ 4.3→4.4), the activation energy for the inversion being very low, so that separation of enantiomers is not possible. However, the formation of quaternary ammonium compounds, e.g. the neuromuscular blocking agent atracurium besylate (4.5), or the formation of an amine oxide, e.g. N-ethyl-N-methylaniline N-oxide (4.6), results in the formation of a chiral centre. In contrast to trivalent derivatives of nitrogen, trivalent pyramidal sulphur derivatives have a higher energy of activation for inversion and the rate is slow enough that the individual enantiomers are relatively stable. Examples of drug molecules containing

Introduction to the principles of drug design and action

126

chiral sulphur and phosphorus centres include the non-steroidal anti-inflammatory prodrug sulindac (4.7) and the phosphamide mustard pro-drug cyclophosphamide (4.8). Compounds which do not possess a chiral centre in their structure may also exist in enantiomeric forms as a result of an axis or plane of chirality. Such systems occur less frequently in compounds of pharmaceutical interest. Atropoisomerism (Greek, atropos inflexible) is a term used to characterise stereoisomers which are chiral due to

Drug chirality and its pharmacological consequences

127

hindrance of rotation about a single bond, e.g. suitably substituted biphenyl derivatives (4.9). In this case rotation about the carbon-carbon bond linking the two phenyl rings is restricted by the steric effect of the substitutents resulting in configurational stability. Examples of interest include the hypnotic methaqualone (4.10, 4.11) and the male antifertility agent gossypol (4.12). The presence of adjacent double bonds as found in allenes also gives rise to enantiomerism, e.g. structures (4.13) and (4.14). In the case of these compounds the substituents lie in intersecting planes and the two structures are non-superimposable, such molecules possess a chiral axis. This type of isomerism is found in the naturally occurring antibiotic mycomycin (4.15). Steric crowding in a molecule may also give rise to enantiomeric structures as a result of distortion, e.g. the helicenes. The simplest helicene consists of six ortho fused benzene rings (4.16) and as a result of steric crowding the molecule is not planar but helical. The

Introduction to the principles of drug design and action

128

molecules exist in either left or right handed spirals in the same way that a corkscrew, or spiral staircase, may be either left (4.17) or right (4.18) handed. Thus, chirality may arise as a result of helicity in a structure. While the above example has little

Drug chirality and its pharmacological consequences

129

significance in medicinal chemistry it is important to appreciate that a number of biologically important macromolecules exist as helical structures. For example the αhelix of proteins, composed

of L-α-amino acids, is right handed and the two polynucleotide strands of the DNA double helix wind around a common axis with a right-handed twist. In the case of these molecules not only are the individual building blocks, i.e. the amino acids and nucleotides chiral but the biopolymers themselves exhibit chirality. 4.2.1 Nomenclature and Designation of Stereoisomers The classical method of distinguishing between a pair of optical isomers makes use of their unique property of rotation of the plane of plane polarised light. Those isomers which rotate light to the right being termed dextrorotatory, indicated by a (+)-sign

Introduction to the principles of drug design and action

130

before the name of the compound while those which rotate light to the left are termed laevorotatory indicated by a (−)-sign. In the older literature the letters d- and l- are also used to indicate (+)- and (−)- enantiomers respectively. The use of these lower case letters gives rise to confusion with the upper case D and L designation for configuration and should be avoided. A racemic mixture, a 1:1 mixture of enantiomers, is indicated by a (±)-sign before the name of the compound. It is important to appreciate that this form of designation yields information concerning a physical property of the material but does not provide information concerning the three dimensional spatial arrangement, or absolute configuration, of the molecule. Also considerable care is required when using the direction of rotation as a stereochemical descriptor as both the magnitude and direction of rotation may vary with the conditions used to make the determination, e.g. temperature, solvent, analyte concentration etc. For example the antimicrobial agent chloramphenicol (4.19) contains two chiral centres and therefore four stereoisomeric forms are possible. The active isomer has the R,R-absolute configuration (see p. 107 for convention), but this stereoisomer is dextrorotatory when the optical rotation is determined in ethanol and laevorotatory in ethyl acetate. Similarly the active S-enantiomers of the 2arylpropionic acid (4.20) non-steroidal anti-inflammatory agents, e.g. ibuprofen (4.21), are dextrorotatory as the free acids but the corresponding sodium salts are laevorotatory. Additional complications arise if the drug material is a mixture of two diastereoisomers, e.g the β-lactam antimicrobial agent moxalactam (latamoxef) is a mixture of two epimers which are both laevorotatory. Their designation, based on the configuration of the side chain chiral centre and optical rotation are (−)-(R)- and (−)(S)-moxalactam (4.22). In this

Drug chirality and its pharmacological consequences

131

case the designation of the material by optical rotation is meaningless and provides no information concerning the stereochemical composition of the material, i.e. single isomer or mixture. Once the three dimensional structure of a stereoisomer has been determined, by for example X-ray crystallography, then the absolute configuration of a molecule may be indicated by the use of a prefix letter to the name of the compound. Two systems are currently used, the R/S or Sequence Rule nomenclature of Cahn, Ingold and Prelog and the older D/L system of Fischer and Rosanoff. One of the major problems in organic chemistry is the representation of three dimensional structures on two dimensional sheets of paper, the relationships between stereoisomers can best be seen and understood by the use of molecular models. The Fischer projection, devised by the carbohydrate chemist Emil Fischer, is a common method for two dimensional representations of three dimensional structures. In Fischer projections the structure is drawn in a vertical rather than horizontal form with the lowest numbered carbon atom, in standard nomenclature terms, or the most highly

Introduction to the principles of drug design and action

132

oxidised end of the chain, drawn at the top. At each chiral centre along the main axis of the molecule the vertical bonds project back away from the reader while the horizontal bonds project up towards the reader. In the case of glyceraldehyde (2,3dihydroxypropanal) the simplest carbohydrate, containing one chiral carbon atom the individual enantiomers are drawn thus (4.23):

The chiral carbon atom is regarded as being in the plane of the paper and those groups which are bonded horizontally, i.e. the H and OH project up towards the reader and those bonded vertically, i.e. the CHO and CH2OH project back away from the reader. Thus the above structures represent (4.24):

The structure of glyceraldehyde with the secondary hydroxyl group drawn on the right of the Fischer projection was designated as having the D configuration and that with the secondary hydroxyl on the left the L configuration. At the time this representation of structure was developed there were no methods for the determination of the three dimensional nature of molecules and the observed optical rotations of the two enantiomers were arbitrarily assigned as D-(+) and L-(−). At this time the letters d and 1 were used to indicate the direction of rotation rather than (+) and (−), and this combination of both upper and lower case letters to define both the physical property and the shape of the molecule, as pointed out above, continues to add to the confusion associated with the study of stereochemistry. It was not until the 1950s that it was possible to show that the optical rotation assignment in fact corresponded to the structures drawn which was highly fortuitous. Stereoisomers of compounds which can be related to D-glyceraldehyde by synthesis, are given the D-configuration,

Drug chirality and its pharmacological consequences

133

irrespective of the observed direction of rotation of polarised light and compounds related to L-glyceraldehyde are given the L-configuration. For example (+)-glucose (4.25), (−)-2-deoxyribose (4.26) and (−)-fructose (4.27) having the terminal configuration of (D)-(+)-glyceraldehyde, are assigned to the D-series. In the case of the amino acids the reference compounds used are D-(+)- and L-(−)-serine (4.28). The use of this system presents a number of problems particularly if there is more than one chiral centre in the molecule. Thus the amino acid L-threonine (4.29) may be related to L-serine at carbon 2 and D-glyceraldehyde at carbon 3. In the case of the αamino acids the α-carbon atom is used to define the stereochemistry and the majority of natural amino acids have the L-configuration at this centre. D-amino acids are however found in a number of peptide antibiotics, e.g. bacitracin, penicillins etc.

Introduction to the principles of drug design and action

134

In an attempt to overcome the difficulties associated with the D/L system Cahn, Ingold and Prelog devised their Sequence Rule system. Using this method the substitutent atoms attached to a chiral centre are ranked in order of priority which is based on their atomic number. The higher the atomic number the greater the priority. If a decision on priority cannot be made on the basis of the atoms directly attached to the chiral centre then the atoms two bonds away are considered. This process is continued along a substituent until all the priorities have been assigned. The molecule under examination is then viewed from the side opposite to the group of lowest priority and if the priority sequence, highest to lowest, is to the right (i.e. clockwise) then the centre is of the Rabsolute configuration (Latin rectus, right) and if to the left (i.e. anticlockwise) the Sabsolute configuration (Latin sinister, left). In the case of glyceraldehyde (4.23) the priority order of the groups is: HO(highest), -CHO, -CH2OH, H (lowest). The aldehyde group has a higher priority than the primary alcohol as the aldehydic carbon atom is considered to be bonded to two oxygen atoms, one “real” and one “ghost” or “phantom” oxygen so that the carbonoxygen double bond is taken into account. The application of these rules to the enantiomers of glyceraldehyde is illustrated below (4.23).

Thus D-(+)-glyceraldehyde has the R-absolute configuration using the Cahn, Ingold, Prelog sequence rules and L-(−)-glyceraldehyde has the S-absolute configuration.

Drug chirality and its pharmacological consequences

135

The naturally occurring catecholamines, (−)-noradrenaline (4.30) and (−)adrenaline (4.31) have been stereochemically related, by chemical degradation studies, to D-(−)-mandelic acid (4.32) and therefore these two compounds are assigned the Dconfiguration. In the case of noradrenaline (4.30) and adrenaline (4.31), and related chiral derivatives of phenylethylamine, the convention regarding the presentation of Fischer projections with the lowest numbered carbon atom at the top is not strictly applied. These agents are conventionally drawn “upside down” as Fischer projections as shown in the structures (4.30) and (4.31). Redrawing the Fischer projections of (4.30) and (4.31) to a form suitable for assigning the configuration yields structure (4.33), and examination of the sequence indicates that the D-enantiomers of both catecholamines correspond to the R-absolute configuration. One of the major problems with stereochemical nomenclature is the continued use of both the above systems for designation of absolute configuration and also the use of the physical descriptors (+) and (−). The potential problems associated with the use of the

Introduction to the principles of drug design and action

136

physical descriptors has been presented above. The reason the D/L system continues to be used is essentially biochemical. For example D-(+)-glucose (4.25), the enantiomer of glucose found in biological systems, could be known as (2R, 3S, 4R, 5R)-2,3,4,5,6pentahydroxyhexanal or (2R, 3S, 4R, 5R)-aldohexose, which does not take into account the ring structure of the molecule and the two anomeric forms. It is obviously simpler to refer to D-(+)-glucose. Also the naturally occurring amino acids, are of the L-configuration and the application of the R/S system results in a lack of consistency within the series. For example L-Serine (4.28) has the S-absolute configuration while L-cysteine (4.34) has the R-configuration as a result of the presence of the sulphur atom. Additional complexities may also arise in the nomenclature of semisynthetic products, as in some cases both systems are used to define the structure of the molecule. For example the absolute stereochemistry of the 6-aminopenicillanic acid and 7-aminocephalosporanic acid nucleii have been determined and defined in terms of the R/S system but the addition of a side chain, e.g. ampicillin (4.35) and cephalexin (4.36), may result in the introduction of an additional chiral centre, which in the case of these two compounds is frequently

Drug chirality and its pharmacological consequences

137

defined in terms of the D/L system. Thus, the British Pharmacopoeia (1993) defines ampicillin as (6R)-6-(α-D-phenylgrycylamino) penicillanic acid (4.35) and cephalexin as 7-α-D-phenylglycylamino-3-methyl-3-cephem-4-carboxylic acid (4.36). The side chain chiral centre being denoted by the D/L system and only in the case of ampicillin is the stereochemistry of the ring system indicated and then for only one of the three chiral centres. Within the literature the two possible diastereoisomers arising from the introduction of the side chain of such compounds are frequently referred to in terms of D and L. It is important to appreciate that the stereochemical designations, R and S, are defined by a set of arbitary rules and that with respect to biological activity the relevant feature is the three dimensional spatial arrangement of the functionalities within the molecule. A change in one functional group may result in an alteration of the configurational designation but have no influence on the relative orientation of the functionalities required for biological activity with respect to one another. For example the active enantiomers of the 2-arylpropionic acid NSAIDs have the Sconfiguration (4.20) which corresponds to the R-configuration of the 2aryloxypropionic acid herbicides (4.37). Similarly in the case of the β-blockers the

Introduction to the principles of drug design and action

138

active agents of the arylethanolamine series have the R-configuration (4.38) whereas those of aryloxypropanolamine series have the S-configurational (4.39) designation.

The metabolism of a drug may also result in an alteration of configurational designation with no change in the spatial arrangement of the functionalities. For example fonofos (4.40*), a cholinesterase inhibitor, undergoes oxidation to yield fonofos-oxon (4.41) which is also active. As a result of the sequence rule designations the R-enantiomer of fonofos yields the S-enantiomer of fonofos-oxon and (S)-fonofos yields (R)-fonofosoxon. In the case of fonofos this change in designation is important as the activity and toxicity of the R-enantiomer is greater than that of the S-isomer, whereas the situation is reversed for fonofos-oxon, i.e. S>R. Without an appreciation of the structures of the individual enantiomers it would appear that the activity of the oxygen derivatives showed the reverse stereoselectivity to the sulphur series which is obviously not the case.

Drug chirality and its pharmacological consequences

139

4.2.2 The nomenclature problem in generic names A major problem in therapeutics is the lack of readily available information on the stereochemical identity or composition of a chiral drug in standard reference works. In the majority of cases it is impossible to determine if the material being used is a single enantiomer, a racemic mixture, a mixture of diastereoisomers or some other possibility. It is frequently the case that the (±)-prefix is used to indicate that the material is a racemic * The designation applied to structure (4.40) may appear to be incorrect, but in the Sequence Rule the participation of d-orbitals in bonding is neglected for assignment of designation, e.g. the bonds of sulphur atoms in sulphoxides are regarded as single.

mixture, but if the compound in question contains two chiral centres in its structure then four stereoisomeric forms are possible, i.e. two pairs of enantiomers and hence two racemic mixtures. Which of the two possible racemates is the drug or is it a mixture of all four stereoisomers? The use of the (±)-prefix in this case does not specify the composition of the material. There is therefore a need within drug nomenclature to provide a system of generic names which will indicate if a compound may exist in more than one stereoisomeric form and also the nature of the material used, i.e. single isomer or mixture. An examination of the current British National Formulary (No. 31, 1996) indicates a number of agents listed under the headings Lev or Levo, Dex or Dextro, e.g. levamisole, levodopa, levomepromazine, dexamethasone, dexamphetamine, dexfenfluramine, dextromethorphan, indicating that the material is a single stereoisomer. However, for the remaining agents there are no indications of the stereochemical nature of the material. The extention of the above approach to nomenclature to include prefixes such as rac, for racemic mixtures, diam, for mixtures of diastereoisomers and mep, for mixtures of epimers has been proposed. 4.2.3 Prochirality Atoms which are bonded to two identical groups and to two other different groups are said to be prochiral. For example if either of the two methylene group hydrogen atoms in ethanol (4.42) were replaced by another group, e.g. deuterium, then the carbon atom (C1) becomes chiral and two enantiomeric forms are possible (4.43). If ethanol (4.42) is viewed from the side opposite the hydrogen atom indicated ** then the sequence of groups about C1 i.e. HO, CH3, H, is anticlockwise. If the molecule is viewed from the side opposite the hydrogen indicated * then the sequence of groups is reversed, i.e. clockwise. In terms of their molecular environments these two hydrogen atoms are not equivalent, the carbon atom C1 is prochiral and the two hydrogen atoms are said to be enantiotopic. If H** is arbitrarily preferred over H* then an R-designation is obtained and H** is designated pro-R and H* as pro-S (4.44). Differentiation of enantiotopic groups may be of considerable significance in biochemistry and metabolism (see Section 4.3).

Introduction to the principles of drug design and action

140

4.3 BIOLOGICAL ACTIVITY That enantiomers can exhibit different biological activities has been appreciated for over a century. One of the first reported observations of the differential physiological actions

of stereoisomers was that of Piutti, who in 1886 isolated the enantiomers of the amino acid asparagine (4.45) and reported that the (+)-enantiomer tasted sweet and the (−)enantiomer was bland. Similar observations have been reported for other amino acids and the enantiomers of the D-series taste sweet, whereas those of the L-series are either tasteless or bitter. Enantiomers may also exhibit different odours the (−)enantiomer of carvone (4.46) smells of spearmint whereas (+)-carvone has an odour of caraway. The (+)-enantiomer of the related terpene limonene (4.47) smells of orange and the (−)-enantiomer of lemon.

Drug chirality and its pharmacological consequences

141

The differential pharmacological activity of drug enantiomers was shown in the early years of the present century when the English pharmacologist Cushny demonstrated that (−)-hyoscyamine was more potent than the (+)-enantiomer and that (−)-adrenaline had much greater activity than its (+)-antipode. In order to rationalise the observed differences in pharmacological activity between enantiomers Easson and Stedman, in 1933, suggested a three point fit model between the more active enantiomer and its receptor. The enantiomer on the left (4.48) is involved with three simultaneous bonding interactions with complementary sites on the receptor surface. Whereas that on the right (4.49) may only take part in two such interactions. Alternative orientations of the enantiomer on the right (4.49) to the receptor site are possible but only two interactions may take place at any one time. According to the Easson-Stedman model the more potent enantiomer is involved with a minimum of three intermolecular interactions with the receptor surface whereas the less potent isomer may interact at two sites only. Thus the “fit” of the enantiomers to the receptor are different as are their binding energies.

Introduction to the principles of drug design and action

142

The Easson-Stedman model is supported by data derived from an examination of the activity of (−)-(R)-adrenaline (4.31). The three points of attachment of (R)-adrenaline being the secondary amino group, the catechol ring system and the alcohol hydroxyl group. Comparison of the activity of (R)- and (S)-adrenaline and that of the achiral desoxy compound N-methyldopamine indicates that the activity of the S-enantiomer and the achiral compound are the same. In the case of (S)-adrenaline the hydroxy group is orientated in an unfavourable position for three simultaneous interactions with the receptor and only a two point interaction is possible. Similarly, Nmethyldopamine may also interact at two points, with the result that the activity is similar to that of (S)-adrenaline and much less than that of the R-enantiomer. Similar data has been obtained for the corresponding enantiomers and achiral derivatives of (R)-noradrenaline (4.30) and (R)-isoprenaline for both α and β adrenoreceptor activity. In 1948 Ogston, unaware of the Easson-Stedman hypothesis, proposed a similar three point attachment model in order to rationalise the results from enzymatic studies using prochiral substrates. In the case of a compound CABBD (4.50) the two B groups are

Drug chirality and its pharmacological consequences

143

enantiotopic and may be differentiated on interaction with an enzyme active site such that only one of the groups undergoes transformation. Ogston proposed that the substrate interacts with three sites on the enzyme but that only one of the complimentary sites to the enantiotopic groups B is involved with the biochemical transformation. If reaction can only occur at site B” then group B* in the substrate, but not group B, is converted in the product, i.e. the groups B and B* are not sterically equivalent. Transformations of this type are relatively common in biochemistry and in drug metabolism. For example the synthesis of (−)-(R)-noradrenaline (4.30) from dopamine (4.51), mediated by dopamine-β-hydroxylase, proceeds with total stereoselectivity, i.e. is stereospecific.

Similar specificity is shown by this enzyme in the metabolism of other substrates, e.g. (+)-(S)-α-methyldopamine (4.52) to (−)-(1R,2S)-α-methylnoradrenaline (4.53). The antihypertensive agent α-methyldopa (4.54) is marketed as the single L-enantiomer corresponding to the S-configuration using the sequence rule designation. This agent undergoes decarboxylation, mediated by dopa decarboxylase, to yield (+)-(S)-αmethyldopamine (4.52), which then undergoes, dopamine β-hydroxylase mediated oxidation to (1R,2S)-α-methylnoradrenaline (4.53), the active agent. As (+)-(S)-αmethyldopamine, is chiral the two hydrogen atoms on the β-carbon atom are said to be diastereotopic rather than enantiotopic.

Introduction to the principles of drug design and action

144

The above models are very useful but relatively simplistic representations of what may in fact occur during the drug, or substrate, interaction with a receptor, or enzyme,

and assumes that the drug has to adopt a particular orientation in relation to the receptor site. It is possible that the less active, or potent, enantiomer may also be involved in three intermolecular interactions with the receptor resulting from additional interactions with the biomolecule which do not occur with the more active enantiomer. It is also feasible to propose that the interactions do not necessarily need to be attractive, e.g. the interactions could be both attractive and repulsive. In addition the interaction between the drug and the receptor/enzyme target may result in conformational changes in both the target macromolecule and the ligand. Thus, the final interaction model may be fairly complex and both the stereochemistry and conformational flexibility of the ligand need to be taken into account. 4.3.1 Terminology used in the pharmacological evaluation of stereoisomers The differential biological activity of a pair of stereoisomers has given rise to additional terminology. Thus, the stereoisomer with the higher receptor affinity, or activity, is termed the Eutomer and that with the lower affinity, or activity, the Distomer. The ratio of affinities, or activities, of the two stereoisomers, a measure of the stereoselectivity is known as the Eudismic Ratio and its logarithm as the Eudismic Index (EI):

Drug chirality and its pharmacological consequences

145

where KEu and KDis are equilibrium dissociation constants for the eutomer and distomer respectively. A plot of Eudismic Index versus pKEu, for a homologous or congeneric series of isomeric pairs of compounds generally yields a straight line. The slope of which is positive and is known as the Eudismic Affinity Quotient (EAQ) which is a quantitative measure of the stereoselectivity within the compound series for a particular biological effect. It is important to appreciate that such terminology applies to a particular activity of a drug. For example in the case of a dual action drug the eutomer for one activity may be the distomer for another, or the enantiomers may be equal in activity. In the case of the β-blocking drug propranolol the eutomer for β-blocking activity is the enantiomer of the S-absolute configuration, which is between 40 to 100 fold more potent than its antipode, depending on the test system used. In contrast the enantiomers of propranolol have similar activities with respect to their membrane stabilising properties. There are also examples where both enantiomers of a drug are marketed with different therapeutic indications. In the case of propoxyphene the dextrorotatory enantiomer of the 1S, 2R-configuration is available as dextropropoxyphene (4.55) an analgesic agent and levopropoxyphene (4.56), with the 1R, 2S-configuration, as an antitussive. In the case of this example not only are the molecules mirror image related but so are their trade names DARVON® (dextropropoxyphene) and NOVRAD® (levopropoxyphene).

4.3.2 “Purity” of enantiomerically pure drugs The determination of the eudismic ratio of a pair of stereoisomers obviously depends on the availability of enantiomerically pure compounds. The reported eudismic ratios for the stereoisomers of a particular compound may vary widely within the literature. Whilst data of this type would be expected to vary from one laboratory to another an important contributory factor is probably associated with the enantiomeric purity of the materials examined particularly for the less active isomer. Early investigations on

Introduction to the principles of drug design and action

146

the activity of the enantiomers of chloroquine, for example, indicated that there were no differences in terms of toxicity or efficacy. Subsequent investigations have indicated that the “individual enantiomers” used in the initial study were little better than racemates. As the eudismic ratio increases then the significance of a small quantity of the eutomer as an impurity of the “inactive” distomer also increases. When isoprenaline was initially resolved the reported ratio of activities (R/S) was approximately 12. Further experimentation and improved resolution, in this case repeated fractional crystallisation, resulted in a 1000 fold difference in activity. The influence of relatively small quantities of stereoisomeric impurities on eudismic ratio may be illustrated by a recent report of the activity of the stereoisomers of formoterol a β2-selective agonist. Formoterol (4.57) has two chiral centres and therefore exists in four stereoisomeric forms, the two chiral centres being positioned α and β to the aliphatic nitrogen atom. An examination of the activity of the four isomers on the relaxation of airway smooth muscle, indicated a relative order of potency of αR,βR>αS,βR≈αS,βS> αR,βS. In an early report the eudismic ratio for the enantiomers αR,βR/αS,βS was determined to be 14. A more recent investigation reported the same relative order of isomeric potency but a eudismic ratio αR,βR/αS,βS of 50. In the later study the distomer, the αS,βS-enantiomer, was contaminated with 1.5% of the active αR,βR-isomer. Reduction in the “active impurity” to less than 0.1% resulted in an increase in eudismic ratio αR,βR/αS,βS to 850 and similar reductions of the “impurity” in the αS,βR-and αR,βS-stereoisomers resulted in an altered order of relative potency to αR,βR> αS,βR≈αR,βS>αS,βS. Further increases in the “purity” of the inactive isomer may result in an increase in eudismic ratio. The degree of enantiomeric purity is frequently not specified in the pharmacological literature, or alternatively, is presented in terms of optical rotation which is not a very sensitive technique at levels of contamination of a few percent. In such cases analytical methodology with an increased sensitivity for enantiomeric analysis is more appropriate, e.g. chromatographic methods using chiral stationary phases. The limitations of optical rotation determinations may be illustrated by a consideration of the BP 1993 monograph on naproxen. Naproxen (4.58) is a nonsteroidal anti-inflammatory drug marketed as the single dextrorotatory S-enantiomer. The BP requires the optical rotation of the material, determined in chloroform, to be between +63.0 and +68.5° which based upon the

Drug chirality and its pharmacological consequences

147

published specific rotation corresponds to a stereochemical purity of between 95.5 to 103.7%. In comparison the analytical limits for purity determination by volumetric analysis are 98.5–100.5%. Thus, the chemical purity limits are more stringent than those for the stereochemical purity. 4.3.3 Receptor selectivity As pointed out above eudismic ratios are only of significance for a particular biological activity of a drug. For a drug which can act at two or more sites differences in eudismic ratio provides useful information in terms of the stereochemical demands and geometry of the site, a means of comparison between receptors in different tissues and may also be used as a method of distinguishing receptor subtypes. Obviously such comparisons must be made with caution to ensure that potentially misleading factors, e.g. diffusion barriers, tissue uptake and metabolism, are taken into account or controlled as such factors may vary markedly between tissues. The activity of the enantiomers of the neuroleptic agent butaclamol have been investigated with tissue preparations containing D2-dopaminergic, α-adrenergic, 5-HT2 and 5-HT1 serotoninergic, and opioid receptors. The eudismic ratio, (+)/(−), varied markedly with receptor system, (+)-(3S,4aS,13bS)-butaclamol (4.59) being 1250 times more active than the (−)-enantiomer in displacing haloperidol at D2-receptors, 143 times more active at displacing LSD at 5-HT2-receptors and equally active at displacing nalorphine at opioid receptors. The greater eudismic ratio was observed for actions in which the compound showed the greatest potency (see Section 4.3.4). Comparison of the stereochemical discrimination of the enantiomers of noradrenaline by the α1 and α2-adrenergic receptors indicates basic differences

Introduction to the principles of drug design and action

148

between the two receptor subtypes. The eudismic ratios (R/S) obtained being 107 and 480 fold for α1 and α2 receptors respectively. Similar differences are also observed for α-methylnoradrenaline, the eudismic ratios for the 1R,2S/1S,2R enantiomeric pair being α1, 60 and α2, 550. Thus, for phenylethylamine derivatives the steric demands of the α2-receptor are more stringent than those of the α1-receptor subtype.

Differential stereoselectivity has also been observed with agonists at the histamine receptor subtypes. The introduction of a methyl group a to the amino group in histamine results in the chiral molecule α-methylhistamine (4.60). Examination of the activity of the enantiomers of α-methylhistamine at the three histamine receptor subtypes yields eudismic ratios (R/S) of 1, 0.6 and approximately 100 at H1, H2 and H3 receptors respectively. The H1-receptor showing no stereoselectivity, the H2-receptor limited selectivity for the S-enantiomer and the H3-receptor showing marked stereoselectivity. Examination of pD2 values for the R-enantiomer at the three receptor subtypes yields values of 4.54, H1; 3.96, H2 and 8.40 at H3. Thus, (R)-αmethylhistamine is a highly selective H3 agonist and stimulation at H3-receptors would be expected to occur at concentrations 104 times lower than those required for H1 or H2 stimulation. Similar stereoselectivity for the H3-receptor is also observed for α,βdimemylhistamine. In this case the αR,βS-enantiomer (4.61) is 100 fold more active than its αS,βR-antipode and shows 130,000 fold greater selectivity for the H3 receptor than the other two subtypes. Compound (4.61) is the most active chiral agonist known at H3-receptors and shows the greatest receptor subtype selectivity. These two examples illustrate that the introduction of chirality into a critical site in a molecule may result in significant receptor subtype selectivity. 4.3.4 Quantitative structure—activity relationships The significance of stereochemistry with respect to quantitative structure—activity relationships (QSAR) is dependent on the site of the chiral centre within the molecule. Is

Drug chirality and its pharmacological consequences

149

the chiral centre located in a position which will influence the interaction of the drug with the target receptor? A number of situations are possible: (a) The chiral centre is located in a critical position within the molecule such that alteration of the stereochemistry, or structural modification to an achiral analogue results in a marked reduction in activity, e.g. the situation with (R)adrenaline (4.31) referred to previously (p. 114). (b) The chiral centre is located in a critical position within the molecule but the eutomer has enhanced, or the same activity, as an achiral analogue, the distomer being reduced in activity compared to the achiral compound. For example examination of the activity of the acetylcholine analogue (S)-βmethacholine (4.62) on isolated rat intestine yields a pD2 value of 6.8, compared to the value of 7.0 obtained with acetylcholine, whereas, the Renantiomer, the distomer, yields a value of 4.1. In this case it appears that a two point interaction only is required for activity but that the orientation of the methyl group at the chiral centre is critical for activity. In the S-enantiomer, the eutomer, the methyl group is presumably orientated in a non-critical binding region of the receptor, whereas in the R-enantiomer the orientation results in steric repulsion.

Introduction to the principles of drug design and action

150

(c) The chiral centre is in a non-critical position in the molecule such that both enantiomers and the achiral analogue have the same, or similar, activities. Examination of the properties of the H1-antihistamine terfenadine (4.63) in either pharmacological or biochemical assay systems, indicates no difference in activity between the enantiomers. Replacement of the hydroxy group at the chiral carbon atom by hydrogen yields an achiral derivative which has similar activity as the enantiomers of terfenadine. Thus the hydroxyl group is located in a non-critical position for receptor binding.

If the chiral centre is located in a critical region of the molecule then differences in activity between isomers are expected and such differences would be greater for stereoisomers than for homologues, or analogues resulting from relatively simple isosteric replacements. To derive useful data from QSAR studies of chiral compounds each series of stereoisomers should be examined independently. A useful approach for QSAR studies of stereoisomers in a related compound series is Eudismic Analysis, the eudismic index is plotted against the affinity, or potency, of the eutomer and the eudismic affinity quotient, the slope of the line, gives an indication of the stereoselectivity for a particular biological effect (Section 4.3.1). As a general rule the eudismic index is a function of the affinity of the eutomer, the higher the affinity of the drug the greater the degree of complementarity between the drug and its receptor site. Whereas for low affinity compounds the complementarity between the drug and receptor site will be lower and hence the eudismic index will be reduced. For drugs such as terfenadine, i.e. those in which the chirality is not critical for activity, a similarly low ratio would be expected.

Drug chirality and its pharmacological consequences

151

The above relationship, the greater the affinity of the eutomer the greater the eudismic ratio, appears to be common for many series of drugs and is known as Pfeiffer’s rule. Examples of compounds are known which do not follow this generalisation. In these cases: the chiral centre may be in a non-critical site in the molecule; two of the four groups attached to the chiral centre are bioisosteric and therefore, in biological terms at least, are not distinguished; the increased affinity of the distomers is due to additional interactions with the biomolecule which do not occur with the eutomer. 4.4 PHARMACOKINETIC CONSIDERATIONS As many of the processes of drug absorption and disposition involve an interaction between the enantiomers of a drug and a chiral biological macromolecule it is hardly surprising that stereoselectivity is observed during these processes. 4.4.1 Absorption The most important mechanism of drug absorption is passive diffusion through biological membranes a process which is dependent upon the physico-chemical properties of the molecule, e.g. lipid solubility, pKa, molecular size etc. If a chiral drug is absorbed by a passive process then differences between enantiomers would not be expected. However, differences between enantiomers may occur if the drug is a substrate for an active transport or carrier-mediated transport system. Such processes require the reversible combination of a substrate with a biological macromolecule and involve movement against a concentration gradient, expenditure of metabolic energy and may be saturated. Such systems show substrate specificity and hence would be expected to show stereoselectivity. Stereospecific transport systems are known to exist in the gastrointestinal tract for L-amino acids, dipeptides and D-carbohydrates etc. and drugs which are similar in structure to such naturally occurring substrates may be expected to be actively transported. Thus, L-dopa (4.64), L-penicillamine (4.65) and L-methotrexate (4.66) have been shown to be preferentially absorbed from the gastrointestinal tract compared to their D-antipodes which are not substrates and are absorbed by passive

Introduction to the principles of drug design and action

152

diffusion. Such active processes may be expected, in theory at least, to increase the rate rather than the extent of absorption. In fact the bioavailability of D-methotrexate is only 2.5% that of the L-isomer. Many of the β-lactam antibiotics are substrates for the gut dipeptide transport system and as such their absorption would be expected to be stereoselective. The influence of the stereochemistry of the 7-acyl side chain on the absorption of the diastereoisomers of cephalexin (4.36) has been investigated in the rat. Both diastereoisomers are substrates for the carrier mediated transport system with the Lepimer showing a higher affinity than, and acting as a competitive inhibitor for Dcephalexin transport. However, the L-epimer is also more susceptible to the intestinal wall peptidases and cannot be detected in serum, whereas the D-isomer is well absorbed. The drug is marketed as the single D-epimer. Additional biochemical or pharmacological factors may also influence the stereoselectivity of drug absorption. For example the greater oral bioavailability of (−)-(R)-terbutaline (4.67) compared to the less active (+)-S-enantiomer arises as a

Drug chirality and its pharmacological consequences

153

result of stereoselectivity in first pass metabolism and possibly due to the (−)enantiomer increasing membrane permeability.

Differences in absorption may also be observed if the individual enantiomers differ in their effects on local blood flow. For example (−)-bupivacaine (4.68) has a longer duration of action than (+)-bupivacaine following intradermal injection. This difference in activity is due to the vasoconstrictor effects of the (−)-enantiomer reducing blood flow and hence systemic absorption. In vitro both enantiomers have similar potencies. 4.4.2 Distribution Protein binding The majority of drugs undergo reversible binding to plasma proteins. In the case of chiral drugs the products of such binding are diastereoisomeric complexes and individual enantiomers would be expected to show differences in binding affinity. Such differences in binding affinity result in differences between enantiomers in the free, or unbound, fraction which is able to distribute into tissue (Table 4.1). The two most important plasma proteins with respect to drug binding are human serum albumin (HSA) and α1-acid glycoprotein (AGP). In general acidic drugs bind predominantly to HSA, whereas basic drugs bind predominantly to AGP. Differences between enantiomers in plasma protein binding are relatively small (Table 4.1) and in some cases less than 1%. However, such low stereoselectivity in binding may result in much larger differences in the enantiomeric composition of the unbound fraction particularly for highly protein bound drugs, e.g. in the case of indacrinone the free fractions are 0.9% and 0.3% for the (−)-R and (+)-S-enantiomers respectively, i.e. a three fold difference in the free fraction. An extreme example of stereoselectivity in binding is the amino acid tryptophan, the L-enantiomer binding to HSA with an affinity approximately 100 times greater than that of the D-isomer. In terms of drugs, (S)-oxazepam hemisuccinate (4.69) binds to HSA with an affinity 40 times that of the R-enantiomer. However, using bovine serum albumin as a protein source the difference in affinity is only three fold. Such species variation in enantioselectivity in plasma protein binding has also been reported for phenprocoumon and disopyramide.

Introduction to the principles of drug design and action

154

Enantioselectivity in binding may also vary between HSA and AGP. For example in the case of propranolol the binding to AGP is stereoselective for the S-enantiomer, whereas binding to HSA is selective for (R)-propranolol. In whole plasma the binding to AGP predominates and the fraction unbound of the R-enantiomer is greater than that of (S)-propranolol (Table 4.2).

Table 4.1 Stereoselectivity in plasma protein binding. % Unbound Acidic drugs Acenocoumarol (S) 2.0 (R) 1.8 Ibuprofen (S) 0.64 (R) 0.42 Indacrinone (S) 0.3 (R) 0.9 Moxalactam (S) 32 (R) 47 Phenprocoumon (S) 0.72 (R) 1.07 Warfarin (S) 0.9 (R) 1.2 Mephobarbitone (S) 53 (R) 66 Pentobarbitone (S) 26.5 (R) 36.6 Flurbiprofen (S) 0.048 (R) 0.082 Basic drugs Chloroquine (S) 33.4 (R) 51.5 Disopyramide (S) 22.2 (R) 34 Fenfluramine (S) 2.8 (R) 2.9 Methadone (−) 12.4 (+) 9.2 Mexiletine (S) 28.3 (R) 19.8 Tocainide (−) 86–91 (+) 83–89 Verapamil (S) 11 (R) 6.4

Ratio 1.1 1.5 0.33 0.68 0.67 0.75 0.80 0.72 0.59 0.64 0.64 0.96 1.3 1.4 1.0 1.7

Drug chirality and its pharmacological consequences

155

Table 4.2 Stereoselectivity of the plasma protein binding of propranolol enantiomers. Protein source Enantiomer free fraction Ratio R/S R S Whole plasma 0.203 0.176 1.15 HSA 0.607 0.647 0.94 AGP 0.162 0.127 1.28 Stereoselectivity in plasma protein binding may also influence drug clearance for compounds with a low extraction ratio as total clearance is proportional to fraction unbound. In addition stereoselective displacement of drug enantiomers from plasma protein binding sites may give rise to complexities in drug interactions (see Section 4.6.3). Interactions between enantiomers for plasma protein binding sites may also result in pharmacokinetic complications. For example the protein binding of disopyramide is stereoselective and concentration dependent and the pharmacokinetic parameters of the individual enantiomers differ depending if the drug is administered as the racemate or single isomer. Tissue distribution The extent of tissue distribution of a drug depends on both its lipid solubility and relative tissue-plasma protein binding; for example the apparent stereoselective distribution of (S)-ibuprofen into synovial fluid may be explained by differences in protein binding. Relatively few examples of stereoselectivity in tissue binding are known, however this may occur by selectivity in tissue uptake and storage mechanisms. For example there is evidence that the active S-enantiomers of the βblocking drugs propranolol and atenolol undergo selective storage and secretion by adrenergic nerve terminals in cardiac and other tissue. The selective incorporation of the R-enantiomers of some of the 2-arylpropionic acid non-steroidal antiinflammatory agents into lipid has also been observed. The selective distribution of these agents is associated with their metabolism and the formation of “hybrid” triglycerides, the mechanism of which will be discussed below (Section 4.7.5). This selective deposition results in the accumulation of these agents into lipid the toxicological significance of which is unknown. 4.4.3 Metabolism In contrast to other processes involved in drug absorption and disposition, drug metabolism frequently shows marked stereoselectivity. The stereoselective step in metabolism may involve a number of different stages in the enzymic reaction sequence. Thus, the binding of the substrate to the enzyme may be stereoselective and associated with the chirality of the binding site. Selectivity may also be associated with catalysis due to the differential reactivity and orientation of potential target groups with respect to the catalytic site.

Introduction to the principles of drug design and action

156

An examination of the stereochemistry of drug metabolism is of importance as the individual enantiomers of a racemic drug may be metabolised by different routes to yield different products and they are frequently metabolised at different rates. In addition species differences may occur in the metabolism of individual enantiomers and as data derived from animal studies is used to assess potential toxic hazard to man the information may have little relevance. The stereoselectivity of the reactions of drug metabolism may be examined on the basis of: 1. substrate stereoselectivity, i.e. the selective metabolism of one enantiomer over that of the other; 2. product stereoselectivity, i.e. the selective formation of one particular stereoisomer rather than other possible stereoisomers; 3. substrate-product stereoselectivity, i.e. the selective metabolism of one of a pair of enantiomers to produce one of a number of possible diastereoisomeric products. In terms of the stereochemical outcome of metabolic transformations reactions may be divided into five groups as indicated below. 1. Prochiral to chiral transformations In the case of reactions of this type the molecule acquires chirality by metabolism which may take place at a prochiral centre or at a site remote from it. The antiepileptic drug phenytoin (4.70) has a prochiral centre at carbon-5 of the hydantoin ring system and the two phenyl rings are enantiotopic being pro-S and pro-R as indicated (4.70). The major route of metabolism of phenytoin in both animals and man involves aromatic oxidation which in man shows product stereoselectivity for formation of (S)4-hydroxyphenytoin (4.71). In contrast, in the dog oxidation takes place in the pro-R ring to yield (R)-3-hydroxyphenytoin the reaction showing species selectivity in both stereochemistry and regiochemistry (position). It has been pointed out above that sulphoxides may be chiral and therefore the metabolic oxidation of sulphides to sulphoxides will produce chiral metabolites. Cimetidine (4.72) undergoes oxidation at sulphur to yield an optically active sulphoxide (4.73) as a major urinary metabolite. The reaction is product stereoselective for the formation of the (+)-enantiomer, the enantiomeric composition of the material in urine being (+/−) 3:1. Such metabolic transformations may also differ in their stereochemistry depending upon the enzyme system effecting the reaction. For example the model substrate 4tolylethylsulphide (4.74) undergoes oxidation to yield a sulphoxide (4.75) but the reaction produces predominantly the R-enantiomer (>95%) when mediated by the flavin-containing monooxygenase (FMO) and predominantly the S-enantiomer when mediated by the cytochrome P450 system.

Drug chirality and its pharmacological consequences

2. Chiral to chiral transformations

157

Introduction to the principles of drug design and action

158

In transformations of this type metabolism takes place at a site in the molecule which does not alter the chirality of the metabolite relative to that of the drug. Esmolol (4.76) is an ultra short acting, relatively cardioselective β-blocker, which is administered intravenously for the short term treatment of supraventricular arrhythmias and sinus tachycardia. The drug is used as a racemate but the pharmacological activity resides in the enantiomer of the S-configuration, the Risomer being inactive as a β-blocker. The basis of the short duration of action, 10–15 min, is the rapid hydrolysis of the ester functionality by blood esterases. The hydrolysis of this agent shows considerable species variability, e.g. the hydrolysis of the S-enantiomer is faster than that of the R-isomer in the rhesus monkey, rabbit and guinea-pig and shows the reversed stereoselectivity in rat and dog. In man the hydrolysis of both enantiomers occur at similar rates.

Aromatic oxidation of warfarin (4.77) yields the 7-hydroxy metabolite, a reaction which is highly stereoselective for the more active S-enantiomer (ratio S:R:6:1) of the drug. In contrast oxidation at the 6 position of the coumarin ring system shows no stereoselectivity in man. In the rat 7-hydroxywarfarin is the major metabolite but for the R-enantiomer, i.e. the oxidation shows the reverse stereoselectivity compared to man. Recent studies using human DNA expressed cytochrome P450 isoenzymes have indicated that the isoform 2C9 is primarily responsible for the oxidation of (S)warfarin to the 6- and 7-hydroxy compounds whereas isoform 1A2 is involved in the formation of (R)-6-hydroxywarfarin. Warfarin also undergoes oxidation to yield the 4'- and 8-hydroxy derivatives each of which reactions shows a degree of stereoselectivity and it has been proposed that as a result of the regio (positional) and stereoselectivity of oxidation that warfarin could be used as a probe compound for the determination of the isoenzyme composition of hepatic cytochromes P450. 3. Chiral to diastereoisomer transformations Transformations of this type involve the introduction of a second chiral centre into a chiral molecule. Such centres may arise by a Phase I metabolic transformation at a prochiral centre or by a Phase II metabolic transformation by reaction with a chiral conjugating agent. Reactions of the first type include reduction of the prochiral ketone group in warfarin (4.77) to yield a pair of diastereoisomeric warfarin alcohols. In both rat and man the reduction is substrate selective for (R)-warfarin (4.77) and the predominantly formed isomer of the alcohol (4.78) has the S-configuration at the new centre. The

Drug chirality and its pharmacological consequences

159

Phase II or conjugation reactions of drug metabolism are synthetic and involve the combination of the drug, or a Phase I metabolite of the drug, with an endogenous molecule (see

Chapter 1). Many of the endogenous molecules involved in the conjugation reactions are chiral, e.g. D-glucuronic acid, the amino acid glutamine and the tripeptide glutathione, and hence chiral drugs which undergo conjugation with these agents will produce diastereoisomeric products. Oxazepam (4.79) is a chiral benzodiazepine which is used as a racemic mixture. The individual enantiomers of oxazepam are stereochemically unstable and readily undergo racemisation in aqueous media and in contact with glass surfaces. Enantiomeric resolution is only possible when carried out under anhydrous conditions. Both enantiomers of oxazepam undergo conjugation with D-glucuronic acid to yield a pair of stereochemically stable diastereoisomeric conjugates the proportions of which vary between species. In man, dog and rabbit the diastereoisomer produced from (S)oxazepam (4.80) predominates, S/R ratios varying between 2 to 3.4, whereas in the rhesus monkey (R)-oxazepam glucuronide (4.80) is preferentially formed (ratio S/R=0.5). The formation of the stereochemically stable glucuronides, and their direct analysis by high-performance liquid chromatography has facilitated the examination of the stereochemical aspects of disposition of the drug. It is of interest to note that hydrolysis of either diastereochemically pure conjugate results in the formation of the racemic drug. Conjugation with the tripeptide glutathione (GSH; L-glutamyl-L-cysteinylglycine) involves reaction of the nucleophilic sulphur atom of the cysteine residue with

Introduction to the principles of drug design and action

160

electrophilic sites in foreign compounds. The reaction is mediated by the glutathione transferases, a family of isoenzymes with overlapping substrate specificity, found in the cytosolic and microsomal fractions of cells. The mechanism of conjugation with GSH appears to be

a single displacement substitution consistent with an SN2 type reaction and the substrate undergoes Walden inversion. As GSH contains two optically active amino acids in its structure if reaction occurs with a racemic substrate the glutathione conjugates are diastereoisomers. The conjugation of the obsolete chiral hypnotic agent bromoisovalerylurea (4.81) with GSH involves nucleophilic displacement of the bromine atom at the chiral centre and the glutathione conjugates (4.82) have the reverse configurational designation to those in the drug. In the case of αbromoisovalerylurea the reaction is stereoselective for the R-enantiomer of the drug, the cytosolic enzyme(s) showing a three fold greater activity for the R compared to the S-enantiomer. The stereoselectivity of the reaction does vary with isoenzyme such that examination of purified enzyme systems indicates that the isoenzymes of the mufamily show a stereopreference for conjugation of (R)-α-bromoisovalerylurea, whereas those of the alpha-family show a preference for the S-enantiomer. 4. Chiral to achiral transformations

Drug chirality and its pharmacological consequences

161

Transformations in which the chirality of a molecule is lost are relatively unusual. The best known examples involve the oxidation of secondary alcohols to yield the corresponding ketones but the investigation of such reactions is frequently complicated by the stereochemistry of the reverse reaction of reduction. The deamination of amphetamine (4.83) to yield the achiral phenylacetone (4.84) appears to be stereoselective for the R-enantiomer of amphetamine.

Introduction to the principles of drug design and action

162

More recent examples of interest are provided by the 1,4-dihydropyridine calcium channel blocking agents, e.g. nilvadipine (4.85). These agents undergo P450 mediated oxidation to yield the corresponding achiral pyridine analogues (4.86). In the case of nilvadipine this reaction is stereoselective for the (+)-enantiomer in the rat, but for the (−)-enantiomer in dog and man. 5. Chiral inversion Metabolic chiral inversion is a relatively rare transformation and involves the conversion of one enantiomer of a drug to its optical antipode with no other chemical change to the molecule. The reaction was initially observed with the 2-arylpropionic acid NSAIDs, e.g. ibuprofen (4.21) and has since been found to occur with the chemically related 2-aryloxypropionates, which are used as herbicides, e.g. haloxyfop (4.87). In the case of the 2-arylpropionic acids the reaction involves inversion of the relatively inactive

R-enantiomers to their active S-antipodes (4.20), whereas in the case of the 2aryloxypropionates the reaction appears to be the reverse, i.e. the S-enantiomers are converted to their R-antipodes (4.37). This difference in the stereochemistry of the inversion reaction is apparent and arises as a result of the sequence rule designation, the three-dimensional spatial arrangement of the R-2-arylpropionic acids corresponding to that of an S-2-aryloxypropionate. The mechanism of this reaction will be examined in Section 4.7.5.

Drug chirality and its pharmacological consequences

163

4.4.4 Excretion Renal excretion is the net result of glomerular filtration, active secretion and passive and active reabsorption (see Chapter 1). Since glomerular filtration is a passive process differences between enantiomers would not be expected, stereoselective renal clearance may be observed as a result of active secretion, however active reabsorption and renal metabolism may also be significant. Apparent stereoselectivity in renal clearance may also arise as a consequence of stereoselectivity in protein binding rather than active transport. Active renal tubular secretion is thought to be responsible for the differential clearance of the enantiomers of a number of basic drugs with stereoselectivities in the range of 1.0 to 1.8 (Table 4.3). The renal clearance of quinidine has been reported to be four times greater than that of its diastereoisomer quinine. The renal clearance of diastereoisomeric glucuronide conjugates of both ketoprofen and propranolol have also been reported to show stereoselectivity. In both cases renal clearance is selective for the S-enantiomer conjugate of the drug with selectivities of 3.2 and 1.3 fold for propranolol and ketoprofen respectively. Relatively little is known regarding the stereoselectivity of the active processes involved in the biliary secretion of drugs. Differences in the biliary recovery of enantiomers has been reported, e.g. acenocoumarol in the rat, however it is not clear if this is due to stereoselectivity in biliary clearance or as a result of other stereoselective processes.

Table 4.3 Stereoselectivity in renal clearance of basic drugs in man. Drug Stereochemistry Ratio Terbutaline S>R 1.8

Introduction to the principles of drug design and action

Disopyramide Chloroquine Pindolol Metoprolol

(+)>(−) (+)>(−) R>S S>R

164

1.3 1.6 1.2 1.1

4.4.5 Pharmacokinetic Parameters As a result of the above processes compounds administered as racemates rarely exist as 1:1 mixtures of enantiomers in biofluids and tissues, and do not reach their sites of action in equal concentration. The pharmacokinetic profiles of the enantiomers of a racemic drug may differ markedly and hence an estimation of pharmacokinetic parameters, or an examination of drug concentration-effect relationships based on “total” drug, i.e. the sum of the two enantiomer concentrations, present in biological samples may at best yield data of limited value and is potentially highly misleading. The magnitude of the differences between enantiomers in their pharmacokinetic parameters tend to be relatively modest, frequently 1 to 3 fold, compared to those observed in their pharmacodynamic properties. The differences may however be attenuated depending upon the organisational level that the particular parameter characterises, i.e. the whole body (e.g. systemic clearance, volume of distribution, elimination half-life), whole organ (e.g. hepatic clearance, renal clearance) and macromolecular (e.g. intrinsic metabolite formation clearance, fraction unbound). Thus differences in parameters which reflect the whole body level of organisation may be modest, being composed of potentially multiple organ selectives which intern reflect the selectivity of multiple macromolecular interactions. Differences between enantiomers are potentially greatest in these latter parameters which are associated with a direct interaction with a chiral macromolecule. The stereoselectivity of such multiple processes may vary between enzymes, proteins and organs and it is therefore possible that a comparison of parameters that reflect the whole body level of organisation may mask stereoselectivity at an organ or macromolecular level. For example the ratio (R/S) of the plasma half-lifes, systemic clearance and volume of distribution of the enantiomers of propranolol (Section 4.6.2) are 1.01, 1.17 and 1.18 respectively. However, the plasma protein binding of propranolol shows a preference for the S-enantiomer (Table 4.2) whereas the metabolic clearance via 4-hydroxylation is greater for the R-enantiomer (Table 4.4). For drugs which are subject to extensive stereoselective first-pass, or presystemic metabolism, the differential bioavailability of the individual enantiomers may give rise to apparent anomalies in drug-concentration effect relationships with route of administration if the enantiomeric composition of material in plasma is not taken into account. Thus, based on measurements of “total” plasma concentrations verapamil appears to be more effective when given intravenously than orally, whereas propranolol shows the opposite effect. In both cases the likely explanation for the observed effects is stereoselective presystemic metabolism which in the case of verapamil is selective for the more active S-enantiomer and for propranolol the less active R-enantiomer (Sections 4.6.1 and 4.6.2).

Drug chirality and its pharmacological consequences

165

Care is also required in therapeutic drug monitoring of chiral drugs administered as racemic mixtures. The determination of the plasma concentrations of the individual enantiomers of chiral drugs would be advantageous to define the “real” therapeutic range of such compounds. The therapeutic range of “total” tocainide covers a three fold concentration range whereas the enantiomeric composition of the drug in plasma varies up to two fold. Stereochemical considerations may also be of significance for understanding drug interactions both between chiral drugs and a second agent (Section 4.6.3) and also to rationalize differences in the disposition of chiral drugs when given as racemic mixtures or single isomers (Section 4.6.1). 4.5 PHARMACODYNAMIC CONSIDERATIONS As pointed out previously the most important differences between enantiomers occur at the level of receptor interactions, and eudismic ratios are frequently of the order of 100 to 1000 fold. However, it is frequently the case that the “inactive” or less active isomer may contribute to the observed activity of a racemic mixture and a number of possible situations may arise as indicated below. 4.5.1 The pharmacological activity resides in one enantiomer the other being biologically inert There are relatively few examples of drugs which possess one or two chiral centres as part of their structure in which the pharmacological activity is restricted to a single enantiomer the other being totally devoid of activity. In the case of α-methyldopa the antihypertensive activity resides exclusively in the S-enantiomer and this agent is marketed as a single isomer. There are a number of examples, e.g. the β-blockers, where the activity is of the order of one to two orders of magnitude greater but in other actions of these agents stereoselectivity is not observed (Section 4.6.2). For compounds with more than two chiral centres it is frequently found that the configurations of all such centres are fixed requirements or activity/specificity in action is lost, e.g. steroids, ACE inhibitors, e.g. enalapril which has the SSSconfiguration, the SSR isomer being 10−4 fold less active. 4.5.2 Both enantiomers have similar activities Both enantiomers of the antihistamine promethazine (4.88) have similar pharmacological and toxicological properties, and the introduction of the chiral centre in the dimethylaminoethyl side chain results in a 100% increase in antihistaminic potency compared to the non-chiral analogue. In contrast the enantiomers of the 1-aza substituted derivative, isothipendyl (4.89), have similar activities in vitro but in vivo the (−)-enantiomer is ca half as potent as the (+)-isomer and both are less potent than the racemate. The reason for this observation is by no means clear but may be due to differences in drug disposition, e.g. inhibition of metabolism of the (+)-enantiomer by its antipode.

Introduction to the principles of drug design and action

166

Similarly, the enantiomers of flecainide (4.90) are equipotent with respect to antiarrhythmic activity, effect on cardiac sodium channels and show no significant

differences with respect to their pharmacokinetic properties. In the case of flecainide little information is available with respect to the toxicity of the individual isomers but the use of a single isomer would appear not to offer a therapeutic advantage. 4.5.3 Both enantiomers are marketed with different indications The example of dextropropoxyphene (4.55) and levopropoxyphene (4.56) being marketed as analgesic and antitussive agents has been cited previously. Similar differences in activity are found with related opiate derivatives, e.g. dextromethorphan, (+)-3-methoxy-N-methylmorphinan, is a useful antitussive agent which is virtually free from analgesic, sedative or other morphine like effects. Whereas the enantiomer, levomethorphan is a potent opioid with antitussive activity and is addictive. 4.5.4 The enantiomers have opposite effects

Drug chirality and its pharmacological consequences

167

Picenadol (4.91) is a phenylpiperidine analgesic that has both opioid agonist and antagonist activity. The analgesic activity resides entirely in the (+)-3S,4R-enantiomer and the (−)-3R,4S-enantiomer is an antagonist. The racemate exhibits the properties of a partial agonist due to the more potent activity of the (+)-isomer at the µ opioid receptor and the weak antagonist action of (−)-picenadol at the same receptor. The pharmacological activities of the enantiomers of several derivatives of aporphine have been examined and in each case the S-enantiomers appear to be antagonists of their

R-antipodes. (R)-11-Hydroxy-10-methylaporphine (4.92) is a highly selective 5-HT1A agonist, whereas its S-antipode (4.92) is an antagonist at the same receptor. Similarly, (S)-apomorphine (4.93) acts as an antagonist at dopaminergic receptors (D1 and D2) whereas the R-enantiomer is an agonist. In the case of 11-hydroxyaporphine the Renantiomer activates dopamine receptors and the S-enantiomer is an antagonist.

Introduction to the principles of drug design and action

168

A similar though more complex situation arises with 3-(3-hydroxyphenyl)-Npropylpiperidine (3-PPP; 4.94). The initial pharmacological evaluation of this compound was carried out using the racemate and 3-PPP was described as a highly selective presynaptic dopaminergic agonist. Resolution and pharmacological evaluation of the individual enantiomers indicated that the situation was more complex. (R)-3-PPP acts as an agonist at both pre- and postsynaptic dopamine receptors, whereas (S)-3-PPP stimulates presynaptic and blocks postsynaptic receptors. The pharmacological profile observed with the racemate arises from the sum of the activities of the individual enantiomers. (S)-3-PPP has been selected for further evaluation as it appears to influence dopaminergic function in two different ways, i.e. stimulation of the pre- and blockade of the postsynaptic receptors.

The 1,4-dihydropyridines are calcium-channel blockers used for the treatment of angina and hypertension. A number of these agents possess a chiral centre at the 4position of the dihydropyridine ring system and a number of examples are known in which the enantiomers have opposing actions on channel function, e.g. compounds (4.95), (4.96) and (4.97). The S-enantiomers act as potent activators, whereas the Renantiomers are antagonists at L-type voltage-dependent calcium channels. It was thought that the observed effects of the enantiomers of these agents were due to

Drug chirality and its pharmacological consequences

169

interactions at different binding sites. However, it appears that the enantiomers interact with different channel states, open and inactivated, the drug binding sites of which have opposite steric requirements. The situation is further complicated as the Senantiomers of (4.95) and (4.97) are activators at polarised membrane potentials but become antagonists under depolarising conditions. Indeed one author has described these agents as being “molecular chameleons”. 4.5.5 One enantiomer may antagonise the side effects of the other Indacrinone (4.98), a m-indanyloxyacetic acid derivative, is a loop diuretic with uricosuric activity, which has been evaluated for the treatment of hypertension and congestive heart failure. However, following administration of the racemate to man serum urate levels increase. Resolution and pharmacological evaluation of the individual enantiomers indicates that the diuretic and natriuretic activity reside in the (−)-R-enantiomer (4.98) and the uricosuric effects reside in (+)-(S)-indacrinone (4.98). Following administration of the racemate to man the plasma half-life of the Senantiomer ( compared to the R, ) and hence its uricosuric activity is too short to prevent the increase in serum uric acid. Alteration of the enantiomeric composition of the drug from the 1:1 ratio of the racemate by increasing the proportion of the (+)-S-enantiomer resulted in a mixture (S:R:4:1) which was isouricemic and a further increase (S:R:8:1) resulted in a mixture which caused hypouricemia. Hence, in the case of indacrinone the evaluation of the differences in both the pharmacodynamic and pharmacokinetic properties of the individual enantiomers and subsequent manipulation of the enantiomeric composition of the drug results in an improved therapeutic profile.

Introduction to the principles of drug design and action

170

4.5.6 The required activity resides in one or both enantiomers but the adverse effects are predominantly associated with one enantiomer Ketamine (4.99) is a general anaesthetic agent with analgesic properties which does not cause circulatory or respiratory depression. However, its use is restricted by postanaesthesia reactions including hallucinations and agitation; the drug is also the subject of abuse. Both enantiomers have anaesthetic properties but (+)-(S)-ketamine is between three to four fold more potent than the R-enantiomer and has approximately twice the affinity for the opiate receptor. The incidence of the adverse effects, the socalled “emergence reactions”, reported for the drug are greater following the administration of the R-enantiomer than either the racemate or (S)-ketamine. From the available information it would appear that the development of the single enantiomer would be therapeutically beneficial and also reduce the abuse potential of the drug.

4.6 SELECTED THERAPEUTIC GROUPS As pointed out previously the problems associated with drug stereochemistry are not restricted to particular groups of agents but extend across all therapeutic groups. In this section the stereochemistry of some selected therapeutic agents will be examined in an attempt to illustrate some of the complexities which may arise. This is not intended to be an exhaustive compilation but merely to serve as an indication of the potential advantages of stereochemical considerations in pharmacology. 4.6.1 Antiarrhythmic Agents Verapamil Verapamil (4.100) is a calcium channel blocking agent used for the treatment of supraventricular tachyarrhythmias, hypertension and angina. The pharmacodynamic activity of the enantiomers varies quantitatively, with the S-enantiomer being 2.5 to 20

Drug chirality and its pharmacological consequences

171

fold more potent than (R)-verapamil in terms of vasodilation and negative inotropic, dromotropic and chronotropic effects depending on the test system used. An examination of the pharmacological properties of the drug in vivo are also complicated by the formation of norverapamil (the N-desmethyl metabolite) an active metabolite which is reported to have ca 20% of the vasodilation activity of the drug. Verapamil undergoes extensive first-pass, or presystemic metabolism, and based on “total” drug concentrations has a bioavailability of between 20–30%. Examination of the plasma concentration effect relationship, by measurement of the PR interval prolongation following both oral and intravenous administration of the racemic drug indicates that the drug is apparently more potent following intravenous administration. A three fold greater plasma concentration being required following oral administration to produce the same pharmacodynamic effect, i.e. a shift in the dose response curve to the right is observed following oral drug administration. This difference in potency with route of administration is however only apparent and arises due to the differential oral bioavailability of the individual enantiomers of verapamil. Following intravenous administration the plasma concentrations of the less active R-enantiomer are twice those of the active S-enantiomer, whereas following oral administration this ratio R/S is ca 5. This difference in plasma concentration arises as a result of the differential bioavailabilities of the individual enantiomers (R, ≈50%; S, ≈20%), the higher clearance of (S)-verapamil (R/S≈0.57), and the higher volume of distribution of the Senantiomer (R/S≈0.4). The terminal half-lives of the two enantiomers are similar, between 4–5 hours, but not identical. Thus, an investigation of the stereochemical aspects of verapamil disposition explains the apparent anomaly in the concentrationeffect relationship with route of administration. In addition to its use in cardiovascular disease verapamil also has a potential application in the treatment of multidrug resistant tumours. In vitro studies with multidrug resistant tumour cell lines have indicated that verapamil enhances the

Introduction to the principles of drug design and action

172

cytotoxicity of the vinca alkaloids and anthracycline cytostatics, and reverses the resistance to these agents. One mechanism of multidrug resistance is due to a decreased accumulation of the cytotoxic agents as a result of the increased expression of P-glycoprotein (P-170) a membrane transport protein drug efflux pump. Verapamil inhibits this efflux pump by inhibiting the binding of cytotoxic agents and increases the intracellular content of vinblastine and related agents. Studies in vivo have however, been disappointing as the plasma concentrations of the drug required to enhance cytotoxicity cannot be achieved due to the cardiovascular effects of the drug. However, while the activity of (S)-verapamil is greater than that of the R-enantiomer in terms of the cardiovascular effects, both enantiomers have similar effects in terms of their inhibition of the membrane transport pump. It therefore follows that (R)verapamil may be potentially useful as a single isomer drug as higher doses may be used, compared to the racemate, with a reduction in the cardiovascular effects. Disopyramide Disopyramide (4.101) is used, as the racemate, in the treatment of ventricular and atrial arrhythmias and has anticholinergic effects, common side effects include dry mouth and urinary retention. The use of the drug is limited since it reduces cardiac output and left ventricular performance. The antiarrhythmic activity appears to reside predominantly in the enantiomer of the S-configuration, as determined by prolongation of electrocardiogram

QT interval which is between 4 to 5 fold longer following the S- than the Renantiomer. The S-enantiomer is also four to five fold more potent in terms of the anticholinergic activity. There are also pharmacokinetic complications with disopyramide as following the individual administration of the enantiomers to man there are no significant differences in the total clearance, renal clearance or volume of distribution. However, on administration of the racemic mixture (S)-disopyramide has a lower total clearance, renal clearance, volume of distribution and shorter plasma half-life compared to the R-enantiomer. These differences in pharmacokinetics, following administration of the individual enantiomers and the racemate, arise due to enantiomer-enantiomer interactions in plasma protein binding which is also concentration dependent. Tocainide

Drug chirality and its pharmacological consequences

173

Tocainide (4.102) is an orally active antiarrhythmic agent developed from lignocaine. The drug is used as a racemate but the R-enantiomer has three times the activity of (S)-tocainide in a chloroform induced model of fibrillation in the mouse. The plasma half-life of the R-enantiomer at approximately 10 hours is shorter than that of (S)-tocainide with the result that following an intravenous infusion of the drug the ratio of enantiomers (S/R) in plasma increases from ca 1 at 2 min to ca 1.7 after 48 hours. Hence “total” drug plasma concentrations will increase progressively during the infusion but with relatively small changes in pharmacological effect. There is also considerable interpatient variability in the enantiomeric composition of the material in plasma, the S/R ratio varying between 1.3 to 3.8, which is probably associated with variability in metabolism.

4.6.2 β-Blockers The β-adrenoreceptor antagonists may be divided into two chemical groups the arylethanolamine and aryloxypropanolamine derivatives. These agents show a high degree of stereoselectivity with respect to their action at the β-receptor with the pharmacological activity residing in the enantiomers of the R-configuration of the arylethanolamine series and the S-enantiomers of the aryloxypropanolamine group. Examination of the general structures of the active enantiomers of the two series, (4.38) and (4.39), indicates that the three-dimensional spatial arrangement of the active enantiomers are identical inspite of their opposite configurational designations. The stereoselectivity exhibited by these agents may vary markedly, the eudismic ratio for the binding affinity of atenolol enantiomers to the β-receptor being as low as 10 whereas that for pindolol is 1000. Differences in eudismic ratio between β-receptor subtypes have also been observed which indicate that β1-receptors are more sterically demanding than β2-receptors, i.e. higher endismic ratios are observed at β1-receptors than at the β2-subtype. This should not be surprising as there are known to be structural differences between the receptor subtypes. A recent QSAR study has indicated that the differences in enantiomer binding affinity between the two receptor subtypes is associated with higher equilibrium dissociation constants for the distomers at the β2-receptor subtype compared with the β1-subtype. An additional physicochemical property of significance in determining the binding affinity of these agents is their lipophilicity. The addition of the lipophilicity parameter (log P) to the QSAR correlation equations indicated that hydrophobic parameters are of greater significance

Introduction to the principles of drug design and action

174

for drug binding to the β2-receptor than the β1, particularly for the less active distomers for which the binding “fit” would obviously not be expected to be as good as for the eutomers. Thus the “steric” differences observed between the receptor subtypes may arise as a result of increased binding of the distomers to the β2-receptor via hydrophobic interactions. For those β-blockers which show additional pharmacological properties, e.g. the membrane stabilising effects, the enantiomers appear to be equipotent (see below). Of the β-blockers currently available only two, timolol (4.103) and penbutolol (4.104), are marketed as single isomers and being of the aryloxypropanolamine series these agents are available as the S-enantiomers. The remainder are marketed as racemates and in the case of one compound, labetalol (see below), as a mixture of four stereoisomeric forms. In terms of their use in the treatment of hypertension and angina there appears to be relatively little advantage in using single enantiomers particularly as the majority of the adverse effects are related to their pharmacological action and therefore a significant

reduction in side effects is unlikely. There are however a number of reasons why the stereochemistry of the β-blockers should not be neglected as indicated in the examples cited below. Propranolol (4.105) is a lipophilic non-selective β-blocker marketed as a racemate, the S-enantiomer being between 40 to 100 times, depending on the test system used, more potent as a β-blocker than the R-enantiomer. The enantiomers show no differences in activity with respect to the membrane stabilising properties of the drug. Following administration of racemic propranolol to man an examination of plasma concentration effect relationships, based on “total” drug concentrations, results in a

Drug chirality and its pharmacological consequences

175

shift in the dose response curve with route of administration. The drug appears to be between two to three fold more potent following oral dosing than following intravenous administration (the opposite to that observed with verapamil, Section 4.6.1). Propranolol undergoes extensive first-pass, or presystemic, hepatic metabolism which is stereoselective for the less active R-enantiomer, thus the apparent greater potency based on “total” plasma concentrations is a reflection of the increased proportion of the S-enantiomer in the circulating material. Following oral administration of the drug the enantiomeric composition of the material in plasma (S/R) varies between 1 to 4 fold. An additional contributory factor to the increased drug potency following oral compared to iv administration may be the higher plasma concentrations of the active metabolite 4-hydroxypropranolol (4.108). The metabolism and excretion of the enantiomers of propranolol have been examined in some detail. Three main pathways are involved, glucuronidation of the side chain hydroxyl group (4.106), oxidative metabolism of the aliphatic side chain to yield 3-naphthyloxylactic acid (4.107) and aromatic oxidation to 4hydroxypropranolol (4.108), which may undergo glucuronidation and sulphation. Each of these pathways may exhibit stereoselectivity and an examination of the urinary metabolites indicates that aromatic oxidation is selective for (R)-propranolol whereas side chain oxidation and glucuronidation are selective for the S-enantiomer (Table 4.4). However, the situation is slightly more complex and examination of the drug enantiomer clearance and partial metabolic clearance indicates that the metabolism of propranolol is dominated by aromatic oxidation to yield the 4-hydroxy compound, which is highly selective for the R-enantiomer; the partial metabolic clearance via the alternative pathways showing only slight stereoselectivity (Table 4.4). Thus, the enantiomeric composition of the urinary excretion products reflects the increased concentrations of (S)-propranolol available to undergo these transformations.

Introduction to the principles of drug design and action

176

Table 4.4 Fate of propranolol enantiomers in man. Urinary recovery Clearance and partial (% dose) metabolic clearance (L/min) R S S/R R S S/R Propranolol 0.16 0.24 1.50 2.78 1.96+ 0.71 Propranolol glucuronide 5.4 9.6 1.76 0.24 0.27 1.1 Naphthoxylactic acid 7.9 11.5 1.45 0.38 0.31 0.82 0.40 4-Hydroxypropranolol* 19.5 11.6 0.59 0.88 0.35+ * Total of conjugated material, i.e. both glucuronide and sulphate conjugates, both conjugation reactions may also exhibit stereoselectivity. + Significantly different for the two enantiomers of propranolol.

Drug chirality and its pharmacological consequences

177

As pointed out above timolol (4.103) is one of the few β-blockers presently available as a single enantiomer. In addition to its use in the treatment of hypertension and angina timolol is also used for the treatment of wide angle glaucoma. Following administration to the eye significant amounts of the drug are systemically absorbed and cardiovascular and pulmonary side effects have been reported. This systemic absorption is of particular significance for the use of the drug in patients for whom βblocking agents are contraindicated, e.g. those with respiratory disease states, and a number of deaths have been reported following the use of timolol eye drops in asthmatic patients. Using pharmacological test systems for the evaluation of β-blockade (S)-timolol shows marked stereoselectivity in action with eudismic ratios (S/R) of between 50 and 90 depending on the test system used. These large differences in activity reduce to ca three fold when the ocular properties of the drug are examined, e.g. reduction in aqueous humour recovery rate, inhibition of dihydroalprenolol binding in the irisciliary body. The R-enantiomer of timolol has also been shown to reduce intraocular pressure in patients with glaucoma with fewer systemic effects than (S)-timolol. In addition, recent investigations have indicated that (R)-timolol increases retinal/choroidal blood flow, whereas the S-enantiomer decreases it, an unrequired effect. The stereoisomers of timolol therefore represent a possible example of a drug where both enantiomers could be marketed for specific therapeutic indications, the Senantiomer for the treatment of cardiovascular disease states and the R-enantiomer for the treatment of glaucoma. Sotalol (4.109), an arylethanolamine derivative, is a non-selective β-blocker used as a racemate the (−)-enantiomer being 14–50 fold more active, depending on the test system used, than (+)-sotalol in terms of β-blockade. Racemic sotalol also has antiarrhythmic activity, prolonging the duration of the cardiac action potential and increasing ventricular repolarisation time. In terms of antiarrhythmic activity both enantiomers appear to be equipotent and it has been suggested that the single (+)enantiomer may have potential as an antiarrhythmic agent devoid of β-blocking activity. Labetalol (4.110, see Table 4.5), an arylethanolamine derivative, is a dual action drug with combined α and β-blocking activity. Labetalol contains two chiral centres and is marketed as an equal parts mixture of all four possible stereoisomers. Examination of the pharmacological activity (pA2 values) of the four possible stereoisomers (Table 4.5), indicates the β-blocking activity resides in the R,Rstereoisomer, the α1-blocking activity in the S,R-stereoisomer and that the remaining pair are essentially inactive. Labetalol is certainly not one drug with two actions. The R,R-stereoisomer of labetalol, named dilevalol, has been investigated for development as a single isomer β-blocker. However, the development of this compound was stopped following clinical trials in which a small number of patients developed drug induced hepatitis. This adverse effect appears to be of minor significance with respect to labetolol and the reason why the single isomer should

Introduction to the principles of drug design and action

178

Table 4.5 Pharmacological activity of the stereoisomers of labetalol.

1

R, R S, S R, S S, R -

R HO H HO H

2

R H OH H OH

3

R H CH3 CH3 H

4

R CH3 H H CH3

Activity (pA2 values) α1 β1 β2 5.87 8.26 8.52 5.98 6.43 <6.0 5.5 6.97 6.33 7.2 6.37 <6.0

show increased toxicity is at present unknown. Labetalol/dilevalol represents an interesting example indicating that removal of unrequired stereoisomers from a mixture may not be a trivial matter. 4.6.3 Anticoagulants The 4-hydroxycoumarin anticoagulants act by competitive inhibition of the vitamin K dependent step in clotting factor synthesis. Warfarin (4.77), and the related compounds phenprocoumon (4.111) and acenocoumarol (4.112) are used as racemic mixtures and in the case of these agents it is difficult to differentiate their pharmacodynamic effects from their pharmacokinetic properties. For example (R)acenocoumarol is reported to be several times more potent than the S-enantiomer. However, pharmacokinetic studies have indicated that the clearance of (S)acenocoumarol is ten times that of its antipode and at therapeutic doses of the racemate the plasma concentrations of the S-enantiomer are at the limits of detection. The pharmacological activity observed resulting from administration of the racemate is therefore almost entirely due to the R-enantiomer. Animal studies however have demonstrated that (S)-acenocoumarol is intrinsically more potent than the Renantiomer, a result that would be expected as the S-enantiomers of both warfarin and

Drug chirality and its pharmacological consequences

179

phenprocoumon are between two to six times more active than their antipodes depending on the test species used. At normal therapeutic dose levels warfarin is a relatively safe drug and cases of excessive dosage may be managed by the administration of vitamin K. The major problems associated with the use of warfarin arise as a result of drug interactions and warfarin is probably the most extensively investigated drug with respect to stereoselectivity and drug interactions. Some agents selectively interact with (S)warfarin, e.g. phenylbutazone, sulfinpyrazone, whereas others are selective for (R)warfarin, e.g. cimetidine, enoxacin, and other agents, e.g. amiodarone show no stereoselectivity. The majority of the interactions

appear to have a pharmacokinetic rather than a pharmacodynamic basis, the underlying mechanism of which may be fairly complex. For example coadministration of phenylbutazone with racemic warfarin results in an enhanced anticoagulant effect with no significant change in the plasma half-life or clearance of the “total” drug. Examination of the effect of phenylbutazone on the protein binding and clearance of the individual enantiomers of the drug indicated a greater displacement of (R)compared to (S)-warfarin from protein binding sites and a marked reduction of the unbound clearance of the S-enantiomer compared to a much smaller inhibition of (R)warfarin. The resultant effect of these differential interactions is a marked reduction in the total clearance of the more active S-enantiomer and an increase in the clearance of (R)-warfarin. Thus, the increased pharmacodynamic effect results from a combination of stereoselectivity in both displacement of the enantiomers of warfarin from plasma protein binding sites and inhibition of metabolism. The complexity of the situation is

Introduction to the principles of drug design and action

180

emphasized as the total warfarin clearance, i.e. the value obtained using “total” plasma concentrations, does not change. 4.6.4 Antihistamines The H1-blocking activity of the dimethylaminopropyl series of antihistamines, e.g. pheniramine (4.113), chlorpheniramine (4.114) and brompheniramine (4.115), resides in the enantiomers of the S-absolute configuration. The eudismic ratios being approximately 30, 200 and 150 respectively. In the case of chlorpheniramine the therapeutic index for the two enantiomers are 3380 and 25 for S and R respectively, determined in guineapigs. Both chlorpheniramine and brompheniramine have been marketed as the single

S-enantiomers in the USA and it has been reported that the reduced dose of the single isomers compared to the racemate resulted in a corresponding reduction in the sedative effects of these agents. Within the dimethylaminoethyloxyether series of compounds, i.e. derivatives of diphenhydramine (4.116), substitution of one of the phenyl rings or replacement of one ring by a 2-substituted pyridine ring results in the introduction of a chiral centre. The 4-methyl substituted derivative of diphenhydramine shows stereoselectivity in action, the R-enantiomer (4.117) being ca 65 times more potent than its antipode using a guineapig ileum test system. The 2-pyridyl derivative, carbinoxamine (4.118), also shows stereoselectivity in action the S-enantiomer being ca 30 times more potent than the R-isomer. Comparison of the structures of the more potent isomers of these two compounds

Drug chirality and its pharmacological consequences

181

with the eutomers of the pheniramine series, e.g. chlorpheniramine (4.114), illustrates the significance of considering the three dimensional structure of a molecule rather than the configurational designation when comparing biological activity. Methyl substitution at the benzylic carbon results in an additional series of active agents, e.g. clemastine (4.119). In this example the basic dimethylamino group has been replaced by a 2-substituted pyrrolidine ring resulting in the introduction of an additional chiral centre. The activities of all four stereoisomers have been examined and the major determinant of activity is the configuration at the benzylic carbon atom, the pA2 values using guinea-pig ileum being: R,R, 9.45; R,S, 9.40; S,S, 7.99 and S,R, 8.57. The structure of clemastine (4.119) represents the R,R-stereoisomer the Rconfiguration at the benzylic carbon corresponding topologically to the active isomers of the previous series, clemastine is marketed as the single R,R-stereoisomer.

Introduction to the principles of drug design and action

182

4.6.5 Non-steroidal anti-inflammatory drugs 2-Arylpropionic acids The 2-arylpropionic acids (2-APAs) of general structure (4.20) are an important group of non-steroidal anti-inflammatory drugs (NSAIDs). The majority of these agents are used as racemic mixtures even though their major pharmacological activity, inhibition of cyclooxygenase, resides in the enantiomers of the S-absolute configuration, the R-enantiomers being either inactive, or weakly active, in in vitro test systems (Table 4.6). At present naproxen (4.58), flunoxaprofen (4.120), ibuprofen (4.21), in Austria, and ketoprofen, in Spain, are marketed as the single S-enantiomers. The large differences observed in the in vitro activity of the enantiomers of these agents decrease markedly in in vivo test systems and in some cases, e.g fenoprofen (4.121) and ibuprofen, the enantiomers appear to be essentially equipotent (Table 4.6). This difference in enantiomeric activity in vivo and in vitro is in part due to the metabolic chiral inversion of the inactive R-enantiomers to their active S-antipodes in both animals and man. The pharmacological activity of these agents, when

Drug chirality and its pharmacological consequences

183

administered as racemates is therefore closely linked with the stereochemical aspects of their metabolism. The mechanism of the chiral inversion reaction is thought to involve the formation of an acyl-coenzyme A thioester of the R-enantiomer of the 2-APAs. The enzyme mediating this reaction has not been fully characterised but is believed to be an acyl CoA synthetase of the type involved in the metabolism of endogenous fatty acids. Once formed the (R)-2-arylpropionyl-coenzyme A thioester (4.122) undergoes epimerisation of the 2-arylpropionyl moiety to yield a mixture of both possible epimeric acyl-CoA derivatives, which may then undergo hydrolysis to liberate both enantiomers of the 2-arylpropionate. The enzymology of this process is not well understood but chemical synthesis and biochemical investigations on both possible acyl-CoA thioesters have indicated that both stereoisomers undergo the epimerisation reaction. The stereoselective step in the pathway appears to be formation of the (R)-2arylpropionyl-CoA thioester, the S-enantiomers of the substrates being unable to form the thioester. An alternative pathway to the hydrolysis of the acyl-CoA thioesters is acyl transfer of the 2-arylpropionyl moiety to glycerol resulting in the formation of “hybrid” triglycerides, e.g. (4.123) and hence distribution of the drug into adipose tissue. The stereochemistry of the 2-arylpropionic acid moiety found in adipose tissue has been investigated in the rat

Introduction to the principles of drug design and action

184

following administration of both the individual enantiomers and racemic ibuprofen. Administration of both (R,S)- and (R)-ibuprofen resulted in the incorporation of the drug into triglycerides, the enantiomeric composition of the material being R>S in both cases. Total drug lipid levels, following administration of equal doses, were approximately twice as high following the administration of (R)-ibuprofen than the racemate, whereas following the administration of (S)-ibuprofen only trace quantities of the drug could be detected. In vitro studies using (R)- and (S)-fenoprofen (4.121) as substrates and rat hepatocyte and adipocyte preparations, indicated stereoselective incorporation of (R)-fenoprofen into triglycerides and also that (R)- but not (S)fenoprofen inhibited endogenous triglyceride synthesis in vitro. The above investigations indicate that both the (R)- and (S)-2-arylpropionyl moiety may be transferred from the acyl-CoA thioesters to glycerol but that the incorporation of drug depends upon the presence of the R-enantiomer, as would be expected from the stereoselectivity of acyl-CoA formation. The toxicological significance of drug incorporation is not known but it has been suggested that the formation of hybrid triglycerides may result in the accumulation of these agents and possible toxicity due to their effects on normal lipid metabolism and membrane function. Chiral inversion has been reported for a number of the 2-arylpropionic acids in both animals and man, the extent of the reaction appears to be both substrate and species

Drug chirality and its pharmacological consequences

185

dependent. Human pharmacokinetic studies have shown that ibuprofen, fenoprofen and

Table 4.6 Relative activity of the enantiomers of 2-arylpropionic acid NSAIDs in in vitro and in vivo test systems. Compound In vitro Test Test In vivo Ratio S/R Ratio S/R Carprofen >16 IPGS 14 Acute adjuvant >24 IPA induced arthritis Fenoprofen 35 IPA 1 Carrageenin paw oedema; UVE Flurbiprofen 200 IPA 2–16 Guinea-pig 880 Antagonism of anaphylaxis SRS-A Ibuprofen 160 IPGS 1.4 Toxin induced writhing; Pain threshold 1.1 UVE Indoprofen 100 IPGS 20 Carrageenin paw oedema 31 Granuloma pouch 25 Toxin induced writhing Naproxen 130 IPGS 28 Carrageenin paw oedema 70 IPGS 15 Antipyretic activity Pirprofen 6.4 IPGS IPGS, inhibition of prostaglandin synthesis; IPA, inhibition of platelet aggregation; UVE, ultraviolet induced erythema.

Introduction to the principles of drug design and action

186

benoxaprofen undergo significant inversion in man, whereas the reaction either does not occur or is relatively minor for indoprofen, flurbiprofen, ketoprofen and carprofen. In addition to chiral inversion a number of these agents show stereoselectivity in plasma protein binding (e.g. fraction unbound ibuprofen enantiomers S>R) and in other routes of metabolism, e.g. glucuronidation and oxidation. In the majority of cases following administration of the racemic drug the plasma concentrations and areas under the plasma concentration versus time curves (AUC), of the active Senantiomers exceed those of their R-antipodes (e.g. benoxaprofen, carprofen, fenoprofen, flurbiprofen, ibuprofen, indoprofen), but others, e.g. ketoprofen and tiaprofenic acid, show similar plasma concentrations and AUCs. The dispositional properties of these agents are also complicated by enantiomer-enantiomer interactions. For example both enantiomers of ibuprofen show concentration-dependent plasma protein binding and compete for binding sites, which may explain the differences observed in their pharmacokinetic parameters when administered as single enantiomers or as the racemic mixture. For those 2-arylpropionates marketed as racemic mixtures and for which chiral inversion is a significant route of metabolism, the effective dose of the active agent is unknown. In the case of these agents the R-enantiomers act essentially as pro-drugs for their active S-antipodes. The extent of the inversion reaction would also be expected to vary within the population, possibly with disease state and thus any attempt to relate plasma concentrations to clinical effect must take the stereochemistry of the circulating drug into account. The use of the single S-enantiomers of these agents offers a number of advantages: accurate dosing, simplification of pharmacokinetics and concentration-effect relationships and avoidance of the potential problems due to hybrid triglyceride formation and inhibition of fatty acid metabolism. Sulindac Sulindac (4.7), a benzylidine analogue of indomethacin, is a pro-drug (see Chapter 7) the anti-inflammatory activity of which resides in the sulphide metabolite (4.124), the other major metabolite is the sulphone derivative (4.125). Sulindac is used as the racemic

Drug chirality and its pharmacological consequences

187

sulphoxide and little is known concerning the stereoselectivity of the reduction to the active sulphide or the oxidation to the sulphone. The stereochemistry of the sulphide (4.124) oxidation back to sulindac (4.7) has been investigated in vitro and the reaction appears to show stereoselectivity for the formation of the (+)-enantiomer. The stereochemistry of the material in plasma will therefore depend on potential stereoselectivity in sulphoxide reduction, sulphide oxidation and oxidation of the sulphoxide to the sulphone (4.125). It is therefore likely that the enantiomeric composition of sulindac will vary with time but the clinical significance of this is unknown. 4.6.6 Antimicrobial agents β-Lactam antibiotics The majority of the β-lactam antibiotics are semisynthetic agents, the stereochemistry of the 6-aminopenicillanic (6-APA; 4.126) and 7aminocephalosporanic (7-ACA; 4.127) nucleii being determined as 3S, 5R, 6R and 6R, 7R respectively. The introduction of an α-substituted acyl side chain on the 6-APA and 7-ACA nucleii results in the introduction of an additional chiral centre and the formation of two epimers, e.g. ampicillin (4.35), carbenicillin (4.128) and cephalexin (4.36). In the case of ampicillin (4.35) and cephalexin (4.36) the official preparations are those containing the D-configuration in the side chains which correspond to the Rdesignation using the sequence rule system. The differential absorption of the epimers of cephalexin have been referred to previously (Section 4.4.1). The two epimers of

Introduction to the principles of drug design and action

188

ampicillin differ in terms of their aqueous solubility and activity, the epimer of the Dconfiguration in the side chain having two to five fold greater activity, depending on the test system used, than that of the L-epimer. Unlike the above examples carbenicillin is used as an epimeric mixture. The individual epimers of carbenicillin show only slight differences in activity, but more importantly are stereochemically unstable undergoing rapid epimerisation in solution. The contribution of the stereochemical instability to the observed lack of difference in activity is by no means clear, but in the case of carbenicillin separation of the individual epimers would appear to be a futile exercise.

Moxalactam (4.22), a l-oxacephem derivative, is also used as a mixture of two epimeric forms, designated R and S with respect to the acyl side chain chiral centre. The antimicrobial activity of the compound resides predominantly in the R-epimer which is approximately twice as active as (S)-moxalactam depending on the test system used. The two epimers are stereochemically unstable and undergo epimerisation to yield equilibrium mixtures in the ratio R:S of 50:50 and 45:55 in buffer and serum respectively. The rates of epimerisation vary depending on the environment and epimeric form but at 37°, in serum, the half-life of epimerisation is the same for both compounds (1.5 h). Following intravenous infusion of the epimeric mixture to man the serum elimination half-life of the “total” drug is about 2.3 hours,

Drug chirality and its pharmacological consequences

189

and the serum concentrations of the less active S-epimer are approximately twice those of the R-epimer within four hours with a ratio R/S in renal clearance of 1.5. In terms of the relative merits of single isomers versus stereoisomeric mixtures compounds such as moxalactam present considerable problems as the half-life of epimerisation under physiological conditions is only slightly shorter than the serum elimination half-life. It would be difficult under such circumstances to recommend the use of a single stereoisomer. β-Lactam Pro-drugs The poor oral availability of a number of the penicillins and cephalosporins has resulted in the synthesis of lipophilic ester pro-drugs (see Chapter 7). The majority of these are not simple esters but involve the introduction of an acyloxymethyl or acyloxyethyl function into the molecule. These groups undergo rapid enzymatic hydrolysis in vivo to yield the corresponding hydroxymethyl or hydroxyethyl esters which, being hemiacetal derivatives, spontaneously cleave with liberation of the active β-lactam and the corresponding aldehyde. The introduction of an hydroxyethyl function into the promoiety results in an additional chiral centre and therefore a pair of diastereoisomers, e.g. cefuroxime axetil (4.129) and cefdaloxime pentexil (4.130), which may differ in terms of their physicochemical properties and also their susceptibility to enzymatic hydrolysis. Cefuroxime axetil (4.129) undergoes hydrolysis in vivo to yield cefuroxime, acetaldehyde and acetic acid and is used as an equal parts mixture of the two epimers. Following oral administration to man the pro-drug can not be detected in the systemic circulation and shows a bioavailability based on urinary recovery of cefuroxime of between 30 to 50%.

Introduction to the principles of drug design and action

190

Stereoselectivity in the hydrolysis of the pro-drug occurs by both serum and intestinal mucosal esterases isolated from both rat and dog tissue. In all cases the S-epimer is hydrolysed selectivity but the selectivity varies between 2.5 to 14 fold with both tissue and species. Such Stereoselectivity in hydrolysis in the gut may contribute to the observed bioavailability of the liberated cefuroxime in man. Cefdaloxime is poorly absorbed from the gastrointestinal tract and has been esterified to yield the pivaloylethyl pro-drug (4.130). The bioavailability and pharmacokinetics of cefdaloxime have been investigated following the administration

Drug chirality and its pharmacological consequences

191

of the individual and an equal parts mixture of the pro-drug diastereoisomers to experimental animals. Following administration to the dog the bioavailability of cefdaloxime was three times greater after dosing with the S-epimer than the R. It has been reported that a similar situation occurs in man and the S-epimer has been selected for further development. Stereoselectivity in the absorption of diastereoisomeric pro-drugs and therefore the subsequent availability of the drug, may arise as a result of differential solubility at the absorption site, rates of diffusion through the gut wall and enzymatic activity in the intestinal mucosa, liver and blood, and as such the potential problems associated with the introduction of a chiral promoiety into a molecule need to be taken into consideration at the compound design stage. Quinolones The quinolones are synthetic antibacterial agents based on the 1,4-dihydro-4oxopyridine-3-carboxylic acid ring system. An important subgroup of these agents possess a tricyclic fused ring structure with a chiral centre in the saturated ring, e.g. ofloxacin (4.131), flumequine (4.132) and methylflumequine (4.133). The antibacterial activity of these agents has been shown to reside in the enantiomers of the S-absolute configuration, the R-enantiomers being considerably less active than the corresponding racemates, the S-enantiomers having approximately twice the activity of the racemates. In the case of ofloxacin (4.131) the difference in in vitro enantiomeric activity ranges from 8 to 128 fold against both Gram-positive and Gram-negative bacteria. In addition the corresponding non-chiral analogues of methylflumequine (4.133) and ofloxacin (4.131), i.e. structures (4.133) and (4.131) where R1=R2=H, are more active

Introduction to the principles of drug design and action

192

than the R-enantiomers but less active than the racemates. Such data implies steric constraints at the site of action with the orientation of the methyl group attached to the chiral centre hindering the interaction in the case of the R-enantiomers and enhancing the interaction of the S-enantiomers. The target enzyme of the quinolones is believed to be DNA gyrase (bacterial topoisomerase II) and good correlations between the IC50 concentrations for enzyme inhibition and antimicrobial activity, as determined by MIC concentrations, have been obtained for this series of compounds. In the case of ofloxacin the rank order of potencies for enzyme inhibition is identical to that observed for MIC activity, with the S-enantiomer being 9.3 and 1.3 fold more active in terms of enzyme inhibition, than the R-enantiomer and the racemate respectively. As there are similarities between DNA gyrase and mammalian topoisomerase II it is useful to evaluate the activity of the quinolones on the enzyme and hence their effects on mammalian cells. The rank order of potency of ofloxacin isomers against mammalian topoisomerase II is the same as that obtained with DNA gyrase, i.e. S>R,S>R. However, the relative activity of the two enantiomers decreases from 12.4 with DNA gyrase to 1.8 against topoisomerase II. More importantly the S-enantiomer is 6.7 fold more selective than the R-isomer with respect to the DNA gyrase. The non chiral analogue (4.131, R1=R2 =H) of ofloxacin is the least selective of the compounds examined. Thus, the presence and orientation of the methyl group at the chiral centre not only determines the potency of these compounds but also increases their selectivity of action.

Drug chirality and its pharmacological consequences

193

A number of quinolone derivatives have also been developed which are substituted at carbon-7 of the bicyclic ring system and contain a chiral centre in the substituent, e.g. temafloxacin (4.134). In comparison to ofloxacin and related derivatives, differences in

the enantiomeric activities of the 7-substituted compounds are of relatively minor significance. For example the enantiomers of temafloxacin show only small differences in activity in in vitro test systems and possess similar activities against DNA gyrase. This difference in stereoselectivity of action between the two series of quinolones is presumably due to the centre of chirality in the 7-substituted compounds being in a position remote from the critical binding region of these molecules. 4.7 TOXICOLOGY The process of drug safety evaluation is complex, expensive and time consuming involving acute and chronic toxicity testing, mutagenicity and genetic toxicology, reproductive toxicology, carcinogenicity and clinical safety evaluation both pre- and post-marketing. There is also a need to carry out mechanistic and toxicokinetic studies in order to determine the animal exposure to both the drug and metabolites and to aid in the extrapolation of animal data to man. At present there is relatively little published data on the comparative toxicity of single enantiomers versus racemic drugs and even less information arising from clinical studies. However, examples may be cited which are illustrative of aspects of stereochemical considerations in safety evaluation.

Introduction to the principles of drug design and action

194

Fenvalerate (4.135), is a synthetic pyrethroid insecticide which contains two chiral centres, and thus four stereoisomers are possible. Administration of the compound in the diet to a range of animal species resulted in granulomatous changes in the liver, lymph nodes and spleen. Separation and toxicological evaluation of the individual stereoisomers indicated that the toxicity was associated with only one of the four isomers. Subsequent metabolic studies indicated that the toxicity was associated with the formation and disposition of a cholesteryl ester, (R)-2-(4-chlorophenyl)isovaleric acid cholesterylester (4.136), formed by transesterification of the single toxic stereoisomer of fenvalerate. Fortunately the active isomer of fenvalerate may be synthesised stereospecifically. While not a drug this example does indicate that stereochemical considerations may prevent a compound being discarded following an adverse toxicological evaluation. A similar situation in terms of stereoselective toxicity appears to occur with the potassium channel activator cromakalim the activity of which resides in the (−)(3S,4R)-enantiomer (4.137). Administration of high doses of the racemate to the monkey resulted in the development of heart lesions which appear to be associated with the (+)-enantiomer. This compound is now under development as the single (−)enantiomer.

Drug chirality and its pharmacological consequences

195

Some examples of the use of single isomers versus racemic mixtures of relatively “old” compounds in the clinic are also known, e.g. D-penicillamine and L-dopa (4.64). The use of both these compounds as pure enantiomers, rather than their racemates, resulted in a decrease in toxicity. The initial use of racemic dopa for the treatment of Parkinson’s disease resulted in nausea, vomiting, anorexia, involuntary movements and granulocytopenia. The use of L-dopa resulted in halving the required dose, a reduction in toxicity, granulocytopenia was not observed with the single enantiomer, and an increased number of improved patients. Similarly the use of synthetic racemic penicillamine in the USA for the treatment of Wilson’s disease resulted in a number of adverse reactions including nephrotic syndrome, optic neuritis, thrombocytopenia

Introduction to the principles of drug design and action

196

and/or leukopenia and the racemate was withdrawn. In the UK, D-penicillamine was obtained by the hydrolysis

of penicillin and the adverse effects were either reduced or abolished. Animal studies have also indicated that L-penicillamine inhibits growth with weight loss, causes intermittent fits and death, toxicity not observed with the D-enantiomer. Recent investigations have also indicated that the mutagenic activity of L-penicillamine is approximately eight fold greater than that of the D-enantiomer. Thalidomide (4.138) is a compound frequently cited, particularly in the popular press, to support arguments for the development of single isomer drugs. Thalidomide was introduced, as the racemate, into therapeutics in the early 1960s as a sedativehypnotic agent and was used in pregnant women for the relief of morning sickness. However, the drug was withdrawn when it became apparent that its use in pregnancy was associated with malformations, particularly phocomelia (shortening of the limbs), in the offspring. Investigations in the late 1970s using SWS mice indicated that both isomeric forms of the drug are hypnotic agents but that the teratogenic properties of the drug reside in the S-enantiomer. Thus, the argument goes: if the drug had been used as the single R-enantiomer then the tragedy of the early 1960s could have been avoided. However, the situation with thalidomide is much more complex. Rodents are resistant to the teratogenic toxicity of the drug and the mouse is a poor model for teratogenicity testing. Data obtained in a more sensitive test species, New Zealand White rabbits, indicates that both the enantiomers of thalidomide are teratogenic. An additional problem with the drug is its stereochemical stability since the single isomers undergo rapid racemisation in biological media. Thus, even if a single isomer was administered to an experimental animal the other would be formed relatively rapidly. The acute toxicity of thalidomide, as determined by the LD50 test, also presents a complex problem. The individual enantiomers have similar reported LD50 values of approximately 1.0–1.2 g/kg in mice, but the value for the racemate is greater than 5 g/kg, i.e. the racemate is non toxic. In this case it would appear that the administration of the racemic mixture is exerting a protective effect the mechanism of which is unknown. Taken together the above information indicates that the situation with thalidomide is by no means as clear as sometimes implied and the drug is certainly not a good example to cite in support of arguments for single isomer drugs.

Drug chirality and its pharmacological consequences

197

A number of chiral drugs administered as racemates have been withdrawn from use, e.g. the cardioselective β-blocker practolol, the NSAID benoxaprofen, the anticholinergic calcium antagonist terodiline. In the majority of cases the significance of stereochemistry to the adverse reactions is difficult to assess as the information is not available. However, the use of single isomers would have halved the required dose and the adverse reactions may have been reduced as a consequence. 4.8 RACEMATES VERSUS ENANTIOMERS: THE FUTURE As pointed out in the Introduction drug chirality has become a significant consideration for both the pharmaceutical industry and the regulatory authorities. Should all chiral drugs be marketed as single isomers? There are a number of arguments in favour of this approach, e.g. the plasma-concentration-effect relationships are simplified; the pharmacokinetic profile is less complex; there is a reduced potential for complex drug interactions; removal of a potentially interacting or toxic “impurity” resulting in an improved pharmacological profile of the drug and the potential for an increase in therapeutic index. The single enantiomer versus isomeric mixture debate will obviously have a considerable impact on new drug development and there is already evidence which indicates that the number of single isomer drugs/products being presented to the regulatory authorities is increasing. In the late 1980s the US Food and Drug Administration (FDA) issued a statement to the effect that “the Agency is impressed by the possibility that the use of single enantiomers may be advantageous by permitting better patient control, simplifying dose-response relationships and by reducing the extent of interpatient variation in drug response”. Both the FDA and the European Union Committee on Proprietary Medicinal Products (CPMP) have issued formal guidelines for the investigation of chiral active substances, as have authorities in Switzerland, Australia and the Nordic Countries. In Japan no formal guidelines have been issued but stereochemical matters are dealt with via a normal consultation procedure. At present none of the regulatory bodies have an absolute requirement for the development of single isomer drugs; however if a racemate is presented for evaluation then its use must be justified. There are a number of arguments which may be used to support the submission of a racemate: 1) the individual isomers are stereochemically unstable and readily racemise in vitro and/or in vivo; 2) the preparation of the drug as a single enantiomer on a commercial scale is not technically feasible; 3) the individual enantiomers have similar pharmacological and lexicological profiles; 4) one enantiomer is known to be totally inactive and not provide an additional body burden or influence the pharmacokinetic properties of the other; 5) the use of a racemate, or non-racemic mixture of isomers, produces a superior therapeutic effect than either individual enantiomer.

Introduction to the principles of drug design and action

198

An additional valid question is the therapeutic significance of the compound in relation to the seriousness of the disease state and drug adverse reaction profile. In addition to new drug development a number of established drugs, marketed as racemates, have been examined to see if their adverse reaction profile may be improved if used as single stereoisomers. This so-called “Racemic Switch” has at present resulted in a small number of compounds being re-marketed as single isomer preparations in some countries, e.g. the anorectic agent dexfenfluramine in Europe, the antimicrobial agent levofloxacin in Japan (currently undergoing Phase III clinical trials in Europe and the USA) and the NSAIDs dexibuprofen in Austria and dexketoprofen in Spain. In the case of dexfenfluramine (4.139) the racemate had been available for over 25 years and a considerable amount of clinical data had been accumulated. In terms of the pharmacology of the compound, (+)-(S)-fenfluramine (4.139) has between four to five times greater

activity than the R-enantiomer in terms of serotonin receptor activity and reduction in food intake with only twice the toxicity in acute screening tests. However, the Renantiomer does exhibit side effects and it was possible to demonstrate an improved risk-benefit ratio with the single isomer compared to the racemate. The most recently introduced (1996) single isomer drug in the UK is cisatracurium besylate, the 1R, 2R, 1’R, 2’R-isomer (4.140) of the non-depolarizing neuromuscular blocking agent atracurium (4.5). Atracurium contains four chiral centres but due to its symmetrical structure exists as ten isomeric forms. The commercially available material consists of an unequal parts mixture of the ten forms of which the 1R, 2R, 1’R, 2’R-isomer accounts for 15% of the material. The single isomer has similar pharmacodynamic properties to atracurium in terms of onset, duration and recovery of action, with an improved side effect profile with respect to cardiovascular effects and histamine release. Other compounds under examination as racemic switches include the β-blocker sotalol, the antiarrhythmic verapamil and the anaesthetic-analgesic agent ketamine. However, additional compounds presented in alternative formulations allowing different routes of drug administration will probably be marketed in the near future. The resurgance of interest in drug chirality has also indicated other agents which may be candidates for the racemic switch. For example (R)-salbutamol (4.141) is 68 times more active than the S-enantiomer as a β2-agonist. Recent reports indicate that the Senantiomer induces airway hyper-reactivity and may cause adverse effects in asthmatic patients and thus

Drug chirality and its pharmacological consequences

199

salbutamol may be considered for racemic switch to the single R-enantiomer. That such reintroductions of single isomer drugs may not be without problems and may provide unexpected results is illustrated by the example of labetalol/dilevalol referred to previously (Section 4.6.2). In the case of labetalol removal of the isomeric “impurity” was not a trivial matter. 4.9 CONCLUDING COMMENT The material presented in this chapter was selected to provide a background to the biological significance of drug chirality and to highlight the advantages of stereochemical considerations in pharmacology. In the past such quotes as “Warfarin enantiomers should be treated as two drugs” and “(S) and (R)-propranolol are essentially two distinct entities pharmacologically” have appeared in the literature. In the future, as a result of regulatory considerations, the majority of chiral drugs will be introduced as single isomers and a number of compounds currently marketed as racemates will be introduced as single isomers. But for many drugs, currently in use as racemates, relatively little is known regarding the pharmacological or toxicological activities or pharmacokinetic properties of the individual enantiomers. For example there is little published information concerning the effect of novel formulations on enantiomer delivery or bioavailability; the influence of ageing, disease state, gender, or genetic factors on drug enantiomer disposition; the influence of drug interactions

Introduction to the principles of drug design and action

200

with respect to stereoisomers. The results of additional pharmacological and pharmacokinetic investigations of the enantiomers of marketed racemates may result in new indications for “old” drugs, improve the clinical use of these agents and hence result in increased safety and efficiacy. Probably the best take home message for the budding medicinal chemist would be: if you make a chiral compound, finish the job and separate the isomers yourself don’t expect the patient to do it for you. FURTHER READING Aboul-Enein, H.Y. and Wainer, I.W. (eds.) (1997) The Impact of Stereochemistry on Drug Development and Use. New York: Wiley. Ariëns, E.J. (1984) Stereochemistry, a basis for sophisticated nonsense in pharmacokinetics and clinical pharmacology. European Journal of Clinical Pharmacology 26, 663–668. Ariëns, E.J., Soudijn, W. and Timmermans, P.B.M.W.M. (eds.) (1983) Stereochemistry and Biological Activity of Drugs. Oxford: Blackwell. Cahn, R.S., Ingold, C.K. and Prelog, V. (1956) The specification of asymmetric configuration in organic chemistry. Experimenta 12, 81–94. Caldwell, J., Winter, S.M. and Hutt, A.J. (1988) The pharmacological and toxicological significance of the stereochemistry of drug disposition. Xenobiotica 18(suppl 1), 59–70. De Camp, W.H. (1989) The FDA perspective on the development of stereoisomers. Chirality 1, 2–6. Easson, L.H. and Stedman, E. (1933) Studies on the relationship between chemical constitution and physiological action. V Molecular dissymmetry and physiological activity. Biochemical Journal 27, 1257–1266. Eliel, E.L. and Wilen, S.H. (1994) Stereochemistry of Organic Compounds. New York: Wiley. Evans, A.M. (1992) Enantioselective pharmacodynamics and pharmacokinetics of chiral non-steroidal anti-inflammatory drugs. European Journal of Clinical Pharmacology 42, 237–256. Hutt, A.J. and Caldwell, J. (1983) The metabolic chiral inversion of 2-arylpropionic acids; a novel route with pharmacological consequences. Journal of Pharmacy and Pharmacology 35, 693–704. Hutt, A.J. and O’Grady, J. (1996) Drug chirality: a consideration of the significance of the stereochemistry of antimicrobial agents. Journal of Antimicrobial Chemotherapy 37, 7–32. Lehmann, F.A.F. (1982) Quantifying stereoselectivity or how to choose a pair of shoes when you have two left feet. Trends in Pharmacological Sciences 3, 103–106. Patil, P.N., Miller, D.D. and Trendelenburg, U. (1975) Molecular geometry and adrenergic drug activity. Pharmacology Reviews 26, 323–392. Pfeiffer, C.C. (1956) Optical isomerism and pharmacological action—a generalisation. Science 124, 29–31. Rauws, A.G. and Groen, K. (1994) Current regulatory (draft) guidance on chiral medicinal products: Canada, EEC, Japan, United States. Chirality 6, 72–75.

Drug chirality and its pharmacological consequences

201

Smith, D.F. (ed.) (1989) Handbook of Stereoisomers: Therapeutic Drugs. Boca Raton: CRC Press. Stereochemistry in Drug Action. Proceedings of the Third Biochemical Pharmacology Symposium (1988) Biochemical Pharmacology 37, 1–148. Tucker, G.T. and Lennard, M.S. (1990) Enantiomer specific pharmacokinetics. Pharmacology and Therapeutics 45, 309–329. Wainer, I.W. (ed.) (1993) Drug Stereochemistry, Analytical Methods and Pharmacology, 2nd edition. New York: Marcel Dekker. Walle, T., Webb, J.G., Bagwell, E.E., Walle, U.K., Daniell, H.B. and Gaffney, T.E. (1988) Stereoselective delivery and actions of beta receptor antagonists. Biochemical Pharmacology 37, 115–124. Williams, K. and Lee, E. (1985) Importance of drug enantiomers in clinical pharmacology. Drugs 30, 333–354.

5. QUANTITATIVE STRUCTUREACTIVITY RELATIONSHIPS AND DRUG DESIGN JOHN C.DEARDEN and †KENNETH C.JAMES CONTENTS 5.1 INTRODUCTION

168

5.2 HYDROPHOBICITY PARAMETERS

169

5.2.1 Hydrophobic (Hansch) substituent constants

170

5.2.2 Hydrophobic fragmental constants

171

5.2.3 Chromatographic hydrophobicity values

172

5.2.4 Aqueous solubility

173

5.3 ELECTRONIC PARAMETERS

173

5.3.1 Hammett constants

174

5.3.2 Inductive substituent constants

177

5.3.3 Taft’s substituent constants

178

5.3.4 Hydrogen bonding parameters

178

5.3.5 Whole molecule parameters

179

5.4 STERIC PARAMETERS 5.4.1 Substituent parameters

179 179

5.4.1.1 Taft’s steric substituent constants

179

5.4.1.2 Van der Waals dimensions

181

5.4.1.3 Charton’s steric constants

181

5.4.1.4 Sterimol parameters

182

5.4.1.5 Molar refractivity

182

Quantitative structure

203

5.4.2 Whole molecule parameters

182

5.4.2.1 Relative molecular mass (RMM)

182

5.4.2.2 Molecular volume

183

5.4.2.3 Surface area

183

5.4.2.4 The kappa index

183

5.4.2.5 Minimal steric difference

183

5.4.2.6 Molecular shape analysis

184

5.4.2.7 Molecular similarity

184

5.4.2.8 3-D parameters 184 5.5 TOPOLOGICAL PARAMETERS 185 5.6 BIOLOGICAL RELATIONSHIPS 186 5.6.1 Ferguson effect 186 5.6.2 Hansch analysis 187 5.6.2.1 Correlation coefficient 189 5.6.2.2 Regression coefficients 189 5.6.2.3 Standard error of the estimate 189 5.6.2.4 Standard deviation of the coefficient 190 5.6.2.5 F values 190 5.6.2.6 Optimal partition coefficient 191 5.6.2.7 The bilinear relationship 192

Introduction to the principles of drug design and action

204

5.6.2.8 Comparison of slopes and intercepts 192 5.6.3 Free-Wilson analysis 192 5.7 SOME LIMITATIONS AND PITFALLS OF QSAR 195 5.8 MULTIVARIATE ANALYSIS 198 5.8.1 Pattern recognition methods 203 5.9 NEURAL NETWORKS 204 5.10 SUMMARY 206 FURTHER READING 207

5.1 INTRODUCTION Medicinal chemists have tried to quantify relationships between chemical structure and biological activity since before the turn of the century. However, it was not until the early 1960s, through the efforts of Corwin Hansch and his co-workers, that a workable methodology was developed and the subject that was to become known as quantitative structure-activity relationships (QSAR) was born. Since then, thousands of research papers, articles and reviews on QSAR have emerged, with unfamiliar symbols and parameters, and with results which are expressed in a format different from that of traditional medicinal chemistry. It is the object of this chapter to explain these methods of expression, what they are meant to convey, and how the technique may be used in drug design. The traditional method of searching for new medicinal compounds has sometimes been described as chemical roulette. A chemical structure, known to have a particular biological activity, is chosen, and attempts are made to improve it by modifications based on chemical intuition and isosteric considerations (see Section 5.2), until a highly active compound with minimal side-effects is produced. A plan of the probable receptor site is built up as the number of compounds synthesized and tested increases, and the selection of further new compounds becomes progressively more rational. Beckett’s work on analgesics is a classical example of this procedure. By carefully choosing his compounds, he was able to chart a map of the analgesic receptor site (since modified), which is reproduced in Figure 5.1. It can be seen that there is a

Quantitative structure

205

hollow which will accommodate a protruding group, a flat area which will fit a similar flat surface, and a negatively charged site. Methadone ((5.1) X=H) will fit this receptor; it also has a phenyl group which can lie

Figure 5.1 Analgesic receptor site (as proposed by Beckett, A.H. (1956) Analgesics and their antagonists: some steric and chemical considerations. Part 1. The dissociation constants of some tertiary amines and synthetic analgesics; the conformation of methadone compounds. Journal of Pharmacy and Pharmacology 8, 848–859. on the flat surface, and an alkyl chain which will occupy the hollow. Using this approach, one is able to anticipate the shapes of biologically active molecules, and speculate on the types and positions of groups which will bring about the optimal stereochemistry required for activity. Molecular mapping of receptor sites is now carried out with the aid of computer graphics (see Chapter 3). The quantitative structure-activity approach uses parameters which have been assigned to the various chemical groups that can be used to modify the structure of a drug. The parameter is a measure of the potential contributions of its group to a particular property of the parent drug. In the present situation, a steric parameter, which assesses the bulkiness of the group occupying the hollow on the drug receptor, would be appropriate. In a typical procedure, a series of related compounds are examined, and the relevant parameters of their substituent groups compared with the biological activities of the compounds and then, by mathematical procedures, the structures of the most promising derivatives are predicted. Parameters governing several different properties can be employed, but the three most commonly used are steric and electronic parameters and parameters related to partitioning.

Introduction to the principles of drug design and action

206

5.2 HYDROPHOBICITY PARAMETERS Drugs move through an organism, from the site of administration to the site of action, largely by a process of partitioning through lipid membranes. It follows that the partition coefficient (P) of a drug greatly affects the rate at which it reaches the site of action. (It should be remembered that although P is an equilibrium constant, it is defined from the law of mass action as the ratio of the forward and reverse partitioning rate constants.) As P is logarithmically related to free energy, it should be possible to split log P into parameters characteristic of the chemical groups that make up the molecule. Most of the partitioning work in quantitative structure-activity relationships has been based on the 1-octanol-water system. This is because 1-octanol is considered to be a reasonable model of a lipid, in that it has a polar head-group and a long hydrocarbon chain. 5.2.1 Hydrophobic (Hansch) substituent constants The difference in log P between a compound containing a substituent group X and the substituted parent compound (X=H) was defined by Hansch as the hydrophobic substituent constant π, i.e. π=log PX−log PH=log(PX/PH). The subscript H represents the unsubstituted compound and the subscript X represents the derivative in which hydrogen has been replaced by the group X. Values can be used to calculate 1-octanol-water partition coefficients in the same way as Hammett constants can be used to estimate dissociation constants. Thus, the log P value of butan-2-one between 1-octanol and water is 0.32, therefore the log P for . hexan-2-one in the same system should be Values may also be correlated directly with biological activities to give a quantitative structure-activity relationship (QSAR), as will be discussed later. Collander showed that partition coefficients in one solvent system (P1) are related to those in another (P2) by:

(5.1) where k1 and k2 are constants. Partition coefficients for an extensive range of compounds can be found in the literature, together with values of k1 and k2, to convert from one solvent system to another, using Equation [5.1]. It should, however, be noted that the Collander equation holds best when the two solvent systems are similar in nature. Hansch constants for groups attached to aromatic nuclei fall roughly into three categories, depending on the nature of the group already in the ring, and these are: (i) strongly electron-donating groups, (ii) strongly electron-withdrawing groups and (iii) groups lying between these two extremes.

Quantitative structure

207

The differences involved can be seen from Table 5.1, which shows a selection of Hansch substituent constants. It has been recommended that for structure-activity relationships involving substitution on an aromatic ring, π constants based on phenoxyacetic acid should be used for systems in the third group, and constants based on phenol should be used for those in the first group. As an alternative to π values, log P values can be correlated with biological activities. When values are not available, log P values can be determined experimentally, or when the solubility is considerably higher in one solvent than in the other, can be estimated as the ratio of the solubilities in the individual solvents (although this is not acceptable if there is self-association of the compound in either solvent). When π values are not available, they can be determined experimentally. It might be thought that by summation of π values of all the groups in a molecule, the total log P value could be calculated. This is not so, for two reasons. Firstly, π values are substituent

Table 5.1 Hansch substituent constants. (1) Groups attached to non-conjugated systems Group π Group π -OH −1.16 -OH −1.39 (primary) (secondary) -OH −1.43 -OCH3 −0.47 (tertiary) -Cl 0.39 -Br 0.60 −1.19 -I 1.00 -NH2 -COCH3 −0.71 -NO2 −0.85 -CH3, 0.52 CH2, -CH (2) Groups conjugated to aromatic systems Occupying group -COOH -OH -NO2 Entering CH2OH group OCH2COOH CH2COOH -H 0.00 0.00 0.00 0.00 0.00 0.00 3-Cl 0.76 0.68 0.83 0.84 1.04 0.61 4-Cl 0.70 0.70 0.87 0.86 0.93 0.54 0.51 0.49 0.52 0.50 0.56 0.57 3-CH3 4-CH3 0.52 0.45 0.42 0.48 0.48 0.52 3-OH −0.49 −0.52 −0.38 −0.61 −0.66 0.15 4-OH −0.61 – −0.30 −0.85 −0.87 0.11 3-OCH3 0.12 0.04 0.14 – 0.12 0.31 −0.04 0.01 0.08 0.00 −0.12 0.18 4-OCH3

Introduction to the principles of drug design and action

208

3-NO2 0.11 −0.01 −0.05 0.11 0.54 −0.36 4-NO2 0.24 −0.04 0.02 0.16 0.50 −0.39 More comprehensive lists of Hansch substituents constants can be found in: Tute, M.S. (1971) Principles and practice of Hansch analysis: A guide to structure-activity correlation for the medicinal chemist. In Advances in Drug Research, edited by N.J. Harper and A.B.Simmonds, Vol. 6, pp. 1–77. London: Academic Press; Hansch, C. and Leo, A. (1979) Substituent Constants for Correlation Analysis in Chemistry and Biology. New York: John Wiley. constants and must be added to the log P value of a parent compound. Secondly, and more importantly, there is an inherent flaw in π values, in that the values for, say, CH3 and -CH2 are the same. In other words, it is assumed that the log P value of hydrogen is zero, which is incorrect. Consequently, whilst the use of π is perfectly acceptable as a measure of substituent hydrophobicity, it should not be assumed that values are additive. 5.2.2 Hydrophobic fragmental constants In 1974 Rekker obtained hydrophobic constants (f) for a large number of molecular fragments by breaking down the octanol-water log P values of a large number of compounds. These were found to have better additivity than π values, although correction factors were needed for constitutive effects such as proximity of polar groups. Hansch and Leo also devised a set of fragmental constants using a different approach: they measured the log P values of many small molecules (e.g. H2, CH4) and calculated their fragmental constants from these. Both fragmental constant methods of calculating log P are now available in computerised form. A small selection of f values is given in Table 5.2. Using these, the value of log P for n-propanol, CH3(CH2)2OH, is (3×0.20)+(7×0.23)−(2×0.12)−1.64=0.33, which agrees well with the experimental value of 0.34. The result using Hansch constants from Table 5.1 is (0.52×3)−1.16=0.40. This is a simple example; for more complex molecules, numerous corrections have to be included to allow for proximity effects, folding effects, aromaticity, etc. (for details see the first two references cited in Table 5.2). 5.2.3 Chromatographic hydrophobicity values Chromatography is essentially a partitioning process, so Rf values from thin-layer chromatography (TLC), and capacity factors (k) from high-performance liquid chromatography (HPLC), are related to partition coefficients. For reversed phase TLC:

(5.2)

Quantitative structure

209

(5.3) For HPLC,

(5.4) where tc and to are the retention times of the compound and of a non-retained solute respectively.

(5.5) A range of stationary phases can be used, depending on the nature of the compounds to be chromatographed, to give an appropriate range of Rm or log k’ values. For example, a TLC plate may be impregnated with liquid paraffin, and acetone-water mixtures used as

Table 5.2 Fragmentation constantsa. Fragment f 0.20

Fragment

f −1.11

0.23 −1.54 −1.64 −1.82 For hydrocarbon chains, 0.12 (n−1) is subtracted, where n is the number of bonds between carbons and between carbon and hetero atoms excepting hydrogen. a Rekker, R.F. (1977) The Hydrophobic Fragment Constant, pp. 39–106. Amsterdam: Elsevier. A comprehensive account of fragmental constants can be found in: Hansch, C. and Leo, A. (1979) Substituent Constants for Correlation Analysis in Chemistry and Biology. New York: John Wiley. James, K.C. (1986) Solubility and Related Properties. New York: Marcel Dekker. the mobile phase; Rf values can then be extrapolated to zero acetone concentration. HPLC stationary phases can be coated with octanol, or chemically bonded with a range of chemicals.

Introduction to the principles of drug design and action

210

Chromatographic methods generally do not cover a very wide range of hydrophobicity, but have the advantage that compound purity is not so crucial as it is in direct measurement of partition coefficient. 5.2.4 Aqueous solubility Generally, as partition coefficient increases, aqueous solubility decreases, although the relationship is not all that simple because of factors such as the entropy of melting, which partly controls solubility. Nonetheless, aqueous solubility can be used as a measure of hydrophobicity (or rather of hydrophilicity). It should also be noted, as mentioned in Section 5.2.1, that the ratio of solubilities in octanol and water gives a close approximation of the partition coefficient; values are not identical because of concentration effects and mutual solubility of the solvents. 5.3 ELECTRONIC PARAMETERS The negatively charged site on the analgesic receptor described previously suggests that an electron-deficient group on a potential analgesic molecule, positioned so that it will come into contact with the negative site, will help the molecule bind to the receptor. The electron-deficient centre in methadone is provided by the protonated amine group. If the electron-density on the amine group is decreased, its electrostatic attraction for the receptor will become stronger. This can be achieved by attaching an electron-withdrawing group, such as chlorine ((5.1) X=Cl) to the amine group, while an electron-donating group, such as methoxy ((5.1) X=OCH3) will have the opposite effect. Considerations of this sort approach drug design in only a qualitative manner. It is more effective to quantify these qualities; electronic parameters perform this function by giving a value which is a measure of the degree of electron-donating or electron-withdrawing power. The best known electronic parameter is the Hammett substituent constant.

5.3.1 Hammett constants

Quantitative structure

211

In 1940, Hammett introduced his substituent constants to predict equilibrium constants and rate constants for chemical reactions. He reasoned that an electron-withdrawing group, attached to the aromatic ring of benzoic acid, would increase the acid strength of the carboxyl group, and the greater the electron-withdrawing power, the greater the increase in strength. He was therefore able to assign substituent constants (σ) to groups according to their influence on the acid strength of benzoic acid. Hammett’s substituent constant is defined by:

(5.6) Ko represents the dissociation constant of benzoic acid ((5.2) X=H), and Kx that of benzoic acid substituted by the group X. More conveniently, σ can be expressed in terms of Equation [5.7].

(5.7) Thus, considering benzoic acid which has a pKa value of 4.19, and p-toluic acid ((5.2) X =p-CH3) which has a pKa value of 4.36, the change in acid strength brought about by the methyl group (σp–CH3) is equal to 4.19−4.36=−0.17. A small selection of Hammett substituent constants is given in Table 5.3, from which it can be seen that electron-withdrawing groups have positive values, electron-donating groups have negative values, and hydrogen has a value of zero. Scrutiny of Table 5.3 now shows that the analgesic activities of methadone analogues should increase in the order X=OCH3
Introduction to the principles of drug design and action

212

Table 5.3 Electronic substituent constants. Group Hammett Inductive Taft Swain-Lupton constantsa constantsb constantsc constantsd σm σp σ1 σ* F R -H 0.00 0.00 0.00 0.49 0.00 0.00 -CH3 −0.07 −0.17 −0.05 0.00 −0.04 −0.13 −0.05 −0.10 −0.05 −0.10 -C2H5 −0.07 −0.15 -Cl 0.37 0.23 0.47 — 0.41 −0.15 -Br 0.39 0.27 0.45 — 0.44 −0.17 -I 0.35 0.30 0.39 — 0.40 −0.19 -NO2 0.71 0.78 — — 0.67 0.16 -OH 0.12 −0.37 0.25 — 0.29 −0.64 0.12 −0.27 0.25 — 0.26 −0.51 OCH3 -C6H5 0.06 −0.01 0.10 0.60 0.08 −0.08 a Hammett, L.P. (1940) Physical Organic Chemistry, p. 186. New York: McGraw Hill. b Charton, M. (1964) Definition of ‘inductive’ substituent constants. Journal of Organic Chemistry 29, 1222–1227. c Taft, R.W. (1956) Separation of polar, steric and resonance effects. Steric Effects in Organic Chemistry, edited by M.S.Newman, pp. 559– 675. New York: John Wiley. d Hansch, C. and Leo, A. (1979) Substituent Constants for Correlation Analysis in Chemistry and Biology. New York: John Wiley. More comprehensive lists of electronic substituent constants can be found in Hansen and Leo (above) and Tute, M.S. (1971) Principles and practice of Hansch analysis: A guide to structure-activity correlation for the medicinal chemist. In Advances in Drug Research, edited by N.J.Harper and A.B.Simmonds, Vol. 6, pp. 1–77. London: Academic Press.

Quantitative structure

213

Fukata and Metcalf’s results can be expressed in the form of Equation [5.8], which is the equation for the best straight line through the points in Figure 5.2.

(5.8) MLD50 is the minimum dose that kills 50% of the flies treated. The equation was calculated using a mathematical process called regression (least squares) analysis. The

Introduction to the principles of drug design and action

214

Figure 5.2 Insecticidal activities of some diethylsubstituted phenyl phosphates. (Results taken from Fukata, T.R. and Metcalf, R.L. (1956) Structure and insecticidal activity of some diethyl-substituted phenyl phosphates. Journal of Agricultural and Food Chemistry 4, 930–935; and Metcalf, R.L. and Fukata, T.R. (1962) Metasulfapentafluorophenyldiethyl phosphate and metasulfapentafluorophenyl-N-methyl carbamate as insecticides and anticholinesterases. Journal of Economic Entomology 55, 340–341. calculation is not difficult, but is rather protracted. It is, therefore, only since the introduction of computers and electronic calculating machines that these so-called regression equations have been used extensively.

Quantitative structure

215

A regression equation is a convenient way of quantitatively expressing a correlation, but on its own it does not give as much information as a graph, because the graph indicates how many results were considered, how scattered they were around the best fitting line, and how representative of the results the line is. Additional data should therefore be given, the basic minimum being shown in Equation [5.9], which is laid out in the conventional manner. k1 is

(5.9) the intercept, and k2 the coefficient for σ. The correlation coefficient is a number which varies from zero to 1. The higher the number, the better the correlation. What constitutes a satisfactory correlation coefficient depends on the number of results; the greater the number of results, the lower the acceptable correlation coefficient. In quantitative structure-activity relationships, a figure in excess of 0.9 is aimed for. A useful feature of the correlation-coefficient is that it is the square root of the explained variation; for example, a relationship having a correlation coefficient of 0.990 explains 0.9902×100= 98% of the variation between results. Additional statistical information, in particular the F value for the equation, is sometimes quoted. The F value is a numerical indicator of whether or not the relationship expressed by the equation is coincidental; the higher the value, the less likely the relationship is due to chance. The format for expressing F distribution and other data is shown in Equation [5.27], and the interpretation explained in Section 5.6.2.5). The reason for using the logarithm of biological response in Figure 5.2 and Equation [5.8] has thermodynamic origins. The free energy of a transition involving a given molecule is assumed to be the sum of the free energies of its substituent groups. Thus for example, the excess free energy of ionization of p-toluic acid ((5.2) X=pCH3) over that of benzoic acid is equal to the contribution of the p-methyl group. Equation [5.6] uses log (Kx/Ko) instead of free energy because equilibrium constants are logarithmically related to free energy (ΔG) through the van’t Hoff equation [5.10] in which R is the gas constant and T is the temperature. Log (Kx/Ko) and are therefore also additive. Because of this direct relationship, Hammett’s and equivalent equations are said to be linear free energy relationships (LFER). It is therefore logical that the logarithms of biological parameters should also be used in quantitative structureactivity relationships.

(5.10)

Introduction to the principles of drug design and action

216

Hammett substituent constants can be used only for nuclear aromatic substituents and their effects upon side-chain groups in the meta or para position to them. They cannot be used for ortho substituents because of short-range effects such as steric hindrance and intramolecular hydrogen bonding. Numerous other electronic substituent constants have been introduced since Hammett’s original work, many of which have been used in quantitative structure-activity relationships, but only two of them have been used to any great extent. These are the inductive substituent constant and the Taft substituent constant. For information on other constants the reader is referred to the more comprehensive treatises listed at the end of the chapter. 5.3.2 Inductive substituent constants Hammett substituent constants are a measure of both inductive and mesomeric effects. The p-substituent constant (σp) has a greater resonance component than the equivalent meta constant (σm), and the inductive contribution can be calculated from Equation [5.11].

(5.11) σ1 is the inductive substituent constant, and can be used in aliphatic compounds in which the influencing and influenced groups do not form part of a conjugated system. Inductive substituent constants have also been obtained from the dissociation constants of 4-substituted bicyclo(2,2,2)octane carboxylic acids (5.4), and αsubstituted acetic acids. A

small selection of inductive substituent constants is given in Table 5.3. More comprehensive lists can be found in the references cited below the Table and at the end of the chapter. Another set of constants is that devised by Swain and Lupton, who factored the Hammett constant into its inductive and resonance components, and a selection of these is listed in Table 5.3. 5.3.3 Taft’s substituent constants

Quantitative structure

217

Taft’s substituent constants (σ*) are a measure of the polar effects of substituents in aliphatic compounds when the group in question does not form part of a conjugated system. They are based on the hydrolysis of esters and are calculated from Equation [5.12], where k represents the rate constant for the hydrolysis of the substituted compound, and k0 that of the methyl derivative.

(5.12) The bracketed term with the subscript B represents basic hydrolysis and the other, with the subscript A, acid hydrolysis. The factor 2.48 brings the constants on to the same scale as the Hammett constants. The equation depends on the fact that although both basic and acid hydrolysis are sensitive to steric effects, only basic hydrolysis is influenced by polar effects; hence, by subtracting the acid term from the basic term, only the polar effect remains. A limited list of Taft substituents is given in Table 5.3; larger compilations can be found in the literature. Taft substituent constants are different from the others in that methyl, rather than hydrogen, is the standard group, for which the constant is zero. However, they can be compared with other constants by writing the methyl group in the form, CH2-H and identifying it as the group for H. Another substituent constant representing a group X can then be compared by using the Taft constant for CH2-X. Under these circumstances, Taft and inductive substituent constants are approximately related by:

(5.13) 5.3.4 Hydrogen bonding parameters Hydrogen bonding is an important property in drug activity; it contributes to solubility, partitioning and receptor binding. It is, however, very difficult to quantify, and consequently its use in QSARs is usually as an indicator variable. That is, a substituent (or molecule) capable of hydrogen bonding is ascribed a value of 1, and a substituent (or molecule) incapable of hydrogen bonding takes a value of 0. Hydrogen bonding ability is often split into H-bond donor (HD) and H-bond acceptor (HA) ability. Thus -OH, -NH2 and -COOH would have HD values of 1 and HA values of 1, whilst -OCH3, NMe2 and -COOMe would have HD values of 0, but would have HA values of 1. It is generally accepted that halogens are not capable of hydrogen bonding. Hansch and Leo (1979) Substituent Constants for Correlation Analysis in Chemistry and Biology, pp. 49–54, New York: John Wiley) list these values for a wide range of substituents. A number of attempts have been made to devise more quantitative measures of H-bonding ability; to date the best are the solvatochromic parameters of Abraham and co-workers, but these require experimental determination.

Introduction to the principles of drug design and action

218

5.3.5 Whole molecule parameters If the compounds being studied do not have a common parent, or if substituent parameters are not available for some of the substituents, then it is necessary to use whole molecule parameters in QSARs. A further reason is that some parameters are available only as whole molecule parameters. Dipole moment is a common example. It might be thought that the paucity of measured dipole moments would be restrictive, but dipole moments can now readily be calculated using a molecular orbital (MO) package such as MOPAC. Indeed, the availability of such software has meant that a large number of MO parameters can be easily and rapidly calculated. Among those in wide use in QSAR are the energies of the highest occupied molecular orbital (a measure of the ease of electron loss) and of the lowest unoccupied molecular orbital (a measure of the ease of electron gain). Atomic charges are not whole molecule parameters, but they reflect the whole molecule, and cannot be classed as substituent parameters. They, too, are used extensively in QSAR correlations. Others that have been used are superdelocalisabilities and frontier electron densities. Kier and Hall have recently devised an electronic atom index based on molecular topology; they term this the electrotopological index. It is based on counts of valence and sigma electrons for each atom, and takes account also of the influence of other atoms. It has been shown to correlate well with biological activities, and (since it is an atom index) highlights the importance of specific atoms in controlling the activity. 5.4 STERIC PARAMETERS Steric parameters fall into two broad classes—those modelling bulk and those modelling shape. The former are more plentiful and easier to calculate, but are on the whole less useful and informative. Again, we can have both whole molecule and substituent parameters. 5.4.1 Substituent parameters 5.4.1.1 Taft’s steric substituent constants This constant (Es) is a corollary of Equation [5.12]. It depends on the fact that acid hydrolysis is determined almost completely by steric factors, and is defined by Equation [5.14].

(5.14) Es values have been used to examine the effects of substituents X on the antihistamine activities (the biological response, Rb) of some analogues of diphenhydramine (5.5), and Equations [5.15] and [5.16] derived.

Quantitative structure

219

(5.15)

(5.16) The standard errors and correlation coefficients clearly show that the important factor is stereochemical, since the observed results are nearly twice as scattered about the line represented by Equation [5.16], in comparison with Equation [5.15], and Equation [5.15] explains 78% of the variation, while the correlation coefficient of Equation [5.16] is unacceptably low. It can be argued that Equation [5.16] does explain 40% of the variation, and that an improved correlation between log Rb and a combination of the two substituent constants is probable. The three variables can be correlated by multiple regression analysis, and the result expressed in Equation [5.17].

(5.17) The arithmetic is similar to, but more complex than, simple linear regression and can be handled by most computers. Equation [5.17] cannot be expressed on graph paper, but can be represented by a three-dimensional model. When the number of variables exceeds 3, the results cannot be expressed in the form of either graph or model, since all dimensions will have been exhausted. A regression equation is, therefore, the only method of expression which can be used in such situations. Equation [5.17] confirms that the effect is predominantly steric, since introduction of σ has produced only an insignificant improvement in either correlation coefficient or standard error. Es values are given in Table 5.4. The list is limited by the experimental difficulties in obtaining the physico-chemical data upon which Es values are based.

Table 5.4 Taft steric parametersa.

Introduction to the principles of drug design and action

Substituent

Es 0.00 −0.19

Substituent

220

Es −0.07 −0.27

C6H5CH2 −0.38 −0.39 i-C4H9 −0.93 −1.54 c-C6H11 −0.79 a Taft, R.W. (1956) Separation of polar, steric and resonance effects. In Steric Effects in Organic Chemistry, edited by M.S.Newman, pp. 559–675. New York: John Wiley. More comprehensive lists of Taft steric parameters can be found in: Tute, M.S. (1971) Principles and practice of Hansch analysis: A guide to structure activity correlation for the medicinal chemist. In Advances in Drug Research, edited by N.J.Harper and A.B.Simmonds, Vol. 6, pp. 1–77. London: Academic Press. Hansch, C. and Leo, A. (1979) Substituent Constants for Correlation Analysis in Chemistry and Biology. New York: John Wiley. 5.4.1.2 Van der Waals dimensions Van der Waals volume (Vw) and radius (rv) represent the actual dimensions of the group. Since chemical groups are rarely symmetrical, the van der Waals radius depends on the axis along which it is measured, and three types are defined, rv(min), the minimum radius, rv(max), the maximum radius, and rv , which is the distance the group protrudes from the bulk of the parent molecule. Van der Waals radii can be correlated with biological results in the same ways as Es, to which they are linearly related. rv(min) is usually preferred because groups are expected to take up positions which will minimize the degree of steric interaction. Sometimes the mean of the three radii (rv(av)) is used. McGowan and Mellors introduced the characteristic volume Vx, the molar volume at absolute zero, and provided atomic and bond contributions for its evaluation (see Section 5.4.2.2). 5.4.1.3 Charton’s steric constants The principal problem with van der Waals radii and Taft’s Es values is the limited number of groups for which these constants have been determined. Charton introduced a corrected van der Waals radius U, in which the minimum value of van der Waals radius of the substituent group (rv(min)) is corrected for the corresponding radius for hydrogen (rvH), as defined by Equation [5.18].

(5.18)

Quantitative structure

221

They were shown to be a good measure of steric effect by correlation with Es values, and were also found to be rectilinearly related to the rates of esterification of substituted carboxylic acids with methanol and ethanol. Since there were ample data for these reactions, Charton was able to extend his list of constants to 62, by substituting rate constants into the regression equations linking esterification rates to U. The large number of constants available makes this a useful source of steric parameters for QSAR studies. 5.4.1.4 Sterimol parameters A criticism of the steric parameters described so far is that they represent only one aspect of the shape of the group. Thus for example, van der Waals volume represents the total volume of the group, and rv the width along one plane. Sterimol parameters were developed to overcome this weakness. Each chemical group is allocated 5 Sterimol parameters: L, which is the distance the group protrudes from the parent molecule, and B1−B4, which give the widths of the group in four directions, 90° to each other and perpendicular to the axis along which L is measured. Cross sectional dimensions increase from B1 to B4. The procedure is shown diagrammatically in Figure 5.3. 5.4.1.5 Molar refractivity Molar refractivitiy (MR) is defined as:

(5.19) where n is refractive index, M is relative molecular mass and ρ is density. Its units are those of molar volume, and thus it may be regarded as a measure of bulk, although the constitutive component of polarisability is also present, which leads some workers to regard it as a measure of weak electronic interactions. Although it is measured or calculated as a whole molecule parameter, it is strictly additive, and hence substituent MR values are available (see Hansch, C. and Leo, A. (1979) Substituent Constants for Correlation Analysis in Chemistry and Biology, New York: John Wiley.) 5.4.2 Whole molecule parameters 5.4.2.1 Relative molecular mass (RMM) RMM or molecular weight is perhaps the simplest of all steric parameters; it is certainly the easiest to calculate, and has been widely used in QSAR. It may be noted that it is often used, together sometimes with log P, in correlating penetration rates

Introduction to the principles of drug design and action

222

through membranes, which suggests a size restriction on molecules passing through pores.

Figure 5.3 Sterimol parameters. 5.4.2.2 Molecular volume Molar volume is defined as the volume occupied by one mole of a pure compound; it has an electronic component, since strong intermolecular attraction can hold molecules more closely together. Because of this, it requires experimental measurement. Molecular volume can be calculated by summing the van der Waals volumes of the constituent atoms (obtained from the van der Waals radii). A more rigorous method is to use a computer program that rolls a water molecule over the molecular surface defined by the van der Waals radii, to give a cavity surface volume. Perhaps the simplest way of calculating molecular volume accurately is to use the McGowan and Mellors characteristic volume method, which simply sums atomic and bond contributions as follows: C 16.35, H 8.71, O 12.43, N 14.39, F 10.48, Cl 20.95, Br 26.21, I 34.53, S 22.91, P 24.87; for each bond, irrespective of type, subtract 6.56. Thus for NH2COCH3 the value is 2×16.35+12.43+14.39+5×8.71−8×6.56= 50.59 cm3 mol−1. 5.4.2.3 Surface area The computer program mentioned in Section 5.4.2.2, that rolls a water molecule over the molecular surface, can be used to generate an accessible surface area, a parameter widely used in QSAR studies, since it is molecular surfaces that come into contact with solvent and receptor. It can be particularly effective if the contribution of hydrophobic and hydrophilic surface areas can be distinguished.

Quantitative structure

223

5.4.2.4 The kappa index Introduced by Kier in 1985, this parameter is derived from the number of 2-bond fragments (e.g. C-C-C) in a non-hydrogen molecular skeleton. Linear molecules tend to have higher kappa values, as is shown by the values for the isomeric hexanes:

n-Hexane 2-Methylpentane 3-Methylpentane 2,3-Dimethylbutane 2,2-Dimethylbutane

5.000 3.200 3.200 2.222 1.633

The kappa index is thus a shape parameter. It is readily calculated, and requires no experimental measurement. 5.4.2.5 Minimal steric difference This parameter assesses the difference between molecules in terms of the parts which do not overlap when one chemical formula is placed on top of the other. If for example, piperidine (5.6) is compared with pyrrolidine (5.7), the methylene group, surrounded by the dotted circle, will determine the MSD, since this is the only portion which does not overlap. The rules of the calculation are as follows: (i) hydrogen atoms are ignored, (ii) elements in the second period of the Periodic Table have a weighting of 1 (e.g. C), (iii) elements in the third period have a weighting of 1.5 (e.g. S), (iv) elements in higher periods have a weighting of 2 (e.g. Br).

Thus the MSD between piperidine and pyrrolidone is 1, and that between pyrrolidine and indole (5.8) is 4. 5.4.2.6 Molecular shape analysis Devised by Hopfinger in 1980, this method uses a similar principle to that involved in the minimal steric difference method. It uses the concept of common overlap volume between a reference compound (usually the most active) and the other compounds in a

Introduction to the principles of drug design and action

224

series; this common overlap volume is considered to represent the steric requirements of the receptor. The method in effect uses the steric similarity of a set of molecules as a parameter. 5.4.2.7 Molecular similarity A third parameter using the concept of similarity has been developed by Richards, and is incorporated in the TSAR (Tools for Structure-Activity Relationships) software. Based on equations derived by Carbo and by Hodgkin, it allows comparison of the similarity of a set of molecules to a standard (e.g. the most active in a series) on the basis of either electrostatic potential or steric parameters. Although quite new in concept, it is already finding wide application in QSAR analysis. A number of other methods of determining molecular similarity have recently been developed, and are finding use in, for example, the searching of data-bases for the screening of compounds for specified types of drug action. 5.4.2.8 3-D parameters Most of the classical QSAR parameters, with the exception of those that model shape, take no account of conformation or of the fact that most molecules are threedimensional. Nevertheless, since a significant contribution to a molecule’s biological activity arises from its fit and binding to a receptor, molecular three-dimensionality is clearly important. In recent years, therefore, much effort has gone into the examination and development of parameters that reflect that three-dimensionality. Such parameters can be as simple as inter-atomic distances or torsion angles or as complex as the distribution of electrostatic potential around a molecule. One approach that has aroused much interest is that known as CoMFA (comparative molecular field analysis). This involves firstly superimposing the molecules to be studied, within a three-dimensional grid or lattice. This is a simple procedure for most congeneric series, whereby the common features of the molecules can readily be superimposed. For non-congeneric series, with no obvious common features, alignment is much more difficult and more subjective. Hence most CoMFA studies to date have been concerned with congeneric series. A probe atom is then placed at each lattice point in turn, and the steric (Lennard-Jones) and electrostatic (Coulombic) fields exerted by each molecule at each lattice point are then calculated. This results in a large number of data-points, and partial least squares (PLS) statistics is used to determine the minimal set of data-points necessary to distinguish the set of compounds according to their biological activities. The PLS model then has to be cross-validated, for example by the leave-one-out method. If necessary and appropriate, re-alignment of the most poorly predicted compounds can then be carried out, and the above steps repeated. The contoured QSAR coefficients can then be displayed to allow visualisation of regions where electrostatic and/or steric fields have the greatest effect on activity.

Quantitative structure

225

5.5 TOPOLOGICAL PARAMETERS Graph theory is that branch of chemistry dealing with molecular topology, since a molecular structure is described as a graph. Graph theory is particularly concerned with the way atoms are connected in a molecule, and many attempts have been made to relate topology to molecular properties. Of these approaches, the most successful is undoubtedly that of Kier and Hall, who developed a series of topological parameters called molecular connectivities (mχ) from an original concept of Randic. The superscript m denotes the order of the parameter. Zero order connectivity (0χ) is the simplest and is defined by Equation [5.20],

(5.20) where δi is a number assigned to each non-hydrogen atom, reflecting the number of non-hydrogen atoms bonded to it. Thus for 1-butane (5.9),

The first order connectivity (1χ) is derived for each bond by calculating the product of the numbers associated with the two atoms of the bond. The reciprocal of the square root of this number is the bond value. Bond values are summed to give the first order connectivity for the molecule, so that the value of 1-butane is,

The 1χ value for 2-butane is similarly calculated to be 1.732.

Higher order connectivities are calculated by multiplying δi values across appropriate numbers of bonds.

Introduction to the principles of drug design and action

226

Heteroatoms are accounted for by use of the so-called valence-corrected molecular connectivities (mχv). The physical significance of molecular connectivities is not easy to comprehend. They contain much steric information, as the above examples of 1butane and 2-butane show, but are perhaps best regarded as indicators of molecular complexity. A further problem with the use of molecular connectivities in drug design QSAR is that there is no easy way of translating a given mχ value into a molecular feature, as one can readily do with, say, log P. Recent published work is, however, addressing this problem. Another type of topological parameter is information content. Given a molecular graph (i.e. typically the hydrogen-suppressed skeletal molecular structure), an appropriate set A of n elements is derived, based upon certain structural features. The set A is then partitioned into disjoint subsets Ai of order ni; pi is the probability that a selected element of A will occur in the ith subset, and is equal to ni/n. The mean information content IC of an element of A is then defined by Shannon’s relationship:

(5.21) The total information content of the set A is then n times IC. In effect, it is a measure of molecular complexity, and although its derivation may seem somewhat abstruse, it has been found in many instances to be a useful parameter for the correlation of biological activity. Other means of information content have been derived by Basak and co-workers from the Shannon relationship. They have been found useful in, for example, the prediction of molecular similarity. 5.6 BIOLOGICAL RELATIONSHIPS 5.6.1 Ferguson effect Ferguson was probably the first person to connect free energy with biological activity. He abstracted toxicity data from the literature, and noted that in homologous normal aliphatic alcohols, the logarithm of toxic concentration varied with carbon number in the same way as properties which were linearly related to free energy. He suggested that concentration in a body fluid is not a critical factor controlling the biological activity of a drug, and that it is the concentration within the receptor cell which is critical. If the cell contents are in equilibrium with the surrounding body fluid, the partial free energies, or thermodynamic activities, in the two phases will be equal. Ferguson expressed this activity as Ce /Cs for solutions, where Ce is the effective concentration and Cs the solubility. Also, since gas concentrations are measured in terms of their partial pressure, the activities for gases and vapours are given by pe /ps, where pe is the effective vapour pressure and ps the saturated vapour pressure.

Quantitative structure

227

The theory worked well with general anaesthetics. The compounds examined by Ferguson had widely different chemical structures, for example N2O, CHC13, (C2H5)2O, and potencies, the narcotic concentrations being in the range 0.5–100%. However pe /ps remained reasonably constant. It also worked well with the in vitro antimicrobial activities of phenols. However, with homologous series, the thermodynamic activity was frequently found to increase with carbon number. This behaviour shows up as an increase in biological activity as the series is ascended, followed by a sudden abrupt fall, i.e. a cut-off point. An attractive explanation is that at the point at which biological activity falls, the biologically effective concentration would be greater than the solubility of the homologue. A supersaturated solution would therefore be necessary to achieve an effect. 5.6.2 Hansch analysis Hansch explained the cut-off point in certain homologous series by suggesting that the equilibrium conditions required by Ferguson’s theory were not established. The systems with which Ferguson’s approach was successful (e.g. general anaesthetics) were ones in which equilibrium is quickly established. Two stages were postulated in drug action. (a) A ‘random walk’ from the point of administration to the site of action. This will involve passage through a series of membranes, and is therefore related to partition coefficient. It is expressed mathematically as f(P), a function of the partition coefficient. (b) Attachment to the receptor site expressed mathematically as kx, which depends on (i) the shape of the molecule, and hence on the stereochemistry of its substituent groups, (ii) the electron density on and polarisability of the attachment groups. The methods of quantifying the attachment to the receptor site have already been discussed in the form of electronic and steric substituent constants. Biological activity (Rb) will also be dependent on concentration C, so that the complete relationship is

(5.22) f(P) represents a mathematical function of partition coefficient. The random walk involves passage across hydrophilic barriers and lipophilic barriers. Substances with low aqueous solubilities will be impeded (or, if the solubility is sufficiently low, prevented) from crossing hydrophilic barriers and there is a similar connection between low lipid solubilities and ease of crossing lipid barriers. Somewhere between the two extremes, there will be an optimal balance between hydrophilic and lipophilic properties, so that a plot of hydrophilic-lipophilic nature

Introduction to the principles of drug design and action

228

against the likelihood of the molecule completing the random walk would be expected to take the form shown in Figure 5.4. Hydrophilic-lipophilic nature can be expressed in terms of a lipid-water partition coefficient, so that log P, where P is the 1-octanolwater partition coefficient, can be used as the abscissa scale. The plot approximates to a parabola, for which the general equation is

(5.23) and therefore the random walk can be expressed as

(5.24) or alternatively for a group of related compounds as

(5.25)

Figure 5.4 Parabolic dependency of biological response on octanol-water partition coefficient. where k1, k2 and k3 represent the relevant coefficients. The complete biological process can therefore be fitted into equations of the form

Quantitative structure

229

(5.26) The second and third terms on the right hand side represent the random walk and the fourth and fifth terms, the electronic and steric factors governing the attachment to the receptor site. Any of the other electronic and steric parameters could be substituted for σ and es. Hydrophobic effects may also contribute to receptor binding as well as to transport, and electronic and steric effects can, by affecting metabolism, contribute to the transport stage. The factors which control biological activity can be identified by fitting the experimental data into regression equations. Suppose 17 compounds were submitted to a pharmacological test, and that Equations [5.27–5.31] were obtained when the biological responses were correlated with various combinations of Hammett constants, Taft steric parameters and 1-octanol-water partition coefficients. The numbers in brackets in the equations, immediately preceding log P, σ and Es are the standard deviations of the coefficients; their significance is explained in Section 5.6.2.4.

(5.27)

(5.28)

(5.29)

(5.30)

(5.31)

Introduction to the principles of drug design and action

230

Much information can be derived from these equations, as explained below. 5.6.2.1 Correlation coefficient The correlation coefficient of Equation [5.27], because it is close to 1.00, indicates that the relationship represents the experimental results reasonably well, and explains 0.9572 ×100=91.6% of the variation. However, if the steric and electronic parameters are omitted, to give Equation [5.28], the new equation still has a good correlation coefficient. It is doubtful whether the correlation coefficient of Equation [5.27] is significantly better, since Equation [5.27] contains two extra variables. If there were 17 variables for example, r would equal 1.000, irrespective of the data. The strongest evidence comes from the correlation coefficients of Equations [5.30] and [5.31], which are very low, indicating that neither σ nor Es contributes to biological activity. Furthermore, the significantly lower correlation coefficient of Equation [5.29] in comparison with that of Equation [5.28] suggests that the relationship between log Rb and log P is binomial. One of the problems of using the correlation coefficient r as a measure of goodness of fit is that inclusion of more parameters in the equation, be they relevant or not, will always increase r. However, most statistical packages now include the calculation of an r2 value (r2(adj.)) adjusted for the degrees of freedom in the correlation. In the example above, r2(adj.) for Equation [5.27] is 0.930, whereas that for Equation [5.28] is 0.934. This clearly shows that the two additional terms in Equation [5.27] do not improve the correlation. 5.6.2.2 Regression coefficients The coefficients of the variables give support to the evidence given by the correlation coefficients. The coefficients of σ and Es in Equations [5.30] and [5.31] are small, and in comparison with the intercepts and coefficients of log P and (log P)2, suggest that Equations [5.30] and [5.31] represent plots in which the regression lines would be almost parallel with the σ and Es axes. The larger coefficients in log P in Equations [5.27] and [5.28] support the conclusions given by the correlation coefficients, that biological activity is dependent on the hydrophilic-lipophilic nature of the compounds under test. The lower coefficients in (log P)2 might give the impression that the squared term is not important; for example, the coefficient of (log P)2 in Equation [5.27] is only 0.731, in comparison with 2.342 with log P. However, it must be remembered that (log P)2 is usually bigger than log P. Thus if P is of the order of 1000, (log P)2=3 log P, and one would anticipate a correspondingly larger coefficient for log P, as is the case in Equations [5.27] and [5.28]. A similar pitfall occurs when parameters of considerably different magnitude are compared; for example Equation [5.32], in which V is molar volume (molecular weight/density) in cm3 mol−1, suggests that molecular size is not a controlling factor. However, molar volumes are of the order of 102 cm3, while σ has a value less than 1.00. In this light the coefficients are comparable. It is preferable that parameters should be scaled to roughly the same numerical values. This

Quantitative structure

231

has the advantage that coefficients can be readily compared, and also improves the stability of the statistical analysis.

(5.32) 5.6.2.3 Standard error of the estimate The standard error should be as low as possible. Generally standard error decreases as the correlation coefficient increases. The values given in Equations [5.27] to [5.31] support the conclusion drawn from the correlation coefficients and regression coefficients. 5.6.2.4 Standard deviation of the coefficient The figures in brackets following the coefficients represent the standard deviation of the coefficient, which means that if the experiment is repeated the coefficient should lie between these limits; for example, the coefficient for σ in Equation [5.27] should be 0.0361±0.0190. Obviously the higher the standard deviation, the less reliable is the coefficient, and the less is the likelihood that the variable it represents is related to the biological response. The confidence in the term can be assessed by dividing the coefficient by the standard deviation. Thus for the second term on the right hand side of Equation [5.27], the ratio is 2.342/0.105=22.3, which, because it is a high number, suggests that the term is important. When there is doubt whether the ratio can be considered sufficiently high, it may be compared with the limiting Student’s t value. Most statistical text books give tables of these. Some limiting t values are given in Table 5.5, and it can be seen that they depend on the probability level and on the number of degrees of freedom. The probability level is generally taken as 0.05 for QSAR studies. The number of degrees of freedom is (n−m−1), where n is the number of sets of data (17 in the example) and m is the number of variables (4 in Equation [5.27]). is therefore 12, which from Table 5.5 give a t value of 2.179 for a probability of 0.05. Since this is less than 22.3, the term in log P is significant. The same test rejects the σ and Es terms in Equation [5.27]. 5.6.2.5 F values For convenience, F distribution results are given only for Equation [5.27]. The two numbers in the subscript (4 and 12) following the letter F are m and n−m−1, as defined in the previous paragraph, and the 10.9 following is the experimental F value which fits the data. The F value indicates the probability that the equation is a true relationship between the results, and not merely coincidence. If the experimental figure exceeds the limiting value, the relationship is a true one, within the given probability level. Limiting F values can be obtained from statistical tables, from which

Introduction to the principles of drug design and action

232

Table 5.6 has been abstracted. The numbers running along the top represent the first number in the subscript following F, and those running down the left hand side, the second number in the subscript. Thus for Equation [5.27], v1=4 and v2=12 in Table 5.6, giving F=3.26. Table 5.6 is based on the probability of 0.05 (α, 0.05=3.26), therefore there is less than a 1 in 20 chance that the relationship is a coincidence, and hence better than a 19 in 20 chance that the results

Table 5.5 Student t values. Degrees of freedom 12 13 14 15

0.05 2.179 2.160 2.145 2.132

Probability 0.01 3.055 3.012 2.977 2.947

Table 5.6 0.05 Probability points of the F-distribution. v1 1 2 3 4 v2=1 161.4 199.5 215.7 224.6 2 18.5 19.0 19.2 19.2 12 4.75 3.89 3.49 3.26 13 4.67 3.81 3.41 3.18 14 4.60 3.74 3.34 3.11

5 230.2 19.3 3.11 3.03 2.96

are truly related in the manner given. F=10.9 is obviously much greater than 3.26, and it would be of interest to know precisely how good it is. Consultation of a table for 0.001 points of the F distribution gives F4,12α, 0.001=9.63, so that the probability of Equation [5.27] representing a chance relationship is less than 1 in 1000. The mechanism of calculating the statistical parameters of regression, used above, is considered to be outside the scope of this book, which seeks to explain the interpretation, rather than the preparation, of QSAR data. The necessary arithmetic is usually built into the computer program. 5.6.2.6 Optimal partition coefficient An advantage of the parabolic relationship between log Rb and log P is that the biological activity goes through a maximum corresponding to an optimal value of P, which is easily calculated, as shown in the following example. The concentration (HD50) producing hypnosis in 50% of mice by a series of barbiturates (5.10) was found to be related to the octanol-water partition coefficient (P) by:

Quantitative structure

233

(5.33) The change in log 1/HD50 with log P will be zero when the biological effect is at a maximum; therefore the differentiation of log 1/HD50 with respect to log P, and placing the result equal to zero will give the optimal partition coefficient (P0), i.e.

(5.34)

Therefore

Hansch and Fujita suggested that the Ferguson effect was represented by the left hand side of the parabola, where the plot was approximately rectilinear, and that the fall off in biological activity as the homologous series was ascended was a consequence of the change in slope in the region of the maximum. 5.6.2.7 The bilinear relationship In 1976 Kubinyi proposed an alternative to the parabola as a description of biphasic relationships. He observed that many so-called parabolic QSARs appeared to consist rather of straight ascending and descending sections joined by a curvilinear section at the peak. Furthermore the ascending and descending sections often had markedly different slopes. To model this Kubinyi devised his bilinear model:

(5.35)

Introduction to the principles of drug design and action

234

This equation can model even the case in which activity increases and then levels off as log P increases. The equation is now widely used in QSAR, although it is not so easy to use as the quadratic equation, since the β term has to be obtained by iteration. Figure 5.5 shows a data-set which is clearly modelled better by the bilinear equation than by a parabola. 5.6.2.8 Comparison of slopes and intercepts If two sets of compounds are submitted to the same biological test, and yield rectilinear equations having similar slopes, the indication is that they have similar modes of action. Similarly, if two biological tests are applied to the same set of compounds, and are found to yield rectilinear equations with similar slopes, the tests are probably measuring the same response. The compounds giving the greater intercept in the first case are the more potent, and the test giving the greater intercept in the second is more sensitive. Intercepts can be compared only when the equations have similar slopes. 5.6.3 Free-Wilson analysis This is an alternative procedure to Hansch analysis, in that so-called de novo substituent constants based on biological activities are used, rather than physical properties. As one of their examples, Free and Wilson used the antimicrobial activities of some 6-deoxytetracyclines (5.11) against Staphylococcus aureus. The compounds they examined are summarized in Table 5.7, together with their antimicrobial activities. Biological activities can be expressed in terms of the constituent groups in the molecules; for example, Equation [5.36] can be used to describe the antimicrobial activity of the first compound in Table 5.7.

(5.36) µ is the overall average antimicrobial activity for the whole series, and a, b and c and contributions of the groups -R, -X and -Y, respectively. The identity of the terms prefixed

Quantitative structure

235

Figure 5.5 Biphasic relationship fitted better by bilinear equation than by a parabola. by the letters a, b and c can best be explained if it is imagined that the contributions to the total antimicrobial activities made by the groups in position R can be determined experimentally, and are given in column 4 of Table 5.7. a[H] will then be defined by Free-Wilson analysis as the mean of the figures in column 4 involving R=H, minus the means of all the figures in column 4, i.e.

Introduction to the principles of drug design and action

236

(5.37) Similarly a[CH3]=−6.3.

Obviously it is not possible experimentally to determine partial biological activities of this sort, but the calculation given above serves to show that

(5.38) and this relationship applies no matter what numbers are displayed in column 4 of Table 5.7. Similarly,

(5.39) and

(5.40) Table 5.7 Antimicrobial activities of 6-deoxytetracyclines against Staphylococcus aureusa,b. R Supposed X Y Experimental

Quantitative structure

237

Compound H CH3 partial Br Cl NO2 NO2 NH2 CH3CONH biological activity biological activity I — 5.9 — — — — 60 — — 2.2 — — — 21 II III — 1.4 — — — — 15 — — — 50.4 — — 525 IV V — 32.5 — — — — 320 — 29.2 — — — — 275 VI VII — 13.2 — — — — 160 5.2 — — — — 15 VIII — IX — 14.6 — — — — 140 6.0 — — — — 75 X — a Spencer, J.L., Hlavka, J.J., Petisi, J., Krazinski, H.M. and Boothe, J.R. (1963) 6-Deoxytetracylines. V. 7,9-Disubstituted products. Journal of Medicinal Chemistry 6, 405–407. b Free, S.M. and Wilson, J.W. (1964) A mathematical contribution to structure-activity studies. Journal of Medicinal Chemistry 7, 395–399. Table 5.7 yields 10 equations analogous to Equation [5.36], with 9 unknowns, µ, and the contributions of R=-H or -CH3, X=-NO2 or -Br or -Cl and Y=-NO2 or -NH2 or CH3CONH-. µ can be equated to the mean experimental response, and three of the remainder can be eliminated through Equations [5.38–5.40], leaving five unknowns. Calculation of the best values to fit the 10 equations can be carried out using a computer, and gives the substituent constants shown in Table 5.8, from which the antimicrobial properties of new compounds can be predicted. Thus for example, if R=CH3, X=-Cl and Y=-NH2, the predicted antimicrobial activity will be 1606/10−112+84+123= 256. 1606 is the total experimental biological activity. The major weakness of this approach is that it can be used only for relationships which are rectilinear. The technique has been extended to parabolic relationships by introducing terms representing interactions between substituent groups, and by using equations involving both Hansch and Free-Wilson parameters. In recent investigations, activity has been replaced by log activity, which is related to free energy, and therefore additive. Another innovation is that the activity of the unsubstituted compound (in which the substituent is hydrogen) is used as standard, thereby eliminating the need for restricting equations. The Free-Wilson method has another shortcoming, namely that it assumes that the contribution of a substituent to the biological activity is constant. This is often not the case, as Figure 5.4 shows.

Introduction to the principles of drug design and action

238

5.7 SOME LIMITATIONS AND PITFALLS OF QSAR The concept of quantitative structure-activity relationships has been continually under fire since it was first introduced, but most of the shots have been directed against the ways in which the technique has been used, rather than against the overall idea. QSAR is dependent on the accuracy of the biological results which, by their nature, are susceptible to considerable experimental error. There is therefore a built-in scatter which cannot be explained mathematically. The true relationships can be hidden within this scatter (or alternatively, false correlations can evolve) and their failure to fit as closely as is desirable is blamed on biological variation. Accurate biological data are essential for this technique. Concentration or dose data should always be presented in molar units (e.g. mmol kg−1) otherwise they are not comparable. The success of QSAR predictions is highly dependent on the number of results from which they are derived; the greater the number, the more reliable the correlation. Five biological results for every variable on the right hand side of the correlation equation

Table 5.8 Calculated substituent constants for antimicrobial activities 6deoxytetracyclines against Staphylococcus aureusa. R X Y a[H] 75 b[Cl] 84 c[NH2] 123 a[CH3] −112 b[Br] −16 c[CH3CONH] 18 b[NO2] −26 c[NO2] −218 a Free, S.M. and Wilson, J.W. (1964) A mathematical contribution to structure-activity studies. Journal of Medicinal Chemistry 7, 395–399. are generally regarded as a minimal acceptable level because of the risk of chance correlations, and the more this figure is exceeded the better. Correlations derived from the first results emanating from a structure-activity exercise can change considerably when more results come to hand. Coefficients can change, and physico-chemical parameters which were originally considered significant can cease to be important. Many of the earlier publications on QSAR were based on too few results. A further concern is that, even though a QSAR correlation appears to be extremely good, it may not be able accurately to predict the activities of other compounds. This is because the QSAR has been developed using a certain number of compounds (the training set) having certain molecular characteristics. Unless the compounds whose activities are to be predicted by the QSAR (the test set) have similar molecular characteristics, the prediction will not be accurate. This brings us to two basic tenets of QSAR development and application. Firstly, the compounds used in the training set should span a sufficiently wide range of parameter space; this means (i) that a reasonable range of values of any one parameter should be covered (e.g. log P values from 0 to 5) and (ii) there should be, in the training set, compounds showing variation in all the different types of parameter—

Quantitative structure

239

hydrophobic, electronic and steric. An example of a poor training set in this respect would be an homologous series, with compounds differing only in the length of an alkyl chain. (Sadly, this is still sometimes seen.) Such a training set would have two grave faults; (a) there is little or no variation in electronic parameters as alkyl chain length increases, and (b) hydrophobicity and size of alkyl groups are highly collinear, and so hydrophobic and steric effects could not be differentiated. The second basic tenet arising from the above is that QSAR correlations should not be used outside the range of parameter space of the training set. For example, if a training set covered a log P range of 0–5, the QSAR should not be used to predict the activity of a compound with, say, log P=7. The predictivity of a QSAR correlation should always be checked. A subjective way is to test the ability of the QSAR to predict the activities of the compounds not in the training set, and to use one’s judgment in deciding whether or not the predictions were sufficiently accurate. The most usual way in which this is done is to leave out, by random selection, a number of compounds from the training set and to use those as the test set. Up to 50% of compounds can be left out, provided that the remaining compounds form a reasonable training set, in terms of both the number of compounds and the range of parameter space spanned. A more objective way to assess predictivity is to use a cross-validation technique (such as are available in many statistics software packages) to derive a cross-validated r2 value. A common way in which this is done is by the leave-one-out method: one compound is removed from the data-set and a QSAR correlation obtained for the remaining compounds is used to predict its activity. The compound is then returned to the training set, another one is removed, and a second QSAR correlation is obtained. This procedure is repeated until all the compounds in turn have been removed. The cross-validated r2 value (r2(CV)) is then computed. It will be lower than the ordinary r2 value, but should not be too much lower if the QSAR correlation is to be valid for predictive purposes. (Opinions differ on how much lower is acceptable; however, if a particular QSAR had an r2 value of 0.9, then an r2(CV) value of 0.8 would indicate good predictivity of the equation.) It has been suggested that the results generated in QSAR are by their nature correlated, and that random biological responses can be correlated with physicochemical parameters. Bias of this sort can be demonstrated, but the chance of coincidental relationships with correlation coefficients greater than 0.9 is highly improbable provided that the ratio of data-points to parameters is not less than 5:1. Distributions consisting of two clusters of results, as exemplified by Figure 5.6, can yield good linear correlation coefficients, and yet have no predictive value. In much of the early QSAR work, biological data were correlated successively to each parameter, and combinations of a range of physical parameters, on a trial and error basis. A theory of biological action was then evolved, fitting the parameters which gave the best correlation. Whilst this approach can be justified in separating the influence of steric, electronic and distribution effects, its use in comparing parameters depending on the same basic property usually stretches the technique too far. Thus dipole moments have been successfully correlated with biological results which could not be related to Hammett a constants. In such situations it would be tempting to derive a theory in which dipole moment is a unique factor governing biological

Introduction to the principles of drug design and action

240

activity. However, Hammett constants and dipole moments are closely related. Differentiation between the two properties is therefore not permissible. Similarly, workers have found that one electronic parameter has fitted their data, and others did not. A large number of electronic parameters have been listed, many of which can be expressed as linear parameters of an inductive substituent constant F and a resonance substituent R. Strictly speaking therefore, expressions of the form shown in Equation [5.41] should be employed in seeking quantitative structure-activity relationships, but the procedure is impracticable, even if only for the single consideration that a prohibitively large number of results would be required to make the expression meaningful.

(5.41)

Figure 5.6 Distribution consisting of two clusters. Interrelationships between parameters must be eliminated before a correlation can be considered acceptable, for two reasons. Firstly, the use of highly correlated parameters makes mechanistic interpretation of the QSAR difficult; secondly high intercorrelation can render the statistical analysis unstable. Before an expression of the form of Equation [5.42] can be considered valid, linear regressions should be carried out between each pair of parameters. This is best expressed in the form of a correlation matrix, an example of which is given in Table 5.11.

Quantitative structure

241

(5.42) A correlation matrix is specific to the chemical groups and physical constants upon which it is based, and some combinations give rise to greater collinearity than others. Compounds and parameters used in a QSAR exercise should therefore be chosen with care, e.g. by using similarity clustering. Table 5.9 shows eight cluster sets taken from Hansch and Leo (1979); in planning a QSAR exercise, one should ensure that the selected substituents are well distributed across the cluster sets. It is obvious, for example, that molar volume (a steric parameter) in set 3 will increase from left to right, as also will log octanol-water partition coefficients. Similarly, collinearity between Hammett constants and steric constants for set 1 should be anticipated. The substituents in Table 5.9 were chosen to illustrate the procedure by using sets whose behaviour can be predicted. Less obvious interrelationships and larger collections of substituents demand involved mathematical procedures. More comprehensive cluster set tables can be found in the literature. Cluster analysis is discussed briefly in Section 5.8. 5.8 MULTIVARIATE ANALYSIS Multivariate analysis broadly defines a collection of techniques which attempt to reduce complex collections of data to simpler relationships. Such a technique may, for example, be a simple classification procedure, seek to reduce correlated data to a smaller representation of independent variables, or attempt to find common factors within a set of data. Data are usually expressed in the form of a matrix, which is a rectangular array of numbers, enclosed in square brackets. Matrices can be subjected to mathematical manipulations, such as addition and multiplication, but the procedures involve different

Table 5.9 A model cluster set table. Cluster set Substituents 1 -F, -Cl, -Br, -I 2 -NO2, -NO, -CN 3 -H, CH3, C2H5, C3H7 4 -OH, NH2 5 -CH2OH, -C2H4COOH, C3H6COOH 6 -NHCONH2, -NHCOCH3 7 -CONH2, -CONHC2H5, -COOH 8 -OCH3, -OC2H5, -OC3H7 Further cluster sets can be found in: Hansch, C. and Leo, A. (1979).

Introduction to the principles of drug design and action

242

Substituent Constants for Correlation Analysis in Chemistry and Biology. New York, John Wiley. rules from those of classical algebra, and become very protracted when there are a lot of data. Table 5.10 is a typical matrix of QSAR data, and is a series of rows of numbers, each row representing one compound. Each number in a row forms part of a vertical column, and each column represents a specific property, in this case log TM, log BRmax, log kc, Rm or Es. These parameters are defined in Table 5.10. Such a matrix is called an observation matrix, because it contains the raw data upon which the analysis will be based. Observation matrices are rarely used as such; instead, new matrices which are smaller and easier to handle are developed from them. We have a situation in Table 5.10 where there are two dependent and three independent variables, which poses the questions: (a) Are the two dependent variables, BRmax and TM, related? (b) If they are related, can one of them be ignored? (c) If they are not related, which independent variables influence one of the dependent variables, and which influence the other? These possibilities can be investigated by constructing a correlation matrix. The correlation matrix for Table 5.10 is shown in Table 5.11. Parameters are ranged along the top and down the left hand side, and the relevant correlation coefficients crossreferenced in the resulting grid. The correlation coefficient of 0.85 between log TM and log BRmax suggests that the two dependent variables are related, but not sufficiently so to warrant the use of a regression equation. Log TM appears to be highly correlated with both log kc and Tm (correlation coefficients of 0.98 and 0.97 respectively), but one of these relationships could be the result of the high correlation (r=0.95) between log kc and Rm. Es is not highly correlated with any of the other terms, but the correlation coefficients (0.90, 0.81, 0.88 and 0.80) are sufficiently large to suggest there could be relationships between Es and one or more of the other parameters. The existence of a steric contribution to biological activity is therefore uncertain, undoubtedly a result of only five compounds being examined. One would not normally carry out an analysis of this sort with so few data, but they provide an example from which the procedures are easily explained.

Table 5.10 Androgenic activities and QSAR parameters of some testosterone esters. Ester log TM log BRmax log kc Rm Es Formate 0.30 1.60 1.27 0.58 0.00 Acetate 0.48 1.73 1.48 0.46 −1.24 Propionate 0.78 2.10 2.00 0.11 −1.58 Butyrate 0.78 2.31 2.09 −0.09 −1.60 Valerate 0.90 2.02 2.06 −0.26 −1.63

Quantitative structure

243

Results were obtained with groups of castrated rats. Animals were sacrificed at intervals and prostate and seminal vesicles removed and weighed. BRmax (maximum biological response) represents the mean maximum weight of prostate plus seminal vesicles. TM (time of maximum effect) represents the time after injection when BRmax occurred. kc (catalytic constant) is the rate constant for the hydrolysis of the esters, in vitro, with standardised liver homogenate. Rm is a chromatographic parameter derived from the Rf value and logarithmically related to the partition coefficient. Es is a parameter related to the bulkiness of the ester group. Table 5.11 Correlation matrix of androgenic activity data. log TM log BRmax log kc log BRmax 0.85 0.98 0.94 log kc 0.97 0.83 0.95 Rm 0.90 0.81 0.88 Es

Rm

0.80

Correlation matrices can be obtained directly by computer, using the MINITAB statistical software for example. Thus if the data shown in Table 5.10 are entered, and the appropriate command given, the matrix of the Table 5.11 will be displayed. Further information can be obtained from a principal components analysis, which can be carried out by entering the matrix given in Table 5.11 into a suitable software package such as MINITAB. The result is shown in Table 5.12. Mathematical interpretation of the terms involved is outside the scope of this book, but their functions will be explained. For further information, readers are referred to more specialist texts (for example, Armstrong and James, 1996). The five principal components are listed in the left hand column. Each is a combination of the original variables, consisting of five coefficients or eigenvectors, given in the right hand columns. Thus, the first principal component is:

The eigenvalues, given in column two, add together to give the total number of principal components, i.e. 4.57+0.21+0.20+0.02+0.00=5.00, and each eigenvalue, in the second column, expresses the fraction of the total variance which its principal component represents. Thus the first principal component explains (4.57×100)/5=91.4% of the variance. Similarly, the second and third principal components explain 4.2 and 4.0% of the variance respectively, giving a total of 99.6% for the first three components, so that the remaining two principal components are insignificant.

Introduction to the principles of drug design and action

244

In a similar manner, the magnitudes of the eigenvectors are a measure of the importance of the original variables to the principal component. The total sum of the squares of the eigenvectors in each principal component is equal to unity, for example for the first principal component (0.46)2+(0.43)2+(0.47)2+(−0.45)2+(−0.43)2=1.00. The sum of the squares of the eigenvectors is called the communality of the row. The importance of an eigenvector to the row can be assessed by subtracting its square from one, and noting

Table 5.12 Eigenvectors and eigenvalues for androgenic activity data. Principal component Eigenvalues log TM log BRmax log kc Rm 1 4.57 0.46 0.43 0.47 −0.45 2 0.21 −0.03 −0.21 −0.14 0.44 3 0.20 0.37 −0.80 −0.10 −0.46 4 0.02 −0.54 0.18 −0.47 −0.62 5 0.00 −0.60 −0.31 0.73 −0.09

Es −0.43 −0.86 −0.04 −0.27 −0.07

how much the communality deviates from unity. Thus the communality of the second principal component, if the first eigenvector for log TM is removed, is:

from which we conclude that log TM is of insignificant importance to the second principal component, so that 4.2% (0.21×100/5) of the variance of the raw data is explained by a relationship between log BRmax, log kc, Rm and Es. Factor analysis is used for matrices in which the value of each element is the product of two factors, one (f) characteristic of the column, and the other (a) characteristic of the row, giving

where x is the numerical value of the element. Thus for example, for three columns (A, B and C) and three rows (1, 2 and 3),

Factor analysis was originally based on the results of a multisubject examination, in which it was observed that subject scores were roughly related. It was therefore suggested that one or more of the row terms could be ignored, and substituted by a factor e, e.g. ignoring column C,

Principal component 1

Factors

Quantitative structure

245

2 3 The e term represents the error involved in ignoring column C, and is equal to unity minus the communality of the columns which were considered. Cluster analysis is used to classify results into groups or clusters. Two general procedures will be described, as follows: (a) Dendrograms (b) Hierarchic methods. Dendrograms are used to investigate collections of factors to determine which are related and which are independent. A typical example is shown in Table 5.13, which lists ten questions about a collection of shampoos, put to a test panel. The panel was asked to use the shampoos, and give them numerical gradings. The resulting correlation matrix is shown in Table 5.14. A dendrogram of the correlation matrix is plotted in Figure 5.7, in which the ordinate represents correlation coefficients, and the question numbers are equally spaced along the abscissa, not necessarily in numerical order. The highest correlation coefficient in

Table 5.13 Shampoo questionnaire. Assess the following qualities on the given 0 to 7 scale 1. Overall impression 2. Suitability for your hair type 3. Lathering ability 4. Rinsability 5. Cleansing power 6. The condition of your hair 7. The manageability of your hair 8. The feel of your hair 9. The texture of your hair 10. How tangle free your hair was left Table 5.14 Correlation matrix for shampoo evaluation factors. Question number 2 3 4 5 6 7 8 1 0.90 0.47 0.33 0.60 0.81 0.34 0.41 2 0.35 0.33 0.49 0.72 0.33 0.31 3 0.47 0.63 0.36 0.15 0.39 4 0.59 0.29 0.46 0.28 5 0.60 0.41 0.54

9 0.27 0.21 0.22 0.19 0.35

10 0.36 0.31 0.28 0.45 0.42

Introduction to the principles of drug design and action

6 7 8 9

246

0.41 0.43 0.32 0.41 0.35 0.34 0.28 0.16 0.37 0.86

Table 5.14 is 0.90, between questions 1 and 2, and these numbers are placed side by side at a convenient position on the horizontal axis, to form the base of a rectangle of height 0.90. The next highest correlation coefficient (0.86) occurs between questions 9 and 10, but neither 9 nor 10 correlates well with the other questions, so this result is set aside while other, higher correlation coefficients are considered. The next correlation coefficient is 0.81, for questions 1 and 6, and since it has question 1 in common with the first rectangle, this rectangle is plotted alongside with question 1 shared by the two rectangles, which are then joined by a horizontal and two vertical lines, corresponding to the next highest correlation coefficient (0.72 between 2 and 6), as shown in the figure. Before considering the next highest correlation coefficient (0.63 between 3 and 5), it is necessary to note that the next two correlation coefficients are equal (0.60), and are between question 5 and either 1 or 6, both of which belong to the cluster of two rectangles described above. The rectangle between 3 and 5 is therefore drawn alongside the “6,1,2” cluster, to which it is linked by a bar, 0.60 units high. The next correlation coefficient (0.59) is between question 5, which is in the “3,5,6,1,2” cluster, and question 4. A vertical of 0.59 units is therefore drawn, and linked by a horizontal bar to the cluster. For similar reasons, a vertical of 0.54 from question 8, which has the next highest correlation coefficient, is placed for convenience on the other side of the “4,3,5,6,1,2” cluster, and joined to it with a horizontal bar. The next three correlation coefficients involve questions which already form part of the dendrogram, and are therefore ignored, and question 7 is joined

Quantitative structure

247

Figure 5.7 Shampoo dendrogram. Note that the correlation coefficient scale is negative, so that a short rectangle represents a high coefficient and a tall rectangle represents a low coefficient. to the cluster in the same way as question 8. Finally, the “9,10” rectangle, which correlates poorly with all the other questions (r=0.43), is joined up to the cluster which forms the remainder of the questions. The dendrogram suggests that there are three clusters, (a) Between questions 3, 4 and 5, indicating that the subjects associate lather with cleansing, which is not surprising. (b) Linking questions 1, 2 and 6, suggesting that subjects discriminate between shampoos principally for the condition in which it leaves their hair. (c) Between questions 9 and 10, and these correlated very little with the other attributes. The inference is that the descriptions, “texture” and “tangle-free” were associated, and were not related with terms like “condition”, “manageability” and “feel”.

Introduction to the principles of drug design and action

248

5.8.1 Pattern recognition methods Some pharmacological and toxicological data are binary (e.g. a compound may be assessed as being carcinogenic or non-carcinogenic) rather than continuous. Such data are not amenable to multiple regression analysis, and other methods have to be used in their analysis. A number of such methods are available, the most widely used of which is discriminant analysis. A number of physico-chemical and/or structural parameters are generated for each compound, and an appropriate computer program selects the best combination of these that will discriminate between the different classes of biological activity. If n parameters are selected, then the discrimination is by means of a hyperplane in n-dimensional hyperspace. This is difficult to comprehend, and so principal components analysis is often used to reduce the data, and frequently a plot of the first principal component (PCA1) against PCA2 gives good discrimination. An example of this is shown in Figure 5.8. A slightly different approach is exemplified in Figure 5.9, in which active compounds are seen generally to have lower π values than inactive compounds. The situation is rarely as simple as this. Experimental results usually take the form of a continuous spectrum of biological activities, and an arbitrary dividing line, above which the compound is deemed to be active, has to be drawn across the graph of “active” or “inactive” against a dependent variable, such as Hansch’s π value. Overlapping can be rationalised by determining the mean squared distance (MSD); for example, for compounds 1, 2 and 3 in Figure 5.9, if π1 represents the π value of compound 1 and so on.

(5.43) For all six compounds there are 19 other possible combinations of three. These must all be computed, and if A is the number of groups of three, including the suspected cluster, having MSD values equal to or less than the MSD value of the selected group, the probability (P) that a cluster could arise by chance is given by Equation [5.44].

(5.44) The procedure can be extended to more than two dimensions; for example, if a plot of π against σ suggested that three compounds (1, 2 and 3) were the only biologically active molecules in a larger population, the MSD of the three compounds can be calculated from Equation [5.45] and compared with values for the other combinations of three.

Quantitative structure

249

(5.45) Hansch π values and Hammett σ values have a similar range of numbers. When parameters have a significantly different order of magnitude, they must be given equal weight. This can be done by subtracting the mean of each column in the matrix from its individual variables, and dividing by the standard deviation of the column. Such procedures are too protracted for manual calculations, but can be handled by a computer. For example, McFarland and Gans (1986) examined 20 compounds, involving 77 520 clusters of seven compounds. 5.9 NEURAL NETWORKS In recent years neural network (NN) programs have been increasingly used in problem solving, and a number of QSAR applications have been published. NNs model the way that the human brain processes information, and have the potential to be extremely powerful tools. A typical scheme for an NN simulation is shown in Figure 5.10. Input nodes receive information, and pass it to one or more intermediate (hidden) layers of

Figure 5.8 Principal component analysis of skin corrosivity of acids (from Barratt, M.D. (1995) ATLA 23, 111–122) (reproduced

Introduction to the principles of drug design and action

250

with permission). The principal components were derived from log P values, molecular volumes, melting points and pKa values.

Figure 5.9 Cluster graph. “neurons” which process the information. The processed data are then passed to the output nodes. In Figure 5.10 all nodes in one layer are shown connected to all nodes in adjacent layers, but this does not have to be the case, so that the NN can be set up in various ways. The network is trained using a training-set of data, so that it can then make predictions about additional data. The level of training is crucial, and it is possible to overtrain a network. Recent QSAR studies using NNs indicate that they are capable of giving better predictions than are Hansch-type regression equations. 5.10 SUMMARY Quantitative structure-activity relationships can be said to have brought drug design into the computer age. They have introduced a quantitative element into a subject which had hitherto been entirely qualitative. The medicinal chemists of old used their instinct and experience to predict conformational changes which should bring about increased biological activity, but had little idea of the magnitude of those changes. QSAR requires less instinct and experience, and also estimates the extent of the biological activity. Hansch and Free-Wilson analyses are lead-optimizing techniques. This means that it is necessary to begin with a ‘lead’ compound, having the basic pharmacological properties; analogues are prepared and tested, and the resultant data used to plan new compounds which will have enhanced biological activity. The

Quantitative structure

251

technique demands that compounds be selected in a highly rational manner, with a systematic and stepwise progression throughout. Because of this stringent approach, much of the early structure-activity data in the literature are not adaptable to QSAR. These techniques are also self-limiting: as more results become available, the picture becomes clearer, but the number of possible new compounds become more restricted. Powerful new mathematical techniques in which the computer is made to recognize molecular structures are being extended to drug design, and could move the concept of QSAR from lead-optimizing to lead-generating. For example, if through the use of 3D QSAR the receptor features necessary for activity can be identified, it should be possible

Figure 5.10 A typical neural network scheme. to design novel compounds that will complement those receptor features, and which should thus possess the requisite activity. FURTHER READING Armstrong, N.A. and James, K.C. (1996) Pharmaceutical Experimental Design and Interpretation. London: Prentice Hall. Basak, S.C. (1987) Use of molecular complexity indices in predictive pharmacology and toxicology: a QSAR approach. Med. Sci. Res. 15, 605–609. Buisman, J.A.K. (1977) Biological Activity and Chemical Structure. Amsterdam: Elsevier. Charton, M. (Ed.) (1996) Advances in Quantitative Structure-Property Relationships, Vol. 1. Greenwich, CT, USA: JAI Press. Coulson, A.E. (1965) An Introduction to Matrices. London: Longman Group Ltd. Hansch, C. and Leo, A. (1979) Substituent Constants for Correlation Analysis in Chemistry and Biology. New York: Wiley.

Introduction to the principles of drug design and action

252

Hansch, C. and Leo, A. (1995) Exploring QSAR. Vol. 1. Fundamentals and Applications in Chemistry and Biology. Washington DC: American Chemical Society. Hansch, C., Leo, A. and Hoekman, D. (1995) Exploring QSAR. Vol. 2. Hydrophobic, Electronic and Steric Constants. Washington DC: American Chemical Society. James, K.C. (1974) Linear free energy relationships and biological action. In Progress in Medicinal Chemistry, edited by G.P.Ellis and G.B.West, Vol. 10, pp. 205–243. Amsterdam: North Holland Publishing. James, K.C. (1986) Solubility and Related Properties. New York: Marcel Dekker. Kendall, M. (1980) Multivariate Analysis. London and High Wycombe: Charles Griffin & Co. Ltd. Kier, L.B. and Hall, L.H. (a) (1976) Molecular Connectivity in Chemistry and Drug Research. London: Academic Press. (b) (1986) Molecular Connectivity in Structure-Activity Analysis. Letchworth: Research Studies Press. Kubinyi, H. (1993) QSAR: Hansch Analysis and Related Approaches. Weinheim, Germany: VCH. Kubinyi, H. (Ed.) (1993) 3-D QSAR in Drug Design—Theory, Methods and Applications. Leiden, Netherlands: ESCOM Science Publishers. McFarland, J.W. and Gans, D.J. (1986) The significance of clusters in the graphical display of structure-activity relationships. Journal of Medicinal Chemistry 29, 505– 514. McGowan, J.C. and Mellors, A. (1986) Molecular Volumes in Chemistry and Biology. Chichester: Ellis Horwood Ltd. Tute, M.S. (1971) Principles and practice of Hansch analysis: a guide to structureactivity correlation for the medicinal chemist. In Advances in Drug Research, edited by N.J.Harper and A.B.Simmonds, Vol. 6, pp. 1–77. London: Academic Press. Wells, P.R. (1968) Linear Free Energy Relationships. London: Academic Press.

6. FROM PROGRAMME SANCTION TO CLINICAL TRIALS: A PARTIAL VIEW OF THE QUEST FOR ARIMIDEX™, A POTENT, SELECTIVE INHIBITOR OF AROMATASE PHILIP N.EDWARDS CONTENTS 6.1 INTRODUCTION

208

6.2 BACKGROUND

209

6.3 ICI START ON AROMATASE INHIBITION

213

6.4 AROMATASE RESUMED

217

6.4.1 The legacy from anti-oestrogens

217

6.4.2 The potential importance of uninterrupted drug cover

218

6.4.3 Increasing concerns about timeliness

219

6.4.4 Naphthol-lactones, tight binding and in vitro/in vivo relationships

221

6.4.5 The design principles behind ICI 207658—later named Arimidex™

222

6.5 BIS- TRIAZOLE (6.17)—A TIMELY AROMATASE COMPOUND IN DEVELOPMENT

227

6.6 THE SEARCH FOR THE IDEAL BACK-UP CANDIDATE

229

6.7 INTO THE CLINIC

232

FURTHER READING

233

Introduction to the principles of drug design and action

254

6.1 INTRODUCTION Breast cancer is the commonest cancer in women and, despite continuing advances in treatment, each year world-wide an increasing number die from the disease: in Japan the incidence has been increasing by an alarming 10% per annum. In the early stages of the disease, 30–40% of patients respond to hormonal or anti-hormonal therapy. One way to deprive hormone-dependent cancer of its primary mitogens, oestrogens, is to prevent their synthesis—preferably by inhibition of aromatase, the ultimate and biochemically unique enzyme that converts androgens such as testosterone to mitogenic oestradiol. This account provides some of the background to the author’s and ICI’s involvement with hormonal modulation, but attempts mainly to cover the cytochrome P450-dependent enzyme, aromatase (oestrogen synthase, P450arom/NADPH cytochrome P450 reductase), its inhibition, and the way in which the ICI Aromatase Team selected a development compound, was forced by long-term toxicity to abandon it, but was more fortunate in its second choice with the compound ICI 207658, numbered D1033 during early development and later given the name anastrozole. During the development phase of D1033, ICI Pharmaceuticals became Zeneca Pharmaceuticals and the designation ZD1033 was used for the drug which is now called Arimidex*. The medicinal chemistry coverage, in focusing mainly on the author’s team contributions to the programme, is a partial account of work that involved several chemistry teams. 6.2 BACKGROUND An undergraduate course in organic chemistry in the mid 1950s typically made good use of the inspiring work of many groups in the fields of steroid structure determination, conformational analysis, reactivity and synthesis. The potent and multifarious biological properties of such molecules, along with the synthetic challenges presented by, for those times, extremely complex structures, made them synthetic targets for many eminent chemists of the day. The first formal total synthesis of cortisone (6.1), then thought to be a miracle drug, had been briefly reported by Fieser and Woodward in August 1951. However, the race to achieve the first non-trivial synthesis of cortisone had unexpectedly been won by a group of young chemists in Mexico City who were employed by a small, recently-formed company called Syntex Inc. That synthesis, starting with readily available diosgenin—from Mexican yams—had commercial potential from the sale of intermediates as well as the final product and its dihydroderivative, hydrocortisone or cortisol. Cortisone was in great demand for the treatment of severe inflammatory and immunologically-related conditions, as well as for treating Addison’s disease—a previously life-threatening condition caused by a deficiency of cortisol synthesis in the adrenal gland of an afflicted individual. Clearly, our target aromatase inhibitor would have to avoid in vivo inhibition of the cytochrome P450-dependent enzymes involved in cortisol/cortisone synthesis. Indeed

From programme sanction to clinical trials

255

it ideally had to avoid inhibiting any cytochrome P450-dependent enzyme except aromatase. Schenkman and Greim (1993) recently edited a wide-ranging multi-author review of such enzymes. Carl Djerassi was a leading member of that early Syntex group and he was soon hailed by many in academia as one of the promising synthetic chemists of the day. Such limited accolades were eventually dwarfed when he achieved broadly proclaimed ‘immortality’ as the ‘Inventor of The Pill’. As so often even in those days, the title is largely the result of the mass media requirement for (over)simplification. As Djerassi (1992) emphasises * Arimidex is a trademark, the property of Zeneca Limited.

in his autobiography, his part in the invention of ‘The Pill’ was that of main contributor to the search for and discovery of the orally potent progestagen, norethindrone (norethisterone) (6.2); this was achieved during his brief time at Syntex and overlapped the cortisone work. Others had foreseen the potential of such agents and many others were involved in the development and exploitation of this and related compounds—indeed it took more than a decade for such hormonal modulation to gain even limited acceptance as a method of contraception. It is of interest in the present context that norethindrone and other terminal acetylenes such as ethynyloestradiol may owe some of their improved oral activity to irreversible, mechanism-based, covalent inhibition of drug-metabolising cytochrome P450-dependent enzymes in the liver. This possibility and the nature of such enzymes were unknown at the time of those drugs’ discovery, but with the advent of that understanding it is now possible to design inhibitors of P450s based on terminal acetylenes. Potent inhibitors of aromatase have been generated by modifying natural substrates, and close analogues, through the addition of an ethynyl group to C19—the initial site of substrate oxidation by aromatase. Amazingly, it was only in 1982 that norethindrone was shown in vitro to be a rather weak (~2 µM), irreversible inhibitor of aromatase. It is unlikely, however, that this has relevance to its contraceptive use. By the late 1950s many research groups were involved in hormonal modulation. One of those groups was in ICI Pharmaceuticals and its efforts were rewarded with the discovery and successful development of the oestrogen antagonist tamoxifen, or Nolvadex™ (6.4) as ICI named it (Nolvadex is a trademark, the property of Zeneca Limited). A number of structurally related oestrogen agonists had been discovered in the 1930s, one of which, diethylstilboestrol (6.3), remains, somewhat controversially, in use to this day.

Introduction to the principles of drug design and action

256

Nolvadex™ (6.4) was found to be an effective treatment for a substantial proportion of post-menopausal patients with oestrogen-dependent breast cancer. It has in recent years been the largest selling chemically-defined anticancer drug of all time. The author’s first year in research, 1957–8, coincided with three events important to this discourse. First, M Klingenberg and later D Garfinkel independently reported the generation of a new absorption peak at 450 nm when homogenised liver supernatant was exposed to carbon monoxide: pigment 450 (P450) was born, but the function, if any, of this pigment was unknown. Second, K J Ryan reported that androgens incubated with human placental microsomes were converted to oestrogens. This amazing process involves the removal, by then unknown chemical steps, of a very hindered, non-activated methyl group from C10 and a non-activated hydrogen atom from C1 of the androgenic precursors

testosterone or androstenedione. What agent or agents were at work was again unknown. Third, the Swiss pharmaceutical giant, Ciba, a major force in steroid chemistry, pharmacology and drugs, started clinical trials with aminoglutethimide, (AG) (6.5), as a prospective anticonvulsant drug. Those trials and subsequent use under the name Elipten™ (later Orimeten™: Ciba, and now Cytadren™: Ciba-Geigy) revealed multiple adverse side effects, one of the most serious being adrenal insufficiency. A few years after launch it was withdrawn from sale, but as so often in chemotherapy, one person sees a side effect while another sees an opportunity: a medical adrenalectomy might be useful in various adrenal hormone-dependent diseases—including breast cancer where surgical adrenalectomy was an established hormonal manoeuvre. Some years later, that possibility became actuality—AG was shown to be useful in several conditions, including advanced breast carcinoma. It originally was assumed that efficacy flowed from suppression of adrenal pregnenolone synthesis. Rather lowpotency (~26 µM) inhibition in vitro of an adrenal-derived enzyme, P450scc, that converts cholesterol via side chain cleavage to pregnenolone, had long since been demonstrated and adrenal hypertrophy in various species dosed with AG is ascribed to that gland’s attempt to maintain steroidogenic homeostasis. Inhibition of P450scc would in turn limit synthesis of the many other steroids, including oestrogens, which have pregnenolone as a precursor. Scheme 6.1 shows a selection of steroidogenic pathways—unidirectional arrows indicate that one or more of the steps in that pathway involves a cytochrome P450 enzyme. These pathways operate in differing degrees

From programme sanction to clinical trials

257

according to tissue, species, sex, age and in pregnancy and disease and their products elicit a wide variety of responses depending on the target cell type and its environment (Castagnetta et al., 1990). Replacement glucocorticoid was administered with AG during breast cancer therapy in part because of the above findings. Later quantitative studies however showed reduced oestrogen levels but normal or even increased levels of androstenedione in AG-treated patients’ plasma: androstenedione is the main androgenic precursor of oestrone. These

new findings appeared inconsistent with substantial P450scc involvement. Subsequent in vitro studies—twenty-five years after Ryan’s discovery—suggested that efficacy in post-menopausal breast cancer patients resulted mainly from inhibition of the enzyme (or enzymes) that converts androgens to oestrogens. By then the placental enzyme

Introduction to the principles of drug design and action

258

activity was widely known, but, because the enzyme is embedded in microsomal phospholipid membranes, and is functionally dependent on that association, it was not yet well characterised. Crystallisation is likely to be extremely difficult or impossible. We now know that P450arom is a single enzyme—the product of the CYP19 gene on chromsome 15 in man. Gene expression is subject to complex and multifactorial regulation. The enzyme is widespread in humans, males and females, in the brain and the periphery, but it is much less widespread in most other species. It belongs to the cytochrome P450-dependent class of enzymes and is commonly called aromatase or P450arom, but the latter designation refers specifically to the haem-binding protein of the two component enzyme: the second component, a flavoprotein, is reduced nicotinamide adenine dinucleotide diphosphate(NADPH)-cytochrome P450 reductase; the reductase is common to all P450-dependent enzymes. Use in vitro of oxidants other than air, for example Ph-I=O, or H2O2, allows most P450s to function in the absence of the reductase. Interestingly, and relevant to the mechanism (Scheme 6.2), iodosobenzene is ineffective in the final step of aromatisation while hydrogen peroxide allows all three steps to proceed. P450arom, or rather the oestrogens it produces, has differing roles according to the sex of the animal and cell type in which it is expressed. Importantly it is expressed and is functional in producing trophic effects in many breast cancer cell lines—a bloodborn oestrogen supply is not always necessary for growth. Testololactone(Teslac™, Squibb) (6.6), used in breast cancer and originally thought to act via androgenicity, may also owe much of its efficacy to aromatase inhibition. 6.3 ICI START ON AROMATASE INHIBITION In the last years of the 1970s the Fertility Project Team in ICI was again testing compounds for antifertility potential. Some of that effort was devoted to random screening with the end point being prevention of pregnancy in rats. One of the more potent compounds discovered, the N-’benzyl’ imidazole (6.7), was considered to be worthy of further investigation since its structure and overall biological effects did not point to any known mode of action. There was some concern that its fragmentation to a quinone methide might be involved—if that was so, the generation of such a reactive species would make it and any analogues unattractive. The postulated quinone methide is implicated in the lung toxicity of butylated hydroxytoluene (BHT; 3,5ditertiarybutyl-4-hydroxy-toluene), a widely used antioxidant. Also at that time, the Team’s interest in aromatase had been heightened by the results of clinical and biochemical studies in patients receiving AG. Preclinical results being obtained by Angela Brodie and co-workers with 4-acetoxy-androst-4-ene-3,17dione were also encouraging. This steroid was active in vivo, especially in the oestrogen-dependent DMBA (dimethylbenzanthracene) rat tumour model, but interest focused later on the 4-hydroxy compound, 4-OHA (6.8), (formestane), a potent, Ki=10 nM, and time-dependent aromatase inhibitor since licensed and named Lentaron™ by Ciba-Geigy. Structure (6.8) is shown with partial van der Waals’ radii for some ‘atoms’ (actually CH2 and CH3 groups); those ‘atoms’ in the enzyme bound state are

From programme sanction to clinical trials

259

postulated to be in contact with the large, extensively-planar protoporphyrin-IX prosthetic group which is depicted, edge-on, as a thick line. Partial van der Waals’ radii for atoms in the porphyrin are not shown but extend to contact those shown for the steroid. The official start of the Aromatase programme—just a few weeks into the new decade, was contemporaneous with the beginning of another team’s attempt to find an anti-oestrogen working through inhibition of translocation of the oestrogen receptor from cytosol to nucleus, but more of that later. Chemistry started in two main

directions, steroid-based (naturally!) and azole-based: the N-’benzyl’ imidazole (6.7) had by now been shown likely to be an aromatase inhibitor. Our compound collection, together with some standard antifungal agents from ICI Plant Protection, generated a structurally diverse set of leads with some remarkably simple azoles, e.g. N-(mpentanoylbenzyl)imidazole (IC50=2 ng/ml), being very potent inhibitors of human placental microsomal aromatisation—e.g. of testosterone to oestradiol or of androstenedione to oestrone. The literature evidence at the time was consistent with all the aromatase chemistry being carried out by a single enzyme, but this became certain only much later, during the second part of the programme. The multiple steps involved in this conversion were already broadly established from a vast body of work by many academic groups (Brodie et al., 1993), see Scheme 6.2, but some of the finer mechanistic detail remains controversial (Aktar, Njar and Wright, 1993). The lower part of the Scheme and footnote commentary represent the author’s view of the how the final steps might proceed. At this stage we knew from ICI work on antifungal agents which potently inhibit fungal lanosterol-14-methyl demethylase, a P450-mediated reaction very closely related to that performed by aromatase, as well as from literature reports of azoles inhibiting various P450 enzymes, and already rather extensive studies with the N’benzyl’ imidazole (6.7), that superior selectivity versus AG could be the key to a successful drug. Considering how ‘dirty’ AG is by modern standards, this seemed at first an easy target. We soon thought otherwise: the in vivo effective azoles then to hand were all clearly deficient in one or more respects. Surprisingly to us because we were not aware of any connection with P450 enzymes, all the in vivo more potent (but

Introduction to the principles of drug design and action

260

still weakly potent) azoles, including (6.7), caused unacceptable elevation of liver triglycerides at modest multiples of their aromatase-effective doses. Another frequently observed effect in vivo—adrenal enlargement—indicated unwanted inhibition of non-oestrogenic steroid production. No pattern of selectivity could be discerned. Yet another indicator of potentially inadequate selectivity was the increased sleeping times observed after co-administration to mice of hexobarbital with each of the few azoles tested: such effects are probably due to inhibition of liver P450mediated oxidative clearance of this xenobiotic sedative. Multiple high doses of all azoles examined caused increases in liver weight to body weight ratios and elevation of some liver mixed function oxidase (P450) enzymes. These elevated P450 levels can cause increased clearance rates and modify metabolite patterns of hormones, drugs and other natural products and xenobiotics. Both these effects and enzyme-inhibitory effects are present in AG-treated patients—but we decided that none of these effects would be acceptable at the therapeutic dose of our target molecule. Increased P450mediated production of toxic and particularly mutagenic metabolites is one of the consequences of smoking and is implicated in the increased incidence of cancer in smokers. Smoking differs from most drug therapy in causing different P450s to become elevated, but obviously everyone would wish to minimise such risks—even in long-term drug treatment of cancer patients. So how might one achieve selectivity? There are several possibilities: set up screens and throw everything you have at them; or, ideally with the help of precision models—Dreiding etc. and computer modelling, try to use substrate structures and inhibitor structures as a guide to drug design; or, again using modelling, try to understand the reaction(s) using as much detailed information as exists, make educated guesses about

From programme sanction to clinical trials

261

the three-dimensional interactions needed for recognition and mechanism, then design the drug around as many hopefully unique features as possible. Mainly in part two of the programme, we did some of each of these and developed other ideas that will be discussed later. Modelling at various levels should be, and was, an on-going process. The amino acid sequence of human aromatase was not known at

Introduction to the principles of drug design and action

262

the time of our work but became so soon thereafter: its sequence of 503 amino acids shows only about 30% homology with other known mammalian P450s; the latter group are highly homologous. This puts P450arom in a unique category. Despite the implications from the foregoing, several groups have published models of the enzyme based on lipid-free, water-soluble, bacterial enzymes, e.g. P450cam, that have been crystallised and the structures determined at high resolution by X-ray diffraction methods. In the author’s view none of these models is satisfactory and any model is at present highly speculative. Speculative hydrogen bonds indicated in Scheme 6.3 are those used during our work. The peroxy intermediate (6.9), bound to a partial enzyme model, is shown as a stereoscopic pair in Figure 6.1. This binding mode was the basis of essentially all our modelling—despite the (still) speculative nature of such a species. It is shown essentially as we used it except that the amino acid side-chains on the α-helical protein fragment have been updated: we used those in P450cam. Some inhibitors throughout the chapter are drawn as they would appear in such a model— but with the partial α-helix removed and viewed edge on to the porphyrin multi-ring system: these changes allow easier comparisons of our suggested binding modes. Steroidal inhibitors might seem intrinsically to hold better prospects for selectivity, but, as shown in part in Scheme 6.2, Nature’s wide use of P450 enzymes in chopping, trimming and oxidatively modifying this skeleton argues against overconfidence in this intuitive position. Furthermore, steroids typically have other problems such as rapid clearance and poor efficacy by the oral route, especially in the rat—our preferred test species. Effects through receptor interactions are another concern. At a practical level, synthesis of new compounds can be demanding and slow and several other groups—industrial and academic—were known to have a substantial start on us. Counterbalancing this, we had a steroid expert in the team and we thought it worth a try. Probably none of these considerations counted for much in the light of the excitement generated by the ‘translocation’ work yielding some extremely interesting compounds. This new lead, irrespective of mechanism, was anti-oestrogenic in every test that was applied. The observations of apparent translocation of receptors, in response to oestrogens, turned out to be artifactual—they are always predominantly in the nucleus—but this idea nonetheless led to the discovery of the first ‘pure’ antioestrogens (Wakeling 1990).

From programme sanction to clinical trials

263

Figure 6.1 Stereo-pair of proposed (partial) active site of P450arom There is recent evidence to show that these agents are not equivalent to the total absence of oestrogen—the potential outcome of aromatase inhibition. Control of gene transcription is a complex multifactorial process in which the occupied but apparently oestrogenically-inactive receptor still has a role. Ongoing clinical studies may start to tease out some of the therapeutic implications of this complex and still little explored biochemistry. Unsurprisingly, chemistry effort was switched from aromatase. Soon still more effort was required: it was then that the author came to work for the first time on hormonal modulation. 6.4 AROMATASE RESUMED 6.4.1 The legacy from anti-oestrogens A ‘pure’ anti-oestrogen development candidate was chosen in 1985. ICI 182780 (6.11), is a 7α-(long side chain) substituted oestradiol derivative, but many nonsteroidal frameworks were investigated during the programme and most, with appropriate side-chains, yielded potent, ‘pure’ anti-oestrogens. Generally these frameworks, linked to azole rings via short side chains, yielded moderate-to high-potency aromatase inhibitors in the ensuing second phase of that programme. Because of its limited conformational freedom and very high inhibitory potency against human placental aromatase, one such azole, the (racaemic) triazole derivative (6.12), is particularly relevant to computer modelling of the enzyme active site. Several highly potent aromatase inhibitors arose inadvertently during the final stages of the anti-oestrogen work. We were attempting to find a replacement for the

Introduction to the principles of drug design and action

264

potency-enhancing but metabolically-sensitive phenolic hydroxyl group in our pure anti-oestrogens: phenolic compounds bind to the receptor in vitro about 100-fold tighter than non-phenolic analogues. Many alternatives to the phenolic OH group had been tried, but none came even close to matching its ‘magical’ effect. One possible explanation for these dramatic findings is that strong interactions occur between the receptor protein and both the inplane acidic hydrogen and the in-plane oxygen lonepair of the aromatic OH group. With this hypothesis, no benzenoid derivative was likely to match the phenol, but that conclusion need not apply to planar heterocyclic systems: the 4-substituted pyrazole

(6.13) was designed to interact with just such a hypothetical phenol-binding site, as shown, minus double bonds for clarity, in (6.14). Because there is no positional correspondence of atoms in the two differentlyinteracting rings, the design of the pyrazole 4-substituent could not be based on the normal steroid structure. Instead it was designed such that, overall, the molecule possesses a similar outline shape to the steroid skeleton and the hydroxyl group could be positioned roughly to correspond to that in testosterone. The design was a miserable failure—inhibition of radiolabelled oestrogen binding to the receptor was undetectable; a substantial volume deficit in the steroid ring B and C regions may contribute to this result. Inhibition of aromatase in contrast was among the best we had then seen: AR1: IC50=2 ng/ml (see below)! Unfortunately, activity in vivo was not detected at the highest dose examined, 20 mg/kg, and none of several pyrazoles was better, even after intra-peritoneal dosing.

From programme sanction to clinical trials

265

These results illustrate a common problem in chemotherapy—good activity in vitro all too often fails to manifest itself in vivo. Sometimes this can be rationalised, in part, in terms of competitive phase effects: the highly lipophilic N-’benzyl’ imidazole (6.7) binds to albumin and some other macromolecules and is extracted into fat deposits, phospholipid bilayers and other fatty body components. This drastically reduces the free aqueous concentration of the drug in vivo relative to that in vitro, and binding to the enzyme suffers in proportion. More usually poor bioavailability and rapid clearance are the most relevant parameters. The latter effect is undoubtedly relevant to potency in our OI3 test, but much more so in the OI2 test that we relied on increasingly throughout this second phase. These tests involved oral dosing of compound at 12.00 noon on day 3 (OI3) or 4.00 pm on day 2 (OI2) of the lightsynchronised ovarian cycles of female rats; it had been previously determined that suppression of ovarian oestrogen production from mid-afternoon until midnight of day 3 of the 4 day cycles, prevents priming of the hypothalamus for the ovulationtriggering surge of luteinising hormone (LH) on day 4. Ovulation inhibition is the observed end-point. 6.4.2 The potential importance of uninterrupted drug cover On theoretical and practical grounds (occasional non-compliance) there are good reasons for wanting a cytostatic anticancer drug to have a long half-life (t1/2): even the transient presence each day of growth-promoting levels of oestrogen may reduce response rates, quality or duration of effect. Tumour cells can express aromatase and synthesise oestrogens locally so plasma drug and oestrogen levels might give an incomplete picture of intra-tumour drug effectiveness. On the other hand, too long a half-life may result in serious consequences in the event of a severe adverse reaction. This analysis led us, in this case, to aim for an average t1/2 of our target drug, in patients, (assuming the simplest possible kinetics) of at least 12–16 hours and preferably not greater than two days. Because of competitive pressures, this criterion was made the dominant factor at one stage of the programme, despite the fact that predicting t1/2 values in man from data in other species was known to be little better than guesswork. 6.4.3 Increasing concerns about timeliness By the autumn of 1985, as we restarted the programme, the competitive situation in aromatase was intense. Many analogues of AG had been revealed and even pcyclohexylaniline had been shown to possess good in vitro potency—equal to AG with respect to human placental aromatase but substantially less so versus rat ovarian enzyme. We had during the initial phase of the programme tested a few compounds in parallel against rat and human enzymes: AG was seven-fold more potent against rat enzyme, while 1-nonylimidazole was three-fold selective in the reverse sense. Almost no other comparative tests were performed—our limited resources were needed elsewhere—so our interpretation of in vitro (human) to in vivo (rat) potency ratios was always potentially flawed by species differences in enzyme binding; we had to hope,

Introduction to the principles of drug design and action

266

we still hope, that we were not seriously misled, but the reader needs to bear this in mind during apposite parts of structure/activity relationship (SAR) discussions. Some of these new AG analogues showed much improved selectivity for aromatase—sometimes through improvements against the target enzyme, but often through reduced potency against other enzyme(s), typically P450scc. Potency in rats however, where reported, remained disappointing. Another recently reported analogue of AG had the 4-aminophenyl group changed to 4-pyridyl and it was reported to be more selective opposite P450scc. We decided to make a sample for in-house investigation. While we were doing that, and making a number of analogues, we roughly derived the necessary parameters for imides (at that time they were not available from published lists of Allinger’s MM2 force-field parameters) so that we could perform molecular mechanics calculations on such systems: we were thus able to predict that both this new analogue and AG exist very largely with the aromatic rings axially disposed. This interesting prediction led us to perform a slightly modified synthesis aimed at the analogue (6.15), which calculations predicted would exist overwhelmingly in the axial pyridyl conformation, and (6.16) which should have a moderate preference for the equatorial pyridyl conformation. The latter was only 2- to 6-times less potent in AR1 than either the parent or (6.15). No compound in this pyridyl series had sufficient potency in vivo to warrant further interest from us. By far the largest area of competitive activity concerned steroidal inhibitors, particularly of the time-dependent variety, but similarly disappointing in vivo results generally applied here. Even the Brodie compound, 4-OHA (6.8), had been shown to have unwanted androgenic effects and an intra-muscular depot formulation used for human dosage was not always well tolerated. Also by this time, we had noted an association between azole-based antifungal activity and aromatase inhibition. And since many drug companies were or had been active in the antifungal area, we needed to give rapid attention to this potential source of leads. Fortunately for us, another team within ICI Pharmaceuticals had recently completed an

antifungal programme based on inhibition of the multi-enzyme mediated biosynthesis of ergosterol—an essential constituent of fungal cell walls which is not synthesised in mammals. The specific target of that programme had been fungal lanosterol-14-methyl demethylase. As part of a frequently-applied ten-day teratology assessment, they had

From programme sanction to clinical trials

267

seen placental enlargement and effects on foetal development in pregnant rats—all potentially consistent with aromatase inhibition and a common property of the imidazole/ triazole compounds they had explored. This work had heightened awareness and understanding of selectivity issues in the business, and placental enlargement in rats provided the Aromatase Team with a test (PE9) wherein chronic effects of oestrogen depletion (compounds dosed once daily for 9 days) could be compared with similarly chronic effects on other systems in the same test animal. This overcomes problems of differential handling of compound between individuals, sexes or species, all of which in retrospect can be seen to have misled us at some point of the programme. As with all chronic tests, the accumulation of long half-life compounds (in this context, t1/2>ca. 1 day) can present problems, but may also allow such compounds to be identified at an early stage. We felt sure that time was not on our side, so we were well pleased when screening of our antifungal agents soon yielded a compound which was potent (IC50=7 ng/ml) in our aromatase screen (AR1: human placental microsomes; substrate, 40 nM [1,2-3H]androstenedione) and was inconsistently active in OI2/OI3 at 0.25–0.5 mg/kg. Poor aqueous solubility may have led to the inconsistency, but removal from that structure of an ortho-chloro substituent led to the bis-triazole (6.17), which was less active in AR1 (IC50=40 ng/ml), but consistently active in OI2, OI3 and PE9 at 0.2 mg/kg (approx. =ED50). The compound is therefore 20- to 50-times as potent as AG. It was urgently subjected to as detailed an investigation as the perceived time pressure allowed. That time pressure increased substantially during 1986 as the competitive situation grew still more intense. Schering were claiming long-lived oral effects for atamestane (1-methyl-androsta-1,4-diene-3,17-dione) dosed orally at 1 mg/kg to male volunteers, while Ciba-Geigy disclosed that their racemic bicyclic imidazole derivative, CGS 16949A (6.18), is 1000-times as potent as AG, with ED50 in female rats of 30 µg/kg, and inhibition of aromatase was evident in human male volunteers even at 0.3 mg per man! At that time, only in its duration of effect in volunteers, 4–10 hours, did the CibaGeigy compound seem to present room for improvement. This placed still greater emphasis on the half-life requirement of our target drug.

6.4.4 Naphthol-lactones, tight binding and in vitro/in vivo relationships

Introduction to the principles of drug design and action

268

While much of the Team’s early effort went on antifungal leads, we were also finding widespread activity with azoles attached to stilbenes related to (6.3), cis- and trans-2aryltetrahydronapthalenes related to (6.12), and to more speculative frameworks based on computer modelling—such as the naphthol-lactone (6.20) whose synthesis and structure are shown in Scheme 6.3. In vivo, lactones and simple phenol esters such as pivalate esters (archetypal prodrugs) are almost always too rapidly hydrolysed, via enzyme-mediated catalysis, to allow the longevity we demanded. But this lactone is a special case: it is impossible for it to be hydrolysed at pH 7.4—the typical value for blood. It even stays ring closed in very dilute sodium hydroxide—due to thermodynamics, not kinetics! The reason lies in the large increase in steric compression strain that accompanies ring opening. In contrast, the negligible problems generated in the ring-opened form of the synthetic intermediate (6.19), Scheme 6.3, leave this compound highly sensitive to hydrolysis—t1/2 in water at room temperature is ~120 seconds at pH 9: this is 5000-times more reactive than ethyl acetate. The more hindered lactone (6.20) is however rapidly reduced by borohydride, but only to a lactol (hemi-acetal); further reduction under the weakly basic conditions would require ring opening, which, like the ester hydrolysis, essentially does not occur. A lactol ethyl-ether, inadvertently produced in a reaction that had the cyclic ether as its target, was relatively poor in vitro but in vivo it had similar activity to the corresponding lactone. Rather efficient liver cytochrome P450-mediated reoxidation to the lactone seems a likely explanation. The imidazole (6.20), X=Y=CH, is extremely potent in vitro. In a single AR1 test it inhibited the aromatisation of tritium-labelled androstenedione by 74% at 1.25 ng/ml—the lowest concentration tested, while at higher concentrations the figures were: 2.5 ng/ml—95.5% and 5.0 ng/ml—99%. Such figures, if they can be relied on, are indicative of ‘tight binding’—the condition in which, for the simplest case, free drug concentration is significantly depleted from the nominal value by binding to a site, usually the active site, which is present in the test medium at a concentration only somewhat less than or equal to twice the observed 50% inhibitory concentration. Ultimately, half an equivalent of inhibitor—essentially all bound to the target—is the absolute minimum required for 50% inhibition no matter how potent the agent might be (rare, catalytically-active, irreversible inhibitors excepted). In most test situations one cannot rely on 95% inhibition being different from 100% inhibition, but here we are measuring release of tritiated water,

From programme sanction to clinical trials

269

which is easily and completely separable from the precursor, so very small amounts of reaction can quantitatively be measured. Other extremely potent inhibitors show similar responses, while weakly and moderately potent azole-based inhibitors all display classical inhibition curves consistent with the simplest outcome of 1:1 competition between substrate and inhibitor. The sigmoidal appearance of linear-%inhibition versus log-concentration curves tends to obscure modest deviations from ‘normality’, but the theoretical curve for the simplest case can be transformed to a linear function, or, more conveniently, the %inhibition axis can be transformed so that experimental points for the simplest case should be linear and lie on a line which passes through 9.09, 50 and 90.9% inhibition at 0.1-, 1- and 10-times the IC50. Graph paper to this design was generated ‘in house’ some years ago. Data point sets for two tight-binding inhibitors and another set for a borderline case are shown in this format in Figure 6.2, along with three theoretical ‘curves’ (curved/inclined lines). The thinner central curve should fit observations when an enzyme present at 5 nM, acting on a negligible concentration of substrate, is inhibited by a compound with an equilibrium inhibition constant, Ki, equal to 5 nM: the IC50 of 7.5 nM is only a 1.5-fold underestimate of its true dissociation constant. If a second compound binds 1000-fold tighter, i.e. Ki= 5 pM, the thick curve on the left shows that, under the above conditions, the observed IC50 would be 2.5 nM—the limiting condition corresponding to half-an-equivalent referred to above. Tight binding thus limits the apparent potency advantage of the second compound, over the first, to 3-fold rather than the 1000-fold which would be observed with ‘infinitely’dilute enzyme solutions. The thick ‘curve’ on the right for a compound with Ki=100 nM differs only minutely from linearity. Experimental data points shown for three compounds, and other observations, pointed to the presence of roughly 3–5 nM binding sites (not necessarily active enzyme) in our typical AR1 test milieu, so compounds with IC50 values less than ~20 nM could, for the best comparisons, be

Introduction to the principles of drug design and action

270

corrected to non-tight binding values, preferably nowadays by computed data-fitting techniques. Assuming a Km for androstenedione of 40 nM, one can estimate pKi (−log Ki) values for the

Figure 6.2 Effects of ‘tight binding’ on % inhibition. compounds in Figure 6.2; in sequence they are: very approximately 10.3, approx. 9.3 and 8.1. For tight-binding corrected pIC50 (−log IC50) values subtract 0.3 from pKi. The data in Figure 6.2 show that the imidazole (6.20), X=Y=CH, may be the most potent inhibitor yet described. Modelling studies indicated that additional lipophilic substituents/fusions at C4 and/or C5 could produce yet further substantial

From programme sanction to clinical trials

271

improvements—but these ideas were not pursued because ICI 207658 had been identified as a compound with great potential and, more to the point, the improvements we most sought in vivo required an approach with reduced susceptibility to oxidative metabolism at its core. We returned to that task much later. The in vitro binding/inhibitory sequence: imidazol-1-yl>triazol-1-yl>triazol-4-yl, with 10- to 30-fold gaps was a consistent finding in our work; 5-methylimidazol-1-yl may sometimes be superior to its parent, but tight binding usually makes this uncertain. Each of these azoles attached to the naphthol-lactone framework had good potency in OI3 (shorter-term test) with ED50s of 250 to 500 µg/kg, but only the triazol1-yl derivative retained its potency in OI2 (single dose given 20 hours earlier than the time of dosing in OI3). In this series and in most others, the triazol-1-yl compounds were equal or superior to the imidazoles in vivo despite the order of magnitude disadvantage in binding affinity. Advantages were most marked in the OI2 test. We believe relatively easy oxidation of the imidazole ring and/or the linking methylene group to be chiefly responsible for this disparity since the imidazole/triazole activity ratio is at its most extreme with the most robust frameworks. Robustness here is based partly on chemists’ qualitative judgement and partly on observed plasma half-lives. In the naphthol-lactone case the imidazole had equal activity to the triazole but this still suggested easier than desired attack on the framework. Much later that idea was supported by a brief toxicology/pharmacokinetic once-per-day(u.i.d) oral dosing study on the N1-linked triazole, which returned a t1/2 of less than one hour in male rats. The N4-linked triazole behaved worse than the imidazole in terms of its OI3/OI2 ratio (value for N4-linked triazole was ≥8), so it was expected to be rapidly cleared and to produce little toxicity in a similar study, which likewise involved u.i.d oral dosing. In fact it produced unacceptable liver toxicity. It seems likely that binding to haem-linked iron is disfavoured by haem-N(δ-) to azole-N(δ-) repulsion, see (6.21), whereas when this heterocycle is a ligand to other metallo-enzymes (not just ironbased enzymes) such repulsion could be less or even attractive if an alcohol, water or amide ligand is present—as shown for R-OH in (6.22). One should further reflect that the average initial unbound state of the compound in vitro is essentially equal to that of an aqueous solution [though perhaps not for the highly lipophilic N-’benzyl’ imidazole (6.7)], and the nitrogen lone pairs are solvated by hydrogen bonds to water—which is an excellent proton donor; so, since transfer to the enzyme-bound state involves loss of that solvation, the lack of a hydrogen bond or some equivalent in the bound state means that potency must suffer. This seems to apply in our case, but the SAR of the latest bis(4’-cyanophenyl)-methylazole CibaGeigy inhibitors (Lang et al., 1993), in particular the very high in vitro potency of the tetrazol-2-yl derivative CGS 45688, implies either that some H-bond replacement occurs or, perhaps more likely, there are marked conformational energetic/steric effects favouring the additional ring nitrogens in that particular compound 6.4.5 The design principles behind ICI 207658—later named Arimidex The design of the naphthol-lactones, based on molecular modelling a wide selection of the large number of inhibitors then known, had in vitro potency as its dominant feature. But one has good reason to hope that selectivity will increase as potency

Introduction to the principles of drug design and action

272

increases. Testing for selectivity is time and resource consuming and can look at only a very small fraction of relevant P450s, let alone other enzymes. Effects due to receptor binding should also be considered—particularly if the drug is structurally similar to the substrate. The modelling approach by its nature has little, at least initially, to do with avoidance of metabolism, and ‘adding on’ this feature at some later point is seldom likely to be easy. Sustituting fluorine for hydrogen, especially in aromatic rings or as in CF3 or CF2 groups is widely practised and comes closest to a panacea—usually tolerable changes in size and lipophilicity (Edwards, 1994)—but synthetic difficulties and cost are often prohibitive

and, like all supposed panaceas of the author’s experience, it often seems least attractive when it is most needed. Independent of modelling there are useful general principles that can be applied to seek extra selectivity. These may lead to the achievment of established selectivity requirements and perhaps also help avoid unexpected toxicity and nonpharmacologically-related side effects: • without sacrificing much in vitro potency relative to a (notionally) iso-lipophilic comparator—flexibility, more specifically easily accessible conformational space, should be minimised; this is often a basis for improved potency. • without sacrificing much in vitro potency relative to a (notionally) iso-lipophilic comparator—one or more low-flexibility groups may be introduced which have high steric and/or solvation demands. • polar or more specifically hydrogen bonding atoms and groups should be preferred to lipophilic isosteres and they should be well dispersed throughout the molecule; here the potency criterion is more complex due to competitive phase effects: reducing lipophilicity has to be consistent with maintaining adequate in vivo starting state to target-bound state energy differences, steady state assumed. For a set of analogues with positive logPoctanol values this approximates, on a pIC50/logP plot, to being on the high-potency/low-logP side of a line of near unit slope (e.g. with slope near 0.7), drawn through data-points for compounds of substantial current interest. The problem remains how to achieve at least some of these aims while maintaining or preferably improving resistance to oxidative metabolism and other clearance processes. In this regard, one perhaps widely useful group was first recognised as a result of the attempted synthesis of tetralone (6.24), see Scheme 6.4. The tetralone and related naphthol-lactone targets had been conceived and worked on together, but a synthetic

From programme sanction to clinical trials

273

intermediate to the tetralone, (6.23), was early on converted to the imidazole (6.25). In view of the widespread activity in compounds of this type, it was not surprising to find good potency in AR1: IC50=4 ng/ml, but the ED50 in OI3 of 0.5 mg/kg was somewhat surprising. It was more surprising when compared to the results that had been obtained several years earlier with N-4’-cyanobenzylimidazole—one of the compounds that had shown unacceptable liver effects with multiple high doses. That very simple compound had a better IC50 (≤1 ng/ml), but a slightly worse ED50 of 1 mg/kg. We expected the cyanophenyl compound to be metabolically the more robust and so probably should have been the more potent in OI3 and OI2—in line with its higher enzyme inhibitory potency. In OI2 (single doses given 20 hr prior to that in OI3), both were classed as inactive (ED50≥2.5 mg/kg), so their relative potency in this test was unknown; further

work at higher doses was not done because neither had the required ‘duration’ characteristics. The unexpected potency reversal was rationalised by assuming that the 1-cyano-1methyl-ethyl (CME) group was itself not easily attacked and in addition probably conferred some steric and electron-withdrawing protection to the benzenoid ring and its additional substituent. The modest difference in lipophilicity might then be invoked, through an increase in apparent volume of distribution and a lower rate of clearance of unchanged drug, to account for the potency reversal seen in OI3, where duration of action is moderately important. Regarding lipophilicity, the f-values for the CME group are +0.25 for octanol/water (measured) and −0.4 for hexane/water (estimated) compared to −0.40 and −0.99 (both averages of several measured values) respectively for aromatic cyano. The modest octanol/hexane difference for the CME group indicates that its presence in a molecule should not of itself seriously compromise penetration through lipid membranes; this bodes well for oral absorption and rapid distribution. Another factor assisting absorption, good solubility, could arise from the easy rotation of the strongly anisotropic CME group about the Ar-C bond in solution: this will help to keep melting points low since easy rotation in the crystal

Introduction to the principles of drug design and action

274

environment is highly unlikely and entropic factors therefore favour non-crystalline states, e.g. non-glassy melts and solutions. Melting points, together with logPoctanol values, are inversely correlated with log(aqueous solubilities). That the geminal methyl groups, relative to most aliphatic groups, could have substantial resistance to oxidative attack follows from bond energetics—primary C-H bonds are the strongest—but also from a consideration of how the activation energy for hydrogen transfer is influenced by its surroundings. The forming H—OFe bond is highly polar and electron transfer to the extremely electrophilic O=FeIVporphyrin+ • runs ahead of nuclear motion; this increases the fractional positive charge on the methylene group and the migrating H atom. The reaction-generated electric field and changes in fractional charges are opposed by the strong, electron-withdrawing field due to the cyano group (σF≈0.57) and also by the weaker electron-withdrawing field associated with the aryl ring (σF≈ 0.13); see (6.26).

The intervening quaternary carbon atom somewhat distances(insulates) the methyl groups from these field effects, but with an expected ‘transmission coefficient’ of ~0.4, one could still expect substantial protection. Similar lines of argument apply to the benzylic methylene hydrogens with the more electron withdrawing triazole, now with no ‘insulating’ atom, inhibiting oxidation better than imidazole: σF≈0.49 and 0.35 respectively (azole σF data are from the Ph.D. thesis of D J Hall, University of Wales at Bangor, 1990). The size and hydrophilicity of the azole rings also hinder oxidative metabolism of the methylene groups, but attack is speeded by their weak resonance stabilisation of the transition state to the intermediate radicals. Fortunately the lastnamed effect is not dominant in these systems. The effect of the 1-cyano-1-methyl-ethyl (CME) group on the ease of oxidation of the aryl rings is also expected to be substantial: thus for CH2-CN, σF≈0.23 and σp+≈0.16 (more deactivating than an aromatic chloro-substituent: σp+=0.11). The CME group is expected to be equally deactivating and even non-protonated imidazolylmethyl, and more so the triazolylmethyl group, will further reduce rates of electrophilic (oxidative) aryl substitution. Steric hindrance around the cyano gives confidence that hydrolytic or other nucleophilic attack at this group could be minimal, and no easy metabolic release of cyanide is predicted—unlike benzyl cyanide where oxidation at the relatively unhindered benzylic C-H bonds produces a cyanohydrin, which allows easy release of cyanide and potential acute toxicity. Guided by the above ideas on selectivity and previous SAR, we thus set out urgently to synthesise the 3,5-bis-CME analogue of (6.25) and to make triazole equivalents. The triazol-1-yl analogue of (6.25) was disappointing but ICI 207658, (6.27), (see Table 6.1), was very potent in vitro and, more importantly at that time, in

From programme sanction to clinical trials

275

vivo it was equipotent with CGS 16949A in the demanding OI2 test, both having ED50≈15 µg/kg. Clearly this was very exciting! In preliminary studies in male rats, at extremely high multiples of the effective dose in females, ICI 207658 showed no untoward effect; in particular liver triglycerides remained normal. A small increase in liver weight was consistent with the observed induction of mixed function oxidases. Adrenal and other organ weights were the same as controls. The reason for using males for this and the many other compounds investigated is that oestrogens indirectly regulate both adrenal weight and circulating triglyceride concentrations. We hoped that changes in the background levels in males would cause only minor changes in these parameters of central interest. Problems arising from the use of different sexes are usually minor in most species except for rat: males frequently clear compounds faster than females and maximum plasma concentrations, Cmax, are often lower. The effects are multiplicative on AUC (area under curve of plasma-concentration vs. time). Thus simple ratios of effective dose in one sex to side-effect/toxic dose in the other can sometimes mislead selectivity assessments. The same cautions and others are more widely known to apply to comparisons across species. This discussion of selectivity ratios benefits greatly from hindsight and is relevant here mainly to the Team’s first development compound described in Section 6.5. The retrospectroscope also indicates that our clamour for pharmacokinetic and preliminary toxicological studies, which exceeded the capabilities of the appropriate Safety of Medicines Department workgroup to respond, contributed to some insecure conclusions. Most such studies took place later than initial in vivo selectivity studies and were not chiefly driven by the need to better assess selectivity ratios. In the case of ICI 207658, Cmax was lower in males than females by two- to threefold, but half-lives of ~6 hr, dropping to ~4 hr at the end of the multiple-dose study were reported to be essentially the same in both sexes. The results of this preliminary study supported and expanded the basis for the Team’s conclusion that ICI 207658 was a very promising compound. Half-lives in rat usually are much shorter than in man so the above values, while being short of our target range, were not a cause for concern and it was predicted by our Safety of Medicines experts that induction of mixed function oxidases in liver was most improbable at the very low predicted human therapeutic dose. It was subsequent data from studies in dog and pigtailed macaque monkey—producing half-lives of ~8 hr and 7 to 10 hr respectively—that seemed to indicate a remarkably uniform half-life across species and led to a majority view that the compound could not be relied on to achieve our target minimum half-life in patients. Well before any of the pharmacokinetic data were available we had discovered that ICI 207658 occupies a pinnacle of in vivo SAR space; not a single analogue came within an order of magnitude in OI2 potency terms! This is not the place to go into detail so data on just a few compounds are shown in Table 6.1. Well over a hundred analogues were made with small and sometimes larger variations at every locus where change is possible. If you can think up a related structure, we probably made it or tried to make it (one rather obvious analogue is an exception—that involving changing cyano to nitro—we never did attempt to make it, despite it being one of our listed targets for a long time). Even

Introduction to the principles of drug design and action

Table 6.1 Potency of selected azoles. Compound 1-(Rα-methyl)-3-CME-5-Rm number benzene α R Rm

in vitro AR1 IC50 ng/ml triazol-1-yl CMEa (ICI 207658) 4 (6.27) imidazol-1-yl CME ~0.2b (6.28) triazol-3-yl CME 500 (6.29) triazol-1-yl CH2-S-CH3 5 (6.30) triazol-1-yl C(CH3)2-OH 15 (6.31) triazol-1-yl C(=O)-CH3 2 (6.32) triazol-1-yl C(CH3)2-COCH3 ~0.8b (6.33) a CME represents a 1-cyano-1-methyl-ethyl group b estimated values (adjusted for tight binding)

276

in in vivo vivo OI2 OI3 ED50 ED50 mg/kg mg/kg 0.015 0.015 ~0.5 >1 >1 ~0.5 ≥1 >0.5 ~0.2

such small changes as homologating one of the four methyl groups to an ethyl group, or converting two geminal methyls to cyclo-alkyl (3- or 4-membered), or introducing an ortho- or para-bromo substituent, or changing the positions around the benzene ring, etc. Many active compounds were identified, but none was as supremely effective as ICI 207658. Analogues of CGS 16949A (6.18) containing one or two m-CME groups had, consistent with modelling work, significantly inferior AR1-potency. As can be seen even from the very small data-set in Table 6.1, poor in vivo potency was rarely attributable to inadequate enzyme affinity. The early hypothesis concerning the properties of the CME group, now groups, with regard to in vivo handling of the drug stood the test of time—albeit one or two compounds with good and even excellent AR1 figures, but poor OI2 results, remain difficult to explain. Single test results may be wrong, we seldom had good reason to retest, or perhaps the anomalies relate to rat vs. human aromatase selectivity, or perhaps our analysis is flawed. 6.5 BIS-TRIAZOLE (6.17)—A TIMELY AROMATASE COMPOUND IN DEVELOPMENT The antifungal lead had been converted to potential development candidate almost overnight, but that potential had to be assessed. At the time of the discovery of the bistriazole (6.17) no detail of the Ciba-Geigy compound was known, so AG was our yardstick and, because we were building up an ever-increasing body of data, it remained so for the first half of the resumed programme. Against that yardstick the limited in vitro selectivity data was pleasing, particularly with regard to P450scc/P450arom, where a 60-fold improvement over AG was seen. Most of the early efficacy data was very promising—and not just in rats: in monkeys dosed at 0.1

From programme sanction to clinical trials

277

mg/kg for 10 days a near maximum achievable reduction in oestrogens was achieved. Similarly potent effects were not seen in dog—perhaps because levels of testosterone, a precursor of oestrone and oestradiol, increased 5- to 10-fold in a dose-related fashion with drug treatment. Literature reports ascribed the testosterone increases to hypothalamic aromatase inhibition so we tried to use this as a test system. Unfortunately, near maximal testosterone levels were seen only after multiple doses of 1 mg kg−1 day−1. In view of the very high blood levels achieved in dogs this relatively massive required dose seemed unlikely to correspond simply to aromatase inhibition; we therefore placed little reliance on intra-species dog selectivity ratio assessments. It is also possible that dog aromatase differs substantially from the rat and human enzymes. In preliminary seven-day toxicity studies in rat, there were the expected increases in mixed function oxidases and increased liver weights at high doses, but, in the absence of effects on triglycerides, these were acceptable findings. The only slight concern expressed in the proposal for development was an increase in adrenal weights: in male rats at 50 mg/kg, 250-times the OI2 and placental enlargement ED50-doses; in dogs at 10 mg/kg. The problem of rat sex differences in drug handling was substantial since both Cmax and half-life in males were a third of those in females. AUC is therefore an order of magnitude lower and if tissue levels daily fall below some critical threshold for long enough, it is possible for body systems to largely recover from ‘toxic’ effects. Small but significant reductions of the male rat accessory sex organ weights and testosterone and LH plasma levels at all doses down to and including the lowest tested, 0.1 mg/kg, were regarded as toxicologically inconsequential. Since other aromatase inhibitors tested had no such effects, these findings demonstrate a lack of selectivity, but in what way remains uncertain. It may be that, akin to AG, changes in P450mediated rates/routes of hormone catabolism are occurring. Akin to this, interference with barbiturate metabolism in young male rats was evident at 1 mg/kg, but not at 0.1 mg/kg. The Team’s development proposal was accepted by higher management and, only 15 months after restarting the programme, the Team had a compound in full development! With luck and rapid development we might still achieve commercial success—but the Ciba-Geigy compound, seemingly superior in potency and selectivity, was now clearly well ahead in the race. And there were still so many hurdles for the bis-triazole (6.17) to clear. As the toxicity studies with (6.17) proceeded, the tally of adverse findings increased and our understanding of the unusual steroidogenesis in rat adrenal increased, leaving us with concerns about our ability to detect adverse changes relevant to other species, particularly man. In dogs, hypokalaemia was seen at modest doses and adrenal cortex vacuolation was slightly elevated from controls even at 0.5 mg/kg. It seemed certain that inhibition of 11-hydroxylation was to blame, but there was also no doubt that matters were made worse by the progressively higher Cmax and AUC values which follow daily dosing of any long half-life compound. In our dogs the half-life was 2–4 days so substantial accumulation would have occurred. This is the reverse situation to that described above for male rats and emphasises the importance of temporal drug level profiles to safety/selectivity assessments. Such

Introduction to the principles of drug design and action

278

profiles are also very important to some chronic efficacy studies: (6.17) has a half-life in pigtailed monkeys of one day so, barring enzyme induction—which is unlikely at the low doses used—there will be no gaps in drug cover and with chronic dosing a two-fold elevation of Cmax and AUC should occur. As stated previously, it is almost maximally effective with once-daily doses of 0.1 mg/kg. Contrast this with CGS 16949A: we found it to have a half-life of ~5 hr in female rats but less than 2 hr in monkeys; its large advantage over (6.17) in OI2 and still greater advantage (40-fold) in OI3 is reversed in monkeys—they require 0.1 mg/kg every 12 hr to achieve near maximal reduction of oestrogen levels. This competitor compound mirrored our own in steadily revealing its weaknesses throughout the time of the bis-triazole (6.17) development. Our work in rats and dogs revealed increasing selectivity issues and the absence of, from our viewpoint, relevant selectivity data from both oral presentations and publications dealing with CGS 16949A (fadrozole hydrochloride), including human studies, encouraged us in due course to review our priorities. As a business we had had many adverse experiences with long half-life compounds in chronic (6 month and more) toxicity studies. This was not to be an exception. In dogs, serum cholesterol and triglyceride levels were reduced by modest doses of the bis-triazole (6.17) and, by six months, cataracts were seen in the eyes of most dogs dosed at 7.5 mg/kg. The new fibre cells, which in mammals continuously enlarge the eye lens during life, need to synthesise their own cholesterol because they incorporate it in large amounts into their membranes and, being an avascular tissue, they cannot obtain it from the low density lipoprotein (LDL) in plasma. A prolonged shortage of cholesterol in this tissue seems to lead to cataracts. We suspect that at the observed high plasma levels in dogs, (6.17) inhibits one or more of the P450-dependent enzymes that transform lanosterol to cholesterol. The lead was born from a poor fungal lanosterol-14-methyl demethylase inhibitor and died, 18 months into development, due, probably, to inhibition of a canine lanosterol-demethylase. In passing it is worth noting the large number of conformers easily accessible to (6.17) and its multiple chelation possibilities—bidentate and even tridentate. Perhaps these facts contribute to its inadequate selectivity. 6.6 THE SEARCH FOR THE IDEAL BACK-UP CANDIDATE Inhibition of mammalian cholesterol synthesis had no precedent in aromatase inhibitors prior to the findings with (6.17), but we now urgently needed to look at possible successors to (6.17) in this new light. What had we found in that category during those 18 months? Not a lot. We found as expected that we could improve on the original naphthol-lactones (6.20)—but not sufficiently to fall into the presently required category. Potency in OI2 and PE9 had been somewhat improved, those improvements being associated with electron withdrawing substituents at C7. The better substituents, e.g. cyano, should hinder oxidative metabolism of the aromatic ring system and the proximal benzylic methylene. During this synthetic work we were surprised, following nitration or bromination, to observe amongst several products

From programme sanction to clinical trials

279

some substitution at the very sterically hindered C9 position; mostly reaction was at C7. As in the ICI 207658-like series, we made a large number of analogues but failed to make significant headway. We concluded, using an estimated σF value for the phenolic-lactone group, that extra protection of the gemdimethyls and the aromatic system was desirable. But as we saw no practical way to provide it we ceased work on this series. On reflection there were more changes we might have tried. The only blemish on the profile of ICI 207658 was its projected ‘inadequate’ halflife in man; in all in vitro selectivity tests conducted, including now against cholesterol synthesis, it performed superbly. None of a great many analogues was attractive. What else might be done? There were clues to hand: androstenedione had been synthesised with deuterium or tritium in specific locations as part of the aromatase mechanistic studies. Kinetic isotope effects were seen. Of most relevance to us, the 19-trideutero compound in admixture with the non-deuterated parent showed an intermolecular isotope effect kH3/kD3 approaching 3-fold, and the first hydrogen removal is known to be rate limiting. If this ratio, or even half the ratio, applied to the half-life of a deuterated ICI 207658—in at least two of the species used previously—might we not be home and dry? A quick back-of-the-envelope calculation showed that the additional cost of even a tetradeca-deutero compound could be trivial—probably only about one penny/mg— and, with an increased half-life, there was reason to believe the daily dose might be only ~3 mg per patient. We made the three deuterated compounds (6.34), (6.35) and (6.36), henceforth referred to as D2, D12 and D14. The hydrogens attached to the triazole ring were not changed because the SAR and metabolite studies pointed firmly against metabolism in this ring being relevant. Several antifungal triazoles had half-lives in male rats of about a week.

Similarly the hydrogens directly attached to the benzene ring were not replaced because these are rarely subject to primary isotope effects: rate-determining attack takes place initially at carbon, on the π-system, and secondary isotope effects are typically too small for the present purpose. The D2 and D12 compounds were made despite concerns that wherever attack normally might occur, the partially caged substrate would still react at the remaining weakest point, so called ‘metabolic switching’, and so reduce any advantage. The

Introduction to the principles of drug design and action

280

likelyhood of metabolic switching in our target deuterated analogues also carried with it the risk of generating new, longer-lived, or more abundant metabolites with reduced selectivity or increased toxicity. Of course reduced toxicity is also possible: the subject has been reviewed by Pohl and Gillette (1984–5). Intramolecular isotope effects in P450-mediated oxidative reactions, as in nonenzymic haem-based model systems, can be very large: values around 20 are known and 5–10 are normal. In contrast, the isotope effect expressed in kcat for metabolism of phenylethane or α,α-dideuterophenylethane with a rabbit liver-derived P450LM2 enzyme is only 1.28. This indicates that at least one enzymic step with a large ‘commitment to catalysis’ precedes hydrogen abstraction in this case (White et al., 1986). It can be argued that the effect is small in this case because of the activated nature of the secondary, benzylic C-H bonds. It is however by no means exceptional and ICI 207658 contains a related if much deactivated, more hindered and more hydrophilic part-structure. In contrast to this low value a related study on trideuteromethoxy anisole showed an in vivo isotope effect of 10. Studies in vivo generally show less marked substrate dependence with smaller but still some substantial isotope effects. Increases in half-life of 1.5- to 2.5-fold are typical (Blake, Crespi and Katz, 1975). In such clearance processes one is dealing with multi-step events and the oxidative step is normally only partially rate determining. D B Northrop has developed a general equation, Equation [6.1], for the interpretation of isotope effects in multi-step reactions. The maximum rate, Dv, is controlled by the ratio of catalysis, R, which represents the ratio of the rate of the isotope-influenced catalytic step to the rate of the other forward steps contributing to the maximum rate.

(6.1) Octanol/water partition and in vitro aromatase inhibition studies showed the expected equivalence of all isotopic species. The effective size of the more slowly vibrating CD fragment is on average very slightly smaller than a corresponding C-H fragment, but the difference in non-covalent binding properties is well below the detection limit in most biological systems. In vivo potency and limited pharmacokinetic studies with the three compounds, mainly as single agents but sometimes as solid-solution mixtures (to avoid possible differential solubilisation and absorption of individual samples) produced somewhat confusing results. OI2 tests (necessarily using females) in head to head comparisons with ICI 207658 (DO) showed a 3-fold potency improvement for D2 and improvements of 3.5- and 2-fold for D14 on separate occasions. The result for D12 was identical to that of DO. This indicated the benzylic methylene as the main site of oxidation in female rats at very low (2, 5, 10 and 20 µg/kg), near-therapeutic doses. A very limited pharmacokinetic study compared the compounds at 1 mg/kg with historical data. Plasma levels of D12 in female rats were followed to 70 hours post dosing and showed no detectable isotope effect. A similar result applied to D2 from samples taken at 1, 2 and 8 hours post dosing in males—the timepoints at which data were available from the historical study of DO in males. Only D14 showed some

From programme sanction to clinical trials

281

effect: in males followed to 24 hr post dosing: the Cmax and AUC increased by 60% (up to 8 hr), but with no detectable change in half-life. In part because of the limitations of the severely resource-constrained pharmacokinetic study, we tried in a very few animals to use mass spectrometric analysis to follow intra-individual handling of mixtures of compounds. In male rats dosed with a solid-solution of DO with its D2 and D14 analogues, and using the historical data on DO for comparisons, the apparent isotope effects interpreted as halflives were 1.0 to 1.2 for the D2 compound (poor data due to plasma-related peaks) and 2.0 for D14. The same isotope effect, 2.0, was seen for a binary solid-solution of DO and D14 compounds. A similar experiment in one dog yielded an apparent isotopeinduced increase in half-life of 2.1-fold up to 12 hr post dosing, but decreasing beyond this time to an average of 1.7-fold over the full 24 hours of the experiment. Being encouraged by these sighting experiments, but realising their extreme limitations and the possible future need for more extensive work, we carried out a detailed analysis of the likely kinetic scheme for the overall process. Surprisingly, this revealed that results from these, at-first-sight, ideal experiments, involving intraindividual temporal changes in concentration ratios of compounds, cannot be unambiguously interpreted in terms of individual clearance rates or related half-lives. We were therefore left with insufficient solid evidence of benefit from deuteration and the undoubted penalties of increased compound costs, analysis costs and uncertainties with regard to Registration Authority views and delays. The approach was abandoned. The search for a better candidate continued mainly with cis-tetrahydronaphthalenes like (6.12), stilbenes related to stilboestrol (6.3) and corresponding reduced analogues with a 1,2-diarylethane framework. In many cases, compounds with a pyridine ring replacing one of the benzene rings, e.g. (6.37) (racemate), had excellent activity both in AR1 and in OI2. None of these compounds was satisfactory in all respects. In some the half-life was too long—longer than or similar to the bis-triazole (6.17) in the dog was now a near automatic bar to progression—while others had inadequate selectivity: like bis-triazole (6.17), the pyridine (6.37) substantially lowered serum cholesterol levels—by then a totally unacceptable encumbrance. Whether the effects on aromatase and cholesterol synthesis could be separated through resolution was not investigated. Time

Introduction to the principles of drug design and action

282

had almost run out. We had had to progress many compounds, first through larger scale synthesis, then often into semi-chronic and chronic tests before finding them unsatisfactory. Janssen also were now forging ahead with the very impressive but racemic triazole R76713, (6.38). Published information showed potent activity in volunteers, so they too were now far ahead of us and our limited selectivity data on the compound gave us no comfort whatever. We had to make a choice—now. That choice was by now almost inevitable: ICI 207658 was associated with only temporary accumulation following multiple large doses in dogs and it had an excellent selectivity profile. Selectivity had again come close to the top of the Team’s priorities. Once again life comes pretty much full circle and ICI 207658 was entered into development under the number D1033. Long term toxicity studies revealed no significant additional findings to those seen with shorter exposures, but comprehensive pharmacokinetic studies revealed that the preliminary half-life estimates in rat had been in error. Due to a plasma-associated material interfering with the assays, those estimates were generally too long; for example, the initial t1/2 in males is now known to be ~2 hr and a half-life of 2.3 hr was observed after long-term dosing. It is therefore almost certain that the isotope work did achieve its objectives. But had we pursued this line into development we may have been faced with a supra-optimal half-life in patients—see below. 6.7 INTO THE CLINIC Escalating dose studies in male volunteers confirmed our expectation of good absorption, rapid distribution and high bioavailability, so the compound progressed into the first patients. The benefits we had confidently expected to find materialised and with a half-life of two days it fitted our target for optimum use long-term in postmenopausal women. No serious side effects, enzyme induction or inhibition—barring aromatase—have been observed and no indications of a lack of selectivity have been seen at either the 10 mg or 1 mg u.i.d. doses investigated (Plourde, Dyroff and Dukes 1994). Since the lower dose gives >95% inhibition of aromatisation in biochemical studies in patients, and equivalent anticancer efficacy to the higher dose, this smaller quantity, corresponding to approximately 15–20 µg/kg, was chosen as the recommended dose for use of Arimidex in postmenopausal breast cancer. Arimidex was launched in the U.K. on 19th September 1994 and so became the first of the fourth-generation, potent, highly-selective aromatase inhibitors to achieve commercialisation. Mature results from large randomised clinical trials show that Arimidex treated patients have a significant survival benefit over patients treated with another endocrine agent. It is the first aromatase inhibitor to show such an advantage. Acknowledgements The author extends his thanks to Mr Mike Large and Mr Chris Green for their contributions to much of the chemistry described, to his biological colleagues headed by Dr Mike Dukes for their superb work and often heroic efforts, and to the host of

From programme sanction to clinical trials

283

others who make essential contributions to the appraisal of any potential new addition to the drug armamentarium. I also thank Dr Dukes for helpful comments and criticisms of the manuscript. Thanks and apologies are extended to the many other chemists who made contributions to the Team’s endeavours but whose work has here been so scantily reported. FURTHER READING Akhtar, M., Njar, V.C.O. and Wright, J.N. (1993) Mechanistic Studies on Aromatase and Related C-C Bond Cleaving P-450 Enzymes. Journal of Steroid Biochemistry and Molecular Biology 44, 375–387. Blake, M.I., Crespi, H.L. and Katz, J.J. (1975) Studies with Deuterated Drugs. Journal of Pharmaceutical Sciences 64, 367–391. Brodie, A., Brodie, H.B., Callard, G., Robinson, C, Roselli, C. and Santen, R. (eds.) (1993) Recent Advances in Steroid Biochemistry and Molecular Biology: Proceedings of the Third International Aromatase Conference; Basic and Clinical Aspects of Aromatase. Journal of Steroid Biochemistry and Molecular Biology, Volume 44(4–6). Castagnetta, L., D’Aquino, S., Labrie, F. and Bradlow, H.L. (eds.) (1990) Steroid Formation, Degradation and Action in Peripheral Tissues. Annals of the New York Academy of Sciences, Volume 595. Djerassi, Carl (1992) The pill, pygmy chimps, and Degas’ horse: the autobiography of Carl Djerassi. New York: BasicBooks. Edwards, P.N. (1994) Uses of Fluorine in Chemotherapy. In Organofluorine Chemistry: Principles and Commercial Applications, edited by R.E.Banks, B.E.Smart and J.C.Tatlow, pp. 501–541. New York: Plenum Press. Henderson, D., Philibert, D., Roy, A.K. and Teutsch, G. (eds.) (1995) Steroid Receptors and Antihormones. Annals of the New York Academy of Sciences, Volume 761. Lang, M., Batzl, Ch., Furet, P., Bowman, R., Hausler, A. and Bhatnagar, A.S. (1993) Structure-activity Relationships and Binding Model of Novel Aromatase Inhibitors . Journal of Steroid Biochemistry and Molecular Biology 44, 421–428. Plourde, P.V., Dyroff, M. and Dukes, M. (1994) Arimidex®: A potent and selective fourth generation aromatase inhibitor. Breast Cancer Research and Treatment, Special Issue: Aromatase and its Inhibitors in Breast Cancer Treatment, edited by A.M.H.Brodie and R.J.Santen, 30(1), 103–111. Pohl, L.R. and Gillette, J.R. (1984–85) Determination of Toxic Pathways of Metabolism by Deuterium Substitution. Drug Metabolism Reviews 15(7), 1335– 1351. Schenkman, J.B. and Greim, H. (eds.) (1993) Cytochrome P450. Handbook of Experimental Pharmacology, Volume 105. Wakeling, A.E. (1990) Novel Pure Antioestrogens, Mode of Action and Therapeutic Prospects. Annals of New York Academy of Sciences 595, 348–356.

Introduction to the principles of drug design and action

284

White, R.E., Miller, J.P., Favreau, L.V. and Bhattacharyya, A. (1986) Stereochemical Dynamics of Aliphatic Hydroxylation by Cytochrome P450LM2. Journal of the American Chemical Society 108, 6024–6031.

7. PRO-DRUGS ANDREW W.LLOYD and H.JOHN SMITH CONTENTS 7.1 INTRODUCTION

236

7.2 PRO-DRUG DESIGN

240

7.3 APPLICATION TO PHARMACEUTICAL PROBLEMS

241

7.3.1 Patient acceptability

241

7.3.2 Drug solubility

242

7.3.3 Drug stability

243

7.4 PHARMACOLOGICAL PROBLEMS

245

7.4.1 Drug absorption

245

7.4.2 Drug distribution

250

7.4.3 Site-specific drug delivery

253

7.4.4 Sustaining drug action

258

7.5 SUMMARY

259

FURTHER READING

259

7.1 INTRODUCTION Although pharmaceutical companies attempt to design and develop new chemical entities using rational and logical processes, very few of these compounds become clinically useful drugs because unpredictable interactions with biological systems reduces therapeutic efficacy and in many cases leads to undesirable toxicity. An alternative approach to enhance therapeutic activity relies on the chemical modification of known compounds to overcome the undesirable physical and chemical properties using pro-drug design. A pro-drug is a pharmacacologically inactive compound which is metabolised to the active drug by either a chemical or enzymatic process. Some of the early pharmaceuticals were found to be pro-drugs and this has led to the subsequent

Introduction to the principles of drug design and action

286

introduction of the metabolite itself into therapy, particularly in cases where the active metabolite is less toxic or has fewer side effects than the parent pro-drug. The administration of the active metabolite may also reduce variability in clinical response between individuals which is attributed to differences in pharmacogenetics, particularly in disease states. The earliest example of a pro-drug is arsphenamine (7.1) used by Ehrlich for the treatment of syphilis. Later Voegtlin demonstrated that the activity of this compound against the syphilis organism was attributable to the metabolite oxophenarsine (7.2). Arsphenamine was later replaced by oxophenarsine in therapy as the the metabolite was less toxic at the dose required for effective therapy.

(7.1) Other such discoveries have led to the development of complete classes of drug compounds. For example the development of present day sulphonamide therapy evolved from the discovery by Domagk in 1935 that the azo dye prontosil (7.3) had antibacterial activity. Prontosil was subsequently shown to be a precursor which was metabolised to the active agent, p-aminobenezenesulphonamide (7.4), in vivo. This led to the subsequent development of a wide range of therapeutically superior sulphonamides through modification of the aminobenzenesulphonamide molecule.

(7.2) The antimalarial drugs pamaquin (7.5) and paludrine (7.7) are also both converted to active metabolites by the body. Pamaquin is dealkylated and oxidised to the quinone (7.6) which is 16 times more active in vivo than the parent compound whereas paludrine cyclises to give the active dihydrotriazine (7.8) which has structural similarities to the active antimalarial pyrimethamine (7.9).

Pro-drugs

287

(7.3)

(7.4) The dihydrotriazine metabolite, cycloguanil (7.8) has been administered as the insoluble pamoate salt in an oily base through a single intramuscular injection to provide malarial protection for up to several months depending on the particular particle size of the drug substance. During the development of depressants trichloroethanol (7.11) was shown to be the active metabolite of the once used hypnotic chloral hydrate (Noctec®) (7.10). This led to the use of trichloroethanol acid phosphate (7.12) for patients where choral hydrate was found to be either unpalatable or caused gastric irritation.

(7.5)

Introduction to the principles of drug design and action

288

The antiepileptic activities of methylphenobarbitone (Prominal®) (7.13), primidone (Mysoline®) (7.14) and methsuximide (7.15) have also been shown to be related to the plasma levels of active metabolites. The active metabolites are obtained on demethylation of methylphenobarbitone and oxidation of primidone. Methsuximide is also demethylated and at steady state the metabolite of this compound has been shown to be present at 700-fold greater concentrations than the parent drug.

The non-steroidal anti-inflammatory drug sulindac (Clinoril®) (7.16) is also a prodrug which is reduced to the active metabolite (7.17) although some of the inactive sulphone (7.18) is formed by oxidation.

The in vivo hydrolysis of aspirin (7.19) to salicylic acid (7.20) by esterases allows the administration of aspirin in preference to salicylic acid which is more corrosive to the gastrointestinal mucosa.

(7.6) Hexamine (Hiprex®, Mandelamine®) (7.21) is administered as a pro-drug of formaldehyde (7.22) for the treatment of urinary tract infections although it was initially used to dissolve renal stones. An enteric coat is used to protect the pro-drug

Pro-drugs

289

from stomach acid; however on reaching the acidic environment of the urine the formaldehyde is released where it exerts its antiseptic action.

(7.7) Phenylbutazone (Butozolidine®) (7.23) is converted by the body into the two hydroxylated forms, oxyphenbutazone (7.24) and (7.25). The drug is used in therapy under hospital supervision, mainly as an antiinflammatory agent, and this activity resides in form (7.24). However, another use of the drug is as a uricosuric agent, in the treatment of gout, and this action is attributable to the form (7.25). The observation that substitution in the side-chain of phenylbutazone results in enhanced urincosuric action has led to the discovery of several other agents which have this action, in particular sulphinpyrazone (7.26). In addition to those drugs detailed above several drugs which were metabolised to active compounds were initially considered to be pro-drugs but later shown to possess activity themselves. For example, phenacetin (7.27), an analgesic and antipyretic agent, is mainly metabolized in the body to an active metabolite, N-acetyl-paminophenol

Introduction to the principles of drug design and action

290

(7.8) (paracetamol) (7.28), as well as to an inactive metabolite, the glucuronide of 2hydroxyl phenacetin (7.29), in small amounts.

Paracetamol has replaced phenacetin in therapy, since it is usually free from toxic effects associated with phenacetin, e.g. methaemoglobin formation. However, extensive hepatic necrosis may occur when overdoses are ingested since the normal biotransformation pathway (conjugation with glutathione) is then saturated and a highly reactive metabolite is formed which binds irreversibly to hepatic tissue. More recent work has shown that phenacetin itself possesses antipyretic activity and that this activity is not dependent on metabolism to paracetamol.

Pro-drugs

291

7.2 PRO-DRUG DESIGN Most chemically designed pro-drugs are composed of two parts in which the active drug is linked to a pharmacologically inert molecule. The chemical bond between the two parts of the pro-drug must be sufficiently stable to withstand the pharmaceutical formulation of the pro-drug whilst permitting chemical or enzymatic cleavage at the appropiate time or site. After administration or absorption of the pro-drug, the active drug is usually released either by catalysed hydrolysis by the liver or intestinal enzymes or simply by hydrolysis although reductive processes have also been utilized. Pro-drugs are most commonly used to overcome the biological and pharmaceutical barriers which separate the site of administration of the drug from the site of action (Figure 7.1). 7.3 APPLICATION TO PHARMACEUTICAL PROBLEMS The pharmaceutical problems that have been addressed using prodrug design include unpalatibility, gastric irritation, pain on injection, insolubility and drug instability. 7.3.1 Patient acceptability Unpleasant tastes and odours may often affect patient compliance. For example very young children generally require liquid medication since they are usually not amenable to swallowing capsules or coated tablets. Despite the life-threatening toxicity the antibiotic

Introduction to the principles of drug design and action

292

Figure 7.1 THE PRO-DRUG CONCEPT: A diagrammatic representation of the prodrug concept where a pharmaceutically active drug is converted to an inactive compound to overcome pharmaceutical and biological barriers between the site of administration and the site of action. chloramphenicol (7.30) it is still administered orally for the treatment of typhoid fever and salmonella infections. However the drug has an extremely bitter taste and is entirely unsuitable for administration as a suspension to such patients. To overcome this problem orally administered chloramphenicol is usually formulated as the inactive tasteless palmitate (7.31) or cinnamate (7.32) esters. The active parent drug is released from these compounds by esterases present in the small intestine.

Pro-drugs

293

The bitter taste of the antibiotics clindamycin and erythromycin have been similarly masked using the palmitate ester and hemisuccinate ester pro-drugs, respectively. The antimicrobial metronidazole (7.33) is another example of a drug with an unacceptably bitter taste. To overcome this problem it is administered as a suspension of benzoylmetronidazole (Flagel S®) (7.34). Likewise, ethyl dithiolisophthalate (Ditophal®) has replaced the foul smelling liquid ethyl mercaptan for the treatment of leprosy. The odourless inactive diisophthalyl thioester is metabolised to the active parent drug by thioesterases.

7.3.2 Drug solubility The formulation of insoluble compounds for parenteral delivery represents a major problem as the insoluble drug will have a tendency to precipitate on injection in an organic solvent. The solubility of such compounds may be improved by the use of phosphate or hemi-succinate pro-drugs. For example the insoluble glucocorticoids such as betamethasone, prednisolone, methylprednisolone, hydrocortisone and dexamethasone are available for injection as the water-soluble pro-drug in the form of the disodium phosphate (RO.PO32− 2Na+) or sodium hemi-succinate (RO.CO.CH2CH2COO− Na+) salts. The phosphate esters are rapidly hydrolysed to the active steroid by phosphatases, whereas the hemi-succinate salts are less efficiently hydrolysed by esterases, possibly due to the presence of an anionic centre (COO−) near the hydrolysable ester bond. The poorly water-soluble anti-inflammatory steroidal alcohol dexamethasone has been shown to rapidly (t1/2=0 min) liberate the active steroid in vivo when injected as the water-soluble phosphate (7.35).

Introduction to the principles of drug design and action

294

The water-soluble phosphate ester of the anti-inflammatory agent oxyphenbutazone (7.36) is rapidly hydrolysed in vivo and gives higher blood levels of oxyphenbutazone on oral or intramuscular administration than attained on administration of the same doses of the parent drug. Difficulties in the formulation of the anticonvulsant drug phenytoin (7.37) as a soluble injectable dosage form has led to the development of water-soluble pro-drugs which have been shown to have a superior in vivo performance in rats. The pro-drug is prepared by the reaction of phenytoin with an excess of formaldehyde to give the 3hydroxymethyl intermediate (7.38), which is unstable in the absence of excess reagent. Conversion of the intermediate (7.38) to the disodium phosphate ester pro-drug (7.39) gives a water-soluble derivative. This is metabolised in vivo by phosphatases to (7.38), which rapidly breaks down (t1/2=2s) at 37°C, (pH 7.4), to give the active drug, phenytoin.

7.3.3 Drug stability Many drugs are unstable and may either breakdown on prolonged storage or are degraded rapidly on administration. This is a particular problem on oral administration as drugs are often unstable in gastric acid. Although enteric coatings may be used, it is also possible to utilise pro-drug design to overcome this problem. For example, the antibiotic erythromycin is destroyed by gastric acid and, as an alternative to enteric-coated tablets, it is administered orally as a more stable ester.

Pro-drugs

295

The inactive erythromycin estolate (laurylsulphate salt of the propionyl ester), when administered as a suspension, is rapidly absorbed and the propionyl ester converted by body esterases to the active erythromycin. The propionyl ester gives higher blood levels after oral administration on an equi-dose basis than the acetate or butyrate esters. The ethyl succinate ester has also been used. 5-Aminosalicylic acid (mesalazine) is useful in the treatment of ulcerative colitis and to a lesser degree in the management of Crohn’s disease. It cannot be administered orally since firstly, it is unstable in gastric acid and secondly, it would not reach its site of action in the ileum/colon since it would be absorbed in the small intestine. Sulphasalazine (7.40), where mesalazine is covalently linked with sulphapyridine, is broken down in the colon by bacteria to the two components and in this way 5aminosalicylic acid is delivered to the required site of action.

However, sulphapyridine is responsible for the majority of side-effects attributable to this combination and is thought to have little therapeutic activity. An alternative prodrug, osalazine (7.41), consisting of two molecules of 5-aminosalicylic acid has been developed to overcome this problem. Reduction of the azo bond by the colonic microflora therefore liberates two molecules of 5-aminosalicylic acid. Mesalazine has also been administered orally as tablets coated with a pH-dependent acrylic-based resin which disintegrates in the terminal ileum/colon as the environment pH rises above pH 7.

Microbial metabolism of pro-drugs has also been utilised in the delivery of corticosteroids to the colon. Such compounds are generally readily absorbed from the upper gastrointestinal tract and therefore delivered ineffectively to the colon. Administration of corticosteroids, such as dexamethasone (7.42), as glycoside prodrugs overcomes these problems by reducing systemic uptake in the small intestine. Pro-drugs, such as dexamethasone-β-D-glucoside (7.43), are hydrolysed by the specific glycosidases produced by the colonic bacteria and the parent corticosteroid absorbed from the lumen of the large intestine resulting in much higher concentrations in the colonic tissues.

Introduction to the principles of drug design and action

296

More recently macromolecular pro-drugs have been investigated as means of overcoming instability and undesirable systemic uptake. For example, 5aminosalicyclic acid has been linked to poly(sulphonamidoethylene) to give another mesalazine pro-drug known as polyasa (7.44) which has been shown to have less side effects than sulphasalazine and is therefore better tolerated by patients found to be allergic to or intolerant of sulphasalazine. (7.40)

(7.9) Macromolecular prodrugs have also been investigated as a means of reducing degradation of drugs by gastrointestinal enzymes. For example, the coupling of the B chain of insulin to water soluble copolymers such as N-(2hydroxypropyl)methacrylamide or poly(N-vinylpyrrolidone-co-maleic acid) appears to reduce the susceptibility of the insulin B chain to degradation by brush border peptidases in vitro.

Pro-drugs

297

7.4 PHARMACOLOGICAL PROBLEMS There are a number of pharmacological problems which may be addressed by prodrug design. These problems may be either related to pharmacokinetic, pharmacodynamic or toxic properties of the drug. Inappropriate pharmacokinetics may result in an undesirable rate of onset or duration of action of a drug. Poor pharmacodynamics may be a consequence of inefficient or unpredictable drug absorption from the gastrointestinal tract, inappropriate distribution and variable bioavailability as a consequence of presystemic metabolism or the inability to reach the site of action from the systemic circulation, e.g. penetration of the blood brain barrier. Toxic side effects may be due to non-specific drug delivery to the site of action. 7.4.1 Drug absorption Many drugs are either poorly or unpredictably absorbed from the gastrointestinal tract resulting in variation in efficacy between patients. Pro-drug design has been utilised in a number of cases to optimise the absorption of such drugs thereby improving their bioavailability. Many penicillins are not absorbed efficiently when administered orally and their lipophilic esters have been used to improve absorption. However, simple aliphatic esters of penicillins are not active in vivo and therefore activated esters are necessary for release of the active penicillin from the inactive pro-drug. Ampicillin (7.45), a wide spectrum antibiotic, is readily absorbed orally as the inactive pro-drugs, pivampicillin (7.46), bacampicillin (7.47) and talampicillin (7.48) which are then converted by enzymic hydrolysis to ampicillin.

Introduction to the principles of drug design and action

298

The preferred pro-drug is pivampicillin since minimal hydrolysis occurs in the intestine before absorption into the systemic circulation. Pivampicillin, the pivaloyloxymethyl ester, contains an acyloxymethyl function which is rapidly hydrolysed by enzymes to the hydroxymethyl ester. This hemi-ester of formaldehyde, spontaneously cleaves with release of ampicillin and formaldehyde. In a similar manner, bacampicillin and talamipicillin are cleaved and decompose to give ampicillin together with acetaldehyde and 2-carboxybenzaldehyde, respectively. Acyclovir (7.49) has been widely used for the treatment of herpes simplex and herpes zoster infections. This pro-drug is activated through phosphorylation by the viral thymidine kinase to acyclovir monophosphate which is then converted to the triphosphate, which inhibits DNA polymerase, by host cellular enzymes. However the use of this drug has been limited to some extent by low oral absorption; only 20% of a 200 mg dose being absorbed and little improvement being seen with doses above 800 mg. This has led to the development of a range of acyclovir prodrugs including ‘6deoxyacyclovir’ (BW A515U; (7.50)) which has been used for prophylaxis of herpesvirus infections in patients with haematological malignancies. It is well absorbed orally and produces plasma concentrations of the drug which are much higher than those obtained by oral administration of acyclovir. The drug (7.50) is converted to acyclovir in vitro by xanthine oxidise. An alternative orally active pro-drug is valaciclovir (7.51), the L-valyl ester of acyclovir, which is rapidly hydrolysed by first pass intestinal and hepatic metabolism. The mechanism of this biotransformation has yet to be fully elucidated but is thought to be enzymatic in nature.

(7.10)

Pro-drugs

299

More recently famciclovir (7.52) has been licensed in the United Kingdom for the treatment of herpes zoster infections. Famciclovir is an orally absorbed 6-deoxy, diacetyl ester pro-drug of penciclovir (7.53). This pro-drug is rapidly deacetylated and oxidised in the intestinal wall and liver to give a systemic availability of pencyclovir of 77% on oral administration. In vitro studies suggest that aldehyde oxidase, rather than xanthine oxidase, is involved in the conversion of famciclovir to penciclovir in the human liver.

(7.11) Penciclovir is selectively phosphorylated by viral thymidine kinase in the same way as acyclovir. Although penciclovir triphosphate, generated by the phosphorylation of the monophosphate by cellular enzymes, is 100 times less efficient at inhibiting the DNA polymerase from herpes virus it has similar activity to acyclovir. This may in part be explained by the 10- to 20- times greater intracellular stability of penciclovir triphosphate compared to acyclovir triphosphate. Several 2′,3′-dideoxynucleoside analogues such as zidovudine (azidothymidine, AZT) (7.54) and 2′,3′-didehydro-3’-deoxythymidine (D4T) (7.55) have potent antiviral activity against human immunodeficiency virus (HIV). These compounds are phosphorylated intracellularly to the 5′-triphosphate derivatives which inhibits the viral reverse transcriptase. To achieve effective metabolic antagonism against reverse transcriptase the plasma concentration of these compounds must be maintained. However, this has proved difficult because of the rapid elimination and metabolism of these compounds. Furthermore, the undesirable side effects associated with such compounds has been attributed to elevated plasma concentrations of these drugs. In an attempt to overcome these problems and to improve oral bioavailability a number of workers have recently investigated the potential of ester pro-drugs of these compounds. These studies have demonstrated that such prodrugs increase the circulating half-life whilst limiting the elevation of the plasma concentration of the parent nucleoside. Some of the ester pro-drugs were also shown to have higher absolute oral bioavailabilities than the parent nucleoside drug.

Introduction to the principles of drug design and action

300

The use of these nucleoside analogues as antiviral and anti-neoplastic agents is also limited by their absolute requirement for kinase mediated intracellular phosphorylation. Nucleotide phosphates are unable to readily penetrate membranes and therefore have little therapeutic utility. This has led to the development of masked-phosphate prodrugs of anti-HIV nucleoside analogues, such as (7.56), which facilitate intracellular delivery of the bio-active free phosphate. These compounds have been shown to be 25 times more potent and 100 times more selective than the parent nucleosides. Unlike the parent drugs they also retain good activity against kinase-deficient cells. Such strategies also have important implications for the development of much wider ranges of compounds to combat the emergence of resistance to certain nucleoside analogues.

In another example, the antihypertensive effects on oral administration of the angiotensin-converting enzyme inhibitor enalaprilat (7.57) have been improved by conversion to the more efficiently absorbed ethyl ester, enalapril (7.58). In the active form, less than 12% is absorbed whereas the inactive derivative has an improved absorption of between 50% and 75%. The pro-drug enalapril is converted in vivo to the active enalaprilat by hydrolysis in the liver following absorption from the gastrointstinal tract.

Pro-drugs

301

Animal studies have shown that the oral absorption of certain basic drugs may be increased by the preparation of ‘soft’ quaternary salts. The ‘soft’ quaternary salt is formed by reaction between an α-chloromethyl ester (7.59) and the amino group of the drug. The quaternary salt formed is termed a ‘soft’ quaternary salt since, unlike normal quaternary salts it can release the active basic drug on hydrolysis.

(7.12) ‘Soft’ quaternary salts have useful physical properties compared with the basic drug or its salts. Water solubility may be increased compared with other salts, such as the hydrochloride, but more important there may be an increased absorption of the drug from the intestine. Increased absorption is probably due to the fact that the ‘soft’ quaternary salts have surfactant properties and are capable of forming micelles and unionized ion pairs with bile acids etc., which are able to penetrate the intestinal epithelium more effectively. The pro-drug, after absorption, is rapidly hydrolysed with release of the active parent drug as illustrated below.

(7.13) Such an approach has also been utilised to achieve improved bioavailability of pilocarpine on ocular administration. Pilocarpine is rapidly drained from the eye resulting in a short duration of action. The ‘soft’ quaternary salt (7.60) has a lipophilic side-chain which has been shown to improve absorption in rabbits and gives a more prolonged effect at one tenth of the concentration of pilocarpine. The action of this compound has been shown to be due to the release of pilocarpine on hydrolytic cleavage of the ester followed by release of formaldehyde.

Introduction to the principles of drug design and action

302

Topical administration is also used in the treatment of glaucoma with adrenaline (7.61). which lowers the intraocular pressure. Enhanced therapeutic efficacy may be achieved using a more lipophilic prodrug dipivefrin (7.62) which is 100 time more active than adrenaline as a consequence of more efficient corneal transport, followed by deesterification by the corneal tissue and release of adrenaline in the aqueous humor. Consequently lower doses of dipivefrin than adrenaline can be administered to achieve the same therapeutic effect. This offers advantages in reducing the side-effects associated with the use of adrenaline including cardiac effects due to systemic absorption and the accumulation of melanin deposits in the eye.

7.4.2 Drug distribution The modification of a drug to a pro-drug may lead to enhanced efficacy for the drug by differential distribution of the pro-drug in body tissues before the release of the active form. For example, more extensive distribution of ampicillin occurs in the body tissues when the methoxymethyl ester of hetacillin (a 6-side-chain derivative of ampicillin) is administered, than is obtained with ampicillin itself. Conversely, decreased tissue distribution of a drug may occur, as was observed when adriamycin as its DNAcomplex was administered as a pro-drug. Decreased tissue distribution restricts the action of a drug to a specific target site in the body and may therefore decrease its toxic side-effects, resulting from its reaction at other sites. Anticancer drugs can suppress growth in normal as well as neoplastic tissue. Improved selective localization has been achieved using non-toxic pro-drugs which release the active drug within the cancer cell as a result of either the enhanced enzyme activity in the cell or enhancement of reductase activity in the absence of molecular oxygen in hypoxic cells. The pro-drug cyclophosphamide (7.63) is used for the treatment of certain forms of cancer and as an immunosuppressant after organ transplant. It does not possess alkylating properties and consequently is not a tissue vesicant since the electronwithdrawing properties of the adjacent phosphono-function decrease the nucleophilic properties of the β-chloroethylamino-nitrogen atom and prevent formation of the reactive alkylating ethyleniminium ion. The pro-drug requires hepatic mixed-function oxidase-mediated metabolic activation to generate 4-hydroxycyclophosphamide

Pro-drugs

303

(7.64). The 4-hydroxycyclophosphamide exists in equilibrium with its open ring tautomer aldophosphamide (7.65) which undergoes β-elimination to produce the alkylating cytotoxic phosphoramide mustard (7.66) in the target cells.

(7.14) Cyclophosphamide is also metabolised by aldehyde dehydrogenase to the inactive carboxyphosphamide (7.67). Since this reaction provides a detoxification pathway, the effectiveness of cyclophosphamide is found to inversely correlate with the dehydrogenase activity of the target cells. The action of this alkylating species would be expected to be restricted to the target tissue but unfortunately in practice the action of the drug is more widespread and it shows toxicity to normal tissue, one of the apparent effects being alopecia. Recently the organic thio-phosphate pro-drug amifostine has been introduced as a cytoprotective agent to reduce the toxic effects of cyclophosphamide on bone marrow. Amifostine uptake into normal cells occurs by facilitated diffusion and is therefore more rapid than the uptake into tumour cells by passive diffusion. As tumour cells are often hypoxic, poorly vascularised and have a low pH environment they also have reduced alkaline phosphatase activity. Amifostine (7.68) exploits these differences in uptake and enzyme activity to ensure that the pro-drug is only dephosphorylated to the active drug in healthy tissues. The active drug therefore selectively deactivates the

Introduction to the principles of drug design and action

304

reactive cytotoxic species produced by cyclophosphamide in non-tumour tissue without compromising the efficacy of the chemotherapy.

(7.15) In addition, the acrolein produced from (7.65) was initially found to cause bladder trouble. This problem has been overcome by either administration of cyclophosphamide together with an alkyl sulphide (sodium 2mercaptoethanesulphonate, mesna, Uromitexan®) to remove acrolein as it is formed by addition to the β-carbon atom by a Michael reaction, or use of a modified cyclophosphamide (7.69) which does not form acrolein after ring opening.

The anticancer effect of the pro-drug procarbazine (7.70) has also been attributed to to the formation of a cytotoxic species in the target cells. In this case, procarbazine is metabolised by the mixed function oxidase to azoprocarbazine (7.71) which undergoes further cytochrome P450 mediated oxidation to azoxy procarbazine isomers (7.72, 7.73) which liberate the diazomethane alkylating agent (7.74) in the target cells.

Pro-drugs

305

(7.16) A series of other non-toxic nitrogen mustard pro-drugs have also been designed to regenerate the parent alkylating agent in neoplastic tissues by taking advantage of the difference in the level of enzymatic amidase between normal and neoplastic cells. N,N-diallyl-3-(1-aziridino)propionamide (DAAP) is active against certain forms of leukaemia but does not cause leucopenia, a common toxic side-effect observed with other bifunctional alkylating agents. This observation suggests that DAAP is selective in its action against dividing (neoplastic) cells where a high amidase level occurs. 7.4.3 Site-specific drug delivery Pro-drugs have more recently been used to achieve site-specific drug delivery to various tissues. Such pro-drugs are designed to ensure that the release of the active drug only occurs at its site of action thereby reducing toxic side-effects due to high plasma concentrations of the drug or non-specific uptake by other body tissues. This has led to the development of systems for site-specific delivery to the brain and to cancer cells.

Introduction to the principles of drug design and action

306

The blood-brain barrier is inpenetrable to lipid insoluble and highly polar drugs. Although lipophilic pro-drugs may be used to overcome this physiological barrier, the increased lipid solubility may enhance uptake in other tissues with a resultant increase in toxicity. Furthermore, therapeutic levels of such lipophilic pro-drugs can only be maintained if there is a constant plasma concentration. These problems may be overcome by utilising a dihydropyridine—pyridinium salt type redox system. This approach was first used to enhance the penetration of the nerve gas antagonist pralidoxine into the CNS using (7.75) a non-polar pro-drug which crosses the barrier, where it is rapidly oxidized to the active form and trapped in the CNS.

More recently this approach has been developed as a general rationale for the sitespecific and sustained delivery of drugs which either do not cross the blood-brain barrier readily or are rapidly metabolized. Phenylethylamine and dopamine have been used to illustrate the principles involved and in vivo work has been described in animal experiments. The delivery system is prepared by condensing phenylethylamine with nicotinic acid to give (7.76) which is then quaternized to give (7.77). The quaternary ammonium salt (7.77) is then reduced to the 1,4-dihydro-derivative (7.78). The prodrug (7.78) is delivered directly to the brain, where it is oxidized and trapped as the pro-drug (7.77). The quaternary ammonium salt (7.77) is slowly cleaved by enzymic action with sustained release of the biologically active phenylethylamine and the facile elimination of the carrier molecule. Elimination of the drug from the general circulation is by comparison accelerated, either as (7.77) or (7.78) or as cleavage products. This rationale removes excess drug and metabolic products during or after onset of the required action. This is in contrast to normal penetration of the brain by a drug from plasma, where plasma levels must be maintained to produce the required effect and which can cause systemic side-effects.

Pro-drugs

307

(7.17) In animal experiments the anti-inflammatory effect of topically applied hydrocortisone has been increased, and its systemic effects after absorption decreased, by use of the prodrug spirothiazolidine derivative (7.79). These beneficial effects are due to restriction of the action of hydrocortisone within the skin. After absorption, (7.79) is hydrolysed in a stepwise manner with eventual release of hydrocortisone within the skin from the accumulated pro-drug, resulting in a more intense anti-inflammatory effect and a decrease in its rate of leaching into the blood stream to produce systemic effects. The sustained release of hydrocortisone is due to retardation of the intermediate hydrolytic product (7.80) by disulphide formation (7.81) between its thiol group and a thiol group of the skin, followed by a slow breakdown of (7.81) to release hydrocortisone.

Introduction to the principles of drug design and action

308

(7.18) Success in cancer chemotherapy probably lies in utilizing differences in rates of growth between the rapidly-dividing tumour cells and the slower non-cycling normal tissue cells, as evidenced by responsiveness to chemotherapy of leukaemia and the high growth solid tumours. However, a different approach is needed for low growth solid tumours. The blood supply to large solid tumours is disorganized and internal regions may be non-vasculated and the cells, termed hypoxic, deprived of oxygen. Hypoxic tissues are known to have greater powers of reduction than oxygenated areas and the reduced species are expected to be stable in the absence of molecular oxygen, which could theoretically reverse the reduction process. This knowledge has been used in the development of a rationale for targeting drugs to the internal hypoxic regions of solid tumours, these regions being relatively inaccessible to drugs that are rapidly metabolized or strongly bound to tissue components. This approach could provide a selective chemical drug-delivery system when used in combination with treatments likely to be limited by the presence of hypoxic cells. Certain aromatic or heterocyclic nitro-containing compounds can be reduced in an hypoxic environment to produce intermediates which then fragment into alkylating species. The 2-nitro-imidazole compound misonidazole (7.82) is selectively cytotoxic to cultured hypoxic cells. Reduction of the nitro group to the hydroxylamine (RNH2OH) probably occurs, with further fragmentation occurring to give the DNAalkylating species including glyoxal ((CHO)2).

Pro-drugs

309

Nitracine (7.83) is another selective alkylating agent for hypoxic mammalian cells in culture after reduction, although the identity of the active species is unknown. Although nitracine is 105 times more potent than misonidazole in this system, it lacks activity in murine or human xenografted tumours.

Research has also been direct towards the bioactivation of aromatic nitrogen mustards, where the mechanism of action is predictable and the activation step occurs by reduction of a substituent group in the aryl ring. The alkylating ability of the βchloroethylamine side-chain is dependent on the electron density on the nitrogen. The p-nitro substituent in (7.84), by exerting an electron withdrawing effect, reduces the electron density on the nitrogen thereby inhibiting the formation of the alkylating carbonium ion. Reduction of the nitro group in (7.84) in an hypoxic environment removes its electron withdrawing effect and restores the ability of the compound to form the alkylating species via an SN1 reaction pathway. Whether reduction gives the hydroxylamine (7.85), or the amine (7.86) is uncertain, but both species have been calculated to have greater activity than the nitro compound.

The aziridine (7.87) may be activated in a similar manner and has been shown to be selectively toxic to hypoxic cells. It should be noted that the presence of additional groups in the aryl ring may affect the actual electron density on the nitrogen atom, and hence the reactivity of the alkylating species generated, despite the activation process occurring on reduction.

Introduction to the principles of drug design and action

310

Soluble macromolecular pro-drug delivery systems have also been developed to improve the pharmacokinetic profile of pharmaceutical agents by the controlled release of the active agent. It has been suggested that such soluble polymeric carriers have the potential to improve the activity of conventional antitumor agents. Recently the potential of N-(2-hydroxypropyl)-methacrylamide (HPMA) copolymers as carriers for the anti-tumour agent doxorubicin (DOX) has been investigated. Doxorubicin was linked to the polymeric carrier by peptidyl spacers designed to be cleaved by lysosomal thiol dependent proteases, which are known to have increased activity in metastatic tumours. Such conjugates have been shown to have a broad range of antitumour activities against leukamic, solid tumour and metastatic models. Fluorescein labelled HPMA copolymers have beeen shown to accumulate in vascularised stromal regions, particularly in new growth sites in the tumour periphery. Treatment of C57 mice bearing subcutaneous B16F10 melanomas with DOX-HPMA copolymer conjugate improved the treated to control lifespan by three fold with respect to that obtained on aggressive treatment with free doxorubicin. It has been suggested that these macromolecular pro-drugs reduce toxicity by controlled drug release following passive accumulation and retention within solid tumours. Recent research has been directed towards alternative approaches to obtain sitespecific activation of pro-drugs for cancer chemotherapy using antibody-directed enzyme prodrug therapy (ADEPT) (Figure 7.2). The ADEPT approach employs an enzyme, not normally present in the extracellular fluid or on cell membranes, conjugated to an antitumour antibody which localises in the tumour via an antibodyantigen interaction on administration. Once any unbound antibody conjugate has been cleared from the systemic circulation, a pro-drug, which is specifically activated by the enzyme conjugate, is administered. The bound enzyme-antibody conjugate ensures that the pro-drug is only converted to the cytotoxic parent compound at the tumour site thereby reducing systemic toxicity. It has been shown that in systems utilising cytosine deaminase to generate 5-fluorouracil from the 5-fluorocytosine pro-drug at tumour sites, 17 times more drug can be delivered within a tumor than on adminstration of 5fluorouracil alone. The ADEPT approach has been recently investigated as a means of overcoming the side effects of taxol which is an effective treatment for breast cancer but also attacks healthy tissues. The system utilises a β-lactamase enzyme antitumour antibody conjugate

Pro-drugs

311

Figure 7.2 ANTIBODY-DIRECTED ENZYME PRODRUG THERAPY (ADEPT): A diagrammatic representation of the ADEPT approach to cancer chemotherapy which employs an antitumour antibody conjugated to an enzyme. The conjugate is localised at the tumour site via an antibody-antigen interaction and converts a subsequently administered pro-drug into a cytotoxic agent which attacks the tumour. and a pro-drug (PROTAX) which consists of taxol linked via a short chain to cepham sulphoxide. Taxol is selectively released at the tumour site by the localised βlactamase enzyme which is not normally found in any other tissues. In studies on cultured human breast cells it has been shown the prodrug is almost as effective as taxol on cells which have been treated with the enzyme-bound antibody, however PROTAX alone is only a tenth as toxic to cancer cells as taxol and is therefore less likely to harm healthy cells. More recently advances in molecular biology have led to the development of a virus-directed enzyme pro-drug therapy (VDEPT) using suicide genes. Suicide genes encode for nonmammalian enzymes which can convert a pro-drug into a cytotoxic

Introduction to the principles of drug design and action

312

agent. Cells which are genetically transduced to express such genes essentially commit metabolic suicide in the presence of the appropriate pro-drug. Typical suicide genes include herpes simplex thymidine kinase and Escherichia coli cytosine deaminase. Viral vectors are used to carry the gene into both tumour and normal cells. Tumour specific transcription of the suicide gene is achieved by linking the foreign gene downstream of a tumour-specific transcription unit such as the proximal ERBB2 promoter. The ERBB2 oncogene is overexpressed in approximately a third of all breast and pancreatic tumours by transcriptional upregulation of the ERBB2 gene with or without gene amplification. In recent studies a chimeric minigene consisting of the proximal ERBB2 promoter linked to a gene coding for cytosine deaminase has been constructed and incorporated into a double-copy recombinent retrovirus. In vitro studies using pancreatic and breast cell lines have been used to demonstrate significant cell death on treatment of cells which expressed ERBB2 with the viral vector and 5fluorocytosine, whereas cells which did not express ERBB2 were not affected. 7.4.4 Sustaining drug action Pro-drug design has also been successfully used to modify the duration of action of the parent drug by either reducing the clearance of the drug or by providing a depot of the parent drug. The pro-drug bitolterol (7.88), which is the di-p-toluate ester of N-t-butyl noradrenaline (7.89), has been shown in dogs to provide a longer duration of bronchodilator activity than the parent drug. Furthermore, the pro-drug is preferentially distributed in lung tissues rather than plasma or heart so that the bronchodilator effect, following subsequent biotransformation of the pro-drug, is not associated with undesirable cardiovascular effects and is slow and prolonged.

The phenothiazine group of drugs, acting as tranquillizers, have been converted to long acting pro-drugs which are administered by intramuscular injection. Not only is the frequency of administration reduced but the problem associated with patient compliance is also eliminated. Flupenthixol (7.90) when administered as the decanoate ester (7.91) in an oily vehicle for the treatment of schizophrenia is released intact from the depot and subsequently hydrolysed to the parent drug, possibly after penetration of the blood-brain barrier. Maximum blood levels are observed within 11–17 days after injection and the plateau serum levels averaged 2–3 weeks in duration.

Pro-drugs

313

Similarly, perphenazine (7.92) has been used as the enanthrate ester (7.93) and pipothiazine (7.94) as the undecanoate (7.95) and palmitate (7.96) esters. Vasopressin has been used for the treatment of bleeding varicose veins in the lower end of the oesophagus (oesophageal varices), a condition which affects about 1000 individuals annually. The vasoconstrictor action of the drug stops the bleeding, but the action is of short duration and cannot be prolonged by the use of higher doses due to

the development of toxic side-effects. Glypressin, Gly-Gly-Gly-Lys-vasopressin, is an inactive pro-drug of vasopressin and after injection the glycyl residues are steadily cleaved off by enzymic action to release the active drug. A sustained low level of vasopressin is obtained in this manner, which is sufficient to produce the required vasoconstriction effect on portal blood pressure whilst minimizing the possibility of unwanted effects caused by high blood pressure.

Introduction to the principles of drug design and action

314

7.5 SUMMARY The examples given in this chapter illustrate the importance of the pro-drug concept as a means of overcoming pharmaceutical and pharmacological problems encountered during drug development. In addition, recent advances in biotechnology have made it possible to utilise pro-drug design to develop chemical drug delivery systems which provide various means of targetting the delivery of parent drugs to specific sites within the body. Clearly, the increasing demands for more efficacious and less toxic drugs will ensure that prodrug approaches continue to be exploited in the development of future drug substances. FURTHER READING Bagshawe, K.D., Sharma, S.K., Springer, C.J. and Rogers, G.T. (1994) Antibody directed enzyme prodrug therapy (ADEPT). Annals of Oncology 5, 879–891. Bodor, N. and Farag, H.H. (1983) Improved delivery through biological membranes. 11. A redox chemical drug-delivery system and its use in brain-specific delivery of phenylethylamine. Journal of Medicinal Chemistry 26, 313–18. Denny, W.A. and Wilson, W.R. (1986) Considerations for the design of nitrophenyl mustards as agents with selective toxicity for hypoxic tumour cells. Journal of Medicinal Chemistry 29, 879–87. Druzgala, P., Winwood, D., Drewniak-Deyrup, M, Smith, S., Bodor, N. and Kaminski, J.J. (1992) New water-soluble pilocarpine derivatives with enhanced and sustained muscarinic activity. Pharmaceutical Research 9, 372–377. Duncan, R. (1992) Drug polymer conjugates-potential for improved chemotherapy. Anti-Cancer Drugs 3, 175–210. Easterbrook, P. and Wood, M.J. (1994) Successors to acyclovir. Journal of Antimicrobial Chemotherapy 34, 307–311. Harris, J.D., Gutierrez, A.A., Hurst, H.C., Sikora, K. and Lemoine, N.R. (1994) Gene therapy for cancer using tumour-specific prodrug activation. Gene Therapy 1, 170– 175. Huber, B.E., Richards, C.A. and Austin, E.A. (1994) Virus-directed enzyme/prodrug therapy-selectively engineering drug sensitivity into tumors. Annals of New York Academy of Sciences 716, 104–114. Huennekens, F.M. (1994) Tumor targeting: activation of prodrugs by enzymemonoclonal antibody conjugates. Trends in Biotechnology 12, 234–239. McGuigan, C., Sheeka, H.M., Mahmood, N. and Hay, A. (1993) Phosphate derivatives of d4T as inhibitors of HIV. Bioorganic and Medicinal Chemistry Letters 3, 1203– 1206. Riley, T.N. (1988) The prodrug concept and new drug design and development. Journal of Chemical Education 65, 947–953. Seymour, L.W., Ulbrich, K., Steyger, P.S., Brereton, M., Subr, V., Strohalm, J. and Duncan, R. (1994) Tumor tropism and anticancer efficacy of polymer-based doxorubicin prodrugs in the treatment of subcutaneous murine B16F10 melanoma. British Journal of Cancer 70, 636–641.

Pro-drugs

315

Sinkula, A.A. and Yalkowsky, S.H. (1975) Rationale for design of biologically reversible drug derivatives: prodrugs. Journal of Pharmaceutical Sciences 64, 181– 210. Stella, V.J., Charman, W.N.A. and Naringrekar, V.H. (1985) Prodrugs. Do they have advantages in clinical practice? Drugs 29, 455–73. See references to other reviews cited therein. Waller, D.G. and George, C.F. (1989) Prodrugs. British Journal of Clinical Pharmacology 28, 497–507.

8. DESIGN OF ENZYME INHIBITORS AS DRUGS ANJANA PATEL, H.JOHN SMITH and JÖRG STÜRZEBECHER CONTENTS 8.1 INTRODUCTION

262

8.1.1 Basic concept

262

8.1.2 Types of inhibitors

264

8.2 GENERAL ASPECTS OF INHIBITOR DESIGN

271

8.2.1 Target enzyme and inhibitor selection

271

8.2.2 Specificity and toxicity

273

8.3 RATIONAL APPROACH TO THE DESIGN OF ENZYME INHIBITORS

275

8.4 DEVELOPMENT OF A SUCCESSFUL DRUG FOR THE CLINIC

277

8.4.1 Oral absorption

278

8.4.2 Metabolism

279

8.4.3 Toxicity

281

8.5 STEREOSELECTIVITY

281

8.6 EXAMPLES OF ENZYME INHIBITORS AS DRUGS

283

8.6.1 Protease inhibitors

283

8.6.1.1 Serine proteases

284

8.6.1.2 Metallo-proteases

299

8.6.1.3 Aspartate proteases

309

8.6.2 Acetylcholinesterase inhibtors

317

8.6.3 Aromatase inhibitors

320

Design of enzyme inhibitors as drugs

317

8.6.4 Pyridoxal phosphate—dependent enzyme inhibitors

322

8.6.4.1 GABA transaminase (GABA-T) inhibitors

325

8.6.4.2 Peripheral aromatic amino acid decarboxylase (AADC) inhibitors

327

8.6.4.3 Ornithine decarboxylase (ODC) inhibitors

329

FURTHER READING

330

8.1 INTRODUCTION Enzymes catalyse the reactions of their substrates by initial formation of a complex (ES) between the enzyme and substrate(S) at the active site of the enzyme. This complex then breaks down, either directly of through intermediary stages, to give the products (P) of the reaction with regeneration of the enzyme:

(8.1)

(8.2) kcat is the overall rate constant for decomposition of ES into products, k2 and k3 are the respective rate constants for formation and breakdown of the intermediate E' (i.e. kcat= k2k3/(k2+k3)). Chemical agents, known as inhibitors, modify the ability of an enzyme to catalyse the reaction of its substrates. The term inhibitor is usually restricted to chemical agents, other modifiers of enzyme activity such as pH, ultra-violet light and heat being known as denaturising agents. 8.1.1 Basic concept The body contains several thousand different enzymes each catalysing a reaction of a single substrate or group of substrates. An array of enzymes is involved in a metabolic pathway, each catalysing a specific step in the pathway, i.e.

Introduction to the principles of drug design and action

318

These actions are integrated and controlled in various ways to produce a coherent pattern governed by the requirements of the cell. The basis for using enzyme inhibitors as drugs is that inhibition of a suitably selected target enzyme leads firstly to a build-up in concentration of substrate(s) and secondly to a corresponding decrease in concentration of the metabolite(s), one of which leads to a useful clinical response. Where the substrate gives a required response (i.e. agonist) inhibition of a degradative enzyme leads to accumulation of the substrate and accentuation of that response. Build up of acetylcholine by inhibition of acetylcholinesterase using neostigmine is used for the treatment of myasthenia gravis and glaucoma (Equation [8.3]).

(8.3) Where the metabolite has an action judged to be clinically undesirable or too pronounced, then enzyme inhibition reduces its concentration with a decreased (desired) response. Allopurinol is an inhibitor of xanthine oxidase and is used for the treatment of gout. The inhibition of the enzyme decreases conversion of the purines xanthine and hypoxanthine to uric acid, which otherwise deposits and produces irritation in the joints (Equation [8.4]).

(8.4) In the above example an enzyme acting in isolation was targeted but several other strategies may be used with enzyme inhibitors to produce an overall satisfactory clinical response. The target enzyme may be part of a biosynthetic pathway consisting of a sequence of enzymes with their specific substrates and co-enzymes (Equation [8.5]). Here the aim is to prevent, by the careful selection of the target enzyme in the pathway (see Section 8.2.1), the overall production of a metabolite which either clinically gives an unrequired response or is essential to bacterial or cancerous growth.

Design of enzyme inhibitors as drugs

319

(8.5) Sequential chemotherapy involves the use of two inhibitors simultaneously on a metabolic chain (Equation [8.6]) and is employed with the aim of achieving a greater therapeutic effect than by application of either alone. This situation arises when dosage is limited by host toxicity or resistant bacterial strains have emerged. The best known combination is the antibacterial mixture co-trimoxazole, consisting of trimethoprim (dihydrofolate reductase (DHFR) inhibitor) and the sulphonamide sulphamethoxazole (dihydropteroate synthetase inhibitor) although the usefulness of the latter in the combination has been queried.

(8.6) Inhibition of an enzyme on occasions leads to formation of a dead-end complex between the enzyme, co-enzyme and inhibitor rather than straightforward interaction between the inhibitor and the enzyme. 5-Fluorouracil inhibits thymidylate synthetase to form a dead-end complex with the enzyme and coenzyme, tetrahydrofolate, so preventing bacterial growth (Equation [8.7]).

(8.7)

Introduction to the principles of drug design and action

320

Where product build-up progressively decreases the activity of an enzyme on its substrate, then enhancement of product inhibition (Equation [8.8]) can be achieved by inhibiting an enzyme which disposes of that product. S-Adenosylhomocysteine (SAH), the product of methylating enzymes (e.g. catecholamine methyltransferase, COMT) using S-adenosylmethionine (SAM), and an inhibitor of these enzymes, is removed by the hydrolytic action of its hydrolase (SAH’ase). Inhibitors of SAH’ase should allow a build-up of the product, SAH, leading to a useful clinical effect.

(8.8) Inhibitors have been used (see Equation [8.9]) as co-drugs to protect an administered drug with a required action from the effects of a metabolizing enzyme. Inhibition of the metabolizing target enzyme permits higher plasma levels of the administered drug to persist, so prolonging its biological half-life and either preserving its effect or resulting in less frequent administration. Clavulanic acid, an inhibitor of certain βlactamase enzymes produced by bacteria, when administered in conjunction with a βlactamase-sensitive penicillin preserves the antibacterial action of the penicillin towards the bacteria.

(8.9) 8.1.2 Types of inhibitors

Design of enzyme inhibitors as drugs

321

Enzyme inhibiting processes may be divided into two main classes, reversible and irreversible, depending upon the manner in which the inhibitor (or inhibitor residue) is attached to the enzyme. Reversible inhibition occurs when the inhibitor is bound to the enzyme through a suitable combination of van der Waal’s, electrostatic, hydrogen bonding and hydrophobic forces, the extent of the binding being determined by the equilibrium constant, Ki, for breakdown of the EI or EIS complex for classical inhibitors. Reversible inhibitors may be competitive, non-competitive or uncompetitive depending upon their point of entry into the enzyme-substrate reaction scheme. Competitive inhibitors, as their name suggests, compete with the substrate for the active site of the enzyme and by forming an inactive enzyme-inhibitor complex decrease the interaction between the enzyme and the substrate:

(8.10) The rate (υ) of the enzyme-catalysed reaction in the presence of a competitive inhibitor is given by;

(8.11) where Km is the Michaelis constant which is the molar concentration of substrate at . The extent to which the reaction is slowed in the presence of the which inhibitor is dependent upon the inhibitor concentration [I], and the dissociation constant, Ki, for the enzyme-inhibitor complex. A small value for indicates strong binding of the inhibitor to the enzyme. With this type of inhibitor the inhibition may be overcome, for a fixed inhibitor concentration, by increasing the substrate concentration. This fact can be readily established by examination of Equation [8.11].

Introduction to the principles of drug design and action

322

The value for Ki may be obtained by determining the initial rate of the enzymecatalysed reaction using a fixed enzyme concentration over a suitable range of substrate concentrations in the presence and absence of a fixed concentration of the inhibitor. Rearrangement of Equation [8.11] gives

(8.12) A plot of 1/υ against 1/[S], known as a Lineweaver-Burk plot for the two series of experiments, gives two regression lines which cut at the same point on the 1/υ axis (corresponding to 1/Vmax) but cut the 1/[S] axis at values corresponding to −/Km and −1/Km(1+[I]/Ki) in the absence and presence of the inhibitor, respectively, from which Km and Ki can be calculated. Very often the inhibitory potency within a series of inhibitors may be expressed as an IC50 value. The IC50 value is the concentration of inhibitor required to halve the enzyme activity and this value should be used with care when comparing interlaboratory results, since it is dependent on the concentration of substrate used (Equation [8.13]).

(8.13) Non-competitive inhibitors combine with the enzyme-substrate complex and prevent the breakdown of the complex to products (Equation [8.14]). These

(8.14) inhibitors do not compete with the substrate for the active site and only change the Vmax parameter for the reaction. The kinetics for this type of inhibitor are given by

(8.15)

Design of enzyme inhibitors as drugs

323

The extent of the inhibition, by a fixed concentration of inhibitor, cannot be reversed by increasing the substrate concentration (contrast competitive inhibition) since substrate and inhibitor bind at different sites. A Lineweaver-Burk plot of 1/u against 1/[S] gives a straight line which cuts the 1/[S] axis at −/Km and the 1/υ axis at (1+[I]/Ki)/Vmax. Other classes of reversible inhibitors are the uncompetitive and mixed inhibitors where Km and Vmax are both altered. Nearly all reversible inhibitors which have been designed as potential drugs, as well as drugs in current use, are competitive inhibitors, one notable exception being the cardiac glycosides, which are non-competitive inhibitors of Na+, K+, -ATPase. One reason for this is that competitive inhibitors of the enzyme bear some resemblance to the substrate, since they bind at the same site, and this knowledge has provided a starting point in design, whereas the other types bind elsewhere on the enzyme and need not resemble the substrate, so removing an obvious design aspect. A special type of competitive inhibitor is a transition state analogue. This is a stable compound which resembles in structure the substrate portion of the enzymic transition state for chemical change. An organic reaction between two types of molecules is considered to proceed through a high energy activated complex known as the transition state which is formed by collision of molecules with greater kinetic energy than the majority present in the reaction. The energy required for formation of the transition state is the activation energy for the reaction and is the barrier to the reaction occurring spontaneously. The transition state may break down to give either the components from which it was formed or the products of the reaction. The transition state for the reaction between hydroxyl ion and methyl iodide is shown in Equation [8.16]. The transition state shown depicts both commencement of formation of a C– OH bond and the breaking of the C–I bond. Enzymes catalyse organic reactions by lowering the activation energy for the reaction and one view is that they accomplish this by straining or distorting the bound substrate towards the transition state.

(8.16) Equation [8.17] shows a single substrate-enzymatic reaction and the corresponding non-enzymatic reaction where ES‡ and S‡' represent the transition states for the enzymatic and non-enzymatic reaction, respectively, and KN‡ and KE‡ are equilibrium constants, respectively, for their formation. KS is the association constant for formation of ES from

Table 8.1 Some reversible inhibitors used clinically (after Sandler and Smith, 1989). Drug Enzyme inhibited Clinical use

Introduction to the principles of drug design and action

Allopurinol Acetazolamide, methazolamide, dichlorphenamide, ethoxzolamide Trimethoprim, methotrexate, pyrimethamine Cardiac glycosides 6-Mercaptopurine, azathioprine Captopril, enalapril, cilazapril Sulthiame Sodium valproate Idoxuridine Cytosine arabinoside (Ara-C), 5-fluoro-2',5'anhydro-cytosine arabinoside N-(Phosphonoacetyl)-Laspartate (PALA) Indomethacin, ibuprofen, naproxen

324

Xanthine oxidase Gout Carbonic anhydrase Glaucoma, antiII convulsants Dihydrofolate reductase Na+, K+,-ATP’ase Riboxyl amidotransferase Angiotensinconverting enzyme Carbonic anhydrase Succinic semialdehyde dehydrogenase Thymidine kinase and thymidylate kinase DNA, RNA polymerases

Anti-bacterial, anticancer, antiprotozoal agents Cardiac disorders Anti-cancer therapy Anti-hypertensive agent Anti-convulsant (epilepsy) Epilepsy Anti-viral agent Anti-viral and anticancer agent

Aspartate Anti-cancer agent transcarbamylase Anti-inflammatory Prostaglandin synthetase cyclooxygenase I and II Sterol 14αAntimycotic demethylase of fungi

Miconazole, clotrimazole, Ketoconazole, ticonazole Benzserazide AADC (peripheral) Co-drug with Ldopa in Parkinson’s disease Aminoglutethimide, Aromatase Oestrogen-mediated fadrozole, vorozole, breast cancer letrozole Saquinavir HIV protease HIV infections Zidovudine, ddI, HIV reverse HIV infections zalcitabine, transcriptase

Design of enzyme inhibitors as drugs

TIBO derivatives Acyclovir, vidarabine, ganciclovir Naftifine terbinafine Finasteride Mevinolin, pravastatin, synvinolim Adriamycin, etoposide

325

viral DNA polymerase

Herpes infections

fungal squalene epoxidase 5αreductase HMG-CoA reductase Topoisomerase II

Anti-fungals Benign prostatic hyperplasia Hyperlipidaemia Anti-cancer agents

E and S, and KT is the association constant for the hypothetical reaction involving the binding of S‡' to E. Analysis of the relationships between these equilibrium constants shows the KTKN‡=KSKE‡. Since the equilibrium constant for a reaction is equal to the rate constant multiplied by h/kT, where h is Planck’s constant and k is Boltzmann’s constant, then KT=KS(kE/kN), where kE and kN are the first-order rate constants for breakdown of the ES complex and the non-enzymatic reaction, respectively. Since the ratio kE/KN is usually of the order 1010 or greater, it follows that KT KS. This means that the transition state S‡' is considered to bind to the enzyme at least 1010 times more tightly than the substrate.

(8.17) A transition state analogue is a stable compound that structurally resembles the substrate portion of the unstable transition state of an enzymic reaction. Since the bond-breaking and bond-making mechanism of the enzyme-catalysed and nonenzymatic reaction are similar, then the analogue will resemble S‡' and have an enormous affinity for the enzyme compared to the substrate and consequently will be bound more tightly. It would not be possible to design a stable compound which mimics the transition state closely, since the transition state itself is unstable by possessing partially broken and/or made covalent bonds. Even crude transition state analogues of substrate reactions would be expected to be sufficiently tightly bound to the enzyme to be excellent reversible inhibitors. This expectation has been borne out in practice.

Introduction to the principles of drug design and action

326

Design of a transition state analogue for a specific enzyme requires a knowledge of the mechanism of the enzymatic reaction. Fortunately, the main structural features of the transition states for the majority of enzymatic reactions are either known or can be predicted with some confidence. Another class of competitive inhibitor which binds tightly to the enzyme is the slow, tight-binding inhibitor. These may be bound either noncovalently or covalently and are released very slowly from the enzyme because of the tight interaction. The slow binding is a time-dependent process and is believed to be due either to an enforced conformational change in the enzyme structure or reversible, covalent bond formation. Coformycin, methotrexate and allopurinol belong to this class and are useful drugs. Tight binding, where the dissociation from the complex takes days, is not distinguishable in effect from covalent bonding and this type of inhibitor may be classed as an irreversible inhibitor. Compounds producing irreversible enzyme inhibition fall into two groups; active site-directed (affinity labelling) inhibitors and mechanism-based inactivators (kcat inhibitors, suicide substrates). Active site-directed irreversible inhibitors resemble the substrate sufficiently to form a reversible enzyme-inhibitor complex, analogous to the enzyme-substrate complex, within which reaction occurs between functional groups on the inhibitor and enzyme. A stable covalent bond is formed with irreversible inhibition of the enzyme. Active site-directed irreversible inhibitors are designed to exhibit specificity towards their target enzymes, since they are structurally modelled on the specific substrate of the enzyme concerned. In the previous discussion on reversible inhibitors, the potency of an inhibitor was shown to be reflected in the Ki value, which is characteristic of the inhibitor and independent of inhibitor concentration. However, the actual level of inhibition achieved in an enzyme system involves the use of equations into which inhibitor and substrate concentrations, as well as the Km value for the substrate, need to be incorporated. Similarly, the potency of an irreversible inhibitor is given by binding and rate constants which are both independent of inhibitor concentration. This allows a precise comparison of the relative potency of inhibitors, which is necessary in the design and development of more effective inhibitors of an enzyme. Irreversible inhibition of an enzyme by an active site-directed inhibitor can be represented by

(8.18) provided that complex formation between the inhibitor and the enzyme is ignored here for the present time. The reaction is bimolecular, but, since the inhibitor is usually present in large excess of the enzyme concentration, the kinetics for inactivation of the enzyme follow a pseudo first-order reaction. In the general case of a bimolecular reaction between two compounds A and B, the rate of reaction is given by

Design of enzyme inhibitors as drugs

327

(8.19) where k2 is the second-order rate constant, a and b are the initial concentrations of A and B, respectively, and the concentration of product is x at time t. Integration and rearrangement of Equation [8.19] gives

(8.20) In the situation where a b, this simplifies to

(8.21) Since k2a=k1 where k1 is the pseudo first-order reaction rate constant, then

(8.22) A plot of log (b−x) versus t for the reaction as it proceeds, using a known concentration of the inhibitor, gives a regression line with slope=−k1/2.303, from which k1 and k2 may be obtained. In practice in enzyme inhibition reactions it is sometimes found that k1 is not directly proportional to a so that the value of k2 is not constant with a change in the concentration of the inhibitor a. This is due to initial binding of the inhibitor to the active site of the enzyme before the irreversible inhibition reaction occurs.

(8.23) The rate of the inactivation reaction is given by

Introduction to the principles of drug design and action

328

(8.24) where x represents the concentration of the inhibited enzyme (EI), Ki is the dissociation constant for the enzyme-inhibitor complex and k+2 is the first-order rate constant for the breakdown of the complex into products. Integration of Equation [8.24] gives

(8.25) where k1 is the observed first-order rate constant and

(8.26) When Equation [8.26] is written in the reciprocal form

(8.27) A plot of 1/k1 against 1/[I] gives a regression line from which k+2 and Ki may be evaluated, since the intercepts on the 1/k1 and 1/[I] axes give the values for 1/k+2 and −1/Ki, respectively. Many irreversible inhibitors of certain enzymes have previously been recognized in which the range of electrophilic centres normally associated with active site-directed irreversible inhibitors, e.g. −COCH2Cl, −COCHN2, −OCONHR, −SO2F, are absent so that the means by which they inhibited the enzyme was not understood. The action of these inhibitors has now become understandable since they have been characterized as mechanism-based enzyme inactivators. Mechanism-based enzyme inactivators bind to the enzyme through the Ks parameter and are modified by the enzyme in such a way as to generate a reactive group which irreversibly inhibits the enzyme by forming a covalent bond with a functional group present at the active site. On occasion, catalysis leads not to a reactive species but an enzyme-intermediate complex which is partitioned away from the catalytic pathway to a more stable complex by bond rearrangement (e.g β-lactamase inhibitors). These inhibitors are substrates of the enzyme, as suggested by their alternative name, kcat inhibitors, where kcat is the overall rate constant for the decomposition of the

Design of enzyme inhibitors as drugs

329

enzyme-substrate complex in an enzyme-catalysed reaction. Mechanism-based inactivators do not generate a reactive electrophilic centre until acted upon by the target enzyme. Reaction may then occur with a nucleophile on the enzyme surface, or alternatively the species may be released and either react with external nucleophiles or decompose (Equation [8.28]).

(8.28) The ratio of the rate constants i.e. k+4/k+3 gives the partition ratio (r) for the reaction and where this approaches zero the mechanism-based inactivation will proceed with little turnover of the inhibitor and release of the active species as shown in Equation [8.29] where the non-covalent enzyme-inhibitor complex (EI) is transformed into an activated species (EI*) which then irreversibly inhibits the enzyme.

(8.29) Consequently, the reactive electrophilic species, by not being free to react with other molecules in the biological media, has a high degree of specificity for its target enzyme and exhibits low toxicity. The inactivation rate constant for a mechanism-based enzyme inactivation is termed kinact and is a complex mixture of the rate constants k2, k3 and k4 (Equation [8.28]). However the kinetic form of Equation [8.28] and that for active-site directed inhibition are identical so that Equation [8.27] becomes,

(8.30) which, since becomes,

Introduction to the principles of drug design and action

330

(8.31) A plot of t1/2 versus the reciprical of the inhibitor concentration for the inactivation process using various concentration of the inactivator gives a regression line which cuts the y axis at 0.693/kinact and the x axis at −1/KI. The meaning of KI described here and Ki the dissociation for the enzyme—reversible inhibitor complex may not be the same under certain conditions e.g. when k3 becomes rate deterining. Certain criteria need to be fulfilled before an irreversible inhibitor can be classified as a ‘mechanism’ based enzyme inactivator (see Silverman). 8.2 GENERAL ASPECTS OF INHIBITOR DESIGN 8.2.1 Target enzyme and inhibitor selection Occasionally, drugs in current use for one therapeutic purpose have exhibited sideeffects indicative of potential usefulness for another, subsequent work establishing that the newly-discovered drug effect is due to inhibition of a particular enzyme. Although the drug may possess minimal therapeutic usefulness in its newly found role, it does constitute an important ‘lead’ compound for the development of analogues with improved clinical characteristics. The use of sulphanilamide as an antibacterial was associated with acidosis in the body due to its inhibition of renal carbonic anhydrase. This observation led to the development of the currently little used acetazolamide and subsequently the important chlorthiazide group of diuretics although these have a different mode of action. The anticonvulsant aminoglutethide was withdrawn from the market due to inhibition of steroidogenesis and an insufficiency of 11β-hydroxy steroids. Aminoglutethimide, in conjunction with supplementary hydrocortisone, is now in clinical use for the treatment of oestrogen-dependent breast cancer in postmenopausal women due to its ability to inhibit aromatase, which is responsible for the production of oestrogens from androstenedione. Other more potent aromatase inhibitors have subsequently been developed (see Section 8.6.3). Iproniazid, initially used as a drug in the treatment of tuberculosis, was observed to be a central nervous stimulant due to a mild inhibitory effect on MAO. This observation eventually led to the discovery of more potent inhibitors of MAO, such as phenelzine, tranylcypromine, selegiline ((−)-deprenyl) and chlorgyline. Many drugs introduced into therapy following detection of biological activity by pharmacological or microbiological screening experiments have subsequently been shown to exert their action by inhibition of a specific enzyme in the animal or parasite. This knowledge has helped in the development of clinically more useful drugs by limiting screening tests to involve only the isolated pure or partially purified target enzyme concerned and so introducing a more rapid screening protocol. However translation of in vitro potency to the in vivo situation and finally the clinic is thwart with difficulties as will be seen later (also see Chapter 6). The rational design of an enzyme inhibitor for a particular disease or condition in the absence of a lead compound presents a challenging task to the drug designer, since

Design of enzyme inhibitors as drugs

331

selection of a suitable target enzyme is a necessary first step in the process of drug design. A priori examination of the biochemical or physiological processes responsible for a disease or condition, where these are known or can be guessed at, may point to a suitable target enzyme in its biochemical environment, the inhibition of which would rationally be expected to lead to alleviation or removal of the disease or condition. In a chain of reactions with closely packed enzymes in a steady state (see Equation [8.32]), where the initial substrate A does not undergo a change in concentration as a consequence of changes effected elsewhere in the chain, then any type of reversible inhibitor which inhibits the first step of the chain will effectively block that sequence of reactions.

(8.32) It is a general misconception that the overall rate in a linear chain can be depressed only by inhibiting the rate limiting reaction, i.e. the one with lowest velocity at saturation with its substrate. Since individual enzymes will not be saturated with their substrates, the overall rate is determined largely by the concentration of the initial substrate, so that the first enzyme will often be rate limiting, irrespective of its potential rate due to a low concentration of its substrate. Inhibitors acting at later points in the chain of closely bound enzymes may not block the metabolic pathway. If (Equation [8.32]) is considered, competitive inhibition of E2 the reaction will initially decrease the rate of formation of C but eventually the original velocity (υ2) of the step will be attained as the concentration of B rises due to the difference between its rates of formation and consumption. However, selection of a target enzyme within a metabolic chain which does not inhibit the first step may lead successfully to translation of in vitro results, with the isolated target enzyme, to the in vivo situation due to additional changes. These changes relate to an increase in concentration of B which may have secondary effects on the chain due to product inhibition (B on E1) or product reversal either of these effects can slow υ1, so leading to a slowing of the overall pathway. This view is well illustrated by studies on inhibitors of the noradrenaline biosynthetic pathway. These were intended to decrease production of noradrenaline at the nerve-capillary junction in hypertensive patients, with an associated reduction in blood pressure. The selected target enzyme aromatic amino acid decarboxylase (AADC) catalyses the conversion of dopa to dopamine in the second step of the biosynthesis of noradrenaline from tyrosine. Many reversible inhibitors, although active in vitro against this enzyme, fail to lower noradrenaline production in vivo although they may slow decarboxylation of dopa in peripheral tissues. Irreversible inhibitors of AADC successfully lower noradrenaline levels (see later). However, competitive inhibitors have proved useful clinical agents, as examination of Table 8.1 illustrates, especially where the target enzyme has a degradative role on a substrate and is not part of the metabolic pathway in which the substrate is produced.

Introduction to the principles of drug design and action

332

Examples here are the anticholinesterases (Equation [8.3]) and AADC inhibitors as L—dopa protecting agents in the treatment of Parkinson’s disease. Irreversible inhibition progressively decreases the titre of the target enzyme to a low level and the biochemical environment of the enzyme is unimportant. For example α-monofluoromethyldopa is a mechanism-based inactivator of AADC and produces a metabolite which irreversibly inhibits and decreases the level of the enzyme by >99% (see Section 8.6.4.2). This leads to a near complete depletion of catecholamine levels in brain, heart and kidney despite the occurrence of the enzyme in the second step of the noradrenaline biosynthetic pathway as discussed earlier. The production of inhibited enzyme must be faster than the generation of new enzyme by resynthesis to maintain the target enzyme titre at a low level so that dosing is infrequent. For mechanism based inactivators, not only is the turnover rate of the enzyme important because of enzyme resynthesis, and this rate may be 103−105 slower than for natural substrates, but the partition ratio for the reaction should ideally be close to zero when every turnover should result in inhibition. A list of drugs which act by irreversible inhibition of the enzyme is given in Table 8.2. 8.2.2 Specificity and toxicity Inhibitors used in therapy must show specificity towards the target enzyme. Inhibition of closely related enzymes with different biological roles (e.g. trypsin-like enzymes such as thrombin, plasmin and kallikrein), or reaction with constituents essential for the well-being of the body (e.g. DNA, glutathione, liver P-450 metabolizing enzymes) could lead to serious side-effects.

Table 8.2 Some irreversible inhibitors used clinically (after Sandler and Smith 1989). Drug Enzyme inhibited Clinical use Omeprazole H+, K+-ATPase Anti-ulcer agent Sulphonamides Dihydropteroate Anti-bacterial synthetase Iproniazid, phenelzine, MAO Anti-depressant isocarboxazid, tranylcypromine Neostigmine, eserine, dyflos, Acetylcholinesterase Glaucoma, benzpyrinium, ecothiopate, myasthenia gravis tacrine Alzheimers disease Penicillins, cephalosporins, Transpeptidase Antibiotics cephamycins, carbapenems, monobactams Organic-arsenicals Pyruvate Anti-protozoal dehydrogenase agents O-Carbamyl-D-serine Alanine racemase Antibiotic

Design of enzyme inhibitors as drugs

D-Cycloserine Azaserine γ-Vinyl GABA (Vigabatrin) Clavulanic acid, sulbactam

Alanine racemase Formylglycinamide ribonucleotide aminotransferase GABA transaminase β-lactamase

α-Difluoromethylornithine,

L-Ornithine decarboxylase

Selegiline ((—)-deprenyl)

MAO-B

4-Hydroxyandrostendione

Aromatase

5-Fluorouracil

Thymidilate synthetase

333

Antibiotic Anti-cancer Epilepsy Adjuvant to penicillin antibiotic Trypanosomal and other parasitic diseases Co-drug with Ldopa in Parkinson’s disease Oestrogenmediated breast cancer Anti-cancer

Active site-directed irreversible inhibitors are alkylating or acylating agents and would be expected to react with a range of tissue constituents containing amino or thiol groups besides the target enzyme, with potentially serious side-effects. They are mainly used in in vitro studies for labelling of amino acid residues at the active site. Mechanism-based inactivators do not possess a biologically reactive functional group until after they have been modified by the target enzyme and, consequently, would be expected to demonstrate high specificity of action and low incidence of adverse reactions. It is these features which have encouraged their active application in inhibitor design studies. In the situation where the target enzyme is common to the host’s normal cells as well as to cancerous or parasitic cells, chemotherapy can be successful when host and parasitic cells contain different isoenzymes, e.g. DHFR, with that of the parasite being more susceptible to carefully designed inhibitors. Alternatively, the target enzyme may be absent from the host cell. Sulphonamides are toxic to bacterial cells by inhibiting dihydropteroate synthetase, an enzyme on the biosynthetic pathway to folic acid. The host cell is unaffected, since it utilizes preformed folic acid whilst the susceptible bacterial cannot. Sulphonamides (8.1) are toxic to bacterial cells by inhibiting the utilization of p-aminobenzoate (8.2) by dihydropteroated synthetase, an enzyme in the biosynthetic pathway to dihydrofolic acid. Normal and cancerous cells contain the same form of the target enzyme, DHFR, but the faster rate of growth of the tumour cells makes them more susceptible to the effects of

Introduction to the principles of drug design and action

334

an inhibitor. Although side-effects occur, these are acceptable due to the lifethreatening nature of the disease. 8.3 RATIONAL APPROACH TO THE DESIGN OF ENZYME INHIBITORS Once the target enzyme has been identified then usually a ‘lead’ inhibitor has previously been reported or can be predicted from studies with related enzymes. The design process is then initially concerned with optimising the potency and selectivity of action of the inhibitor to the target enzyme using in vitro biochemical tests. Candidate drugs are then examined by in vivo animal studies for oral absorption, stability to the body’s metabolising enzymes and toxic side effects. Since many candidates may fall at this stage further design is necessary to maintain desirable features and design out undesirable features from the in vivo profile. Since an in vivo profile in animal studies is not directly translatable to the human situation studies with human volunteers are also required before a drug enters clinical trials. Computerised molecular modelling is nowadays an essential part of the design process but its relative importance in this process is determined by the state of knowledge concerning the target enzyme (see Chapter 3). Ideally a high resolution crystal structure of the target enzyme with the active site identified by co-crystallisation with an inhibitor provides a knowledge of binding sites on the inhibitor and enzyme and their relative disposition. Furthermore an additional binding site may be identified so that a modified inhibitor using this additional site may be more potent or selective towards its target. Once the enzyme crystal structure is known the mode of binding of inhibitors fortuitously discovered later can be clarified (hindsight) to explain structural features responsible for their mechanism of action. Usually for a newly discovered target the enzyme crystal structure is not known and the 3D-structure of the protein has to be less satisfactorily predicted from either NMR studies of by homology modelling from a related protein of known 3D-structure. For homology modelling the sequence similarity between the two proteins should be at least 30%. Either of these techniques can lead to the identification of prospective binding sites at the enzyme active site and on a lead inhibitor by “docking” the inhibitor at the active site. Observations can lead to further structural modification of the inhibitor to either improve fit or improve potency by taking advantage of additional binding areas such as hydrogen bonding groups or hydrophobic residues on the enzyme. The relative positions of potential binding areas at the active site can provide a pharmacophoric pattern which can be used for de novo inhibitor design. Also,

Design of enzyme inhibitors as drugs

335

searching of 3D structural data bases can provide novel structures, designed for another purpose, with binding groups held in the correct 3D pattern through an appropriate carbon skeleton. If a model of the enzyme active site does not exist, as is usual for a new target enzyme, then design is based on a knowledge of the substrate, a lead inhibitor (perhaps from a related enzyme) and of the mechanism of the catalytic reaction. Molecular modelling may enter into the design process at a later stage. A few selected examples are now given to illustrate this approach. The anti-hypertensive drug captopril (8.48), an inhibitor or angiotensin Iconverting enzyme (ACE), was designed from a knowledge of the substrate specificity and a known lead inhibitor of its target enzyme, together with a guess that the mechanism of action of ACE was similar to that of the zinc metalloprotease carboxypeptidase A about which much was known (see Section 8.6.1.2). Further structural modification gave the related enalaprilat and, from molecular modelling using inhibitor superimposition, cilazaprilat (8.53). Many mechanism-based inactivators of pyridoxal phosphate-dependent enzymes are known, some of which were designed from a knowledge of the mechanism of action of their respective target enzymes. Inhibitors of AADC, histamine decarboxylase, ornithine decarboxylase and GABA- transaminase designed in this way have proved to be useful drugs (see Section 8.6.4). Aspartate proteases, such as renin and HIV-protease catalyse the hydrolysis of their substrates by aspartate ion-catalysed activation of the weak nucleophile water effectively to the strong nucleophile, hydroxyl ion. The hydroxyl ion attacks the carbonyl of the scissile amide bond in the substrate to give a tetrahedral intermediate which collapses to the products of the reaction (Equation [8.33]). HIV-protease is an aspartate protease which cleaves polyproteins formed in viral reproduction to the correct length for viral maturation. Inhibitors of HIV-protease have been designed based on the amino acid sequence around a scissile bond of the polyprotein substrate and the structure of the tetrahedral intermediate. Using the substrate sequence 165–9 (Leu-Asp-Phe-Pro-ILeu) for a particular polyprotein a stable tripeptide analogue possessing a hydroxyethylamine moiety (-CH(OH)-CH2) to resemble the tetrahedral intermediate (-C(OH)2-) has been developed (see Section 8.6.1.3). This compound, Saquinavir (8.79), has IC50=0.4 nM and is now in clinical trials as an agent to prevent the spread of viral infection. Stable amino- and carboxyl terminal blocking groups are present and the hydrophobicity of the proline in the substrate has been increased in the perhydro isoquinoline residue. Other HIV protease inhibitors have been developed for other scissile bonds in the polyprotein substrate using a variety of functions (see Section 8.6.1.3) which simulate the tetrahedral intermediate formed during catalysis. The crystal structure of the protease is now available leading to further designed inhibitors. Modelling with a series of inhibitors by superimposition (matching) of key functional groups, similar areas of electrostatic potential, and common volumes may identify areas i.e. the pharmacophore, with similar physical and electronic properties in the more active members of a series. Whereas this approach is suitable for rigid structures it is less applicable to flexible molecules since the conformation in solution may be different to that required to efficiently bind to the enzyme active site.

Introduction to the principles of drug design and action

336

Alternatively the common conformational space available to a range of active inhibitors can be used to distinguish this from the space available to less active or inactive analogues which may lead to a defined model for the pharmacophore.

More sophisticated methods have more recently been used to correlate a wide range of physicochemical properties with enzyme inhibitory activity and whereas some of these methods merely rationalise structure-activity relationships others may lead to new inhibitor design (see Chapters 3 and 5)).

Design of enzyme inhibitors as drugs

337

8.4 DEVELOPMENT OF A SUCCESSFUL DRUG FOR THE CLINIC The development of an inhibitor from its inception through to clinical trials is thwart with difficulties. After satisfactory in vitro screening of a potent inhibitor for selectivity towards its target enzyme (i.e. little effect on related enzymes) in vivo studies in animals are undertaken to establish that the candidate drug is well absorbed when administered orally, has a low rate of metabolism (long biological half life, ) and is free from toxic side effects. The in vivo studies present a formidable barrier to the development process and many candidate drugs can fall at this stage as has been described in Chapter 6 for the development of an aromatase inhibitor. 8.4.1 Oral absorption Oral absorption of a drug may be improved by chemical manipulation to a biologically inert but more absorbable form of a drug which after absorption is converted by the body’s enzymes to the active parent drug i.e. prodrug, (see Chapter 7). This approach has proved particularly useful for drugs possessing a carboxylic acid group which being in the ionised form at pH7 may not be well absorbed in the small intestine. Examples are ampicillin where well absorbed esters in the form of pivampicillin, becampicillin, talampicillin release ampicillin in the plasma by initial hydrolysis by esterases to an intermediate which degrades in the aqueous media (see Section 7.4.1). The ACE inhibitor enalaprilat (see 8.6.1.2) is well absorbed as its ethyl ester, enalapril, and the enkephalinase A (MEP) inhibitor SCH 32615 (8.3), a dicarboxylic acid, is well absorbed as the acetonide of the glycerol ester, SCH 34826 (8.4). The potent enkephalinase inhibitor thiorphan (8.5) is not active parenterally but the protected prodrug, acetorphan (8.6) is absorbed through the blood-brain barrier and

Introduction to the principles of drug design and action

338

Design of enzyme inhibitors as drugs

339

subsequently converted by brain enzymes to the active drug. The absorption of the antiviral acyclovir (8.7) has been improved as the valine ester, valaciclovir (8.8) and other analogues penciclovir (7.53) and famcyclovir (7.52) are futher improvements. The oral administration of peptide-like enzyme inhibitors may lead to poor absorption due to the polar nature of the peptide backdown as well as degradation losses by intestinal proteases. Consequently high potency with an IC50 value in the low nanomolar range is required for such drugs. Saquinavir (8.82) an HIV protease inhibitor has a low oral absorption (c. 2%) but this is offset by a low IC50 of 0.4 nM. 8.4.2 Metabolism For a reversible inhibitor to be a useful drug it must exist sufficiently long at the site of its target enzyme to exert its therapeutic effect. Since the level of the inhibitor at the site is a function of its plasma level, liver metabolism of the drug in the plasma to biologically inert product(s) leads to dissociation from the site and reversal of the inhibition. The biological half life (t1/2) of an inhibitor in man is not directly related to that obtained from animal experiments although it is usually longer than that observed in the rat. A half life of c. 8 h is an acceptable figure in man although for cancer chemotherapy a longer half life 12–36 h is required to provide adequate drug cover in the event of patient non compliance with the dose regimen. The metabolic processes by which drugs are modified have been considered in Chapter 1 and most of these processes will lead to a shortening of the t1/2 of the inhibitor. The most important of these involves hydroxylation by liver P450 enzymes. This general phenomenon has been previously described for aromatase inhibitors (Chapter 6) where hydroxylation can occur on a vulnerable imidazole nucleus and benzylic –CH2– group with loss of activity. Replacement of imidazole by triazole may lead to a loss in in vitro potency but this is reversed in the in vivo situation due to the greater metabolic stability of the triazole nucleus. Furthermore substitution of vulnerable and groups with electron withdrawing substituents decreases the development and subsequent hydroxylation (see Chapter 6). This chance of approach is also illustrated in the development of fluconazole (8.12). Ketoconazole (8.9), an antifungal agent, has a short t1/2 and is highly protein bound, due to its lipophilic nature, so that less than 1% of the unbound form exists at the site of action. Modification led to UK-46,245 (8.10) which had twice the potency in a murine candidosis model but further manipulation was required to improve metabolic stability and decrease lipophilicity. This was achieved in UK-47,265 (8.11) which has 100 times the potency of ketoconazole on oral dosing. Unfortunately this compound was hepatotoxic to mice and dogs and teratogenic to rats. Alteration of the aryl substituent to 2,4-difluorophenyl gave fluconazole (8.12) which is >90% orally absorbed and has t1/2=30 h. It is used for the treatment of candida infections and as a broad spectrum antifungal. The stability to metabolism of fluconazole could be attributed to possession of the stable triazole nucleus which is not hydroxylated unlike imidazole as groups to hydroxylation by flanking electron well as protection of the withdrawing groups (hydroxyl, triazole, difluorophenyl).

Introduction to the principles of drug design and action

340

Table 8.3 Toxic or side effects exhibited by some enzyme inhibitors as drugs or drug candidates. ACE inhibitors cough; due to build up of bradykinin (controlling PGE2/PGI2) and substance P (tachykinins) NSAIDS renal syndromes; gastrointestinal effects HMG CoA reductase myopathy inhibitors MAO inhibitors hypertensive reaction with tyramine-containing (unselective) foods Cholinesterase abdominal cramps, salivation, diarrhoea inhibitor Steroidogenesis adrenal hormone suppression cytochrome P-450 enzyme inhibitors

Design of enzyme inhibitors as drugs

341

8.4.3 Toxicity Toxic effects may become apparent on chronic dosing during animal pharmacology studies, clinical trials or even after marketing. A well-known example is aminoglutethimide introduced as an anti-epileptic and subsequently withdrawn due to effects on steroidogenesis enzymes leading to a ‘medical adrenalectomy’. It was later reintroduced as an anti-cancer agent for the treatment of breast cancer by oestrogen deprivation to capitalise on this toxic effect. The toxic side effects may merely be a matter of inconvenience or may be more severe (see Table 8.3). Many drugs e.g. cimetidine, erythromycin, ketoconazole, choramphenicol, isoniazid, verapamil, including enzyme inhibitors are non-specific inhibitors of liver cytochrome P-450 enzymes i.e. inhibit many iso-enzyme forms. They consequently affect the metabolism of other drugs given concurrently leading to enhanced levels of these drugs and appearance of toxic effects. Specific inhibitors of P450 isoezymes have a similar effect but this effect is restricted to specific substrates of the particular isoenzyme concerned. Examples include quinolone antibiotics (isoenzyme CYP1A2) and sulphaphenazole (CYP2C8/9). 8.5 STEREOSELECTIVITY The stereochemistry of enzyme inhibitors possessing a chiral centre(s) is usually important in determining their potency towards a specific enzyme and this is a problem to be addressed in the early stages of drug design since it can sometimes be avoided by limiting the studies to achiral compounds. Drug Registration Authorities world-wide are moving towards a requirement that for all drugs the enantiomeric active form must be marketed unless for the racemate the activity of the separate enantiomers is available and enantioselective methods of chemical and biological analysis have been used in both animal and human studies. These requirements take into account the pharmacological consequences of the use of racemic drugs which has been previously described in Chapter 4. Whereas the literature abounds with examples of activity residing mainly in one enantiomer following in vitro studies, very few of these compounds have, as yet, reached the clinic or been subjected to registration requirements and in vivo information is not available from animal studies. Aminoglutethimide (AG) (8.13), a long-established aromatase inhibitor, is used clinically as the racemate in the treatment of breast cancer in post-menopausal women (after surgery) to decrease their tumour oestrogen levels. The (+)(R)- form is about 38 times more potent as an inhibitor than the (−)(S)- form. AG is also an inhibitor of the side-chain cleavage enzyme (CSCC) which converts cholesterol to pregnenolone in the adrenal steroidogenic pathway. Depletion of corticosteroids in this manner requires adjuvant hydrocortisone administration with the drug. Here the (+)(R)- form is about 2.5 times more potent than the (−)(S)- form. For pyridoglutethimide (rogletimide) (8.14), an analogue of AG without the undesirable depressant effect, the inhibitory potency resides mainly in the (+)(R)- form (20 times that of the (−)(S)- form). 1-Alkylation improves potency in vitro but the

Introduction to the principles of drug design and action

342

activity for the most potent inhibitor in the series, the 1-octyl, resides in the (−)(S)form owing to a change in the mode of binding of inhibitor to enzyme. A more selective inhibitor of aromatase than AG is the triazole vorozole (8.15) which is about 1000-fold more potent as an inhibitor. The (+)(S)- form is 32 times more active than the (−)(R)- form, but the very small inhibitory activity of the racemate towards other steroidogenic pathway enzymes, 11β-hydroxylase and 17,20lyase, originates in the (−)-and (+)- forms respectively.

It is of interest that in the benzofuranyl methyl imidazoles (8.16), some of which are 1000 times more potent as aromatase inhibitors in the racemic form than AG, comparable activity lies in both enantiomers. MAO occurs in two forms, MAO-A and MAO-B. The use of MAO inhibitors as antidepressents is complicated by a dangerous hypertensive reaction with tyramine-

Design of enzyme inhibitors as drugs

343

containing foods (the ‘cheese-effect’) which is due to inhibition of MAO-A located in the gastro-intestinal tract which would otherwise remove the tyramine. L-(−) deprenyl (selegiline) (8.17), a selective inhibitor of MAO-B, is widely employed to limit dopamine breakdown in Parkinson’s disease in selective inhibitory dosage. The (−)isomer is much more potent than the (+)-isomer and, since the products of metabolism are (−)-metamphetamine and (+)-metamphetamine respectively, the more potent (+)metamphetamine side-effects are removed from the racemate by use of L-deprenyl. γ-Aminobutyric acid (GABA) transaminase inhibitors allow a build-up of the inhibitory neurotransmitter GABA and are potential drugs in the treatment of epilepsy. The inhibitory action of γ-vinyl GABA (vigabatrin), a drug used clinically in the treatment of this disease, resides mainly in the (S)-enantiomer (see Section 8.6.4.1). 8.6 EXAMPLES OF ENZYME INHIBITORS AS DRUGS 8.6.1 Protease inhibitors Proteases have been classified according to substrate specificity (enkephalinase, collagenase, elastase), substrate size (peptidases, proteinases) or cleavage site on the substrate (aminopeptidases, carboxypeptidases) and localisation (human neutrophil elastase, pancreatic elastase, HIV-protease). However for the purpose of developing protease inhibitors based on the mechanism of the respective enzyme, a classification based on a knowledge of the catalytic function at the active site has proved to be more useful and four subclasses (serine, cysteine, aspartic and metalloproteases) have been identified which catalyze the cleavage of the amide bond linking two amino acids by nucleophilic attack on the scissile carbonyl carbon atom. Little was known of the actual structure of these various enzymes during the early development of enzyme inhibitors. The approach of first identifying a ‘lead’ compound, using models of active sites based on knowledge of substrate specificity, and then optimising its structure has been highly successful in the design and development of potent and selective enzyme inhibitors. Crystal structures are now available for many enzymes. Information from X-ray crystallography studies of enzyme-inhibitor complexes and computer assisted molecular modelling is becoming an important part of the design and development process. 8.6.1.1 Serine proteases Serine proteases form the largest group and occur in the plasma (as coagulation factors and complement components), and the intestine (as cellular proteases). Examples include chymotrypsin, trypsin, elastase, cathepsin G, thrombin and plasminogen activator. The key catalytic element is a serine hydroxyl group and the nucleophile is an intergral part of the enzyme structure and therefore substrates for these enzymes undergo covalent catalysis. The primary sequences of individual serine proteases vary but the active site consists of:

Introduction to the principles of drug design and action

344

(i) a catalytic site where the covalent bond making-bond breaking reactions take place involving three amino acid residues (His-57, Asp-102, Ser-195 for chymotrypsin) known as the ‘catalytic triad’ and the oxyanion hole’ comprised of NH groups of serine and glycine (Ser-195, Gly-193 for chymotrypsin) which stablise the oxyanion of the tetrahedral adduct and, (ii) an extended binding site where noncovalent binding, through hydrogen bonding and hydrophobic forces, occurs between the enzyme and the substrate through the amino acid residues extending on either side of the scissile bond. The most important of these is the interaction between the S1 subsite and the P1 residue (see Figure 8.1 for definitions) as it determines substrate specificity for serine proteases. Modification of the P1 residue may alter the enzyme selectivity of the substrate or inhibitor. The stages of peptide bond hydrolysis illustrated using α-chymotrypsin include (see Figure 8.2): (i) complex formation between the substrate and the extended binding site of the enzyme (ii) formation of a tetrahedral intermediate (a high energy transition state-like intermediate between the substrate and acyl-enzyme) formed by nucleophilic addition of the serine hydroxyl group (Ser-195) to the carbonyl carbon atom of the scissile peptide bond. Hydrogen bond formation of the serine hydroxyl group with the imidazole of His-57 (which is also interacting with Asp-102), increases the nucleophilicity of Ser-195 hydroxyl group. (iii) the proton on the serine hydroxyl group which was transferred to His-57 is shuttled to the nitrogen atom of the C-terminal amine product so aiding the collapse of the tetrahedral intermediate and formation of the acyl enzyme.

Figure 8.1 Definition of binding sites for substrates and inhibitors.

Design of enzyme inhibitors as drugs

345

Figure 8.2 Mechanism of action of serine proteases. (iv) Hydrolysis of the acyl-enzyme by catalytic addition of water produces the N-terminal carboxylic acid fragment of the peptide and the free enzyme which is then ready to repeat the cycle. Elastase inhibitors Human neutrophil and leucocyte elastases are members of a subfamily of proteases released in response to various inflammatory stimuli and capable of degrading a variety of structural proteins including collagen, elastin and fibronectin. Under normal circumstances, elastolytic activity is controlled by natural proteinaceous inhibitors such as α1-proteinase inhibitor (α1-P1) which is present in plasma (guards the lower airways), secretory leukocyte protease inhibitor (SLP1) secreted by mucosal cells (protects the larger airways) and elafin found mainly in the skin and has also been detected in bronchial secretions. The regulation of elastase activity by these natural proteinaceous inhibitors breaks down in a number of pathophysiological states resulting in unrestrained elastolytic activity associated with diseases such as emphysema, rheumatoid arthritus, chronic bronchitus, cystic fibrosis, adult respiratory distress syndrome and glomerulonephritis. The aim of developing human neutrophil elastase inhibitors has been to identify agents to treat diseases associated with one of the most destructive enzymes in the body. The primary structure, x-ray crystal structure and gene sequence for human neutrophil elastase (HNE) have been determined. Natural inhibitors have been produced by recombinant technology and formulations have been developed for aerosol and intravenous administration. Attempts have also been made to develop low molecular weight synthetic inhibitors based on the enzyme’s mechanism. Design of human leucocyte elastase (HLE) inhibitors has been based on computer modelling

Introduction to the principles of drug design and action

346

studies and proposed enzyme-inhibitor complexes using X-ray crystal structure information from HLE or porcine pancreatic elastase (PPE). Structural information for PPE and PPE-inhibitor complexes has been available for some time and initial efforts were directed towards mapping the active-site of the enzyme with synthetic substrates and irreversible enzyme inhibitors such as peptidyl chloromethyl ketones. Elastases have a relatively small S1 subsite and are more dependent than most other serine proteases on interactions of their extended binding pocket as a means of achieving binding of substrates. Most mapping studies have focused on the N-terminus side of the scissile bond. The optimal tetrapeptide recognition sequence for both synthetic peptide substrates and inhibitors has been identified (MeO-Suc-Ala-Ala-Pro-Val-X). Proline, the optimal residue at P2 for both the substrates and inhibitors, is thought to ‘pre-organise’ the ‘backbone’ into a conformation which is complementary to the enzyme. The methoxysuccinyl group increases potency by binding to the S5 subsite and improves the aqueous solubility of the peptide. Various elements of the catalytic machinery, involved in the formation and breakdown of the transition state and reaction intermediates, have been considered for inhibitor design. Affinity label inhibitors combine the concept of the substrate fragment, an affinity fragment which guides the molecule into the active site, and a chemically reactive group which reacts with the essential catalytic groups, usually by alkylation, to block enzyme activity. This type of inhibitor e.g. peptidyl chloromethyl ketone has been the most studied and used to investigate secondary enzyme-inhibitor interactions. Transition-state analogues lack the scissile amide linkage. Peptidyl-aldehydes, boronic acids, -phosphonic acids, and -methyl ketones derivatives all form a complex with the enzyme which resembles the tetrahedral intermediate. Peptidyl aldehydes such as (8.18), which are selective for human leucocyte elastase, borrow the structural features of the substrate elastin and the natural inhibitor α1-P1 to build the appropriate recognition features into the backbone of the inhibitors. Inhibitors of this type lack metabolic stability. Simple aliphatic ketones are poor inhibitors of serine proteases but trifluoromethyl ketones (8.19) are competitive, slow binding, inhibitors of elastase. Introduction of alpha fluorine atoms increases the electrophilicity of the carbonyl group and also stabilises the resulting oxyanion formed with the serine hydroxyl of the tetrahedral complex (now confirmed by X-ray crystallography) which resembles the oxyanion of the enzyme-

Design of enzyme inhibitors as drugs

347

substrate reaction intermediate. The slow binding nature of trifluoroketones is due to a rate-limiting conformational change of a highly stabilised enzyme-inhibitor complex

Introduction to the principles of drug design and action

348

which allows for optimum interaction between the enzyme and inhibitor. Optimal inhibitory activity is exhibited by tri- and tetra-peptide analogues (P1-P4) with lipophilic side-chains. However, in vivo activity in the hamster emphysema model does not correlate with the in vitro potency of these compounds when administered directly into the lung. Acylsulphonamide derivatives exhibit a long duration of action and in vivo activity. ICI 200 880 (8.20) (administered intravenously or by aerosol) is undergoing clinical evaluation in adult respiratory distress syndrome. Information from X-ray crystallography studies of some peptidyl inhibitors bound to HLE and PPE has been useful in designing pyridone-based inhibitors (8.21). Crystal structures indicate a pair of hydrogen bonds between the P3 residue of the inhibitors and the Val216 residue of the enzyme. The carbonyl and amido groups of Val-216 are approximately planar and there is nearly a coplanar arrangement of the reciprocal pair of hydrogen bonding partners (NH, C=O) of the P3 residue of the inhibitor. Maintenance of this coplaner arrangement was considered to be important in the design of non-peptide inhibitors. The S3 subsite of HLE, a relatively shallow area exposed to the solvent, indicated that occupation of this site did not make a significant contribution to the binding affinity of inhibitors. A planar molecular fragment, such as a pyridone ring, is acceptable as the P3 residue of inhibitors. Molecular modelling studies have demonstrated that the pyridone carbonyl and 3-position NH groups could be positioned so that hydrogen bonding interactions could be formed with Val-216. However the pyridone nucleus could not access the enzyme S2 subsite

Design of enzyme inhibitors as drugs

349

which normally requires a hydrophobia group to improve inhibitor binding affinity. Pyrimidinone-containing trifluoromethyl ketones (8.22, Ki=0.1 µM) show a good combination of enzyme selectivity, oral bioavailability and reasonable duration of action. Boronic acid inhibitors (8.23, Ki=6.2 nM) show good inhibitory potency in vitro but not in vivo. Some pentafluroethyl derivatives such as MDL 101 146 (8.24, Ki=25 nM) are orally active, whereas the trifluoromethyl derivatives (8.25, Ki=12 nM) show no oral activity possibly due to the difference in the degree of hydration of the respective electrophilic ketones. The E-enol acetate derivative (8.26) of MDL 101 146 acts as an orally active prodrug. Unlike affinity labels and transition-state analogues, mechanism-based inhibitors of elastase are activated by the catalytic machinery of the target enzyme by two possible mechanisms. With “acyl enzyme” inhibitors the catalytic attack causes the formation of an acyl enzyme which deacylates slowly without irreversibly inactivating the target

Introduction to the principles of drug design and action

350

Design of enzyme inhibitors as drugs

351

enzyme. Studies on the electronic effects of ring substituents on the acylation and deacylation where the transferred group is a benzoyl derivative have shown that, in general, electron donating substituents stabilize the acyl enzyme. Steric factors also contribute to the stabilization process. Inhibitors which act mainly via this mechanism with an appropriate pharmacokinetic profile and selectivity for human elastases have yet to be developed. An alternative approach has been the development of compounds where the catalytic attack causes the formation of chemically reactive intermediates which irreversibly modify the enzyme by forming stable covalent bonds with a different functional group at or near the active site. This type of inhibitor has the advantage of being relatively chemically inert until activation by the target enzyme. Bacterial penicillin binding proteins, beta-lactamases and serine proteases all hydrolyse an amide bond via formation of a tetrahedral intermediate. The benzyl ester of the beta-lactamase inhibitor clavulanic acid is a weak inhibitor of HLE. Extensive screening of chemical derivatives have identified cephalosporin sulphone (8.27), as a potent beta-lactam elastase inhibitor. Bacterial penicillin binding proteins require a beta configuration at C-7, whereas mammalian elastases prefer the alpha configuration at C-7. Small alpha-orientated substituents such as chloro or methoxy are preferred at C-7 with the sulphone derivatives showing the highest inhibitory potency for elastase. Masking of the free carboxyl group at C-4 of the cephalosporins with ester, amide or ketone substituents also increases inhibitory activity for elastase. Modelling studies show the C-4 substituents to be positioned around the S1'-S2' sites of elastase. The shape and lipophilicity of C-4 substituents also contribute to elastase inhibitory potency by altering the reactivity and structural reorganisation ensuing from betalactam cleavage which is essential to the enzyme inactivation mechanism. C-4 tertButyl ketones of 7 alpha-chlorocepham series are equal in potency with the ester derivatives and the tert-butyl ketones of 7 alpha-methoxycepham (8.28, Ki= 21 nM, t1/2=75 h) show high inhibitory potency combined with hydrolytic stability which is greater than the ester, thioester or amide (8.29, Ki=75 nM, t1/2=25 h) analogues. Introduction of an acyloxy substituent at the C-2 position further increases potency. Xray crystallographic data from cephalosporin A bound to the porcine enzyme shows inhibition of elastase to be initially reversible, followed by a time-dependent irreversible inhibition resulting from alkylation of His-57 by the dihydrothiazine ring of the inhibitor. Crystallographic data, biochemical studies and structural characterisation of enzyme-inhibitor complexes and biproducts indicates that the mechanism operating through the enzyme-inhibitor complex involves beta-lactam ring opening by the catalytic Ser-195 residue, expulsion of a leaving group at the 3'position of the cephem moiety and

Introduction to the principles of drug design and action

352

binding to the His-57 residue of the enzyme catalytic triad (see Figure 8.3). Cephem 4ketones, with no adequate leaving group attached to the dihydrothiazide ring, behave as poor substrates rather than inhibitors and are slowly completely hydrolysed by HNE. A novel mechanism of enzyme inhibition has been suggested for benzisothiazolone inhibitors. The inhibitors inactivate the enzyme by a suicide mechanism but the enzyme is then able to regain its full activity. The lead compound for benziothiazolone inhibitors, KAN 400 473, (8.30, Ki=15 nM), could not be detected in human blood after an incubation time of less than 1 minute. An isopropyl substituent at the 4position

Design of enzyme inhibitors as drugs

353

Figure 8.3 β-Lactam inhibition of elastase (after Navia et al. (1987) Nature (Lond.), 327, 79).

significantly improves inhibitory potency and metabolic stability in human blood (Ki= 0.3 nM, t1/2=45 min). Introduction of a methoxy group in the 6-position further enhanced blood stability (t1/2=260 minutes) probably due to the increased reactivity of the benzisothiazolone carbonyl by the electron-donating 6-methoxy group. The 2,6dichlorobenzoate leaving group was optimum for potency and the compound retained stability in human blood. These inhibitors, however, have poor in vivo activity due to low hydrophilicity. Compounds such as WIN 64733 (8.31, Ki=0.014 nM) and WIN 63759 (8.32, Ki=0.013 nM), with aqueous solubilizing substituents show good pharmacokinetic properties and specificity for HNE. Replacement of

Introduction to the principles of drug design and action

354

mercaptotetrazole (see 8.30) with a diethylphosphate as leaving group significantly increases in vivo activity (8.33, Ki=0.035 nM). Thrombin inhibitors Thrombin plays a central role within the coagulation cascade initiating not only fibrin clotting but exerting several cellular effects, too. The serine proteinase thrombin is a member of the trypsin family which attack peptide bonds following Arg or Lys residues. Therefore, inhibitors occupying the active site must possess or imitate the basic aminoor guanidinoalky side chain of Lys and Arg. As will be described later, in extensive biochemical and pharmacological studies thrombin inhibitors were shown effective as anticoagulants and antithrombotics. The main criteria for a low-molecular weight

thrombin inhibitor to be an ideal anticoagulant are high selectivity and systemic bioavailability after oral application. Thrombin is not present in an active form in blood but is formed from prothrombin after activation of the coagulation cascade, whereas its substrates (fibrinogen, thrombin-activatable clotting factors) are permanently present. Consequently, inhibitors to be of therapeutic value must be present in the plasma at adequate concentrations to immediately neutralize the thrombin generated upon massive activation after vascular injury. It has been calculated that a pulse of thrombin is formed reaching a peak of about 200 nmol/l. However, immediately after initiation of the coagulation cascade the concentration of thrombin will be lowered by endogenous

Design of enzyme inhibitors as drugs

355

inhibitors. To be effective in anticoagulation the plasma concentration of a potent inhibitor should be at least 100 nmol/l. X-ray crystal structures of complexes between thrombin and several inhibitors and substrate analogues have been solved providing the basis for rational drug design (Figures 8.4 and 8.5). Besides the primary specificity binding site to which the basic P1 amino acid of substrates is bound, there are two further important binding sites: the hydrophobic aryl-binding site and the anion-binding exosite, also called fibrinogen recognition site. The aryl-binding site located close to the active site is occupied by Phe at P9 of the fibrinopeptide A sequence, it is important in the binding of inhibitors of small size. The anion-binding exosite was discovered first from the crystal structure of the complex between thrombin and the naturally occurring thrombin inhibitor hirudin isolated from the medicinal leech Hirudo medicinalis. Four Arg and five Lys residues but also hydrophobic amino acids contribute to this positively charged region, involved in both the recognition of the substrates fibrinogen and thrombin receptor but also in the binding of thrombomodulin and some natural inhibitors. Three main types of inhibitors have been developed which potently inhibit thrombin. These include peptide inhibitors based on natural substrates, arginine analogues, and benzamidine-derived compounds. Peptide chloromethyl ketones, aldehydes, esters or amides which possess the thrombin-sensitive Gly-Val-Arg sequence of the natural substrate fibrinogen and those resembling the Pro-Arg cleavage sites of factor XIII and prothrombin, are less effective inhibitors. However, extending of the Pro-Arg sequence with a D-Phe at P3 position gives effective inhibitors such as the chloromethylketone H-D-Phe-Pro-Arg-CH2Cl (PPACK, (8.34)), the aldehyde H-D-Phe-Pro-Arg-H (8.35) and the boronic acid derivative Ac-D-PhePro-Arg-B(OH)2 (DuP 714, (8.36)). The D-Phe at P3 resembling Phe at P9 of the fibrinopeptide A sequence occupies the aryl binding site. The chloromethylketone (PPACK, 8.34) is the most powerful and most selective irreversible inhibitor of thrombin known, with a second order rate constant three to five

Figure 8.4 View of the active site cleft of thrombin, displayed with its Connolly

Introduction to the principles of drug design and action

356

dot surface; molecular surfaces are colored dark blue, red, and light blue if donated by basic, acidic, or other residues, respectively. The bound PPACK molecule (8.34) is shown in red, with its Arg side chain disappearing into the primary specificity binding site. The prominent cleft, running from left to right, is where the substrate polypeptide chain would bind. The insertion loop which partially occludes the active site gives the thrombin molecule its selectivity. The aryl-binding site is located to the left close to the insertion loop; the anion binding exosite is to the right. (The Figure was courteously provided by Drs. W.Bode and M.Stubbs).

Figure 8.5 Active-site region of bovine NAPAPthrombin (light green) superimposed with the experimentally determined inhibitors NAPAP (blue, (8.45)), argatroban (yellow, (8.41)), and PPACK (red, (8.34)). In contrast to the extended peptide-like PPACK molecule, the nonpeptidic inhibitors bind in compact, U-shaped

Design of enzyme inhibitors as drugs

357

conformation. (The Figure was courteously provided by Drs. W.Bode and M.Stubbs). orders of magnitude higher than that for the inhibition of other trypsin-like proteases, such as factor Xa, plasmin, urokinase, plasma and glandular kallikrein. Binding of (8.34) in the thrombin-inhibitor complex is shown in Figure 8.4. After i.v. application, the D-Phe-Pro-Arg-derived inhibitors (8.34, 8.35, 8.36) exhibited anticoagulant effects in various animal experiments, however, oral bioavailability is low (<10 %). Low selectivity with respect to the fibrinolytic system would limit the therapeutic use of the aldehyde

(8.35) and the boronic acid derivative (8.36). Methylation of the terminal nitrogen of the parent compound gives Me-D-Phe-Pro-Arg-H (efegatran; (8.37)) with improved stability and selectivity. Exchange of boroArg at P1 by methoxypropylboroGly enhanced both oral bioavailability and selectivity of inhibition. The derivative Z-DPhe-Pro-boro-Mpg-OPin (8.38), lacking the basic guanidino function of Arg which was thought essential for binding to the primary specificity site, is a potent thrombin inhibitor with nanomolar Ki. Transformations of the Arg carboxyl in D-Phe-Pro-Arg peptides, such as introduction of nitrile, fluoroalkylketones, phosphonic acid or a-keto carbonyl residues do not improve selectivity and oral bioavailability. Remarkable improvement of the pharmacokinetic properties—both absorption and half life in circulation—was obtained with inogatran (8.39), containing agmatine at P1 and an N-terminal acetic acid residue. Another type of inhibitor has been developed from synthetic Nα-arylsulfonylated arginine ester-type substrates. Potent reversible inhibitors of thrombin are DAPA (8.40) and argatroban (OM 805, MQPA; (8.41)) with Ki values in the 20–40 nM range.

Introduction to the principles of drug design and action

358

Argatroban is the only direct thrombin inhibitor used in clinical practice so far. The drug has been successfully used in man as an antithrombotic agent instead of heparin.

Design of enzyme inhibitors as drugs

359

Argatroban is not sufficiently absorbed after oral administration, however, replacement of the guanidino group of Arg by amino-substituted heterocycles enhanced cell permeation. Benzamidine derivatives are effective competitive reversible inhibitors of thrombin. 4-Amidinophenylpyruvic acid (APPA; (8.42)) is an outstanding inhibitor with respect to oral bioavailability (up to 80%). Despite its low selectivity and affinity (Ki for thrombin 6.5 µM) anticoagulant and antithrombotic effects could be demonstrated in vivo. Selective inhibitors of thrombin were found among derivatives of amino acids containing a benzamidine moiety at the side chain. The Nα-tosylated piperidides of 3amidinophenylalanine (3-TAPAP; (8.43)) and of 4-amidinophenylalanine (4-TAPAP; (8.44)) had potencies and binding properties similar to those of the arginine derivative

Introduction to the principles of drug design and action

360

argatroban (8.41) (Figure 8.5). Nα-Naphthylsulfonyl-glycyl-4-amidinophenylalanine piperidide (NAPAP; (8.45)), with a glycine spacer at the α-amino substituent, binds tightly to thrombin. It was the first synthetic thrombin inhibitor with nanomolar Ki. Both anticoagulant and antithrombotic effects correspond to its pronounced antithrombin

Design of enzyme inhibitors as drugs

361

activity, however, it is not enterally absorbed and rapidly eliminated from the circulation. Variations of the basic structures were performed to improve the pharmacokinetic properties, however, most of them led to a drastic loss in inhibitory activity, without clear improvement in oral bioavailability. From the X-ray crystal structure of the NAPAP-thrombin complex (Figure 8.5) it was obvious that the NAPAP molecule is bound nearly perfectly to thrombin, so that there is only limited space for additional substituents. Despite this, the NAPAPthrombin crystal structure was used for modelling of several new derivatives. Changing Gly at P2 to Asp (CRC 220) enhanced the antithrombin potency but not bioavailability. A remarkable compound was developed by changing the benzamidine moiety to 3-amidinopiperidine. Ro 46–6240 (8.46) inhibits thrombin with a Ki of 0.3 nM and has outstanding selectivity, never found with benzamidines. However, (8.46) has no improved pharmacokinetic properties. From the X-ray structure of its complex with thrombin, 3-TAPAP (8.43) was chosen as a promising model for the synthesis of new inhibitors allowing different substitutions without drastic influence on antithrombin activity. Introduction of a Cterminal piperazide opened the way to introducing quite different substituents on the second nitrogen.

Introduction to the principles of drug design and action

362

Several derivatives of 4-guanidinophenylalanine were also synthesized. The Nαdansylated piperidide (S-2581; (8.47)) has moderate antithrombin activity, however, introduction of 4-guanidinophenylalanine into peptides does not result in potent inhibitors. Two classes of synthetic peptides were designed after solving the X-ray crystal structure of the thrombin-hirudin complex. Hirudin exerts a concerted binding of its N-terminal and C-terminal domains to the active site and the anion-binding exosite of thrombin. Several analogues of the C-terminal anionic hirudin “tail” were synthesized. The so-called hirugens inhibit thrombin-induced fibrinogen clotting in micromolar concentrations but not the cleavage of synthetic peptide substrates. Also singlestranded DNA molecules (aptamers), selected from a pool of oligodeoxyribonucleotides, interact with the anion-binding exosite of thrombin. Very potent bifunctional thrombin inhibitors with nanomolar Ki values based on the Cterminal sequence of hirudin and the D-Phe-Pro-Arg sequence were designed using

Design of enzyme inhibitors as drugs

363

Gly residues as spacer. Also introduction of NAPAP as N-terminal part was proposed. The so-called hirulog is very effective antithrombotically in animal models. The drug was well tolerated in man and entered Phase III trials. Hirugens and hirulogs composed of more than 10 amino acids do not show any enteral absorption as is also the case with most of the synthetic thrombin inhibitors designed so far. The knowledge on thrombin inhibitors accumulated so far shows that the desirable goal in the development of new thrombin inhibitors for therapeutic use should not only be the design of more active compounds but the improvement of the pharmacokinetic properties of known types of inhibitor, especially with regard to oral bioavailability. 8.6.1.2 Metallo-proteases Metallo-proteases are a group of enzymes which possess a catalytically essential zinc atom at the active site. Much of the information regarding the catalytic mechanism of metalloproteases has been based on active site models of carboxypeptidase A and thermolysin. Here, four ligands—two histidine residues, a glutamic acid residue and a water molecule—are arranged in a tetrahedral fashion around the zinc atom. A catalytically functional glutamyl residue (Glu-270 for carboxypeptidase A and Glu143 for thermolysin) is also present at the active site. In general all metalloproteases use the same basic mechanism for peptide bond cleavage. The water bound as the fourth zinc ligand is hydrogen bonded to the carboxylate of the glutamyl residue during the resting state of the enzyme. The water is displaced from the zinc by the carbonyl group of the amide bond to be cleaved in the peptide substrate, through a five coordinate transition state, during formation of the enzyme-substrate complex. The carbonyl group of the scissile bond of the substrate is then attacked by water to give a tetrahedral intermediate. The zinc atom polarizes the carbonyl making it more susceptible to nucleophilic attack and stabilizes the tetrahedral intermediate. Decomposition of the tetrahedral intermediate results in bond cleavage. The structures of the zinc-inhibitor complexes, determined using thermolysin-inhibitor complexes, have shown thiol (2-benzyl-3-mercaptopropanoyl-L-alanylglycinamide), carboxylate (L-benzylsuccinic acid) and phosphinyl (phosphoramidon) inhibitors to be tetrahedrally coordinated to zinc with displacement of a water molecule. The structure of the thermolysin-hydroxamic acid inhibitor complex differs in that the hydroxamate moiety (CONHOH) forms a bidentate complex with the zinc through the carbonyl oxygen and the hydroxyl group, so that a penta-coordinate complex is formed between three ligands from the enzyme and two from the hydroxamate group. The active sites of metallo-proteases also contain a varying number of residues responsible for substrate recognition which differ significantly between individual metallo-proteases. Examples of metallo-proteases include leucocyte collagenase, membrane metalloendopeptidase (neutral endopeptidase) and angiotensin-converting enzyme (ACE). Approaches used in the design and development of inhibitors of zinc metalloproteases include leads from biproduct inhibitors (incorporating features of both cleavage products from the enzyme catalysed reaction), introduction of a zinc-binding ligand (e.g. thiol, hydroxamate) and modification of substrates. Angiotensin converting enzyme inhibitors

Introduction to the principles of drug design and action

364

Angiotensin converting enzyme (ACE) cleaves a dipeptide from angiotensin I (a decapeptide) at the carboxyl terminal to generate angiotensin II (an octapeptide) which has hypertensive activity and stimulates the release of aldosterone (Equation [8.34]). ACE also catalyzes the hydrolysis of the vasodilator bradykinin (a nonapeptide). ACE inhibitors are therefore important antihypertensive agents.

(8.34) During the initial stages of development of ACE inhibitors, little was known concerning its structure and characteristics other than that it was a zinc-dependent enzyme. In the early work a model of the active site of ACE was tentatively constructed, based on information from the well known structure of carboxypeptidase A, where a zinc moiety at the active site forms a complex with the scissile amide bond of the peptide substrate and a positively charged residue binds to the negatively charged C-terminal carboxyl group of the substrate. Unlike Carboxypeptidase A which cleaves a single amino acid from the C-terminus, ACE cleaves off a dipeptide. ACE does not show specificity for C-terminal hydrophobic amino acids, indicating that its active site does not have a hydrophobic pocket corresponding to that possessed by carboxypeptidase A. Two sites, corresponding to side chain substituents of the two terminal amino acid residues of the substrate (AA1 and AA2, see 8.48) were included in the model and have since been validated by structure-activity studies. D-Benzylsuccinic acid, a biproduct inhibitor of carboxypeptidase A, was introduced into the structure of a dipeptide and became the starting point for the design of ACE inhibitors. Proline, the C-terminus amino acid in naturally occurring peptide inhibitors such as the nonapeptide SQ 20881 (Glu-Trp-Pro-Arg-Pro-Gln-IlePro-Pro), became the natural choice as the C-terminal amino acid in synthetic inhibitors. Structural requirements for binding to the active site were determined by using simple peptide inhibitors. Zinc-binding ligand. Succiny-L-proline, the first bi-product inhibitor, has a weak inhibitory activity for ACE. Replacement of the carboxyl group of succinyl-L-proline by a sulphydryl function as zinc chelating group produced an inhibitor, captopril (8.48), which was 1650-fold more potent and orally active. An alkyl chain length of 2 carbons, between the amide carbonyl and zinc-complexing ligand, is optimum for efficent binding. Other zinc-binding ligands such as N-carboxylate and phosphate ions also effective. AA1 substituent. The presence of an α-substituent contributes to inhibitory potency but is not essential. Methyl or benzyl substituents increase the inhibitory potency of succinyl-L-proline, but a cyclohexyl residue results in loss of potency. These substituents may contribute to increased binding, but the increase in inhibitory potency is more likely to be due to the restriction in conformation introduced at the chiral centre.

Design of enzyme inhibitors as drugs

365

AA2 substituent. Binding is enhanced by a free C-terminal carboxyl group and esterification reduces inhibitory potency. Succinyl-L-proline is the most effective, although other C-terminal aromatic amino acids and leucine residues are also acceptable but are less efficent than proline. The superiority of L-proline is probably due to its rigid structure which may lock the carboxyl group into a favourable conformation for interaction with the positively charged residue (probably protonated arginine) at the active site of the enzyme. AA1-AA2 amide bond. The presence of an amide bond in the correct position is critical for binding of the inhibitor. Analogues which either lack the amide bond or in which the amide group is displaced have significantly reduced inhibitory potency.

Development of other inhibitors has also been based on the concept of biproduct inhibition. Optimization of orientations of the C-terminus carboxyl and amide carbonyl for binding to the enzyme by incorporating some features of succiny-Lproline into an Ala-Pro ‘backbone’ led to the development of thiol-free inhibitors such as enalaprilat (8.49). This had a better side effect profile than captopril but was not active when given orally. The ethyl ester enalapril is well absorbed and is subsequently hydrolysed in vivo

Introduction to the principles of drug design and action

366

to the active inhibitor enalaprilat. Enhancement of enzyme-inhibitor interactions by introduction of bulky hydrophobic groups at the C-terminal produced inhibitors such as quinaprilat (8.50), ramiprilat (8.51) and spiraprilat (8.52). A methyl sustituent on the P1' residue enhances potency of inhibitors with proline at the C-terminal (e.g.

Design of enzyme inhibitors as drugs

367

enalaprilat), and is built into compounds with large hydrophobic residues at the Cterminus such as cilazaprilat (8.53), lisinopril (8.54), benzeprilat (8.55) and perindoprilat (8.56). With these inhibitors, the larger group appears to be sufficient to fill the S1'-S2' pockets and give the right orientation to the remaining part of the molecule for interaction with the enzyme. The phosphonate containing inhibitor, fosinoprilat (8.57), as well as ceranapril (8.58), has features of both a bi-product inhibitor and the transition-state of the enzyme catalysed reaction. With the exception of captopril and lisinopril, all inhibitors in clinical use have to be given as ester prodrugs for oral bioavailability. Apart from their use as antihypertensives, ACE inhibitors are proving to be particularly useful as an adjunct with diuretics or digoxin in the treatment of heart failure. They also appear to have a particular role in reducing blood pressure in patients with diabetic nephropathy.

Introduction to the principles of drug design and action

368

Membrane metalloendopeptidase inhibitors Membrane metalloendopeptidase (MEP, also known as enkephalinase and neutral endopeptidase) is involved in the deactivation of the enkephalin pentapeptides and other peptides and hormones including atrial natriuretic peptide (ANP), Substance P, cholecystokinin, bradykinin and chemotactic peptide. MEP cleaves the enkephalin pentapeptides (Tyr-Gly-Gly-Phe-Leu/Met) at the Gly3-Phe4 bond and much of the earlier work has focused on inhibition of enkephalin degradation (to allow a build up of the pentapeptides in vivo) in search for compounds as potential non-addictive analgesics. Although many compounds which have been developed are potent MEP inhibitors in vitro, the hope of a therapeutically useful analgesic remains to be realised. Attention has now shifted to another pharmacological aim. The loss of biological activity of ANP is the result of cleavage by MEP at the Cys7-Phe8 bond. Inhibitors which prolong the biological activity of ANP have a potential therapeutic role in the treatment of hypertension and congestive heart failure. A model of the MEP binding site, based on information from the earlier development of inhibitors of related zinc metallo-proteases such as carboxypeptidase A, thermolysin and ACE has been developed. Using MEP inhibitors with the general structure X-AA1-AA2 (X=zinc complexing ligand, AA1 and AA2=amino acid corresponding to P1' and P2' positions respectively, see 8.59), the optimal amino acid

Design of enzyme inhibitors as drugs

369

requirements for S1' and S2' subsites, zinc complexing ligands of varying affinity together with the importance of other binding sites have been determined.

Zinc binding ligands. Suitable amino acids or short peptides containing terminal zinc liganding groups such as thiol, carboxyl, phosphoramidate or hydroxamate all show inhibitory activity for MEP. From studies on hydroxamic acids, the position of the zinc-binding group seems critical, the optimal inhibitory potency being obtained when the zinc-binding ligand is separated by a single carbon from the chiral centre of the AA1 residue. S1' subsite. This site can accommodate large groups such as cyclohexyl and biphenyl moieties, but optimum inhibitory activity has been observed when benzyl is the side chain substituent on AA1. Introduction of a methyl-, methoxy- or aminosubstituent on the phenyl ring does not affect inhibitory potency but nitro- and dimethylamino-substituents reduce potency. The presence rather than the absolute configuration of the AA1 is important, indicating flexibility within the region of the active site containing the S1' subsite and zinc. However, the (S)-isomers exhibit greater inhibitory potency. S2' subsite. The binding requirements of this subsite have been established using the dipeptides Phe-Y or Tyr-Y. Compounds without a side-chain on AA2 (Phe-Gly, TyrGly) or substitution on the alpha carbon with methyl group (Phe-Ala, Tyr-Ala), an aromatic ring or a large hydrophobic residue all show good inhibitory activity, βalanine or γ-aminobutyric acid in the AA2 position also increase inhibitory activity. The (S)-isomer is the preferred configuration at this subsite. Positively charged arginine residue. This binds with the C-terminal ionized carboxyl of AA2. The presence of a free terminal carboxylate group in AA2 therefore increases binding between the enzyme and substrate or inhibitor. Inhibitors containing

Introduction to the principles of drug design and action

370

a sulphonic acid instead of a carboxy group e.g. m-aminobenzenesulphonic derivatives show a high degree of inhibitory potency and selectivity for MEP. Hydrogen bond doner group. This binds with the terminal amide (peptide) linkage. Evidence that the amide group of the peptide bond between AA1 and AA2 is hydrogen bonded to the active site of the enzyme comes from the observation that Nmethylation of the peptide link in the dipeptides Phe-Gly, Phe-Ala or Phe-Leu, leads to 100-fold reduction in inhibitory activity. The amino acid sequence of MEP has now also been determined. It consists of 749 amino acids spanning the cell membrane and includes a 27-amino acid residue cytoplasmic domain, a 13-amino acid residue hydrophobic domain and a large extracellular domain containing the active site. The three zinc-coordinating residues have been identified as His-583, His-587 and Glu-646 and Glu-584 as the residue involved in the acid-base catalytic mechanism occuring at the active site. The use of the sulphydryl group as the zinc-binding ligand inserted into dipeptides was shown to be optimal for binding with the active site and led to the development of the first potent inhibitor of MEP, thiorphan (8.59, Ki=4.7 nM). Thiorphan was also found to be a relatively efficient inhibitor of ACE (Ki=150 nM), retroinversion of the amide bond (retrothiorphan) increased selectivity for MEP (MEP, Ki=6–10 nM, ACE Ki>10 µM). The N-Carboxyalkyl-based MEP inhibitors SCH 32615 (8.60) and SCH 39370 (8.61) were developed from concepts similar to those used in the development of N-carboxyalkyl ACE inhibitors. Two aromatic amino acid residues occupying the S1 and S1' subsites combined with β-alanine or γ-aminobutyric acid at AA2 enhanced MEP inhibitory potency and selectivity over ACE. It has been proposed that the Ncarboxyalkyl group serves to bind the zinc and the β-alanine residue is a critical component in determining selectivity for MEP as significant ACE inhibitory activity is observed when alanine is present as AA2. More conformationally constrained molecules, based on γ-aminobutyric acid in the AA2 position combined with cycloleucine at AA1 led to the development of

Design of enzyme inhibitors as drugs

371

candoxatrilat (8.62, UK 69578), where the (+)-enantiomer is 30-fold more potent than the (−)-enantiomer. Phosphoramidon, a phosphoryl dipeptide of microbial origin, inhibits both thermolysin and MEP and has formed the basis for the development of specific phosphoryl inhibitors of MEP. A phosphonic acid dipeptide containing a βalanine residue (8.63) has shown selectivity for MEP. N-Phosphonomethyl dipeptide inhibitors such as CGS 24592 (8.64) were based on the observation that the ACE inhibitors fosinopril (8.57) and ceranapril (8.58) tend to be longer acting than other carboxylic acid or thiol-containing analogues. It was noted that CGS 24592 (8.64)

Introduction to the principles of drug design and action

372

underwent a very slow hydrolysis in bicarbonate solution to the derivative (8.65) which exhibited unexpected inhibitory potency for MEP (IC50=15 nM). The structure represented a significant

departure from other MEP inhibitors which contain a modified di- (or tri-) peptide backbone, with a critical secondary amide bond and a zinc-chelating ligand. Modification of the C-terminal carboxylic acid functionality of (8.65) to a tetrazole led to a highly potent, non-peptide MEP inhibitor CGS 26303 (8.66).

Design of enzyme inhibitors as drugs

373

As with ACE inhibitors, these MEP inhibitors are not well absorbed orally, which limits their potential therapeutic usefulness. To improve pharmacokinetic profiles, the inhibitors have been further developed as prodrugs such as sinorphan (prodrug of (S)thiorphan), SCH 34826 (a lipophilic ester of 8.59), UK 79 300 (an indanyl ester of (+)isomer of (8.62)) and CGS 25462 and CGS 26393 (the aminomethyl phosphonate derivatives of 8.64 and 8.66 respectively). A different approach to improving potential therapeutic efficacy in the development of non-addictive analgesics has been the realisation of combined inhibitors of more than one enzyme in a single inhibitor. Kelatorphan (8.67) inhibits MEP, aminopeptidase N and dipeptidylaminopeptidase, the enzymes involved in inactivation at different points of the enkephalin pentapeptides in the CNS. A variation of this approach has been the concept of covalently linking two different types of inhibitor in a ‘prodrug’. An aminopeptidase N (APN) inhibitor and a MEP inhibitor have been linked by a thioester or a disulphide bond in order to increase the hydrophobicity, and so absorption, of each molecule (8.68). Hydrolysis or reduction, respectively, leading to the release of the two active inhibitors, occurs once the compound has passed the blood brain barrier.

Introduction to the principles of drug design and action

374

Combined inhibitors such as the mercaptoalkyl derivatives alatrioprilat (8.69) and glycoprilat (8.70) display both MEP and ACE inhibitory activity and are being assesed for their therapeutic potential in the treatment of cardiovascular diseases. 8.6.1.3 Aspartate proteases HIV protease inhibitors Two genetically distinct subtypes, HIV-1 and HIV-2, of human immunodeficiency virus (HIV) have been identified. Reverse transcriptase inhibitors such as AZT have had limited success because of emergence of viral resistance and drug toxicity. Blockade of the virally encoded protease, which is critical for viral replication, has become a major target in the search for an effective anti-viral agent and several inhibitors are now under development. HIV-1 protease catalyzes the conversion of a polyprotein precursor (encoded by gag and pol genes) to mature proteins needed for the production of infectious HIV particle. A highly conserved triad, Asp-Thr(Ser)-Gly, in the viral enzyme which is also found in mammalian proteases belonging to the aspartic acid family, suggested a

Design of enzyme inhibitors as drugs

375

similar mechanistic class for HIV protease. This has now been confirmed by elucidation of the crystal structure of the native HIV protease and the HIV protease complexed with aspartyl protease inhibitors. There are however significant structural differences between the retroviral and classical aspartyl proteases such as renin. Mammalian and fungal aspartyl proteases generally comprise of 200 amino acids and consist of two homologous domains with the key catalytic triad occuring twice. The structure of HIV protease has been identified by X-ray crystallographic methods as a homodimer comprising of two identically folded subunits (each comprising of 99 amino acids). Each subunit contributes one of the two conserved aspartates (Asp 25 and Asp 125) to the single hydrophobic active site cavity. It is believed that during hydrolysis, a water molecule attacks the carbonyl carbon of the peptide bond of the substrate while the carbonyl oxygen accepts the proton from one of the catalytic aspartic acid residues leading to the formation of a tetrahedral transition state. Catalytic studies have suggested that in the transition state, one of the aspartic acid residues exists in the neutral form whereas the other residue is negatively charged. However, the protonation state of the protease aspartic acid residues in the complex with its inhibitors remains controversial. After the formation of the transition state, two conformationally flexible flaps (one per subunit) close around the substrate. HIV-protease cleaves the polyprotein precursor at eight different sites, of which Tyr-Pro and Phe-Pro residues (occuring as P1-P1' at three of the cleavage sites of HIV1), are of particular interest in relation to the development of inhibitors. The amide bonds N-terminal to proline are not hydrolysed by mammalian aspartic proteases and therefore offer selectivity for the viral enzyme. Leu-Ala, Leu-Phe, Met-Met, and PheLeu are also found at HIV-1 cleavage sites. The amino-acid sequences flanking the cleavage have been divided into three classes. Class 1: Phe-Pro or Tyr-Pro at P1-P1' Class 2: Phe-Leu at P1-P1' and Arg at P4 Class 3: Gln or Glu at P2' Studies using oligopeptides have shown that seven residues spanning P4-P3' are required for specific and efficient hydrolysis of the P1-P1' amide bond and crystallographic data suggests multiple hydrogen bonding to the backbone of inhibitors spanning this site and close van der Waals contact for the P3-P3' side chains. Incorporation of a transition-state mimic into substrate analogues has been one of the strategies used in the development of enzyme inhibitors. Substitution of the scissile amide bond with non hydrolyzable dipeptide isosteres in appropriate sequence context has also proved to be successful in the development of potent renin inhibitors. A number of such dipeptide isosteres (inserted into a heptapeptide template spanning P4P3' and which mimic the tetrahedral intermediate of peptide hydrolysis) have been evaluated. Hydroxyethylene (8.71), dihydroxyethylene (8.72) and hydroxyethylamine (8.73) isosteres provide the greatest intrinsic affinity for the enzyme. The order of affinity of other isosteres for HIV-1 protease has been established as difluoroketones (8.74)=statine (8.75)> phosphinate (8.76)>reduced amide isostere (8.77). The principle structural features in most transition state analogues designed to inhibit HIV protease is the critical hydroxyl group shown by x-ray analysis to bind both aspartic acid groups.

Introduction to the principles of drug design and action

376

Pepstatin A (Iva-Val-Val-Sta-Ala-Sta), a natural product, contains two residues of the amino-acid statine. It is a non-specific inhibitor of aspartic acid proteases and inhibits several retroviral proteases, including the hydrolysis of both polyprotein and oligopeptide

Design of enzyme inhibitors as drugs

377

substrates by HIV-1 protease. The concentration of inhibitor required to inhibit HIV-1 protease are significantly higher than those required for mammalian or fungal aspartic proteases. The structure of H-261 (8.78) mimics the cleavage sequence of the renin substrate angiotensinogen (Leu-Val). It is also non-specific and inhibits both HIV-1 (Ki=5 nM) and HIV-2 (Ki=35 nM) protease. Analogues incorporating the cyclohexyalanine-Val hydroxyethylene isostere, U-81749 (8.79, Ki=70 nM) and the dihydroxyethylene isostere of cyclohexylalanine-Val, U-75875 (8.80, Ki<1 nM) both show potent antiviral activity in cell cultures.

Adaptation of the hydroxyethylamine dipeptide isostere to mimic the Phe-Pro site has produced inhibitors with selectivity for the retro viral protease (8.81, Ki=0.66 nM). Conversion of the proline to a decahydroisoquinoline nucleus has been very successful

Introduction to the principles of drug design and action

378

in the development of the potent selective HIV protease inhibitor saquinavir (8.82, Ki< 0.12 nM) which has recently gained licensing approval in USA for clinical use. Unlike compound JG-365 (8.83), where the crystal structure has shown a preferance for the (S)-hydroxyl enantiomer of the isostere fragment of the molecule, the (R)configuration is preferred for saquinavir (R-enantomer IC50=0.4 nM, S-enantomer IC50=>100 nM). X-ray crystallography studies have shown that the hydroxyl group is located between the aspartic acids in both JG-365 and saquinavir, but the adjacent methylene groups fit in a different manner into the active site. SC 52151 (8.84, IC50=6.3 nM), based on hydroxyethylurea isostere has oral bioavailability. L 735 524 (8.85, IC50=0.36 nM), which is a combination of a

Design of enzyme inhibitors as drugs

379

hydroxyethylene isostere and a hydroxyethylamine isostere, is also orally active. The sulphonamido moiety, in the novel (R)-hydroxyethyl sulphonamides isostere (8.86), has also been used to replace the P1'P2' amide linkage of the inhibitor (8.87, Ki=1 nM).

Introduction to the principles of drug design and action

380

Symmetrical inhibitors (8.88, IC50=0.2 nM; 8.89, Ki=0.8 µM) capitalize on the unique symmetry of the homodimeric enzyme. Unlike transition-state analogues, the

Design of enzyme inhibitors as drugs

381

Introduction to the principles of drug design and action

382

stereochemistry of the two hydroxyl groups is not significant. Modifications to evaluate the effect of polar heterocyclic end groups led to the nonsymmetrical inhibitor A77003 (8.90, IC50<1 nM). Improved oral bioavailabilty was obtained with A 80987 (8.91, Ki= 0.25 nM) where the methylamide groups had been replaced by esters. Penicillin-derived symmetrical dimers (8.92) have been identified as good lead structures from screening programmes. A hybrid of a penicillin derived structure and statine isostere shows good inhibitory activity (8.93, Ki=0.25 nM). The antipsychotic agent haloperidol (8.94, Ki=100 µM) was identified as a weak inhibitor through a computational search of a structural database based on a complementary shape of the HIV-1 protease active site. The 1,3-dithiolane analogue (8.95, K=15 µM) exhibits greater inhibitory potency.

Design of enzyme inhibitors as drugs

383

Introduction to the principles of drug design and action

384

Design of enzyme inhibitors as drugs

385

8.6.2 Acetylcholinesterase inhibitors Acetylcholine is the chemical transmitter released at the nerve endings in the parasympathetic and motor nervous systems following a nervous impulse. After a response from the tissue the acetylcholine is removed by hydrolysis to inert products by acetylcholinesterase (see Equation [8.35]) in the proximity. Inhibitors of acetylcholinesterase allow a build up of acetylcholine at the nerve endings so that a more prolonged effect is produced which is useful in the treatment of myasthenia gravis, a disease associated with the rapid fatigue of muscles, as well as in the treatment of glaucoma where stimulation of the ciliary body improves drainage from the eye and decreases intra-ocular pressure. A more recent potential use has been in the treatment of Alzheimers disease and senile dementia of the Alzheimer’s type (SDAT).

(8.35) Inhibitors of acetylcholinesterase fall into two groups: the reversible carbamate inhibitors such as eserine (physostigmine (8.96)), neostigmine (8.97) and benzylpyrinium (8.98) and the irreversible organophosphorous inhibitors, dyflos

Introduction to the principles of drug design and action

386

(8.99) and ecothiopate (8.100). The carbamates carry a positive charge and are bound at the anionic site (carboxylate ion) of the enzyme and correctly positioned to form a carbamyl enzyme with the serine hydroxyl group at the esteratic site (see Equation [8.36]). The carbamyl enzyme is only slowly decomposed (t1/2=~20 min) and in the presence of excess inhibitor the enzyme is partially locked up in this form so that its activity towards the substrate acetylcholine is decreased. Dilution or removal of excess inhibitor leads to a shift in the steady-state inhibition level with an increase in activity of the enzyme.

(8.36) The organophosphorus compounds rapidly react with the enzyme to form a stable phosphoryl enzyme and the enzyme is irreversibly inhibited (see Equation [8.37]).

Design of enzyme inhibitors as drugs

387

(8.37) The organophosphorus compounds have a long duration of action in the body after a single dose of the drug and enzyme activity only returns after synthesis of fresh enzyme. Due to dangers of overdosage, as well as handling, they are little used except for treatment of glaucoma where the other less toxic cabamate drugs have not proved satisfactory in a particular therapy.

Volatile organophosphorus compounds such as sarin (8.101) and tabun (8.102) have been prepared for use as nerve gases in war and other less volatile compounds have been used as insecticides for the spraying of crops. Inhibition of the mammalian or insect enzyme leads to a build-up of acetylcholine and death from accumulated acetylcholine poisoning. Much research has been carried out to find antidotes, for nerve gas poisoning, which could be distributed to the population in the event of war. One of these discoveries, pyridine-2-aldoxime mesylate (pralidoxime (8.103)) has been successfully used, in conjunction with atropine to block the action of acetylcholine on receptors, in the treatment of accidental poisoning during crop spraying. Pralidoxime is considered to complex at the anionic site where it is firmly held by electrostatic attraction in the correct spacial configuration for attack by the oxime anion on the phosphorus atom with displacment of the inhibitor residue from the enzyme. There is evidence that Alzheimer’s disease and senile dementia of the Alzheimer’s type (SDAT) are associated with dysfunction of normal cholinergic neurotransmission in the brain leading to learning and memory deficiencies. Examination of patients with these diseases has shown reduced levels of ChAT (acetyl-Co A: choline Otransferase), acetylcholinesterase and the muscarinic receptor sub type M1. ChAT is responsible for the synthesis of acetylcholine in the cerebral cortex and it has been postulated that by inhibiting acetylcholinesterase in the brain the associated build up

Introduction to the principles of drug design and action

388

of acetylcholine will enhance the cognitive function. Three main types of inhibitors have been used which lead to improvement in SDAT patients. The carbamate inhibitors exemplified by physostigmine (8.96) which itself has a poor pharmacokinetic profile but this is improved in the N-heptyl derivative heptylphysostigmine (eptastigmine). Tacrine has been approved for use for Alzheimer’s patients but is non-specific in its action by also inhibiting plasma butyrylcholinesterase leading to adverse peripheral effects. Its use also leads to hepatotoxicity as a significant side effect. Recent work has centred on the third group of inhibitors, the benzylpiperidines. These are not quaternised and so are able to penetrate the blood brain barrier as the unionised base form and are potent selective inhibitors of acetylcholinesterase with good pharmacokinetic properties. Compounds in clinical trial are E-2020 (8.104) and TAK-147 (8.105).

8.6.3 Aromatase inhibitors Aromatase belongs to a group of cytochrome P-450 enzymes responsible for hydroxylation processes in the body. It contains a Fe3+ -haem catalytic site which, after reduction to Fe2+, binds and activates oxygen, leading to initial insertion of two hydroxyl groups on the C-19 (methyl) carbon of its substrates androstenedione and testosterone. A further hydroxylation occurs, followed by aromatization to oestrone

Design of enzyme inhibitors as drugs

389

and oestradiol, respectively, accompanied by elimination of water and formate by a mechanism only partially understood. The steroidogenic pathway (see Figure 8.6) from cholesterol to the substrates of aromatase commences in the adrenals with the action of the cytochrome P-450 enzyme, cholesterol side chain cleavage enzyme (CSCC), producing pregnenolone which is then isomerized by another enzyme to progesterone. Progesterone is converted by 17 α-hydroxy: 17,20-lyase (P450 17), another P-450 enzyme, to androstenedione which can be reduced by a dehydrogenase to testosterone. Aromatase is located mainly in fatty tissue in postmenopausal women and mainly in ovarian tissue in premenopausal women. After diagnosis of a breast tumour, it is removed by surgery and this is followed by a course of chemotherapy to reduce new tumour growth or suppress metastasis in other parts of the body. Mammary tissue contains oestrogen receptors, and depending on their concentration the patient can be categorized as either oestrogen receptorpositive (ER+) or negative (ER−). About one-third of the cases of breast cancer in women are hormone-dependent, the major hormone involved in supporting the growth of the tumours being oestradiol. The categorization can determine the type of chemotherapeutic treatment employed. The first line drug for use in the treatment of mammary cancer in postmenopausal women with (ER+) and (ER−) tumours is tamoxifen. This is an oestrogen receptor antagonist which, by competing with oestradiol for the receptor, can reduce the ability of oestradiol to stimulate tumour growth. Tamoxifen has weak oestrogenic activity and compounds ICI 164,384 and ICI 182,780 without this effect are now in clinical trial.

Introduction to the principles of drug design and action

390

Figure 8.6 Steroidogenesis pathway. Tamoxifen-resistant tumours (ER+) are sometimes amenable to treatment with a second line drug which is an aromatase inhibitor. This reduces the plasma level of

Design of enzyme inhibitors as drugs

391

circulating oestradiol available to the tumour tissue by inhibiting the action of aromatase, present in the fatty tissue, on androstenedione. The non-steroidal aromatase inhibitor aminoglutethimide (8.106) is in current clinical use for the treatment of (ER+) breast cancer in postmenopausal women. On chronic administration of the drug, the already low plasma oestrogen level present in elderly women is further rapidly lowered and maintained, enabling a success rate in terms of remission or stabilization of about 33% (unselected patients) or 52% (ER+ patients). Aminoglutethimide was initially introduced into therapy as an anti-epileptic drug, but after initial withdrawal due to noted side-effects of adrenal insufficiency it was reintroduced into cancer chemotherapy due to its potential effect for interrupting the steroidogenic pathway to oestrogen production. Subsequent work showed that it was a potent, competitive, reversible inhibitor of aromatase with a weaker effect on the CSCC enzyme (which accounts for its effects on adrenal hormone production). Aminoglutethimide is co-administered with hydrocortisone to supplement decreased production of 11β-hydroxysteroids due to its effect on CSCC. Side-effects associated with use of the drug are ataxia, dizziness and lethargy, due to its sedative nature. These effects, which can lead to patient non-compliance, decrease after several weeks’ administration of the drug. Consequently, more specific inhibitors without these side-effects have been sought. Several anti-fungal agents based on imidazole e.g. ketoconazole, econazole were known at this time which inhibit the fungal P450 14α-demethylase enzyme. They are inhibitors of aromatase but have a wide spectrum of activity against other P450 enzymes in the steroidogenic chain. Several potent specific inhibitors of aromatase containing an imidazole or triazole nucleus (increased in vivo stability) have subsequently been developed. Fadrozole (8.107), (+)- vorozole (8.108), letrozole (8.109) and arimidex (6.27) (achiral) are now in clinical trial. These compounds are 400–1000 fold more potent than aminoglutethimide and have no CNS effects. Fadrozole also inhibits the 18-hydroxylase enzyme responsible for aldosterone production at doses much higher than used clinically; this side effect has been designed out in the more selective letrozole. Mechanism-based inactivators of aromatase are known and these are based on the androstenedione (substrate) skeleton. 4-Hydroxyandrostenedione (8.110) is in clinical trial as a intramuscular injection (formestane) given once weekly and is a specific irreversible inhibitor of the enzyme although the mechanism is not clear. It has to be administered parenterally since it is rapidly metabolised by first-pass metabolism following oral administration. It has been reported to produce a 30% complete or partial tumour regression with disease stabilisation in a further 15% of patients. Other steroidal irreversible inhibitors in trial include plomestane (8.111) and exemestane (8.112). Recent views are that breast tissue is capable of synthesising oestrogens mainly from the action of a sulphatase on oestrone sulphate. The oestrone produced provides oestradiol by the action of a 17β-hydroxysteroid dehydrogenase. Inhibitors of steroid sulphatase are being developed as potential adjuvants to aromatase inhibitors to further deplete oestrogen levels. One of these compounds emate (8.113) is an irreversible inhibitor of the sulphatase.

Introduction to the principles of drug design and action

392

8.6.4 Pyridoxal phosphate-dependent enzyme inhibitors Enzymes using pyridoxal phosphate as coenzyme catalyse several types of reactions of amino acid substrates, such as (1) transamination to the corresponding α-ketoacid, (2) racemization, (3) decarboxylation to an amine, (4) elimination of groups on the β-and γ-carbon atoms, (5) oxidative deamination of ω-amino acids. The coenzyme is bound to the enzyme by formation of an aldimine (Schiff-base) with the ω-amino group

Design of enzyme inhibitors as drugs

393

Introduction to the principles of drug design and action

394

of a lysine residue. The first step in the reaction with the amino acid substrate is an exchange reaction to form an aldimine with the α-amino group of the amino acid (see Equation [8.38]. Either by hydrogen abstraction (transamination, racemization) or by decarboxylation, a negative charge is developed on the α-carbon atome and this is distributed over the whole conjugated cofactor system. Protonation then occurs on either the α-carbon atom (decarboxylation, racemization) or on the carbon atom adjacent to the pyridine ring (transamination) as shown in Equation [8.38]. The direction of the fission which occurs is dictated by the nature of the protein at the active site so that a specific enzyme catalyses a particular type of reaction. Information has recently become available on the crystal structure of several of these enzymes and the role of their active site residues.

(8.38)

Design of enzyme inhibitors as drugs

395

At one time, several irreversible inhibitors of several pyridoxal phosphate-dependent were known but their mechanism of action was not clear since they did not possess the electrophilic centres present in the active site directed irreversible inhibitors known at that time. Later, when a new class of inhibitor, the mechanism-based enzyme inactivator, became known their inhibition mechanism became predictable from the well established mechanism of action of these enzymes. The next step for design was to manipulate the amino acid substrate structure of a suitable target enzyme in such a manner as to obtain maximal exploitation of the enzyme’s machinery. The inhibitors act as substrates of the enzyme but their structure is such that they either (1) divert the electron flux from the α-carbanion formed away from the coenzyme moiety, or (2) using the normal electron flux either give rise to reactive species or generate a stable substrate-cofactor which binds strongly to the enzyme active site. All these mechanisms can lead to irreversible inhibition of the enzyme. Mechanism-based inactivators of many pyridoxal phosphate-dependent enzymes are known but only a few target enzymes and their inactivators of therapeutic interest will be discussed here. 8.6.4.1 GABA transaminase (GABA-T) inhibitors γ-Aminobutyric acid (GABA) is considered as the main inhibitory neurotransmitter in the mammalian central nervous system. There has been much interest recently in the design of inhibitors of the pyridoxal phosphate-dependent enzyme, α-ketoglutarateGABA transaminase. This enzyme governs the levels of GABA in the brain (see Equation [8.39]). Inhibitors of the enzyme would allow a build-up of GABA and could be used as anticonvulsant drugs for the treatment of epilepsy.

(8.39) γ-Acetylenic GABA (8.114) is a time-dependent inhibitor of GABA-T but also inhibits other pyridoxal phosphate-dependent enzymes, γ-vinyl GABA (vigabatrin, 8.115) acts in a similar manner through its (S)-enantiomer but has a more specific action. Vigabatrin has shown promise as a drug for the treatment of epilepsy. Studies on drug-resistant epileptic patients have indicated that additional therapy with vigabatrin reduces seizures by over 50% in more than half the population studied without develpment of tolerance. Halomethyl derivatives of GABA have also been described as inhibitors of GABAT. The fluoromethyl derivative (8.116) is the best time-dependent inhibitor and the inactivation is accompanied by elimination of fluoride ion. Shortening of the chain of (8.116) to give the β-alanine derivatives (8.117) and (8.118) produced inhibitors with

Introduction to the principles of drug design and action

396

similar kinetic constants. However, in vivo, (8.118) was almost 100-fold more active than vigabatrin but showed unexplained delayed toxicity after a single administration and was not further developed. The mechanism of action of the GABA-T inhibitors based on GABA and bearing either an unsaturated function or a leaving group has not yet been clearly elucidated. The initial postulated mechanism was that these two groups of inhibitors formed a Schiff-base with pyridoxal phosphate, followed by loss of the α-carbon proton. With the unsaturated derivatives the electron flow that followed was towards the coenzyme moiety to give the vinylimine (8.119), whereas with the fluoromethyl derivatives the electron flow was away from the coenzyme and accompanied by loss of

fluoride ion to give the enimine (8.120) (see Equations [8.40] and [8.41]). The electrophilic centres developed in the conjugated systems by the normal or abnormal electron flow react with a nucleophile at the active site of the enzyme. More recent work with other pyridoxal phosphate-dependend enzymes has suggested an alternative mechanism, which is llustrated in Equation [8.42] for the fluoromethyl derivatives. Here the enimines dissociate from the pyridoxal phosphate to give an enamine which then recombines with the lysine of the active site. The co-factor is then attacked by the electrophilic centre of the enamine to yield a stable complex at the active site which leads to irreversible inhbition of the enzyme.

Design of enzyme inhibitors as drugs

397

(8.40)

(8.41)

(8.42) Another potent irreversible inhibitor of GABA transaminase is gabaculine (8.121), which is a naturally occurring neurotoxin isolated from Streptomyces toxacaenis. Although not of clinical application, this inhibitor is interesting since it is considered

Introduction to the principles of drug design and action

398

to inhibit the enzyme by a different mechanism to that previously described for suicide inactivators. Gabaculine acts as a substrate and is converted in the normal manner to the ketimine (8.122) (Equation [8.43]). This then aromatizes under the influence of a basic group to form a stable enzyme-bound pyridoxamine derivative and the enzyme is inactivated.

(8.43) 8.6.4.2 Peripheral aromatic amino acid decarboxylase (AADC) inhibitors Noradrenaline is synthesized at the nerve endings of the postganglionic fibre by a series of reactions from tyrosine (see Equation [8.44]). Inhibitors of AADC have been synthesized as potential antihypertensive drugs on the basis that a decrease in the biosynthesis of noradrenaline would deplete noradrenaline stores at the nerve endings and lead to a decrease in blood pressure. Although many reversible inhibitors of the enzyme are known from in vitro studies (e.g. methyldopa (8.123)) only a few exert an antihypertensive action in vivo and probably by an alternative mechanism since inhibition of the first step in the pathway is not involved (see Section 8.2.1). However, this work led to the discovery of the inhibitors carbidopa (8.124) and serazide (8.125) which have proved useful as adjuvants in the treatment of Parkinson’s disease with Ldopa. L-Dopa penetrates into the basal ganglia of the brain where it is decarboxylated to the active agent, dopamine. Large doses of L-dopa are required in therapy since it is depleted in the plasma by peripheral AADC to dopamine which is readily removed by monoamine oxidase. Combination of L-dopa with serazide or carbidopa leads to decreased metabolism of L-dopa so that smaller effective

Design of enzyme inhibitors as drugs

399

(8.44)

doses can be used in therapy which have fewer side-effects than large doses. Necessary features of these inhibitors are that they do not penetrate the blood-brain and interfere with the decarboxylation of L-dopa to dopamine in the brain and, for the reason given above, neither do they reduce the synthesis of endogenous amines in the peripheral tissues. AADC is a pyridoxal phosphate-dependent enzyme and serazide and carbidopa are potent pseudo-irreversible inhibitors of the enzyme. They probably function by binding to pyridoxal phosphate as carbonyl group reagents.

Introduction to the principles of drug design and action

400

Several mechanism-based inactivators of AADC have been described and the αmonofluoromethyl (8.126) and α-difluoromethyl (8.127) derivative of dopa deserve special mention here. These compounds are time-dependent irreversible inhibitors of the enzyme. During one turnover of the inhibitor by the enzyme, one equivalent each of CO2 and F− is released and the inhibitor binds in a 1:1 ratio to the enzyme-cofactor complex. It is not clear whether the reaction pathway occurs through the mechanism shown in Equation [8.41] or in Equation [8.42]. α-Difluoromethyldopa is comparable to carbidopa and effectively protects exogenous dopa against decarboxylation. It inhibits brain AADC only at high concentrations. α-Monofluoromethyldopa effectively inhibits AADC centrally as well as peripherally and the resulting depletion of peripheral catecholamines produces antihypertensive effects which can be reversed by i.v. infusions of dopamine. 8.6.4.3 Ornithine decarboxylase (ODC) inhibitors Naturally occurring polyamines such as putrescine, spermidine and spermine are required for cellular growth and differentiation. Spermidine and spermine are derived in human-type cells from putrescine. Putrescine is synthesized by decarboxylation of ornithine, catalysed by the pyridoxal phosphate-dependent enzyme ornithine decarboxylase (ODC) (Equation [8.45]). ODC has a very short biological half-life and its synthesis is stimulated ‘on demand’ by trophic agents and controlled by putrescine and spermidine levels. ODC has been considered a suitable target enzyme for the control of growth in tumours and disease caused by parasitic protozoa.

(8.45)

α-Difluoromethylornithine (eflornithine; (8.128)) is a mechanism-based inactivator of the enzyme and irreversibly inhibits the enzyme by the general mechanism previously depicted with elimination of a single fluoride ion to produce a conjugated electrophilic imine (c.f. 8.120) which reacts with the nucleophilic thiol of Cys-390. A further fluoride ion is then eliminated which, following transaldimation with Lys-69, leads to

Design of enzyme inhibitors as drugs

401

the species Cys-390-S-CH=C(NH2)-(CH2)3-NH2 which loses ammonia and cyclises to (2-(1-pyrroline)methyl) cysteine. Eflornithine has low toxicity in animals and has shown antineoplastic and antiprotozoal actions in clinical trials. E-2-(fluoromethyl) dehydroornithine (8.129), is another derivative which is a mechanism-based inactivator of the enzyme. The methyl or ethyl esters are effectively hydrolysed at the higher cellular concentrations attained due to improved absorption and are 10 time more effective than eflornithine or (8.129) in decreasing ODC activity in animal tissues, an effect which is long lasting. The methyl ester is also more active than eflornithine against trypanosomiasis and malaria in mice. FURTHER READING Aldridge, W.N. (1989) Cholinesterase and esterase inhibitors and reactivation of organophosphorus inhibited esterases. In Design of Enzyme Inhibitors as Drugs, edited by M.Sandler and H.J.Smith, Vol. 1, pp. 294–313. Oxford: Oxford University Press. (a) Bode, W., Huber, R., Rydel, T.J. and Tulinsky, A. (1992) X-Ray crystal structures of human α-thrombin and of the human thrombin-hirudin complex. (b) Powers, J.C. and Kam, C.-M. (1992) Synthetic substrates and inhibitors of thrombin. (c) Stone, S.R. and Maraganore, J.M. (1992) Hirudin interactions with thrombin. In Thrombin—Structure and function, edited by L.J.Berliner, (a) pp. 3–61, (b) pp. 117–159, (c) pp. 219–256. New York, London: Plenum Press. Brodie, A. (1994) Aromatase inhibitors. In Design of Enzyme Inhibitors as Drugs, edited by M.Sandler and H.J.Smith, Vol. 2, pp. 424–438. Oxford: Oxford University Press. Claeson, G., Scully, M.F., Kakkar, V.V. and Deadman, J. (eds.) (1993) The design of synthetic inhibitors of thrombin, Vol. 340 of Advances in Experimental Medicine and Biology. De Lombaert, S., Blanchard, L., Tan, J., Sakane, Y., Berry, C. and Ghai, R.D. (1995) Biorganic and Medicinal Chemistry Letters 5, 145–150. Demuth, H.U. (1990) Recent development in inhibitory cysteine and serine proteases. Journal of Enzyme Inhibition 3, 249–278. Drake, P.L. and Huff, J.R. (1994) HIV protease as an inhibitor target for treatment of AIDS. In Advances in Pharmacology, edited by J.T.August and M.W.Anders, Vol. 25, pp. 399–454. New York: Academic Press. Edwards, P.D. and Bernstein, P.R. (1994) Synthetic inhibitors of elastase. Medicinal Research Reviews 14, 127–194. Fisher, J.F., Tarpley, W.G. and Thaisrivongs, S. (1994) HIV protease inhibitors in Design of Enzyme Inhibitors as Drugs, edited by M.Sandler and H.J.Smith, Vol. 2, pp. 226–289. Oxford: Oxford University Press. Fournie-Zaluski, M.-C., Coric, P., Turcaud, S., Lucas, E., Noble, F., Maldonado, R. and Roques, B.P. (1992) “Mixed Inhibitor-Prodrug” as a new approach towards systemically active inhibitors of enkephalin—degrading enzymes. Journal of Medicinal Chemistry 35, 2473–2481.

Introduction to the principles of drug design and action

402

Hlasta, D.J. and Pagani, E.D. (1994) Human leukocyte elastase inhibitors. In Annual Reports in Medicinal Chemistry, edited by J.A.Bristol, pp. 195–204. New York: Academic Press Inc. John, R.A. (1995) Pyridoxal phosphate-dependent enzymes. Biochimica et Biophysica Acta, Protein Structure and Molecular Enzymology 1248(2), 81–96. Jung, M.J. and Danzin, C. (1989) New developments in enzyme-activated irreversible inhibitors of pyridoxal phosphate-dependent enzymes of therapeutic interest. In Design of Enzyme Inhibitors as Drugs, edited by M.Sandler and H.J.Smith, Vol. 1, pp. 257–293. Oxford: Oxford University Press. Leonetti, G. and Cuspidi, C. (1995) Choosing the right ACE inhibitor. Drugs 49, 516– 535. Markwardt, F. and Stürzebecher, J. (1989) Inhibitors of trypsin-like enzymes with a physiological role. In Design of Enzyme Inhibitors as Drugs, edited by M.Sandler and H.J.Smith, pp. 619–649. Oxford: Oxford University Press. Patel, A., Smith, H.J. and Sewell, R.D.E. (1993) Inhibitors of enkephalin-degrading enzymes as potential therapeutic agents. In Progress in Medicinal Chemistry, edited by G.P.Ellis and D.K.Luscombe, Vol. 30, pp 327–378. Amsterdam: Elsevier Science Publ. Rabasseda, X., Mealy, J. and Castaner, J. (1995) TAK-147. Drugs of the Future 20, 248–250. (a) Sandler, M. and Smith H.J. (1989) Introduction to the use of enzyme inhibitors as drugs. (b) Frick, L. and Wolfenden, R. (1989) Substrate and transition-state analogue inhibitors. (c) Shaw, E. (1989) Active-site-directed irreversible inhibitors. (d) Tipton, K. (1989) Mechanism-based inhibitors. In Design of Enzyme Inhibitors as Drugs, edited by M.Sandler and H.J.Smith, Vol 1, (a) pp. 1–18; (b) pp. 19–48; (c) pp. 49–69; (d) pp. 70–93. Oxford: Oxford University Press. Schwartz, J.C., Gros, C., Duhamel, P., Duhamel, L., Lecomte, J.M. and Bralet, J. (1994) “Atriopeptidase” (EC 3.4.24.11) inhibition and protection of atrial natriuretic factor. In Design of Enzyme Inhibitors as Drugs, edited by M.Sandler and H.J.Smith, Vol. 2, pp. 739–754. Oxford: Oxford University Press. Silverman, R.C. (1988) Mechanism-based enzyme inactivation: chemistry and enzymology, Vol. 1, pp. 3–23. Boca Raton: CRC Press, Inc. (a) Slater, A.M., Timms, D. and Wilkinson, A.J. (1994) Computer-aided molecular design of enzyme inhibitors. (b) Luscombe, D.K., Tucker, M., Pepper, C.J., Nicholls, P.J., Sandler, M. and Smith, H.J. (1994) Enzyme inhibitors as drugs: from design to the clinic. In Design of Enzyme Inhibitors as Drugs, edited by M.Sandler and H.J.Smith, Vol. 2, (a) pp. 42–64; (b) pp. 1–41. Oxford: Oxford University Press. Stubbs, M.T. and Bode, W. (1993) A player of many parts: the spotlight falls on thrombin structure, Thrombosis Research 69, 1–58. Vanden Bossche, H. (1992) Inhibitors of P450-dependent steroid biosynthesis: From research to medical treatment. Journal of Steroid Biochemistry and Molecular Biology 43, 1003–1021.

9. THE CHEMOTHERAPY OF CANCER DAVID E.THURSTON and SYLVIA G.M.J.LOBO CONTENTS 9.1 INTRODUCTION

333

9.2 TERMINOLOGY

334

9.3 METASTASES

334

9.4 CELL GROWTH CYCLE

335

9.4.1 Excessive cell proliferation

335

9.4.2 Loss of tissue-specific characteristics

335

9.4.3 Invasiveness

335

9.4.4 Metastasis

336

9.5 MECHANISMS OF TUMOUR FORMATION 9.5.1 Internal factors

336 336

9.5.1.1 Mutation

336

9.5.1.2 Addition or loss of genetic material

336

9.5.1.3 Changed gene expression

336

9.5.2 External factors

336

9.5.2.1 Viruses

336

9.5.2.2 Chemicals

337

9.5.2.3 Radiation

338

9.5.3 Hereditary factors

338

9.6 TREATMENT

338

9.6.1 Surgery

338

9.6.2 Radiation therapy (Radiotherapy)

339

Introduction to the principles of drug design and action

404

9.6.3 Photodynamic therapy

339

9.6.4 Immunotherapy and vaccines

339

9.6.5 Chemotherapy

340

9.6.5.1 Achievements of chemotherapy

340

9.6.5.2 Discovery of drugs and their preclinical evaluation

340

9.6.5.3 Accessibility of tumour cells to drugs

340

9.6.5.4 Achieving selective toxicity

341

9.6.5.5 Limiting the toxicity of chemotherapeutic agents

341

9.6.5.6 Overview of the mode of action of chemotherapeutic agents

342

9.6.5.7 Drug resistance

343

9.6.5.8 Combination chemotherapy

343

9.6.5.9 Adjuvant chemotherapy

344

9.7 ANTIMETABOLITES

344

9.7.1 DHFR Inhibitors (antifolates)

344

®

344

9.7.1.1 Methotrexate (Maxtrex ) 9.7.2 Purine antimetabolites

346

9.7.3 Pyrimidine antimetabolites

347

9.8 DNA INTERACTIVE COMPOUNDS 9.8.1 Cross-linking agents

348 348

9.8.1.1 Nitrogen mustards

348

9.8.1.2 Aziridines

352

9.8.1.3 Methanesulphonates

352

9.8.1.4 Triazeneimidazoles

353

9.8.1.5 Imidazotetrazinones

354

9.8.1.6 Nitrosoureas

354

9.8.1.7 Metal complexes

356

The chemotherapy of cancer

405

9.8.1.8 Carbinolamines

356

9.8.1.9 Cyclopropanes

357

9.8.1.10 Procarbazine

358

9.8.2 Intercalating agents

359

9.8.2.1 Anthracyclines

359

9.8.2.2 Anthracenes

360

9.8.2.3 Phenoxazines

361

9.8.3 Topoisomerase inhibitors

362

9.8.3.1 Ellipticine

362

9.8.3.2 Camptothecin

362

9.8.3.3 Etoposide

363

9.8.4 DNA Cleaving agents

363

9.8.4.1 The Bleomycins

363

9.8.4.2 The Enediynes

365

9.9 ANTITUBULIN AGENTS

365

9.9.1 Vinca alkaloids

365

9.9.2 The Taxanes

366

9.10 MISCELLANEOUS AGENTS 9.10.1 Hydroxyurea

367 367

9.10.2 Mithramycin 368 9.10.3 Mitotane 368 9.11 IMMUNOSUPPRESSIVE AGENTS 368 9.11.1 Azathioprine 368 ®

9.11.2 Cyclosporin (Neoral ) 369

Introduction to the principles of drug design and action

406

9.12 ANTI-HORMONAL AGENTS 369 9.12.1 Breast cancer 369 9.12.1.1 Anti-oestrogens 370 9.12.1.2 Aromatase inhibitors 370 9.12.2 Prostatic cancer 371 9.12.2.1 Oestrogen therapy 371 9.12.2.2 LHRH analogues 371 9.12.2.3 Anti-androgens 372 9.13 ENZYMES 374 9.14 BIOLOGICALS 375 9.14.1 Interferon alpha 375 9.14.2 Tumour necrosis factor 376 9.14.3 Interleukin 376 9.14.4 Growth factors 377 9.15 PRODRUGS AND DRUG TARGETING APPROACHES

377

9.15.1 Bioreductive prodrugs 377

The chemotherapy of cancer

407

9.15.2 Estramustine 378 9.15.3 Photoactivated prodrugs (Photodynamic Therapy, PDT)

378

9.15.4 Antibody—drug conjugates 379 9.15.5 ADEPT 379 9.15.6 GDEPT 380 9.16 NEW RESEARCH TOOLS 380 9.17 FUTURE POSSIBILITIES 381 9.17.1 Gene targeting 381 9.17.1.1 Antigene (Macromolecules and small molecules)

381

9.17.1.2 Antisense oligonucleotides 382 9.17.1.3 Ribozymes 382 9.17.2 Oncogene—product inhibitors 382 9.17.3 Gene therapy 382 9.17.4 Growth factor and signalling pathway modulators

383

9.17.5 Resistance inhibitors 383 9.17.6 DNA—Repair inhibitors 384

Introduction to the principles of drug design and action

408

9.17.7 Telomerase inhibitors 384 9.17.8 Antimetastic agents 384 9.17.9 Blood-flow modifying agents 384 9.17.10 Vaccines 385 9.17.11 Chemopreventative agents (“Neutriceuticals”) 385 9.18 ANTI-EMETICS 385 FURTHER READING 386

9.1 INTRODUCTION Cancer is a disease in which the control of growth is lost in one or more cells leading to a solid mass of cells known as a tumour. A growing (primary) tumour will often become life-threatening by obstructing vessels and/or organs. However, death is most often caused by spread of the primary tumour to numerous other sites in the body which makes surgical intervention impossible. Other types of cancers such as leukaemia involve a build-up of large numbers of cells in the blood. In the first three decades of this century cancer accounted for less than 10% of all UK deaths, infectious diseases being the main cause of mortality. Whilst dramatic progress has been made in controlling infections, similar progress has not been made with the treatment of cancer. Improved diet, living conditions and health care have increased the average life-span to the point where cancer, which is a disease of advanced years (70% of new cases of cancer in the UK occur in those over 60), has become more prevalent. Consequently, about 300,000 new cases of cancer occur each year in the UK. The annual number of deaths from cancer of all types is approximately 160,000 which constitutes approximately 25% of all UK deaths. Statistics show that approximately one in three of the population will suffer from some form of cancer during their lives and one in four will die from the disease. Furthermore, 1 in 10,000 children will be diagnosed annually as suffering from cancer, which means that there are 1300 new cases each year. It is thought that exposure to an ever increasing number of chemicals (carcinogens) in both the environment and the diet may be significant. Occupational factors are thought to account for 6% of cancers, while life-style and diet may account for up to 30%. Genetic predisposition is also a factor in some types of the disease.

The chemotherapy of cancer

409

9.2 TERMINOLOGY A tumour or neoplasm is an abnormal tissue mass or growth which results from neoplasia, a state in which the control mechanisms governing cell growth are deficient, leading to cell proliferation. Cancers are generally named according to the type of tissue in which they arise; for example, sarcoma describes those neoplasms occurring in mesodermal tissue which includes connective tissue, bone and muscle. Osteosarcoma refers to bone cancer, and tumours of the epithelial tissues such as the mucous membranes and glands (including cancers of the breast, ovary and lung) are referred to as carcinomas. Cancers of the blood or haemopoietic tissue are generally known as blastomas. These can involve lymphoid, erythroid or myeloid cells which generally fall into the sarcoma category. Leukaemias describe those cancers which originate in leucocytes, and may be myeloid, lymphatic or monocytic; in addition, these particular cancer types may be chronic or acute. Bone marrow cell tumours are referred to as myelomas, and in multiple myeloma (the most common bone marrow cancer) a clone of plasma cells is involved. Neoplasia of erythroid stem cells is known as primary polycythaemia. The reticuloendothelial system is also susceptible to tumourogenesis. Lymphosarcoma is cancer of the lymphoid cells, whereas Hodgkin’s disease is an example of a lymph adenoma which, although it mainly affects reticulum cells, can extend to eosinophils, fibroblasts and lymphocytes. 9.3 METASTASES Metastasis is the ability of solid tumours to spread to new sites and establish secondary cancerous growths. Tumour cells may easily penetrate the walls of lymphatic vessels and distribute to draining lymph nodes. Cancer cells may also invade blood vessels directly, since capillaries have weak thin walls which offer little resistance. A tumour may also spread across body cavities from one organ to another; e.g. stomach to ovary. Most patients who die of cancer do so as a consequence of metastasis to vital organs. At the point of clinical recognition of cancer, curative surgical or radiological treatment is only possible if metastasis of the primary tumour has not occurred. Therefore, early diagnosis is essential. Since about 50% of malignant tumours have metastasised prior to diagnosis, the condition is often beyond the reach of curative surgery or radiotherapy alone. It is in these cases that systemic chemotherapy can often help to reduce the total tumour mass. 9.4 CELL GROWTH CYCLE Cancer is a disease of cells characterised by a reduction or loss of effectiveness in the normal cellular control and maturation mechanisms that regulate multiplication. A schematic diagram of the cell cycle is shown below:

Introduction to the principles of drug design and action

410

There are four main criteria common to all cancers. 9.4.1 Excessive cell proliferation This usually results in the formation of a tumour. Normal adult tissues do not grow but maintain a steady number of cells. In some tissues, e.g. liver, this is achieved without proliferation because there is little cell loss. In the bone marrow, however, a steady number of cells are maintained by a fast rate of cell division balanced by a fast rate of cell loss. Often it requires only a slow increase in the rate of proliferation to gradually outgrow normal controlled cellular populations. 9.4.2 Loss of tissue-specific characteristics In the early stages of tumour growth cancer cells often resemble the original cells from which they are derived. Later, tumour cells lose the appearance and function of these tissues. 9.4.3 Invasiveness This is the ability to grow into adjacent tissue. The tumour not only expands in size but also infiltrates surrounding tissue. When nerve-endings are affected pain is experienced. 9.4.4 Metastasis One of the major obstacles to successful cancer treatment is the ability of cancer cells to move around the body (metastasise) and develop into new tumours elsewhere.

The chemotherapy of cancer

411

9.5 MECHANISMS OF TUMOUR FORMATION It is now accepted that cancer is a “genetic” disease resulting from changes to the sequence information (and its expression) in specific genes. A number of ways in which these changes can be brought about by either internal, external or hereditary factors are described below. 9.5.1 Internal factors Tumour formation may result from changes to the genetic information brought about by malfunction of the normal DNA processing systems within the cell. 9.5.1.1 Mutation Genetic mutations can take several forms. In a “point” mutation, only one base is altered, and the new codon that results will cause insertion of a different amino acid into that particular position of the protein. Should the protein be a growth factor, then tumourogenesis could result. In a “translocation” mutation, an entire DNA sequence is moved from one part of a gene or chromosome to another. Again, the loss of the proteins corresponding to the two original DNA sequences or the presence of the new protein may lead to tumourogenesis. The original genes are known as “protooncogenes”; that is, genes which will not cause cancer unless suitably activated (i.e. by translocation to form an “oncogene”). The proto-oncogene/oncogene theory has been proved in the case of Burkitt’s Lymphoma and Chronic Myelogenic Leukaemia (CML) in which the precise sequences involved in the translocations have been identified. 9.5.1.2 Addition or loss of genetic material During normal DNA handling processes such as repair, DNA bases may be inadvertently added or deleted. This can have a similar effect to point mutations in altering codons and ultimately modifying the structure of growth factors and control proteins. 9.5.1.3 Changed gene expression A problem with gene expression may occur, such as uncontrolled expression or amplification. Should growth factors or proteins that are responsible for receptor formation be involved, then tumourogenesis may result.

Introduction to the principles of drug design and action

412

9.5.2 External factors 9.5.2.1 Viruses A link between viruses and cancer has been recognised since 1911, when Peyton Rous demonstrated that avian spindle cell carcinoma could be transmitted from one bird to another by a cell-free filtrate containing the virus (which now bears Rous’ name—the Rous Sarcoma Virus). Since then, other viruses have been identified which are linked to human cancers. Viruses may be either RNA retroviruses such as Human-T-Cell Leukaemia Virus (HTLV-1) or DNA viruses. It is believed that RNA viruses contain DNA-polymerases which facilitate the production of double-stranded viral DNA. On being incorporated into the host DNA, the viral DNA may cause transformation to a cancerous state via a number of different mechanisms including production of an oncogene from an existing gene, damage to a tumour suppresser gene or insertion of a completely new gene. For example, HTLV-1 introduces a gene known as tax which results in the overexpression of interleukin-2. This can lead to adult T-cell lymphomas and leukaemias with an increase in the number of activated lymphocytes, although these may take years to develop in susceptible individuals. HTLV-1 is endemic in South East Asia and the Caribbean; in the Far East it is also associated with nasopharyngeal cancers. Epstein-Barr Virus (EBV) infects 90% of the world’s population and is considered to be an “initiator” of cancer as opposed to a specific cause. For example, 90% of Burkitt’s lymphoma cells test positive for EBV which allows infected lymphocytes to become immortal, leading to a potentially cancerous state. Burkitt’s lymphoma is endemic in those parts of Africa with chronic malaria suggesting that the latter may be a co-factor in lymphoma development. Hepatocellular carcinoma has been linked with Hepatitis B Virus (HBV), and is endemic in Southeast Asia and tropical Africa. The risk of tumour formation is greatest in those who are infected from an early age, and males are four times more likely to develop the cancer than females. It is believed that the X-gene in HBV codes for proteins which promote transcription. There are fifty different types of Human Papilloma Virus (HPV), and HPVs 16 and 18 have been linked to cervical cancer. The virus produces several proteins, some of which enhance mitosis while others interfere with P53 (a tumour suppresser gene) or modify the interaction between cellular proteins and transcription factors. 9.5.2.2 Chemicals There is now convincing evidence that certain chemicals in the environment, encountered through the diet, lifestyle or occupation may be responsible for some cancers. The link between cigarette smoke and lung cancer is now well-established, and it is also known that carcinogenic polycyclic aromatic hydocarbons (PAHs) are formed in overcooked fried or barbecued meat. Carcinogenic amines are formed in the stomach as a result of the bacterial degradation of nitrites used as preservatives in meat and fish, and the potent carcinogenic aflatoxins are found in peanut butter as a

The chemotherapy of cancer

413

result of the fungal infection of peanuts during growth. Occupation-associated cancer is not a feature of the nineteenth century alone as, in 1775, Sir Percival Pott noted the high frequency of scrotal skin cancer amongst young chimney sweeps. Infrequent washing meant that the tarry deposits produced by burning coal were in contact with the skin for long periods of time. More recently, vinyl chloride used by workers in the plastics industry has been associated with angiosarcoma of the liver. Furthermore, employees in the furniture industry have been prone to develop naso-pharyngeal malignancies brought about by the inhalation of particulate matter carrying organic compounds during leather and wood polishing processes. Most of these organic carcinogens are thought to work by covalently modifying DNA (either before or after metabolism). In addition to these organic carcinogens, certain dusts and minerals are known to cause cancer; for example, the link between asbestos and pleural and peritoneal tumours is well established. 9.5.2.3 Radiation Malignancies have been linked with exposure to α and β particles or X-rays which are known to damage DNA by fragmentation through the formation of free radicals. A link between nuclear fall-out from atomic bombs and cancer was firmly established after the Hiroshima bombing in World War II. It has also been postulated that children living close to nuclear power stations are at a higher risk of contracting leukaemias and brain tumours. The danger of the escape of radioactive materials from nuclear reactors was highlighted by the Chernobyl incident in Russia which caused widespread contamination of the food chain. It is also known that a build-up of radon gas produced by certain types of granite can endanger the occupants of houses built from this material. Radon is a naturally occurring radioactive gas that, once inhaled, enters the bloodstream and delivers radiation to all tissues; bone marrow is particularly sensitive and so leukaemias predominate. In the UK, local councils have been obliged to offer grants to affected householders so that buildings can be structurally modified to improve ventilation. Very recently it has been proposed that electromagnetic radiation from overhead power lines may be associated with childhood leukaemias, and that microwaves produced by mobile telephones held to the ear may cause brain tumours. In the former case, it has been suggested that rather than the electromagnetic radiation itself causing cancer, the high electric fields generated by power lines may concentrate radioactive radon gas into local pockets. Recently, the UK Government has been sufficiently concerned about these suggestions to fund a detailed investigation. 9.5.3 Hereditary factors A number of genes have now been identified that, if inherited, can predispose individuals to certain types of cancer. For example, two genes, BRCA1 and BRCA2, have recently been identified and sequenced by UK and US researchers. These genes are inherited and are associated with breast cancer. Other genes associated with colon and bowel tumours are known to be inherited. This has lead to the introduction of diagnostic screening with subsequent genetic counselling for affected individuals.

Introduction to the principles of drug design and action

414

9.6 TREATMENT Cancer treatment often encompasses more than one approach, and the strategy adopted is largely dependent on the nature of the cancer and how far advanced it is. 9.6.1 Surgery The surgical removal of tumours is confined to those considered to be solid (for example breast, lung or colon tumours) as opposed to the leukaemias. If a tumour is small and reasonably well defined it can usually be removed by surgery. However, there is often additional treatment with chemotherapy or radiotherapy to try and eliminate any cells that may have either remained behind or metastasised. Increasingly, radiotherapy or chemotherapy may be administered prior to surgery, in order to shrink the tumour and facilitate its removal. 9.6.2 Radiation therapy (Radiotherapy) Radiation therapy utilises X-rays or radiopharmaceuticals (radionuclides) which act as sources of γ-rays. X-Rays are delivered locally in a highly focused beam to avoid damage to healthy tissue, and there is still intensive research into the most effective treatment regimes in terms of the duration and frequency of exposure. The main radionuclides in use include cobalt-60, gold-198 and iodine-131. Gold-198 concentrates in the liver, and iodine-131 is used to treat thyroid cancers as iodine accumulates in this gland. A significant proportion of tumour cells are hypoxic (lack oxygen) and are thus less sensitive to damage by irradiation. Therefore, prior to and during radiation therapy, oxygen is often given to try and sensitise the cells. Radiosensitising drugs such as metronidazole have also been administered prior to treatment to try and improve the therapeutic outcome of radiation therapy. A process known as high linear energy transfer (HLET) has also been applied to the irradiation of hypoxic cells. In this procedure, the tumour is irradiated with neutrons (heavier than X or γ-rays) which decay to a-particles, the latter causing cell damage in an oxygenindependent manner. 9.6.3 Photodynamic therapy Photodynamic therapy (PDT) is a relatively new form of cancer treatment that is currently being evaluated in clinical trials. PDT involves the sytemic administration of a photosensitiser such as the haematoporphyrin derivative Photophrin® which is found to be selectively retained by malignant cells. A few days after administration of this agent, the tumour is irradiated with an intense light source of an appropriate wavelength which excites the photosensitiser. Upon decay to its ground state, available oxygen is transformed into singlet oxygen which is highly cytotoxic and damages the tumour. There is also evidence that, by damaging endothelial cells, PDT restricts tumour blood flow. Laser light sources are now well-developed, and the use of flexible optical fibres means that tumours in inaccessible parts of the body including the GI tract can be easily reached through key-hole surgery techniques and

The chemotherapy of cancer

415

endoscopy; ovarian and stomach tumours have been treated experimentally in this way. Other less expensive non-laser light sources have been recently marketed and new types of photosensitisers are being developed. The use of PDT is therefore likely to escalate in the future. 9.6.4 Immunotherapy and vaccines The aim of immunotherapy is to stimulate the body’s natural response to fighting the cancer. Several neoplasms, including some types of breast cancer, have been found to possess specific tumour antigens, and this has led to the development of monoclonal antibodies specific for some tumour types. Although little success has so far been achieved by treating patients with antibodies alone, research is still ongoing into the development of vaccines that may either prevent tumour formation or modify the growth of established tumours. In the latter case, there has been recent publicity over the use of a vaccine to prolong the life of melanoma patients. Tumour-specific antibodies have also been used for drug targeting by attaching them to either drugs or enzymes (e.g. ADEPT), and these strategies are discussed later. 9.6.5 Chemotherapy 9.6.5.1 Achievements of chemotherapy Despite the limitations discussed below, cancer chemotherapy has made remarkable progress. Nitrogen mustards were the first agents to be clinically used, although their predecessors were initially used in warfare. Cisplatin, which was also discovered serendipitously, has provided a major advance in the treatment of testicular and ovarian carcinomas. In the latter case, the cure rate is very high for patients with early diagnosis. Taxol is a more-recently discovered natural product that is showing great promise in the chemotherapy of lung and refractory ovarian cancers. Similarly, the synthetic agent temozolomide is proving to be highly effective for melanomas and brain tumours (particularly those occurring in children). 9.6.5.2 Discovery of drugs and their preclinical evaluation Most clinically used anticancer drugs were discovered either through chance (e.g. cisplatin and the nitrogen mustards) or through screening programmes (e.g. vinblastine and taxol). Only recently, since a more detailed knowledge has been acquired of the fundamental biochemical differences between normal and tumour cells, has rational drug design become possible (e.g. Marimastat®). A combination of the power of screening techniques and rational drug design has recently been realised with the widespread trend towards the use of combinatorial chemistry to rapidly generate large numbers of molecules of diverse structure for screening. New agents are nearly always evaluated initially in in vitro tumour cell lines. Unfortunately, this only measures the cytotoxicity of an agent and provides no indication of whether it is likely to have useful in vivo antitumour activity. It can, however, indicate whether the agent has selective cytotoxicity towards a particular

Introduction to the principles of drug design and action

416

tumour type which may suggest which in vivo experiments to carry out. Human tumour xenografts represent the most successful animal model in which human tumour fragments are transplanted into rodents. However, even these models do not necessarily correspond to equivalent tumours in humans, as numerous differences exist including the integrity of the blood supply to the transplanted tumour and general biochemical species differences. For example, a toxicity not detected in rodents may be serious enough in man to prevent clinical use of an agent. Such toxicities are often a consequence of interspecies variations in metabolism. Despite these problems, most drugs in clinical use today (with the exception of hormonal agents) were introduced as a result of activity demonstrated in animal models. It is worth noting that, sometimes, drugs that are active in humans show no effect in animal models. For example, hexamethylmelamine [2,4,6-tris(dimethylamino)-1,3,5-triazine] is the lead compound for the “melamine” class of antitumour agents and is metabolised to a carbinolamine species in man with activity against carcinomas of the bronchus, ovary and breast. However, it exhibits only minimal activity when tested in rodent models, as this species fails to carry out the crucial metabolic step. 9.6.5.3 Accessibility of tumour cells to drugs The accessibility of anticancer drugs to tumour cells varies. Whilst leukaemia cells are fully exposed to drugs in the blood stream, solid tumours have a less-reliable blood supply. Small tumours are usually reasonably well-supplied and are more susceptible to drug action than large tumours which frequently have poor capillary access, particularly in their centres which can be hypoxic. The degree of accessibility of the chemotherapeutic agent therefore explains the greater sensitivity of small primary tumours and early metastases to chemotherapy and highlights the importance of early diagnosis and treatment. It is noteworthy that brain tumours are particularly resistant to chemotherapy as few drugs are capable of crossing the blood brain barrier. 9.6.5.4 Achieving selective toxicity The development of more-effective chemotherapeutic agents is critically dependent upon the discovery of exploitable biochemical differences between normal and tumour cells. Such differences should allow a more rational approach to drug design rather than relying on the empirical manner in which many of the present day drugs have evolved. Examples of such a fundamental difference are presently limited, but perhaps the best known is the discovery that some lymphoid malignancies are dependent on an exogenous supply of asparagine whereas healthy cells can synthesise their own; this led to the clinically-useful agent asparaginase (see Section 9.13). A more recent example is the development of agents such as Marimastat® that target metalloproteinase enzymes crucial for the process of metastasis. There is hope that complete selectivity may one day be achieved through gene targeting in which genetic differences between tumour cells and healthy cells are exploited (see Section 9.17.1). Indeed, there is speculation that some of the DNA-binding agents such as the mustards that bond covalently to GC sequences of DNA may exploit the fact that some oncogenes are particularly GC-rich. However, a more common view at present is that

The chemotherapy of cancer

417

most clinically-useful anticancer drugs (including the mustards) are generally cytotoxic but are more damaging to faster growing cells. Although it is true that certain types of cancer cells grow faster than other tissues in the body, in the majority of tumours the rate of cell division is still slower than that of normal bone-marrow, skin epithelium or the mucosa of the mouth. This explains the consistent pattern of side effects accompanying chemotherapy which are dose-limiting in practice. 9.6.5.5 Limiting the toxicity of chemotherapeutic agents It has proved possible to limit bone marrow toxicity by exploiting cell “kinetic” differences between normal and tumour stem cells. Stem cells constitute the smallest, yet most important, compartment in a proliferating system. They are capable of an indefinite number of divisions and are responsible for maintaining the integrity and survival of a cell population. Thus, the increase in size of a lymphoma cell population compared with a normal stem cell population can be explained by the fact that only 20% of the bone marrow stem cells are usually in active cycle at any one time, the remaining 80% being in the resting phase (G0). The dividing marrow stem cell population may be significantly reduced in size by chemotherapy, however within 3–4 days the remaining stem cells can move from G0 into active cycle. This means that very high doses of drugs may be given for 24–36 hour periods interspersed with adequate recovery intervals. Studies of the cell kinetic patterns of tumour growth have suggested a classification for cytotoxic agents based on their ability to reduce the stem cell population of normal bone-marrow and lymphoma cells in mice. Two classes of antitumour agents are distinguished:

Class Cell-cycle-specific agents which kill cells in only one phase of the cycle; e.g. S-Phase (the period of DNA synthesis): 61: mercaptopurine, cytosine arabinoside and methotrexate; or M phase (mitosis): vinblastine and vincristine. An increased dose of these drugs will not kill more bone marrow stem cells than killed by the initial dose. Class Non-cell-cycle-specific agents that kill cells at all phases of the cell cycle; e.g. cyclophosphamide, melphalan, chlorambucil, cisplatin, 2: BCNU, CCNU, 5-fluorouracil, actinomycin D and daunorubicin. An increased dose of these drugs will increase the number of bone marrow stem cells killed. Although cell killing can occur in all phases, it is possible that some agents are more active in a given phase of the cycle. With a few exceptions, such as the cumulative toxicities associated with adriamycin (cardiac), bleomycin (pulmonary) and cisplatin (renal), common toxicities of anticancer drugs are usually reversible within 2–3 weeks. Mucositis (associated with actinomycin D, adriamycin, bleomycin, methotrexate, 5-fluorouracil and daunorubicin) is reversible over a period of 5–10 days, reflecting the rapid recovery of normal tissues. Nausea and vomiting, which accompany the administration of certain

Introduction to the principles of drug design and action

418

drugs, may be partially overcome with modern anti-emetic agents such as the 5HT3 antagonists (see section 9.18). 9.6.5.6 Overview of the mode of action of chemotherapeutic agents Different mechanisms of action include interference with metabolism (e.g. methotrexate), interruption of cell division (e.g. the effect of vinblastine and taxol on the spindle), steroid receptor antagonism (e.g. effect of tamoxifen on the oestrogen receptor) or the inhibition of steroid biosynthesis (e.g. aromatase inhibitors). The recently developed Marimastat® works by targeting metalloproteinease enzymes involved with metastasis. A large number of clinically-useful antitumour agents have DNA as their target. Some agents block the synthesis of DNA (e.g. 6-mercaptopurine), whilst others act by becoming incorporated into DNA and then interfering with its function (e.g. 6-thioguanine). However, the majority of clinically-useful anticancer agents interact directly with DNA. As the primary genetic material, DNA has many desirable characteristics as a drug target. It is an active participant in a wide range of biological processes and disruption is likely to have a detrimental effect on cell growth. Furthermore, drugs targeted to nucleic acids act at the earliest possible stage of gene expression and should be highly efficient on a molar basis. DNA-interactive drugs exert their effect by a number of different mechanisms. DNA cross-linking may occur on the same strand (intrastrand cross-linking) or between DNA strands (interstrand cross-linking). Interstrand cross-linking is observed with alkylating agents such as the nitrogen mustards; cisplatin is an example of an agent that causes intrastrand cross-links. An important feature of intrastrand crosslinking is that it bends the DNA at the adduct site. Intercalation is a type of noncovalent interaction in which planar molecules insert between the stacked base-pairs of DNA (e.g. mitoxantrone). The intercalating moiety is inserted perpendicular to the DNA helix, causing the bases to separate vertically. This lengthens the DNA and distorts the sugar-phosphate backbone. Unwinding of the helix at the intercalation site disrupts the action of enzymes such as RNA polymerases and the topoisomerases which control DNA supercoiling. A further mechanism involves groove-binding in which the agent interacts either covalently or non-covalently in the minor or major grooves. The two grooves of DNA differ in electrostatic potential, hydration and hydrogen bonding characteristics, as well as in width and depth. Most DNA-binding control proteins are thought to interact in the major groove as it is more information-rich than the minor groove. Experimental agents such as the lexitropsins and other netropsin/distamycin analogues bind non-covalently in the minor groove, whereas analogues of CC-1065 and the pyrrolobenzodiazepines bind covalently. The minor-groove binding distamycinmustard conjugate, tallimustine, is presently in clinical trials. Most of the mustards and the recently marketed temozolomide (see Section 9.8.1.5) act in the major groove. Finally, some agents can interact with DNA (by one of the above mechanisms) and then cause strand scission (strand breakage) at the binding site usually through the production of free radicals (e.g. bleomycin and doxorubicin). As in the case of DNA cleavage by ionising radiation, the reaction is usually oxygen dependent and the sugars or bases of the sugar-phosphate backbone are attacked. If the breaks occur in close

The chemotherapy of cancer

419

proximity but on opposite strands then double-strand breakage occurs. Alkylation of DNA bases or of the phosphate backbone can also result in strand scission, but this is oxygen independent. 9.6.5.7 Drug resistance The development of drug resistance may seriously interfere with treatment. Initial selective cytotoxicity towards a tumour can sometimes be followed by a rapid recovery in tumour growth, and this resistance may increase after each administration. Drug resistance has been observed in a number of drug-sensitive tumour types including breast, choriocarcinoma and lymphoblastic leukaemias. It has been demonstrated that in some resistant tumour cells there is an altered expression of particular proteins; for example, glutathione transferase production may increase. This enzyme is believed to be responsible for resistance to alkylating agents by catalysing their covalent interaction with glutathione rather than DNA. Another mechanism of resistance involves the active transport of anticancer drugs out of cells. The discovery of the multidrug resistance gene (MDR) has led to an understanding of how tumours can become resistant to a wide variety of agents, and future therapies may even include strategies to down-regulate the MDR gene as a means to enhance the effectiveness of existing anticancer drugs. This problem of resistance has been overcome in several tumour types by using a combination of different cytotoxic agents as described below. 9.6.5.8 Combination chemotherapy Most attempts at treating tumours with single agents have been disappointing. A single drug kills the population of cells that is most sensitive to it and leaves a resistant fraction unharmed and still dividing. This led, in 1960, to the first use of a combination of drugs for treating testicular tumours. The “cocktail” principle was then rapidly extended to other tumour types. Each drug included in a particular combination should be active as a single agent but have different toxic (dose-limiting) side-effects from the others. Multiple drug therapy also enables the simultaneous attack of different biological targets thus enhancing the effectiveness of the treatment. The successful application of this technique to the treatment of acute lymphoblastic leukaemia is illustrated below with the comparative response rates of single agents and various combinations:

Drugs Methotrexate (M, see Section 9.7.1.1) Mercaptopurine (MP, see Section 9.7.2) Prednisone (P) Vincristine (V, see Section 9.9.1) Daunorubicin (D, see Section 9.8.2.1) P, V P, V, M, MP

% Complete Remission 22 27 63 57 38 94 94

Introduction to the principles of drug design and action

P, V, D

420

100

A variety of combination schedules using different drugs and dosage regimes are now available and, in many cases, are accepted as superior to single drug therapy. Furthermore, many cancers may already be disseminated at the time of clinical detection, and so combination chemotherapy may be commenced concurrently with local treatment (e.g. radiotherapy, surgery) to maximise benefit for the patient. The micrometastases associated with the primary tumour are often very sensitive to chemotherapy since they have a good blood supply that facilitates drug access. They are also less likely to develop drug resistance than an older tumour. 9.6.5.9 Adjuvant chemotherapy In treating cancer, it is sometimes necessary to co-administer other agents that may enhance the activity of the anticancer drug or counteract any side-effects produced. Common side-effects include nausea, vomiting and other gastrointestinal disturbances, hair loss and myelosuppression. To counteract the gastrointestinal problems, antiemetics will usually be administered (see Section 9.18). Myelosuppression is more problematic in that it can lead to an increased risk of infection, and so antibiotic and/or antifungal therapy may be required. 9.7 ANTIMETABOLITES Antimetabolites function by blocking crucial metabolic pathways essential to cell growth. Selectivity is based on the concept that some cancer cells can be faster growing than many normal cell populations with the exception of cells such as those in the bone marrow or parts of the GI tract. While this may be true in the case of leukaemias, older solid tumours often have a very small fraction of cells in active growth. All antimetabolite agents in current clinical use, including the antifolates and the purine and pyrimidine antimetabolites, interfere with DNA synthesis. 9.7.1 DHFR inhibitors (Antifolates) 9.7.1.1 Methotrexate (Maxtrex®) Tetrahydrofolic acid is produced by the action of the enzyme dihydrofolate reductase (DHFR) on dihydrofolic acid (9.1), and is required for the synthesis of thymine which becomes incorporated into DNA. Slight modification of the structure of folic acid produced the lead antimetabolite aminopterin; methotrexate (9.2), which was shown to be more selective, followed in the 1950s. Methotrexate binds more strongly to the active site on DHFR than the natural substrate by a factor of 104 due to the presence of an amino rather than a hydroxyl moiety which increases the basic strength of the pyrimidine ring.

The chemotherapy of cancer

421

The most basic centre in the methotrexate molecule is at the N1 and adjacent C2NH2 position as confirmed by 13C NMR measurements at C2. Examination of the drug-enzyme complex by X-ray diffraction has shown that the pyrimidine ring is situated in a lipophilic cavity with the cation of N1/C2-NH2 binding to an aspartate-26 anion of the enzyme. Other binding points revealed by X-ray include hydrogen bonding between C4-NH2 and the carbonyl groups of both Leu-4 and Ala-97, and ionic interactions between the α-COOH of the glutamate residue and the basic side chain of Arg-57. The p-aminobenzoyl

residue lies in a pocket formed on one side by the lipophilic side chains of Leu-27 and Phe-30, and, on the other side, by Phe-49, Pro-50 and Leu-54. A neighbouring pocket lined by Leu-4, Ala-6, Leu-27, Phe-30 and Ala-97 accommodates the pteridine ring. The nicotinamide (NADPH) portion of the fully extended co-enzyme lies sufficiently close to the pteridine ring to facilitate transfer of a hydride anion from the pyridine nucleus to the C6-position. These studies reveal that methotrexate occupies the reverse position at the active site of the enzyme compared to the substrate. Although the p-aminobenzoate and glutamate portions of both are identically bound, with dihydrofolate N-1 is unbound, C2-NH2 and C4-OH bind only to water molecules, N3 is hydrogen bonded to Asp-26, N5 is unbound and N8 interacts with Leu-4 via van der Waals forces. This results in the substrate being comparatively loosely bound, a surprising consequence of the differences in position and strength of the most basic centres in the substrate and inhibitor molecules. In lymphoblastic leukaemia, methotrexate produces a remarkable remission of symptoms for about a year, however the cells eventually develop resistance by increasing the production of DHFR. Choriocarcinoma, a fast growing tumour of pregnancy with a previously high death rate, is rapidly and completely cured with methotrexate. Dramatic cures have also been obtained in Burkitt’s lymphoma which is a highly malignant carcinoma of the lymph glands. Disadvantages of the use of methotrexate include its adverse effect on the production of red blood cells which leads to macrocytic anaemia. Gastrointestinal ulceration and potential damage to the

Introduction to the principles of drug design and action

422

kidneys and liver also require careful monitoring. In high dose intermittent schedules the adverse effect on bone marrow can be relieved by the periodic administration of the calcium salt of N5-formyltetrahydrofolic acid (9.3, folinic acid, leucovorin) which enables blockade of tetrahydrofolic acid production to be bypassed (folinic acid “rescue” therapy). Many derivatives of methotrexate have been synthesised with a view to reducing its toxicity. The phenyl substituted 3’,5’-dichloro derivative is significantly less toxic, perhaps due to its ease of metabolism to 3’,5’-dichloro-7-hydroxymethotrexate which is equally effective at inhibiting DHFR. Analogues containing a fluorine atom have also been

synthesised so that their interaction with the DHFR enzyme may be studied by FNMR both in vitro and, more recently, in vivo. 9.7.2 Purine antimetabolites Purine analogues inhibit a later stage in DNA synthesis than DHFR inhibitors. Their major problem is a lack of selective toxicity, since purines are involved in many cellular processes apart from nucleic acid synthesis. 6-Mercaptopurine (9.4, 6-MP) is well-established in the treatment of childhood leukaemias, especially chronic myelocytic leukaemia where the remission rate is about 50%. The free base form is converted by sensitive tumour cells to the ribonucleotide 6-mercatopurin-9-yl (MPRP) which results from interaction of the base with 5phosphoribosyl transferase. Resistance to 6-MP usually arises due to loss of this enzyme within the tumour. Although MPRP inhibits several enzymatic pathways in the biosynthesis of purine nucleotides including the conversion of inosine-5’-phosphate to adenosine-5’phosphate, the main inhibitory action appears to occur at an earlier stage when 5’phosophoribosylpyrophosphate is converted into phosphoribosylamine by phosphoribosylpyrophosphate amido-transferase. Allopurinol may be used as an adjuvant therapy to inhibit xanthine oxidase mediated degradation of 6-MP to thiouric acid which may cause renal damage. Another cytotoxic drug used for treating myeloblastic leukaemia, 6-thioguanine (9.5), is also metabolised to the 9-(1’-ribosyl-5’-phosphate) by tumour cells. However, in contrast to MPRP, this does not inhibit an enzyme but is further phosphorylated to the triphosphate and then incorporated into DNA as a “false” nucleic acid. The lack of

The chemotherapy of cancer

423

selectivity towards tumour cells is due to rapid incorporation of 6-thioguanine into the DNA of bone marrow cells. Another analogue, 8-azaguanine exerts its anticancer action in a similar way.

9.7.3 Pyrimidine antimetabolites Cytarabine (9.6, ARA-C) and 5-fluorouracil (9.7, 5-FU) are two of the best known pyrimidine antimetabolites. Cytarabine is one of the most effective single agents available for treating myeloblastic leukaemia, achieving a remission rate of 25%. When combined with 6-thioguanine or the cytotoxic antibiotic daunorubicin, remission rates are increased to 50%. Other combinations that include vincristine have led to claims of 70% remission, and some long term survivals of myeloblastic leukaemia have been reported. A disadvantage of cytarabine therapy arises from its rapid hepatic deamination by cytosine deaminase to give an inactive uracil derivative. This short half-life is counteracted by continuous infusion methods of administration. The rapid deamination has led to the quest for pyrimidine nucleoside deaminase inhibitors which might be co-administered, although this approach has not yet met with success. Deaminase-resistant O-acyl derivatives have also been prepared which can be hydrolysed by esterases. The agent fluoro-2’,5’-anhydrocytosine arabinoside is a novel cyclonucleotide that is hydrolysed non-enzymatically and has significant antitumour activity against stomach and pancreatic adenocarcinomas. Of the many halogeno-pyrimidines investigated, only fluoro derivatives have any appreciable antitumour activity. 5-Fluorouracil (5-FU) is highly effective as a 5% cream in treating skin cancer, and has also found clinical use in the palliative treatment of certain solid tumours such as those of the GI tract, breast, and pancreas. 5-FU is initially metabolised to the 2’-deoxyribonucleotide, 5-fluoro-2’-deoxyuridylic acid (FUdRP), which is a potent inhibitor of thymidylate synthetase. The latter causes the transfer of a methyl group from the co-enzyme methylenetetrahydrofolic acid to deoxyuridylic acid which is converted to thymidylic acid and incorporated into DNA. 5-FU has been shown to have an affinity for thymidylate synthetase several thousand times greater than that of the natural substrate. This remarkable effect is associated with the unique properties of the fluorine atom whose van der Waals radius compares favourably with that of hydrogen although the bond strength is considerably greater. Additionally, the high electronegativity of fluorine affects the electron distribution, conferring a lower pKa on the molecule compared to uracil. These two features combine to enable FUdRP to fit the active site of the enzyme extremely well although the fluorine cannot be removed thus effectively inhibiting the enzyme. Further studies

Introduction to the principles of drug design and action

424

have suggested that a nucleophilic sulphydryl group at the active site forms a covalent bond to FUdRP leading to a ‘dead end’ adduct of the enzyme, co-enzyme and 5-FU. Structure-activity studies

have shown that the increased size but lower electronegativity of other halogen atoms reduce activity. It has been postulated that the high selectivity of 5-FU, especially in skin treatments, may reflect the fact that certain types of cancer cells lack the relevant enzymes to degrade it. 9.8 DNA INTERACTIVE COMPOUNDS A large number of clinically-useful anticancer drugs exert their effect by interacting with DNA. Some agents intercalate between the base pairs of DNA, whereas others alkylate in either the minor or major grooves. Other types cross-link the DNA strands together in either an “intra” or “interstrand” fashion. As with other classes of anticancer drugs, the selective toxicity towards cancer cells may arise solely from the difference in growth rate of populations of cancer cells compared to normal cells, which also explains their toxicity toward bone marrow and cells of the GI tract. However, there is speculation that some DNA-interactive agents may selectively target certain DNA sequences (for example GC-rich regions) which may inhibit the growth of cancer cells. The agents below are categorised according to their mechanism of action. 9.8.1 Cross-linking agents 9.8.1.1 Nitrogen mustards The mustards were developed further as chemical warfare agents during World War II as an advance over sulphur mustard gas [S(CH2CH2Cl)2] that made its debut in World War I. Clinicians observed that the leukocyte count dropped in victims surviving sulphur mustard exposure. However, whilst some leukaemias proved responsive, any benefit was greatly outweighed by the toxic effect of the compound. The bioisosteric

The chemotherapy of cancer

425

nitrogen mustards, which had an improved therapeutic index, were introduced into the clinic by Goodman in 1946. The first aliphatic example was mechlorethamine hydrochloride (9.8, mustine) which was found to effectively depress the white blood cell count and was used for treating certain leukaemias.

The high chemical reactivity of the nitrogen mustards and their vulnerability to attack by a wide range of nucleophilic centres accounts for the observed toxicities. Under physiological conditions, the aliphatic nitrogen mustards undergo initial internal cyclisation through elimination of chloride to form cyclic aziridinium (ethyleneiminium) ions. This cyclisation involves intramolecular catalysis through neighbouring group participation and results in a positive charge on the nitrogen which is delocalised over the two adjacent carbon atoms. This cation, although relatively stable in aqueous biological fluids, is highly strained and reacts readily with any nucleophile. The clinically-useful antitumour activity is thought to result from attack of the N7-atom of a guanine residue in the major groove of DNA. The process is then repeated with the second β(2)-chloroethyl group and a second guanine N7atom located on the opposite strand. This gives rise to an interstrand cross-link that effectively locks the two strands of DNA together.

Introduction to the principles of drug design and action

426

Mode of action of the aliphatic nitrogen mustards Theoretically, this type of alkylation resembles a SN2 process since the ratecontrolling step is the bimolecular reaction between the cyclic aziridinium ion and the nucleophile which involves simultaneous bond formation and breakage. The preceding step, involving iminium ion formation, is a fast unimolecular process. In practice it is difficult to draw sharp distinctions between the contributions of SN1- and SN2-type mechanisms. The aromatic nitrogen mustards were introduced in the 1950s and are milder alkylating agents. The aromatic ring acts as an electron-sink, withdrawing electrons from the nitrogen atom and thus discouraging aziridinium ion formation. These analogues are sufficiently deactivated that they can reach their DNA target sites before being degraded by reaction with other nucleophiles. This means that the aromatic mustards can be taken orally which is a significant advantage. During early StructureActivity Relationship (SAR) studies it was found that direct attachment of a carboxyl group to the aromatic ring improved solubility but reduced activity due to the additional electron-withdrawing effect. The carboxyl group was thus electronically insulated from the aromatic ring by a number of methylene groups, three proving optimal, giving rise to chlorambucil (9.10) which is one of the slowest acting and least toxic nitrogen mustards, effective in chronic lymphocytic leukaemia, malignant lymphomas and carcinoma of the breast and ovary. Like mustine, it is often administered in conjunction with other drugs such as vinblastine and procarbazine which increase remission rates. It has been proposed that the central nitrogen atom of an aromatic mustard is not sufficiently basic to form a cyclic aziridinium ion since the nitrogen electron pair is delocalised by interaction with the π electrons of the aromatic ring. Alkylation is therefore thought to proceed via an SN1 mechanism, with normal carbonium ion formation (resulting from chloride ion ejection) providing the rate-determining step.

The chemotherapy of cancer

427

Mode of action of aromatic mustards Melphalan (9.9, phenylalanine mustard) was synthesised with a view to introducing a degree of selectivity based on the concept that attachment of an amino acid residue (phenylalanine) to the nitrogen mustard might facilitate selective uptake by tumour cells in which rapid protein synthesis occurs. Since melphalan prepared from Dphenylalanine is much less active than material prepared from the L-form, it has been postulated that melphalan may be conveyed into cells by the L-phenylalanine active transport mechanism. Although it not clear whether this is the case, melphalan is widely used in multiple myeloma, breast and ovarian carcinoma and in the rare condition of macroglobulinaemia. A further attempt to produce more-selective mustards was based on the concept that some tumours are thought to possess high concentrations of phosporamidases. Cyclophosphamide (9.11) is the most successful mustard to result from this work and it has now been in use for over 20 years. The design of this mustard prodrug was based on the concept that the P=O group should decrease the availability of the nitrogen lone pair in an analogous manner to the phenyl ring of the aromatic mustards, thus deactivating the molecule to nucleophilic attack. It was postulated that the P=O group would be removed by phosphoramidases, thus releasing the nitrogen lone pair and restoring the electrophilicity of the molecule. Although shown to be a clinicallyuseful drug, it was later demonstrated that activation is not due to enzyme-catalysed hydrolysis of the P=O group but to oxidation by liver microsomal enzymes. After 4hydroxylation, the molecule fragments

to give the highly electrophilic acrolein and phosphoramide mustard, although the production of normustine [HN(CH2CH2Cl)2] has also been observed.

Introduction to the principles of drug design and action

428

Mechanism of action of cyclophosphamide Cylophosphamide has a wide spectrum of activity ranging from malignant lymphomas and lymphoblastic leukaemia to carcinomas of the bronchus, breast, ovary and various sarcomas. It can cause myelosuppression and haemorrhagic cystitis of the bladder which is thought to result from the excretion of electrophilic acrolein in the urine. This problem has been partly overcome by the co-administration of 2mercaptoethanesulphonate (9.13, MESNA) which acts as a “sacrificial” nucleophile, forming a non-reactive water soluble adduct that is eliminated safely in the urine. The slow rate of in vivo hydroxylation of cyclophosphamide in man has led to the synthesis of a 4-hydroperoxy derivative which spontaneously yields the 4-hydroxy metabolite after administration. Ifosfamide (9.12) is an analogue of cyclophosphamide but is not technically a nitrogen mustard due to the translocation of one chloroethyl moiety to another position within the molecule. Unlike cyclophosphamide, it can only be administered intravenously but it has a similar spectrum of activity. Despite much research and an in-depth understanding of the chemistry of the nitrogen mustards, their precise mechanism of action at the molecular level is still unknown. It is clear that these molecules “staple” the two strands of DNA together via covalent interactions in the major groove, and it can be demonstrated that enzymes such as RNA polymerase are blocked by mustard-DNA adducts. It can also be shown from both electrophoresis-based experiments and molecular modelling that mustard adducts cause distortion of the DNA around the binding site which can be transmitted through a number of base-pairs (the “teleomeric” effect). Therefore, DNA processing may be affected at a point distant from the adduct site. Apart from the “kinetic” explanation for the antitumour activity of the mustards, an alternative view is that their GC-selectivity may be involved. For example, it is known that some of the gene sequences associated with Burkitt’s lymphoma are particularly GC-rich, and that this disease is highly responsive to cyclophosphamide. This view has driven the design of

The chemotherapy of cancer

429

mustard analogues such as tallimustine (a nitrogen mustard conjugated to a netropsin analogue) which has enhanced DNA sequence recognition properties. Tallimustine, which is presently in clinical trial, binds in the minor groove and spans five base pairs, recognising a 5’-GAAAT sequence. In addition to MDR-related resistance, the effect of nitrogen mustards can be significantly reduced by an increase in concentration of glutathione in the cell. The highly nucleophilic glutathione forms adducts with mustards which are then no longer able to react with DNA. Cells also become resistant to nitrogen mustards by carrying out repair operations whereby the mustard adducts are excised and the damaged DNA resynthesized. Repair inhibitors that could be co-administered with mustards to enhance their clinical effectiveness are under active development. 9.8.1.2 Aziridines Rather than form aziridinium ions as reactive intermediates, thiotepa (9.14) and related analogues have an aziridine ring already incorporated in their structure. Ring-opening of the aziridines with nucleophiles is slower compared with the fully-charged aziridinium ions of the mustards. However, depending upon the pKa of the aziridine nitrogen, there is likely to be significant protonation at physiological pH, meaning that, in practice, the aziridinium ion may be the reactive species. Thiotepa itself has been used in the treatment of ovarian and breast carcinoma. It is injected intramuscularly, intravenously or directly into the tumour mass or administered by intrapleural or peritoneal infusion. A substituted benzoquinone ring has also been employed as an “anchor” for the aziridine groups. In the experimental agents AZQ (9.15) and BZQ (9.16), the aziridine moieties are deactivated by the withdrawal of electrons from the nitrogen into the quinone carbonyl groups via the ring. These molecules are employed as bioreductive prodrugs, as reduction of the quinone ring to either the semiquinone or the hydroxyquinone species reverses the electron flow and raises the pKa of the nitrogen. This allows activation of the aziridine rings to the corresponding aziridinium ions via protonation.

9.8.1.3 Methanesulphonates Busulphan (9.17, 1,4-di(methanesulphonyloxy)butane) is the best known example of an alkyl dimethanesulphonate with significant antitumour activity.

Introduction to the principles of drug design and action

430

These compounds are known to cross-link DNA; the methanesulphonyloxy moieties act as leaving groups after attack by nucleophilic sites on DNA. From a mechanistic viewpoint, the methanesulphonate groups should participate in an SN2type alkylation

reaction. The mode of action has been investigated by SAR studies that have revealed that unsaturated analogues of known stereochemistry such as the corresponding butyne and trans-butene derivatives are inactive, whilst the cis-butene derivatives retain activity. It is thought that the activity of the cis-analogue and the more-flexible saturated busulphan depend on their ability to form a cyclic derivative by 1,4bisalkylation of suitable nucleophilic groups. 1,4-di(7-guanyl)butane has been identified as a product of reaction between busulphan and DNA suggesting that this drug acts as an interstrand cross-linking agent in a similar manner to the nitrogen mustards. However, a study of the structure of urinary metabolites suggests that cysteine residues in certain proteins are also alkylated. Busulphan causes significantly less nausea and vomiting than other DNA crosslinking agents and is thus more acceptable to patients. It is highly effective in chronic granulocytic leukaemia where it can keep patients almost symptom-free for long periods of time. Unfortunately, it does have a profound toxic effect on granulocytes and megakaryocytes which requires careful monitoring. Nitrogen mustard analogues have also been synthesised in which their β-chloroethyl groups have been replaced by sulphonate esters. 9.8.1.4 Triazeneimidazoles Dacarbazine (9.18, 5-(3’,3’-dimethyl-1-triazenyl)imidazole-4-carboxamide, DTIC) was one of several triazenes originally evaluated as potential inhibitors of purine biosynthesis. Although it was found to have a wide spectrum of activity ranging from malignant lymphomas to melanomas and sarcomas, it was later established that its mechanism of action is not associated with inhibition of purine biosynthesis. Instead it was demonstrated that demethylation occurs in vivo to afford the corresponding 5amino derivative and a transient methyl diazonium ion which has been shown through radiolabeling experiments to methylate DNA at guanine N7 positions. Since triazenes are liable to photochemical decomposition, the infusion bottle must be protected from light during intravenous administration. The irritant properties of the compound also preclude contact with skin and mucous membranes. Combinations of DTIC with adriamycin, bleomycin and vinblastine have also been employed.

The chemotherapy of cancer

431

9.8.1.5 Imidazotetrazinones Temozolomide (9.19) is a relatively new alkylating agent that entered Phase II clinical trials in 1996 and will be licensed for use in melanoma and brain tumours. It is similar in its spectrum of activity and cross-resistance profile to the nitrosoureas, and also causes a dose-limiting bone marrow suppression. Temozolomide was developed from the experimental agent mitozolomide (9.20) and, like DTIC, acts as a source of CH3+ carbocations. This agent has a novel mechanism of action; after positioning itself in the major groove, nucleophilic attack of a water molecule on the lactam carbonyl initiates a degradation process that affords an amino imidazole derivative, N2 and CO2 in addition to the CH3+ ion. Formation of these smaller stable molecules presumably acts as a driving force for the degradation process.

It has now been established that the methyl carbocations generated methylate the N7positions of guanine bases in the major groove. Furthermore, it is known that there is a preference to alkylate guanines occurring in the centre of runs of three or more guanine bases. It is possible that, as with the nitrogen mustards, the selectivity of temozolomide might be due to its ability to target guanine rich gene sequences associated with some tumours. 9.8.1.6 Nitrosoureas The nitrosoureas are known to alkylate DNA and cause both interstrand cross-links and monoadducts at a number of different sites. The screening of a large number of nitrosourea analogues established the structural unit for optimal activity as the 2chloroethyl-N-nitrosoureido group. The most significant property of the nitrosoureas

Introduction to the principles of drug design and action

432

is their activity towards cancer cells in the brain and cerebrospinal fluid, the so-called sanctuary sites. This is due to the relatively high lipophilicity of these molecules compared to other agents. Two examples of clinically-useful nitrosoureas are carmustine (9.21, N,N’-bis(2-chloroethyl)-N-nitrosourea, BiCNU) which is administered intravenously and lomustine (9.22, N-(2-chloroethyl)-N’-cyclohexyl-Nnitrosourea, CCNU) which is administered orally.

Although these molecules possess chloroethyl fragments, their activity is not associated with aziridinium ion formation as in the mustards, since the corresponding nitrogen atom is part of a urea structure and so the electron pair on the nitrogen is not available to participate in a cyclisation reaction. Although the mechanism of action has not been firmly established, it is thought that the alkylation of nucleic acids proceeds via generation of a chloroethyl carbonium ion. The alkyl isocyanate fragment also formed is thought to carbamoylate the amino groups of proteins.

Mode of action of nitrosoureas Although their lipophilic properties render them particularly suitable for treating intracerebral tumours, BiCNU and CCNU are also active in malignant lymphomas and carcinomas of the breast, bronchus and colon. Carcinoma of the GI tract, which is notably intractable to drug treatment, also responds to the nitrosoureas. Unfortunately, the nitrosoureas cause severe bone marrow toxicity which is usually dose-limiting. The discovery of antitumour activity in the naturally occurring nitrosourea streptozotocin led to the synthesis of the β-chloroethyl analogue, chlorozotocin (9.23) which has significant antileukaemic activity but with reduced bone-marrow toxicity.

The chemotherapy of cancer

433

This initiated the design of other mono- and disaccharide analogues, whilst association of the 2-chloroethyl-N-nitrosoureidomethyl moiety with an aminopyridine ring has provided other experimental agents active against brain tumours.

9.8.1.7 Metal complexes Cisplatin is often quoted as one of the greatest successes in the development of chemotherapeutic agents. It has pronounced activity in testicular and ovarian cancers, however, as with many other clinically-useful drugs, it was discovered by serendipity rather than by design. In the 1960s, Rosenberg and co-workers observed that passing an alternating electric current through platinum electrodes in an electric cell containing E. coli led to arrest of cell division. The cause of the cytostatic effect was eventually traced to platinum complexes formed electrolytically at concentrations of only 10 parts per million in the presence of ammonium salts and light. cis-Diaminedichloroplatinum (9.24, cisplatin) was identified as one of the most active complexes, and the cis isomer was shown to possess significant antitumour activity. Cisplatin was introduced clinically in the UK in 1979. Cisplatin has been shown to form intrastrand cross-links in the major groove of DNA with preferential interaction between (guanine N7)-(guanine N7), (guanine N7)(adenine N7) (in both cases with the bases adjacent to one another) and (guanine N7)(guanine N7) (with one base inbetween the alkylated guanines). Based on gel electrophoresis and NMR studies, these intrastrand cross-links are known to kink the DNA at adduct sites, a phenomenon that can now be observed directly by Atomic Force Microscopy. In some cisplatin-resistant cell lines the adducts are rapidly repaired, and it is thought that the DNA repair enzymes recognise the distortion around the adduct site. Interestingly, the configurational isomer, trans-platin has significantly less antitumour activity. Clinical trials of cisplatin in combination with vinblastine have produced complete remission in 59% of patients with testicular cancer, and 30% remission in ovarian carcinoma. Other tumours which respond to cisplatin include squamous cell carcinoma of the head and neck, bladder carcinoma and refractory choriocarcinoma. Cisplatin suffers from dose-limiting adverse effects including leukopenia, extreme nausea, and

Introduction to the principles of drug design and action

434

renal dysfunction at higher doses. Another disadvantage is the high cost of the drug due to the platinum content. Carboplatin (9.25) is an analogue of cisplatin incorporating a cyclobutyl-substituted dilactone ring. It is better tolerated than cisplatin in terms of GI toxicity, nephrotoxicity and neurotoxicity. However, myelosuppression is more pronounced than with cisplatin. A new cisplatin analogue (JM 216) that can be administered orally is presently in clinical trials. 9.8.1.8 Carbinolamines The carbinolamine group [-NR-CH(OH)-] is an electrophilic moiety found in a number of synthetic and naturally-occurring compounds known to interact with DNA. Trimelamol (9.26), which contains three carbinolamines, was developed from the clinically-active

hexamethylmelamine and pentamethylmelamine analogues which do not contain carbinolamines themselves but are converted during oxidative metabolism. Trimelamol was designed to have the carbinolamine moieties already in place. PhaseII trials carried out in the early 1990s indicated that trimelamol is active in refractory ovarian cancer, and is less emetic and neurotoxic than pentamethylmelamine. A number of different mechanisms of action have been suggested for trimelamol including DNA alkylation or the possibility that it can act as a nucleic acid antimetabolite. Although the precise mode of action has not yet been proven, trimelamol is known to be a reasonably potent DNA interstrand cross-linking agent although the exact sites of alkylation have not been well-defined. One of the

The chemotherapy of cancer

435

drawbacks of trimelamol is its poor solubility in a number of physiologically compatible solvents making it difficult to formulate. Some analogues of trimelamol that partially overcome this problem have been developed and may enter clinical trials in the near future. Anthramycin (9.27) is an example of a naturally occurring carbinolaminecontaining compound that also interacts with DNA. These compounds, known as the pyrrolo[2,1-c][1,4]benzodiazepines (PBDs), are perfectly shaped to fit the minor groove of DNA where they covalently attach to the N2-position of guanine. The PBDs are sequence-specific, recognising a purine-guanine-purine sequence and bonding to the central guanine. They have been shown to possess antitumour activity in the clinic although they suffer from side-effects including cardiotoxicity and bone-marrow suppression. The cardiotoxicity is thought to be due to the phenolic -OH group that can be converted to quinone species capable of producing free radicals with the potential to damage heart muscle. Other PBDs which lack a C9-phenolic group are not cardiotoxic. One of the most active analogues known as sibiromycin possesses a sugar moiety at C7 which is thought to enhance cellular uptake. There is interest in developing an anthramycin analogue suitable for clinical trials, and recent synthetic studies have succeeded in joining two PBD molecules together to make novel minor-groove interstrand cross-linking agents that span up to seven basepairs and recognise a unique purine-GATC-pyrimidine sequence. The design of these new cross-linking agents was based on the concept that new generations of alkylating and cross-linking agents should recognise longer sequences of DNA potentially leading to a higher selectivity for cancer cells coupled with reduced toxicity. These new cross-linking agents are unique in that they appear to cause negligible distortion of the DNA upon binding thus avoiding detection by the DNA-repair enzymes and potentially reducing the possibility of drug resistance. 9.8.1.9 Cyclopropanes Adozelesin (9.28), carzelesin (9.29) and bizelesin are synthetic analogues of (+)CC1065, an extremely cytotoxic antibiotic isolated from Streptomyces zelensis in 1974. It was

Introduction to the principles of drug design and action

436

shown to bind in the minor groove of DNA to 5’-Pu-N-T-T-A or 5’-A-A-A-A-A sequences (N=any other base), alkylating the N3-position of adenine (A italicised). As in adozelesin, the DNA-reactive part of (+)CC-1065 was shown to be a cyclopropane moiety attached to the para-position of a dihydroquinone. It was postulated and later established that attack of the cyclopropane ring by an adenine N3 is driven by the energy released in aromatising the dihydoquinone to a phenol. Unfortunately, (+)CC-1065 is extremely toxic in vivo and has an unusual delayed lethality due to liver toxicity. Adozelesin, U77779 (bizelesin) and U80244 (carzelesin) were synthesised in an attempt to reduce this toxicity, and all three compounds have recently been studied in clinical trials. It is noteworthy that carzelesin is a prodrug; the cyclopropane-dihydroquinone system is generated in situ from the chloromethyl substituent after release of the phenolic -OH moiety through hydrolysis of the carbamate group. Although adozelesin and carzelesin are monoalkylating agents, bizelesin is a minor groove cross-linking agent. 9.8.1.10 Procarbazine Procarbazine (9.30), N-(1-methylethyl)-4-[(2-methylhydrazino)methyl]benzamide, is a hydrazine derivative that was first synthesised as a mono-amine oxidase inhibitor. It was later shown to have significant activity in lymphomas and carcinoma of the bronchus.

The chemotherapy of cancer

437

Metabolic N-oxidation occurs to give azoprocarbazine, and it is believed that subsequent rearrangement produces either methyl diazonium or methyl radicals which act as DNA methylating agents towards guanine residues.

9.8.2 Intercalating agents These agents are flat in shape, usually consisting of three or four fused aromatic rings. They work by inserting between the base pairs of DNA perpendicular to the axis of the helix. Once inserted, they are held in place by interactions including hydrogen bonding and van der Waal’s forces. Intercalation can be detected in naked DNA by an increase in helix length which can be evaluated as an increase in viscosity using sedimentation values or as a change in mobility of DNA fragments in an electrophoresis gel. Some intercalators with arrays of functional groups at either end of the molecule protruding into both the minor and major grooves are known as “threading agents”. A number of different mechanisms of action have been ascribed to intercalatingtype drugs. It can be demonstrated in the laboratory that some intercalators and threading agents block transcription and interfere with other DNA processing enzymes. Many intercalating agents have a preference for GC-rich sequences and, as with alkylating agents, this has been suggested to account for their selectivity. However, some intercalators are known to “trap” complexes between topoisomerase enzymes and DNA leading to strand cleavage. Others are known to chelate metal ions and produce DNA-cleaving free radicals, or to interact with cell membranes. Some examples of the different classes of intercalating agents are given below. 9.8.2.1 Anthracyclines The anthracyclines (sometimes known as the anthraquinones) are a group of antitumour antibiotics first isolated from Streptomyces peucetius. They are the best known family of intercalating agents, consisting of a planar anthraquinone nucleus attached to an amino sugar. Doxorubicin (9.32, Adriamycin) is one of the most important anticancer drugs available because of its broad spectrum of activity. It plays a significant role in the treatment of solid tumours such as carcinoma of the breast, lung, thyroid and ovary, as well as soft-tissue carcinomas. Daunorubicin (9.31, Daunomycin) is an important agent in the treatment of acute lymphocytic and myelocytic leukaemias. Although doxorubicin and daunorubicin are both natural

Introduction to the principles of drug design and action

438

products, semi-synthetic analogues have also been developed including epirubicin (9.33), idarubicin and ametantrone. Four mechanisms have been suggested for the mode of action of the anthracyclines, although there is still controversy about the relative importance of each one. Firstly, the planar ring system can be inserted perpendicular to the long axis of the double helix with the amino sugar appearing to confer stability on the adduct through hydrogen bonding

interactions with the sugar phosphate backbone. Intercalation is known to interfere with DNA processing and transcription, and can be regarded in some cases as a point mutation. Secondly, the anthracyclines are known to form complexes with topoisomerase enzymes leading to strand breaks. Thirdly, binding to cell membranes has been observed, and it has been suggested that this could alter membrane fluidity and ion transport, and disturb various biochemical equilibria in the cell. Lastly, generation of semiquinone species can lead to free radical or hydroxyl radical production (in a fashion similar to that proposed for mitomycin C, see Section 9.15.1) leading to DNA and cellular damage. Radical formation may be mediated by chelation of divalent cations such as calcium and ferrous ions by the phenolic and quinone functionalities, and is believed to be responsible for the cardiotoxicity observed with the anthracyclines.

The chemotherapy of cancer

439

9.8.2.2 Anthracenes These compounds are based on the anthracene nucleus and have three rings rather than the four present in the anthracyclines. Mitoxantrone (9.34) and bisantrene (9.35) are examples of anthracenes that have been reasonably successful in the clinic. Both agents are used clinically for childhood and adult myelogenous leukaemia, nonHodgkins lymphoma and breast cancer. One advantage of these drugs is that they have very low cardiotoxicity compared with the anthracyclines. Mitoxantrone binds to GCrich sequences whereas bisantrene has little base-pair specificity. Like the anthracyclines, there is evidence that DNA is cleaved although the mechanism is not thought to be linked to the generation of reactive oxygen species.

9.8.2.3 Phenoxazines Dactinomycin (9.36), a chromopeptide antibiotic, was isolated from Streptomyces parvulus in the 1940s and used initially as a potent bacteriostatic agent although it was found to be far too toxic. The clinically-useful antitumour activity of dactinomycin was not observed until ten years later when it was tried with great success in the treatment of Wilm’s tumour (a kidney tumour in children) and a type of uterine cancer. The molecule consists of a tricyclic phenoxazin-3-one chromophore with two identical pentapeptide lactones attached. The mode of action of dactinomycin appears, to some extent, to be dependent upon its concentration, with either blockade of DNA synthesis occurring or inhibition of DNA-directed RNA synthesis preventing chain elongation. X-Ray crystallography studies have shown that the phenoxazone ring intercalates preferentially between GC base pairs where it can interact with the N2amino groups. The peptide moieties position themselves in the minor groove and participate in extensive hydrogen bonding and hydrophobic interactions with functional groups in the floor and walls of the groove, thus providing significant stabilisation of the adduct and blockage of RNA polymerase. It also causes singlestrand DNA breaks in a similar manner to adriamycin, either through radical formation or by interaction with topoisomerase. Dactinomycin is used in combination with vincristine in the treatment of Wilm’s tumour, and combined with methotrexate in the treatment of gestational choriocarcinoma. It is also used in some testicular sarcomas and in Kaposi’s sarcoma, a tumour associated with AIDS patients. Tumour resistance

Introduction to the principles of drug design and action

440

to dactinomycin is believed to be due to both reduced uptake and active transport of the drug out of the tumour cells.

9.8.3 Topoisomerase inhibitors DNA topoisomerases are a family of enzymes responsible for the cleavage, annealing and topological state (e.g. supercoiling) of DNA. There are two types of topoisomerase enzymes, I and II. Topoisomerase I enzymes are capable of removing negative supercoils in DNA without leaving damaging nicks. They break one strand of DNA only, with the free phosphate residue becoming attached to a tyrosine residue on the enzyme. The complex then rotates, relieving the supercoiled tension of the DNA, and the two ends are then resealed. Topoisomerase II enzymes cleave double-stranded DNA, passing a complete duplex strand through the cut, followed by resealing of the original strands. The inhibition of these enzymes is believed to be involved in the mode of action of some intercalators such as the anthracyclines and anthracenes, although there has been controversy over whether the drugs bind to the enzyme prior to complex formation or after the enzyme-DNA complex has formed. Amsacrine (9.38) is an example of an acridine that is thought to work by topoisomerase inhibition. It is used in the therapy of myelogenous leukaemias, advanced ovarian carcinomas and lymphomas. In addition to the anthracyclines, some new families of agents such as the ellipticines and camptothecins (see below) have been discovered that specifically inhibit topoisomerase.

The chemotherapy of cancer

441

9.8.3.1 Ellipticene Ellipticine (9.37) is an example of a plant alkaloid that exerts its antitumour action through intercalation and inhibition of the topoisomerase II enzyme. In in vitro cytotoxicity studies, ellipticine is particularly active against nasopharyngeal carcinomas. 9.8.3.2 Camptothecin Camptothecin (9.39) is a plant alkaloid with a unique five-ring system that exerts antitumour activity through inhibition of topoisomerase I.

9.8.3.3 Etoposide

Introduction to the principles of drug design and action

442

Podophyllotoxin extracted from the American mandrake rhizome is another example of a natural product with anticancer activity. Etoposide (9.44) is a semisynthetic glucoside of epipodophyllotoxin, and is used clinically to treat small cell bronchial carcinoma for which it is claimed to be one of the most effective compounds known. Its mechanism of action involves inhibition of topoisomerase II. Many other semisynthetic derivatives, such as podophyllic acid ethyl hydrazide, have been prepared and tested against selected tumours. Some of these appear to have a direct effect on DNA, inducing single-strand cleavage. Antitumour activity has been reported in some leukaemias and lymphomas, oat cell carcinoma of the bronchus and in malignant teratomas.

9.8.4 DNA Cleaving Agents 9.8.4.1 The bleomycins The bleomycins (9.40) are a group of closely related natural products that exert their antitumour activity by binding to DNA in a sequence selective manner followed by strand cleavage. The preparation know as Bleomycin sulphate consists of a mixture of the glycopeptide bases (e.g. A2, A2I, B1–4 etc.) with A2 as the predominant component.

The chemotherapy of cancer

443

The mixture is obtained from Streptomyces verticillus, and the individual molecular weights are in the region of 1300. The bleomycins have attracted interest because they tend to accumulate in squamous cells and are therefore suitable for inclusion in regimens to treat tumours of this cell type in the head, neck and genitalia. Bleomycin sulphate has also been used in Hodgkin’s disease and testicular carcinomas. It may be given intravenously, intramuscularly or subcutaneously. Unlike most anticancer drugs, it is only slightly myelosuppressive, and dose-limiting toxicity is confined to the skin, mucosa and lungs. Enzymes in most other tissues rapidly deactivate the bleomycins, probably as a result of deamination or peptidase activity. However, the bleomycins cause erythema, pain and hypertrophic changes in the skin in areas where there is a lot of keratin. As a result, ulceration in these areas and pigmentation of the nails may occur. In addition, pulmonary fibrosis occurs in 5%–15% of patients. A carboxylic acid derivative, bleomycin acid, has been prepared and the substitution of various amino moieties in the molecule has enabled the synthesis of over 100 analogues. For example, Pepleomycin is a derivative with less tendency to cause pulmonary fibrosis. Despite the size and complexity of the molecule, particularly with regard to the number of chiral centres, the first total synthesis of bleomycin was reported in 1982. Within the bleomycin molecule there are three distinct regions which are believed to contribute towards its mechanism of action. First, the heterocyclic bithiazole moiety (top right as drawn) is thought to intercalate with DNA. Electrostatic attraction of the highly basic sulphonium ion to the phosphate residues in DNA stabilise the adduct. Once bound, the second domain (top left) which consists of a β-hydroxyhistidine, a β-aminoalanine and a pyrimidine forms an iron (II) complex which interacts with oxygen to generate free radicals leading to single and double-

Introduction to the principles of drug design and action

444

strand breaks. Currently, it is not clear whether the activation of this complex is selfinitiating or the result of enzyme catalysis. The third region of bleomycin (bottom left) is glycopeptidic in nature and, while having no direct antitumour activity of its own, may contribute to either drug uptake by tumour cells or provide additional stabilising hydrogen bonding interactions with DNA or associated histone proteins. 9.8.4.2 The Enediynes A new class of DNA-cleaving agent known as the enediynes is in the development stage. These compounds are based on natural products such as esperamycin and calicheamycin whose structures contain two triple bonds in close proximity. These molecules interact with DNA and then undergo a unique thiol-mediated cyclisation (the Bergman cyclisation) during which the triple bonds rearrange to form an aromatic ring. This process causes a proton to be abstracted from a sugar in the DNA backbone leading to strand cleavage. The basis for any tumour cell selectivity with the enediynes is not entirely clear, and it is likely that they will be developed mainly as “warheads” to be attached to other sequence-selective DNA-targeting agents. 9.9 ANTITUBULIN AGENTS 9.9.1 Vinca Alkaloids Plants have traditionally been an invaluable source of medicinal compounds, including anticancer drugs. Following a screening program for agents with potential hypoglycaemic properties, two alkaloids, vinblastine and vincristine, which occur as minor constituents of the Madagascar periwinkle (Vinca rosea), were isolated and shown to reduce white blood cell count. These agents have subsequently played an important role in cancer chemotherapy. The two Vinca alkaloids have complex but similar chemical structures. Vincristine (9.42) is more widely used than vinblastine (9.41) but the plant produces the latter in approximately 100-fold greater quantity. Fortunately, vinblastine may be converted to vincristine by a simple chemical step involving oxidation of the methyl to a formyl group. Furthermore, since these Vinca alkaloids have proved so useful in therapy, efforts

The chemotherapy of cancer

445

have been directed towards the design of new analogues, and the synthesis of vinblastine itself was reported in 1979. Research has also been carried out into cell culture techniques and the use of immobilised plant enzymes as a means to produce the alkaloids more efficiently. Compounds representing the two halves of the structure of the dimeric Vinca molecules occur in much higher proportions in the plant extract, and the possibility of linking these at the appropriate positions and with the correct stereochemistry, has also become feasible. The vinca alkaloids are cell-cycle-specific agents that block mitosis by metaphase arrest. Their cytotoxic effects appear to result from binding to the microtubules. These structures were first characterised in the cytoplasm over 20 years ago and are comprised of two main proteins, the α and β tubulins (Mr c 55 000), which form the microtubule scaffolding upon which many of the dynamic internal processes in living cells, including cell division, depend. Microtubules are long tubular structures of about 25 nm in diameter which form the major component of the mitotic spindle apparatus responsible for the movement of chromosomes during cell division. Binding of the Vinca alkaloids to the tubulins interferes with microtubule assembly causing damage to the mitotic spindle apparatus and preventing chromosomes from travelling out to form daughter cell nuclei. However, the basis for the tumour cell selectivity of these compounds is less clear. Vinblastine sulphate is included on a weekly basis in several drug regimens for treating Hodgkin’s disease, disseminated carcinoma of the breast, choriocarcinoma and testicular carcinoma. Vincristine sulphate, despite its similarity in structure to vinblastine, has a different spectrum of both antitumour activity and side-effects. Several drug combinations include vincristine for the treatment of acute lymphoblastic and myeloblastic leukaemias, Wilm’s tumour, rhabdomyosarcoma, neuroblastoma, retinoblastoma, soft tissue sarcomas and disseminated cancer of the breast, testes, ovaries and cervix. Although neurotoxicity can be pronounced, a relatively low bonemarrow toxicity renders it suitable for combination with drugs that cause greater bonemarrow depression. Vindesine sulphate (9.43) is a semi-synthetic product derived from vinblastine, although its spectrum of activity corresponds more closely to that of vincristine. Efforts towards the synthesis of new analogues are generally directed at reducing the

Introduction to the principles of drug design and action

446

neurotoxicity (peripheral parasthesia and autonomic neuropathy) associated with the vinca alkaloids. 9.9.2 The Taxanes ®

Paclitaxel (9.45, Taxol ) is a highly complex tetracyclic diterpene found in the bark and needles of the Pacific yew tree Taxus brevifolia. The cytotoxic nature of extracts of Taxus brevifolia was first demonstrated in 1964; pure taxol was isolated in 1966 and its structure published in 1971. However, taxol has only recently appeared in the clinic, over 30 years since its discovery. The mechanism of action of this agent also involves the microtubules. There is an equilibrium between the microtubules and tubulin dimers, and the assembly and disassembly of dimers is governed by cell requirements. Taxol is thought to promote microtubule assembly, shifting the equilibrium in favour of the polymeric form of tubulin and reducing the critical concentration of the non-polymerised form by stabilising the microtubule complex. There are also some reports of taxol acting as an immunomodulator and activating macrophages to produce interleukin-1 and tumour necrosis factor (see section 9.14).

The main indication for taxol is in the treatment of refractory ovarian cancer, although clinical trials are currently underway to assess its therapeutic value in breast cancer. Side effects include myelosuppression, peripheral neuropathy, cardiac conduction defects (arrhythmias), alopecia, muscle pain, nausea and vomiting. The barks from a very large number of yew trees would be required to provide a single course of treatment for an ovarian cancer patient, and the problems associated with producing sufficient quantities of this agent account, in part, for its delay in arriving in the clinic. Fortunately, semi-synthesis is now possible by extracting baccatin (taxol without the ester group) in large amounts from the leaves (a renewable resource) of a related species, Taxus baccata. The ester side chain (shown to the left of the molecule as drawn above) can now be made synthetically and then joined on to baccatin to provide taxol. There is also interest in the total synthesis of taxol, with the first route being reported in the early 1990s. Taxol itself has relatively poor water solubility, and so there is interest in producing analogues with improved solubility profiles. A related compound called taxotere has also generated interest.

The chemotherapy of cancer

447

9.10 MISCELLANEOUS AGENTS 9.10.1 Hydroxyurea Hydroxyurea (9.46) is thought to exert its antitumour action in the S-phase of the cell cycle where it inhibits ribonucleotide reductase. This has the effect of depleting the deoxynucleoside triphosphate pool. An advantage of hydroxyurea is that it can be taken orally. It is generally used in combination chemotherapy with other agents in the treatment of melanoma and chronic myelogenous leukaemia.

9.10.2 Mithramycin Mithramycin (9.47) is a chromomycin antibiotic isolated from Streptomyces plicatus. It is believed to exert its antitumour activity by forming a complex with magnesium ions and interfering with DNA-directed RNA synthesis. Mithramycin is clinically useful in the treatment of testicular cancers refractory to standard chemotherapy. However, it seems to have a greater role in reversing hypercalcaemia (associated with malignant disease) through its action on osteoclasts. 9.10.3 Mitotane Mitotane (9.48) is a congener of DDT and is used in the treatment of adrenal carcinoma in which it causes tumour regression and alleviates some of the symptoms associated with excess steroid production.

Introduction to the principles of drug design and action

448

9.11 IMMUNOSUPPRESSIVE AGENTS 9.11.1 Azathioprine Azathioprine (9.49) is a prodrug of 6-mercaptopurine (Section 9.7.2) that contains an imidazoyl “protecting” group. The immunosuppressive effect is believed to be due to the disruption of nucleic acid metabolism at stages where cell proliferation occurs in response to antigen exposure. Azathioprine is therefore useful in the treatment of leukaemias. Allopurinol is often given concurrently to inhibit xanthine oxidase which would otherwise metabolically inactivate the released mercaptopurine. 9.11.2 Cyclosporin (Neoral®) Cyclosporin (9.50) is an immunosuppressive agent used to facilitate the acceptance of bone marrow grafts. It can be used as a therapeutic option in the treatment of leukaemias. The drug is expensive and some therapeutic regimens include concomitant ketoconazole (an antifungal agent) which, by inhibiting the liver enzymes involved in the metabolism of cyclosporin, promotes circulating levels of the agent.

The chemotherapy of cancer

449

9.12 ANTI-HORMONAL AGENTS It has long been recognised that tumours derived from hormone-dependent tissues are themselves dependent on the same hormone. This has been demonstrated by the remissions observed in premenopausal breast cancer following ovariectomy and in prostatic cancer following orchiectomy. 9.12.1 Breast Cancer Oestrogens act as promoters rather than initiators of breast tumour development and can also facilitate tumour invasiveness by stimulating the production of proteases which can degrade the extracellular matrix. Oestrogen receptors (ERs) can be detected in 60–80% of human breast cancers. They consist of specific oestrogen-binding proteins (termed oestrophilins) located in the nucleus of oestrogen responsive breast cells. Oestradiol diffuses into the nucleus, where it binds to an unoccupied receptor site to form an oestrogen-receptor complex. This complex then binds to genomic DNA and stimulates mRNA synthesis which in turn stimulates protein synthesis at the ribosome and subsequent cell division. Oestrogen receptors have been found to be related to the sensitivity of the tumour to anti-endocrine treatment in that approximately 60% of ER positive breast cancer patients respond to endocrine therapy.

Introduction to the principles of drug design and action

450

9.12.1.1 Anti-oestrogens After surgery with associated radiation therapy to remove the tumour mass, antiendocrine therapy is initiated to prevent the growth of metastases. Tamoxifen (9.51) is used as a first line anti-oestrogen; it competes with oestrogen for ER so preventing oestrogen activation and subsequent tumour growth. One third of non-selected postmenopausal patients respond and the rate is higher (50–60%) for ER-positive tumours. Tamoxifen is very well tolerated with only a few side effects related to its weak agonist action. However, it is speculated that tamoxifen use could be associated with an increased risk of endometrial cancer. Newer antagonists with reduced oestrogenic activity e.g. alkylamide analogues of oestradiol (9.53, ICI 164, 384 and 9.54, ICI 182,780) and pyrrolidine-4-iodotamoxifen (idoxifene) have been introduced. Tamoxifen-resistant tumours are sometimes amenable to treatment with a second line drug such as an aromatase inhibitor or occasionally a progestin (e.g. medroxyprogesterone acetate). 9.12.1.2 Aromatase inhibitors Androstenedione and testosterone are converted by the cytochrome P450 enzyme aromatase to oestrone and oestradiol, respectively, as the final step in the steroidogenesis pathway from cholesterol. Selective inhibition of aromatase would lead to reduced oestrogen plasma levels without affecting other hormones produced by the steroidogenesis pathway. Aminoglutethimide (AG) was the first clinically useful aromatase inhibitor to be discovered and is still used despite undesirable CNS effects and a lack of target enzyme specificity. The latter probelm is associated with effects on other cytochrome P450

The chemotherapy of cancer

451

enzymes in the pathway. Several potent and more-specific inhibitors free from CNS effects have been introduced in recent years. These include the reversible inhibitors, fadrozole, Arimidex®, letrozole, vorozole and the irreversible inhibitors, 4hydroxyandrostendione, plomestane and exemestane. These agents are discussed in Chapter 6 and Section 8.6.3 in more detail. 9.12.2 Prostatic cancer Prostatic cancer is promoted by the androgen dihydrotestosterone (DHT) derived from testosterone by the action of 5α-reductase. Prostatic cancer is mainly hormonedependent and is usually well developed on presentation so that survival rates are low and treatment is aimed at increasing the time of survival and the quality of life. Surgical removal of the prostate or testes (orchidectomy) are now less prevalent treatments having been largely replaced by endocrine therapy (oestrogens) and in recent years by treatment with anti-androgen or luteinising-hormone releasing hormone (LHRH) analogues. 9.12.2.1 Oestrogen therapy Oestrogen therapy (diethylstilboestrol, DES) acts by inhibiting the hypothalamicpituitary system through a negative feed-back mechanism resulting in a fall in the secretion of luteinising hormone (LH) from the pituitary and a subsequent decrease in testosterone synthesis by the testes (see Figure 9.1). This “chemical castration” has now lost favour due to cardiovascular complications and the feminising side effects associated with oestrogen. 9.12.2.2 LHRH analogues The introduction of LHRH analogues has provided an alternative to oestrogens and orchidectomy in the treatment of advanced prostate cancer but without significant side-effects. Naturally occurring LHRH (Pyr-Glu-His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-

Introduction to the principles of drug design and action

452

Gly-NH2) has a short half-life (and a pulsitile action on the receptor) but by substituting the amino acid at the 6th position, deleting the amino acid at the 10th position, and adding an ethylamide group to the proline residue at the 9th position, a synthetic analogue (leuprorelin, Prostop SR®) is produced that has a greatly increased potency together with prolonged activity and a non-pulsitile action on the receptor. The initial effect of LHRH agonists is stimulation of the secretion of LH and follicle stimulating hormone (FSH) leading to elevations of serum testosterone to 140–170% of basal levels within several days. Continuous administration, however, leads to dramatic inhibitory effects through a process of “down regulation” of LHRH pituitary membrane receptors, and a reduction in gonadal receptors for LH and FSH, resulting in suppression of testosterone secretion comparable to surgical castration. Thus chronic administration causes the pituitary gland to become refractory to additional stimulation by endogenous LHRH and testicular androgen production is prevented. The most commonly used LHRH agonists are leuprolide (D-Leu6-DES-Gly-NH210LHRH ethylamide), buserelin (D-Ser(But)6-DES-Gly-NH210-LHRH ethylamide, Suprefact®) and goserelin (Ser-(But)6-AZ-Gly-NH210-LHRH, Zoladex®). A specific side-effect of LHRH analogue treatment is a transient worsening of symptoms (including increased bone pain) during the first week of therapy as a result of the initial testosterone surge, with approximately 3–17% of patients affected. This has led to co-administration of anti-androgens before or with the first LHRH analogue to prevent this effect. Other side-effects of these agents include atrophy of the reproductive organs, loss of libido, and impotence. As they are only able to cause a decrease in testicular androgens whilst leaving adrenal androgen production unaffected, the use of anti-androgens in combination with LHRH agonists has been studied and found to give greater survival rates than LHRH agonists alone. 9.12.2.3 Anti-androgens Anti-androgens inhibit the binding of dihydrotestosterone and other androgens to the androgen receptors in target tissues. Target cells are located in all areas of the body that depend on androgens, e.g. the male genital skin, the seminal vesicles, the prostate, fatty tissues and breast tissue as well as the hypothalamus and the pituitary. Antiandrogens bind to the androgen receptor, creating a receptor-(anti-androgen) complex which is unstable and transient. Therefore, androgen-dependent gene transcription and protein synthesis are not stimulated. However, anti-androgens are capable of blocking the tropic effect of all androgens, not only in the prostate but also in the hypothalamus and pituitary. These agents have both central (hypothalamus/pituitary) and peripheral (prostate) effects, and the most extensively studied is cyproterone acetate (9.55, Cyprostat®). In the target prostatic cell, it acts as a competitive inhibitor of the binding of DHT to androgen receptors and centrally, it lowers LH and, therefore, plasma testosterone levels, due to its progestin-like activity. A Phase III study has revealed objective responses in 33% of patients with advanced prostate cancer and stabilisation of the disease in 40%. Lethal cardiovascular events were slightly higher with this drug than with DES.

The chemotherapy of cancer

453

Non-steroidal anti-androgens have limited central activity but significant and potent peripheral effects, displacing testosterone and DHT from the androgen receptor, not only in the prostate, but also at the level of the hypothalamus. This latter blocking effect leads to an increase in the release of LHRH and subsequent LH production, leading to a slow but gradual rise in serum testosterone levels to overcome the blockade which can then stimulate prostatic tumour growth. Flutamide (9.56, Drogenil®) was the first non-steroidal anti-androgen to be developed. Flutamide is a prodrug, the active metabolite being hydroxyflutamide (9.57) which acts

Introduction to the principles of drug design and action

454

by inhibiting the uptake of testosterone or the nuclear binding of testosterone and DHT to the androgen receptor. Hydroxyflutamide has peripheral and central activity on all androgen target cells, although gynaecomastia is a common side-effect with approximately 61% of patients being affected, 10% of whom suffer severely. Other side-effects include nausea, vomiting and diarrhoea. As flutamide is a pure antiandrogen, it does not inhibit gonadotrophin production by the pituitary and so gonadal and adrenocortical steroidogenesis continues unabated. This results in a normal or elevated serum testosterone level which allows patients to retain their libido. Molecular modelling studies of hydroxyflutamide have attributed its greater binding affinity to the dominant conformation in which the NH bond is hydrogen bonded to the hydroxyl function. The related compounds, Anandron (9.58) and Casodex (9.59) (see later), have similar conformations.

Figure 9.1 Control of androgen levels by the Hypothalamus-Pituitary axis. Nilutamide (anandron) is another non-steroidal anti-androgen that has some structural similarities to flutamide, although it does not require metabolic activation. It is well absorbed and has a much longer half-life than flutamide (45 hours, compared to 5–6 hours for flutamide), permitting a once-daily dosage. It is centrally and peripherally active, and is therefore associated with a rise in plasma LH and subsequently plasma

The chemotherapy of cancer

455

testosterone. Nilutamide is particularly useful in patients who are intolerant to flutamide. Casodex® is the most recent anti-androgen to be studied. It is well absorbed and has a half-life of 5–7 days. It blocks androgen receptors peripherally and centrally and is associated with increased plasma testosterone levels in men. Casodex® is currently undergoing clinical investigation as a monotherapy and also in combination with other agents for advanced prostate cancer. The results of studies to date suggest that it causes few significant side-effects. 9.13 ENZYMES Asparaginase (crisantapase, Erwinase®) is produced by Erwinia chrysanthemi and is a 133,000 molecular weight tetrameric protein used in the treatment of acute lymphoblastic leukaemia. The mechanism of action is based on the fact that these particular tumour cells have very low levels of asparagine which is required for cell growth; instead they must obtain this amino acid exogenously. Healthy cells, on the other hand, can synthesis their own asparagine. The strategy involves the administration of crisantapase which reduces the concentration of asparagine in the body by converting it to aspartic acid and ammonia, thereby removing asparagine from the protein synthesis cycle. Whereas healthy cells can rapidly synthesise their own supply of asparagine, the tumour cells succumb to the reduced levels in their environment. Resistance to the drug develops when the tumour cells begin to synthesise their own asparagine. 9.14 BIOLOGICALS 9.14.1 Interferon alpha (Intron A®, Roferon A®, Walferon®) Interferon was discovered in 1957 by scientists at the National Institute for Medical Research (UK). It is a glycoprotein induced in response to viral infections and is usually effective only in species in which it is produced. It was first investigated as an antiviral agent and shown to have a broad spectrum of activity. Most interferons produce an 80–100% reduction in the incidence of experimentally-induced common colds, and their use is also being examined further in treating chronic hepatitis B, papilloma viruses (warts) and virus infections associated with immunosuppressed patients following renal transplantation. A new system of nomenclature for the interferons has been devised. To qualify as an interferon a substance must be a protein which exerts virus non-specific anti-viral activity in homologous cells through cellular metabolic processes involving synthesis of both RNA and protein. The preferred abbreviation for interferon is IFN. Each interferon is then designated by the animal of origin, e.g. human: HuIFN, murine: MuIFN, bovine: BovIFN. The interferons are next classified into types according to antigen specificities, e.g. α, β and γ, which correspond to the previous designations of leucocyte, fibroblast and type II immune, respectively. It was thought that previous

Introduction to the principles of drug design and action

456

type names were misnomers as both leucocytes and fibroblasts can produce each of the two types (α and β) of interferon. α and β Interferons are usually stable in acid media whilst γ interferons are acid-labile. Properly documented differences in molecular size appear to be useful parameters for characterization until more stringent criteria such as amino acid sequence or monoclonal antibody recognition are forthcoming. Molecular weight designations are indicated as HuIFN-α (18 K), MuIFN-β (39 K), etc. Interest in the potential anticancer activity of interferon first arose following encouraging results obtained in Sweden while treating osteogenic sarcoma and myelomatosis. In 1981 doctors in Yugoslavia, using human leucocyte interferon preparations injected directly into tumours, reported substantial improvements or total remissions in cancers of the head and neck. Although the mechanism of antitumour activity is unknown, either a direct effect on malignant cells or a stimulation of the host’s immune system have been postulated. Despite the intense interest generated by the interferons when first discovered, a considerable time elapsed before the compounds were brought into clinical use, the major difficulty being commercial production of sufficient quantities. Large doses are required for treatment, and initially only minute amounts of varying levels of purity became available from human tissue culture methods. However, from the early 1980s there were significant advances in production techniques, including the development of recombinant DNA technology, which allowed biosynthetic interferons of high purity to be made available in clinically-useful quantities. The three principal types of interferon, namely α-IFN produced by leucocytes and other lymphoid (lymphoblastoid) cells, β-IFN from fibroblasts and γ- (or immune) IFN became available initially. More recently, lymphoblastoid interferon, Wellferon (a complex mixture of α-interferons and recombinant interferon) and Intron A (a pure αinterferon) have become available. α-Interferon is now licensed for the treatment of Kaposi’s sarcoma, and certain blood cancers. Combination chemotherapy with interferon appears to be promising with some treatment regimes. For example, combinations of Intron A® with doxorubicin, cisplatin, vinblastine, melphalan and cyclophosphamide have been evaluated in ovarian, cervical, colorectal and pancreatic carcinomas. The interferons have some toxicity problems; dose-related side-effects include influenza-like symptoms, lethargy and depression. Myelosuppression, affecting granulocytes, may occur along with cardiovascular problems such as hypotension or hypertension and arrhythmias. In 1980, researchers at Upjohn showed that a new group of 6-phenylpyrimidine derivatives caused the body to produce interferon. It has now been demonstrated that the same drugs protect animals against viruses and also improve their defences against tumour cells, although this has not yet been demonstrated in humans. It has also been suggested that interferon stimulates prostaglandin synthesis, and this may help to explain how interferon inhibits cell growth. 9.14.2 Tumour necrosis factor Tumour Necrosis Factor (TNF) is a glycoprotein produced by macrophages, monocytes and natural killer cells, and is partly responsible for tumour cell lysis.

The chemotherapy of cancer

457

Phase I and II studies, administering the agent either intravenously or intramuscularly, have been carried out on TNF with mixed results. Part of the problem is ensuring that a high enough concentration of TNF reaches the tumour site, but this has to be balanced against adverse effects which include hypotension and cardiotoxicity. Transient fever has also been observed in some patients, as have haematological disturbances, the latter being reversible on cessation of treatment. 9.14.3 Interleukin (Aldesleukin®, Proleukin®) Interleukin-1 has been shown to possess both direct and indirect antitumour effects. Its main use in the treatment of cancer is for the protection of bone marrow cells from the deleterious effects of radiation and chemotherapy. However, it can also release a cascade of haemopoietic growth factors. Interleukins are proteins that occur naturally in the body; there are several classes which are being studied for their role in inflammatory and immunomodulatory processes. A recombinant interleukin, interleukin-2 (IL-2), is used clinically in metastatic renal cell carcinoma, where it is administered by intravenous infusion. Unfortunately, the response rate is less than 50% and there is a comprehensive toxicity profile, with one of the most common adverse effects being capillary leakage which leads to hypotension and pulmonary oedema. Investigations are also underway to determine whether IL-2 can enhance the efficacy of tumour vaccines. 9.14.4 Growth factors Growth factors or cytokines are proteins which affect cell growth and maturation. Recombinant technology has allowed the production of large amounts of cytokines and there are several in clinical trials. Haemopoietic growth factors have found a use in counteracting the myelosuppressive side effects associated with many anticancer agents. Granulocyte colony stimulating factor (G-CSF) and granulocyte-macrophage colony stimulating factor (GM-CSF) increase the circulating number of neutrophils, eosinophils and macrophages by inducing inflammation, and it has been shown that some tumour cells possess receptors for these CSFs. The dosing regimen is usually once daily, with side effects including influenza-like symptoms. Erythropoietin has also been shown to be clinically useful in treating certain types of malignant anaemia, and is naturally produced by the body in response to hypoxia. Recombinant technology has now been applied to the production of erythropoietin. Inhibiting growth factors can lead to useful antitumour activity, and known inhibitors include octreotide (Sandostatin®) which is in clinical use, and suramin (a polysulphonated naphthylurea). These two compounds are analogues of somatostatin, a naturally-occurring growth hormone. Octreotide is administered subcutaneously and is useful in controlling symptoms, but does not always cause tumour reduction. Suramin binds proteins extensively due to its polyanionic nature. Early clinical trials with suramin were hampered by severe toxicities (renal and liver dysfuntion, adrenal insufficiency and peripheral neuropathy), but with suitable dosage adjustments clinical trials are continuing in breast and prostate cancer. The clinical use of lymphokine (a

Introduction to the principles of drug design and action

458

protein or glycoprotein produced by a lymphocyte) and natural killer cells is also under investigation. 9.15 PRODRUGS AND DRUG TARGETING APPROACHES A number of different strategies are available for targeting cytotoxic agents to tumour sites or for activating them inside or near a tumour. Some examples of these are described below. Also see the section on Gene Targeting in 9.17.1. 9.15.1 Bioreductive prodrugs Mitomycin C (9.60) is a naturally-occurring antitumour antibiotic, considered to be the prototype bioreductive alkylating agent. The three components of the mitomycin molecule essential for its mode of action are the quinone, aziridine and carbamate moieties. It is thought that initial reduction of the quinone (one-electron reduction yields a semiquinone, whilst a two-electron reduction gives the hydroquinone) leads to transformation of the heterocyclic nitrogen from an amido to an amino form which facilitates elimination of the β-methoxide ion. Tautomerisation of the resulting iminium ion and loss of the carbamate group then creates an electrophilic centre which is susceptible to attack by a nucleophilic DNA base. Nucleophilic attack of the aziridine moiety by a nucleophile on the opposite strand of DNA also occurs, leading to an interstrand cross-link. It is now

known that the main mode of DNA interaction for mitomycin involves cross-linking two guanine-N2 groups within the minor groove. However, the most important feature of mitomycin is the bioreductive “trigger” that is required before cross-linking can take place. It is known that the centres of some tumours, particularly older ones of larger size, are hypoxic due to a poor blood supply. The bioreductive conditions that exist at the centre of these tumours is thought to explain the tumour cell selectivity of mitomycin, which is successfully used to treat solid tumours such as those of the colon, lung and breast. The concept of bioreductive activation has caused great interest, and a number of other compounds have been designed based on this mechanism of action (see Chapter 7).

The chemotherapy of cancer

459

9.15.2 Estramustine Estramustine (9.61) is a nitrogen mustard that exerts its action by alkylating DNA. The rationale behind this prodrug is that by linking the mustard to oestradiol, the hormone component of the molecule may preferentially transport the drug to those cells which bear oestrogen receptors. Once at the site, the mustard-hormone conjugate may hydrolyse so that the mustard fragment is taken into the cell.

9.15.3 Photoactivated prodrugs (Photodynamic Therapy; PDT) There is considerable interest in the concept of Photodynamic Therapy which involves the administration of a non-toxic prodrug that can be activated selectively at the tumour site by light of a specific wavelength. This approach has been in use for a number of years for the treatment of psoriasis (PUVA treatment) using 8methoxypsoralen (9.62). This agent

is relatively non-toxic until exposed to UV light when it cross-links DNA at thymine sites causing distortion of the DNA helix. This concept has now been extended to the treatment of cancer where it has been discovered that porphyrin-type molecules are selectively taken up by some tumours. Photofrin® is an example of a commercially available agent for Photodynamic Therapy. The drug is administered and time allowed for it to circulate and to be taken up by the tumour. Light is then administered locally to the tumour usually from a laser source. The use of surgical lasers with flexible optical fibres means that tumours in inaccessible places can be irradiated using endoscopy and modern techniques in “keyhole” surgery to minimise trauma to the patient. So far, tumours in the GI tract and on

Introduction to the principles of drug design and action

460

the ovaries have been treated by PDT. An intense non-laser light source has recently been marketed for this type of therapy and many new types of prodrug are under development. 9.15.4 Antibody-drug conjugates This type of therapy involves the administration of a cytotoxic agent attached to an antibody. Although, in principle, the antibody should guide the drug selectively to the tumour cells, little success has been achieved in practice. One problem with this approach is that, once bound to the tumour cell, the drug-antibody conjugate may not be internalised and so little cytotoxic affect is achieved. For example, studies with vinblastine-antibody conjugates demonstrated that although doses of the agent could be lowered thus reducing toxicity to some degree, the overall efficacy of the treatment was not a significant improvement over vinblastine treatment alone. 9.15.5 ADEPT ADEPT is an acronym for “Antibody-Directed Enzyme Prodrug Therapy” (see also Section 7.4.3), and is an advance over the use of antibody-drug conjugates alone. This treatment is based on the idea that a prodrug is usually converted into its active form by some enzymatic process. In ADEPT, an antibody-enzyme conjugate is initially administered to the patient which leads to the localisation of enzyme at the tumour site. The prodrug is then administered so that conversion to the active cytotoxic agent is only achieved when it comes into contact with enzyme at the tumour site. Significant progress has been made with the development of this approach, despite the fact that the choice of enzyme is limited to those not generally found in the body. Glutamase and nitro reductase enzymes are presently being studied, and a number of potential prodrugs have been developed. Initial trials have produced very promising results. 9.15.6 GDEPT Gene-Directed Enzyme Prodrug Therapy is related to ADEPT in that an enzyme localised at the tumour site releases a cytotoxic agent from its prodrug form. However, instead of delivering the enzyme to the tumour via an enzyme-antibody conjugate, it is delivered using a gene therapy approach. For example, research is presently being carried out into the use of viral vectors to introduce genes coding for specific enzymes into tumours and organ systems. This is an “enabling” technology in that it could be used to deliver any type of drug to a particular organ for any disease state including cancer. 9.16 NEW RESEARCH TOOLS The search for new anticancer drugs has been enhanced by the development of a wide variety of new research tools. For example, advances in molecular biology allow new

The chemotherapy of cancer

461

genes to be identified and then sequenced with great rapidity. An astonishing number of cancer-related genes such as BRCA1 and BRCA2 have already been discovered, and the list is increasing rapidly. As a result of identifying these genes, it is likely that gene targeted agents can be developed (see below). There is even a commercially available mouse strain known as Oncomouse® that has certain oncogenes incorporated into its genome. These mice will typically die of the gene-related tumour within a known period of time, and so the effect of novel therapeutic agents on life span can be evaluated. Once cancer related genes and their corresponding proteins (i.e. receptors or enzymes) have been identified, it is possible to use X-ray crystallography and NMR techniques to elucidate the precise 3-dimensional shape of the protein both in the solid state and in aqueous solution. Computer-aided molecular modelling may then be used to design ligands that can interact with specific parts of the protein to modify its function (see Chapter 3). Significant advances have also been made in screening technology. For example, with DNA-interactive agents it is not only possible to study the sequence-selectivity of DNA-binding using in vitro “footprinting” assays, but “in vivo footprinting” techniques have recently become available which allow individual drug binding sites in the nucleus of living cells to be identified. Progress in the use of robotics to handle multi-well plates means that “high-throughput” screening can be used to evaluate large numbers of compounds in a short period of time. A new synthesis technology known as “Combinatorial Chemistry” has developed in parallel to take advantage of high throughput screening. With combinatorial techniques, molecular fragments are joined together in a sequentially random fashion to provide a mixture of a large number of molecules (possibly 1000s) within a short period of time. The use of a large number of fragments provides “molecular diversity” in which “libraries” of molecules of widely different 3-dimensional shapes are produced. The key to success with this technology is a “tagging” procedure in which ingenious techniques are used to ensure that each new compound has a unique and traceable “label” even though it may be in a mixture of hundreds or even thousands of other molecules. These libraries can be passed through high-throughput screens and the structure of any one compound in the mixture can be traced through the tag or by the history of the synthesis. The active molecule can then be re-synthesised on a larger scale for further evaluation. 9.17 FUTURE POSSIBILITIES Cancer research is funded through government bodies and initiatives, charitable organisations and a number of pharmaceutical companies with strategic intents in this area. Although all researchers, whether industrial or academic based, would prefer to see funding for cancer research increased, it is often argued that the problem is a shortage of good ideas for new agents and therapies rather than a short-fall in funding. The present lack of effective drugs and treatments, given the resources that have been channelled into cancer research world-wide since President Nixon declared a “War on Cancer” in 1971, would tend to support this hypothesis. A few examples of new

Introduction to the principles of drug design and action

462

approaches under active development are briefly described below. The exciting new concepts of ADEPT and GDEPT have already been described above. 9.17.1 Gene targeting Once a gene has been identified and associated with a particular disease such as cancer, it should be theoretically possible to develop agents capable of selectively targeting either the gene itself or the equivalent messenger RNA, making use of the type of interactions (i.e. hydrogen bonding) that allow nucleic acid strands to pair together. The binding of such agents should prevent transcription by mechanisms including the direct inhibition of RNA transcriptase. The same objective might be achievable through the use of small lower molecular weight ligands. Selectivity is the most important issue here, as it has been calculated that it may be necessary to actively recognise up to 15 to 20 base pairs of DNA in order to selectively target one gene in the entire human genome (approximately 100,000 genes). However, in practice, sufficient selectivity to minimise side-effects might be acceptable. A large number of oncogenes have now been identified and sequenced, and the technology to design and produce molecules with DNA-recognition properties is rapidly advancing. One overriding advantage of the antigene strategy is that inhibition of just one gene will prevent the formation of numerous copies of the mRNA transcript and the corresponding protein. Agents that target DNA or the corresponding messenger RNA are known as antigene or antisense agents, respectively. Ribozymes are a special family of oligonucleotides that target and bind to mRNA but then induce cleavage. 9.17.1.1 Antigene (Macromolecules and small molecules) The macromolecular antigene approach utilises oligonucleotides that interact with double-stranded DNA to form a so-called “triple helix”. The oligonucleotide, which is typically 15–20 base pairs in length, lays in the major groove, held in place mainly by hydrogen bonding interactions. Although a great deal of effort has been put into this area, two major problems exist. Firstly, recognition of DNA is presently limited to runs of cytosines or thymines, which are presently insufficient to successfully target clinically-important gene sequences. Secondly, in practice, oligonucleotides are difficult to deliver to the DNA of cells. In addition to their cost, other problems include stability, poor pharmacokinetics and cell penetration, and rapid metabolism. A great deal of effort has been made to stabilise oligonucleotides towards enzymatic degradation or metabolism, and this has been achieved by chemically modifying the backbone phosphate groups. Strategies to enhance the stability of oligonucleotideDNA complexes have also been pursued; for example, tethering oligonucleotides to intercalating moieties. Designer proteins are also being used to target gene sequences, and one group has succeeded in using phage display technology to produce a protein that can specifically target the bclabl oncogene. In the small molecule area (e.g. molecules of <1500 in molecular weight), agents have now been produced that can recognise up to seven or more base pairs of DNA, and this is an intense area of research. The lexitropsins, the CC-1065 analogues, and

The chemotherapy of cancer

463

the pyrrolobenzodiazepines are three well-known examples of families of DNAinteractive agents that are being developed to recognise extended sequences of DNA. Small molecules have an advantage in that they lack many of the problems associated with oligonucleotides or proteins such as poor stability and delivery. 9.17.1.2 Antisense oligonucleotides An antisense oligonucleotide is a relatively short length of single-stranded DNA (e.g. 15–20 base pairs) with a sequence complementary to a region of the target mRNA. Hybridisation with the RNA then interferes with the process of translation. However, a similar set of problems exist to those described above with regard to delivery, stability and cost. In addition, this approach is not likely to be as efficient as the antigene strategy in which inhibition of just one copy of the gene will prevent many more copies of mRNA from being produced. More recently, a number of research groups have started to search for small molecules capable of interacting selectively with RNA. 9.17.1.3 Ribozymes Ribozymes are similar to antisense oligonucleotides in that they target and bind to mRNA. However, once bound, they cause cleavage of the mRNA fragment. The basepair sequence of the ribozyme allows it to adopt a hair-pin structure when associating with RNA. An enzyme known as RNase H is involved which induces RNA cleavage after ribozyme binding. 9.17.2 Oncogene-product inhibitors Following on from the use of antigene or antisense agents, it may prove possible to develop small ligands capable of binding to oncoproteins (proteins produced from an oncogene), rendering them inactive. Once an oncoprotein has been identified and purified, its 3-dimensional structure can be obtained by either X-ray crystallography or NMR, and molecular modelling then used to help design low molecular weight antagonists. This approach can be supplemented by the use of combinatorial chemistry and high throughput screening. However, it is not likely to be as efficient as the antigene or antisense strategies where inhibition of just one copy of the gene will prevent many copies of the protein from being produced. 9.17.3 Gene therapy Strategies for gene therapy in the treatment of cancer include the use of genetic tagging to identify fragments of a tumour remaining after standard treatments, vaccines containing vectors capable of expressing cytokines, the insertion of genes which activate drugs (e.g. GDEPT), and the insertion of tumour suppresser genes or the replacement of malfunctioning ones. Delivery of genes to the tumour site is generally achieved using viral vectors such as the adenovirus, although there may be problems with infection if the virus is not suitably attenuated. Alternatively, naked

Introduction to the principles of drug design and action

464

DNA can be delivered directly to the tumour, and there is intense effort to develop delivery systems capable of this. There are over 100 clinical trials in progress at the present time, and a number of biotechnology companies are investing heavily in this area. 9.17.4 Growth factor and signalling pathway modulators It is currently accepted that some tumour growth factors are autocrine (affect the secreting cell) or paracrine (affect adjacent cells) hormones. Autocrine growth factors have been studied in greater depth than paracrine ones, and several neuropeptides have been identified, including gastrin releasing peptide (found in small cell lung carcinoma), bradykinin, cholecystokinin, galanin, neurotensin and vasopressin. In response to this, novel therapeutic areas have arisen for targeting these molecules. Considerable effort is currently being put into the discovery and development of inhibitors of various signalling pathways associated with cell growth. For example, Substance P analogues have been shown to block the growth promoting effects of some of the neuropeptides listed above as has substance P(6–11). Many pathways are directly associated with specific oncogenes; for example, ras farnesyltransferase is an attractive target for the treatment of ras-dependent tumours, and a number of compounds have already been identified for development. Protein kinase C and the tyrosine kinases are also involved in signalling pathways. Bryostatin, a natural product obtained from a marine organism, activates protein kinase C and is currently under clinical investigation, as are the tyrophosphitins which disrupt the function of tyrosine kinases. Other approaches are also being considered; for example, monoclonal antibodies could be used to remove the growth factors produced. There is also interest in identifying agents that can selectively initiate apoptosis (programmed cell death) in tumour cells as opposed to healthy ones. 9.17.5 Resistance inhibitors Drug resistance provides a major obstacle to cancer chemotherapy, and resistance may develop to more than one class of agent. Mechanisms include the induction of the MDR (multiple drug resistance) protein which can actively transport different classes of drug out of a cell, an increase in glutathione production which can “neutralise” alkylating agents, or an enhancement of metabolism and/or DNA repair. It is possible that a clinically-useful modification of drug resistance might be achievable through the use of “chemoenhancing” agents capable of sensitising tumour cells to a drug by depressing the resistance pathway. Alternatively, “chemoprotecting” agents may be able to protect healthy cells during chemotherapy. The former approach has been demonstrated experimentally by co-administration of verapamil (a calcium-channel blocker) which can enhance the effect of some anticancer drugs. Also, for use in radiotherapy, “radiosensitising” agents have been identified that can increase the sensitivity of tumours to radiation. There is a corresponding interest in “radioprotective” agents that can protect surrounding healthy tissue during radiotherapy treatment.

The chemotherapy of cancer

465

9.17.6 DNA-repair inhibitors The efficiency of DNA-binding agents such as the nitrogen mustards and cisplatin is known to be reduced by the numerous repair systems that normally protect DNA from damage by carcinogens, radiation and viruses. DNA repair is mediated by a remarkable set of enzymes that first recognise the damage incurred and then set out to repair the lesion, usually by excising the damaged segment of DNA and resynthesising a new one. The mechanism by which these enzymes work is an active area of research, and the first crystal structure of a repair enzyme has only just become available. There is presently interest in developing agents capable of inhibiting DNA repair as, theoretically, they could enhance the therapeutic value of the DNA-binding class of anticancer drugs. However, presumably there is an associated risk that carcinogens and other sources of DNA damage might be rendered more potent. The poly ADP ribosylation (PADPR) of damaged DNA is thought to be associated with the attraction of repair enzymes to the site, and so there has been a considerable effort to develop PADPR inhibitors. So far, a number of potential inhibitors have been identified, some based on simple benzamide analogues. One such agent is presently in clinical trials. 9.17.7 Telomerase inhibitors Telomeres are repeat sequences of DNA found at the end of chromosomes. They are thought to be crucial for cell division during which they signify the end of a chromosome molecule. A number of telomeres are lost during each cycle of cell division, and this is thought to act as a type of “biological clock” leading to a natural cell death when the telomeres have been depleted. Over 80% of different tumour types have been shown to express telomerase and this is thought to offer an explanation for the immortality of cancer cells. Telomerase enzymes have now been purified, assays have been established, and a number of telomerase inhibitors have been identified. It is thought that telomerase inhibitors may have selective toxicity towards tumour cells as telomerase is not expressed to the same level in healthy cells. 9.17.8 Antimetastatic agents A primary tumour is not often the direct cause of death of a patient as it can be removed by surgery or treated with radio- or chemotherapy. It is usually the secondary tumours (metastases) that lead to death as they become too dispersed throughout the body to make further treatment, particularly surgery, effective. Attention has recently been focused on the mechanism by which tumour cells move around the body, either via the blood or the lymph system, and establish themselves in new locations. There is an intense effort to develop drugs that can interfere with the metastatic process. For example, one drug in late clinical trials known as Marimastat® inhibits metalloproteinase enzymes which are thought to be important for metastases and tumour development.

Introduction to the principles of drug design and action

466

9.17.9 Blood-flow modifying agents A good blood supply is important for the growth of a primary tumour and can be even more important for the establishment of a metastatic tumour at a distant location. The enzyme angiogenin has recently been identified as important for establishing new microvasculature at the site of a metastatic tumour. The structure of angiogenin has now been established from X-ray crystallographic studies, and an effort is currently underway to design inhibitors. 9.17.10 Vaccines Vaccination against cancer is an active area of research, with the objective of either activating or inducing a host response to tumour-associated antigens. Patients with melanoma have already been treated with a melanoma vaccine using BCG as an adjuvant; while there was evidence to suggest an immunological response in some patients, efficacy was poor overall. Tumours associated with viruses such as Hepatitis B and Human Papilloma Virus are also targets for a vaccination approach to cancer treatment or prophylaxis. With Hepatitis, a combination of hepatitis immunoglobulin and heat inactivated Hepatitis B vaccine produces greater protection than vaccination with hepatitis immunoglobulin alone. Research in the future may be directed towards avoiding tumour cell or antigen vaccines in favour of using antibodies. This eliminates the risk associated with passing on live tumour cells and also avoids the need to purity antigens. 9.17.11 Chemopreventive Agents (“Neutriceuticals”) As part of a general trend towards preventative medicine, there is growing interest in agents capable of antagonising the effects of carcinogens ingested in the diet or inhaled in the air. Various different mechanisms can be envisaged for agents of this type. For example, it should be possible to “neutralise” carcinogens in the GI tract and/or prevent their absorption. Alternatively, it may be possible to selectively enhance the metabolism of some carcinogens. Although highly attractive, one problem with chemoprevention is that clinical trials would need to be conducted over many years, as it is possible that cancer deaths in old age might be associated with carcinogens ingested in the first few decades of life. It is also difficult to devise in vitro screens for chemopreventive agents. Although it is possible to screen for compounds that enhance metabolism, there is no guarantee that faster metabolism, even for selected carcinogens, will lead to prevention of carcinogenesis. Indeed, some carcinogens are activated by metabolism. A number of lead compounds of diverse structure (i.e. flavensids, terpenes) have originated from epidemiological studies relating to diet. For example, soy beans are known to contain chemopreventive agents, as a soy-rich diet can lead to a lower incidence of bowel cancer. Broccoli, curry powder and orange peel are also known to contain chemopreventive agents, and compounds thought to be the active agents have been isolated and characterised.

The chemotherapy of cancer

467

9.18 ANTI-EMETICS The role of anti-emetics in cancer chemotherapy is of great importance since many of the cytotoxic agents in clinical use cause profound nausea and vomiting. Uncontrolled vomiting may outweigh the benefits of treatment and also lead to poor patient compliance. Traditionally, drugs such as metoclopramide have been used, but with the addition of the 5HT3 antagonists such as ondansetron and granisetron, the incidence and severity of emesis has been substantially reduced. FURTHER READING Bishop, J.M. and Weinberg, R.A. (eds.) (1996) Molecular Oncology. New York: Scientific American, Inc. Browne, M.J. and Thurlby, P.L. (eds.) (1996) Genomes, Molecular Biology and Drug Discovery. London: Academic Press Ltd. Culver, K.W., Vickers, T.M., Lamsam, J.L., Walling, H.W. and Seregina, T. (1995) Gene therapy of Solid Tumours. British Medical Bulletin 51, 192–204. Dalgleish, A.G. (1994) Viruses and Cancer. British Medical Bulletin 47, 21–46. Dobrusin, E.M. and Fry, D.W. (1992) Protein Tyrosine Kinases and Cancer. Annual Reports in Medicinal Chemistry 27, 169–178. Hochhauser, D. and Harris, A.L. (1991) Drug Resistance. British Medical Bulletin 47, 178–196. Larson, E.R. and Fischer, P.H. (1989) New Approaches to Antitumour Therapy. Annual Reports in Medicinal Chemistry 24, 121–128. Lee, M.D., Ellestead, G.A. and Borders, D.B. (1991) Calicheamicins: Discovery, Structure, Chemistry and Interaction with DNA. Accounts of Chemical Research 24, 235–243. Lemoine, N.R. and Cooper, D.N. (eds.) (1996) Gene Therapy. Oxford: BIOS Scientific Publishers Ltd. Macdonald, F. and Ford, C.H.J. (1997) Molecular Biology of Cancer. Oxford: BIOS Scientific Publishers Ltd. Miller, A.D. (1992) Human Gene Therapy Comes of Age. Nature (London) 357, 455– 460. Mulligan, G.C. (1993) The Basic Science of Gene Therapy. Science 260, 926–932. Nature (Supplement to issue 6604) (1996) Intelligent Drug Design 384, 1–26. Neidle, S.J. and Waring, M.J. (eds.) (1993) Molecular Aspects of Anticancer DrugDNA Interactions. London: The Macmillan Press. Pratt, W.B. and Ruddon, R.W. (eds.) (1979) The Anticancer Drugs. Oxford: Oxford University Press. Pullman, B. (1991) Sequence Specificity in the Binding of Antitumour Anthracyclines to DNA. Anti-cancer Drug Design 7, 95–105. Scientific American (Special Issue) (1996) What You Need to Know About Cancer 275, 4–167. Silverman, R.B. (1992) The Organic Chemistry of Drug Design and Drug Action. London: Academic Press Inc.

Introduction to the principles of drug design and action

468

Suffness, M. (1993) Taxol: From Discovery to Therapeutic Use. Annual Reports in Medicinal Chemistry 27, 305–314. Vousden, K.H. and Farrell, P.J. (1994) Viruses and Human Cancer. British Medical Bulletin 50, 560–581. Workman, P. (ed.) (1992) New Approaches in Cancer Pharmacology: Drug Design and Development. London: Springer-Verlag. Yarnold, J.R., Stralton, M. and McMillan, T.J. (eds.) (1996) 2nd Edition Molecular Biology for Oncologists. London: Chapman and Hall.

10. NEUROTRANSMITTERS, AGONISTS AND ANTAGONISTS ROBERT D.E.SEWELL, RICHARD A.GLENNON, MALGORZATA DUKAT, HOLGER STARK, WALTER SCHUNACK and PHILIP G.STRANGE CONTENTS 10.1 OVERVIEW

388

ROBERT D.E.SEWELL

388

10.1.1 Introduction

388

10.1.1.1 Neurotransmitters

388

10.1.1.2 Receptors

390

10.1.1.3 Neuronal signal effectors

391

FURTHER READING

391

10.2 SEROTONIN RECEPTORS AND LIGANDS

391

RICHARD A.GLENNON AND MALGORZATA DUKAT

391

10.2.1 Introduction

391

10.2.2 Subpopulations of 5-HT Receptors

392

10.2.3 Receptor populations and ligands

394

10.2.3.1 5-HT1A

395

10.2.3.2 5-HT1B

396

10.2.3.3 5-HT1D

398

10.2.3.4 5-HT1E

400

10.2.3.5 5-HT1F

400

10.2.3.6 5-HT2

400

10.2.3.7 5-HT3

403

Introduction to the principles of drug design and action

470

10.2.3.8 5-HT4

403

10.2.3.9 5-HT5, 5-HT6 and 5-HT7

407

10.2.4 Possible applications of 5-HT agonists and antagonists

407

10.2.5 Epilogue

408

FURTHER READING

408

10.3 HISTAMINE RECEPTORS

410

HOLGER STARK AND WALTER SCHUNACK

410

10.3.1 Introduction

410

10.3.2 H1-Receptors

411

10.3.3 H2-Receptors

415

10.3.4 H3-Receptors

417

FURTHER READING

421

10.4 DOPAMINE RECEPTORS

422

PHILIP G.STRANGE

422

10.4.1 Introduction

422

10.4.2 Dopamine agonists

423

10.4.2.1 Aminotetralins and related compounds

423

10.4.2.2 Aporphine alkaloids

425

10.4.2.3 Ergot alkaloids

425

10.4.2.4 Benzazepines

427

10.4.2.5 Miscellaneous structures

427

10.4.2.6 Selective agonists

428

10.4.3 Dopamine antagonists

429

FURTHER READING

432

Neurotransmititers, agonists and antagonists

471

10.1 OVERVIEW Robert D.E.Sewell 10.1.1 Introduction Over recent years, our concepts about neurotransmitters have been transformed firstly, because the number of putative transmitters has increased to upwards of 40 candidates and secondly, because our understanding of neurotransmitter function has widened to cover a more diverse range of effects. The aim of this chapter, is not to embrace the medicinal chemistry of all the neurotransmitters and their agonists and antagonists, but more to concentrate on three selected candidates, namely serotonin, histamine and dopamine. Amongst these three examples, there are receptor types which typify both of the major functional classes of neuronal receptor (i.e. ionotropic and metabotropic receptors). 10.1.1.1 Neurotransmitters Neurotransmitters are a varied assortment of substances implicated in the transfer of signals across chemical synapses. These neuronal elements consist of narrow clefts which separate presynaptic nerve terminals from receptors located on postsynaptic membranes found on subsequent neurones, muscles or glands. In addition to postsynaptic receptors, there are receptors sited presynaptically on the nerve terminals themselves. These are sometimes designated as autoreceptors; capable of responding to released neurotransmitter, providing a negative feedback function concerned with regulation of transmitter release. Other neuronal synapses exist which rely on electrical rather than chemical transmission and this represents a very fast mode of communication between cells. The chemical neurotransmitters (Figure 10.1) may be classed into three major categories: (1) Simple amino acids like glutamate, γ-amino butyric acid (GABA) and glycine account for transmission at a high proportion of CNS synapses. They are relatively fast communicators and occur at the highest levels of all the neurotransmitter groups in brain tissue (micromoles per gram of tissue). (2) The amine neurotransmitters are composed of ‘classical’ transmitters such as acetylcholine, serotonin (5-hydroxytryptamine—5-HT) histamine and the catecholamines (dopamine and noradrenaline). Also included in this group are purine neurotransmitters like adenosine and adenosine triphosphate (ATP). All the neurotransmitters in the group are found in moderate concentrations in neuronal tissue (nanomoles per gram of whole brain tissue) and tend to exert a slower modulatory type of action. (3) The neuropeptides occur at the lowest neuronal levels of all three groups (
Introduction to the principles of drug design and action

472

might be more appropriately considered as local neuromodulators.

Figure 10.1 Examples of neurotransmitters. It is not uncommon for neurones to contain more than one neurotransmitter, a typical example being neuropeptides occurring in combination with amines in one and the same neurone. 10.1.1.2 Receptors Neurotransmitter effects are mediated via receptors located in neuronal membranes and these are thought to be of two functional classes: Metabotropic receptors Metabotropic receptors are linked to intracellular proteins which transduce signals across the cell membrane. These proteins are known as G-proteins (so-called because they are coupled to the guanosine nucleotides GDP or GTP). They are comprised of three subunits (α, β, and γ) of which the α subunit possesses GTP-ase activity. The DNA sequences coding for the majority of G-protein coupled metabotropic receptors

Neurotransmititers, agonists and antagonists

473

do have some sequence homology so they are viewed as a superfamily and over 100 constituent members have been cloned. Binding of a neurotransmitter or an agonist with a metabotropic receptor often stimulates the formation of an intracellular second messenger by means of an effector enzyme. The second messengers include adenosine 3',5'-cyclic phosphate (c-AMP), inositol phosphates, diacylglycerol and arachidonate. The overall mechanism invariably involves an amplification process since a single neurotransmitter/agonistreceptor complex may activate several G-protein molecules in turn to generate many secondary messenger molecules intracellularly. There appear to be subtypes of virtually all metabotropic receptors which differ either in location and/or second messenger coupling, but they all possess seven trans-membrane hydrophobic spanning regions in their peptide sequences, with an extracellular amino terminus and an intracellular C-terminus. The membrane spanning portions of receptor proteins form α-helices and possess similarities within each receptor group so it is probable that the agonist binding site resides at least partly in the membrane spanning region. Ionotropic receptors Ionotropic receptors are linked directly to ion channels in the neuronal membrane. They are responsible for transient (submillisecond) increases in the conductance of specific ions in each instance and this gives rise to rapid synaptic transmission. There are several examples which include the nicotinic acetylcholine receptor, the 5-HT3 receptor and receptors for the amino acids such as GABAA, glycine and N-methyl-Daspartate (NMDA). Ionotropic receptors are comprised of 4–5 protein subunits in a complex linked to an ion channel. The basic structure of each subunit consists of a protein which loops in and out of the neuronal membrane. In the case of the nicotinic receptor, which has been comprehensively studied, there are 5 subunits (α2, β, γ, δ) each traversing the membrane a total of five times. Four of the spanning segments are hydrophobic in nature and are considered as truly integral transmembrane domains. The fifth segment has only one face of its α-helix structure which is hydrophobic while the other face is hydrophilic and, along with counterparts in the other 4 subunits, constitutes the lining of the ion channel interfacing with an aqueous environment. When an agonist binds to this pentameric receptor, a conformational change occurs in the complex which allows the passage of ions through the channel. 10.1.1.3 Neuronal signal effectors Both metabotropic and ionotropic receptors are coupled to effectors involving enzymes or ion channels and the modulation of cytoplasmic Ca2+ concentration features in several of these systems. Thus, ionotropic receptors can regulate cytoplasmic Ca2+ concentration either directly by conducting Ca2+ itself, or via permeability to monovalent cations which depolarise the membrane as Na+ influxes and this subsequently activates Ca2+ channels in the proximity of the receptor. Metabotropic receptors may activate slower biochemical effector processes which modulate Ca2+ concentration intracellularly through the following G-proteins:

Introduction to the principles of drug design and action

474

(1) Gs which catalyses the conversion of adenosine triphosphate (ATP) to cAMP by stimulating adenylate cyclase whilst concomitantly activating Ca2+ channels. (2) Gi/Go which inhibits both adenylate cyclase and Ca2+ channels whilst simultaneously activating K+ channels. (3) G11/Gq which activates phospholipase C to catalyse the hydrolysis of phosphatidylinositol biphosphate (PIP2) to 1,4,5-inositol triphosphate (IP3) and diacylglycerol (DAG). IP3 may cause Ca2+ release from such intracellular storage sites as the endoplasmic reticulum. DAG on the other hand, stimulates the enzyme protein kinase C (PKC) which is capable of phosphorylating other cellular or membrane proteins, enzymes and ion channels. If ion channel proteins are phosphorylated, this may alter ion fluxes albeit in a slower manner than that generated by ionotropic receptors.

FURTHER READING Cascieri, M.A., Fong, T.M. and Strader, C.D. (1995) Molecular characterization of a common binding site for small molecules within the transmembrane domain of Gprotein coupled receptors. Journal of Pharmacological and Toxicological Methods 33, 170–185. Gilman, A.G. (1987) G proteins: transducers of receptor-generated signals. Annual Review of Biochemistry 56, 615–649. Nicholls, D.G. (1994) Proteins, transmitters and synapses. Oxford: Blackwell Scientific Publications. Strader, C.A., Fong, T.M., Tota, M.R. and Underwood, D. (1994) Structure and function of G-protein-coupled receptors. Annual Review of Biochemistry 63, 101– 132. Strader, C.A., Fong, T.M., Graziano, M.P. and Tota, M.R. (1995) The family of Gprotein-coupled receptors. FASEB 9, 745–754. Unwin, N. (1993) Nicotinic acetylcholine receptor at 9A resolution. Journal of Molecular Biology 229, 1101–1124. Watson, S. and Girdlestone, D. (1995) Receptor and ion channel nomenclature. Supplement (1995) Trends in Pharmacological Sciences. 10.2 SEROTONIN RECEPTORS AND LIGANDS Richard A.Glennon and Malgorzata Dukat 10.2.1 Introduction Serotonin (5-hydroxytryptamine, 5-HT; 10.1) was discovered about 50 years ago and was the subject of extensive investigation during the 1950s and 1960s. Although it was

Neurotransmititers, agonists and antagonists

475

implicated in numerous physiological functions and pathological and psychopathological conditions, research was hampered by a lack of suitable pharmacological tools with which to study serotonergic function, and by problems associated with interpretation of pharmacological results which seemed to vary depending upon the particular pharmacological model being used or physiological response being monitored. During the late 1960s and into the 1970s, the pace of 5-HT research slowed to a crawl. With the advent of radioligand binding techniques, particularly the rapid filtration method, and with the development of new psychopharmacological assay techniques and methodologies, there was renewed interest in centrally acting agents and neurotransmitters, in general, and in 5-HT in particular. Interest in 5-HT was further heightened with the discovery in late 1979 of two different populations (i.e., 5-HT1 and 5-HT2) of central 5-HT binding/receptor sites. For the first time, it was thought possible to explain some of the confounding results reported in the years before; that is, different populations of 5-HT receptors might mediate different responses and, further, might behave differently toward various serotonergic agents. Studies during the 1980s and into the 1990s focussed broadly on, (i) discovery and characterization of additional populations of 5-HT receptors, (ii) identification of the functional significance of these receptors, and (iii) development of novel subtypeselective agonists and antagonists (Herndon and Glennon 1993). Radioligand binding techniques accounted for the early discovery of several new populations of 5-HT receptors; indeed, the number of proposed populations of sites began to increase at an alarming pace. Questions were raised concerning whether so many different populations of 5-HT receptors could actually exist, or whether they simply represented tissue or species homologues, minor variants of known 5-HT receptors, or artifacts resulting from the selectivity (or, more accurately, the non-selectivity) characteristics of the serotonergic agents being used to investigate the receptors. Subsequently, molecular biology confirmed that many of these populations represented different receptors and that certain others were species homologs. Such studies also resulted in the discovery of additional populations of sites. 10.2.2 Subpopulations of 5-HT receptors At least 15 populations of 5-HT receptors have now been cloned (Table 10.1). Transmembrane homology amongst these populations ranges from about 30% to

Introduction to the principles of drug design and action

476

>90%. As might be expected, low homology is associated with distantly related populations of 5-HT receptors whereas high homology is typically found between 5HT receptors that display similar pharmacological profiles and second messenger systems. Likewise, species

Table 10.1 5-HT Receptor populations, second messengers and useful agents. Family Second Selective Selective subtype messengera agonistb antagonistb 5-HT1 5-HT1A (−)cAMP 8-OH DPAT WAY-100,135 (10.3) (10.9) — 5-HT1B (−)cAMP CP-93,129 (10.11) 5GR 127935c (−)cAMP Sumatriptanc HT1Dα (10.12) (10.17) 5GR 127935c (−)cAMP Sumatriptanc HT1Dβ (10.12) (10.17) (−)cAMP — 5-HT1E 5-HT1F (−)cAMP — 5-HT2 5-HT2A IP3 DOXd (10.19) Ketanserine (10.18) IP3 DOXd (10.19) SB 204741 5-HT2B (10.25) d 5-HT2C IP3 DOX (10.19) Ketanserine (10.18) 5-HT3 5-HT3 Ion channel mCPBGf (10.28) Zacopride 2-Me 5-HT (10.32) Tropisetron (10.29) (10.35) Ondansetron (10.37) 5-HT4 5-HT4 (+)cAMP SC 53116 SB 204070 (10.38) (10.39) SB 207710 (10.40) GR 113808 (10.42)

Neurotransmititers, agonists and antagonists

477

5-HT5 5-HT5A 5-HT5B

? ?

— —

— —

5-HT6

(+)cAMP

—

—

5-HT6 5-HT7 5-HT7 (+)cAMP — — 5-HT1 receptors are negatively coupled to adenylate cyclase [(−)cAMP] whereas certain other receptors are positively coupled [(+)cAMP]. The 5-HT2 family of receptors is coupled to phosphoinositide hydrolysis (IP3). b See text for discussion and additional examples. These agents are not necessarily specific for a given population of receptors but are merely useful agents that display some selectivity. The (−) symbol indicates that no selective agents have yet been identified. c No high-affinity agents have yet been found to discriminate between the two subpopulations. d DOX refers to 1-(2,5-dimethoxy-4-X-phenyl)-2-aminopropane where X=−CH3 (DOM), −Br (DOB), or −I (DOI). e Ketanserin binds with lower affinity at 5-HT2B receptors than at 5HT2A/2C receptors. f mCPBG=metachlorophenylbiguanide. a

homologues usually display >90% homology. Populations of 5-HT receptors belong both to the G-protein and ion channel superfamilies of neurotransmitter receptors; the former category can be further categorized according to their second messenger coupling. Table 10.1 shows some of the best studied and more recent populations of 5HT receptors and their second messenger systems. 5-HT receptor nomenclature can be rather bewildering; fortunately, this nomenclature has been recently revised (Martin and Humphrey 1994; Hoyer et al. 1994). Nevertheless, older literature must be read very carefully. For example, 5-HT1B receptors were once thought to exist in various animal species, including human; the term “5-HT1B” now refers only to rodent 5-HT1B receptors, whereas humans possess 5-HT1D receptors in corresponding anatomical locations. In fact, two different populations of 5-HT1D receptors have now been identified: 5-HT1Dα and 5-HT1Dβ; 5HT1Dβ receptors appear to be the human counterpart (i.e., species homolog) of rodent 5-HT1B receptors. An even more complex situation exists with the 5-HT2 family of receptors. Because 5-HT1C receptors were found to display greater sequence homology, similar second messenger coupling, and related pharmacology to what were once termed 5-HT2 receptors, they have been renamed 5-HT2C receptors. Thus, the term “5-HT1C” is no longer used. Correspondingly, and for purpose of distinction, the original 5-HT2 receptors have been renamed 5-HT2A receptors. Today, the term “5HT2” is used only to refer to the 5-HT2 family of receptors. For several years, there

Introduction to the principles of drug design and action

478

was no 5-HT2B receptor population; 5-HT2F receptors were renamed 5-HT2B receptors shortly after their initial characterization to fill the obvious gap. To make matters even more confusing, 5-HT2 (now 5-HT2A) receptors, have been demonstrated to exist both in a high-affinity (5-HT2H) and low-affinity (5-HT2L) state. During a period of controversy, when it was thought that these two affinity states might represent distinct populations of sites, the terms 5-HT2A and 5-HT2B were introduced to describe 5-HT2H and 5-HT2L binding behavior. Thus, certain terms, such as 5-HT2A receptors for example, have different meanings or refer to different populations of receptors depending upon their chronological occurrence in the literature. As a further note of caution, care must be given to what constitutes a site-selective agent. Certain agents were once considered quite selective for a given population of receptors. However, with the proliferation of 5-HT (and other neurotransmitter) receptor populations, and with more extensive investigation of older agents, it is now recognized that many of these agents are considerably less selective than once supposed. Table 10.1 lists some of the agents typically used to investigate the various populations of 5-HT receptors; although these agents are listed as being “selective” agonists or antagonists, no claims are made, or intended, as to their absolute selectivity. Rather, these are merely relatively selective or semi-selective agents that display some selectivity for that particular population of receptors versus most (but not necessarily all) other populations of 5-HT receptors. The interested reader is referred to the primary literature to obtain more information about the selectivity of a specific agent; for comparative binding data on ligands at various populations of 5-HT receptors, see Hoyer et al. (1994) and Ziffa and Fillion (1992). The present chapter will focus primarily on some of the more widely used or standard serotonergic agents, and on some of the more recent agents. For the most part, the earlier work, including the medicinal chemistry and structure-activity relationships of serotonergic agents (Glennon et al. 1991; Glennon and Dukat 1993; King 1994), has been already reviewed and further information about uncited material can be found in these review articles. Serotonergic signal transduction pathways have also been reviewed. 10.2.3 Receptor populations and ligands Not so long ago, it was thought that the chemical class to which a serotonergic agent belonged might be readily related to selectivity. For example, indolealkylamines, like 5-HT itself, were considered to be nonselective agents, aminotetralins were 5-HT1A ligands (and, more specifically, 5-HT1A agonists), arylpiperazines were 5-HT1B agonists, and benzoate esters and benzamides were 5-HT3 antagonists. None of these generalizations has held up over time. Most serotonergic agents can be conveniently categorized into one of several different large chemical families (Glennon and Dukat 1993); however, few (if any) of these families contain only agents that are exclusively agonists or antagonists for a specific population of 5-HT receptors. Selectivity, and functional activity, seem more controlled by pendent substituent groups rather than by the chemical class to which the agent belongs. Subtle structural changes can result in significantly different pharmacological consequences. Some of the general structureactivity and structure-affinity relationships (SAR and SAFIR) of various chemical

Neurotransmititers, agonists and antagonists

479

classes will be discussed. Because space does not allow a detailed discussion of these relationships, only selected examples will be provided. More extensive reviews of this topic can be found in the literature (e.g. Glennon et al. 1991; Glennon and Dukat 1993). 10.2.3.1 5-HT1A Because 5-HT1A receptors were one of the first to be identified, there is considerable information available on 5-HT1A ligands. 5-Carboxamidotryptamine (5-CAT; 10.2) has seen application as a 5-HT1A agonist; although this agent is non-selective, it was once used to classify a receptor population as belonging to the 5-HT1 family. Interestingly, this generality is no longer valid in that although 5-CAT binds at most populations of 5-HT1 receptors, it displays low affinity for 5-HT1E receptors; 5-CAT also binds at certain other populations of non-5-HT1, 5-HT receptors. Nevertheless, it is still a relatively widely used agent. Two of the most important structure types include the aminotetralins and the arylpiperazines. The standard 5-HT1A agonist, 8-OH DPAT or 8-hydroxy-2-(dipropylamino)tetralin (10.3), remains one of the more selective serotonergic agents available; its selectivity is likely related to the observation that the intact indole nucleus of 5-HT is not a requirement for activation of 5-HT1A receptors. Although racemic 8-OH DPAT is the most commonly used form of the agent, it has been demonstrated that R(+)8-OH DPAT is a full agonist whereas its S(−)-enantiomer is a partial agonist. Incorporation of a 5-fluoro group further shifts the functional activity of this compound; S(−)5-F 8-OH DPAT or S(−)UH-301 is a potent 5-HT1A antagonist. Tritiated 8-OH DPAT is the radioligand of choice for binding studies. No single chemical class of agents has been as extensively investigated for serotonergic activity as the arylpiperazines (reviewed Glennon and Dukat 1993). N1(Aryl)piperazines can be divided into two broad classes: the N4-unsubstituted or shortchain arylpiperazines, and the long-chain arylpiperazines (LCAPs). Short-chain, and especially N4-unsubstituted arylpiperazines are notoriously non-selective serotonergic agents with different binding profiles and functional activities; in fact, certain agents may be agonists or partial agonists at one receptor population and antagonists at another. The long-chain arylpiperazines seem much more selective for 5-HT1A versus other populations of receptors, and selected examples bind at 5-HT1A receptors with Ki values as high as 0.1 nM. It might be noted, however, depending upon the particular agent, that certain of these agents also bind with high affinity at dopamine, αadrenergic, and/or at members of the 5-HT2 family of receptors. The general structure of the arylpiperazines is given by: Ar-PIPERAZINE-(CH2)n-Terminus. The aryl (Ar) group can vary widely amongst phenyl, substituted phenyl, naphthyl, and heteroaryl. The terminus is usually an amide (or imide) or an aromatic group, and there appears to be considerable bulk tolerance. Thus, different combinations of functionalities can result in a vast array of structures that retain affinity for 5-HT1A receptors. With regard to n, two to four methylene groups appear optimal. However, the length of this chain can influence selectivity and, furthermore, the nature of the amide substituent can influence optimal chain length. For example, n=4 is optimal when the

Introduction to the principles of drug design and action

480

terminus is a heteroarylamide, but when the terminus is an alkylamide, n=2 is optimal. Electronic distribution in the amide region also plays a role in 5-HT1A affinity. Typical examples of this class of agents include the agonists or partial agonists buspirone (10.4), gepirone (10.5), and ipsapirone (10.6); several very low efficacy partial agonists, including BMY 53857 (10.7) and NAN-190 (10.8), have also been reported and have been widely used as 5-HT1A antagonists, but only WAY-100,135 (10.9) is considered a silent antagonists. A final, although less well investigated, class of agents that has been explored is the aryloxyalkylamines, such as β-adrenergic antagonists propranolol (10.10) and pindolol. Though these agents bind at 5-HT1B and/or β-adrenergic receptors with greater affinity than they display for 5-HT1A receptors, they were among the first compounds shown to behave as 5-HT1A anatagonists. Interest with these agents continues (Langlois et al. 1993), and mechanistic studies that have implicated βadrenergic involvement for a particular action solely on the basis of antagonism by propranolol or pindolol may be in need of re-investigation. 10.2.3.2 5-HT1B Arylpiperazines were once the mainstay of 5-HT1B receptor research because they were considered selective for this population. With the realization that these agents are far less

Neurotransmititers, agonists and antagonists

481

Introduction to the principles of drug design and action

482

Neurotransmititers, agonists and antagonists

483

selective than originally reported, their use has diminshed. Another agent that played a pivotal role in 5-HT1B research was RU-24969; however, this agent displays only several-fold selectivity for 5-HT1B versus 5-HT1A receptors. A newer series of agents developed by Pfizer, e.g. CP-93,129 (10.11), is related to RU-24969 but seem to be somewhat more selective. Interest in 5-HT1B receptors dampened once it was shown that rodent 5-HT1B receptors are structurally distinct from their species homolog, the human 5-HT1Dβ receptors. These two receptor populations display a high (>90%) degree of homology and most agents that bind at 5-HT1B receptors also bind at 5HT1Dβ receptors. As a consequence, attention was refocused on the latter population. Interestingly, aryloxyalkylamines such as propranolol are amongst the few agents that can differentiate between these two populations of receptors in that they bind with high affinity at 5-HT1B receptors, but with >100-fold lower affinity at 5-HT1Dβ receptors. 10.2.3.3 5-HT1D Sumatriptan (10.12) is considered a prototypical 5-HT1D agonist. Structure-affinity relationships for the binding of 5-HT1D ligands at bovine receptors were described prior to cloning of human 5-HT1Dα and 5-HT1Dβ receptors. Much less is known about the SAFIR for binding at human 5-HT1D receptors. But, several new agents have been recently introduced including an oxadiazole (10.13), carbazole (10.14), and the 5(pyridylamino)indole (10.15). One of the problems facing sumatriptan and many other 5-HT1D agonists is their high affinity for 5-HT1A receptors; 10.12–10.15 typically display <50-fold 5-HT1D selectivity. ALX-1323 (NOT; 10.16) is a new agonist with high affinity and >300-fold selectivity for 5-HT1Dβ receptors relative to 5-HT1A receptors. Recently, a novel series of 5-HT1D antagonists has been reported; typical examples include piperazines (10.17), where R=5-methylisoxazol-3-yl and 2-methyl4-(N,N-dimethylcarboxamido)phenyl (Clitherow et al. 1994).

Introduction to the principles of drug design and action

484

Neurotransmititers, agonists and antagonists

485

10.2.3.4 5-HT1E Human 5-HT1E receptors were first decribed in brain homogenates six years ago using radioligand binding techniques. The receptors were not fully characterized at that time but it was demonstrated that even minor molecular modification of the 5-HT structure resulted in reduced affinity. For example, O-methyl 5-HT and 5-CAT (10.2), agents typically binding with high affinity at most populations of 5-HT1 receptors, displayed low affinity for these receptors. Several years later, several groups independently cloned 5-HT1E receptors and demonstrated binding profiles similar to those which had been reported earlier. Methiothepin serves as a nonselective 5-HT1E antagonist. At this time, no 5-HT1E selective agents have been reported. 10.2.3.5 5-HT1F This is the most recent population of human 5-HT1 receptors to be identified. No selective agonists or antagonists are currently available. 10.2.3.6 5-HT2 The first two populations of 5-HT receptors to be identified were the 5-HT1 and 5-HT2 receptors. It is now recognized that these populations actually consist of subpopulations. However, although this was realized early on for 5-HT1 receptors, recognition of subpopulations of 5-HT2 receptors did not occur until the late 1980s and was not fully appreciated until into the 1990s. Consequently, much of the early work on “5-HT2” receptors may in fact reflect results that can now be dissociated into 5HT2A, 5-HT2B, and 5-HT2C. Nevertheless, it is only very recently that work has been initiated on attempting to develop agents with selectivity for 5-HT2 receptor subpopulations. Much of the early work on 5-HT2 agonists and antagonists was previously reviewed (Herndon and Glennon 1993; Glennon et al. 1991). Ketanserin (10.18) was one of the first, and is still one of the most widely used, 5-HT2 antagonists. Tritiated ketanserin has been the radioligand of choice for investigating 5-HT2 receptors, but it appears that ketanserin binds with lower affinity at 5-HT2B receptors than it displays for 5HT2A or 5-HT2C receptors. Additionally, ketanserin has been variously reported to bind with as little as 2-fold to as much as 140-fold selectively for 5-HT2A versus 5-HT2C receptors. Various other 5-HT2 antagonists have been described. Additional SAFIR and binding hypotheses have been suggested to account for the binding of various antagonists at 5-HT2A receptors. α-Methylserotonin, although not particularly selective, has been employed as a 5-HT2 agonist. The DOX series of compounds represent another widely used group of 5-HT2 agonists; these are typified by DOB and DOI (10.19, where X=Br and I, respectively). [125I] DOI is available for use as a radioligand. In certain functional

Introduction to the principles of drug design and action

486

Neurotransmititers, agonists and antagonists

487

studies, 1-(3-chlorophenyl)piperazine (mCPP; 10.20) has been reported to be a 5-HT2C (or 5-HT2B/2C) agonist but a 5-HT2A antagonist; however, it also binds at other populations of 5-HT receptors. With the recent reclassification of 5-HT2 receptors has come attempts to develop new agents with subtype selectivity. Spiperone (10.21) and AMI-193 (10.22) display >1,000-fold selectivity for 5-HT2A versus 5-HT2C receptors. MDL 100,907 (10.23) has also been reported to bind with 200-fold higher affinity at 5-HT2A versus 5-HT2C receptors. However, these results were reported prior to the discovery of 5-HT2B receptors. SB 200646A (10.24) was the first reported 5-HT2C-selective antagonist, but was later

Introduction to the principles of drug design and action

488

shown to also be a 5-HT2B antagonist; structural modification subsequently led to the development of the 5-HT2B-selective antagonist SB 204741 (10.25). SDZ SER-082 (10.26) and SB 206553 (10.27) are other examples of a 5-HT2B/2C versus 5-HT2Aselective antagonists. Thus, even though there are a great number of agents that have been termed “5-HT2 antagonists”, the identification of subtypes of 5-HT2 receptors has initiated a search for more selective agents. Even while this search progresses, molecular biological and functional studies have identified new species homologs of 5-HT2 receptors and have also raised the possibility of additional members of the 5HT2 family of receptors.

10.2.3.7 5-HT3 The 5-HT3 population of receptors was first studied using isolated peripheral tissue preparations and it was several years before a suitable radioligand was identified and 5-HT3 receptors were characterized in the brain. Two of the most commonly used 5HT3 agonists are (3-chlorophenyl)biguanide (mCPBG; 10.28) and 2-methyl 5-HT (10.29). The N,N,N-trimethyl quaternary salt of 5-HT (5-HTQ; 10.30) also seems to be a selective 5-HT3 agonist. The SAR of 5-HT3 agonists has not been well investigated. In contrast, numerous 5-HT3 antagonists have been reported and a detailed discussion of their SAR, although beyond the scope of this chapter, has recently appeared (King 1994). The first useful 5-HT3 antagonist, MDL 72222 (10.31), resulted from the observations that metoclopramide and cocaine are weak 5-HT3 antagonists. MDL 72222 was subsequently shown to possess those basic features important for 5-HT3 antagonist activity i.e. arylcarbonyl linker-basic side chain. Many of the early agents were aryl-substituted benzoate esters and benzamides, but structurally related agents were also developed. Some of the older and more widely used 5-HT3 antagonists include: zacopride (10.32), renzapride (10.33), zatosetron (LY 277359; 10.34), tropisetron (ICS 205–930; 10.35), granisetron (BRL 43694; 10.36), and ondansetron (10.37). A significant amount of structural latitude is permitted, particularly in the basic side chain. This has resulted in the development of hundreds of 5-HT3 antagonists. Tritiated tropisetron, zacopride, granisetron, and related compounds have been used in radioligand binding studies.

Neurotransmititers, agonists and antagonists

489

10.2.3.8 5-HT4 A population of receptors originally identified in primary cell cultures of mouse embryo colliculi, and subsequently investigated in detail using peripheral functional assays, has been shown to exist in the brain and has been termed 5-HT4. 5-HT4 receptors were very recently cloned and display <50% homology in their transmembrane domains with other

Introduction to the principles of drug design and action

490

5-HT receptors. Two isoforms or splice variants, 5-HT4L and 5-HT4S, that vary only in the length and sequence of their C-terminal chain, have been identified. Many 5-HT3 ligands had been shown to be structurally unique relative to other serotonergic agents;

Neurotransmititers, agonists and antagonists

491

this was probably not unusual given that 5-HT3 receptors are the only population of ion channel receptors within the 5-HT family. Interestingly, however, the first agents used to investigate 5-HT4 receptors were some of these same agents, even though 5HT4

receptors represent G-protein receptors. Certain 5-HT3 agonists were found to behave as 5-HT4 antagonists whereas others even acted as (partial) agonists. More selective agents have now been identified including the agonist SC 53116 (10.38), and the antagonists SB 204 070 (10.39), SB 207 710 (10.40), RS-23597–190 (10.41), GR 113808 (10.42), and SDZ 205,557 (10.43). [3H]GR 113808 and [125I]SB 207 710 have been used as radioligands. A 5-HT4 agonist pharmacophore model has been proposed and new agonists, derivatives of carbazimid-amide (10.44) have been synthesized on the basis of this model.

Introduction to the principles of drug design and action

492

Neurotransmititers, agonists and antagonists

493

Introduction to the principles of drug design and action

494

10.2.3.9 5-HT5, 5-HT6, and 5-HT7 These three populations represent the newest families of 5-HT receptors to be identified; (see Glennon and Dukat (1995) and Lucas and Hen (1995) and references therein for a recent overview). Both a rat 5-HT5A and 5-HT5B receptor, but only the human 5-HT5A receptor, have been cloned. Rat and human 5-HT6 as well as human, rat, mouse, and hamster 5-HT7 receptors have been cloned. To date, no selective ligands have been identified. Interestingly, the 5-HT1 ligand 5-CAT (10.2) binds at 5HT7 receptors with high affinity and the 5-HT1A agonist 8-OH DPAT (10.3) also binds at 5-HT7 receptors. Certain typical and atypical neuroleptic agents and tricyclic antidepressants bind at 5-HT6 and 5-HT7 receptors suggesting that these receptor populations may play a role in the mechanism of action of psychotherapeutic agents. 10.2.4 Possible applications of 5-HT agonists and antagonists Experimental reports suggest that 5-HT receptors in the CNS may be implicated in the control of several basic physiological functions and behaviours which include food intake, thermoregulation, sexual behaviour, aggression, panic attacks as well as sleep, circadian rhythm, cognitive processes, cardiovascular disorders and schizophrenia (see Glennon and Dukat 1995). Agonists and antagonists for 5-HT receptor sub-types may therefore represent an array of possible therapeutic targets for the medicinal chemist (see Leonard 1994). Thus, it has been proposed that agonist stimulation of 5-HT1A receptors may suppress central 5-HT activity to increase the feeding response and this might have potential in the drug treatment of anorexia nervosa. 5-HT2 agonists, in contrast, have the opposite effect in that they suppress appetite in a 5-HT2 antagonist reversible manner. Moreover, 5-HT1A agonists tend to induce hypothermia whereas 5HT1B and 5-HT2 stimulation produces hyperthermia in animal studies. Even though investigations in humans suggest that 5-HT2 receptors are perturbed during depression and this malfunction tends to be normalized during effective antidepressant therapy in some patients, the most frequent current treatment relating to these effects, involves indirect receptor stimualtion via inhibition of neuronal reuptake of 5-HT itself. Both experimental and clinical data support the view that 5-HT1A agonists or partial agonists (e.g. buspirone), in addition to 5-HT2 and 5-HT3 antagonists, may have potential as drug treatments for anxiety. Though there are few reports that 5-HT3 antagonism is effective in the treatment of schizophrenia, it has been proposed that antagonists at 5-HT3 receptors may be valuable in the management or treatment of drug abuse in humans. Migraine is largely thought to be due to vasodilatation in brain tissue probably in association with localised inflammation. In this context, 5-HT1D receptors are located on both cerebral and extracerebral blood vessels and are stimulated by the agonist sumatriptan to include vascular constriction and reduced release of inflammatory mediators. This mechanism is thought to be responsible not only for migraine prophylaxis but also in the alleviation of the headache and nausea associated with a migraine attack once it has been established. Other clinical applications for 5-HT receptor ligands include: antagonists at 5-HT3 (eg. ondansetron) which are employed to treat chemotherapy-induced emesis and certain newer antidepressants (eg. trazodone and mianserin) which may share a

Neurotransmititers, agonists and antagonists

495

common 5-HT2 mechanism of action. Moreover, 5-HT2A agonists (eg. DOM and DOB) possess hallucinogenic properties. 10.2.5 Epilogue At least 15 different types of 5-HT receptors (Table 10.1), and numerous species homologs, have now been cloned. Several other, less well-investigated populations of 5-HT receptors have also been described in the literature. This has become an enormous challenge for medicinal chemists who have attempted to formulate and understand the SAR and SAFIR of various agents in order to develop newer compounds with subtype selectivity. This chapter, though not comprehensive, has described some of the most useful agents having evolved to date. In the past, serendipity accounted for the discovery of some of the most useful compounds in the armamentarium of serotonergic agents (including, for example, 8-OH DPAT and ketanserin). Futhermore, the design of selective agents was often hampered by confounding pharmacological results stemming from interference by populations of 5HT receptors that had yet to be discovered. With advances in molecular biology, the identification and characterization of novel receptor populations, and an increased awareness of SAR and SAFIR, newer agents are being designed with greater rationale. The availability of the amino acid sequences of 5-HT receptors populations from molecular biological studies is spurring molecular modeling investigations of the receptors themselves. Increasingly, these graphics models, coupled with site directed mutagenesis and the investigation of chimeric receptors, are being employed to explain the binding of ligands to 5-HT receptors, by identifying or implicating specific amino acid residues that may be contributing to binding interactions, and to aid the design of newer agents. The ever increasing knowledge of receptor populations and the results of studies described above, bode well for things to come and should eventually allow the design of high-affinity site-selective agents. FURTHER READING Adham, N., Kao, H.-T., Schechter, L.E., Bard, J., Olsen, M. and Urquhart, D. et al. (1993) Cloning of another human serotonin receptor (5-HT1F): A fifth 5-HT1 receptor subtype coupled to the inhibition of adenylate cyclase. Proceeding of the National Academy of Sciences of the U.S.A. 90, 408–412. Aghajanian, G.K. (1995) Electrophysiology of serotonin receptor subtypes and signal transduction pathways. In Psychopharmacology: The Fourth Generation of Progress, edited by F.E.Bloom and D.J.Kupfer, pp. 451–460. New York: Raven Press. Andersen, K., Liljefors, T., Gundertofte, K., Perregaard, J. and Bogeso, K.P. (1994) Development of a receptor interaction model for serotonin 5-HT2 receptor antagonists. Predicting selectivity with respect to dopamine D2 receptors. Journal of Medicinal Chemistry 37, 950–962. Boess, F.G. and Martin, I.L. (1994) Molecular biology of 5-HT receptors. Neuropharmacology 33, 275–317.

Introduction to the principles of drug design and action

496

Bonhaus, D.W., Bach, C., DeSouza, A., Salazar, F.H.R., Matsuoka, B.D. and Zuppan, P. et al. (1995) The pharmacology and distribution of human 5-hydroxytryptamine2B (5-HT2B) receptor gene products: comparison with 5-HT2A and 5-HT2C receptors. British Journal of Pharmacology 115, 622–628. Briejer, M.R., Akkermans, L.M.A., Lefebvre, R.A. and Schuurkes, J.A.J. (1995) Novel 5-HT2-like receptor mediates neurogenic relaxation of the guinea pig proximal colon. European Journal of Pharmacology 279, 123–133. Buchheit, K.-H., Gamse, R., Giger, R., Klein, F., Kloppner, E. and Pfannkuche, H.-J. et al. (1995) The serotonin 5-HT4 receptor 1. Design of a new class of agonists and receptor map of the agonist recognition site. Journal of Medicinal Chemistry 38, 2326–2330. Choudhary, M.S., Sachs, N., Uluer, A., Glennon, R.A., Westkaemper, R.B. and Roth, B.L. (1995) Differential ergoline and ergopeptine binding to 5-hydroxytryptamine 2A receptors: Ergolines require an aromatic residue at position 340 for high affinity binding. Molecular Pharmacology 47, 450–457. Clitherow, J.W., Scopes, D.I.C., Skingle, M., Jordan, C.C., Feniuk, W. and Campbell, I.B. et al. (1994) Evolution of a novel series of [(N,N-dimethylamino)propyl]- and piperazinylbenzanilides as the first selective 5-HT1D antagonists. Journal of Medicinal Chemistry 37, 2253–2257. Forbes, I.T., Ham, P., Booth, D.H., Martin, R.T., Thompson, M. and Baxter, G.S. et al. (1995) 5-Methyl-1-(3-3-pyridylcarbamoyl)-1,2,3,5-tetra-hydropyrrolo[2,3-f] indole: A novel 5-HT2C/5-HT2B receptor antagonist with improved affinity, selectivity, and oral activity. Journal of Medicinal Chemistry 38, 2524–2530. Forbes, I.T., Jones, G.E. and Murphy, O. (1995) N-(1-Methyl-5-indolyl)-N’-(3methyl-5-isothiazolyl)urea: A novel, high-affinity 5-HT2B receptor antagonist. Journal of Medicinal Chemistry 38, 855–857. Forbes, I.T., Kennett, G.A., Gadre, A., Ham, P., Hayward, C.J. and Martin, R.T. et al. (1993) N-(1-Methyl-5-indolyl)-N’-(3-pyridyl)urea: The first selective 5-HT1C antagonist. Journal of Medicinal Chemistry 36, 1104–1107. Gerald, C., Adham, N., Kao, H.-T., Olsen, M.A., Laz, T.M. and Schechter, L.E. et al. (1995) The 5-HT4 receptor: molecular cloning and the pharmacological characterization of two splice variants. EMBO Journal 14, 2806–2815. Glennon, R.A. and Dukat, M. (1993) 5-HT Receptor ligands—Update 1992. Current Drugs: Serotonin 1, 1–45. Glennon, R.A. and Dukat, M. (1995) Serotonin receptor subtypes. In Psychopharmacology: The Fourth Generation of Progress, edited by F.E.Bloom and D.J.Kupfer, pp. 415–429. New York: Raven Press. Glennon, R.A. and Westkaemper, R.A. (1993) 5-HT1D receptors: A serotonin receptor population for the 1990s. Drug News Perspectives 6, 390–405. Glennon, R.A., Hong, S.-S., Dukat, M., Teitlek, M. and Davis, K. (1994) 5(Nonyloxy)tryptamine: A novel high-affinity 5-HT1Dβ serotonin receptor agonist. Journal of Medicinal Chemistry 37, 2828–2830. Glennon, R.A., Ismaiel, A.M., Chaurasia, C. and Titeler, M. (1991) 5-HT1D serotonin receptors: Results of a structure-affinity investigation. Drug Development Research 22, 25–36.

Neurotransmititers, agonists and antagonists

497

Glennon, R.A., Westkaemper, R.B. and Bartyzel, P. (1991) Medicinal chemistry of serotonergic agents. In Serotonin Receptors Subtypes, edited by S.Peroutka, pp. 19– 64. New York: Wiley-Liss. Herndon, J.L. and Glennon, R.A. (1993) Serotonin receptors, agents, and actions. In Drug Design for Neuroscience, edited by A.P.Kozikowski, pp. 167–212. New York: Raven Press. Hoyer, D., Clarke, D.E., Fozard, J.R., Hartig, P.R., Martin, G.R., Mylecharane, E.J., et al. (1994) International Union of Pharmacology classification of receptors for 5hydroxytryptamine (serotonin). Pharmacology Reviews 46, 157–203. Ismaiel, A.M., Arruda, K., Teitler, M. and Glennon, R.A. (1995) Ketanserin analogues: The effect of structural modification on 5-HT2 serotonin receptor binding. Journal of Medicinal Chemistry 38, 1196–1202. Ismaiel, A.M., De Los Angeles, J., Teitler, M., Ingher, S. and Glennon, R.A. (1993) Antagonism of the 1-(2,5-dimethoxy-4-methylphenyl)-2-aminopropane stimulus with a newly identified 5-HT2 versus 5-HT1C-selective antagonist. Journal of Medicinal Chemistry 36, 2519–2525. Johnson, M.P., Bacz, M., Kursar, J.D. and Nelson, D.L. (1995) Species differences in 5-HT2A receptors: cloned pig and rhesus monkey 5-HT2A receptors reveal conserved transmembrane homology to the human rather than the rat sequence. Biochemica et Biophysica Acta 1236, 201–206. King, F.D., Brown, A.M., Gaster, L.M., Kaumann, A.J., Medhurst, A.D., Parker, S.G. et al. (1993) (±)3-Amino-6-carboxamido-1,2,3,4-tetrahydrocarbazole: A conformationally restricted analogue of 5-carboxamidotryptamine with selectivity for the 5-HT1D serotonin receptor. Journal of Medicinal Chemistry 36, 1918–1919. King, F.D., Jones, B.J. and Sanger, G.J. (1994) 5-Hydroxytryptamine-3 Receptor Antagonists. Boca Raton: CRC Press. King, F.D. (1994) Structure activity relationships of 5-HT3 receptor antagonists. In 5Hydroxy tryptamine-3 Receptor Antagonists, edited by F.D.King, B.J.Jones and G.J.Sanger, p. 1–44. Boca Raton: CRC Press. Kuipers, W., van Winjngaarden, I., Kruse, C.G., ter Horst-van Amstel, M., Tulp, M.Th.M. and Ijzerman, A.P. (1995) N4-Unsubstituted N1-arylpiperazines as highaffinity 5-HT1A receptor ligands. Journal of Medicinal Chemistry 38, 1942–1954. Langlois, M., Bremont, B., Rouselle, D. and Gaudy, F. (1993) Structural analysis by the comparative molecular field analysis method of the affinity of β-adrenoceptor blocking agents for the 5-HT1A and 5-HT1B receptors. European Journal of Pharmacology (Molecular Pharmacology Section) 244, 77–87. Leff, P. and Martin, G.R. (1988) The classification of 5-hydroxytryptamine receptors. Medical Research Reviews 8, 187–202. Leonard, B.E. (1994) Serotonin receptors—where are they going? International Clinical Psychopharmacology 9(Suppl.), 7–17. Leonhardt, S., Herrick-Davis, K. and Titeler, M. (1989) Detection of a novel 5-HT receptor subtype (5-HT1E) in human brain: interaction with GTP-binding protein. Journal of Neurochemistry 53, 465–471. Lucas, J.J. and Hen, R. (1995) New players in the 5-HT receptor field: genes and knockouts. Trends in Pharmacological Sciences 16, 246–252.

Introduction to the principles of drug design and action

498

Macor, J.E., Blank, D.H., Fox, C.B., Lebel, L.A., Newman, M.E. and Post, R.J. et al. (1994) 5-[(3-Nitropyrid-2-yl)amino]indoles: Novel serotonin agonists with selectivity for the 5-HT1D receptor. Variation of the C3 substituent on the indole template leads to increased 5-HT1D receptor selectivity. Journal of Medicinal Chemistry 37, 2509–2512. Martin, G.R. and Humphrey, P.P.A. (1994) Receptors for 5-hydroxytryptamine: Current perspectives on classification and nomenclature. Neuropharmacology 33, 261–273. Misztal, S., Bojarski, A., Mackowiak, M., Boksa, J., Mokrosz, J.L. (1992) Effect of the terminal amide fragment on 5-HT1A and 5-HT2 receptor affinity for N-[3-(4aryl-1-piperazinyl)propyl] derivatives of 3,4-dihydroquinolin-2(1H)-one and its isomeric isoquinolones. Medicinal Chemistry Research 2, 82–87. Nozulak, J., Kalkman, H.O., Floersheim, P., Hoyer, D., Schoeffter, P. and Buerki, H.R. (1995) (+)cis-4,5,7a,8,9,10,11,11a-octahydro-7H-10-methylindolo[1,7bc][2,6]naphthyridine: A 5-HT2C/2B receptor antagonist with low 5-HT2A receptor affinity. Journal of Medicinal Chemistry 38, 28–33. Orjales, A., Alonso-Cires, L., Labeaga, L. and Corcostegui, R. (1995) New (2methoxyphenyl)piperazine derivatives as 5-HT1A receptor ligands with reduced α1-adrenergic activity. Synthesis and structure-affinity relationships. Journal of Medicinal Chemistry 38, 1273–1277. Peroutka, S.J. (1994) Molecular biology of serotonin (5-HT) receptors. Synapse 18, 241–260. Sanders-Bush, E. and Canton, H. (1995) Serotonin receptors: Signal transduction pathways. In Psychopharmacology: The Fourth Generation of Progress, edited by F.E.Bloom and D.J.Kupfer, pp. 431–441. New York: Raven Press. Saudou, F. and Hen, R. (1994) 5-HT receptor subtypes: Molecular and functional diversity. Medicinal Chemistry Research 4, 16–84. Schmidt, C.J. and Fadayel, G.M. (1995) The selective 5-HT2A receptor antagonist MDL 100, 907 increases dopamine efflux in prefontal cortex of the rat. European Journal of Pharmacology 273, 273–279. Schreiber, R., Brocco, M. and Millan, M.J. (1994) Blockade of the discriminative stimulus effects of DOI by MDL 100,907 and the atypical antipsychotics clozapine and risperidone. European Journal of Pharmacology 264, 99–102. Serotonin receptors: Molecular genetics and molecular modeling. Special issue of Medicinal Chemistry Research 3, 269–418. Shih, J.C., Chen, K. and Gallaher, T.K. (1995) Molecular biology of serotonin receptors. In Psychopharmacology: The Fourth Generation of Progress, edited by F.E.Bloom and D.J.Kupfer, pp. 407–414. New York: Raven Press. Stjernlof, P., Ennis, M.D., Hansson, L.O., Hoffman, R.L., Ghazal, N.B. and Sundell, S. et al. (1995) Structure-activity relationships in the 8-amino-6,7,8,9-tetrahydro3H-benz[e]indole ring system. Effects of substituents in the aromatic system on serotonin and dopamine receptor subtypes. Journal of Medicinal Chemistry 38, 2202–2216. Street, L., Baker, R., Castro, J.L., Chambers, M.S., Guiblin, A.R., Hobbs, S.C. et al. (1993) Synthesis and serotonergic activity of 5-(oxadiazolyl)tryptamines: Potent agonists for 5-HT1D receptors. Journal of Medicinal Chemistry 36, 1529–1538.

Neurotransmititers, agonists and antagonists

499

Ziffa, E. and Fillion, G. (1992) 5-Hydroxytryptamine receptors. Pharmacology Reviews 44, 401–458. 10.3 HISTAMINE RECEPTORS Holger Stark and Walter Schunack 10.3.1 Introduction The story of histamine (10.45, specific histamine numbering is shown on this structure) began in the early 1900s with classical pharmacological investigations concerning its physiological and pathophysiological effects. Histamine has been recognized as an important chemical messenger communicating information from one cell to another. A large variety of cell types including smooth muscles, endocrine and exocrine glands, blood cells and cells of the immune system of mainly vertebrates respond to histamine stimuli with large differences from species to species. New aspects were brought into this

field of research by the discovery that histamine does not only act as a local hormone but also acts as a neurotransmitter. Although histamine is widely distributed within mast cells in almost all mammalian peripheral tissues it plays an important role in the mammalian brain displaying powerful neuromodulatory, immunomodulatory, and neurotransmitter effects. Histamine itself does not cross the blood-brain barrier. Physiologically, it is produced by decarboxylation of L-histidine, mainly by specific histidine decarboxylase. This enzyme could be selectively inactivated by (S)-αfluoromethylhistidine (10.46) being a “suicide” substrate. Histamine catabolism occurs along two alternative pathways. One metabolic route is via methylation by the specific histamine N-methyltransferase to Nτmethylhistamine. The other one is the oxidative deamination by diamine oxidase to imidazole acetaldehyde and further oxidative and coupling products. The first pathway seemed to be the only one to operate in the mammalian brain. Therefore, the methylated metabolite is the most important catabolite for investigations on central histamine levels for neurotransmitter function. According to their chronological order of discovery three subtypes of histamine receptors are pharmacologically accepted at present: histamine H1-, H2-, and H3receptors. The effects following direct stimulation of these G-protein coupled receptors are manifold and depend on species and tissue. H1-receptors are coupled to

Introduction to the principles of drug design and action

500

the phosphatidylinositol cycle, H2-receptors to adenylate cyclase; the signaling system for H3-receptors is unknown so far. Whereas H1- and H2-receptors are characterized by classical pharmacology as well as by molecular biology in different species the cloning of the H3-receptor is at present still under investigation. H1- and H2-receptors are postsynaptically located. The H3-receptor was identified at first as a presynaptically located autoreceptor inhibiting the synthesis and release of histamine in histaminergic neurons. Later on, the newly found function of H3-heteroreceptors modulating the release of a number of different neurotransmitter (noradrenaline, acetylcholine, dopamine, serotonin, neuropeptides) gave further hints for therapeutic indications of H3-receptor ligands. 10.3.2 H1-receptors Histamine possesses two basic moieties: the primary nitrogen on the side-chain (Nα) protonated under physiological conditions (pKa1=9.73) and the protonable aromatic imidazole nucleus (pKa2=5.91). The first H1-receptor agonists developed followed the minimal structural requirements having an aromatic ring and an ethylamine side-chain. This approach was more or less successful with N-methyl-2-(2-pyridyl)ethanamine (betahistine, 10.47) and 2-(2-thiazolyl)ethanamine (10.48). Although they show less than 30% activity compared to the endogenous ligand (histamine 100%) they were used as pharmacological tools for a long time. Recent developments in the class of 2substituted

histamine derivatives resulted in 2-(3-bromophenyl)histamine (10.49) and its trifluoromethyl analogue (10.50). These compounds have 112% and 128% relative

Neurotransmititers, agonists and antagonists

501

activity, respectively. Compounds (10.49) and (10.50) make the H1-receptor— mediated biological functions in experimental characterization much more precise than former tools. Most people have allergic and inflammation reactions on their mind when thinking of histamine. Compounds preventing this response to histamine are the first so called “antihistamines”. These H1-receptor antagonists are among the most widely used medications in the world. In contrast to histamine these agents are generally lipophilic compounds due to aromatic moieties. A small chain connects this aromatic ring structure to a protonable basic center, in most cases a tertiary nitrogen. One of the most potent and selective antagonists used pharmacologically is mepyramine (pyrilamine, 10.51). In radiolabelled form (3H, 11C) it can also be used for binding characterizations in vitro as well as in vivo. One of the most useful [125I]iodinated H1receptor antagonists is iodobolpyramine (10.52) which could be easily prepared in a reaction with a radiolabelled precursor. Many other compounds have been developed subsequently of which chlorpheniramine (10.53) and diphenhydramine (10.54) are two examples.

Introduction to the principles of drug design and action

502

Chlorpheniramine shows a stereoselective effect presenting higher antagonist activity with its (S)-(+)-enantiomer (eutomer). These H1-receptor antagonists of the first generation have a number of side effects which led to further developments of neuroleptics (chlorpromazine), antidepressants (amitriptyline, doxepin) or anticholinergic agents (trihexyphenidyl, biperiden). The most common central sideeffect of lipophilic H1-receptor antagonists is sedation. This property is now exploited with diphenhydramine and promethazine (10.55) as a main indication. Sedation and

Neurotransmititers, agonists and antagonists

503

atropine-like side-effects led to the next generation of H1-receptor antagonists. Ketotifen (10.56) is a potent H1-receptor antagonist and possesses antagonist activity to other allergy-inducing agents such as bradykinin or serotonin. This is an example of an antihistamine with additional

Introduction to the principles of drug design and action

504

antiallergic activity. Terfenadine (10.57) and astemizole (10.58) are not capable of readily penetrating the blood-brain barrier. Due to this lack of central effects in

Neurotransmititers, agonists and antagonists

505

therapeutic dosage as a consequence of pharmacokinetic differences, (10.57), (10.58) and other newer antihistamines (loratidine, cetirizine, levocabastine) may have advantages in their side-effect profile. 10.3.3 H2-receptors Structural similarities of the H2-receptor agonists dimaprit (10.59) and amthamine (10.60) to histamine (10.45) are obvious. In addition to the protonated side-chain nitrogen the structures are capable of making hydrogen bonds and of undergoing a 1,3-prototropic tautomerism like the imidazole moiety. The isothiourea and the aminothiazole moieties may be considered as bioisosteres of imidazole in this particular case. These compounds are roughly in the same activity range as histamine. Elongation of the alkyl chain and variation of the amino functionality to a strongly basic substituted guanidine group led to a strong increase in H2-receptor agonist activity. Impromidine (10.61) which contains two imidazole fragments is 48 times more potent than histamine. The imidazolylpropylguanidine group seems to be responsible for agonist binding whereas the other part of the molecule (cimetidine part, cf. 10.63) contributes additional binding. This binding is improved by a diarylalkyl structure such as in arpromidine (10.62). On cardiac H2-receptors this compound is about 100 times more potent than histamine, but shows different activity

Introduction to the principles of drug design and action

506

Neurotransmititers, agonists and antagonists

507

with different preparations of H2-receptors giving a hint for H2-receptor subpopulations. Thereby, the “fluoropheniramine” partial structure of (10.62) (cf. 10.53) also incorporates H1-receptor antagonist activity. The partial structure of impromidine (10.61) that enhances activity is the main structural element of cimetidine (10.63), the first world-wide marketed H2-receptor antagonist. The guanidine group is substituted by a cyano group which reduces the basic properties of this moiety and leads to a highly polar group. The basic imidazole ring connected by a flexible chain to a polar group which is uncharged under physiological conditions leads to H2-receptor antagonists. The imidazole moiety seems to be responsible for the inhibition of cytochrome P450-dependent reactions leading to unwanted side-effects with co-medications in man. Therefore, new developments replaced the imidazole ring bioisostere with a basic substituted furan (ranitidine, 10.64) and thiazole (famotidine) rings or a basic substituted phenoxyalkyl moiety (roxatidine acetate, 10.65; iodoaminopotentidine, 10.66; zolantidine, 10.67). H2Receptor antagonists are used for the treatment of conditions associated with gastric hyperacidity such as peptic ulcer disease or reflux oesophagitis. The new compounds have a high therapeutic index showing low incidence of side-effects. One of the most potent H2-receptor antagonists is iodoaminopotendine (10.66). This compound and a related azido derivative were used in radiolabelled form for autoradiographic localization, specific binding studies and photoaffinity labelling of cerebral H2receptors. All these compounds except (10.65)

Introduction to the principles of drug design and action

508

possess a cyanoguanidine, nitroethenediamine or comparable “urea equivalents” as a structural alternative for the polar group. Therefore, they do not readily cross the blood-brain barrier. No important central effects could be detected with these drugs in vivo. Their physicochemical properties have to be taken into account. Drug design to optimize the partition coefficient, ionization constant, and molecular size with retention of H2-receptor antagonist activity led to zolantidine which is capable of penetrating into the central nervous system. Hallucinatory effects of some H2-receptor antagonists when given in high dosage and the control of nociceptive responses may be clarified by behavioural experiments with zolantidine. 10.3.4 H3-receptors The function of histamine as a neurotransmitter, in addition to its autacoid function, was strengthened by the discovery of the histamine H3-receptor. Histamine shows selectivity for the H3-receptor displaying a higher activity at H3- rather than at H1- and H2-receptors (H3, pD2 7.2; H2, pD2 6.0; H1, pD2 6.85). All H3-receptor ligands with high affinity known so far possess an imidazole ring. Any replacement study leads to a dramatic decrease in affinity. Even substitution by small methyl groups on the

Neurotransmititers, agonists and antagonists

509

imidazole ring decreases H3-receptor affinity (cf. 10.63). On the other hand, methylation on the side-chain increases agonist activity in particular cases. Methylation on the side-chain nitrogen resulted in the H3-receptor agonist Nαmethylhistamine (10.68). Although this compound is often used in tritium-labelled form in H3-receptor binding studies it shows some remarkable affinity at H1- and H2receptors. The use of the chiral methylated histamine derivative [3H](R)-αmethylhistamine should be favoured over other commercially available [3H]labelled H3-receptor agonists for these investigations. The reason for this preference is that the selectivity of side-chain methylated histamine derivatives were increased by (R)-αmethylhistamine (10.69), now the standard agonist for histamine H3-receptors, and was

furthermore increased with (αR,βS)-α,β-dimethylhistamine (10.70). Despite their structural similarity these compounds show low activity at H1- and H2-receptors but compared to histamine they exhibit 15 and 18 times the H3-receptor agonist activity, respectively, displaying impressive receptor selectivity. The side chain-branched histamine derivatives show a high degree of stereoselectivity. In all cases the enantiomer with the same relative configuration in the α-position as L-histidine is the eutomer. The eudismic ratio of (10.69) compared to its (S)-enantiomer (distomer) is about 130. Even the imidazolylmethylpyrrolidine derivative (immepyr, 10.71) shows this stereoselectivity. A comparable compound containing an achiral piperidinylmethyl moiety (immepip) was also developed recently. One of the most potent H3-receptor agonists so far is the imidazolylethylisothiourea derivative (10.72, imetit). Replacement of the side-chain nitrogen by the polar isothiourea moiety, cationic under physiological conditions, leads to improved receptor activity. The

Introduction to the principles of drug design and action

510

sulfur atom in imetit is not critically important and may be replaced by oxygen or methylene. Recent studies with molecular modelling methods showed that all these H3-receptor agonists could be superimposed on one pharmacophore model displaying similar molecular interaction patterns. Unfortunately all these ligands display similar physicochemical properties. They are extremly hydrophilic compounds which do not easily cross the blood-brain barrier and reach the central nervous system (CNS), the area with the highest H3-receptor density. It is notable that the well investigated (R)-αmethylhistamine (10.69) is a good substrate for the inactivating enzyme histamine methyltransferase. Clinical trials were not as promising as the first pharmacological experiments. Pro-drugs of (10.69) were designed to prepare lipophilic compounds which could easily penetrate biological membranes besides not being a substrate for the inactivating enzyme. Compound (10.73) is the lead structure for pro-drugs of (10.69). The liberation of the active drug (10.69) depends on chemical hydrolysis of the azomethine bond. Depending on the substitution pattern of the benzophenone promoiety

the compounds could be targeted to central or peripheral tissues. High plasma and CNS levels of the biologically highly active (R)-α-methylhistamine could be achieved by this approach for a prolonged duration. Histamine derivatives with larger substituents on the side-chain Nα nitrogen lead to partial agonists, and further increasing the size of the substituents leads to histamine H3-receptor antagonists. This transition from agonist to antagonist was extensively shown with histamine derivatives, but the same is true for derivatives of imetit (10.72). Clobenpropit (10.74) is in vitro one of the most potent H3-receptor antagonists obtained by this approach. The first compounds detected as H3-receptor antagonists were compounds from the line of H2-receptor antagonists. Burimamide (10.75) which was used as the first “selective” H2-receptor antagonist for the characterization of H2receptors was ironically also used for the characterization of H3-receptors later on. The H1-receptor agonist betahistine (10.47) shows some, while the H2-receptor agonists impromidine (10.61) and arpromidine (10.62) show high H3-receptor antagonist activity. Once more the imidazole structure like that in histamine seems to be an essential component of highly potent H3-receptor ligands. With the optimization of the thiourea derivative (10.75) a new rigid piperidino and a cyclohexyl moiety was

Neurotransmititers, agonists and antagonists

511

introduced. The resulting thioperamide (10.76) was the first highly potent and selective H3-receptor antagonists. Therefore, (10.76) is now the standard antagonist. A general structural pattern for activity was developed using different series of antagonists. A nitrogen-containing heterocycle (mainly imidazole) is connected via a chain to a polar group; this structure seems to be essential for a potent antagonist interaction with H3-receptors. A lipophilic residue, linked to the polar group

by a spacer seems to enable the molecule to reach additional binding areas, e.g. a hydrophobic pocket on the receptor, so that the H3-receptor antagonist activity of the

Introduction to the principles of drug design and action

512

resulting molecule increases. This structural pattern was used to design a new H3receptor radioligand with high potency and selectivity. In contrast to the former antagonists possessing a polar group easily able to form hydrogen bonds, the resulting iodoproxyfan (10.77) has an ether functionality. This structure is suited for hydrogen bonding to a limited extent only. This seems necessary to lower the unspecific binding of a useful radioligand. [125I]lodoproxyfan fulfills all criteria for a radioligand such as high activity, selectivity and specificity as well as saturable and reversible binding. Although radioligands are useful for different pharmacological experiments the development of therapeutically acceptable drugs is a different prospect. Recently, a new class of imidazolylalkyl carbamates were developed possessing high H3-receptor antagonist activity in vivo following oral administration. N-Fluorophenyl carbamate (10.78) is one potent lead in this series. Depending on the substitution pattern of the phenyl ring the pharmacokinetic properties could be varied with retention of the antagonist activity. This approach seems to be very promising in developing the first potentially useful H3-receptor antagonists. But the therapeutic indication is not totally clear at present. Different psychic disorders or diseases, e.g. epilepsy, stress, food intake, sleeping, vertigo or Morbus Alzheimer’s disease, seem potential targets for H3receptor antagonists. Influencing the histaminergic neurotransmitter system seems to be an attractive new

approach for numerous diseases. The new ligands should improve our knowledge of the physiological and pathophysiological interactions of different neurotransmitters and show new possibilities for the treatment of different diseases.

Neurotransmititers, agonists and antagonists

513

FURTHER READING Buschauer, A., Schunack, W., Arrang, J.-M., Garbarg, M., Schwartz, J.-C. and Young, J.M. (1989) Histamine Receptors. In Receptor Pharmacology and Function, edited by M.Williams, R.M.Glennon and P.B.M.W.M.Timmermans, pp. 293–347. New York and Basel: Marcel Dekker, Inc. Cooper, D.G., Young, R.C., Durant, G.J. and Ganellin, C.R. (1990) Histamine Receptors. In Comprehensive Medicinal Chemistry: The Rational Design, Mechanistic Study & Therapeutic Application of Chemical Compounds, edited by C.Hansch, pp. 323–421. Oxford and New York: Pergamon Press. Hill, S.J. (1990) Distribution, Properties and Functional Characteristics of Three Classes of Histamine Receptor. Pharmacological Reviews 42, 45–83. Krause, M., Rouleau, A., Stark, H., Luger, P., Garbarg, M., Schwartz, J.-C. and Schunack, W. (1995) Synthesis, X-ray Crystallography, and Pharmacokinetics of Novel Azomethine Prodrugs of (R)-α-Methylhistamine: Highly Potent and Selective Histamine H3-Receptor Agonists. Journal of Medicinal Chemistry 38, 4070–4079. Leurs, R. and Timmerman, H. (1992) The Histamine H3-Receptor: A Target for Developing New Drugs. Progress in Drug Research 39, 127–165. Lipp, R., Stark, H. and Schunack, W. (1992) Pharmacochemistry of H3-Receptors. In The Histamine Receptor, Ser. Receptor Biochemistry and Methodology, edited by J.-C. Schwartz and H.L.Haas, Vol. 16, pp. 57–72. New York and Basel: WileyLiss, Inc. Schwartz, J.-C., Arrang, J.-M., Garbarg, M., Pollard, H. and Ruat, M. (1991) Histaminergic Transmission in the Mammalian Brain. Physiological Reviews 71, 1–51. Zingel, V., Leschke, C. and Schunack, W. (1995) Developments in Histamine H1Receptors Agonists. In Progress in Drug Research 44, 49–85. 10.4 DOPAMINE RECEPTORS Philip G.Strange 10.4.1 Introduction Dopamine receptors have been very important targets for drug design by medicinal chemists partly because of the involvement of dopamine systems in important physiological functions and partly because dopamine receptors are important targets for drug action. In the brain dopamine systems are involved in the control of movement and certain aspects of behaviour, in the pituitary dopamine is important in the control of the secretion of prolactin and melanocyte stimulating hormone (αMSH), in the cardiovascular system dopamine is important in the control of blood pressure and heart rate and in the eye dopamine is important for the control of certain aspects of visual function. In these systems the actions of dopamine are mediated by binding to receptors and blockade or activation of these receptors can offer therapy for certain disorders. For example, dopamine antagonists have been shown to be

Introduction to the principles of drug design and action

514

important in the treatment of schizophrenia, whereas dopamine agonists have been shown to be of use in the treatment of the brain disorder Parkinson’s disease, in the therapy of excessive prolactin secretion and certain prolactin secreting tumours and in the therapy of cardiovascular disorders. For these reasons a very wide range of compounds has been synthesised that bind to receptors for the neurotransmitter dopamine. The development of the concept of multiple dopamine receptors has added further impetus to this drug discovery programme. Based on studies of the actions of dopamine using pharmacological and biochemical techniques it became apparent in the late 1970’s that the concept of a single receptor for dopamine was insufficient to explain the information emerging and it was suggested that there were two receptors for dopamine that were termed D1 and D2. These had different pharmacological and biochemical properties some of which are summarised in Table 10.2. The concept of two dopamine receptors survived until the techniques of molecular biology were applied to the dopamine receptors. This showed that there were at least five dopamine receptor subtypes (D1–D5) which have different structural and functional properties and different localisations in tissues. Some of their properties are summarised in Table 10.3. This rather complicated picture can be simplified by the realisation that on the basis of structural and functional properties these five receptor subtypes can be grouped into two subfamilies: D1/D5 which have properties similar to those of the pharmacologically defined D1 receptor and D2/D3/D4 which have properties similar to those of the pharmacologically defined D2 receptor. The two subfamilies are therefore now termed the D1-like and D2-like subfamilies. In the subsequent discussion when a receptor is referred to as D1 this will imply that this is the receptor subtype defined by gene cloning whereas when the receptor has only been defined by pharmacological or biochemical analyses the nomenclature D1-like/D2-like will be used. This emerging understanding of the multiple subtypes of dopamine receptors has importance for the way the activities of potential new dopamine receptor directed drugs are assayed. In early studies of these compounds animal behavioural tests were used e.g. the induction of stereotyped behaviour or the induction of turning in rodents; these tests detect dopamine receptor activity but do not distinguish compounds with selectivity for different receptor subtypes. The definition of D1-like and D2-like receptors from

Table 10.2 Dopamine receptor subtypes defined on the basis of biochemical and pharmacological studies. D1 (D1-like) D2 (D2-like) selective agonists SKF 38393 (10.97) quinpirole (10.96) selective antagonists SCH 23390* sulphide* biochemical response cAMP↑ cAMP↓, K+ channel↑, Ca2+ channel↓ The data in the Table are based upon the suggestion of Kebabian and Calne (1979) but have been expanded to include more recent information (Vallar and Meldolesi 1989). The original classification was into D1 and D2 receptor subtypes but, as discussed in the text, with

Neurotransmititers, agonists and antagonists

515

the advent of the isoforms defined by molecular biology (Table 10.3) these should be termed D1-like and D2-like receptors. *Formula in Table 10.4. biochemical and pharmacological studies enabled compounds to be assayed for their interaction at the two subclasses using activity based assays e.g. the stimulation of adenylate cyclase for the D1-like receptors or using ligand binding assays. Most recently the availability of cloned genes for the five receptor subtypes has offered the prospect of the assay of selective substances for their activities against each subtype expressed in a suitable cell host using ligand binding assays or activity based assays. The use of these different assay systems poses certain problems in the definition of the selectivity of compounds directed at the different receptors. This is particularly acute for the agonists where there are two qualities of an agonist that are potentially of interest; its affinity for the receptor and its ability to stimulate a response. In defining selectivity between receptors in activity based assays these two quantities are not always defined or separated and this can lead to confusion. Even in ligand binding assays there is potential confusion in extracting the relevant parameters for defining selectivity. Therefore in the discussion below “activity” will be referred to for agonists which is a broad definition of selectivity based on the tests used in the particular publication cited and should be taken only as a guide to actual selectivity. For antagonists there are fewer problems of this nature although the emerging phenomenon of “inverse agonism” may complicate matters. 10.4.2 Dopamine agonists It was in the late 1950’s that an independent role for dopamine as a neurotransmitter was postulated and this led to the development of models for the assessment of dopamine agonism mostly based on animal behavioural tests. From these tests it became clear that a number of naturally occurring or semisynthetic substances possessed dopamine agonist activity. Notable among these were the aporphine alkaloids e.g. apomorphine (10.79) and the ergot alkaloids e.g. ergotamine (10.80). At the same time synthetic programmes were initiated to obtain dopamine agonists with greater potency or activity and some of the important chemical classes of agonists will be considered below. 10.4.2.1 Aminotetralins and related compounds A major synthetic effort has been devoted to compounds related to dopamine but where the dopamine molecule is “locked” in to a rigid structure. The best known examples of these are the aminotetralins. These have been synthesised in a variety of analogues with different hydroxyl substitution patterns on the aromatic ring. The 5,6and 6,7-dihydroxy

Introduction to the principles of drug design and action

516

aminotetralins (ADTN’s) (10.81, 10.82) may be considered to be equivalent to the dopamine molecule, frozen into one or other of its two principal conformations (termed a and β conformations, 10.83, 10.84). The possibility of using these compounds to determine the active conformation of dopamine attracted much interest but it has not proven possible to draw any firm conclusion from the results of the studies. The aminotetralins have activities at both the D1-like and D2-like subfamilies of dopamine receptors and it has been possible to draw some broad conclusions about the structure activity relationships involved. Compounds with two hydroxyl groups show the highest activity and the 5,6- and 6,7-congeners both have high activity. Substantial activity is retained in the monohydroxy compounds, for example the 5-hydroxy aminotetralins (eg. 10.85). The monohydroxy equivalents of dopamine (tyramines) also have some activity as dopamine agonists supporting this idea. Compounds lacking hydroxyl groups eg N,N-dipropyl-2-aminotetralin and phenylethylamine also exhibit weak agonist activity so these hydroxyl groups are not essential for agonist

Neurotransmititers, agonists and antagonists

517

action. The amino group of the aminotetralins has been derivatised in a number of cases and this has shown that the addition of two alkyl groups enhances activity both at D1-like and D2-like receptors (eg.

10.86). Activity increases with increasing alkyl chain length up to the di N-propyl derivatives which have the highest activities and the di N-butyl compounds which have less activity. It seems that there may be some additional site on the receptor with steric constraints which is occupied by these alkyl groups and which enhances agonist activity. A series of 5-hydroxy aminotetralins has been synthesised where the N-n-propyl, N-phenylethyl and N-n-propyl, N-thenylethyl congeners are extremely potent agonists with substantial D2-like selectivity (10.87, 10.88). This suggests that particular groups larger than n-propyl can enhance potency further when attached to the aminotetralin structure. Generally the aminotetralins do not show great selectivity among the different D2-like receptors but 7-hydroxy-N,N′ di n-propyl aminotetralin (10.89) has some selectivity for the D3 receptor mostly based on data from ligand binding assays. 10.4.2.2 Aporphine alkaloids The aporphine alkaloids contain the dopamine structure in a rigid conformation. R(−)apomorphine (10.79) has been shown to have D1-like and D2-like agonist activity although it tends to be a partial agonist at the D1-like receptors. The replacement of the N-methyl group with an N-propyl group in R(−)-N-propyl norapomorphine (10.90) increases the affinity for the D2-like receptors relative to the D1-like receptors. 10.4.2.3 Ergot alkaloids A number of ergot alkaloids with dopamine agonist activity have been isolated from the crude mixture of natural products known as ergot. Typical examples of these are

Introduction to the principles of drug design and action

518

Neurotransmititers, agonists and antagonists

519

ergotamine (10.80) and α-ergocriptine (10.91) which contain the D-lysergic acid structure linked by an amide bond to a cyclic peptide moiety and these have come to be called “ergopeptines”. Modification of these natural products by chemical synthesis has provided substances with better selectivity for dopamine receptors and notable here is bromocriptine (10.92) which has potent D2-like receptor activity but D1-like receptor antagonistic properties. It is used in the treatment of Parkinson’s disease and excessive prolactin secretion. A large number of semisynthetic “ergolines” exist where the structure is based on the lysergic acid structure and the peptide side chain of the ergopeptines has been eliminated. These have been shown to possess potent agonist activity at the D2-like receptors eg. lergotrile (10.93), pergolide (10.94), lisuride (10.95). Some of these compounds also possess significant agonist activity at the D1-like receptors eg

Introduction to the principles of drug design and action

520

pergolide, whereas others possess little or no agonist activity and in fact may be D1 partial agonists or antagonists depending on the test system eg lisuride (10.95).

Much synthetic work has ensued in order to identify the part of the ergoline that is responsible for the dopamine activity. Among the compounds synthesised are a group of partial ergolines including quinpirole (10.96) which is a very selective D2-like agonist. 10.4.2.4 Benzazepines The benzazepine nucleus has been used to provide another series of molecules some of which are potent and selective D1-like agonists eg SKF 38393 (10.97), fenoldopam

Neurotransmititers, agonists and antagonists

521

(10.98). Other analogues exhibit activity at both the D1-like and D2-like receptors. Benzazepines also provide selective D1-like antagonists (see below). 10.4.2.5 Miscellaneous structures The naphthoxazine PHNO (10.99) has been synthesised and is one of the most potent D2-like receptor agonists available with little ability to bind to or activate D1-like receptors. N-n-Propyl-3-(hydroxyphenyl) piperidine (3-PPP) (10.100) is an example of a compound

where the 3(R)-stereoisomer has D2-like receptor agonist activity whereas the 3(S)isomer has variable intrinsic activity depending on the receptor preparation. This can be seen as a preferential ability to act as an agonist at presynaptic autoreceptors but to behave as an antagonist at postsynaptic receptors. The compounds have little D1-like receptor activity.

Introduction to the principles of drug design and action

522

10.4.2.6 Selective agonists Compounds exist, as indicated above, that have the ability to act selectively as agonists on D1-like receptors eg SKF 38393 (10.97) or D2-like receptors eg quinpirole (10.96), PHNO 10.99). These compounds do not, however, show clear selectivity for the different isoforms comprising the two subfamilies with the exception of 7-OH DPAT (10.89) as mentioned earlier. In Table 10.3 some data are given for agonists from ligand binding studies illustrating this point. A second area where selectivity has been claimed for dopamine agonists is at autoreceptors. These are the receptors on dopamine nerve terminals or cell bodies that mediate inhibition of neurotransmitter synthesis, release or cell firing. It was found that although these receptors exhibited a pharmacological profile consistent with a D2-like receptor the autoreceptors were more sensitive to agonists compared to postsynaptic receptors. This would be consistent with autoreceptors being D2-like receptors but with a larger amplification (spare receptor ratio). This can at

Table 10.3 Dopamine receptor subtypes defined on the basis of molecular biological studies. D1-like D2-like receptor isoform D1 D5 D2 D3 D4 446 477 414/443 400 419 amino acids in human receptor pharmacological properties agonist binding (Ki nM) (−)-apomorphine (10.79) 0.7 — 0.7 32 4 dopamine 0.9 0.9 7 4 30 quinpirole (10.96) 1900 — 4.8 24 30 7-OH DPAT (10.89) 5000 — 10 1 650 SKF 38393 (10.97) 1 0.5 150 5000 1000 antagonist binding (Ki nM) haloperidol* 80 100 1.2 7 2.3 chlorpromazine* 90 130 3 4 35 clozapine* 170 330 230 170 21 raclopride 18000 — 1.8 3.5 2400 remoxipride 240000 — 300 1600 2800 (−)-sulpiride* 45000 77000 15 13 1000 SCH 23390* 0.2 0.3 1100 800 3000

Neurotransmititers, agonists and antagonists

523

The different dopamine receptor isoforms can be distinguished structurally on the basis of the sizes of the predicted third intracellular loops and C-terminal tails (Civelli et al. 1993; Strange 1991). The D1like receptors both have short third intracellular loops and long Cterminal tails whereas the D2-like receptors each have long third intracellular loops and short C-terminal tails. The D1-like and D2-like subgroups can also be distinguished on the basis of amino acid homologies. The data for Ki values shown in the Table are derived from Seeman and Van Tol (1994) and Hacksell et al. (1995) using ligand binding. There is some variability in the values derived from different studies and this is a particular problem for agonists where the complexities of agonist binding studies need to be considered. The selectivity of clozapine for D4 over D2 receptors (Seeman and Van Tol 1994) has not been found in all subsequent studies (Strange 1991; Hacksell et al. 1995). *Formula in Table 10.4. least in part explain the apparent autoreceptor selectivity of 3(S)-3-PPP (10.100). If this compound is a partial agonist at the D2-like receptors then at receptors with a large amplification clear agonist activity will be seen but at receptors with a lower amplification factor antagonist activity may be exhibited. More recently compounds have also been synthesised with apparent auto receptor antagonist selectivity (see below). 10.4.3 Dopamine antagonists In the early 1950’s the phenothiazine, chlorpromazine, (Table 10.4) was discovered to have the ability to induce in humans a state of indifference without loss of consciousness and it began to be used as an anti psychotic drug. In the late 1950’s the butyrophenone series of drugs was discovered e.g. haloperidol (Table 10.4) and these were shown to have anti psychotic activity. It was eventually found that a prominent action of these drugs was to inhibit various actions of dopamine and it became clear that one of their principal activities was as dopamine antagonists. The phenothiazines and butyrophenones

Introduction to the principles of drug design and action

524

Table 10.4a The major classes of dopamine antagonists.*

are D2-like antagonists and have varying abilities as D1-like antagonists. They also show varying abilities to act as serotonergic, muscarinic, histaminergic and adrenergic antagonists. Extensive synthetic programmes have been performed with the aim of

Neurotransmititers, agonists and antagonists

525

developing more selective drugs with different structures and in Table 10.4 some of the key structural classes of dopamine antagonist are shown together with an indication

Introduction to the principles of drug design and action

Table 10.4b The major classes of dopamine antagonists.*

526

Neurotransmititers, agonists and antagonists

527

of their selectivity. Of note here are the benzazepines which are selective D1-like antagonists and the substituted benzamides which are selective D2-like antagonists. While some of the compounds shown in Table 10.4 may show some selectivity between D1-like and D2-like receptors they do not in general have any marked abilities to

discriminate the individual members of the two subfamilies. Table 10.3 gives some data from ligand binding studies on the dissociation constants for some antagonists at the different receptor subtypes defined by gene cloning. It can be seen that there are some drugs such as raclopride that show low affinity for some receptor isoforms but there is at present no substance available that has a clear selectivity for a single receptor isoform. It was claimed that clozapine had a clear selectivity for the D4 receptor subtype but subsequent work has not reported the same selectivity. Behavioural evidence for preferential actions of the compounds (+)-UH 232 (10.101) and (+)-AJ 76 (10.102) as antagonists at dopamine auto receptors has been presented but the relation of these findings to the different dopamine receptor isoforms is unclear. These compounds do show some selectivity for binding to the D3 dopamine receptor but the selectivity is not great. Extensive new synthetic programmes are currently in progress to develop compounds with clear selectivity for the different dopamine receptor isoforms and some selective agents are emerging (see, for example, Patel et al., 1996). FURTHER READING Beaulieu, M., Itoh, Y., Tepper, P., Horn, A.S. and Kebabian, J.W. (1984) N,Ndisubstituted 2-aminotetralins are potent D2 dopamine receptor agonists. European Journal of Pharmacology 105, 15–21. Cannon, J.G. (1983) Structure activity relationships of dopamine agonists. Annual Reviews of Pharmacology and Toxicology 23, 103–130. Cavero, I., Massingham, R. and Lefevre-Borg (1982) Peripheral dopamine receptors, potential targets for a new class of antihypertensive agents. Life Sciences 31, 939– 948; 1059–1069.

Introduction to the principles of drug design and action

528

Civelli, O., Bunzow, J.R. and Grandy, D.K. (1993) Molecular diversity of the dopamine receptors. Annual Reviews of Pharmacology and Toxicology 32, 281– 307. Hacksell, U., Jackson, D.M. and Mohell, N. (1995) Does the dopamine receptor subtype selectivity of antipsychotic agents provide useful leads for the development of novel therapeutic agents? Pharmacology Toxicology 76, 320–324. Hauth, H. (1979) Chemical aspects of ergot derivatives with central dopaminergic activity. In Dopaminergic Ergot Derivatives and Motor Function, edited by K.Fuxe and D.B.Calne, pp. 23–31, Oxford: Pergamon Press. Hogberg, T. (1991) Novel substituted salicylamides and benzamides as selective D2 receptor antagonists. Drugs of the Future 16, 333–357. Johansson, A.M., Arvidsson, L.E., Hacksell, U., Nilsson, J.L.G., Svensson, K. and Hjorth, S. et al. (1985) Novel dopamine receptor agonists and antagonists with preferential action on autoreceptors. Journal of Medicinal Chemistry 28, 1049– 1053. Kebabian, J.W. and Calne, D.B. (1979) Multiple receptors for dopamine. Nature (Lond.) 277, 93–96. Leff, P. (1995) Inverse agonism: theory and practice. Trends in Pharmacological Sciences 16, 256. Leysen, J.E. and Niemegeers, C.J.E. (1985) Neuroleptics. Handbook of Neurochemistry 9, 331–361. Martin, G.E., Williams, M., Pettibone, D.J., Yarborough, G.G., Clineschmidt, B.V. and Jones, J.H. (1984) Pharmacolgic profile of a novel potent direct-acting dopamine agonist, (+)-PHNO. Journal of Pharmacology and Experimental Therapeutics 230, 569–576. Patel, S., Marwood, R., Emms, F., Marsten, D., Leeson, P.D., Curtis, N.R., Kulagowski, J. and Freedman, S.B. (1996) Identification and pharmacological characterisation of [125I] L 750,667 a novel radioligand for the D4 dopamine receptor. Mol. Pharmacol. 50, 1658–1664. Seeman, P. (1980) Brain dopamine receptors. Pharmacological Reviews 32, 229–313. Seeman, P. (1987) Dopamine receptors in brain and periphery. Neurochemistry International 10, 1–25. Seeman, P. and Van Tol, H.H.M. (1994) Dopamine receptor pharmacology. Trends in Pharmacological Sciences 15, 264–270. Seeman, P., Watanabe, M., Grigoriadis, D., Tedesco, J.L., George, S.R., Svensson, U. et al. (1985) Dopamine D2 receptor binding sites for agonists. Molecular Pharmacology 28, 391–399. Seiler, M.P. and Markstein, R. (1982) Further characterisation of structural requirements for agonists at the striatal dopamine D1 receptor. Molecular Pharmacology 22, 281–289. Sibley, D.R. and Creese, I. (1983) Interaction of ergot alkaloids with anterior pituitary D2 dopamine receptors. Molecular Pharmacology 23, 585–593. Sorensson, C., Waters, N., Svensson, K., Carlsson, A., Smith, M.W. and Piercey, M.F. et al. (1993) Substituted 3-phenyl piperidines: new centrally acting autoreceptor antagonists. Journal of Medicinal Chemistry 36, 3188–3196.

Neurotransmititers, agonists and antagonists

529

Strange, P.G. (1991) Interesting times for dopamine receptors. Trends in Neurosciences 14, 43–45. Strange, P.G. (1992) Brain Biochemistry and Brain Disorders. Oxford: Oxford University Press. Strange, P.G. (1994) Dopamine D4 receptors: curiouser and curiouser. Trends in Pharmacological Sciences 15, 317–319. Vallar, L. and Meldolesi, J. (1989) Mechanisms of signal transduction at the dopamine D2 receptor. Trends in Pharmacological Sciences 10, 74–77. Waddington, J.L. and O’Boyle, K.M. (1987) The D1 dopamine receptor and the search for its functional role: from neurochemistry to behaviour. Reviews in the Neurosciences 1, 157–184. Wikstrom, H., Sanchez, D., Lindberg, P., Hacksell, U,., Arvidsson, L.E., Johansson, A.M. et al. (1984) Resolved 3-(3-hydroxyphenyl) N-n-propylpiperidine and its analogues: central dopamine receptor activity. Journal of Medicinal Chemistry 27, 1030–1036. Wolf, M.E. and Roth, R.H. (1987) Dopamine autoreceptors. In Dopamine Receptors, edited by I.Creese and C.M.Fraser, pp. 45–96. New York: Alan R.Liss.

11. DESIGN OF ANTIMICROBIAL CHEMOTHERAPEUTIC AGENTS EDWARD G.M.POWER and A.DENVER RUSSELL CONTENTS 11.1 INTRODUCTION

435

11.2 PRODUCTION OF CHEMOTHERAPEUTIC AGENTS

436

11.3 MECHANISM OF ACTION OF CHEMOTHERAPEUTIC AGENTS

438

11.3.1 Inhibitors of cell wall synthesis

438

11.3.2 Membrane-active agents

442

11.3.3 Inhibitors of protein synthesis

442

11.3.4 Inhibitors of nucleic acid synthesis

442

11.3.5 Antibacterial folate inhibitors

445

11.3.6 Conclusions and comments

447

11.4 BACTERIAL RESISTANCE TO CHEMOTHERAPEUTIC AGENTS

447

11.4.1 Enzyme-mediated resistance

447

11.4.1.1 β-lactamases

447

11.4.1.2 Aminoglycoside-modifying enzymes

449

11.4.1.3 Chloramphenicol-inactivating enzymes

451

11.4.2 Outer membrane barrier

452

11.4.3 Transferable resistance

452

11.4.4 Conclusions and comments

453

11.5 DESIGN OF β-LACTAM ANTIBIOTICS

454

Design of antimicrobial chemotherapeutic agents

531

11.5.1 Penicillins

454

11.5.2 Cephalosporins

457

11.5.2.1 Structure-activity relationships

457

11.5.2.2 Pharmacokinetic properties

459

11.5.3 β-Lactamase stability

461

11.5.4 β-Lactamase inhibitors

462

11.5.4.1 β-Lactams as inhibitors

462

11.5.4.2 Naturally occurring β-lactamase inhibitors

462

11.5.4.3 Synthetic β-lactamase inhibitors

464

11.5.4.4 Structure-activity relationships in βlactamase inhibitors

464

11.5.4.5 β-lactamase inducers

464

11.5.4.6 Mutual pro-drugs

465

11.5.5 Other β-lactam ring systems

465

11.5.5.1 1-Oxacephems

465

11.5.5.2 Penems

465

11.5.5.3 Nocarcidins

466

11.5.5.4 Monobactams

466

11.5.5.5 Carbacephems

466

11.6 DESIGN OF OTHER ANTIBACTERIAL AGENTS

467

11.6.1 Aminoglycoside-aminocyclitol antibiotics

467

11.6.2 Tetracyclines

469

11.6.3 Macrolides

471

11.6.4 Chloramphenicol

472

11.6.5 Folate inhibitors

473

11.6.6 Quinolones

474

11.7 DESIGN OF ANTIFUNGAL AGENTS

476

Introduction to the principles of drug design and action

532

11.7.1 Polyene antibiotics

476

11.7.2 Imidazole derivatives

478

11.7.3 Griseofulvin

479

11.7.4 Flucytosine

480

11.7.5 Other membrane-active compounds

480

11.7.6 Cell wall-active compounds

481

11.7.7 Novel antifungal agents

483

11.7.8 Conclusions and comments

483

11.8 DESIGN OF ANTIVIRAL AGENTS 11.8.1 Mechanisms of inhibition

483 484

11.8.1.1 Amantadines

484

11.8.1.2 Nucleoside analogues

484

11.8.1.3 Foscarnet

487

11.8.1.4 Interferons

487

11.8.2 Mechanisms of resistance

488

11.9 OVERALL CONCLUSIONS AND COMMENTS

489

FURTHER READING

490

11.1 INTRODUCTION The design of a new chemotherapeutic agent, suitable for clinical use in the treatment of human infections, must take into account two aspects above all others. First, the drug must possess high antimicrobial activity and secondly it must be non-toxic to human tissue. Paul Erlich’s early concept of a selectively toxic ‘magic bullet’ is thus just as true in the modern world as when it was first propounded. A host of antimicrobial agents has been examined and many shown to be effective inhibitors of micro-organisms in vitro; unfortunately, several of these have been found to be harmful to human tissues and consequently have no role to play in chemotherapy. This chapter will thus concentrate on several of those chemotherapeutic agents that have proved their worth when employed internally (usually orally or parenterally). Antimicrobial compounds that are used for their disinfectant, antiseptic or preservative qualities will not be dealt with here.

Design of antimicrobial chemotherapeutic agents

533

Two important properties of any chemotherapeutic agent were mentioned briefly above. Additionally, any drug must ideally have a broad spectrum of activity, with a rapid bactericidal or other microbicidal action. Some bacteria produce enzymes that can inactivate or modify antibiotics, and insusceptibility of a drug to such degradation or modification could result in its playing an important part in therapy. Likewise, some bacteria possess an outer membrane that acts as a permeability barrier to the entry of some, but not all, antibiotics. Drugs that can readily penetrate this barrier might again be expected to be of possible clinical importance. These two aspects are considered in greater detail later (Sections 11.4.1 and 11.4.2). In addition to being non-toxic, an antibiotic should not cause any hypersensitive reactions, such as those induced in a minority of patients by the penicillins and, to a lesser extent, the cephalosporins. It should, however, be readily absorbed to give high blood and tissue levels and it should be stable to gastric acid. Binding to serum proteins should be of a low order so that high concentrations of the drug are freely available in the plasma. In urinary infection, high urine levels are desirable but some antibiotics are excreted so rapidly that, on occasion, it may be necessary to delay the rate of excretion. These and other pharmacological properties are considered further where appropriate. Finally, the design of any chemotherapeutic agent must involve a consideration of its chemical and physical properties, since these will be of paramount importance to the pharmacist responsible for formulating a suitable product. Such properties include its aqueous solubility, and its stability in solution at different pHs and temperatures. These aspects are considered to be outside the scope of the present chapter. 11.2 PRODUCTION OF CHEMOTHERAPEUTIC AGENTS The demonstration in the early 1920s that lysozyme possessed antibacterial activity, the accidental discovery of ‘penicillin’ by Fleming in the late 1920s and the finding that the azo dye prontosil owed its antibacterial properties to the release in vivo of sulphanilamide, all stimulated research towards the development of chemotherapeutic agents. Fleming, in 1929, published the results of his chance finding that a Penicillium mould caused lysis of staphylococcal colonies on an agar plate. He also showed that the filtrate of a culture of the mould, growing in a liquid medium, possessed significant activity against Gram-positive bacteria and Gram-negative cocci, although most other types of Gram-negative organisms were resistant. In retrospect, it was indeed fortunate that the original plates of staphylococci did not contain organisms that produced an inactivating enzyme (β-lactamase: Section 11.4.1) otherwise the mould would not have shown any activity and could well have been discarded! It is, of course, highly unlikely with today’s knowledge and techniques that benzylpenicillin (the product obtained from the Penicillium mould) would have been lost to mankind, but rather that its introduction into medicine would have been delayed. It is now known that benzylpenicillin is thermolabile. At the time of Fleming’s discovery, however, great difficulty was experienced in extracting the antibiotic from the culture medium, as this property was not appreciated. Later studies by Florey,

Introduction to the principles of drug design and action

534

Chain and their colleagues at Oxford University, and in the USA, used a cold solvent extraction method that succeeded in extracting and purifying the elusive active principle. Extensive research on the composition of the culture media and on induction of mutants of the mould resulted in conditions that gave enhanced antibiotic yields. These facts are all relevant to the subsequent development of antibiotics, since the commercial production of benzylpenicillin stimulated world-wide efforts into examining soil samples with a view to obtaining other antibiotics from moulds residing there. Studies at Rutgers (by Waksman and colleagues) and elsewhere were responsible for the development of streptomycin, the tetracyclines and chloramphenicol. These and many other antibiotics—some of them clinically useful— were obtained from Streptomyces species, which comprise the most prolific source of antibiotics. In the 1950s, a Cephalosporium mould isolated off the coast of Sardinia was found to produce significant antibiotic activity and was sent to Oxford University where, as a result of the efforts of Abraham and his colleagues, the birth of the cephalosporins resulted. Important though these early findings are, they do not in themselves present any coherent pattern in the deliberate design of an antibiotic. This development had to await further discoveries in antibiotic production and an improved knowledge of mechanism of action, bacterial resistance and pharmacokinetics. Towards the end of the 1950s Chain, in collaboration with scientists at Beecham Research Laboratories, observed the production of 6-aminopenicillanic acid (6-APA (11.1)) in media in which Penicillium chrysogenum was growing when phenylacetic acid (C6H5·CH2·COOH; phenylethanoic acid) was omitted. Phenylacetic acid is the precursor of the side-chain of benzylpenicillin ((11.2) R =C6H5CH2CO) and 6-APA is the nucleus to which a side-chain is attached. This finding has had far-reaching effects since it has become the starting point in the deliberate design and synthesis of a family of penicillins having different or improved properties from existing members. This aspect is considered in more detail in Section 11.5.1. 6-APA is a naturally occuring antibiotic. In contrast, the nucleus (7aminocephalosporanic acid, 7-ACA (11.3)) of the cephalosporin group does not occur

Design of antimicrobial chemotherapeutic agents

535

naturally. Studies at Oxford revealed that C. acremonium produced more than one antibiotic: ‘cephalosporin P’ (active against Gram-positive bacteria, but subsequently shown not to be a cephalosporin) and ‘cephalosporin N’ (with activity against Gramnegative bacteria and later found to be a penicillin). Further studies disclosed that ‘cephalosporin N’ was actually contaminated with a true cephalosporin, cephalosporin C (11.4) which showed a high degree of stability to staphylococcal β-lactamase. Cephalosporin C is converted to 7-ACA by appropriate chemical means. Like 6-APA, 7-ACA can be considered as being the starting point in the development of newer antibiotics. The cephalosporins are discussed in Section 11.5.2. The production of semi-synthetic β-lactam antibiotics has formed part of an exciting era in chemotherapy. In the meantime, other semisynthetic antibiotics, e.g. some tetracyclines, have been described and some antibiotics, notably chloramphenicol and the quinolones, have been totally synthesized chemically. These are considered in later Sections. 11.3 MECHANISM OF ACTION OF CHEMOTHERAPEUTIC AGENTS The great majority of studies of the mechanism of action of chemotherapeutic agents have involved investigations of individual compounds or of drugs within a particular group, e.g. tetracyclines. A considerable amount of information is now available (Table 11.1) and will be summarized in this section, since in the space available only a bare outline can be presented. Where possible, ways in which this knowledge can be used to further the design of new antibiotics will be discussed. 11.3.1 Inhibitors of cell wall synthesis The bacterial cell wall is a complex structure (see also Section 11.4.2 and Figure 11.1). Differences occur between the walls of Gram-positive and Gram-negative bacteria, but they all contain a basal peptidoglycan (murein, mucopeptide). This consists of the amino sugars N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc) to which are attached amino acids, some in the unnatural Dconfiguration. In brief, peptidoglycan synthesis involves the stepwise addition of amino acids to MurNAc, the linking to GlcNAc to form a linear polymer and finally a cross-linking (transpeptidation, via a transpeptidase enzyme) of the linear polymers to form a rigid structure, in which the degree of cross-linking varies. A simplified example of cross-linked peptidoglycan is given in Figure 11.2(a). D-cycloserine (11.5), a structural analogue of D-alanine (11.6) inhibits two enzymes (a racemase and

Introduction to the principles of drug design and action

536

a synthetase) involved in the synthesis of the D-alanyl-D-alanine dipeptide. β-lactam antibiotics inhibit the cross-linking (transpeptidation) reaction (see Figure 11.2(b)).

Table 11.1 Mechanisms of action of antibacterial agents. Effect Example(s) Comments Inhibition of cell D-cycloserine Competitive inhibition of wall synthesis alanine racemase and synthetase β-lactams Inhibition of transpeptidases Binding to PBPs (specific) Glycopeptides Binding to peptidoglycan precursor Effect on the Polymixins Affect outer membrane of cytoplasmic Gram-negative bacteria membrane also Ionophores Specific cation Polyenic conductors: non-selective antibiotics Bind to membrane sterols in fungi (bacteria unaffected) Inhibition of protein Streptomycin Inhibits initiation stage synthesis Tetracyclines Inhibits binding of aminoacyl-tRNA to 30S ribosomal subunit Chloramphenicol Inhibits peptidyl transferase Erythromycin Inhibits translocation Puromycin Binds to peptidyl transferase: nonselective Inhibition of RNA Actinomycin D Binds to double stranded synthesis Rifampicin DNA Inhibits DNA polymerase Inhibition of DNA Mitomycin C Covalent linking to DNA synthesis Quinolones Effect on DNA gyrase Novobiocin Effect on DNA gyrase Inhibition of Sulphonamides Competitive inhibitors of tetrahydrofolate Trimethoprim dihydropteroate synthesis synthetase (see also text) Inhibits dihydrofolate reductase

Design of antimicrobial chemotherapeutic agents

537

Figure 11.1 Outer layers of (a) Gram-positive and (b) Gram-negative bacteria.

Introduction to the principles of drug design and action

Figure 11.2 Cross-linked peptidoglycan (a) and role of transpeptidase (b) in Staphylococcus aureus. MurNAc, Nacetylmuramic acid; GlcNAc, N-

538

Design of antimicrobial chemotherapeutic agents

539

acetylglucosamine; AGLA, respectively L-alanine, D-glutamine, L-lysine, Dalanine; (Gly)5. 5 molecules of glycine. Considerable progress has been made in understanding the action of β-lactam antibiotics at the molecular level. Bacterial cell membranes (inner membranes) contain several proteins, known as penicillin binding proteins (PBPs), which are associated with specific enzyme activity (Table 11.2), with which β-lactam antibiotics may combine.

Table 11.2 Function of penicillin-binding proteins (PBPs) in Escherichia coli. PBP Enzyme activity Function Result of Example(s) of βinhibiting lactam inhibiting 1A Transglycosylase, Cell wall Lysis Benzylpenicillin, transpeptidase growth cephalosporins 1B Transglycosylase, Cell wall Lysis Benzylpenicillin, (a, b, transpeptidase growth most d) cephalosporins 2. Transglycosylase, Initiation of Oval cells Mecillinam, transpeptidase cell wall imipenem growth 3. Transglycosylase, Septum Filaments Many transpeptidase formation, cephalosporins, cell division piperacillin, aztreonam 4. Carboxypeptidase Regulation ? Benzylpenicillin, endopeptidase of crossampicillin, linking imipenem Cefoxitin 5. Carboxypeptidase Regulation ? of crosslinking 6. Carboxypeptidase Regulation ? Cefoxitin of crosslinking 7/8 Unknown ? ? Penems Most of these antibiotics bind to only one or two PBPs and, very importantly, the morphological effects induced by various β-lactams are determined by the PBP to which they bind predominantly (examples are provided in Table 11.2). PBPs 1B, 2 and 3 appear to be the most important in E. coli, since binding to PBPs 4, 5 and 6 has no adverse effect on the cells. Thus, a future possibility might well be to design a combination of two β-lactam antibiotics which have different PBP specificity in order

Introduction to the principles of drug design and action

540

to achieve a synergistic result, β-lactams that bind to PBP 1 induce rapid cell lysis or sphaeroplasts (osmotically fragile forms), whereas those binding predominantly to PBP3 induce filamentation. β-lactams binding to PBP2 induce the formation of spherical, osmotically stable forms. These PBPs are associated with specific enzyme activity, as depicted in Table 11.2. Changes in PBPs may be associated with bacterial resistance to β-lactam antibiotics, e.g. (a) a laboratory strain of E. coli resistant to mecillinam possessed a PBP2 with lower affinity for this antibiotic; (b) a hospital isolate of Neisseria gonorrhoeae resistant to benzylpenicillin was found to possess PBPs 1 and 2 with a lower affinity for this drug than sensitive cells; (c) penicillin-resistant strains of pneumococci have been found to possess an altered PBP with reduced affinity for the drug. Interaction of a β-lactam with a ‘penicillin-sensitive enzyme’ (PSE) involves binding of the antibiotic with a serine group in the enzyme by opening of the β-lactam ring system. The penicilloyl (or cephalosporyl)-enzyme complex is more stable than the complex of enzyme and natural substrate (D-alanyl-D-alanine), and the enzyme is consequently trapped in an inactive form, and is thus unable to fulfil its normal role in peptidoglycan synthesis. The glycopeptides (vancomycin, teicoplanin, ristocetin) also inhibit peptidoglycan synthesis by binding to the D-alanyl-D-alanine terminus of various petidoglycan precursors, preventing the transglycosylation step by which glycan units are polymerized within the peptidoglycan. Unlike β-lactam action, the transglycosylase enzyme is not inhibited, but the complex of vancomycin with the dipeptide prevents the substrate from interacting with the active site of the enzyme. 11.3.2 Membrane-active agents The term ‘membrane-active agent’ is generally taken to mean an agent that affects the cytoplasmic membrane in micro-organisms. Gram-negative bacteria, however, also possess an outer membrane (Section 11.4.2) which may act as a penetration barrier to some drugs. The polymyxins, for example, cause the leakage of intracellular constituents by damaging the cytoplasmic membrane of Gram-negative bacteria, but they also disrupt the outer membrane lipopolysaccharide. They are highly toxic to mammalian cells. Polyenic antibiotics combine with sterols in the cytoplasmic membrane of yeasts, fungi and mammalian cells. However, nystatin and amphotericin B have some selective action against fungi because they exhibit a greater binding to ergosterol than to cholesterol (Section 11.7.1). Ionophoric drugs facilitate the passage of specific inorganic cations across the cytoplasmic membrane of Gram-positive bacteria, e.g. valinomycin and monactin are K+-conducting ionophores. Unfortunately, they show a lack of specific toxicity, as they exert the same effect on mammalian membranes also. 11.3.3 Inhibitors of protein synthesis Bacterial ribosomes have different characteristics (sedimentation coefficient of 70, i.e. are 70S, with 50S and 30S subunits) from those of mammalian ribosomes (80S, with

Design of antimicrobial chemotherapeutic agents

541

60S and 40S subunits). Most antibacterial antibiotics that are clinically important inhibitors of protein synthesis have a preferential effect on 70S ribosomes. For example, chloramphenicol affects 70S ribosomes, but not 80S. In contrast, the tetracyclines inhibit protein synthesis on isolated 70S and 80S ribosomes. The tetracyclines bind to the 30S subunit, as does streptomycin. In contrast, erythromycin binds entirely to the 50S subunit. The reason for their selective inhibition of bacterial protein synthesis in vivo resides in the energy-dependent active transport system, present in bacterial but not mammalian cells, that transports these antibacterial agents into bacteria. Bacterial protein synthesis, i.e. peptide chain extension, is carried out on 70S ribosomes. In brief, an amino acid, activated and attached to its transfer RNA (tRNA) binds to the acceptor site of the ribosome. The peptide in the donor site is transferred, as a result of peptidyl transferase activity, to the new amino acid at the acceptor site. Subsequently, translocation of the extended peptide from the acceptor site to the donor site takes place with loss of the preceding tRNA and movement of the ribosome relative to messenger RNA (mRNA) which specifies the next amino acid. The action of some antibiotics is shown in Table 11.1. 11.3.4 Inhibitors of nucleic acid synthesis Inhibitors of nucleic acid synthesis fall into two main categories: (a) those that inhibit the synthesis of purine and pyrimidine nucleotides, e.g. azaserine, a glutamine analogue, although this is not selectively toxic as it is also harmful to mammalian cells; (b) those that ihibit synthesis at the polymerization level. This involves the appropriate polymerase, with nucleic acid as a template, and is the stage where condensation of nucleoside triphosphates into a polynucleotide chain, with joining by 3’,5’-phosphodiester linkages takes place. Some drugs are intercalating agents (e.g. acridines) and inhibit DNA synthesis and DNA-dependent RNA synthesis in whole cells and in cell-free systems. Another intercalating agent, actinomycin D, has a selective action on RNA synthesis. Mitomycin C cross-links to DNA but, like actinomycin D, is also toxic to mammalian cells. Quinoline (11.7) derivatives based on 4-quinolone (11.8), 8-aza-4-quinolone (11.9), 2-aza-4-quinolone (11.10) and 6,8-diazo-4-quinolone (11.11) have been examined for antimicrobial activity. Ciprofloxacin (11.12) has been shown to be the most active 4quinolone against most aerobic Gram-negative bacteria. Nalidixic acid (11.13) has no

Introduction to the principles of drug design and action

542

Design of antimicrobial chemotherapeutic agents

543

effect on Gram-positive bacteria, but many other members are inhibitory including newer trifluorinated derivatives such as sparfloxacin: see also Section 11.6.6. The 4-quinolone derivatives inhibit the supercoiling of bacterial DNA by acting on DNA gyrase (topoisomerase II), an enzyme that nicks double-stranded DNA, introduces negative supercoils and then seals the nicked DNA. DNA gyrase is composed of four subunits, two A monomers and two B monomers. Human cells possess a topoisomerase II which, like bacterial DNA gyrase, is able to cut and seal double-stranded DNA. However, this mammalian enzyme is made up of two subunits and does not possess any supercoiling activity; its action must, therefore, be substantially different from that of bacterial DNA gyrase, which is a satisfactory explanation for the selective action of the 4-quinolone derivatives. One other point about their mechanism of action deserves comment. It has been found that concentrations well in excess of inhibitory levels reduce the bactericidal activity of nalidixic acid and some other members, e.g. norfloxacin, cinoxacin (11.14) and pipemidic (11.15) and piromidic (11.16) acids. Paradoxically, this is believed to arise as a consequence of an inhibition of RNA/protein synthesis at these high drug concentrations. Support for this contention has been obtained by using these drugs in the presence of a known inhibitor, rifampicin (11.17), of RNA synthesis, when their bactericidal effect is abolished. However, ofloxacin and ciprofloxacin are, to some extent, resistant to

Introduction to the principles of drug design and action

544

rifampicin antagonism in E. coli, since their bactericidal effects were only partially abolished. These findings have yet to be explained fully. The 4-quinolones are unaffected by plasmid-mediated resistance mechanisms and thus the frequency of clinical resistance, though increasing, is still low. Rifampicin (11.17), a useful antitubercular antibiotic and a member of the rifamycin group, does not interact with DNA but binds to and inhibits DNA-dependent RNA polymerase with a consequent specific action on RNA synthesis. RNA polymerase from rifampicin-resistant bacteria is not inhibited by the drug. Rifampicin does not bind to or inhibit mammalian DNA polymerase, and hence has a selective toxic action. 11.3.5 Antibacterial folate inhibitors Sulphonamides (11.18) act by competitively inhibiting dihydropteroate synthetase, an enzyme involved in the production from p-aminobenzoate (PAB (11.19)) of dihydropteroate (11.20) during dihydrofolate (11.21) biosynthesis. This inhibitory effect is reversed by excess PAB. It has also been found that these drugs can replace PAB as a substrate, so that they become incorporated in a false dihydropteroate or dihydrofolate.

Design of antimicrobial chemotherapeutic agents

545

Introduction to the principles of drug design and action

546

Bacteria must synthesize folate because they are unable to absorb preformed folate. PAB, an essential constituent in folate metabolism, and sulphonamides have similar chemical structures. Diaminopyrimidines structurally dissimilar from folic acid have been known for some time to be antagonists of folic acid. The best known example, trimethoprim (11.22), is a potent inhibitor of dihydrofolate reductase (DHFR), the enzyme responsible for converting dihydrofolate (11.21) to tetrahydrofolate (11.23) in E. coli but not in man. Other antifolate inhibitors of DHFR are described later in Table 11.6, from which it can be seen that tetroxoprim (11.24) also possesses significant antibacterial activity, whereas pyrimethanine is an antimalarial agent and methotrexate is used in cancer chemotherapy. Since sulphonamides and trimethoprim appear to inhibit sequential changes in tetrahydrofolate synthesis, it was a logical step to use a combination of an appropriate sulphonamide (sulphamethoxazole) with trimethoprim as an antibacterial mixture. Such a combination undoubtedly shows synergism in vitro but not necessarily in vivo. This aspect is discussed more fully in Section 11.6.5. 11.3.6 Conclusions and comments Mechanisms of antibiotic action are increasingly being understood at the molecular level. This is important for two reasons: first, such drugs have proved to be of considerable importance in elucidating the stages in microbial synthetic processes; secondly, knowledge of how a drug acts should lead logically to the development of more potent, selectively toxic chemotherapeutic agents. 11.4 BACTERIAL RESISTANCE TO CHEMOTHERAPEUTIC AGENTS The resistance of bacteria to chemotherapeutic agents has long posed a problem and will continue to do so for the foreseeable future. Much is known about the mechanisms of bacterial resistance, and this section will provide a summary of the available information, because the design of new antibiotics is, to a considerable extent, linked to methods of combating this problem. Bacterial resistance to antibiotics may be either intrinsic, i.e. a natural property of the organism often associated with cellular impermeability, or be acquired, either by mutation in sensitive cell populations or by resistance transfer from one cell to another. Space does not permit a full discussion of resistance expression at the underlying biochemical level, but a summary is provided in Table 11.3. This table demonstrates that bacteria can present resistance to a chemotherapeutic agent by virtue of many different mechanisms. Some of these mechanisms, in particular enzyme-mediated resistance (Section 11.4.1), impermeability (11.4.2) and target enzyme affinity (11.6.5) will be considered in this chapter.

Design of antimicrobial chemotherapeutic agents

547

11.4.1 Enzyme-mediated resistance Resistance of bacteria to β-lactam antibiotics may be associated with enzymes termed β-lactamases. Some bacteria are capable of producing enzymes that modify streptomycin or other aminoglycoside antibiotics. A description, therefore, of the various types of enzymes is important in understanding the development and design of antibiotics. 11.4.1.1 β-lactamases β-lactamases occur widely in nature. They are produced by various Gram-positive and Gram-negative bacteria. In Gram-positive organisms such as staphylococci and Bacillus species, β-lactamase is an inducible enzyme, with low concentrations of various β-lactam antibiotics acting as appropriate inducers, and is released extracellularly. In contrast, β-lactamases of Gram-negative bacteria are usually constitutive, but are induced by high concentrations of appropriate inducing agents in organisms such as Pseudomonas aeruginosa and Enterobacter cloacae. The βlactamases of Gram-negative bacteria are

Table 11.3 Expression of resistance to antibiotics. Expression of Example(s) Comments resistance Enzymatic Some β-lactam See 11.4.1.1 inactivation antibiotics See 11.4.1.3 Chloramphenicol Some β-lactam Enzymatic See 11.4.1.1 antibiotics trapping See 11.4.1.2 Some Enzymatic See 11.4.2 aminoglycoside modification Reduced ability of cells to antibiotics Bacterial take up drugs impermeability1 Some β-lactam Plasmid-mediated decreased antibiotics drug accumulation Aminoglycoside Difficulty in entering Gramantibiotics negative cells Tetracyclines, chloramphenicol, fusidic acid Hydrophobic antibiotics: novobiocin, actinomycin D, erythromycin,

Introduction to the principles of drug design and action

548

rifampicin

Antibiotic efflux Tetracyclines

Energy-dependent efflux of accumulated drugs Decreased affinity β-lactam antibiotics Altered PSEs/PBPs of target enzymes Trimethoprim Altered dihydrofolate Sulphonamides reductase Altered dihydropteroate synthetase Alteration in Streptomycin Protein S12 component of binding site 30S ribosomal subunit determines sensitivity or resistance Erythromycin Ribosomes from resistant cells have lower affinity, resulting from enzymatic methylation of adenine in 23S rRNA Glycopeptides Acquired ligase produces altered peptidoglycan precursors with lower affinity. 1 Depends on chemical nature of drug and on type of organism intracellular, being located in the periplasm situated between the inner and outer membranes, and are less potent enzymes than those produced by Gram-positive bacteria. They may be chromosomally or plasmid-mediated (Section 11.4.3). The effects of β-lactamases generally on susceptible penicillins and cephalosporins are shown in Figures 11.3 and 11.4 respectively. Susceptible penicillins are converted to the corresponding penicilloic acid (11.25) which is inactive; this results from an opening of

Design of antimicrobial chemotherapeutic agents

549

Figure 11.3 Effect of β-lactamases on susceptible penicillins.

Figure 11.4 Effect of β-lactamases on susceptible cephalosporins. the β-lactam ring (see Figure 11.3). The situation with susceptible cephalosporins is more complex. Opening of the β-lactam ring again occurs (11.26) and this is accompanied by expulsion of the group at R2 (except in cephalexin, where R2 is H). The molecule finally breaks up into fragments (see Figure 11.4). Extensive studies have shown that several types of β-lactamases exist among Gram-negative bacteria. These have been classified into several different groups on the basis of their substrate profile (the rate at which different β-lactam substrates are inactivated) and whether their activity is inhibited by p-chloromercuribenzoate (PCMB) and cloxacillin. A simplified summary of the different types is presented in Table 11.4. With several of the newer cephalosporins, Vmax values for hydrolysis by chromosomal cephalosporinases are extremely low. However Km values demonstrate that these antibiotics have very high affinity for these enzymes. It has ben proposed that ‘enzyme trapping’ in the periplasm, without drug hydrolysis, could be a resistance mechanism, but this is likely only where considerable amounts of β-lactamase are

Introduction to the principles of drug design and action

550

produced and the organism is poorly permeable to the antibiotic and the drug itself poorly diffusible. 11.4.1.2 Aminoglycoside-modifying enzymes β-lactam antibiotics are destroyed by β-lactamases. In contrast, aminoglycoside antibiotics are not totally inactivated by plasmid-encoded enzymes but rather modified in the outer regions of the resistant cell and are thus not bound to the target (ribosome). Only a small proportion of the external aminoglycoside need be modified for resistance to be expressed. Aminoglycoside-modifying enzymes are of three types: (i) acetyltransferases (AAC), which transfer an acetyl group from acetylcoenzyme A to susceptible -NH2 groups in the antibiotic; (ii) adenylyltransferases (AAD), which transfer adenosine monophosphate (AMP) from adenosine triphosphate (ATP) to susceptible -OH groups in the antibiotic; (iii) phosphotransferases (APH), which phosphorylate susceptible -OH groups, ATP acting as the source of phosphate.

Table 11.4 Types of beta-lactamases. Group of Preferred Inhibited by: Representative 1 enzyme substrate Clavulanic EDTA enzymes acid 1. Cephalosporin − − AmpC from Gram negatives 2a Penicillins + − Penicillinases from Gram positives 2b Penicillins, + − TEM-12, TEM-2, cephalosporins SHV-1 from Gram negatives + – TEM-3 to TEM-26 2be Penicillins, cephalosporins monobactams 2br Penicillins +/− − TEM-30 to TEM-36 2c Penicillins, + − PSE-1, PSE-3, PSEcarbenicillin 4 2d Penicillins, +/− − OXA-1 to OXA-11 carbenicillin 2e Cephalosporin + − Inducible cephalosporinases from Proteus vulgaris

Design of antimicrobial chemotherapeutic agents

2f

Penicillins, cephalosporins, carbapenems

+

551

–

NMC-A from Enterobacter cloacae, Sme-I from Serratia marcescens 3 Most β-lactams, – + L1 from including Xanthomonas carbapenems maltophilia, CcrA from Bacteroides fragilis 4 Penicillins − ? Penicillinase from Pseudomonas cepacia 1 Based on the Bush-Jacoby-Medeiros scheme (1995) 2 Plasmid-encoded β-lactamases (TEM, PSE, OXA, SHV)

Kanamycin (11.27) can be modified in at least six different ways. Thus, the development of aminoglycosides such as amikacin (Section 11.6.1 (11.28)), with considerably greater resistance to many of these enzymes, is a significant improvement to the currently available range of antibiotics.

Introduction to the principles of drug design and action

552

11.4.1.3 Chloramphenicol-inactivating enzymes Some bacteria, notably R+ Gram-negative rods and staphylococci containing transducible plasmids (Section 11.4.3) can produce an enzyme, chloramphenicol acetyltransferase, that acetylates the hydroxyl groups in the side-chain of chloramphenicol (11.29) to produce initially 3-acetoxychloramphenicol (11.30) and finally 1,3-diacetoxychloramphenicol (11.31) which is inactive. In S. aureus, the acetyltransferase is an inducible enzyme, whereas in Gram-negative bacteria the enzyme is constitutively synthesized. The design of new chloramphenicol derivatives (in which the terminal hydroxyl is replaced by fluorine) might have been a major advance, since these do not function as substrates for chloramphenicol acetyltransferase. Unfortunately, these derivatives were found to be toxic.

Design of antimicrobial chemotherapeutic agents

553

11.4.2 Outer membrane barrier Another way in which bacteria show resistance to antibiotics is by an intrinsic type of resistance associated, in Gram-negative bacteria especially, with the cell wall. The cell envelope of such organims is considerably more chemically complex than the envelopes of Gram-positive bacteria, which consist predominanatly of peptidoglycan and teichoic acids (Figure 11.1a). Figure 11.1b presents a simplified version of the cell envelope of Gram-negative cells. The outer membrane (OM) acts as a barrier to the entry of many high molecular weight hydrophilic molecules, whereas low molecular weight (less than about 550– 650) hydrophilic antibiotics can pass through hydrophilic channels, known as porins, which consist of OM protein molecules. Removal of much of the outer membrane lipopolysaccharide by exposure of cells to ethylenediamine tetraacetate (EDTA), a chelating agent with a strong affinity for envelope Mg2+, renders the organisms sensitive to several types of hydrophobic antibiotics. This is believed to result from a spontaneous re-orientation of the phospholipid molecules, with the head-groups on the outside, therby forming a classical type of membrane structure. The major OM proteins of E. coli include lipoprotein, OM proteins (Omp) A, C and F and protein A. OmpC and OmpF are porins. D-cycloserine, a low molecular weight hydrophilic antibiotic, passes across the OM by porin transport. Chloramphenicol, also of low molecular weight, crosses the outer membrane via OmpF pores. β-lactam antibiotics show an inverse correlation between hydrophobicity and the concentration diffusing across the membrane into the periplasm. Tetracycline itself utilizes the OmpF porin; however, as drug hydrophobicity increases, E. coli becomes increasingly resistant. Minocycline is a hydrophobic tetracycline, but compensates for its poor entry into Gram-negative bacteria by being a more effective inhibitor of protein synthesis. Hydrophobic aminoglycosides are thought to increase penetration by interacting with the OM, causing an increase in permeability, a phenomenon termed ‘self-promoting’ uptake. Porin-deficient Salmonella typhimurium cells are resistant to cephaloridine, a βlactam that normally enters Gram-negative bacteria readily. The porins of Ps. aeruginosa were claimed to be much larger than those of other Gram-negative bacteria (see above) with an exclusion limit of about 9000. It would, therefore, be logical to predict that Ps. aeruginosa is more sensitive to antibiotics than other Gram-negative bacteria, which is the exact opposite of known fact. Current opinion is, however, that the porins in this organism are smaller than those in other Gram-negative bacteria and have an exclusion limit of ca. 300. 11.4.3 Transferable resistance Genetic information can be transferred from one bacterial cell to another by three ways: transduction (phage-mediated), transformation (absorption of free DNA by ‘competent’ cells) and conjugation. The last is generally accepted to be the most common mechanism in the spread of bacterial resistance and thus will be the only one considered here. In the late 1950s it was observed during a dysentery outbreak in Japan that a Shigella strain was resistant to four chemically unrelated antibiotics. It was unlikely

Introduction to the principles of drug design and action

554

that resistance to all four drugs in this strain arose by spontaneous mutation. Other multiple-resistant strains of Gram-negative bacteria were also isolated and it was proposed that blocks of resistance genes were being transferred between bacteria of different species. For this transfer to occur, cell-to-cell contact is made and DNA from one cell, the donor, is transferred to another cell, the recipient, via a conjugation tube. It is known that some bacteria carry small, autonomously replicating extrachromosomal elements known as plasmids. Some of these plasmids mediate their own transfer by conjugation (conjugative plasmids). Non-conjugative plasmids can also be cotransferred by a process known as mobilization. Many of these plasmids carry resistance genes and, therefore, transfer of this DNA from donor to recipient cells will specify production in the latter of, for example, drug-inactivating or modifying enzymes or of additional membrane proteins that prevent uptake of tetracyclines. Plasmids may be considered as being the vectors (carriers) of resistance genes, but the genes may be located on a discrete transposable element called a transposon, which is a DNA element that can insert into several sites on a genome. Sometimes known as ‘jumping genes’, transposons are capable of moving from plasmid to plasmid or from plasmid to chromosome. In staphylococci, phages act as important vectors in the spread of transposons. One or more drug resistance determinants may be located on a transposon (Table 11.5); furthermore, a transposon can enter a different species and remain stable therein, even if the vector is lost. The reason for this is that the transposon can be incorporated into the host’s chromosome or resident plasmid. This is an important point, because it helps to explain the spread of resistance genes in widely different genera. 11.4.4 Conclusions and comments Bacterial resistance to antibiotics is often achieved by the constitutive possession or inducibility of drug-inactivating or -modifying enzymes. This problem can, at least to some extent, be overcome by designing new drugs (e.g. by chemical alteration of an existing molecule) that (a) are insusceptible to this enzyme attack or (b) will inactivate the enzyme concerned thereby protecting enzyme-labile antibiotics that, in the absence of the enzyme, would be highly active antibacterially. Some degree of success has been achieved in both aspects, as discussed in Section 11.5.

Table 11.5 Examples of chromosomally and plasmid-mediated resistance to antibiotics. Chromosomally Plasmid-mediated Transposons2 mediated resistance resistance1 only

Design of antimicrobial chemotherapeutic agents

Methicillin Quinolones Rifampicin

1 2

Aminoglycosideaminocyclitols β-lactams Tetracyclines Sulphonamides Trimethoprim Chloramphenicol Erythromycin Fusidic acid

555

Single, e.g. ampicillin, chloramphenicol, tetracycline Multiple, e.g. ampicillin+ streptomycin+ sulphonamide

Resistance may also be chromosomally mediated Multiple resistance genes may be carried on a transposon

Another problem concerns the lack of penetration of many drugs into Gramnegative bacteria. On the basis of current knowledge, it would seem logical that any design of new agents should at least consider the need for hydrophilic, low molecular weight compounds that can penetrate the OM of these cells via aqueous channels (porins). In this context, the development of amino acid analogues with antibacterial activity is worthy of consideration. These penetrate poorly into bacteria but when part of or attached to small peptides are transported into cells via relatively non-specific permeases. One such example is alaphosphin (L-alanyl-L-1-aminoethyl-phosphonic acid, alafosfalin (11.32)) which is rapidly accumulated by, and concentrated within, bacteria, where it is converted to L-1-aminoethylphosphonic acid (11.33) which acts as an inhibitor of peptidoglycan synthesis (Section 11.3.1). Alaphosphin belongs to a group of compounds, the phosphonopeptides, which are peptide mimics with Cterminal residues that simulate natural amino acids. Their mechanism of action results from transport into the bacterial cell followed by release of the alanine mimetic. These agents were considered as being an important concept in designing new antibacterially active compounds, but unfortunately these findings do not appear to have been followed by the development of any significant new drugs. Despite extensive research, the design of clinically effective antimetabolites of this type has generally been disappointing. 11.5 DESIGN OF β-LACTAM ANTIBIOTICS It was pointed out in Section 11.2 that 6-APA (11.1) and 7-ACA (11.3) formed the starting point in the development of new penicillins and cephalosporins, respectively. This Section will continue this consideration and will also explain current thinking in the design of new β-lactam antibiotics (or of combinations of β-lactams) that can effectively counter the unwanted effects of β-lactamases (Section 11.4.1.1) and bacterial impermeability (Section 11.4.2). In addition, pharmacokinetic aspects will be considered where relevant.

Introduction to the principles of drug design and action

556

11.5.1 Penicillins The original penicillins, benzylpenicillin (penicillin G, (11.2)) and phenoxymethylpenicillin (penicillin V, (11.34)) suffered clinically in the context that both were narrow spectrum, β-lactamase-labile antibiotics. The former is also acidlabile and is thus normally administered by injection, whereas penicillin V is acidstable and is thus given orally. The first semisynthetic penicillin produced from 6APA of any clinical consequence was methicillin (11.35) which, although inactive against Gram-negative

bacteria, possesses significant activity against β-lactamase staphylococcal producers. Its intrinsic potency is, however, less than that of benzylpenicillin against non-βlactamase staphylococci, and acid instability precludes the oral usage of methicillin.

Design of antimicrobial chemotherapeutic agents

557

Methicillin was soon followed by ampicillin (11.36), and this was another significant advance, because ampicillin was the first semisynthetic penicillin to possess marked activity against Gram-negative organisms (although Ps. aeruginosa is resistant). Ampicillin is stable to acid and is administered orally or by injection, but is susceptible to the β-lactamases produced by S. aureus and most Gram-negative bacteria. There then followed many other important new semisynthetic penicillins, such as cloxacillin (11.37) (the first oral, β-lactamase stable penicillin, but again without significant action on Gram-negative cells); its derivative flucloxacillin (11.38) claimed to give higher blood levels; carbenicillin (11.39) the first penicillin with activity against Ps. aeruginosa and its derivative ticarcillin (11.40); amoxycillin (11.41) with a similar spectrum to ampicillin, but which is much better absorbed; temocillin (11.42) with a longer half-life allowing twice-daily dosage; and several others. The design of all the semisynthetic penicillins has had one common goal: to achieve, by the introduction of a different R group (11.2), a new antibiotic with an improved spectrum of activity and/or enhanced stability to β-lactamases. This deliberate design concept, therefore, has achieved some notable successes. The above examples all illustrate development of new penicillins by substitution at the 6-position in the molecule. Position 3 is also uniquely important, since the introduction of various groups here has led to the design of new esters (‘pro-drugs’: see Chapter 7)

Introduction to the principles of drug design and action

558

which are hydrolyzed by enzyme action after absorption from the gut mucosa to give the active antibiotic. Esters (pivampicillin, talampicillin and bacampicillin: see (11.43), (11.44) and (11.45), respectively) at position 3 of the ampicillin molecule break down in vivo to produce higher blood levels of ampicillin than would be obtained if ampicillin itself had been given at an equivalent concentration. Carbenicillin is not absorbed when given orally but esters (carfecillin (11.46), carindacillin (11.47)) in the side-chain at position-6, when given orally will hydrolyse in vivo to give a similar blood level to that obtained with an equivalent dose of carbenicillin given intramuscularly. Thus, ‘prodrugs’ form a useful development in βlactam design. Substituted penicillins, such as piperacillin (11.48), azlocillin (11.49) and mezlocillin (11.50), appear to combine the spectra and degree of activity of ampicillin and carbenicillin. Mecillinam (11.51) a 6β-amidinopenicillin, has limited activity against Gram-positive bacteria but is active against Gram-negative organisms. It binds preferentially to PBP2.

Design of antimicrobial chemotherapeutic agents

559

Thus, alterations in the molecule (and especially at positions 6 and 3) can produce penicillins with changes in microbiological and/or pharmacological properties.

Substitution at other sites has also been examined in the quest for improved design: removal of the sulphur atom of the thiazolidine ring usually leads to a reduction of activity, although the oxapenicillin (clavam) clavulanic acid (Section 11.5.4.2) is an important β-lactamase inhibitor. Additionally, the carbapenems (Section 11.5.5.2), in which sulphur is isosterically replaced by a methylene group but which have a double bond in the 5-membered ring, may possess significant activity. Substitution at the C-5 position reduces antibacterial activity, and substitution at the C-2 locus produces penicillins with activity against Gram-positive but not Gram-negative bacteria. 11.5.2 Cephalosporins Current research on cephalosporins and cephamycins (methoxycephalosporins) is proceeding at a bewildering pace. The cephalosporins may be considered as semisynthetic derivatives of 7-ACA (11.3). Several of the early (first generation) cephalosporins differed more in their pharmacokinetic than in their antibacterial properties. Subsequent developments have been to improve: (i) antibacterial activity, especially in the context of increasing resistance to βlacta-mases produced by Gram-negative bacteria, although it must be added that decreased enzyme lability is sometimes paralleled by a reduction in antibacterial potency; (ii) pharmacokinetic properties by making appropriate substitutions in the molecule, especially at positions 3 and 7.

Introduction to the principles of drug design and action

560

11.5.2.1 Structure-activity relationships The cephalosporins (Δ3-cephalosporins) and penicillins are structurally related in that the β-lactam ring is fused to different rings. The position of the double bond in Δ3cephalosporins is very important, since Δ2-cephalosporins (11.52), irrespective of the composition of the side-chains, are not significantly antibacterial. In contrast, Δ2penicillins are highly active against Gram-positive and Gram-negative bacteria (Sections 11.5.4.2 and 11.5.5.2). The 7α-methyl cephalosporin derivatives have a greatly reduced antibacterial activity, whereas the introduction of a 7α-methoxy group gives compounds (cephamycins) with high antibacterial activity, and possessing considerable stability to most β-lactamases. There is, however, a rapid decrease in activity as the size of the ether group is increased. Oxacephems (oxacephamycins; see Section 11.5.5.1) have been produced by synthetic means and may have high antibacterial activity, including β-lactamase stability. An example of the interplay of various factors in antibacterial activity is demonstrated by the following findings. 7α-Methoxy substitution of cefuroxime, cefamandole and cephapirin gives reduced activity agaist E. coli because of a lower affinity for PBPs and not because of reduced permeability. In contrast, similar substitution in cefoxitin enhances activity because of greater penetration through the OM barrier rather than an increased affinity for PBPs. Cephalosporins are generally less sensitive than penicillins to inactivation by staphylococcal β-lactamase, but may be susceptible to β-lactamases produced by some Gram-negative bacteria. Additional information is provided in Sections 11.4.1.1, including Table 11.4, and 11.5.3. Various characteristics involved in the activity of ceftazidime are depicted in (11.53).

Design of antimicrobial chemotherapeutic agents

561

Introduction to the principles of drug design and action

562

The 3-acetoxymethyl compounds cephalothin (11.54), cephacetrile (11.55) and cephapirin (11.56) have different 7-acyl groups, which are monosubstituted acetamido groups, and have similar antibacterial activity. They are active against Gram-positive bacteria and against β-lactamase-negative Gram-negative organisms. 11.5.2.2 Pharmacokinetic properties The 3-acetoxymethyl compounds are converted in vivo by esterases to the antbacterially less active 3-hydroxymethyl derivatives and are excreted partly as such. The rapid excretion means that such cephalosporins have a short half-life in the body. Replacement of the 3-acetoxymethyl group by a wide variety of groups has rendered other cephalosporins much less prone to esterase attack. For example, cephaloridine (11.57) has an internally compensated betaine group at position 3 and is metabolically stable. It gives higher and more prolonged blood levels than cephalothin. Cephalosporins such as 3-acetoxymethyl derivatives (11.54), (11.55) and (11.56), cephaloridine (11.57) and cefazolin (11.58) are inactive when given orally. For good oral absorption, the 7-acyl group must be based on phenylglycine and the amino group must remain unsubstituted. At position 3, the substituent must be small, non-polar and stable, with a methyl substituent considered desirable (although this can decrease antibacterial activity). Earlier examples of oral cephalosporins are cephalexin (11.59), cefaclor (11.60) and cephradine (11.61). Although cephalexin has some degree of resistance to β-lactamases produced by Gram-negative bacteria, none of these oral cephalosporins possess a significant degree of resistance to these enzymes. More recent oral cephalosporins such as loracarbef (11.62), cefixime (11.63), cefpodoxime (11.64) and ceftibuten (11.65) show increased stability to Gram-negative βlactamases. Cefpodoxime is an absorbable ester (see Chapter 7; ‘pro-drugs’). During absorption, esterases remove the ester side-chain, liberating the active substance into the blood. Cefixime and

Design of antimicrobial chemotherapeutic agents

563

ceftibuten are non-ester drugs and loracarbef is a new oral carbacephem (sulphur at position 1 replaced by carbon). Loracarbef is highly active against Gram-positive organisms, including staphylococci. The others are characterized by activity against Gram-negative and -positive organisms. None of the drugs is active against Ps. aeruginosa. Parenterally administered cephalosporins which are metabolically stable and which are resistant to many types of β-lactamases include cefamandole (11.66), cefotaxime (11.67), cefoxitin ((11.68) (which has a 7α-methoxy group) and cefuroxime (11.69). An esterified derivative of the latter, cefuroxime axetil, retaining all the proerties of the parent molecule, is also now available for oral administration. Other cephalosporins include ceftazidime (see (11.53) where the roles played by various components are indicated), ceftriaxone (11.70), ceftizoxime (11.71), cefpirome (11.72), cefepime (11.73) and cefsulodin (11.74). Cefsulodin is a narrowspectrum drug only used for pseudomonal infections. The development of new cephalosporins continues apace and those listed are meant to serve as examples since space precludes an exhaustive list.

Introduction to the principles of drug design and action

564

Side-chains containing a 2-aimnothiazolyl group at R’ (e.g. cefotaxime, ceftriaxone, ceftizoxime and ceftazidime) yield cephalosporins with enhanced affinity for PBPs of Enterobacteriaceae and streptococci.

Design of antimicrobial chemotherapeutic agents

565

11.5.3 β-lactamase stability This short section attempts to summarize the facts presented in Sections 11.5.1 and 11.5.2. β-lactam antibiotics, both penicillins and cephalosporins, with simple side-chains (Ar-αCH2-CO-NH-) are usually sensitive to β-lactamases, whereas the incorporation of the α-carbon atom into an aromatic ring (e.g. methicillin) increases resistance.

The presence of an additional substituent, e.g. methoxy, in the 6-(penicillins) or 7α(cephalosporins) position greatly increases β-lactamase resistance, although intrinsic antibacterial activity may be decreased. Cefoxitin (11.68), however, has a broad spectrum of activity and is very resistant to β-lactamases. Newer 7αmethoxycephalosporins have been described that have the same spectrum of activity as cefoxitin but are more active. The 1-oxacephem antibiotic latamoxef (moxalactam: (11.75) see Section 11.5.5.1) has oxygen instead of sulphur at position 1, which would tend to make it less chemically stable and more enzyme-labile; however, the presence of the 7α-methoxy group, as in cefoxitin (11.68) stabilizes the molecule. Absence of sulphur or oxygen in the fused ring results in increased resistance to βlactamases (see Section 11.5.4; loracarbef (11.62)). Generally, changes at C2 and C3 in a penicillin or cephalosporin do not affect resistance to a β-lactamase. 11.5.4 β-lactamase inhibitors An exciting concept in antibiotic therapy is the possibility of using β-lactam inhibitors clinically. This is not a revolutionary idea since many of the older penicillins were able to inhibit the enzyme (see Section 11.5.4.1). The development of clavulanic acid (see Section 11.5.4.2) bridged the gap between theoretical desirability and actual practice, and the introduction in 1981 of an antibiotic mixture (amoxyclav; Augmentin®, Beecham Research Laboratories) consisting of clavulanic acid with the β-lactamase sensitive penicillin, amoxycillin, provided the clinician with a new weapon in his armoury against microbial infection. A combination of clavulanic acid with the broad-spectrum penicillin ticarcillin has also been introduced. Other combinations are listed below (Section 11.5.4.3). 11.5.4.1 β-Lactams as inhibitors Some of the earlier penicillins (e.g. cloxacillin and methicillin) and cephalosporins (such as cephalosporin C) were found to inhibit Bacillus cereus β-lactamase and later

Introduction to the principles of drug design and action

566

were shown to be active against some β-lactamases elaborated by Gram-negative bacteria. This inhibition was competitive in nature, and marked potentiation of a βlactamase-sensitive β-lactam antibiotic could be achieved in vitro. The problem, nevertheless, was two-fold: (a) high concentrations of inhibitor were necessary, (b) no single antibiotic then available was able to inhibit a wide range of β-lactamases. Thus, it was then not possible to design for clinical use an antibiotic mixture consisting of a β-lactam and a β-lactamase inhibitor. 11.5.4.2 Naturally occurring β-lactamase inhibitors The β-lactamase-inhibitory properties of cephalosporin C (described above, Section 11.5.4.1), itself produced by a micro-organism, stimulated a search for other naturally occurring β-lactamase inhibitors. In principle, this technique has involved testing culture fluids in which Streptomyces species have been growing for their ability to inhibit the β-lactamase produced by a specific strain of Klebsiella aerogenes. Research investigations at Beecham Research Laboratories in the UK and studies elsewhere, notably in the United States and Japan, have shown the production of β-lactamase inhibitors in the culture fluids of Streptomyces olivaceus and Streptomyces clavuligerus. Three β-lactamase-inhibiting acidic substances, termed the olivanic acids (general structure (11.76)) have, with some difficulty, been isolated from the culture fluids of Streptomyces olivaceus. These possess potent activity against various types of βlactamases, when they act as inhibitors, and are also broad-spectrum antibiotics in their own right. The olivanic acids, characterized as closely related members of a new class (1-carbapenems (11.77)) of fused β-lactams, are analogues of penicillins or clavulanic acid (11.78) where sulphur or oxygen, respectively, has been replaced by a methylene group. An antibiotic thienamycin (11.79) with a similar structure to the olivanic acids has been isolated from Streptomyces cattleya. It is of interest to note that thienamycin, a very broad-spectrum antibiotic, is often a poor β-lactamase inhibitor (see Section 11.5.5.2). It would thus seem logical for a member of the olivanic acid group to be utilized as a β-lactamase inhibitor in combination with a β-lactamase-sensitive antibiotic in the design of a new antibiotic mixture. Unfortunately, the olivanic acids are produced in low yields and, as mentioned above, there have been problems with their isolation. Attention has thus been focused on other types of β-lactamase inhibitors. One of these, clavulanic acid, was sufficiently promising for a comprehensive investigation to be undertaken. Clavulanic acid (a clavam) (11.78) is a fused bicyclic compound containing a β-lactam ring; it is similar in structure to the penicillins except that it contains oxygen in place of sulphur, i.e. an oxazolidine, instead of a thiazolidine, ring. It is produced in higher yields than the olivanic acids, but has a poor antibacterial action; it is, however, a potent inhibitor of staphylococcal β-lactamase and of βlactamases produced by Gram-negative bacteria, in particular those with a ‘penicillinase’ rather than a ‘cephalosporinase’ type of action. Clavulanic acid inhibits β-lactamases of Group 2 (Table 11.4) but is a poor inhibitor of Groups 1, 3 and 4. Clavulanic acid effects a progressive inhibition of β-lactamase, the initial effect probably being competitive in nature, this being followed by a phase of rapid

Design of antimicrobial chemotherapeutic agents

567

inactivation. Studies with different types of β-lactamases have demonstrated that this inhibitor is a kcat inhibitor (see Chapter 8) acting in a competitive and irreversible manner.

11.5.4.3 Synthetic β-lactamase inhibitors Penicillanic acid derivatives are also known to inhibit β-lactamases. Penicillanic acid sulphone (11.80) is a β-lactamase inhibitor that protects ampicillin from hydrolysis by staphylococcal β-lactamase and some, but not all, of the Gram-negative types depicted in Table 11.4. It is however, a less active inhibitor than clavulanic acid. Tazobactam (11.81) is a penicillinic acid sulphone derivative marketed as a combination with piperacillin (11.48). Alone, it has poor intrinsic antibacterial activity but is comparable to clavulanic acid in inhibiting β-lactamase activity. Sulbactam (11.82) is a semisynthetic 6-desaminopenicillin sulphone structurally related to tazobactam. It is an effective inhibitor of many β-lactamases and is also active alone against certain Gram-negative bacteria. It is combined with ampicillin (11.36) for clinical use. 6Bromopenicillanic acid (11.83) inhibits some types of β-lactamases. It is essential that the β-lactamase inhibitor and the penicillin have similar pharmacokinetics, i.e. absorption rates, to arrive together at the site of action in the body. 11.5.4.4 Structure-activity relationships in β-lactamase inhibitors The β-lactam ring of β-lactamase inhibitors appears to mimic the β-lactam ring of substrate molecules, fitting closely into the catalytic centre of the enzyme. Amino derivatives of clavulanic acid have potent inhibitory activity. A variety of β-lactam molecular types show high activity against cell-free β-lactamases, but not when used against whole cells of Gram-negative bacteria. The reason for this poor effect against periplasmic β-lactamase is their low OM penetrability (see Section 11.4.2). Clavulanic acid is a small hydrophobic molecule. Strongly acidic olivanic acids penetrate the outer membrane less well than clavulanic acid.

Introduction to the principles of drug design and action

568

11.5.4.5 β-lactamase inducers In some bacterial species, exposure to certain β-lactams (e.g. cefoxitin, imipenem) may induce the formation of large amounts of chromosomally mediated Class I

‘cephalosporinase’ type β-lactamases (see Table 11.4). These may antagonize the activity of a second β-lactam used simultaneously. Continued production of the enzyme is necessary for a continued decrease in susceptibility to the second β-lactam. 11.5.4.6 Mutual pro-drugs A pro-drug (see Section 11.5.1 and Chapter 7) is an inactive compound that is converted in vivo to an active form. Pro-drugs of β-lactams are usually esters which are broken down by mammalian esterases. A potentially exciting development has been the synthesis of linked esters of penicillins and β-lactamase inhibitors to produce what are termed mutual pro-drugs. These must be well absorbed and the two active constituents released in equal amounts. One problem is that the maximum antibacterial activity is not necessarily achieved at a 1:1 ratio. It has been suggested that it might even be possible to develop a mutual pro-drug (an ester of a penicillin with a βlactamase inhibitor) to be given in combination with the pro-drug of the non βlactamase inhibitor moiety. This is an interesting theoretical possibility, but it might itself pose many practical problems in formulation. 11.5.5 Other β-lactam ring systems For several years, investigations have been carried out on modifications of β-lactam drugs in order to improve and extend their antibacterial activity or to alter their

Design of antimicrobial chemotherapeutic agents

569

pharmacokinetic properties. The result has been the development of an impressive array of β-lactam antibiotics with new ring systems which can be of value clinically in their own right, or may serve as starting points for the design of still more important antibiotics, or, again, may be potent β-lactamase inhibitors (see Section 11.5.4). 11.5.5.1 1-Oxacephems A highly active 1-oxacephem (latamoxef (11.75)) has been obtained semisynthetically from penicillin. The sulphur atom in the cephalosporin dihydrothiazone ring is isosterically replaced by an oxygen atom. Latamoxef shows similarities to other βlactam antibiotics, e.g. a 7α-methoxygroup (as in cefoxitin), a p-hydroxybenzyl group (amoxycillin), an a-carboxylic acid group (carbenicillin) and a 3-(1-methyltetrazol-5ylthiomethyl) substituent (cefamandole). Latamoxef is an effective kcat inhibitor of some β-lactamases and a competitive inhibitor of others. 11.5.5.2 Penems In the penems, the double bond in the dihydrothiazine ring of the cephalosporins has been replaced in the corresponding (thiazolidine) ring of the penicillins. In the carbapenems (11.77), a methylene group has replaced the -S- atom at position 1 in the penicillin molecule. Examples have already been dealt with (olivanic acids, thienamycin) although it must be noted that, depite its high activity and β-lactamase resistance, thienamycin (11.79) suffers the disadvantage of being chemically unstable. An N-formimidoyl derivative, imipenem (11.84), overcomes this problem. Imipenem has a broad spectrum, covering most Gram-positive and -negative aerobic and anaerobic bacteria and is highly resistant to β-lactamases (although it may act as an inducer: see Section 11.5.4.5). Imipenem is administered intravenously with cilastatin, which acts as a specific, competitive inhibitor of the enzyme, dehydropeptidase-I, that metabolizes imipenem in the kidney. A new derivative, meropenem (11.85) containing a dimethylcarbamoylpyrrolidine ring, has been shown in vitro to be more active than imipenem (11.84). Of particular note is its activity against Ps. aeruginosa. 11.5.5.3 Nocarcidins A novel group of β-lactam antibiotics, the nocarcidins (11.86) have been isolated from a strain of nocardia. This group has been characterized into seven closely related compounds (viz. nocarcidins A-G), with nocarcidin A (11.86) the most active. In vitro, nocarcidin A is less active than carbenicillin against Gram-negative bacteria and has no effect on Gram-positive organisms. In vivo, however, nocarcidin A is more active than carbenicillin because the potency of the former is increased in the presence of neutrophils, one type of phagocytic cell.

Introduction to the principles of drug design and action

570

11.5.5.4 Monobactams The monobactams are monocyclic β-lactam antibiotics produced by bacteria. They have been isolated from bacteria using as test organism a strains of Bacillus licheniformis which is specific for, and highly sensitive (100 ng ml-1) to, molecules containing a β-lactam ring. On the basis of the novel nucleus (3-aminomonobactamic acid, 3-AMA (11.87) possessed by these antibiotics a potent monobactam has been synthesized. This is known as aztreonam (11.88). It is highly active against most Gram-negative bacteria and is very stable to most types of β-lactamases, including staphylococcal, although interestingly- it is without effect on the growth or viability of S. aureus strains. Its lack of effect on staphylococci is believed to result from its predominant effect on PBP3 in Gram-negative organisms since this PBP is absent from staphylococci. 11.5.5.5 Carbacephems The newest class of β-lactam antibiotics comprises the carbacephems. An oral highly active compound (loracarbef (11.62)) has been described already. The sulphur in the

Design of antimicrobial chemotherapeutic agents

571

cephalosporin dihydrothiazine ring is replaced by a methylene group (cf. carbapenems, Section 11.5.5.2). Loracarbef is highly active against Gram-positive bacteria, including staphylococci, but shows poor activity against Gram-negative bacteria. 11.6 DESIGN OF OTHER ANTIBACTERIAL AGENTS Whilst it is probably true to state that β-lactam antibiotics occupy the greatest attention in the treatment of bacterial infecctions, it would be incorrect to imply that other antimicrobial agents do not have an important role to play. Several other groups of antibiotics exist, notably the aminoglycosides, tetracyclines, macrolides, polymyxins, lincomycins, rifamycins, quinolones together with chloramphenicol, glycopeptides and bacitracin. In recent years, advances have also been demonstrated in several of these groups, resulting in the production of new antibiotics and these are considered in this section. 11.6.1 Aminoglycoside-aminocyclitol antibiotics The general structure of the 2-deoxystreptamine-containing antibiotics (the aminoglycoside antibiotics) is shown in (11.89), together with examples of the different drugs. Streptomycin (11.90) does not contain 2-deoxystreptamine. The more important aminoglycoside antibiotics are kanamycin (11.27), gentamicin (11.91), tobramycin (11.92), amikacin (11.28), sisomicin (sissomicin) (11.93) and netilmicin (11.94). As a group, the aminoglycoside antibiotics are bactericidal to Gram-negative bacteria and to staphylococci. Two of the major problems, however, have been the development of resistance of Gram-negative bacteria, often by virtue of drugmodifying enzymes (acetyltransferases, adenylylating enzymes and phosphotransferases: see Section 11.4.1.2) and the toxicity associated with a most important member, gentamicin, which has necessitated careful monitoring of blood and body fluid levels. Desirable properties of the newer (post-gentamicin) types have included increased antimicrobial activity, including improved activity against resistant strains, enhanced pharmacokinetic properties or a reduction in, or freedom from, toxicity. Aminoglycoside antibiotics have several sites at which chemical substitution can be made. Alteration in the 3’ position of kanamycin B to give 3’-deoxykanamycin B (tobramycin) (11.92) changes the activity spectrum. Amikacin has a 4-amino-2hydroxybutanoyl group on the amino group at position 1 in the 2-deoxystreptamine ring,

Introduction to the principles of drug design and action

572

Design of antimicrobial chemotherapeutic agents

573

and this enhances the resistance of the molecule to modification by some, but not all, types of aminoglycoside-modifying enzymes. Amikacin is thus effective against several resistant strains because fewer sites on the molecule are modified. Netilmicin (N-ethylsisomycin (11.94)) is a semisynthetic derivative of ssisomicin that is less susceptible to some types of bacterial enzymes. 11.6.2 Tetracyclines The tetracyclines are no longer used to the same extent as they were in the past. The most imortant members of this group are oxytetracycline (11.95), tetracycline (11.96), doxycycline (11.97), clomocycline (11.98), chlortetracycline (11.99), demethyltetracycline (11.100), methacycline (11.101), and minocycline (11.102). The microbiological spectrum tends to be similar, with cross-resistance between the individual compounds, except for minocycline. Minocycline is active against some tetracycline-resistant strains of Gram-negative bacteria and against tetracyclineresistant staphylococci. Against the former, it appears to enter the cells more readily and may be excreted less rapidly. An isosterically related, chemically synthesized derivative, thiacycline (11.103), is more active than minocycline against tetracycline-resistant strains. This agent is thought not to inhibit protein synthesis, but exerts a bactericidal effect causing nonspecific damage to the bacterial cytoplasmic membrane. Although some toxicity problems have become apparent in the context of its possible clinical use (due to its non-selectivity), thiacycline could form a useful fore-runner for a new group of highly active tetracycline antibiotics. Structure-activity studies in the tetracyclines have shown that inhibitory activity is increased significantly by chlorination at position 7 (e.g. (11.99)). Conversely, decreased potency occurs with epimerization of the 4-dimethylamino group (as in 4epitetracycline) or with ring opening, e.g. in isotetracycline or apo-oxytetracycline. The carboxamide group at the C-2 position of the tetracycline molecule appears to be essential for transport into E. coli cells. In acid conditions, tetracycline hydrochloride is converted to the inactive 5a,6anhydro and epi-anhydro compounds. Chlortetracycline (with a 7-chloro substituent)

Introduction to the principles of drug design and action

574

is more resistant to acid-catalysed decomposition through 5α,6-anhydro compound formation, as

are (11.100), (11.101) and (11.102) where the tertiary hydroxyl group has been modified or is absent. Recently, a new group of tetracycline analogues, the glycylcyclines, have been discovered. The glycylcyclines are novel tetracyclines substituted at the C9 position of the molecule with a dimethylglycylamido side-chain (11.104), (11.105). The glycylcyclines possess activity against organisms expressing resistance to the older tetracyclines mediated by the determinants that encode tetracycline efflux proteins. The glycylcyclines therefore represent a significant advance within the tetracycline group of antibiotics because they are not recognised by the efflux proteins that recognise older members of this class. These drugs also possess activity against organisms expressing the tetM and tetO determinants which probably mediate resistance to the older tetracyclines by

Design of antimicrobial chemotherapeutic agents

575

modifying the tetracycline binding site on the bacterial 30S ribosomal subunit. Therefore the affinity, or mechanism of binding of the glycylcyclines to ribosomes modified by the TetM or TetO proteins, is presumably sufficient to prevent aminoacyltRNA binding despite expression of the resistance determinant in the cell.

11.6.3 Macrolides The most important of the antibacterial macrolide antibiotics is erythromycin (11.106). This is active generally against Gram-positive bacteria and some Gram-negative ones, including Legionella pneumophila. A major effort has been made to synthesize macrolides derived from erythromycin or other naturally occurring compounds. The new macrolides are semisynthetic molecules that differ from the original compounds in the substitution pattern of the lactone ring system. The macrolides consist of a large lactone ring containing 12–16 atoms to which are attached one or more sugars. The most recent clinically useful compounds include roxithromycin (11.107), the methoxyethoxy-methyloxime derivative of erythromycin), clarithromycin (11.108), (the methyl derivative of erythromycin), azithromycin (11.109), deoxo-aza-methyl-homoerythromycin, (the only 15-membered macrolide) and dirithromycin (11.110), a new derivative of

Introduction to the principles of drug design and action

576

Design of antimicrobial chemotherapeutic agents

577

erythromycin. Azithromycin shows good activity against Gram-negative bacteria. The macrolides tend to be unstable and inactived by acidic media and are thus administered in enteric-coated tablets or as the more acid-stable esters and ester salts, e.g. ethyl estolate and succinate. Efforts have also been made to overcome these problems, resulting in new molecules such as miocamycin (11.111), the diacetyl derivative of midecamycin (11.112) and rokitamycin (11.113), the butyryl ester of leucomycin A5 (11.114). 11.6.4 Chloramphenicol Chloramphenicol (11.29) possesses a broad spectrum of activity against Gram-positive and Gram-negative bacteria, and acts by inhibiting the peptidyl transferase reaction in protein synthesis. The active form is the D-threo isomer; the L-erythro, D-erythro and L-threo isomers do not inhibit protein synthesis and are all inactive as antibacterial agents. Some bacteria possess an enzyme, chloramphenicol acetyltransferase, that can inactivate the antibiotic (Section 11.4.1.3). This can pose a clinical problem, and the practical use of

chloramphenicol for systemic infections is reduced because of its tendency to cause blood dyscrasias, notably aplastic anaemia. Nevertheless, it remains an important drug, e.g. in the treatment of H. influenzae meningitis. Chloramphenicol is administered orally as the tasteless palmitate, which is hydrolysed to chloramphenicol in the gastrointestinal tract. The highly water-soluble chloramphenicol sodium succinate comprises the injectable form; this acts as a prodrug, and chloramphenicol is rapidly liberated, although it has been stated that only ca. one-half of the antibiotic in blood is in an active form and that plasma

Introduction to the principles of drug design and action

578

concentrations are lower than those achieved with a comparable oral dose of chloramphenicol. 11.6.5 Folate inhibitors It was found in 1948 that alkyl- or phenyl- substituted diaminopyrimidines had an antifolate action. ‘Small molecule’ inhibitors (collectively, diaminopyrimidine derivatives) were thus studied extensively, leading to the development of compounds that were highly active against bacteria or protozoa and also to agents that were selectively toxic for the parasite rather than the host (mammalian) cells. In 1965 it was demonstrated that trimethoprim (11.22) inhibits dihydrofolate reductase (DHFR) and that its specificity of action is at the molecular level (Table 11.6). Unsubstituted diaminobenzylpyrimidines ((11.115) R1=R2=R3=H) bind poorly to E. coli DHFR; the introduction of a single methoxy group (R1 (11.115)) improves binding to some extent, whereas two methoxy groups (R1 and R2 (11.115)) improve it still further, and three methoxy groups (R1, R2 and R3, as in trimethoprim) produce a highly selective potent antibacterial agent, which binds much less strongly to human DHFR than to the bacterial enzyme. Diaminobenzylpyrimidines with good antibacterial activity can be obtained if a methoxy group is retained at R1 and R3 and a methoxyethoxy or methoxymethoxy group introduced at R2. One of the most active compounds is 2,4diamino-5-(3’,5’-dimethoxy-4’-methoxyethoxybenzyl) pyrimidine, known as tetroxoprim (11.24).

Table 11.6 Inhibitors of dihydrofolate reductase (DHFR) in clinical use. Concentrations1 binding to DHFR in: Type of compound Compound E. coli Rat Specific use Benzylpyrimidine Trimethoprim 0.50 40000 Antibacterial Benzylpyrimidine Tetroxoprim 8.00 40000 Antibacterial Benzylpyrimidine Ormetroprim 5.00 34000 Antiprotozoal Benzylpyrimidine Diaveridine 10.00 7000 Antiprotozoal Diaminopteridine Triamterene 170.00 150 Diuretic Phenylpyrimidine Pyrimethamine 250.00 70 Antimalarial Conjugated Methotrexate Anticancer pteridine 1 As 50% inhibitory concentration (I50×10−8 M), the concentration that affects 50% binding to DHFR.

Design of antimicrobial chemotherapeutic agents

579

The rationale for combining trimethoprim with a sulphonamide (as in co-trimoxazole) is based upon the in vitro assertion that the mixture has a markedly increased activity in comparison to either drug alone, i.e. a ‘sequential blockade’ in which a sulphonamide acts as a competitive inhibitor of dihydropteroate synthetase and trimethoprim inhibits DHFR. It has, moreover, often been stated that the use of such a combination reduces the risk of the emergence of resistance. Arguments about the clinical use of co-trimoxazole, however, continue and it has also been shown that the two components bind simultaneously to DHFR rather than causing the sequential blockade in the metabolic pathway referred to above. Sulphamethoxazole has been found to bind to purified E. coli DHFR, thereby preventing the conversion of dihydrofolate to tetrahydrofolate. Furthermore, synergism between trimethoprim and sulphamethoxazole is not demonstrated against sulphonamide-resistant bacteria and it is uncommon for an organism with acquired trimethoprim resistance to show sulphonamide sensitivity. Resistance to trimethoprim, by plasmid mediation, results from an insusceptible target site, viz. an altered DHFR. About 20000 times as much trimethoprim is needed to inhibit the plasmid-encoded enzyme. Future design of diaminopyrimidines as drugs may well have to concentrate on this aspect. 11.6.6 Quinolones The quinolone antimicrobial agents have been one of the fastest growing group of drugs in recent times. To date, more than 10000 different analogues have been synthesized. They are unusual in being totally synthesized chemically. This means that various side-chains can be altered and the resulting analogues tested for their antibacterial properties. Nalidixic acid (11.13) was derived as a by-product of chloroquine synthesis and is regarded as the progenitor of the newer quinolones. The wealth of compounds and information on structure-activity relationships precludes extensive coverage in the space allowed here but some examples will be highlighted to emphasise the importance of design on microbiological and pharmacokinetic properties. The most important develpment was the introduction of a fluorine substituent at C6 (see (11.7)) which led to a great increase in potency and spectrum of antibacterial activity compared to nalidixic acid (11.13). Ciprofloxacin (11.12) and norfloxacin (11.116) are examples of compounds with a fluorine at C-6, often termed collectively fluoroquinolones. These agents show excellent activity against Gram-negative bacteria, good activity against some Gram-positive bacteria but no action against anaerobes. Subsequent design of new quinolones has largely retained the C-6 fluorine moiety, with changes at other sites focusing on increasing activity against Grampositive cocci and anaerobes. An important early discovery was that the side-chain at position N-1 of the basic quinolone ring had a substantial effect on potency. Ciprofloxacin (11.12) possess a cyclopropyl moiety at N-1. Norfloxacin (11.116) is identical to ciprofloxacin in all respects except for an ethyl rather than cyclopropyl side-chain at N-1, yet ciprofloxacin is more active against Gram-negative and positive bacteria. Ciprofloxacin and norfloxacin belong to the second generation of quinolones.

Introduction to the principles of drug design and action

580

The third and most recently developed generation of quinolones has maintained many of the properties of the second generation. Some compounds (e.g. lomefloxacin (11.117)) differ in having sufficiently long half-lives to allow once daily dosing. Lomefloxacin has a second fluorine atom at C-8, an ethyl group at N-1 (cf. norfloxacin (11.116)) and a methyl group on the piperazinyl group at C-7. There are, however, adverse photosensitivity reactions now being recognised with this compound. Two other third generation quinolones are worthy of note. Sparfloxacin (11.118), a difluorinated quinolone, shows enhanced activity against Gram-positive cocci and anaerobes, whilst retaining high activity against Gram-negative bacteria. Sparfloxacin possesses a cyclopropyl side-chain at N-1 (cf. ciprofloxacin (11.12), a second fluorine at C-8 (cf. lomefloxacin (11.117)), an amino group at C-5 and a substituted piperazinyl ring. Temafloxacin (11.119) differs from ciprofloxacin in possessing a difluorophenyl side-chain at N-1 (a trifluorinated quinolone) and a methyl substituent on the piperazinyl side-chain at C-7. Unfortunately, severe haemolytic and nephrotoxic reactions occurred unexpectedly after the marketing of this drug, leading to its subsequent withdrawal. A novel avenue of antibiotic research involves the synthesis of dual action quinolone-cephalosporin hybrids. Simultaneous administration of drugs is used occasionally to broaden the spectrum of activity or to counteract resistance. If two such drugs could be combined in one molecule (prodrug), with release of the active constituents after

metabolism, problems associated with absorption could be reduced. Though experimental quinolone-cephalosporin hybrids have been synthesized, the data available at the time of writing is inconclusive and the clinical applicability of such hybrids is not yet clear.

Design of antimicrobial chemotherapeutic agents

581

11.7 DESIGN OF ANTIFUNGAL AGENTS The design of antifungal agent poses problems different from those associated with the design of antibacterial drugs. Whereas lack of human toxicity with both types of drug is a factor of paramount importance, the differences in structure of, and some biosynthetic processes in, the fungal cell mean that antibacterial antibiotics are usually without action against fungi. Fungal infections are normally less virulent in nature than are bacterial or viral infections. Furthermore, the fungal cells are eucaryotic, as are human cells, and consequently difficulties arise in designing appropriate chemotherapeutic drugs. One possible target is the fungal cell wall, and considerable advances have been made in understanding its structure and biosynthesis. Another is the cytoplasmic membrane. The ideal antifungal agent should be fungicidal, have a broad spectrum, be in a suitable form for oral and intravenous administration and should have adequate penetration of body fluids; in addition, no resistance should develop during therapy. No drug currently in clinical use satisfies all of these criteria. The most important chemotherapeutic antifungal agents are the macrolide polyenic antibiotics, imidazole derivatives, flucytosine and griseofulvin, a very small number of useful agents in comparison to the very large number of antibacterial antibiotics. 11.7.1 Polyene antibiotics The great majority of polyene antibiotics are produced by Streptomyces species, with nystatin (11.120) the first to be isolated. The macrolide ring of the polyenes is larger than that of the other macrolide group (exemplified by erythromycin (11.106)) and contains a series of conjugated double bonds. An ultraviolet absorption spectrum enables a polyenic antibiotic to be classified, on the basis of the number of olefinic (alkenylic) bonds present, into trienes, tetraenes, pentaenes, hexaenes and heptaenes. Generally, antifungal activity increases with the number of conjugated olefinic bonds, although solubility decreases from the tetraenes to the heptaenes. The polyenes act by combining with the cytoplasmic membrane; this is achieved by an interaction between the antibiotic and membrane sterol, disrupting membrane integrity. Consequently, only those organisms containing sterol in the membrane are sensitive. The

Introduction to the principles of drug design and action

582

antifungal activity of the polyenes can be reduced by the addition of sterols or in the presence of sterol-complexing agents. Very few polyenes are used clinically, and two of the most important are nystatin and amphotericin B ((11.121) R=H). They show activity against yeasts and fungi but not bacteria. The therapeutic usefulness of the polyenes as a group is limited by their solubility and stability and especially by their toxicity. Polyene methyl esters have been synthesized and this is a definite advance. For example, amphotericin B methyl ester ((11.122) R=CH3) is water-soluble and can be administered intravenously as a solution, whereas amphotericin B is insoluble in water (possibly due to zwitterion formation) and is formulated as a colloidal form with sodium deoxycholate. The two forms have equal antifungal activity, but much higher peak serum levels are obtained with the ester which is also considerably less toxic. However, neurological effects have been noted with the methyl ester. More recently, an ascorbate salt has been described, which is water-soluble, of similar activity and less toxic. A new approach has been to deliver amphotericn B in a liposomal formulation, which is thought to reduce toxicity. Future design of polyenic antibiotics might be related to an improved understanding of the nature of their interaction with sterols. In this context, it is pertinent to record the desired property of increased interaction between polyene and fungal membrane ergosterol and decreased interaction between polyene and mammalian membrane cholesterol. Filipen has a high affinity for fungal membrane sterol and for cholesterol and thus has cytotoxic and haemolytic properties, rendering it too toxic for clinical use.

Design of antimicrobial chemotherapeutic agents

583

11.7.2 Imidazole derivatives The antifungal imidazoles comprise a large and diverse group of compounds. Some have antibacterial properties (e.g. metronidazole (11.123) which is of importance in treating anaerobic bacterial infections), others are antihelminthic agents (such as mebendazole) and some, notably clotrimazole (11.124), miconazole (11.125), ketoconazole (11.126) and econazole (11.127) are potent antifungal agents. Two newer agents are fluconazole (11.128) and itraconazole (11.129) with less potential for toxicity than the other agents. None of these agents is fungicidal in action, which may be regarded as a drawback,

Introduction to the principles of drug design and action

584

Design of antimicrobial chemotherapeutic agents

585

though no clinical disadvantage has been observed. The imidazoles have resulted from the synthesis and testing of many hundreds of derivatives rather than from any planned programme of designing new agents based on a knowledge of the structure and biosynthesis processes of the fungal cell. The synthetic imidazoles are generally hydrophobic, and their use is often limited by toxicity. They interact with unsaturated fatty acids in the fungal cell membrane, although the exact mechanism of action is still unclear, because mammalian cells also contain unsaturated fatty acids. Selective toxicity probably results from their selective inhibition of the demethylation of the 14α-lanosterol in ergosterol biosynthesis. 11.7.3 Griseofulvin Griseofulvin (11.130) was isolated from the mould Penicillium griseofulvum in 1939, but because of its lack of antibacterial activity was not then investigated further. Several years later, it was found to be a potent antifungal antibiotic, albeit with a fungistatic rather than a fungicidal action, with significant activity against dermatophytes (Trichophyton, Epidermophyton and Microsporum spp.) but not against Cryptococcus, Aspergillus or Candida spp. or against bacteria. Successful antifungal therapy of certain conditions requires adequate penetration of nail keratin. Orally administered griseofulvin is deposited in the deeper layers of the skin, in hair and in keratin of the nails, and is used in the treatment of fungal infections of the skin, hair and nails caused by susceptible organisms. Griseofulvin is not totally absorbed when given orally, and one method of increasing absorption is to reduce the particle size of the drug.

Introduction to the principles of drug design and action

586

11.7.4 Flucytosine Flucytosine (5-fluorocytosine (11.131)) has a relatively narrow spectrum of activity, with yeasts (including Candida and Cryptococcus) being most sensitive. It acts as a competitive antimetabolite for uracil in the synthesis of yeast RNA and it also interferes with thymidylate synthetase, thus affecting DNA synthesis. Once inside the fungal cell, flucytosine is deaminated to 5-fluorouracil (11.132) which cannot itself be used because of (a) its poor penetration into fungi, and (b) its toxicity to human cells. This intracellular deamination is important because the 5-fluorouracil replaces uracil in the fungal RNA, thereby inhibiting RNA synthesis. Furthermore, Candida albicans converts flucytosine to 5-fluorodeoxyuridine monophosphate (FUdRMP)) which inhibits thymidylate synthetase and thus DNA synthesis.

It seems logical, at least on paper, to propose that such information, coupled with the known mechanisms whereby resistance (which may be a problem) to flucytosine arises, might lead to the devlopment and design of more potent antifungal inhibitors. Sub-inhibitory concentrations of amphotericin enhance the fungicidal properties of flucytosine against C. albicans, with a markedly increased incorporation of 5fluorouracil into RNA. It is interesting to speculate that polyenic-induced membrane damage to the fungal cell membrane is responsible for an increased intracellular uptake of flucytosine with a consequent increased deamination to 5-fluorouracil. The effect is most pronounced in flucytosine-resistant C. albicans. Certainly the mixture has proved to be of value in experimental chemotherapy in animals. One problem always associated with flucytosine, however, is the risk of bone-marrow depression. 11.7.5 Other membrane-active compounds The synthetic thiocarbamates, of which tolnaftate (11.133) is an example, have been found to inhibit squalene epoxidase. The enzyme converts squalene, the first lipophilic intermediate in the sterol pathway, to the 2,3-oxide. Thiocarbamates have been used to treat dermatophyte infections for many years. Interestingly, tolnaftate (11.133) inhibits epoxidase from C. albicans, but is inactive against whole cells, presumably due to inability to penetrate the cell wall. The synthetic allylamines also inhibit squalene epoxidase. They are highly selective for the fungal sterol biosynthetic pathway and often have fungicidal effects against a broad spectrum of fungi. Terbinafine (11.134) is the best-selling non-azole drug and is normally used to treat skin, nail and hair infections. However, there are signs that

Design of antimicrobial chemotherapeutic agents

587

terbinafine (11.134) may have a broader spectrum of activity with potential for treating systemic infections too. The morpholines are characterized by a 2,6-dimethylmorpholine ring with a bulky N-substituent. Amorolfine (11.135) has been developed as a topical antifungal. Amorolfine

inhibits Δ8–7 isomerase and Δ14 reductase, causing hyperfluidity of the membrane with an irregular deposition of chitin. Treated cells accumulate sterols. Amorolfine (11.135) has a broad spectrum of activity, but is not absorbed orally, and is undergoing clinical trials as a topical agent. 11.7.6 Cell wall-active compounds The polyoxins and nikkomycins (neopolyoxins) are peptido-nucleosides, structural analogues of a cell wall precursor, uridine diphosphate N-acetylglucosamine. They are highly specific competitive inhibitors of chitin synthase, causing marked morphological changes leading to cell lysis. The polyoxins (e.g. polyoxin D (11.136)) were isolated from Streptomyces cacoi and the closely related nikkomycins (e.g. nikkomycin Z (11.137)) from S. tendae. Nikkomycins tend to be more active against whole cells, presumably due to better transport into the cell. Despite the promise of these compounds, little in vivo testing seems to have been performed, though some reports have shown significant activity against Histoplasma capsulatum and some yeasts and dermatophytes. The recent discovery that many fungi produce multiple chitin synthases of varying susceptibilies to these compounds highlights the problems of designing adequate agents, particularly for medically important fungi. Many fungal walls contain glucose polymers known as glucans joined through aand β- (1,3) or (1,6) linkages. The echinocandins (e.g. echinocandin B (11.138)) are cyclic hexapeptides in which all of the amino acid residues contain hydroxyl groups.

Introduction to the principles of drug design and action

588

All possess a lipophilic side-chain (derived from linoleic acid). Cilofungin (11.139), a semi-synthetic derivative of echinocandin B (11.138), was developed for clinical use but was withdrawn due to toxic effects. These compounds have two major drawbacks as potential clinical agents: (a) a narrow spectrum of activity, and (b) poor solubility in pharmacologically acceptable solvents. It is also apparent that these compounds inhibit the assembly of β-1,3 and β-1,6 linked glucans and not α-linked glucans. Thus, they are highly active

Design of antimicrobial chemotherapeutic agents

589

Introduction to the principles of drug design and action

590

against Candida, which contains large quantities of β-linked glucans, but are inactive against Cryptococcus, which contains mainly α-1,3-glucan. 11.7.7 Novel antifungal agents The increasing population of immunocompromised patients (those with AIDS, those undergoing organ transplants, or those receiving anticancer chemotherapy) is at risk from disseminated fungal infections. Amphotericin B (11.121) is the only fungicidal agent widely used for treatment in this group of patients, but unfortunately it is toxic for many. Novel fungicidal agents with little or no toxicity are therefore required. Attempts are being made to rationally design drugs targeted at specific fungal enzymes. Such enzymes include the DNA topoisomerases, which are essential in maintaining the topological structure of DNA and are involved in such processes as DNA supercoiling and DNA replication. Topoisomerases are found in both humans and fungi and the crucial question is whether the biochemical differences are sufficient to allow selective targeting with minimal toxicity. Two experimental agents are under investigation. The aminocatechol A-3253 (11.140) is approximately ten-fold more active against Candida topoisomerase I than human topisomerase I. The compound also inhibits β-1,3-glucan synthesis (see Section 11.7.6). A second experimental agent, the isothiazoloquinolone A-75272 (11.141), targets topoisomerase II and possesses equivalent activity against the fungal and mammalian enzymes. Nevertheless, such compounds may serve as useful starting points in the rational design of antifungal agents based on the different biochemical properties of fungal enzymes.

11.7.8 Conclusions and comments Studies on antifungal chemotherapy continue to lag behind those on antibacterial agents. As further information is gleaned as to the structural complexity of and biosynthetic processes in fungal cells, it is hoped that a more logical design of powerful new antifungal compounds can be achieved. Certainly, new azole derivatives can be expected to become increasingly important, but approaches such as that described in Section 11.7.7 may produce clinically useful alternatives. 11.8 DESIGN OF ANTIVIRAL AGENTS The design of antiviral agents presents yet another problem. Since viruses literally ‘take over’ the machinery of the infected human cell, an antiviral agent must show a

Design of antimicrobial chemotherapeutic agents

591

remarkable degree of selective toxicity to inhibit the viral cell without having concomitant action on the human cell. In contrast, the metabolism of pathogenic bacteria is sufficiently different from that of human host cells to render these microorganisms sensitive to inhibitors (e.g. penicillins) which have little or no effect on the metabolism of the host. 11.8.1 Mechanisms of inhibition The genetic information for viral reproduction resides in its nucleic acid (RNA or DNA). The viral particle (virion) does not contain the enzymes required for its own reproduction and after entry into the host cell the virion either uses the enzymes already present or induces the formation of new ones. Unlike bacteria, viruses multiply by synthesis of their separate components, followed by assembly. 11.8.1.1 Amantadines The amantadines exert a concentration-dependent effect: a selective strain-specific inhibition of replication of influenza A viruses at relatively low concentrations (<10 µM) and a non-specific inhibition of virus infection at concentrations >100 µM. The selective action at low concentrations does not involve a direct virucidal effect or prevent adsorption of virus to cells, but infection is inhibited at an early, pre-synthesis stage. Amantidine hydrochloride (11.142) has a very narrow spectrum and its use is usually restricted to prevention of influenza A. 11.8.1.2 Nucleoside analogues Methisazone, idoxuridine and cytaribine inhibit DNA, but not RNA, viruses. Idoxuridine (IUD: 5-iodo-2’-deoxyuridine (11.143)) is a thymidine analogue which is incorporated into the viral DNA in place of the natural substrate. Cytaribine (cytosine arabinoside; Ara-C (11.144)) is active against variola; it does not prevent absorption of the virus, or its penetration or synthesis of viral DNA, and appears to inhibit synthesis of viral proteins. Because of its toxicity, idoxuridine is unsuitable for systemic use, and it is restricted to topical treatment of herpes-infected eyes. Cytaribine is significantly more toxic than idoxuridine. Other nucleoside analogues are at least as active as idoxuridine, e.g. adenosine arabinoside (Ara-A; vidaribine, (11.145)). Ribavirin (1-β-D-ribofuranosyl-1,2,4triazole-3-carboxamide (11.146) is a synthetic nucleoside with a broad spectrum of activity, inhibiting both RNA and DNA viruses. Acycloguanosine (acyclovir) is a nucleoside analogue (11.147) which becomes activated only in infected host cells. It is active against herpes viruses and less so against

Introduction to the principles of drug design and action

592

varicella zoster (shingles). Initially, acyclovir was marketed as an ophthalmic product, but later an intravenous formulation was introduced for treatment in life-threatening herpes infections in immunocompromised patients. In brief, acyclovir is active only against replicating viruses. It is activated in infected cells by a herpes-specific enzyme, thymidine kinase. This enzyme initiates conversion of acyclovir initially to a monophosphate and host kinases subsequently convert this to the antiviral triphosphate, which inhibits the action of viral (but not, to the same extent, host cell) DNA polymerase. The triphosphate is incorporated into the nascent DNA strand at its growing point causing chain termination. This is highly significant when considering the potential for toxicity. The failure of the molecule to be incorporated into mature host DNA reduces the risk of mutation in subsequent rounds of DNA replication. Human cytomegalovirus (CMV) is only weakly inhibited by acyclovir but is more sensitive to the related acyclic nucleoside 7-(2-hydroxyl-1-(hydroxymethyl)-

Design of antimicrobial chemotherapeutic agents

593

ethoxy)methyl- guanine (ganciclovir (11.148)). CMV does not encode thymidine kinase

but possesses a protein with an amino acid sequence similar to known protein kinases which suggests a role in protein phosphorylation for the virus gene product. An important aspect of acyclovir action is its obligatory chain termination. This aspect has been exploited in other drugs. The first drug shown to be effective for the treatment of AIDS (caused by HIV, a retrovirus) was the nucleoside analogue 3'-azido-2',3'-dideoxythymidine (zidovudine, AZT (11.149)). The 3' hydroxyl group which provides the attachment point for the next nucleotide in the growing DNA chain during reverse transcription is replaced by an azido group. Thus this molecule is also an obligate chain terminator. AZT is incorporated in place of thymidine and is an extremely potent inhibitor of HIV replication. However, AZT is relatively toxic since it is converted to the triphosphate by cellular enzymes and therefore is also activated in uninfected cells. Two further analogues, 2',3'-dideoxycytidine (ddC (11.150)) and 2',3'dideoxyinosine (ddI (11.151)), which are both obligate chain terminators, have been developed for HIV. Unfortunately, these also lack selectivity and cause side-effects in man. Other potent dideoxy analogues are in development such as 2'-deoxy-3'thiathymidine (3TC (11.152)) and 2',3'-dideoxy- 2',3'-didehydrothymidine (D4T (11.153)) and it is hoped that they may provide more selectivity. However, resistance to these drugs can build up due to mutations. Sorivudine (E-5-(bromovinyl) arabinofuranosyluracil or BVaraU (11.154)) is the most potent inhibitor of varicella zoster virus described to date. The compound is activated by the virus thymidine kinase but the precise mechanism of inhibition of DNA synthesis is unknown. Unfortunately, a metabolite formed from its degradation, 5-bromovinyl uracil, has led to serious clinical problems when administered to patients being treated with the anti-tumour drug 5-fluorouracil (5-FU (11.155)). Fialuridine (2'-fluoro-5-iodo-β-D-arabinofuranosyl uracil, FIAU) is active against hepatitis B, but the liver is also the major target for toxicity, and several patients enrolled in a trial of this compound died from liver failure.

Introduction to the principles of drug design and action

594

Famciclovir (11.156), the pro-drug of penciclovir (7-(4-hydroxy-3hydroxymethylbut-1-yl) guanine (11.157)) is active against varicella zoster and herpes simplex. Nucleoside analogues are often poorly bioavailable via the oral route, and with penciclovir the bioavailability was so low as to preclude its oral use. Famciclovir overcomes this problem and is metabolised in the body to produce penciclovir. The action of penciclovir is similar to that of acyclovir in that it is converted to the active triphosphate form by a herpes-specific thymidine kinase, inhibiting viral DNA synthesis. It is not, however, an obligate chain terminator and could therefore possibly be incorporated into cellular DNA. Penciclovir triphosphate has relatively weak activity against varicella zoster and herpes simplex, but has a longer half-life and achieves higher levels of the triphosphate form than acyclovir. Valaciclovir (11.158), the valine ester pro-drug of acyclovir, has been developed recently to increase the oral bioavailability of acyclovir. Valaciclovir breaks down rapidly to produce acyclovir and valine. Plasma levels of the drug are three to five times higher using Valaciclovir than those using oral acyclovir. Valaciclovir has been shown to be more effective than acyclovir in resolving zoster-associated pain, and it is equally effective against genital herpes but requires less frequent dosing. 11.8.1.3 Foscarnet Foscarnet (sodium phosphonoformate) is a pyrophosphate analogue (11.159) which has potent activity against herpes simplex cold sores and is non-toxic when applied to the skin. It inhibits herpes DNA polymerase.

Design of antimicrobial chemotherapeutic agents

595

11.8.1.4 Interferons The phenomenon of ‘viral interference’ means that one virus greatly modifies the response of the host to infection with a second, immunologically distinct virus. The term

interferon is applied to a class of basic polypeptides of molecular weights ca. 20000– 30000 induced by viruses but the term is also applied to materials (molecular weights ca. 87000) not induced by viruses. Interferons are thus produced by the host cell in response to the virus particle, the viral nucleic acid and non-viral agents (e.g. natural and synthetic polypeptides). The interferon system involves the induction of an antiviral protein (interferon) by a well-defined inducer and its subsequent interaction with the virus, leading to the development of an antiviral state. For many years, yields of interferons from eukaryotic cells were disappointingly low. However, the application of recombinant DNA technology and cell culture technology has allowed the production of large quantities of interferon. This work is potentially one of the most exciting applications of molecular biology to the design of potent antiviral (and anticancer) agents. Extensive clinical trials have shown interferon to have limited usefulness to date. Although shown to be antiviral in animal models, results in humans have been disappointing. It is currently licensed in the USA for treating hairy cell leukaemia, Kaposi sarcoma and refractory condyloma acuminata. A number of studies are currently in progress, many to determine the effectiveness of interferons in combination with other antiviral agents.

Introduction to the principles of drug design and action

596

11.8.2 Mechanisms of resistance Until relatively recently, few antiviral drugs were available and little was known about their mode of action. Now, not only are more drugs in clinical use but also information about their mechanism of antiviral activity has improved substantially, allied to which is the increasing depth of knowledge about resistance mechanisms. Many of the features of bacterial resistance, viz. plasmid-encoded insusceptibility, reduced drug uptake, drug efflux or drug inactivation, do not apply to viruses. Drug resistance of viruses usually results from point mutations which lead to alterations in proteins (enzymes) that normally either activate drugs or are inhibited by them. Activation of Ara-A to the triphosphate, Ara-ATP, occurs in both herpes-infected and uninfected cells, although the herpes DNA polymerase is a target for Ara-A, and this contributes to its selectivity. Resistant mutants specify DNA polymerases which are less susceptible to Ara-ATP than the enzyme specified by the wild-type virus. Acyclovir triphosphate is a more potent inhibitor of herpes simplex DNA polymerase than of the cellular enzyme. Acyclovir resistance arises as a result of mutations in the thymidine kinase or DNA polymerase genes. Two thymidine kinase mutations have been recognised for herpes simplex. One produces strains deficient in thymidine kinase, resulting in reduced or no phosphorylation of acyclovir. However, thymidine kinase-deficient mutants tend to be less pathogenic, though immunocompromised patients may still be at risk. The second mutation results in altered binding properties for acyclovir. DNA polymerase mutations may confer only marginal shifts in drug susceptibility, but the mutant virus can retain its virulence. Ganciclovir is up to 100 times more active than acyclovir against CMV. Resistance to ganciclovir has been associated with point mutations in the catalytic domain of the phosphotransferase gene, which results in no phosphorylation of ganciclovir. Foscarnet-resistant mutants of CMV and herpes simplex have been found clinically. Resistance is due to mutation of the viral DNA polymerase gene. Interestingly, foscarnet-resistant herpes simplex tends to be susceptible to acyclovir, suggesting that the drugs may have different binding sites on the viral DNA polymerase. Amantadine prevents uncoating of influenza A, blocking release of viral RNA into the cytoplasm. Resistance is due to a mutation in the M2 (matrix) protein. Prolonged treatment of HIV with AZT results in resistance to the drug. Various mutations in the viral reverse transcriptase gene are responsible for resistance. The number and combinations of the various mutations relate directly to the level of resistance achieved, and there tends to be no cross-resistance to other dideoxynucleotide analogues. Drug combinations are now in use to attempt to overcome this problem, and results from ongoing trials are awaited with interest. 11.9 OVERALL CONCLUSIONS AND COMMENTS In many instances (e.g. β-lactams), the development of chemotherapeutic agents for use in the treatment of bacterial, fungal or viral infections has generally followed an enlightened empirical pattern, based upon experimental testing and observations of series of compounds, rather than upon theoretical design.

Design of antimicrobial chemotherapeutic agents

597

With antibacterial drugs, for example, screening of moulds or Streptomyces for the production of antibiotics or enzyme inhibitors has often been followed by the isolation, by further chemical means, of material that is suitable for the design of additional drugs. With increasing knowledge about their mechanism of action, and of bacterial resistance to them, at the molecular level, more sophisticated drugs should be designed with, for example, improved intracellular penetration, enhanced enzyme resistance or increased binding at the sensitive site in the cell. Some areas are worthy of future consideration. The first utilizes permease exploitation and in principle has the concept of fooling a transport system into accepting an antibiotic as if it were the natural substrate (see Section 11.4.4). The antibiotic may act as a carrier for a smaller molecular weight ‘warhead’ function which is selectively released only at its intracellular target site. A second relies on suicide inhibitors (see chapter 8) initially being recognised and processed by the target enzyme as if they were a natural substrate. This is followed by molecular rearrangement of the inhibitor to produce a reactive ‘warhead’ species which irreversibly inhibits the enzyme. A third area, which is particularly appropriate in hospital settings where the possibility of bacterial resistance is greater, involves the construction of hybrid molecules which are metabolized to produce two different types of antibacterial agents. Hybrid cephalosporin-quinolone molecules have been synthesized which produce the active cephalosporin and quinolone moieties when administered. It is too early to say whether such constructs will be useful in a clinical setting. Adherence to surfaces, mediated by bacterial surface appendages, is essential for bacterial colonization. Adherent bacteria are less sensitive to antibiotics and to natural host defence systems, although several existing antibiotics will affect adhesion. Competitive inhibition of bacterial adhesion can be achieved in vitro, but problems remain in vivo. Selective interference with adherence could be a useful concept in drug design, although it has yet to be demonstrated that such an approach would be clinically useful. Antifungal compounds have often been developed following chance observations, the screening of antibacterial agents or of substrates that have been discarded because of a lack of effect on bacteria. Modifications of chemical structure and an empirically based testing procedure have usually followed. Information on the mechanism of action of antifungal compounds has increased considerably, as has that on the underlying reasons for tissue toxicity. Taken with an improved knowledge of the fungal structure and molecular biology, these facts lead to the hope that more effective, less toxic antifungal agents can be designed. Antiviral drugs have often proved to be disappointing, the major problem being the achievement of a selectively toxic effect on the infecting virus without a concomitant harmful effect on the host. However, the introduction of acyclovir and newer nucleoside analogues as antiviral agents augurs for a brighter future, though the dramatic changes in biological properties induced by apparently simple changes in these molecules means that, for the moment, each compound must be judged on its own merits.

Introduction to the principles of drug design and action

598

FURTHER READING Bowden, K., Harris, N.V. and Watson, C.A. (1993) Structure-activity relationships of dihydrofolate reductase inhibitors. Journal of Chemotherapy 5, 377–388. Bugg, C.E., Carson, W.M. and Montgomery, J.A. (1993) Drugs by design. Scientific American 269, 92–98. Dreyfuss, M.M. and Chapela, I.H. (1994). Potential of fungi in the discovery of novel, low-molecular weight pharmaceuticals. Biotechnology 26, 49–80. Greer, J., Erickson, J.W., Baldwin, J.J. and Varney, M.D. (1994) Application of the three-dimensional structures of protein target molecules in structure-based drug design. Journal of Medicinal Chemistry 37, 1035–1054. Hamilton-Miller, J.M. (1994) Dual-action antibiotic hybrids. Journal of Antimicrobial Chemotherapy 33, 197–200. Hooper, D.C. and Wolfson, J.S. (1993) Quinolone Antimicrobial Agents, 2nd edn. Washington DC. American Society for Microbiology. Horan, A.C. (1994) Aerobic actinomycetes: a continuing source of novel natural products. Biotechnology 26, 3–30. Hunter, P.A., Darby, G.K. and Russell, N.J. (1995) Fifty years of antimicrobials: past perspectives and future trends. Society for General Microbiology, Symposium 53. Cambridge. Cambridge University Press. Hunter-Cevera, J.C. and Belt, A. (1994) Bacteria as a source of novel therapeutics. Biotechnology 26, 31–47. Kuntz, I.D. (1992) Structure-based strategies for drug design and discovery. Science 257, 1078–1082. Rossmann, M.G. and McKinlay, M.A. (1992) Application of crystallography to the design of antiviral agents. Infectious Agents and Disease 1, 3–10. Russell, A.D. and Chopra, I. (1996) Understanding antibacterial action and resistance, 2nd edn. London: Chapman & Hall. Vaara, M. (1992) Agents that increase the permeability of the outer membrane. Microbiological Reviews 56, 395–411. Whittle, P.J. and Blundell, T.L. (1994) Protein structure-based drug design. Annual Review of Biophysics & Biomolecular Structure 23, 349–375. Yarborough, G.G., Taylor, D.P., Rowlands, R.T., Crawford, M.S. and Lasure, L.L. (1993) Screening microbial metabolites for new drugs—theoretical and practical issues. Journal of Antibiotics 46, 535–544.

12. RECOMBINANT DNA TECHNOLOGY: MONOCLONAL ANTIBODIES FREDERICK J.ROWELL and JAMES R.FURR CONTENTS 12.1 RECOMBINANT DNA TECHNOLOGY 492 12.1.1 Introduction 492 12.1.2 Principles of recombinant DNA technology

492

12.1.2.1 Identification and isolation of the required gene

493

12.1.2.2 Modification of the isolated gene prior to insertion

493

12.1.2.3 Insertion into the cloning vector 494 12.1.3 Production of polypeptides using recombinant DNA technology

496

12.1.3.1 Somatostatin 496 12.1.3.2 Insulin 497 12.1.3.3 Human growth hormone (HGH) 498 12.1.3.4 Lymphokines and monokines 499

Recombinant DNA technology

599

12.1.3.5 Purity of genetically engineered proteins

499

12.1.3.6 Gene therapy 499 12.2 MONOCLONAL ANTIBODIES 500 12.2.1 Introduction 500 12.2.1.1 The immune system 500 12.2.1.2 Antigens 501 12.2.1.3 Antibody structure and classes 501 12.2.2 Monoclonal antibodies 501 12.2.3 Application of monoclonal antibodies 506 12.2.3.1 Analytical 506 12.2.3.2 Diagnostic 506 12.2.3.3 Therapeutic 506 12.2.3.4 Humanised antibodies and alternative production methods

507

FURTHER READING 507

Introduction to the principles of drug design and action

600

12.1 RECOMBINANT DNA TECHNOLOGY 12.1.1 Introduction For hundreds of years mankind has utilised micro-organisms to produce a whole range of natural products which we can use (e.g. ethanol, organic acids, dextrans, antibiotics). Micro-organisms are extremely easy to cultivate and large scale culture results in high yields of the product required which can then be purified and utilized. Natural products can also be extracted from plant tissue. Biologically active compounds from animals can be isolated from the appropriate organ or tissue but as these are extremely potent compounds, they are normally only present in small quantities and large amounts of the appropriate tissue are required to obtain useful quantities of the product. This is a particular problem with compounds of human origin due to lack of cadavers and to the possibility of contamination of the resulting product with human viruses such as hepatitis and the AIDS virus. For proteins extracted from animals and used in humans such as insulin derived from pigs, since the protein is not chemically identical to the equivalent human protein, its use may evoke an immune response leading to sensitisation of the patient. Somatostatin, a hormone that inhibits the secretion of pituitary growth hormone, required half a million sheep brains to be processed to yield about 5 mg of the product. Today, using recombinant DNA technology, the same amount of hormone corresponding to the human protein can be harvested from 9 litres of a culture medium in which has been grown a micro-organism possessing the inserted human somatostatin gene. It is therefore now possible to produce, in large quantities, a whole range of biologically active polypeptides of identical composition to those found naturally in humans or any other living organism using recombinant DNA technology (often termed genetic engineering). 12.1.2 Principles of recombinant DNA technology The insertion of a human gene into, say, a bacterial cell can only be achieved through techniques that enable the gene to replicate within the cell so that all the progeny derived from the original cell possess the inserted gene and that the product defined by the genetic code from the inserted gene is produced via the normal processes of transcription and translation within the cells of the recipient or host cells. The technique is to insert the gene into extranuclear DNA molecules such as plasmids (extrachromosomal loops of DNA) and bacteriophages (viruses that utilise bacteria as their hosts). They are easily isolated from cells and opened up so that the new gene can be covalently attached to the open strand of DNA, the loop reformed, and the plasmid or bacteriophage containing the new gene re-inserted into the bacterium or other host cell. The agents which open and reform the DNA loops are specific enzymes termed endonucleases and ligases respectively. The six main steps in the genetic engineering process (see Figure 12.1) are therefore;

Recombinant DNA technology

601

(1) Isolating the gene for the protein to be synthesised. (2) Opening the cloning vector (the plasmid or bacteriophage). (3) Covalently attaching the DNA corresponding to the new gene into the open loop of the cloning vector. (4) Reforming the bonds within the enlarged DNA to regenerate the loop. (5) Re-inserting the enlarged vector into the host. (6) Culturing the mutant or chimaeric host cell to enable it to replicate and produce the required protein which is isolated from the culture medium. 12.1.2.1 Identification and isolation of the required gene If the amino acid sequence of the protein to be synthesized by recombinant DNA technology is known then its complementary sequence of codons (triplet sequences of nucleotides, the sequence of each codon corresponding to the amino acid sequence in the protein) can be synthesized. If the protein consists of many amino acids, as is the usual case, then this approach is impractical and the approach is limited to synthesis of the sequences of codons unique to the gene for the required protein. In addition the synthesis incorporates a radioactive tag, usually in the form of 32P into the sugarphosphate backbone of the synthetic DNA strips. These radioactive fragments will now bind to the complementary codons on the DNA corresponding to unique codon sequences of the required gene. This provides us with the means of identifying the location of the required gene within a multitude of DNA molecules obtained from synthesis of mRNA mixtures using the enzyme reverse transcriptase and nucleotides. It is assumed that cells or tissues which are producing the required protein will also have a high concentration of the messenger RNA (mRNA) coding for the protein since this contains the transferred genetic message which is read at the ribosome during the synthesis of the protein. mRNA from these target cells or tissues is extracted and the minute amounts of mRNA thus isolated can be transformed into the genetic DNA code for the protein and finally larger quantities of this key intermediate product can be synthesised using a second enzyme called DNA polymerase. In practice the process described is more complicated since firstly, single stranded mRNA must be used and is formed from the double stranded naturally occurring form and secondly, the mRNA coding for the required gene will be embedded within a much larger segment of mRNA coding for a variety of other genes. Hence this large mRNA fragment has to be cut into smaller pieces using specific enzymes and this process may cut the required gene into smaller segments in the process whereas only the complete intact sequence corresponding to the code for the protein is required. 12.1.2.2 Modifications of the isolated gene prior to insertion Having isolated the DNA sequence coding for the required gene it is necessary to ensure that it is in a form which can be successfully incorporated into the DNA of the cloning vector. This requires four key preliminary modifications of the gene.

Introduction to the principles of drug design and action

602

Firstly, sequences of DNA known as introns which are interruptions of the code found in mammalian genes must be removed enzymatically since their presence leads to incorrect translation of the spliced gene by the bacterial or viral cloning vector during protein synthesis. Secondly a signal has to be incorporated at the beginning of the DNA code for protein to signal to the enzyme RNA polymerase to initiate the transformation of mRNA from the synthetic DNA code for the protein. Likewise the correct signals must be attached to the gene’s DNA code to instruct the vector’s ribosome to start and stop the gene’s synthesis during the translation process. A diagrammatic representation of the resulting vector plasmid is illustrated in Figure 12.2. Thirdly an ancillary DNA code for a marker gene (e.g. the gene for resistance to tetracycline) is spliced adjacent to the gene coding for the required protein. This serves as a means of detecting whether the total sequence has been successfully incorporated in to the host’s DNA. Finally the synthesized double stranded DNA carrying all the modifications listed must be treated with enzymes to expose protrusions of bases at the ends of the DNA duplex that can form bonds with complementary protrusions on the ends of the opened circle of the bacterial or viral DNA to which the foreign DNA is to be attached. 12.1.2.3 Insertion into the cloning vector The extrachromosomal circular DNA molecules found in bacteria can be separated from the rest of the chromosomal material in the cell by agarose gel electrophoresis. They can be opened up by breaking bonds between specific pairings of nucleotides in the DNA molecule using enzymes termed restriction enzymes (or endonucleases). Different restriction enzymes break bonds selectively between different pairs of nucleotides as shown in Table 12.1. In order that the correct orientation of the bases on the end of the opened plasmid and the complementary bases on the ends of the gene to be inserted occurs, it is necessary to use the appropriate endonuclease. Production of the mutually complementary sequences at the exposed ends of the strands of vector and foreign DNA produces cohesive or “sticky ended” strands since they will tend to aggregate together due to formation of complementary hydrogen bonds between them. The nicks in the sugar phosphate backbone are now sealed using a DNA ligase so the foreign gene becomes an integral part of the plasmid’s genetic material (Figure 12.1). The next process is the insertion of the chimaeric plasmid into the host bacteria. Bacteria can take up free extracellular DNA by a process termed transformation. The rate of uptake is slow but this can be enhanced by allowing the process to take place at low temperatures (0–5°C) in the presence of calcium ions. It is necessary to identify those cells which have taken up the enlarged plasmid. This is achieved by use of the marker gene such as that for antibiotic resistance. Successful incorporation of this gene together with the gene for the protein should have produced a bacterium which exhibits resistance to the antibiotic tetracycline. Hence bacteria carrying this resistance gene will survive in a culture medium containing quantities of antibiotic which is fatal to non-resistant bacteria.

Recombinant DNA technology

603

Table 12.1 Recognition sites and end products of endonuclease activity. Enzyme Recognition site Cleavage product EcoR I

Hae III

*↓=Cleavage points.

Figure 12.1 Insertion of foreign DNA into a bacterial plasmid.

Introduction to the principles of drug design and action

604

Resistant cells can then be isolated, propagated and tested to determine whether the required protein is being synthesized by the cells. This is achieved by exposing dishes on which cultures are growing to cellulose nitrate filters. If the cultures are producing the required protein this protein will be adsorbed onto the surface of the filter. Subsequent exposure of the filter to protein-specific antibodies carrying a radioactive label produces radioactive patches corresponding to the location of cultures producing the required protein. A major problem that can occur is the proteolytic degradation of the synthesised polypeptide. This can be prevented by fusing the synthetic gene to the gene of a larger protein associated with the plasmid, e.g. beta-galactosidase or beta-lactamase. The fusion protein is resistant to proteolysis and the required polypeptide can then be cleaved off and isolated. A number of proteins undergo post-translational modification such as modifications of signal or leader amino acid sequences and glycosylation of the protein. An example of how this is achieved for insulin is discussed in Section 12.1.3.2. Choice of the host cell is important since different host organisms have differing capacities to perform post translational changes and may differ in their efficiency of recombinant protein production. Currently the most popular hosts are Escherichia coli, Bacillus subtilis, yeast, and cultured cells of higher eukaryotes such as insect and mammalian cells. For proteins such as insulin which require posttranslational modifications and require formation of disulphide bonds to achieve the active product, E. coli is not the vector of choice. In that case the yeast Saccharomyces cerevisiae has been successfully used as the host. Also in contrast to E. coli, coproduction of pyrogens and endotoxins is not a problem with S. cerevisiae. 12.1.3 Production of polypeptides using recombinant DNA technology 12.1.3.1 Somatostatin Somatostatin is a small polypeptide (14 amino acids long) and the gene is relatively easily synthesised. The synthetic gene is illustrated in Figure 12.3. It should be noted that the initiation amino acid, methionine, preceded the NH2 terminal amino acid of somatostatin and that the COOH-terminal amino acid is followed by two stop codons. An Eco RI

Recombinant DNA technology

605

Figure 12.2 Diagrammatic representation of a vector plasmid.

Figure 12.3 The synthetic somatostatin gene (Reproduced by kind permission from Old and Primrose (1985), p. 60). cleavage site is at one end of the gene and a Bam HI cleavage site at the other, thereby providing cohesive termini to facilitate its insertion at these sites in the plasmid vector. The plasmid used is the artifically created plasmid pBR 322 which has been completely sequenced. About 20 endonucleases cleave this plasmid at known sites. Two antibiotic resistance markers are associated with this plasmid: tetracycline resistance (Tc r) and ampicillin resistance (Apr). Small peptides like somatostatin are rapidly degraded by E. coli and it is necessary to fuse the somatostatin gene to the beta-galactosidase gene for protection. This is achieved by inserting the betagalactosidase gene together with the lac control region adjacent to the somatostatin gene. The lac control region comprises a promoter site, an operator site which “switches on” the adjacent structural genes, and the ribosomal binding site. Thus we have all the elements necessary for successful transcription and subsequent translation. The initial and final plasmid is illustrated in Figure 12.4. Note that the Bam HI site is in the Tcr marker and cutting and gene insertion results in the loss of tetracycline resistance.

Introduction to the principles of drug design and action

606

The hybrid protein produced at the ribosome consisting of the beta-galactosidase protein fused to the somatostatin is treated with cyanogen bromide (CNBr). This cleaves at the methionine residue, which lies between the two molecules, yielding somatostatin plus beta-galactosidase fragments. The use of CNBr cleavage is limited to those peptides not possessing methionine as part of their sequence. The somatostatin is detected by immunoassay. Similar techniques have been used for the synthesis of other smaller peptides such as endorphins and enkephalins, which are considered to be opioid peptides. 12.1.3.2 Insulin Insulin is an excellent example of how the problems of post-translational modification have been overcome. The protein secreted contains a signal sequence of amino acids at the N-terminus. During passage through the membrane these are cleaved off so that the pro-insulin formed consists of three chains (A, B, and C). Insulin is formed by the removal of the C chain by proteases. This leaves the A and B chains of insulin in a stable tertiary structure held together by the two disulphide bonds formed when the molecule originally folded as pro-insulin. Bacteria cannot bring about these processes. Thus the approach used is to construct separately the genes for the A and B chain, insert them separately into pBR plasmids and then add the elements of the lac control and beta-galactosidase gene, followed by cloning into the two separate bacterial strains. Thus one culture is producing a hydrid A chain and another a hybrid B chain. Separate cleavage with CNBr frees the two polypeptide chains which, after purification, can be joined by disulphide bonds. Genetically engineered human insulin has now replaced porcine insulin in use.

Recombinant DNA technology

607

Figure 12.4 Diagrammatic representation of the steps involed in generating a recombinant plasmid for the bacterial synthesis of somatostatin (Reproduced by kind permission from Emery (1984), p. 163). 12.1.3.3 Human growth hormone (HGH) Initially, the only source of HGH was human pituitary tissue which was removed at autopsy. HGH from this source was insufficient for clinical treatment. The peptide contains 191 amino acids and although chemical synthesis of the gene is feasible, it is not an easy process. It was therefore necessary to prepare cDNA (copy DNA) by extracting mRNA from pituitary tissue. It was found that there was an Hae III cleavage site in the sequence coding for amino acids 23 and 24. Treatment with Hae III yielded the larger fragment (amino acids 24–191) which could then be combined with the smaller chemically synthesised fragment (amino acids 1–23) which was preceded by the initiating amino acid methionine. This was now inserted into the plasmid next to an appropriate promoter, etc., that is required for successful transcription and translation. The product, however, is not completely identical to the natural hormone as it contains an extra NH2-terminal, methionine, which could induce an immunological reaction. An alternative approach is to insert the entire cDNA

Introduction to the principles of drug design and action

608

sequence into a SV (Simian virus) 40, using monkey kidney cell tissue culture as the host. The HGH excreted by this method is identical to that found in the pituitary gland. 12.1.3.4 Lymphokines and monokines These are families of proteins that have been shown to exhibit antiviral effects, together with the enhancement of elements of the immune system with resultant anticancer effects. These proteins regulate the cellular parts of the immune system. Macrophages produce monokines and the T cells and B cells produce lymphokines. For example alpha, beta and gamma interferons are produced by the leukocytes, fibroblasts and activated T cells, respectively. Natural yields of lymphokines and monokines are low but gentically engineered human versions are now available. 12.1.3.5 Purity of genetically engineered proteins A variety of analytical methods is used to ensure that products resulting from genetic engineering are fit for human use. These include tests for: (1) Identity: polyacrylamide gel electrophoresis, isoelectric focusing, chromatography (particularly reverse phase HPLC), peptide mapping (in which the protein is digested under controlled conditions by protease enzymes and the HPLC profile of the resulting polypeptide fragments serves as a finger print for the parent protein), amino acid analysis, and spectroscopy. (2) Purity: chromatography, spectroscopy, assays for host DNA, assays for pyrogens and other residual cellular proteins derived from the outer membranes of the host organism, particularly for products to be administered chronically or in high doses. (3) Potency of the product: bio-assay against a national or international reference preparation. This ensures that the product has the required biological activity. 12.1.3.6 Gene therapy The above applications aim to counteract the deficiency of a natural protein through its substitution by injection of the protein synthesized outside the body through genetic engineering. Some diseases are due to defects in the patient’s genes, and examples of such diseases are listed in Table 12.2, together with the deficient gene. This deficiency may be manifested in the lack of production of a hormone or factor, synthesis of an inactive enzyme, or synthesis of a malfunctioning receptor. It is now possible in some cases to synthesize the normal gene and to insert it into a vector using the processes described above, to produce vectors which express the functioning human gene.

Recombinant DNA technology

609

Table 12.2 Some diseases possibly amenable to treatment by gene therapy. Disease state Defective gene Cancer-melanoma HLA-B7 Cystic fibrosis cystic fibrosis transmembrane regulator Growth hormone deficiency growth hormone Haemophilia factor VIII and factor XI Hypercholesteremia low density lipoprotein receptor Phenylketonuria phenylalanine hydroxylase Transplantation of this vector into the patient so that the vector produces the required gene product in the patient has been attempted with some success for cystic fibrosis. Alternatively the DNA for the normal gene could be introduced into the patient’s own cells via genetic engineering. 12.2 MONOCLONAL ANTIBODIES 12.2.1 Introduction Antibodies are proteins that are designed to bind specifically to foreign or antigenic molecules or microorganisms which invade higher living organisms. This specific binding initiates a range of in vivo processes designed to neutralise the adverse biological activity of the invading molecules or micro-organisms and expedite their elimination from the body. It is their specific binding with relatively high affinity to antigens which has found exploitation in many areas of biological sciences. It has been the ease of production of antibodies to a wide variety of antigenic species coupled with the ability to produce a single type of antibody of constant specificity and composition through monoclonal antibody technology, that has led to the use of antibodies as potent biological targeting agents for in vivo use and as diagnostic agents. 12.2.1.1 The immune system Higher animals possess a highly sophisticated immune system. Substances that activate the immune system are known as antigens. Two kinds of effector mechanisms mediate the immune response to antigens. One response is mediated by a population of lymphocytes known as T lymphocytes (T cells). These T cells act in conjunction with a second set of lymphocytes termed B lymphocytes (B cells) to ensure that antibodies are only produced to foreign molecules and micro-organisms. It is the function of activated B cells to produce antibodies and of the T cells to police the antibody production process so that antibodies are only produced to invading foreign molecules and micro-organisms.

Introduction to the principles of drug design and action

610

Each B cell has on its surface a unique receptor. A small fraction of the B cells which normally circulate in the blood and lymph will fortuitously bind to patches on the surface of the foreign molecule or micro-organism (called epitopes). The T cell screens the resulting B cell-molecular complex and if the foreign molecule does not possess the marker flags on its surface which identify it as “self” then the T cells signal to the complexed B cells to activate division of these B cells through release of activator molecules (particularly interleukins). The activated B cells now rapidly increase in number and secrete antibodies which possess antibody binding sites, the structure of which is identical to the receptors on the pre-activated cells which themselves bound to the foreign molecules or micro-organisms. Hence these antibodies will bind to the same foreign molecules or epitopes on the surface of larger invading micro-organisms. This is the basis of the humoral immune response. It should be noted that each activated B cell will produce a series of identical daughter cells or clones each of which secretes the same unique antibody. In practice a number of B cells are activated for each antigen and consequently a range of different cloned B cells is generated and hence a variety of antibodies each recognizing different epitopes will be present in the antiserum of the animal exposed to the foreign antigen. The resulting antiserum is termed a polyclonal antiserum. 12.2.1.2 Antigens An antigen is any substance which can elicit an immune response in an animal which possesses a functional immune response. Proteins which are foreign to the animal are generally highly immunogenic and will stimulate the production of a range of different antibodies, each being specific for a particular determinant (or epitope) associated with the surface of the protein. These antibodies will bind to that particular portion of the protein only. Polysaccharides and nucleic acids are less immunogenic than proteins, even though they may have a high molecular weight. Low molecular weight molecules (Mr
Recombinant DNA technology

611

these variable regions lie three hypervariable segments. The VH and LH are folded in such a way as to bring together the hypervariable regions together to form a groove or cavity into which the epitope fits. The antigen-antibody binding is highly specific due to the stereochemical complementarity which is required coupled to complementary hydrophobic and/or ionic interactions between the amino acids in the binding site and the contact groups on the surface of the epitope. The carboxyl terminal shows far less sequence variation between antibodies and is termed the constant region (C) of the antibody molecule. Each species of animal will produce identical constant regions for each subclass of antibody. Cleavage of the antibody molecule can be achieved with proteolytic enzymes such as pepsin and trypsin into fragments that retain antigen binding properties (termed Fab fragments) and fragments that do not (termed Fc fragments). The major humoral antibody is IgG. Its structure is illustrated in Figure 12.5. 12.2.2 Monoclonal antibodies It was noted in Section 12.2.1.1 that an activated B cell will subdivide to produce a clone of identical daughter cells each of which secretes identical antibodies. If a single activated B cell could be isolated and cloned then the resulting antibodies derived from a single clone of daughter cells is termed a monoclonal antiserum. Unfortunately it is not feasible to produce monclonal antobodies via this route as the quantities of antibodies thus obtained are limited since culture and growth of the cell line will rapidly result in cell death due to the mortality of B cells. It is therefore necessary to render activated B

Figure 12.5 Diagrammatic representation of a molecule of human IgG.

Introduction to the principles of drug design and action

612

cells immortal and this is achieved by fusing the genetic material from the required B cells with a cancerous B cell from the same species since cancerous cells are immortal and the resulting hybrid cells should contain the genes for production of the required antibody and for immortality. Large scale tissue culture of the resulting fused cells should enable production and harvesting of monoclonal antibodies on commercial scales. In practice mice or rats are used and the most common type of mouse used for monclonal antibody production is an inbred strain known as BALB/C. The cancerous B cells required for fusion are myeloma cells. Kohler and Milstein in 1975 demonstrated that mouse myeloma cells could be fused with B cells taken from the spleen of immunized mice and the resultant hybrids produced antibody. The technique is called somatic cell hybridization and the cell product termed a hybridoma (Figure 12.6). The five major steps involved in monclonal antibody formation via this process are as follows: (1) Immunization of the selected animals (usually BALB/C mice) with immunogen. (2) Isolation of the spleen from a hyperimmunized animal and fusion of B cells from the spleen with myeloma cells from the same species. (3) Culture of the resulting hydidoma cells in (HAT) medium. (4) Selection of single clones of immortal cells secreting the required antibody. (5) Scale up of the culture process to produce the required monclonal antibody. The initial stage is to repeatedly inject the antigen into a group of about six animals, often in conjunction with immunostimulants such as Freund’s adjuvant (complete for the first immunization followed by incomplete for subsequent ones). Immunizations take place at intervals of about a week and with successive immunizations there is an increased stimulation of B cell clones within the animal responding to the antigen. The presence of

Figure 12.6 Principle of hybridoma formation.

Recombinant DNA technology

613

high concentrations of antibody to the antigen can be demonstrated by taking blood samples from the animal and analysing the sample. In this way the animal giving the best response to the antigen can be identified. A suitable source of these B cells is the spleen from which they are harvested after the animal showing the best response has been sacrificed. The fusion partners (the myeloma cells) are now well established cell lines and many of these lines have mutated to being non-secretors or better still, non-synthesisers of antibody. The latter is the ideal partner for the antigen-stimulated B cell. The cell fusion is normally carried out using large numbers of the two cell types, as the rate of successful fusion is low (about 1 in 105 cells). The inclusion of 50% polyethylene glycol (PEG, Mr 140– 4000) and about 5–10% dimethyl sulphoxide (DMSO) in the solution creates a favourable medium for membrane-membrane fusion between cells to occur. Due to the low fusion rate and of the probability that fusion will occur between identical cells in addition to that between B cells and myeloma cells, the fused cells are transferred to a medium which only allows the required successful fusions to flourish. The standard technique is to use myeloma cells which have lost the capacity to synthesize hypoxanthine-guanine-phosphoribosyl-transferase (HGPRT), one of two enzymes required for the eventual synthesis of DNA in the cell. The other pathway leading to DNA synthesis is the salvage pathway and it can be blocked by addition of the chemical aminopterin to the cell culture medium. If the transferred cells from the fusion step are placed in a medium containing aminopterin (A) then only cells containing HGPRT can survive. The gene for this enzyme will be derived from the activated B cells and use of this pathway for DNA production is also encouraged by the presence of hypaxanthine (H) and thymidine (T) in the medium since these compounds are utilised in the synthesis by the enzyme. The medium is termed HAT medium due to the presence of these three additives in it. Unfused cells and fused cells containing genetic material from only the B cells will not survive in the HAT medium as the unfused myeloma cells die due to inability to synthesize DNA while the other cells eventually die out. Hybridoma cells are able to proliferate as they contain the genes for immortality and for HGPRT. As antigens contain many epitopes, it follows that the resulting hydridoma cells will secrete a number of antibodies corresponding to these epitopes. It is therefore necessary to screen for clones which produce antibodies to the required epitope. This is achieved in two steps; firstly single hydridoma cells are isolated and then these are screened to identify which are producing the required antibodies. Two strategies can be used for cloning cells: (a) Limiting dilution: the hydrid suspension is diluted so that the volume put into each culture well contains on average a single cell. No growth will occur in those wells that receive no cells and those wells that receive more than one cell may result in antibodies of two specificities being found. The clones that develop may be broken up and the dilution process and cloning repeated several times to ensure monoclonality. (b) Solid medium: a solid medium based on agarose gel has been developed which permits the cells to divide on the surface to form visible colonies. These can be broken up and the process repeated to

Introduction to the principles of drug design and action

614

obtain pure clones of hydridoma cells which can then be transferred to a suitable culture medium to test for antibody production. The clones are screened for antibody production by using a suitable assay technique which can initially detect the class of Ig being produced, since IgG or IgM are commonly produced. This test uses a second antibody reagent which specifically binds to IgG or IgM molecules and this binding interaction can be monitored. The next step is to determine whether any of the clones are secreting antibody to any of the epitopes assocaited with the antigen. Once again the use of a second antibody reagent is employed (see Figure 12.7). The test uses the antigen or epitope immobilised on a convenient surface such as the wells of microtitre plates. If the culture supernatant contains antibodies from the antigen then these will bind to the epitope on the surface. Addition of a labelled

Figure 12.7 Use of anti-antibodies to detect synthesis of antibody by hybridoma cells. antibody, which specifically binds to the class of antibody identified in the first screen, to the wells will result in the formation of a second antibody complex linked to the first murine antibody attached to the epitope on the surface. After washing the plate to remove excess second reagent, the presence of the bound second reagent in the complex can be determined from the presence of label in the well(s). Once the clones secreting the required antibody have been identified those cells are transferred to large culture vessels where their propagation can take place and monclonal antibodies subsequently harvested. A variety of culture systems can be employed ranging from conventional animal cell culture in bottles to use of hollow fibre systems onto which cells adhere and through which the culture medium is pumped. The latter systems offer advantages since the secreted antibodies can be

Recombinant DNA technology

615

collected in the medium emerging from the fibres and higher yields of antibody can be obtained (g/l compared with mcg/l for conventional culture systems). Alternatively, the murine hydridoma can be propagated in BALB/C mice where the concentration in ascites fluid can reach 5–20 mg/ml. A diagrammatic representation of monoclonal antibody production is shown in Figure 12.8. Monoclonal antibodies are chemically homogeneous but may not be truly monospecific as they can react with antigens that share an epitope or can react with epitopes that are structurally closely related.

Figure 12.8 Diagrammatic representation of monoclonal antibody production. 12.2.3 Application of monoclonal antibodies There has been a dramatic increase in the use of monoclonal antibodies and the greatest expansion has been in the field of therapy where they have been used to target drugs to specific tissues. However they will continue to find an ever increasing role as analytical and diagnostic reagents due to their reproducible properties of specificity

Introduction to the principles of drug design and action

616

and affinity for specific antigens, and from the ability to produce them in commercial quantities. 12.2.3.1 Analytical Because of their specificity, monoclonal antibodies are used extensively in the assay of serum and urine levels of hormones (e.g. detection of human chorionic gonadotrophin in the urine of pregnant women), drugs, enzymes, etc. They are also finding application in assays for environmental contaminants such as pesticides and dioxins, in soil, water and air samples. A wide range of immunoassay procedures has been developed which employ a variety of end points. Of particular importance are noncompetitive, nonisotopic assays based on fluorescence and spectrophotometric end points, the latter resulting from use of enzyme labels to rapidly convert substrates to coloured products. Such assays are used for therapeutic drug monitoring where fully automated assays provide results within seconds or minutes per sample. 12.2.3.2 Diagnostic Monoclonal antibodies have been used to identify specific antigens associated with cell surfaces. This has led to the development of rapid tissue typing techniques prior to transplantation surgery, the classification of cells and sub-populations (particularly the T cell family), cell-cell interactions and differentiation, the biochemistry of cell surfaces, the classification of micro-organisms and the detection of tumour markers associated with certain types of tumour. 12.2.3.3 Therapeutic One of the most exciting potential uses of monoclonal antibodies is in cancer. The imaging and location of metastases using radiolabelled antibodies has been used with some success, but the use of antibodies as targeting agents for drugs, plant and bacterial toxins is an area into which tremendous effort is being directed. Combinations of antibodies joined together through their Fc regions are also being investigated for therapeutic use. These so called bi-specific antibodies work by targeting the second piggy-backed antibody to the required site within the body, such as a cancer cell, where the targeting antibody binds. The accompanying antibody is specific to a second cell such as a T cell which it captures. The resulting accumulation of T cells on the surface of the cancer cell may result in its destruction through initiation of the immune response. Another theraputic application of monoclonal antibodies is as rescue agents where they are injected into the blood of a patient who has present in their body a potentially lethal amount of some toxic agent. This can be the result of a drug overdose, e.g digoxin, or can result from the bite of a spider or snake, or can result from bacterial toxins such as those associated with blood poisoning. The injected antibody rapidly binds to the toxic agent which is inactive in the resulting complex. In this form the neutralised toxin is eventually excreted from the body.

Recombinant DNA technology

617

Vaccines derived from genetic engineering also promise to provide us with non pathogenic vectors which have been genetically modified to express on their surface one or more antigenic epitopes from viruses, bacteria and organisms which currently produce widespread illness and mortality such as measles, turberculosis, diphtheria, poliomyelitis and so on. Thus a single vaccination to a polyvaccine from such an engineered vector would be highly cost effective and safer to use. Alternatively, instead of using live recombinant vectors it is possible to produce and isolate recombinant surface antigen proteins for use as vaccines. Examples of recombinant vaccines include recombinant BCG, an avirulent bovine tubercle bacillus, and the recombinant surface antigen protein for hepatitis B. 12.2.3.4 Humanised antibodies and alternative production methods It can be appreciated that if murine antibodies are repeatedly used as in vivo therapeutic agents in man then they will evoke an immune response as they are foreign proteins. In order to overcome this problem a variety of alternatives have been explored to make murine antibodies more like human antibodies, to produce human antibodies by genetic engineering, and by use of in vitro methods which do not involve animals at all. It was noted in Section 12.2.1.3 that antibodies can be split by pepsin into two fragments, one of which retains the antibody binding capacity while the other corresponds to the constant region of the molecule. Digestion of the specific murine antibody and isolation of the Fab fragment can therefore be achieved. In a similar manner human IgG can be divided into its Fc and Fab fragments. It is posible to splice together the Fab fragment of the murine antibody with the isolated non-antigenic Fc fragment of the human IgG to produce a hydrid or chimaeric antibody which possesses the required specificity but with reduced antigenicity. Since the amino acid sequences can be determined for both the Fc region of human Ig and the Fab region of the murine antibody, it is possible to synthesise the codon sequence corresponding to the entire chimaeric molecule. When the appropriate promoter and initiator codons have been added to the gene it is possible to insert it into an appropriate host via a plasmid or bacteriophage, with subsequent translation producing a genetically engineered human IgG of the required epitope specificity. A more recent approach involves randomly synthesising codons within the Fab region spliced onto the codon sequence for the human Fc region. The specificty of the resulting antibody is noted and a library of codon sequences built up correlating sequence to specificty. In this way it is hoped that tailor made antibodies of required specificity will be synthesied via genetic engineering from the library information. FURTHER READING Brown, T.A. (1991) Essential Molecular Biology, Volume 1 A Practical Approach. Oxford: IRL Press. Emery, A.E.H. (1984) An Introduction to Recombinant DNA. Chichester: John Wiley and Sons Ltd.

Introduction to the principles of drug design and action

618

Fanger, M.W. and Guyre, P.M. (1991) Bispecific antibodies for targeted cytotoxicity. Tibtechnology 9, 375–380. Holliger, P and Winter, G. (1993) Engineering bispecific antibodies. Current Opinion in Biotechnology 4, 446–449. Hudson, L. and Hay, F.C. (1989) Practical Immunology, 3rd ed. Oxford: Blackwell Scientific Publications. Hurle, M.R. and Gross, M. (1994) Protein engineering techniques for antibody humanisation. Current Opinion in Biotechnology 5, 428–433. Lerner, R.A., Kang, A.S., Bain, J.D., Burton, D.R. and Barbas, C.F. (1992) Antibodies without immunisation. Science 258, 1313–1315. Old, R.W. And Primrose, S.B. (1985) In Principles of Gene Manipulation, edited by N.G.Carr, J.L.Ingraham and S.C.Rittenberg. Oxford: Blackwell Scientific Publications. Paliwal, S.K., Nadler, T.K. and Regnier, F.E. (1993) Rapid process monitoring in biotechnology. Tibtechnology 11, 95–101. Pezzuto, J.M., Johnson, M.E. and Manasse, H.R. (1993) Biotechnology and Pharmacy. New York: Chapman and Hall. Tomlinson, E. (1992) Impact of new biologies on the medical and pharmaceutical sciences. Journal of Pharmacy and Pharmacology 44, Supplement 1, 147–149. Walker, J.M. and Gingold, E.B. (1993) Molecular Biology and Biotechnology, 3rd edn. Cambridge: The Royal Society of Chemistry.

13. BIO-INORGANIC CHEMISTRY AND ITS PHARMACEUTICAL APPLICATIONS DAVID M.TAYLOR and DAVID R.WILLIAMS CONTENTS 13.1 INTRODUCTION

510

13.2 THE IMPORTANCE OF TRACE ELEMENTS IN HUMANS

512

13.3 METAL COORDINATION CHEMISTRY

516

13.3.1 General aspects

516

13.3.2 Metal-ligand specificity

517

13.3.2.1 Complex stability

517

13.3.2.2 The hard and soft acid and base approach (HSAB)

518

13.3.2.3 The HSAB principle in biochemistry and pharmacy

518

13.3.2.4 In vivo complexing and metabolic specificity

520

13.4 PHARMACOLOGICAL AND PHARMACEUTICAL CONSIDERATIONS

521

13.5 METALS AS THE MODUS OPERANDI OF DRUGS

522

13.5.1 Metals as drugs

522

13.5.1.1 Bismuth in the treatment of peptic ulcer

522

13.5.1.2 Metals as anti-cancer agents

523

13.5.1.3 Copper and rheumatoid arthritis

525

Introduction to the principles of drug design and action

620

13.6 DRUGS WHICH EXERT THEIR EFFECTS VIA METAL COMPLEXATION OR CHELATION

528

13.6.1 Metal chelation in antimicrobial activity

528

13.6.2 Metal ion removal

529

13.6.2.1 Removal of copper in Wilson’s disease

529

13.6.2.2 Removal of iron in haematochromatosis

530

13.6.2.3 Treatment of exogenous metal poisoning

531

13.7 METAL-DEPENDENT SIDE EFFECTS OF DRUGS

532

13.7.1 Metal ion sequestration by a metabolite

533

13.7.2 Sequestering drug-non-specific metal ion interactions

534

13.8 TRACE ELEMENT SUPPLEMENTATION

534

13.8.1 Iron, zinc and copper supplementation

535

13.9 CONCLUDING REMARKS

536

FURTHER READING

537

13.1 INTRODUCTION The human body contains about 50 of the elements in the Periodic Table, about half of these are either essential or beneficial to life, while the remainder are adventitious having been introduced from local dietary or environmental influences. Ten of these essential elements, oxygen, carbon, hydrogen, nitrogen, calcium, phosphorus, sulphur, potassium, sodium and chlorine, are present in amounts ranging from tens of kilogrammes to a few tens of grammes, and these so called bulk elements are contained in the proteins, fats and carbohydrates that are the building blocks for the organs and tissues. The major emphasis of this review will rest on the remaining twenty or so of the essential and beneficial elements which, because they are generally present in very small quantities, are often called trace elements. The total mass of these elements in the human body is less than 50 grammes, yet they must be present in the correct concentrations and forms if the individual is to enjoy a healthy life. Of the adventitious elements, some can be described as pollutants arising from human activities, especially from the industrial developments of the past two

Bio-Inorganic chemistry and its pharmaceutical applications

621

centuries, while the remainder enter the human body simply because they are present naturally in drinking water or in the plant and animal tissues which make up our food. The concentrations of the adventitious elements vary from person to person, depending on their geographical environment and their eating habits. There are strict criteria which must be fulfilled before a trace element is classified as essential—it must be present in all healthy tissues and it must cause reproducible symptoms of ill-health if it is excluded from the diet. The essential and beneficial trace elements and their locations in the Periodic Table are shown in Figure 13.1, while Table 13.1 lists the amounts of the thirty essential and beneficial elements found in the human body. The selection of these elements from the 80 stable elements in the Periodic Table has been critically dependent upon the composition of the earth’s surface. Life on this planet probably began about five billion (~5×109) years ago from primitive cells which evolved in the ancient oceans utilizing biochemicals synthesised on the surfaces of sand particles on tidal beaches. Such evolution was based on the elements readily available in the ancient sea bed and the primitive oceans and any life forms dependent on less readily available elements would have been bred out many millions of years ago. Thus the composition of the human body broadly resembles the elemental composition of the ancient sea bed and the primitive oceans. These are the lighter elements of the Periodic Table, since, when the planet was formed from a cloud of dust particles, the middleweight and heavier elements were compacted by gravity into the mantle and core of the earth, respectively. The evolution of life has not been a smooth process. The earliest cells evolved under the highly reducing atmosphere of water vapour, hydrogen sulphide, ammonia and methane that was present at the beginning of terrestrial time. Around two billion years ago the early cell probably contained only about 100 different protein molecules, as

Figure 13.1 The Periodic Table indicating those elements that are known, or firmly believed, to be essential (bold type) or beneficial (italics) elements.

Introduction to the principles of drug design and action

622

Table 13.1 The masses of the essential and beneficial elements occurring in a 70 kg reference person (adapted from Taylor and Williams 1995). Element Mass grammes moles Hydrogen 7000 3500 Carbon 12600 1050 Nitrogen 2100 75 Oxygen 45500 1425 Phosphorus 700 22.5 Sulphur 175 5.5 Chlorine 105 3.0 Sodium 105 4.6 Potassium 140 3.6 Calcium 1050 26 Lithium 0.0007 0.0001 Boron 0.01 0.0009 Fluorine 0.8 0.02 Magnesium 35 1.4 Silicon 1.4 0.05 Vanadium 0.02 0.0004 Chromium 0.005 0.0001 Manganese 0.02 0.00036 Iron 4.2 0.075 Cobalt 0.0007 0.00001 Nickel 0.01 0.0002 Copper 0.11 0.0016 Zinc 2.3 0.035 Arsenic 0.014 0.0002 Selenium 0.02 0.003 Bromine 0.2 0.0025 Molybdenum 0.005 0.00005 Tin 0.03 0.0002 Iodine 0.013 0.001 Barium 0.016 0.00012 compared to many thousands of proteins found in modern cells; it also contained a range of metal ions, some of which fulfilled structural or osmotic roles while others acted as catalysts. Magnesium would have been particularly good in this latter role,

Bio-Inorganic chemistry and its pharmaceutical applications

623

since it is known to catalyse condensation reactions and to have been present in high concentrations in the primeval oceans. The heaviest metals used in evolution are those from the first transition series of the Periodic Table, and such metals as manganese, iron, cobalt and copper all existed in their lower oxidation states in primitive cells. Then, relatively suddenly, perhaps about 1.8 billion years ago, blue-green algae in the oceans began to produce oxygen in sufficient quantities to slowly convert the Earth’s reducing atmosphere into the present oxygen-containing one. This had the effect of raising the oxidation states of the aforementioned transition metals and of releasing the previously insoluble cuprous ores into the biosphere as the more soluble cupric salts. At the same time iron and manganese were immobilised as their higher oxidation states, Fe3O4, FeOOH, Mn3O4 and MnO2. The atmospheric ozone content also began to rise, reaching ~1%, which was sufficient to screen out the harmful effects of the sun’s ultraviolet radiation which was apparently more destructive to aerobic rather than to anaerobic systems. As a result of these dramatic environmental changes, organisms that had evolved using ferrous ions as oxygen carriers became vulnerable to oxidation and probably died out. However, biochemical systems involving cupric complexes and oxygen were now possible. Thus, because of the chequered history of the four metals, iron, manganese, cobalt and copper, in their different oxidation states—some being originally rejected by nature and only relatively recently incorporated into essential biochemical mechanisms—it is to be expected that in complex multicellular species, such as Homo Sapiens, serious disturbances in biochemical reactions may arise from changes in metal balances and from challenges to optimal concentrations of such metals caused by environmental factors or by pharmacological intervention. 13.2 THE IMPORTANCE OF TRACE ELEMENTS IN HUMANS Homo Sapiens is a complex multicellular species that depends for its health upon the correct functioning of ~1015 cells of many different, yet interdependent, types. Thus, the maintenance of good health will also depend on the supply of all 30 essential or beneficial elements in adequate, but not excessive, quantity and in a chemical form that is utilisable within the body. The principle of biphasic response, Figure 13.2, reminds us that while an insufficiency of an element may lead to ill health by preventing the optimal function of some biochemical process, an excess may cause serious, even life-threatening toxic reactions. Trace element insufficiency or imbalances of elemental intake arise either because there has been some imbalance lower down in the food chain from soil to plant to animal to man, or because the element has been ingested in a chemical form that cannot readily be transferred from the lumen of the gastrointestinal tract into the plasma. Such lack of bioavailability may arise because the element is present in a highly insoluble form, or because it has formed complexes with constituents of the gastrointestinal tract contents, or perhaps with a pharmaceutical, which carry a pattern of electrical charge that limits or prevents transfer across the mucosa.

Introduction to the principles of drug design and action

624

Figure 13.2 A generalised dose-response curve for an essential or beneficial element, illustrating the principle of the biphasic response. At very low concentrations there may be insuffient metal for optimal biological effect, while at higher levels the same metal may be toxic. The transition metals, which collectively weigh less than 10 g are particularly vulnerable to interactions that may produce the effects of a deficiency. For example: 1. Each of these trace elements may be present in an organ or body fluid in only microgramme quantities, yet drugs which can interact with them may be administered in milligramme or greater quantities—the same pharmaceuticals may have little influence upon the 1.7 kilogrammes of calcium or the 42 grammes of magnesium but they may easily render a few microgrammes of copper or zinc biochemically inert. 2. Poorly controlled industrial activities have resulted in the release of elements not normally present into our environment or diet, for example lead, cadmium and mercury. Such metals can compete with essential ones for important biochemical binding sites in vivo. 3. Modern food processing, preservation and packaging techniques, as well as the advent of convenience foods, and even junk-food diets, have all influenced the spectrum of trace element and metal complexing

Bio-Inorganic chemistry and its pharmaceutical applications

625

chemical ligand intakes into humans. 4. Advances in medicine and pharmacy have extended human lifespan towards ninety or more years and long-continued dietary or therapeutic practices that cause even very minor disturbance of trace elements may over years lead to serious deficiency or imbalance. The importance of this can be seen in the treatment of conditions such as arthritis, hypertension, gout or minor cardiac disorders which may require medication to be continued for perhaps twenty or more years. 5. As has already been mentioned, our basic biochemistry evolved under anaerobic conditions and even in to-day’s aerobic atmosphere we often still require metals such as manganese, iron, cobalt and copper in their lower oxidation states. Thus, drugs which interfere with oxidationreduction (redox) mechanisms may lead to biochemical inactivation of such metals. Fortunately, trace element balances can usually be easily restored if deficiencies or excesses are recognised early enough. The correction of trace element deficiency is no simple task and a series of criteria need to be satisfied; the elements must be supplied in solution at the correct pH to avoid precipitation, in an appropriate oxidation state and in a form that is bioavailable. It may be the dispensing pharmacist—patient relationship that first discusses such trace element dependent side effects. Thus it is important for pharmacists to have some understanding of trace element biochemistry. In fact, there is no known aspect of biochemistry in which trace metals are not involved. Table 13.2 contains a partial listing of conditions associated with trace element imbalance—a full list would require many pages (see Fraústo da Silva and Williams 1991; Taylor and Williams 1995). For a more detailed discussion of the normal and pathological biochemistry of trace elements the reader is referred to some of the general texts listed in the references (Fraústo da Silva and Williams 1991; Hay 1984; Taylor and Williams 1995; Williams 1976) but a few of the more important roles of such elements are mentioned here. • Some elements are essential as cations or anions for the maintenance of osmotic pressures and to neutralise species of opposite charge. Sodium and potassium ions and even the humble proton fall into this category. Nature has evolved a very specific biochemistry to maintain the concentrations of these elements at just the correct levels. For example, in blood the pH is buffered at ~7.4 despite severe challenges from the metabolic production of bicarbonate ions etc.; changes of even 0.1 of a pH unit are seriously life-threatening. Similarly, the sodium and potassium concentrations in blood are strictly maintained at 140 and 4 mmol.dm−3, respectively; whereas inside cells the concentrations are 20 and 95 mmol.dm−3. Thus specific mechanisms have evolved to keep sodium out of cells while encouraging the influx of potassium. Similarly, chloride is kept outside cells, whereas intracellular fluid is rich HPO42−. • More than 1000 metalloenzymes are known to date; about half contain zinc either as part of the active site or in a structural role, while others

Introduction to the principles of drug design and action

626

require manganese, molybdenum, copper, nickel or selenium for their activity. • Cobalt, once a very important element in primitive cells, now appears to plays an important role only in vitamin B12 and in allowing free radical chemistry at low redox potentials. • Copper, although less than 1 g is found in the average person, has been linked to very many low molecular mass metal complexes and to caeruloplasmin, metallothionein and other proteins in blood plasma and in tissues. Copper is now recognised as playing many roles in vivo, for example, copper oxidases play important roles in scavenging free radicals and in removing excess neurotransmitter amines and in the development of hormonal messengers such as adrenaline.

Table 13.2 Examples of some metal associated disorders. Element Disorder associated Disorder associated with with a deficiency an excess Essential and beneficial metals Calcium Bone deformities Atherosclerosis, Cobalt (Rickets), tetany cataracts, gall stones Anaemia, anorexia Cardiomyopathy, hypothyroidism, polycythaemia, cancer Copper Anaemia, Menke’s Wilson’s disease, “kinky hair” syndrome intestinal and liver inflammation, haemolysis, hyperglycaemia Contact dermatitis, Chromium Defective glucose allergy, lung cancer metabolism, hyperlipidaemia, corneal opacity Iron Anaemias Haematochromatosis, siderosis, cardiac failure, cancer None yet recognised Manic depression Lithium Anaesthesia Convulsions Magnesium Molybdenum Growth depression, Anaemia, “gout-like” keratinization defects, lesions, persistent hyperpurinaemia dysentery (sheep and cattle) Manganese Skeletal deformities, Ataxia, liver cirrhosis, testicular dysfunction psychiatric disorders

Bio-Inorganic chemistry and its pharmaceutical applications

Selenium

Liver necrosis, endemic cardiomyopathy (Kesham disease), osteoarthropathy (Kashin’s Beck disease), muscular dystrophy (sheep and cattle), membrane malfunction Zinc Anorexia, dwarfism, anaemia, hypogonadism, hyperkeratosis, acrodermatitis enteropathica, depressed immune response, teratogenic effects. Non-essential, non-beneficial elements Aluminium None yet recognised Antimony

None yet recognised

Beryllium

None yet recognised

Cadmium

Reduced growth

Lead

None yet recognised

Mercury

None yet recognised

Thallium

None yet recognised

Radioelements e.g. None yet recognised radium uranium, thorium, plutonium and radioisotopes

627

Teratogenesis, foetal toxicity, hepatic and renal damage, cancer, blind staggers in cattle.

Hyperchronic anaemia, metal fume fever at high doses

Bone disorders, encephalopathy, dialysis dementia Hepatic and cardiac damage, cancer Skin and lung irritation, granuloma, cancer Renal, hepatic, cardiac, skeletal and haematological disorders Anaemia, encephalitis, renal damage Encephalitis, neuropsychiatric disorders, renal damage Central nervous system damage, loss of vision or hearing Cancer induction, genetic effects, cataract

Introduction to the principles of drug design and action

628

of essential and beneficial elements • Several metals play important roles through redox reactions in facilitating electron flow to enable high speed transport of signals in nerves, and in storing energy or information.

13.3 METAL COORDINATION CHEMISTRY 13.3.1 General aspects Since many biochemicals and pharmaceuticals are often near ideal ligands for transition metal ions it is appropriate here to summarise briefly the basic aspects of the chemistry of metal complexation. For a full discussion of this topic the reader is referred to the works given in the references (Huheey 1978; Stenlake 1979; Taylor and Williams 1995). A ligand may be defined as a chemical which has a pair of electrons that can be donated to a vacant orbital in a metal ion. Important biochemical/pharmacological donor groups are RS−, −NH2, −COO− and −O−, in addition other groups such as −PO43−, −NO3− etc. can form dative covalent bonds by donating electrons into vacant orbitals in a metal ion. Metal coordination is biochemically important because it may mask the normal chemical properties of the metal or alter the properties of the ligand. The classic example of this phenomenon is the relatively non toxic substance potassium ferrocyanide (K4Fe(CN)6), (see Figure 13.3) which is made by mixing together two very toxic solutions of potassium cyanide and ferrous cyanide. The underlying chemistry of this was explained more than a century ago by Werner, who postulated the following principles.

Bio-Inorganic chemistry and its pharmaceutical applications

629

Figure 13.3 Werner’s complex. Metals have two types of valency; first, a primary and ionizable valency, that is satisfied by negative ions such as Cl−, NO3−, SO42−, CN−, and a secondary and nonionizable (covalent) one. For each metal there is a fixed number of secondary valencies—called the coordination number; for Fe2+ the coordination number is 6, while for Cu2+ it may be either 4 or 6, and coordination numbers ranging up to 8 are found with other metals. The secondary valencies are directional in nature radiating out from the central metal ion towards the corners of a regular tetrahedron or a regular octahedron with coordination numbers or 4 and 6, respectively. In potassium ferrocyanide the complex iron-cyanide moiety, [Fe(CN)6]4−, ionises as a complete unit and no highly toxic CN− ions are produced. Large ligands can be attached to the metal ion by two or more bonds providing that their spacing is sufficient to accommodate the spatial distribution of the secondary valencies of the metal ion. The terms bi-, tri-, hexa- or polydentate are used to describe ligands with 2,3,6, or many points of attachment. Often rings are formed which are called chelate rings, 5- and 6- membered rings are most stable and hence are those most commonly encountered. Chelation may increase bond strengths by factors even as great as one millionfold; such polydentate ligands as the porphyrin rings in haemoglobin or vitamin B12, form exceptionally stable complexes which fix Fe2+ and Co2+, respectively, in their lower oxidation states and facilitate the biochemical functions of the molecules. The directional nature of the secondary bonds of a metal ion enables them to act as templates to hold specific configurations; such templates probably played a major role in biochemical evolution and they are important to-day in, for example, the production of chiral isomers.

Introduction to the principles of drug design and action

630

In biochemistry and pharmacology complexing ligands can be used to hold a metal in an unfavourable valence state, to neutralize a charge to enable a metal complex to pass through a cell membrane, to produce exceptionally stable bonding by multiple chelation and to assist in the development of ligand drugs designed to react with a specific ion in vivo, for example to detoxify and remove a polluting metal. 13.3.2 Metal-ligand specificity 13.3.2.1 Complex stability Metal ions and ligands show a definite order of affinity for each other. One method of quantifying the degree of tightness of the binding is to use mass-action equilibrium data for the reactions between the metal and a ligand, or ligands, to calculate a formation constant, K or β. Thus for the reaction of one atom of a metal M with one molecule of a ligand L the equilibrium is:

(13.1) where K1 is the formation constant. A second constant K2 describes the stability of the 2:1 complex ML2. A cumulative constant, β=K1.K2, is used to describe the overall reactions:

(13.2) When considering whether the modus operandi of a new or existing pharmaceutical may involve metal ion complexing it is helpful to be able to predict which metal iondrug interactions may occur and how strong they will be. Such predictions can be made using the hard and soft acid and base approach. 13.3.2.2 The hard and soft acid and base approach (HSAB) This approach assumes: a) that if a bond exists between two atoms, one will play the role of an acid and the other a base, and b) that electrons hold the bonded atoms together. The acid is taken to be a Lewis acid-type species (atom, molecule or ion) that has vacant accommodation for electron pairs and the base has the tendency to give up electron pairs to the acid. A typical acid-base reaction is:

Bio-Inorganic chemistry and its pharmaceutical applications

631

(13.3) The bond A–B may be any chemical bond, e.g. [(H3N)5Co–NH3]3+, C2H5–OH or H– OH. Thus we may theoretically dissect any species into an acid and a base fragment, irrespective of whether either fragment could exist in isolation. Softness arises from the electron mobility or polarizability of a species. If the electrons are easily moved, the species is soft, if they are firmly held it is hard. Other descriptions of a soft base could include such terms as polarizable, easily oxidized, loosely held valence electrons: these are all associated with a low charge density on the base. A hard acid has a high charge density, is of small size and, usually, does not contain unshared electrons in its valence shell. Naturally, a hard base and a soft acid are the converse of these descriptions. The single principle underlying the HSAB approach is that a strong bond is formed by a hard acid combing with a hard base, or a soft acid combining with a soft base. Hard-soft bonds are weak. Some examples of physiologically or pharmacologically important hard and soft acids and bases are shown in Table 13.3. 13.3.2.3 The HSAB principle in biochemistry and pharmacy The HSAB concept Pearson 1963) that strong bonds are formed only between hardhard or soft-soft components is widely seen in chemistry, biochemistry and pharmacy and many examples are to be found in the suggestions for further reading. One of the earliest examples of this concept at work is found in the observations made by Berzelius (1779–1848) who, in the early years of the 19th Century, noted that some metal ores occurring on the Earth’s surface were carbonates or oxides, while others were sulphides. This can be explained on the HASB principle, since hard acids, e.g. Mg2+, Ca2+, Al3+ or Fe3+ form strong bonds with hard bases such as O2−, CO32− or SO32−, whereas softer acids, e.g. Cu+/2+, Pb2+ or Ag+, prefer soft bases such as S2−. Any hard acid-soft base or soft acid-hard base compounds would have been so unstable that they would have hydrolysed away many millions of years ago. Many transition metal ions which are essential to biochemistry are shown as ‘borderline’ in Table 13.3, but through chemical symbiosis they can exhibit both hard and soft properties. This is especially prevalent when a metal has more than one oxidation state. Symbiosis is

Table 13.3 Hard and soft acid and base (HSAB) classification of some physiologically or pharmacologically important metals and ligands (Extracted from Pearson, R.G. 1963). Hard Soft Borderline Acids

Introduction to the principles of drug design and action

H+, Li+, Na+, K+ Mg2+, Ca2+, Mn2+ N3+, Cl3+, Cr3+, Co3+, Fe3+ As3+, Si4+ VO2+, MoO3+

HX (hydrogen bonding molecules) Bases H2O, OH−, F−, CH3CO−, PO43−, SO42−, Cl−, CO32−, NO3−, ROH, RO−, R2O, NH3, RNH2

632

Cu+, Tl+, Hg+, Fe2+, Co2+, Ni2+, CH3Hg+ Cu2+, Zn2+, Pb2+, 2+ 2+ 2+ 4+ Cd , Pt , Hg , Pt Sn2+, Bi2+, SO2, NO+ Co(CN)52− Tl3+, BH3, RS+, RSe+, I+, Br+, HO+, RO+ I2, Br2, chloranil, quinones, O, Cl, Br, I, N, RO, RO2 M0 (metal atoms)

R2S, RSH, RS− I−, SCN−, S2O32−, CN−, RNC, CO, C6H6, H−, R−

C6H5NH2, N3−, Br−, NO32−, SO32−, N2

a process whereby one hard, or soft, base on a metal ion attracts other hard, or soft, bases to join it. Metal ions are often in a state of dynamic equilibrium between two oxidation states and the lower state can be stabilized by adding soft ligands and vice versa. However, if very hard or very soft ligands are added the metal will be completely anchored in one oxidation state thus preventing the natural biochemical process (e.g. a redox reaction). This, of course, “poisons” the system and many of the best known poisons are acids or bases that are so strongly held to the active sites of an enzyme that the site is effectively blocked. Cadmium ions and organic mercurials are good examples of soft acid poisons. In addition to site blocking, poisoning by heavy metal ions may result in deleterious structural changes in the protein, or even its precipitation. Soft acids, such as the heavy metal ions, bind strongly to sulphur groups and thus may effectively rob the organism of important sulphur-containing proteins, for example those containing cysteine residues. Very soft base poisons may effectively deprive the body of metal ions. For example, cyanides, sulphides and trivalent arsenic compounds, exert their toxic effect by attaching to the metals in metalloporphyrins and metalloenzymes and at high concentrations they may remove the metal ion entirely from the protein. Conversely, non-poisonous or inert materials are needed for artificial prostheses which have to be inserted surgically. Thus, pure metals, or mixtures of pure metals (alloys), which, if dissolved, would give soft acids are often chosen, e.g. gold, silver, tantalum. Because such metals yield soft ions there is a negligible tendency for them to form complexes with hard bases, such as water, carbonate or biological amines. Some of the earliest chelating, or sequestering agents were used clinically for the treatment of excessive metal ion accumulation. This principle of HSAB matching of acids and bases to achieve strong bonds is illustrated in Table 13.4.

Bio-Inorganic chemistry and its pharmaceutical applications

633

Table 13.4 Some of the earliest sequestering agents used for treating metal excesses (from Fiabane and Williams 1977 with permission). Ligand Donor Metals HSAB atoms mobilised classification Dimercaptopropanol 2 S As, Au(I), Soft [BAL] Hg(I,II) EDTA 4 O, 2N Pb(II), Co(II) Borderline etc. D-Penicillamine S, N, O Cu(I,II) Borderline/soft Desferrioxamine Several Fe(III) Hard O 13.3.2.4 In vivo complexing and metabolic specificity Replacement of zinc in carboxypeptidase with the isomorphous cobalt(II) ion leads to increased peptidase activity. Why then does the human body contain only zinc carboxypeptidase and no cobalt carboxypeptidase? This question is often asked and the answer reveals two levels of sophistication. The natural abundance of the elements and, hence, the dietary contents of our ancestors and ourselves excludes some elements which may appear to be feasible on HSAB grounds and on the basis of strict isomorphous replacement. The importance of this natural abundance factor is further illustrated by the Fe2+-containing porphyrins. The order of preference for the central metal ion as determined (a) thermodynamically is Ni2+ Co2+>Fe2+>Zn2+ and (b) kinetic measurements have shown the order to be Cu2+>Co2+ >Fe2+>Ni2+. Thus human haemoglobin could equally well have contained cobalt in place of iron, but iron is more than 1000 times more abundant in the hydrosphere than cobalt, thus mammalian haemoglobins are iron-based. The whole process by which an element, or an energy rich compound of it, is absorbed from the diet and incorporated in the correct oxidation state into a metalloenzyme or other active centre and finally excreted from the body involves very many complexing reactions. These involve many steps including the absorption of the metal from the gastrointestinal tract, its transport across many membranes and the buffering and redox reactions necessary to ensure that its concentration and oxidation state are appropriate at each stage. Although cobalt can apparently replace zinc in carboxypeptidase in vitro, it certainly cannot follow through all the other biochemical processes that are necessary to insert it into the enzyme in vivo. If this chain of metal insertion into enzymes is broken, the organism cannot thrive, or multiply. Sometimes, a contaminating metal can survive through the chain and be inserted into the enzyme but cannot be excreted and accumulates in the body as a poison. Cadmium is one such example; it can be incorporated into metalloenzymes and other metalloproteins but accumulates there as a poison. There are, of course, usually exceptions to prove the rules; some organisms require elements that are not required by Homo Sapiens, for example protochordates use vanadium instead of iron in their oxygen transport system.

Introduction to the principles of drug design and action

634

Two key concepts run through the discussions in this section, selectivity and specificity. Normal human biochemistry is usually very specific. However, when man seeks to disturb normal biochemistry, for example by introducing a ligand drug designed to mobilize a metal ion, or to disturb some aberrant metal-ligand interaction, in order to treat disease, the best that can be hoped for is selectivity. There will inevitably be some side-reactions (not necessarily manifesting themselves clinically) because more than the one desired metal will be complexed or because less than 100% of the drug will be absorbed from the intestine and transported to the site of action. Examples of this will be discussed later. 13.4 PHARMACOLOGICAL AND PHARMACEUTICAL CONSIDERATIONS The concentrations of many essential and beneficial metal ions in vivo are controlled within narrow limits, some being complexed to high molecular mass species such as proteins and others to low molecular mass ligands such as amino- or carboxylic acids. In some metalloenzymes the metal ion is held inertly and cannot be removed without dismantling and rebuilding the whole molecule, while in others the metal ion is reversibly complexed and is in equilibrium with low molecular mass species and aquated metal ions. Table 13.5 shows some examples of these metalloproteins. The non-inertly bound portion of the metal is in a state of labile equilibrium which can readily be disturbed by internal or external causes so that the organism can no longer function normally. As is indicated in Table 13.2 a number of disease states have been associated directly with changes in the concentrations of trace metals in tissues and body fluids. Drugs whose modus operandi is through electron donation may well complex metals, either as their central mechanism or as a side effect. Similarly chelating agents have been used as drugs designed specifically to sequester metals, especially toxic ones such as those shown in Table 13.4, and to mobilise metals from labile protein complexes into low molecular mass forms. Invariably the matching of ligands to metals is in keeping with the principles of the HSAB concept. The late development of knowledge of this subject has arisen, in part, from the fact that most of these trace metal complexes are normally present at concentrations below those measurable by even the most sophisticated analytical techniques. Further, any attempt to concentrate the species up to analytical levels completely upsets the equilibria shown in Table 13.5 so that unrepresentative species (usually the least soluble) precipitate or are extracted. During the last three decades, computer simulation has been used to analyse metals in biofluids at normal biological concentrations. In its simplest form, computer speciation accepts that constants are invariant and just as the bioavailability of aspirin can be calculated from its protonation constant, so the computer can combine the effects of literally thousands of such constants, solubility products, etc. along with total metal and ligand concentrations to compute the distribution of species in biofluids at steady state.

Bio-Inorganic chemistry and its pharmaceutical applications

635

Table 13.5 Inert and labile metal-protein systems in vivo (from May and Williams 1980 with permission). Inert and/or Labile and thermodynamically thermodynamically nonreversible reversible Iron Haemoglobin Myoglobin (Ferritin) Transferrin↔LMM Fe3+ complexes↔[Fe(H2O)6]3+ Copper Caeruloplasmin Serum albumin↔LMM Cu2+ (Metallothionein) complexes↔[Cu(H2O)6]2+ Zinc α2-macroglobulin Serum albumin↔LMM Zn2+ Metallothionein complexes↔[Zn(H2O)6]2+ LMM=Low molecular mass complexes. Table 13.6 Predominant copper complexes identified in blood plasma by computer speciation analysis. Complex species Percentage of total Cu(II) in the low molecular mass fraction Cu(II)-Histidine19 Glutamine* Cu(II)-Histidine2* 15 Cu(II)-Histidine14 Threonine* Cu(II)-Histidine11 Serine* *All these species are anions. The analysis of the concentration and distribution of the different chemical forms (species) of a metal in a system is known as speciation. As an illustration Table 13.6 shows the principal species in which copper(II) exists in normal blood plasma. Computer simulations of metal speciation in biofluids, such as plasma, saliva, intestinal juices or milk, often run to more than 10,000 species. They require a databank of well-validated constants and analytical data and a medium-sized computer: nowadays many speciation programmes can be run on a high performance personal computer. It is important that as far as is possible computer simulations should be validated by experimental data. Many of the principles discussed earlier in this chapter have been revealed, or retrospectively checked out using computer speciation.

Introduction to the principles of drug design and action

636

13.5 METALS AS THE MODUS OPERANDI OF DRUGS Some drugs are metal compounds in which the metal is essential for their desired pharmacological effect; others are ligands which exert their effects by interaction with an endogenous metal at their site of action. The use of metals as drugs can be traced back for thousands of years, for example zinc oxide unguents for wound healing or solutions of rust in acid wine for the treatment of anaemia were used in pre-Christian times. Paracelsus in the 16th Century made the general introduction of some heavy metals into the materia medica. 13.5.1 Metals as drugs More than a dozen metals have been used as drugs during the past few hundred years, including such highly-toxic elements as arsenic, mercury and lead; Table 13.7 lists some of those which still find applications in healthcare. A few of these applications are discussed below. 13.5.1.1 Bismuth in the treatment of peptic ulcer The heavy metal bismuth has been widely used for many years for the treatment of gastric or duodenal ulceration. One of the most prescribed of these agents is a solution of tri-potassium dicitratobismuthate, containing also colourant and emollients, at pH 10. Intragastric fibre-optic colour photography suggested that the bismuth acted as a cytoprotectant by coating the ulcerated area with a precipitate which protected the raw

Table 13.7 Some metals present in pharmaceuticals administered to humans. Element Compound Prescribed as—and use Aluminium Hydroxide Aludrox—antacid Silicate Kaolin—antidiarrhoeal Antimony Gluconate Pentostam—antilieshmaniasis Bismuth Tripotassiumcitrato De-Nol—antacid, antiulcer Boron Boric acid Monphytol—antifungal Cobalt Cyanocobalamin Cobalin-H—pernicious anaemia Iron Glycine sulphate Ferrocontin continus—iron deficiency Gold Thiomalate Myocrisin—antiarthritic Magnesium Sulphate Epsom salts—laxative Platinum Dichlorodiammine Cisplatin—anticancer agent Selenium Sulphide Selsun—seborrhoeic dermatitis Silver Sulphadiazine Flamazine—infected leg ulcers Tin Fluoride Toothpaste—anticaries Zinc Sulphate Solvazinc—proven zinc deficiency

Bio-Inorganic chemistry and its pharmaceutical applications

637

surface from further attack from the gastro- or intestinal juices. Computer speciation analysis subsequently showed that in the acid environment of the stomach the soluble charged Bi(citrate)23− is converted into insoluble precipitates of bismuth oxychloride and insoluble bismuth citrate. Computer simulation also led to other bismuth formulations which produce cytoprotective bismuth patches over ulcers at sites having more neutral pH values. However, optimisation of such ulcer therapies is difficult because of the vagaries of ulcer origins and their response to therapy. A new aspect to ulcer therapy with bismuth compounds was given by the discovery in the early 1980s that the intestinal bacterium Helicobacter pylori probably plays a major role in the induction of gastric ulcers, and even gastric cancer. Bismuth is itself bacteriocidal, but its effect is weak, however, combination of bismuth with antibiotics such as amoxicillin or tetracycline can raise the success rate for ulcer therapy to about 80%, i.e. by about four times that achievable with bismuth alone. Bismuth oxide or bismuth sub-gallate are also used in combination with other agents in antiseptic preparations for the treatment of haemorrhoids. 13.5.1.2 Metals as anti-cancer agents A number of metals, such as gallium, hafnium and palladium, have been shown in animal experiments to be able to slow down the growth of tumours, but none has yet proved sufficiently effective to justify clinical trials. However, an important, but accidental, discovery in 1964 led to the development of an exciting new class of inorganic cytotoxic drugs for cancer chemotherapy. Studies of the effects of electrical currents on the growth of microorganisms led to the discovery that the simple presence of a platinum electrode in a culture, without any electric current being applied, led to severe growth disturbance. The active species were recognised to be tiny concentrations of platinum complexes which were formed in the culture. Tests of one such complex, cis-dichlorodiammineplatinum(II) (cisplatin) (13.1), in animal tumour systems showed that this was a very potent cytostatic agent.

Introduction to the principles of drug design and action

638

Unfortunately, it also exhibited such high general toxicity, especially nephrotoxicity, that any clinical application appeared unlikely. However, despite this disadvantage a Phase I clinical trial was commenced which confirmed the high anti-tumour effectiveness of cisplatin and showed that some of the more general toxic effects could be avoided by judicious hydration of the patient. Cisplatin is a planar, electrically neutral complex that is able to cross cell membranes. Its cytoxic action arises by losing the two adjacent chloride ions to form platinum chelates with two nitrogens from the purine and pyrimidine bases in the desoxyribonucleic acid (DNA) chain to create an intrastrand link within the DNA that interferes with the replication of the DNA when the cell next attempts to divide. The formation of the intrastrand link by cisplatin in a single strand of the DNA contrasts to the formation of an interstrand link between the two DNA strands of the α-helix which commonly occurs with other types of alkylating agent, such as those based on nitrogen mustards (see Chapter 9). This difference in the mode of action of cisplatin and the nitrogen mustards (13.2) probably lies in the separation between the two chloride ions in the molecules. In cisplatin the distance between the two chloride ions is 0.33 nm, while the separation between the chloride ions at the ends of the arms of the nitrogen mustards is 0.80 nm; these distances correspond perfectly to the space required to form inter- and intrastrand linkages, respectively. Cisplatin was introduced clinically in the UK in 1979 and rapidly became a first line drug for use, either alone or in combination with vinblastine or other cytoxic agents, in the treatment of ovarian and testicular cancer and also lung cancer. However, cisplatin suffers from serious disadvantages; it must be infused intravenously; it causes severe nausea and vomiting, being one of the most powerful emetics known; it causes leukopenia and renal dysfunction cannot be entirely avoided. Further, some tumours are beginning to develop resistance to the drug. Much research is being concentrated on reducing these disadvantages. The attachment of a 1,1-dicarboxycyclobutane moiety to the diammineplatinum(II) to produce the cis-(1,1-dicarboxycyclobutane)diammineplatinum(II) (13.3) derivative has removed some of these problems and introduced a second generation agent, carboplatin, into clinical use.

Bio-Inorganic chemistry and its pharmaceutical applications

639

More recently, an orally active platinum drug has been developed. This substance bisacetatoamminedichloro(cyclohexylamine)platinum(IV), code named JM216 (13.4), is metabolised in vivo to produce, as the main plasma metabolite, the ammine(cyclohexylamine)dichloroplatinum(II)(JM118) (13.5) derivative, which is thought to be the active cytotoxic moiety. The orally active drug JM216 is expected to enter clinical practice in the late 1990s. Another fascinating new compound, which is still in the research phase, is transammine(cyclohexylamine)dichlorodihydroxyplatinum(IV). This compound, which appears to form interstrand crosslinks in DNA, is the first trans-platinum complex to show selective anti-tumour activity in vivo. 13.5.1.3 Copper and rheumatoid arthritis This condition afflicts about 5% of the UK population and is particularly prevalent amongst the elderly. The origins of the disease are unknown but it appears to involve disturbances in the chain of reactions which control the patient’s autoimmune response. Some of the weaker links in this chain appear to involve an inadequate supply of copper. This element is so widespread in nature than copper deficiency in

Introduction to the principles of drug design and action

640

humans is rare, however, there is much evidence to suggest that localised imbalances of tissue copper may be related to rheumatoid arthritic conditions. The first reports that administration of low molecular mass copper complexes have a beneficial effect on rheumatoid arthritis appeared more than fifty years ago (see Sorenson 1982 in the further reading list). The mechanisms by which copper acts against rheumatic disease are still unclear although factors such as free radicals, prostaglandin balance and lysyl oxidase activity have been investigated. Further, despite many years of research, no copper-containing anti-rheumatoid arthritic drug has yet found widespread use in clinical practice. Nevertheless, a review of the many research reports concerning the role of copper complexes in vivo gives a valuable general insight into the considerations involved in the search for a metal drug. Paradoxically, the abundance of information may be a hindrance, rather than a help to our understanding and to the development of effective new copper-containing drugs. The copper in normal human blood plasma occurs in four fractions (see Table 13.5). About 90% of the total plasma copper (~17 mmol dm−3) is complexed in a thermodynamically irreversible manner in the copper-oxidase caeruloplasmin. Some, however, is always attached to other protein-binding sites as part of a labile equilibrium between the remaining three copper fractions. These also include low molecular mass complexes and aquated ions. Apart from caeruloplasmin, albumin is by far the largest of the remaining fractions. It is noteworthy that although the metalloprotein (caeruloplasmin) fraction is not in equilibrium with the other three fractions, in certain circumstances a steady state may be set up whereby the metal is exchanged between them. The metal uptake and release is then associated with the formation and degradation of the protein. Typically, this occurs within cells, thus metalloproteins can serve as homeostatic ion reservoirs. However, the metal in these proteins is not immediately available for other biochemical purposes. It must, therefore, be recognised that unless the metal ion fractions in labile equilibrium are rapidly replenished, a deficiency state could arise even in the presence of apparently high total concentrations of the metal ion in the tissue fluids. Sorenson and his colleagues (1982) have reviewed the results of treatment in some 1500 arthritic patients treated with several experimental copper compounds which produced a significant reduction of inflammation. These agents exhibited anti-ulcer activity which is significant in view of the fact that gastrointestinal irritation is often a limiting factor in anti-arthritic therapy. Sorenson also reviewed a great deal of additional evidence concerning copper complexes as anti-arthritic drugs; therefore only two pertinent findings will be mentioned here. First, copper-deficient rats are significantly more susceptible to carrageenan-induced oedema (an experimental model for arthritis) than control animals, thus establishing that a minimum level of copper in the tissues is necessary for the control of inflammation. Second, speculation about the therapeutic efficiency of copper bracelets in rheumatoid arthritis has been placed on a more scientific basis: the dissolution of copper in human sweat has been quantified, its dermal penetration and its positive beneficial effects demonstrated in a clinical trial. Screening studies following intravenous administration of copper compounds have revealed a striking correlation between the reduction of inflammation and the amount

Bio-Inorganic chemistry and its pharmaceutical applications

641

of the metal entering the plasma compartment. The nature of the ligand attached to the copper in the administered complex does not influence the results, rather, it is the increased availability of ions per se which affords protection against the inflammation. Whatever, the mechanism of its action the inflamed tissues appear to be able to acquire copper ions from the labile equilibrium system in the plasma once the administered complex has dissociated and the metal ions have been re-distributed, proportionally, amongst the serum albumin and the naturally occurring low molecular mass ligands. Designing copper complexes for the control of inflammation depends on a knowledge of the differences between healthy and diseased tissue at the molecular level. As has already been noted, such knowledge is still incomplete. However, whatever its true mechanism of action, the observed positive effects of copper on inflammation does suggest that a valid way forward is research aimed at increasing the supply of copper to sites of injury. The rationale for designing rheumatoid arthritis agents which reduce inflammation in this way depends on two fundamental assumptions: I. the therapeutic effect of copper administration arises from an increase in the total labile copper concentration in relevant body compartments, such as synovial fluid, and that II. this increase is fostered by the formation of complexes in plasma that can diffuse across membranes into the synovial fluid. Computer simulation predicts that the supply of copper to the tissues will be enhanced simply by increasing the labile copper concentration in the plasma itself. However, the direct injection of copper into the plasma is fraught with problems. Intravenous injection of <1 mg dm−3 copper may produce toxic symptoms, even haemolysis of erythrocytes. Such levels may also exceed the normal binding capacity of the serum albumin, albeit perhaps only locally, leading to non-specific protein binding and an abnormal tissue distribution of the metal. Although it is possible to alleviate these undesirable effects by slow infusion, by subcutaneous injection or by using chelators to help attain equilibrium binding, such a means of administering copper would not be an appropriate regime for the routine of treatment for a disease as prevalent as rheumatoid arthritis. Oral administration of the same compounds is a possible alternative to intravenous injection, but this route is much less effective due to the fact that 30% or less of dietary copper is absorbed into the blood stream from the gastrointestinal tract. An alternative strategy could be based on the controlled liberation of endogenous reserves of copper from the liver or other tissues. For short-term therapy aimed at correcting a localised copper imbalance, endogenous rather than exogenous sources seem to offer a simple solution. There are three possible ways of achieving this objective: a) by equilibrium competition for labile protein-bound copper, b) by decreasing the affinity of serum albumin for copper by allosteric effects, and c) by extracting copper from inert metalloproteins.

Introduction to the principles of drug design and action

642

It has been suggested that penicillamine may act via the last mechanism in stimulating copper excretion in the treatment of patients suffering from Wilson’s disease (see below). To bypass some of the difficulties of endogenous mobilisation, attention has been drawn to the possibility of copper supplementation by dermal application. The area behind the ear has been suggested as an application site since, due to its low keratin content the skin in this region is about four times as permeable as skin elsewhere. However, the complexes selected must be able to penetrate the dermis and pass, via the lymph, into the plasma, thus similar considerations to those discussed above for intestinal absorption apply. Dermal absorption does offer the advantage that there are no homeostatic control mechanisms to overcome. Thus, complexes need not be so specific for copper(II), nor so powerful. The risks of gastrointestinal irritation by the metal ion are also avoided. For these reasons topical application of copper complexes for the treatment of rheumatoid arthritis appears to be attractive. 13.6 DRUGS WHICH EXERT THEIR EFFECTS VIA METAL COMPLEXATION OR CHELATION 13.6.1 Metal chelation in antimicrobial activity Many microorganisms are critically dependent on one or more metals for their growth and a number of the more effective antimicrobial drugs act by denying the organism the use of such metals (Taylor and Williams 1995). The chelating agent 8-hydroxyquinoline, oxine (13.6), has the advantage over many other antimicrobial agents of acting rapidly and also possessing fungicidal properties. It was one of the first such agents to be shown to exert its action via chelation. The lack of antimicrobial action with derivatives in which the ligand donor groups were blocked with –O–methyl, ≡N–methyl, or isomeric hydroxyderivatives— suggested that the activity of oxine involved chelation. Traces of ferrous or ferric iron were shown to be required for activity and evidence suggests that the mode of action is inside the cell, or at least within the cytoplasmic membrane, since derivatives of oxine having increased hydrophilic properties and also able to chelate iron, e.g. 8hydroxyquinoline-5-sulphonic acid, are not antibacterial. The oxine-iron complex, as a result of rearrangement of the orbitals of the Fe3+ ion, is able to catalyse the oxidation of thiol groups in lipoic acid, an essential co-enzyme required by bacteria for the oxidative decarboxylation of pyruvic acid. The importance of the lipophilic properties is illustrated by the activity of halogenated derivatives such as 5,7-diiodo-8hydroxyquinoline and 5-chloro-8-hydroxy-7-iodoquinoline against the organisms causing bacterial dysentery. The entry of the anti-tubercular agent isonicotinic acid hydrazide, isoniazid, (13.7), is mediated by the formation of a lipid-soluble copper chelate. The activity of other antitubercular drugs, including thiacetazone (13.8) and ethambutol (13.9) are also dependent

Bio-Inorganic chemistry and its pharmaceutical applications

643

on copper(II) chelation. The antiviral activity of methisazone (13.10) appears to arise from copper complexing to the ring carbonyl oxygen and the middle nitrogen atom of the thiosemicarbazide side chain. Such chelates have been shown to interact with nucleic acids. The tetracyclines (13.11) constitute a group of important agents for treating systemic bacterial infections. High values for formation constants and the presence of hard basic groups, such as hydroxyl anions and tertiary amino moieties, indicate a readiness to complex Ca2+ and Mg2+. Much evidence suggests that tetracycline owes its antibacterial activity to its ability to complex Mg2+ in the bacterial cell membrane. The increased lipophilicity of the Mg2+-tetracycline complex facilitates concentration

Introduction to the principles of drug design and action

644

in the bacterial cell where it blocks protein biosynthesis by interfering with the binding of aminoacyl-t-RNA to ribosomal receptors. 13.6.2 Metal ion removal In metal storage diseases, such as Wilson’s disease and haematochromatosis, the symptoms can often be alleviated and the progress of the disease slowed down by treatment with a chelator with a high affinity for the metal concerned. Similarly, exogenous metal poisoning can also be treated by appropriate chelators. 13.6.2.1 Removal of copper in Wilson’s disease Wilson’s disease is an idiopathic condition characterised by an inability to use copper for caeruloplasmin synthesis with the result that there is massive overloading of albumin and the low molecular mass ligands with copper. Dietary copper becomes deposited in

excessive amounts in the brain, liver, eyes and other tissues causing neurological symptoms and cirrhosis of the liver leading to death relatively early in life. The progression of the disease can be markedly slowed by treatment with chelating agents such as D-penicillamine (13.12), or triethylenetetramine (TREN) (13.14). The success of D(−)-penicillamine, a hydrolysis product of some penicillins, as a copper mobilising agent in Wilson’s disease, as well as in the treatment of rheumatoid arthritis, is interesting because this agent does not release copper from serum albumin. Neither does it degrade caeruloplasmin and release its vast stores of copper; it may possibly liberate the metal from liver metallothionein binding sites, but it certainly does markedly increase the urinary excretion of copper. Although D-penicillamine does not apparently disturb serum albumin copper, it does mobilize zinc from this protein and this, together with bone marrow depression, are recognised side effects of D-penicillamine therapy. These are serious side-effects and D-penicillamine is now often replaced by the tridentate ligand triethylenetetramine(TREN) (13.14) in the treatment of Wilson’s disease.

Bio-Inorganic chemistry and its pharmaceutical applications

645

13.6.2.2 Removal of iron in haematochromatosis The massive storage of iron encountered in either primary or secondary haematochromatosis is another condition that is amenable to treatment with chelators. The secondary haematochromatosis which results from the need for repeated blood transfusions in the treatment of the genetic disease thalassaemia (sickle cell anaemia) is a major clinical problem in some tropical countries. For the last three decades the fungal siderophore desferrioxamine (Desferal) (13.15) has been the only selective iron chelator available for clinical use. This is an extremely powerful ligand for the chelation of Fe3+ and it has the advantage of having little affinity for other essential metals such as copper, zinc, calcium or magnesium. Desferal forms hard acid-hard base complexes with iron in which the iron is bound more strongly than in the FeEDTA complex. X-ray diffraction studies suggest that the complex formed involves iron(III) bound to three −N(OH)CO− groups.

Desferal is very expensive and must be administered by slow intravenous infusion. These are serious disadvantages when the pressing clinical need is to be able to treat large numbers of young people in poor countries. There is great interest in developing inexpensive, orally active chelators for iron removal and derivatives of 3hydroxypyridine-4-one appear to offer the desired properties. One such derivative, originally code-named L1 (13.16), has recently entered clinical practice under the name Deferiprone. This agent is able to mobilise iron from the main iron-storage protein ferritin, possibly because the molecule is small enough to enter into the socalled tunnels in the ferritin molecule and to directly chelate the ferric iron deposited there. The hydroxypyridones have also been shown to mobilise iron from the degradation product of ferritin, haemosiderin, that accumulates in the tissues of thallassaemia patients. The iron so mobilised is excreted from the body mainly via the urine.

Introduction to the principles of drug design and action

646

13.6.2.3 Treatment of exogenous metal poisoning Acute or chronic poisoning by lead, or other heavy metals, remains an important clinical problem. The polyaminopolycarboxylic acid chelator ethylenediamineN,N,N',N'-tetraacetic acid (EDTA)(13.17) has a high affinity for lead and this agent has been shown to be a reasonably efficient chelator for lead in vivo. However, EDTA suffers from some disadvantages. First, as a charged ion ETDA4−, the form in which the drug exists in plasma, it is unable to pass through lipid membranes to reach lead deposited in cells; second EDTA is also able to complex essential ions such as Mn2+ and Zn2+ and its prolonged use can lead to a deficiency of these metals. For some purposes EDTA is administered as the disodium salt because the tetrasodium salt is too alkaline. The disodium salt of EDTA, disodium edetate, is also used to reduce blood calcium levels in hypercalcaemia. The hard basic groups, −COO− and −NH2, of the

tetradentate EDTA form a stable water-soluble complex with the hard Ca2+ ions which is excreted via the kidney. Disodium edetate has also been used for treating limeburns on the cornea and for restoring the K+/Ca2+ balance in cardiac arrhythmias accidentally induced by digoxin. “Decorporation therapy” by regular infusions of disodium edetate is also used as an rather equivocal form of treatment for patients with atherosclerosis; the rationale being that the chelator mobilises calcium from the atherosclerotic plaques thus improving general blood flow. Some spectacular results have been claimed by protagonists of this form of therapy! For the treatment of poisoning with lead or mercury the administration of the disodium calcium EDTA complex is preferred as this prevents the chelation and excretion of the essential endogenous calcium. An analogue of EDTA, diethylenetriaminepentaacetic acid (DTPA) (13.18) is employed in the nuclear industry for the treatment of the rare cases of human contamination with plutonium or americium. The problem created by the charged nature of the EDTA4− and DTPA5− species which limits their transport across cell membranes and, thus, their access to intracellular metal deposits may, in theory at least, be circumvented by treatment with two ligands. If a ligand that exists in electrically neutral form in the blood, for example D-penicillamine, is administered first, followed by EDTA administration, then intracellular deposits of a metal such as lead may be mobilised and transported into the

Bio-Inorganic chemistry and its pharmaceutical applications

647

blood stream as neutral D-penicillamine complexes and then quantitatively eliminated in the urine following interaction with EDTA—such a regimen of using two chelators is known as Synergistic Chelation Therapy. Mercury poisoning has also been treated successfully with D-penicillamine, although the N-acetyl derivative (13.13) which has a softer basic group is considered to be more effective. Poisoning by mercury, arsenic, gold or antimony may be treated with dimercaptopropanol (dimercaprol (13.19) or British Anti-Lewisite (BAL)), one of the first chelators to be used clinically. Dimercaprol provides soft sulphydryl groups that bind these soft metals forming water-soluble complexes. Dimercaprol must be injected in an oil suspension which has a number of disadvantages, and the water-soluble derivative dimercaptopropane sulphonate, dimeval (13.20), has been used as an alternative drug for mercury and arsenic poisoning. 13.7 METAL-DEPENDENT SIDE EFFECTS OF DRUGS A prima facie case can usually be made linking a metal ion interaction with almost any drug in the pharmacopoeas and several therapeutic examples of such links have already

Introduction to the principles of drug design and action

648

been discussed. However, the interaction with the metal ion is not part of the desired effect and may lead to unwanted side effects. Examples of such effects would be a drug that in the presence of an endogenous metal ion is extensively inactivated, or a substance that produces a metabolite that complexes with, and inactivates, an essential metal, or an agent designed to accelerate the excretion of an unwanted metal but which also enhances the excretion of an essential metal. It has been postulated that thalidomide (13.21), the drug which when taken by pregnant females as a tranquilliser caused so many birth abnormalities in the 1960s, could have produced its effects of limb-shortening through hydrolysis in the embryo of the peptide-like bonds in the molecule, to produce complexing moieties that sequestered the Ca2+ ions that were essential for the normal development of the limbbuds. 13.7.1 Metal ion sequestration by a metabolite Many anti-tubercular drugs chelate metal ions and, as mentioned earlier, this can enhance their biological activity by rendering them more bioavailable. However, whether such metal chelation is a characteristic of their activity or not, the administration of a ligand drug is likely to interfere with normal trace element behaviour. Such interactions need to be characterized and quantified lest topping-up therapy is necessary. Ethambutol ((+)-2,2'-ethanediyldiimino)-bis-1-butanol) (13.9), an important antitubercular drug, is one such example. While it has been postulated that the mode of action of ethambutol may involve the formation of a ternary complex involving copper(II) ions and RNA, clinical studies have shown that it does lead to an increase in urinary zinc excretion.

Solution studies have been used to measure the formation constants for the complexing of ethambutol and its principal metabolite, 2,2'-(ethanediyldiimino)-bis-1butane carboxylic acid (EDBA), with a range of metal ions essential to humans. Insertion of these data into a computer simulation of the interactions in blood plasma and showed that a dose of 25 mg ethambutol/kg body weight, which produces an EDBA concentration of 1.25× 10−6 mol.dm−3, raised the concentrations of low molecular mass zinc complexes by up to one third, the new complex Zn-EDBA0 being electrically neutral. However, up to ten times this dose of ethambutol has no effect on

Bio-Inorganic chemistry and its pharmaceutical applications

649

the metal ion speciation. Clearly, the clinical disturbances reported during ethambutol therapy are not due to the drug itself, but to its metabolite EDBA. Further, such undesirable side-effects as peripheral vision loss may be explained by the ability of the neutral zinc complexes formed to cause migration of the metal from relatively zincrich areas in the eye. 13.7.2 Sequestering drug-non-specific metal ion interactions Ethylenediaminetetraacetic acid (EDTA) is the most widely used synthetic chelating agent used in vivo, and it also finds extensive use in pharmaceutical formulations and in industrial and domestic applications, including washing powders. In the USA, EDTA is marketed under almost forty different brand names. As mentioned in Section 13.6., EDTA is used in medicine for the treatment of lead or calcium excess. The disodium salt is usually used because it increases the solubility of the compound. Intravenous infusion is necessary since EDTA4− as a charged ion is poorly absorbed from the gastrointestinal tract making oral treatment largely ineffective. For decorporation therapy quite large doses are administered, up to ~50 mg/kg body weight or ~3 g/day and the treatment may need to be repeated on many times. The first recorded clinical use of EDTA was as the nickel complex which was administered as a treatment for adenocarcinoma, but unfortunately, the agent was almost quantitatively excreted in unchanged form in the urine. This presumably reflects the very high formation constant of the Ni2+-EDTA4− complex and the fact that no metal ions could be released in vivo to exert any anti-tumour effect. Over the last thirty years or so more than two million treatments with EDTA have been administered in the USA alone, either for calcium removal (EDTAHNa3-Limclair) or lead removal (EDTAH2Ca-Ledclair). The agent is not specific for calcium, lead or any similar metal, and it is not without side-effects, including damage to the kidneys at high doses. It is not yet known if all the undesirable effects of EDTA in vivo are due to metal chelation, but prolonged therapy does lead to hypocalcaemia and, to a lesser extent, depletion of essential metals like zinc and manganese. Speciation analysis suggests that a ratio of 3:1 EDTAH2Ca:EDTAH2Zn ought to reduce any need for topping up therapy. In the case of the EDTA homologue DTPA (13.18) which has found use in the treatment of human contamination with plutonium or americium DTPAHZn2 has been shown to be less toxic than DTPAHCa2, presumably because there is no depletion of essential metals such as Zn, Mn or Mo. 13.8 TRACE ELEMENT SUPPLEMENTATION It has been mentioned several times in the foregoing discussion that there may be occasions when treatment with a ligand drug may produce a depletion of essential metals that requires the introduction of a topping up therapy. Such topping up therapy is, of course occasionally necessary to correct other forms of metal deficiency, for example the iron deficiency associated with secondary haematochromatosis.

Introduction to the principles of drug design and action

650

Metal supplementation sounds deceptively simple, but in fact many factors must be taken into consideration in selecting or developing a satisfactory preparation. Oral administration is the most commonly used route of supplementation, but it is also the one presenting the most difficulties. Most of the essential and beneficial metals required by the human body are absorbed from the upper part of the small intestine, a region where the pH ranges from ~6 to ~8, and the metals must reach the absorptive sites on the intestinal mucosa in a soluble and absorbable form. However, at this pH in the aqueous environment of the small intestine most multi-valent metal ions react almost quantitatively with water to form hydrolysed, insoluble, and thus nonabsorbable, hydroxides and oxides. Other reactions occur with the numerous complexing ligands present in the intestinal contents to form metal complexes the major fraction of which may be electrically charged and also non-absorbable; the fraction of electrically net neutral and thus absorbable species may well represent only a tiny fraction of the total metal which enters the gastrointestinal tract. In formulating metal supplements a suitable complexing ligand may be added with a view to enhancing the proportion of soluble, neutral metal complexes which are formed near the absorptive surfaces in the intestine. However, because such complex formation is usually very pH dependent the choice of ligand is not always easy. For example, when ferrous iron reacts with ascorbic acid the total percentage of neutral complexes formed in the pH range 5 to 7.5 is >60%, whereas with galacturonic acid that percentage was <30%, thus ascorbic acid appears to be the better ligand for facilitating iron absorption. 13.8.1 Iron, zinc and copper supplementation These are the most prevalent metals in vivo, as mentioned earlier, and supplementation treatments can be traced back thousands of years into pre-Christian times when solutions of rust in acid wine were used for anaemia and zinc oxide unguents for wounds or skin conditions. To-day, the Pharmacopoeias list more than 40 preparations for iron, but only 3 for zinc and none for copper. Modern oral supplementation therapy involves a combination of scientific approaches to which certain psychological and commercial factors have to be added. These considerations: a) aim to increase the flow of metal complexes from intestine to blood by increasing the concentration of lipid-soluble, low molecular mass complexes present in the intestinal fluids; For example, iron preparations may contain Fe(II) in association with complexing ligands such as ascorbate, malate, fumarate, gluconate or amino acids that promote the formation of neutral complexes at pH ~6 −7; gluconate may be used for the same purpose for metals such as Cu, Zn, or Co; b) tend to favour iron(II) rather than iron(III) compounds, since the former can be up to ~17 orders of magnitude more soluble (Ksp Fe(OH)2=10−15.1, Fe(OH)3=10−38.7); c) cause least irritation to the gastrointestinal tract; for this reason ferrous sulphate is not an agent of choice; d) appear to avoid approaches advocated by hundreds of years of folklore

Bio-Inorganic chemistry and its pharmaceutical applications

651

medicine, so as not to undeservedly undermine confidence in either the product or its prescriber; e) use metal complexes that are capable of being patented. Irritation of the gastro-intestinal tract can be a problem, especially in ironsupplementation. Such irritation tends to follow the Irving-Williams series of complex stability for divalent ions, i.e. MnZn. This indicates why oral supplementation with copper is exceedingly difficult and recourse is made to absorption via the skin from copper-impregnated gels used as a dressing—a modern equivalent of the copper bracelet from which the metal can be solubilised by the amino acids in sweat that form neutral lipophilic complexes which can penetrate the epidermis. Chelation approaches have been used to overcome the problems of gastrointestinal irritation by iron, thus preparations containing ferrous gluconate, ferrous fumarate, ferrous succinate or ferrous glycine sulphate are available. In contrast zinc supplementation presents few problems, administration of daily doses of 150 mg Zn2+ as zinc sulphate being well tolerated over periods of a few weeks. When oral iron preparations are not tolerated by the patient, or there is a need to rapidly increase iron levels, the metal may be administered by injection. A widely used injection form of iron is an iron(III) sorbitol (a reduction product of glucose) citric acid complex, but ferric gluconate and Fe(OH)3-dextran complexes are also used. 13.9 CONCLUDING REMARKS The aim of this chapter has been to indicate, hopefully in a convincing manner, that all agents, in one way or another, interact with metal ions—as modus operandi, as sideeffects as counterions (anions or cations), as competing electron acceptors for drug active groups, as possible impurities or even as trace element supplements. The late development of this subject has been due to the scarcity of analytical techniques capable of measuring the very low metal concentrations involved, and also due to the predominance of organic chemistry in pharmacological and pharmaceutical research with the result that the fact that life is an inseparable combination of inorganic and organic chemical reactions is often overlooked. One might comment on the success, or otherwise, of isosteric modifications in drug design. Bearing in mind the odds against evolutionary processes, unassisted by modern technology, selecting Homo Sapiens to be the advanced organism that we are with perfectly balanced biochemistry, it would appear that drug designers with all their armamentarium of computers and laboratory equipment directed at putting the right isoelectronic group in the right place for ideal drug activity, would be well placed to achieve perfection. What then is missing? The fact is that each active group is not moved from one position to a nearer optimal one without knock-on effects. The same electrons on the groups that interlock with the active sites in order to produce the desired pharmacological effects also attract HSAB bases such as protons and metal ions, each being quantifiable in terms of a formation constant, K. Indeed the target site for a drug’s activity can only react with

Introduction to the principles of drug design and action

652

the active groups if the protons or metal ions are displaced. The pK values, however, vary with the position, the energetics of the molecule and the solvent. It is humbling to note that in nature biochemical reactions do not occur as the discrete sequences that we depict as equations involving letters, numbers or other symbols, but rather that every reaction between elements or molecules is an interaction of packets of electron energy. Since these packets are often quite diffuse in their nature, it is not surprising that biochemical interactions are influenced, or are under the influence of, other packets of energy in their immediate environment. Future successes in drug design, especially in designing agents for treating the more intransigent human diseases, will depend not only molecular modelling and creative synthetic developments, but also on a sound understanding of all the interactions of the drug with both the inorganic and the organic components of the body. FURTHER READING Albert, A. (1979) Metal-binding substances. In Selective Toxicity, 6th edn., pp. 385– 442. London: Chapman and Hall. (a) Daniels, T.C. and Jorgensen, E.C. (1977) Physicochemical properties in relation to biological action. Chelation and Biological action, (b) Schultz, H.W. (1977) Surfactants and Chelating Agents. In Textbook of Organic, Medicinal and Pharmaceutical Chemistry, edited by C.O.Wilson, O.Gisvold and R.F.Doerge, 7th edn., (a) pp. 50–56, (b) pp. 222–246. Philadelphia—Toronto: Lippincott. Fiabane, A.M. and Williams, D.R. (1977) In Principles of Bio-inorganic Chemistry, Monographs for Teachers No. 31 pp. 82–107. London: Royal Society of Chemistry. Fraústo da Silva, J.J.R. and Williams, R.J.P. (1991) The Biological Chemistry of the Elements. Oxford: Clarendon Press. Huheey, J.E. (1978) Coordination Chemistry—theory, structure and mechanisms. In Inorganic Chemistry, Principles of Structure and Reactivity, 2nd edn., pp. 332– 488. New York—London: Harper. Hay, R.W. (1984) Bio-inorganic Chemistry, p. 210. Chichester: Ellis Horwood. May, P.M. and Williams, D.R. (1980) The inorganic chemistry of iron metabolism. In Iron in Biochemistry and Medicine, edited by A.Jacobs and M.Worwood, Vol. II, pp. 1–27. London: Academic Press. May, P.M. and Williams, D.R. (1981) Role of low molecular weight copper complexes in the control of rheumatoid arthritis. In Metal Ions in Biological Systems, edited by H.Siegel, Vol. 12, pp. 283–317. New York: Marcel Decker. Miller, S. and Orgel, L.E. (1973) The Origins of Life. London: Chapman Hall. Pearson, R.G. (1963) Hard and soft acids and bases. Journal of the American Chemical Society 85, 3533–3539. Sorenson, J.R.J. (1982) Inflammatory Diseases and Copper, p. 662. New Jersey: Humana. Stenlake, J.B. (1979) Metal Chelation. In Foundations of Molecular Pharmacology, Vol. 2, The Chemical Basis of Drug Action, pp. 86–99. London: Athlone Press. Taylor, D.M. and Williams, D.R. (1995) Trace Element Medicine and Chelation Therapy. London: The Royal Society of Chemistry.

Bio-Inorganic chemistry and its pharmaceutical applications

653

Williams, D.R. (1976) Introduction to Bio-inorganic Chemistry. Illinois: C.C.Thomas. Williams, D.R. (1983) Historical outline of the biological importance of trace metals. Journal of Metabolic Diseases Suppl. 1, 1–4. Williams, D.R. and Halstead, B.W. (1982–83) Chelating agents in medicine. Journal of Toxicology and Clinical Toxicology 19, 1081–1115.

INDEX A-3253, 483 A-75272, 483 A-80987, 312 AADC, 267, 273, 327 inhibitors, 327 Absorption, 2, 122 7-ACA, 153, 437, 454, 457 Ac-D-Phe-Pro-Arg-B(OH)2, 293 ACE, 249, 267, 276, 300 (R)- and (S)-Acenocoumarol, 148 Acetazolamide, 272 Active site-directed inhibitors, 268 Acetorphan, 278 Acetylcholine, 317, 389 receptor, 390 Acetylcholinesterase, 262, 274, 317 inhibitors, 317–20 α-Acetylenic GABA, 325 N-Acetyltransferases, 19, 467 α1-Acidglycoprotein, 124 Acridines, 443 Acrosoxacin, 443 Actinomycin D, 342, 439, 443 Active tubular secretion, 23 Acycloguanosine, 484 Acyclovir, 246, 279, 484, 488 Acylation, 17 Acetyl-CoA: choline O-transferase, 319 Acyl CoA synthetase, 151 Adenosine, 389 arabinoside, 484 3',5’-cyclic phosphatase, 390 triphosphate synthase, 90 Adenylate cyclase, 391 Adenylylating enzymes, 467 ADEPT, 256, 379 Adozelesin, 357 Adrenaline, 108, 250 (R)-Adrenaline, 115 Adriamycin, 250, 359 ADTN’S, 423 Aerosols, 39 AGP, 124 AIDS, 486

Index

655

(+)-AJ 76, 432 Alafosfalin, 454 Alanine racemase, 274 Alatrioprilat, 309 Allopurinol, 262, 369 Allylamines, 480 Aludrox, 523 ALX-1323, 398 Alzheimers disease, 317, 319 3-AMA, 466 Amantidine, 484, 489 Ametantrone, 359 AMI-193, 401 4-Amidinophenyl pyruvic acid, 296 Amifostine, 251 Amikacin, 450, 467 3-Aminobactamic acid, 466 p-Aminobenzenesulphonamide, 237 p-Aminobenzoate, 274, 445 γ-Aminobutyric acid, 325 7-Aminocephalosporanic acid, 153, 437 Aminocyclitol antibiotics, 467 L-1-Aminoethylphosphoric acid, 454 Aminoglutethimide, 17, 20, 211, 272, 282, 322, 370 Aminoglycoside antibiotics, 467 modifying enzymes, 449 6-aminopenicillanic acid, 155, 437 Aminopterin, 344, 503 5-Aminosalicylic acid, 244–5 Amorolfine, 480 Amoxycillin, 8, 34, 455, 462 Amoxyclav, 462 Amphetamine, 131 Amphoteracin (B), 477, 480, 483 Ampicillin, 6, 8, 34, 109, 246, 453, 455 Amsacrine, 362 Amthamine, 415 Analgesic receptor, 169 Anandron, 373 Anastrozole, 209 Androgens, 371–72 Androstenedione, 212, 370 Angiogenin, 384 Angiotensins I and II, 300 Angiotensin-converting enzyme, 249, 267, 276, 300 ANP, 304 Anthracenes, 360 Anthracyclines, 359 Anthramycin, 357 Anti-androgens, 372 Antibacterial agents, 467 Antibody-directed enzyme prodrug therapy, 256, 379

Index

Antibody drug conjugates, 379 Anti-emetics, 385, 407 Antifolates, 344, 445 Antifungal agents, 476 Antigene agents, 381 Antigens, 500–1 Anti-hormonal agents, 369 Antimetabolites, 344 purines, 346 pyrimidines, 347 Antimicrobial agents, 434–90 Anti-oestrogens, 370 Antisense oligonucleotides, 382 Antiserum, 500 Antitubilin agents, 364 Antiviral agents, 483–4 6-APA, 153, 437, 454 Apomorphine, 423, 425, 429 (S)-apomorphine, 138 APPA, 295, 297 Ara-A, 484, 488 Ara-C, 347, 484 Arachidonate, 390 Argatroban, 295–7 Arimidex, 208, 322, 371 Aromatase, 208–9, 274, 282, 320 inhibition, 208, 320, 370 Aromatic amino acid decarboxylase, 267, 273, 327 Arpromidine, 415, 419 Arsphenamine, 237 Asparaginase, 374 Aspartate proteases, 86, 276, 309–17 Aspirin, 9, 20, 33, 239 Astemizole, 414 Atamestane, 220 ATP synthase, 90 Atracurium besylate, 100, 164 Atrial natriuretic peptide, 304 Atropoisomerism, 101 Augmentin, 462 8-Azaguanine, 346 Azaserine, 274 Azathioprine, 368 Azidothymidine, 247 Aziridines, 352 Aziridinium ion, 349 Azithromycin, 471 Azlocillin, 456 Azothioprine, 17 AZQ, 352 AZT, 247, 486, 489 Aztreonam, 466

656

Index

Bacampicillin, 456 Baccatin, 367 Bactereophage, 492 Bacterial resistance, 447 BAL, 532 Bambuterol, 42 B cells, 500 BCNU, 342, 354 Becampicillin, 246 Benoxaprofen, 153 Benzeprilat, 303 Benzpyrinium, 274, 317 Benzyl penicillin, 6, 436, 441, 454 D-Benzylsuccinic acid, 301 Betahistine, 411, 419 Betamethasone, 242 Bioavailability, 6, 30–1, 34 Bile, 6, 20, 25 Bioequivalence, 50 Bio-inorganic chemistry, 509–37 Bioreductive prodrugs, 377 Bisantrene, 360 Bismuth, 522–3 Bitolterol, 258 Bizelesin, 357 Bleomycin, 342–3, 363 Blood-brain barrier, 11 B lymphocytes, 500 BMY 53857, 396 Boronic acid inhibitors, 288 Bovine pancreatic trypsin inhibitor, 68, 83 Bradykinin, 304 BRCA (1,2), 380 Breast cancer, 208, 272, 320, 367, 369, 371 British Anti-Lewisite, 532 BRL 43694, 403 Bromocryptine, 426 (S)-α-Bromoisovalerylurea, 131 6-Bromopenicillanic acid, 464 2-(3-Bromophenyl) histamine, 412 (R)-and (S)-Brompheniramine, 149 Brookhaven Protein Data Bank, 68 Bryostatin, 383 Buccal absorption, 47 Bupivacaine, 124 Burimamide, 419 Buserelin, 372 Buspirone, 396 Busulphan, 17, 352 Butaclamol, 431

657

Index

(+)-(3S, 4aS, 13bS)-Butaclamol, 119 BVaraU, 486 BW A515U, 246 BZQ, 352 Caeruloplasmin, 521, 526 Cahn-Ingold-Prelog convention, 105 Calmodulin, 93 c-AMP, 390–1 Camptothecin, 362 Cancer-related genes, 380 Cancer chemotherapy, 331 Candoxatrilat, 307 Captopril, 276, 301 Carbamates, 317 Carbacephems, 466 Carbamyl enzyme, 317 0-Carbamyl-D-serine, 274 1-Carbapenems, 463 Carbenicillin, 153, 455 Carbidopa, 327 Carbinolamines, 356 (S)-Carbinoxamine, 147 Carbonic anhydrase II, 267 Carbon tetrachloride, 15 Carboplatin, 356, 524 5-Carboxamido-tryptamine, 395 Carboxypeptidase A, 300, 304, 520 Carfecillin, 456 Carindacillin, 456 Carmustine, 354 Carprofen, 153 5-CAT, 395, 400, 407 Carvone, 11 Carzelesin, 357 Casodex, 373 Catecholamine methyltransferase, 264 (+)-CC-1065, 357 CCNU, 342, 354 cDNA, 498 Cephacetrile, 459 Cefaclor, 459 Cefamandole, 458, 460 Cefazolin, 459 Cefdaloxime pentexil, 157 Cefepime, 460 Cefixime, 459 Cefpirome, 460 Cefsulodin, 460 Ceftizoxime, 460 Cefotaxime, 460

658

Index

659

Cefoxitin, 462, 464 Cefpodoxime, 459 Ceftazidime, 460 Ceftibuten, 459 Ceftriaxone, 460 Cefuroxime, 157, 458, 460 Cell cycle specific agents, 341 Cell growth cycle, 334 Cell wall-active compounds, 481 Cephalexin, 110, 155, 459 Cephaloridine, 459 Cephalothin, 459 Cephamycins, 274, 457 Cephalosporins, 274, 437, 449, 454, 457 Cephalosporin C, 438, 462 sulphone, 290 Cephapirin, 458 Cephradine, 459 Ceranapril, 303, 307 Cetirizine, 414 CFC’s, 40 CGS 16949A, 220 CGS 24592, 307 CGS 25462, 308 CGS 26303, 308 CGS 26393, 308 Charge transfer, 54 Charton’s steric constants, 181 ChAT, 319 Chelation, 517 Chiral inversion, 132 transformations, 128–9 Chloral hydrate, 238 Chlorambucil, 342, 348–9 Chloramphenicol, 102, 242, 437, 439, 448, 451, 472 stearate, 34 acetyl transferase, 472 inactivating enzymes, 451 Chlorgyline, 272 N-(2-Chloroethyl)-N’-cyclohexyl-N-nitrosourea, 354 Chlorofluorocarbons, 40 Chlorozotocin, 355 Chlorpheniramine, 412 (R)-, (S)-Chlorpheniramine, 149 Chlorpromazine, 17, 429 Chlortetracycline, 469 Chlorthiazide, 272 Cholesterol, 212 side chain cleavage enzyme, 211, 282 Cilastatin, 465 Cilazaprilat, 276, 303 Cilofungin, 481

Index

Cimetidine, 20, 127, 416 Cinoxacin, 494 Ciprofloxacin, 443–4, 474 Cisatracurium besylate, 164 Cisplatin, 340, 342, 356, 523 Clarithromycin, 471 Clavulanic acid, 264, 274, 290, 457, 462–3 (R,R)-Clemastine, 150 Clindamycin, 242 Clobenpropit, 419 Clomocycline, 469 Cloning vector, 492, 494 Clotrimazole, 478 Cloxacillin, 8, 455, 462 Clozapine, 429, 431 CMC, 34 CMV, 485 Coated tablets, 38 Cobalin-H, 523 Cobalt, 514 Coformycin, 268 Colloidal carriers, 42 Combination chemotherapy, 343 COMFA, 184 Complexation, 34 COMT, 16, 264 Conformation, 61, 66, 69 Controlled release dosage forms, 44 Coordination complexes, 516 number, 517 Copper, 514 complexes, 522, 525–6 Correlation coefficient, 176, 189 matrix, 200 Cortisone, 209 Co-trimoxazole, 263, 474 CP-93, 131, 398 Critical micelle concentration, 34 (−)-(3S, 4R)-Cromakalim, 161 Cross-linking agents, 348 Cyanocobalamin, 523 Cycloguanil, 238 p-Cyclohexylaniline, 219 Cyclooxygenase (I, II), 267 Cyclophosphamide, 101, 250, 342, 350 Cyclopropanes, 357 D-Cycloserine, 274, 438–40 Cyclosporin, 369 Cyproterone acetate, 372 Cytaribine, 347, 484 Cytochrome P450, 13, 19, 127, 320, 370 reductase, 13, 209, 212

660

Index

Cytoplasmic membrane target, 476 Cytosine arabinoside, 341, 484 Cytosine deaminase, 256–7 DAAP, 253 Dacarbazine, 352 Dactinomycin, 361 DAPA, 295 Dapsone, 19 ddC, 486 ddI, 486 Daunomycin, 359 Daunorubicin, 342–3, 347, 359 D-Cycloserine, 274 Debrisoquine, 9 Deferiprone, 531 Dehydropeptidase 1, 465 Dimethyltetracycline, 469 De-Nol, 523 6-Deoxyacyclovir, 246 (−)-Deprenyl, 272, 274, 283 Dermal absorption, 528 DES, 371 Desferal, 530 Desferrioxamine, 530 Dexamethasone, 242, 244 β-D-glucoside, 244 (+)-(S)-Dexfenfluramine, 164 Dextropropoxyphene, 117, 136 DHFR, 263, 274, 344, 445–6 inhibitors, 344, 473 DHT, 371 Diacylglycerol, 390–1 Diaminopyrimidines, 473 cis-Diaminedichloroplatinum, 356 Diastereoisomers, 99 Diaveridine, 474 Diazepam, 9, 12 Diazoxide, 19 Dichloralphenazone, 20 Dichlorodiammine, 523 cis-Dichlorodiammineplatinium (II), 523 2’,3’-Didehydro-3’-deoxythymidine, 247 Diethyldithiocarbamate, 15 Diethylstilboestrol, 210, 371 α-Difluoromethyldopa, 328 α-Difluoromethylornithine, 274 Dihydrofolate reductase, 263, 274, 344, 445 inhibitors, 344 Dihydrofolic acid, 344 Dihydropteroate synthetase, 263, 274, 445

661

Index

1,4-Dihydropyridines, 138 Dihydrotestosterone, 371–2 Dihydroxy amino tetralins, 423 Dimaprit, 415 Dimercaprol, 532 Dimercaptopropanol, 532 Dimethylbenzanthracene, 213 (αR, βS)-αβ-Dimethylhistamine, 418 Diphenhydramine, 149, 412 Dipivefrin, 250 Dipole-dipole interactions, 54 Dirithromycin, 471 Disopyramide, 142 DMBA, 213 DNA, 68, 492, 524 cross-linking, 342 cleaving agents, 363 groove binding, 342 interactive drugs, 342, 348 intercalation, 342, 359 sequence code, 493 repair inhibitors, 384 DNA gyrase, 159, 267, 444 DNA ligase, 494 DNA polymerase, 246, 488, 493 DNA, RNA polymerases, 267 DNA topoisomerases, 483 DNA viruses, 337 Dissociation constants, 3–4 Dissolution, 32 Distomer, 116 Distribution, 9, 124 Disulphiram, 15 DOB, 400, 407 DOI, 400 DOM, 407 L-Dopa, 123, 161, 273 Dopamine, 115, 253, 273, 389, 422, 429 agonists, 423 antagonists, 429 receptors, 422, 432 Dosage forms, 29, 35 Doxorubicin, 256, 343, 359 Doxycycline, 469 DPIs, 41 Drug absorption, 2, 245 carriers, 47 chirality, 95 delivery, 28 handling, 1 stability, 34 targetting, 377

662

Index

Drug powder inhalers, 41 D4T, 247, 486 DTIC, 352, 354 DTPA, 532 DuP 714, 293 Dyflos, 274, 317 E-2020, 319 Echinocandins, 481 Econazole, 322, 478 Ecothiopate, 274, 317 EDTA, 530–1, 534 Efegatran, 295 Eflornithine, 329 Elastase, 285 inhibitors, 285 Electron donor-acceptor interaction, 63 Electronic parameters, 173 Electrostatic interactions, 53, Elimination half-life, 22 Ellipticene, 362 Emate, 322 Emulsions, 36 Enalapril, 249, 278 Enalaprilat, 249, 278, 302 Enantiomers, 98–9 Endonucleases, 494 Enediynes, 364 Enkephalinase A, 278 Enkephalins, 304, 389 Enoxacin, 443 Enteric coating, 38 Enterohepatic cycling, 25 Entropy, 56 Enzyme Inhibitors (see Inhibitors of-) active site directed, 268 affinity labelling, 268 competitive, 265 design, 271, 275 irreversible, 268 mechanism-based, 268 metabolism, 279 non-competitive, 266 oral absorption, 278 reversible, 267 slow, tight binding, 268 stereoselectivity, 281 suicide substrates, 268 toxicity, 281 transition state analogues, 266 types, 264

663

Index

Epirubicin, 359 Epitetracycline, 469 Epitope, 500–1, 504 Epoxide hydrases, 15, 20 α-Ergocriptine, 426 Ergolines, 426 Ergopeptines, 426 Ergosterol, 477 Ergot alkaloids, 425 Ergotamine, 423, 426 Erythromycin, 242–3, 439, 448, 453, 471, 476 estolate, 243 Erythropoietin, 374 Es, 179 Esmolol, 128 Eserine, 274 Estramustine, 378 Ethambutol, 528, 533 Ethyl dithiolisophthalate, 242 Ethylenediaminetetracetate, 531 Ethyleneiminium ion, 349 Etoposide, 366 Eudismic Index, 116 Eudismic Ratio, 116 Eutomer, 116 Excretion, 20, 133 Exemestane, 322, 371 f, 171 Fadrazole, 322 Fab fragments, 501, 507 Famciclovir, 247, 279, 487 Famotidine, 416 Fc fragments, 501, 507 Fenoldopam, 427 Fenoprofen, 151, 153 Fentanyl, 7 Fenvalerate, 160 Ferguson effect, 186 Ferrocontin continus, 523 Fialuridine, 486 FIAU, 486 Filipen, 477 First pass metabolism, 6, 18, 47 Fischer plane projection, 105 Flamazine, 523 Flavin-containing monooxygenase, 127 Flecainide, 136 Flucloxacillin, 8, 455 Fluconazole, 280, 478 Flucytosine, 480

664

Index

Flumequine, 443 (R)- and (S)-Flumequine, 158 (S)-Flunoxaprofen, 151 Fluoro-2’,5’-anhydrocytosine arabinoside, 347 5-Fluorocytosine, 258 E-2-(Fluoromethyl) dehydroornithine, 329 (S)-α-Fluoromethylhistidine, 411 N-Fluorophenyl carbamate, 420 Fluoroquinolones, 474 5-Fluorouracil, 256, 263, 342, 347, 480, 486 Flupenthixol, 258, 429 Fluphenazine, 429 Flurbiprofen, 153 FMO, 127 Folate inhibitors, 473 Folinic acid, 345 Follicle stimulating hormone, 372 Fonofos, 111 Fonofos-oxon, 111 Formestane, 213 Formoterol, 118 Formulation, 31, 35 Foscarnet, 487–8 Fosinopril, 307 Fosinoprilat, 303 Free energies, interaction, 56, 77 Free-Wilson analysis, 192 FSH, 372 5-FU, 256, 263, 342, 347, 480, 486 Fungal cell wall target, 476 Fusidic acid, 448, 453 F values, 176, 190 GABA, 13, 325, 389 receptor, 390 Gabaculin, 325 GABA-T, 274, 283, 325 inhibitors, 325 GABA transaminase, 274, 283, 325 inhibitors, 325 Ganciclovir, 485, 488 Gastrointestinal transit, 47 GCPRs, 59, 89 GDEPT, 380 GDP, 390 Gene-directed enzyme prodrug therapy, 380 Gene targetting, 381 Gene therapy, 382, 499 Gentamicin, 467 Gepirone, 396 Glaucoma, 317

665

Index

666

Glomerular filtration, 23 Glucuronide, 16 Glutathione, 17, 130 Glutethimide, 20 Glyceraldehyde, 106 Glyceryl trinitrate, 6, 7, 46 Glycoprilat, 309 Glycylcyclines, 470 Glypressin, 258 Gold thiomalate, 523 Goserelin, 372 Gossypol, 102 G-proteins, 390, 393 GR 113808, 405 Granisetron, 385, 463 Griseofulvin, 20, 32, 479 Growth factors, 377, 383 GSH, 130 GTP, 310 Guanine nucleotide-coupled receptor proteins , 59, 89 H-261, 311 Haemoglobin, 521 Haloperidol, 317, 429, 431 Halothane, 15, 25 Haloxyfop, 132 Hammett substituent constants, 173 Hansch analysis, 187 Hansch substituent constants, 170, 173 Hard acid, 518–9 Hard base, 518–9 Hard gelatin capsule, 37 Henderson-Hasselbach equation, 3 Heparin, 9 Hepatitis B, 486 Hetacillin, 250 Herpes, 246, 484, 487 Hexamine, 239 hGH, 87, 498 HGPRT, 503 Hin recombinase, 98 Hippurates, 17 Hirudin, 68, 84 Histamine, 410, 421 receptors, 410, 421 Histidine decarboxylase, 411 HIV, 247, 486 protease, 267, 276, 309 reverse transcriptase, 248, 267, 486, 489 HMG-Co A reductase, 267 HSA, 124

Index

667

HSAB concept, 517, 519 5-HT, 39, 389 agonists, 391–408 antagonists, 391–408 receptor, 392 5-HT3 receptor, 390 antagonists, 342, 385 5-HTQ, 403 Human cytomegalo virus, 485 Human growth hormone, 87, 498 Human immunodeficiency virus, 247, 486 Human serum albumin, 124 Hybridoma cells, 502 Hydrallazine, 19 Hydrocortisone, 254 Hydrogen bond, 54 parameters, 178 Hydrophobic effect, 56 fragmental constants, 171 parameters, 169 substituent constants, 170 5-Hydroxyaminotetralins, 424 4-Hydroxyandrostenedione, 8, 213, 219, 322, 371, 424 Hydroxyflutamide, 372 (R)- and (S)-3-(3-Hydroxyphenyl)-Npropylpiperidine, 137 (S)-4-Hydroxyphenytoin, 127 8-Hydroxyquinoline, 528 Hydroxytamoxifen, 370 5-Hydroxytryptamine, 389, 391 Hydroxyurea, 367 Hyoscine hydrobromide, 7 Hyperconjugation, 63 Hypoxic tissues, 255, 341 Hypothalamus/pituitary axis, 372, 374 Ibuprofen, 104, 132, 153 (S)-Ibuprofen, 151 Ibuterol, 42 IC50, 265 ICI 164384, 320, 370, ICI 182780, 217, 320, 370 ICI 176334, 373 ICI 200880, 287 ICI 205930, 403 ICI 207658, 208 Idarubicin, 359 Idoxifene, 370 Idoxuridine, 484 IFN, 375–6 Ifosfamide, 350

Index

668

IgG, 501 Imetit, 419 Imidazole antifungals, 478 Imidazotetrazinones, 354 Imipenem, 464–5 Imipramine, 6, 9 Immepyr, 418 Immune system, 500 Immunosuppressive agents, 368 Impromidine, 415, 419 (R)- and (S)-Indacrinone, 139 Indoprofen, 153 Induction, 20 Inductive substituent constants, 177 Inhibition (metabolism), 19 Inhibitors of: AADC, 267, 273, 327 ACE, 249, 267, 276 Acetylcholinesterase, 262, 274, 317–20 Alanine racemase, 274 Angiotensin-converting enzyme, 249, 267, 276 Aromatase, 208, 274, 282, 320 Aromatic amino acid decarboxylase, 267, 273, 327 Aspartate proteases, 276, 309–17 Aspartate transcarbamylase, 267 Carbonic anhydrase II, 267 Carboxypeptidase A, 300, 304 Catecholamine methyltransferase, 264 Cholesterol side chain cleavage enzyme, 211, 282 Cyclooxygenase I and II, 267 Dehydropeptidase 1, 465 DHFR, 263, 344, 445 Dihydrofolate reductase, 263, 344, 445 Dihydropteroate synthetase, 263, 445 DNA gyrase, 267, 444 DNA, RNA polymerases, 267 Elastase, 285 Enkephalinase A, 278 GABA-T, 325 GABA transaminase, 274, 283 Histidine decarboxylase, 411 HIV protease, 267, 276, 309 HIV reverse transcriptase, 248, 267 HMG-CoA reductase, 267 Δ8–7 Isomerase, 481 β-Lactamase, 264, 274, 290 Lanosterol-14-methyl demethylase, 214, 220, 322, 479 MAO, 272, 283 MAO-B, 274, 283 Membrane metalloendopeptidase, 278, 304–9 MEP, 278, 304–9 Metalloproteases, 299–304

Index

Monoamine oxidase, 272 Na+K+-ATP’ase, 267, 274 ODC, 274, 329 Oestrogen sulphatase, 322 L-Ornithine decarboxylase, 274, 329 P450arom, 208, 274, 282, 320 P450scc, 211, 282 Prostaglandin synthetase, 267 Protease inhibitors, 283–99 Pyridoxal phosphate-dependent enzymes, 322 Pyruvate dehydrogenase, 274 ras-farnesyl transferase, 383 5α-Reductase, 267 Δ14 Reductase, 481 Ribonucleotide reductase, 367 Riboxyl amidotransferase, 267 Serine proteases, 284 Squalene epoxidase, 267, 480 Succinic semi-aldehyde dehydrogenase, 267 Telomerase, 384 Thrombin, 292 Thymidine kinase, 267 Thymidylate kinase, 267 Thymidylate synthetase, 263, 274, 480 Topoisomerases, 267, 444, 483 Transpeptidase, 274 Viral DNA polymerase, 267 Xanthine oxidase, 262 Injections, 8 Inogatran, 295 Inositol phosphates, 390–1 Insulin, 497 Interferons, 375, 487 Interleukin, 376 Intermolecular forces, 51, 79 Intramolecular forces, 61 Introns, 493 Iodoaminopotentidine, 416 Iodoproxyfan, 420 Ionophoric drugs, 439, 442 Ionotropic receptors, 390 Ion channels, 391 Ion-pairs, 34 Iproniazid, 272, 274 Ipsapirone, 396 Irreversible inhibitors, 268 Isocarboxazid, 274 Δ8–7-Isomerase, 481 Isoniazid, 16–7, 19, 528 Isothipendyl, 136 Itraconazole, 478 IUD, 484

669

Index

JG 365, 312 JM 118, 525 JM 216, 356, 525 Kadin, 523 KAN 400473, 291 Kanamycin, 450, 467 Kappa index, 183 Kcat inhibition, 268 Kelatorphan, 308 Ketamine, 140 Ketanserin, 400, 408 Ketoconazole, 280, 322, 418 Ketoprofen, 133 Ketotifen, 413 Ki, 265 L1, 531 L 735524, 312 Labetalol, 146 β-Lactam antibiotics, 454–67 β-Lactamase, 264, 274, 290, 447, 450 Lanosterol 14α-demethylase, 214, 220, 322, 479 Latamoxef, 462, 465 Lentaron, 213 Lergotrile, 426 Letrozole, 322, 371 Leucovorin, 345 Leuprolide, 372 Levocabastine, 414 Levopropoxyphene, 117, 136 Lewis acid (base), 55 LH, 371 LHRH, 371 Ligand binding, 69, 516 Lignocaine, 6 Limonene, 113 Lineweaver-Burk plot, 266 Liposomes, 42 Lisinopril, 303 Lisuride, 426 Log P, 170 Lomefloxacin, 443, 474 Lomustine, 354 Loracarbef, 459 Loratidine, 414 Lung delivery, 38 Luteinising hormone, 371 Luteining-hormone releasing hormone, 371 LY 277359, 403

670

Index

Lymphokines, 499 D-Lysergic acid, 426 Macrolide antibiotics, 471 MAO, 272, 283 MAO-A, 283 MAO-B, 274, 283 Marimastat, 340, 342, 384 mCPBG, 403 mCPP, 401 MDI, 39–40 MDL 72222, 403 MDL 100907, 401 MDL 101146, 288 MDR, 343, 383 Mebendazole, 478 MEC, 30, 45 Mechanism-based inactivators, 268 Mechlorethamine, 348, 350 Mecillinam, 456 Medroxyprogesterone, 370 Melphalan, 342, 348, 350 Membrane-active compounds, 480 Membrane metalloendopeptidase, 304–9 MEP, 278, 304–9 Me-D-Phe-Pro-Arg-H, 295 Mephenytoin, 19 Meprobamate, 16, 20 Mepyramine, 412 2-Mercaptobenzothiazole, 16 6-Mercaptopurine, 341–3, 346, 368 Meropenem, 466 Mesalazine, 244 Mesna, 252, 350 Metabolism, 12, 126 Metabotropic receptors, 380 Metal associated disorders, 515 complex stability, 517 coordination chemistry, 516 ligand specificity, 517 Metallo-proteases, 299–304 Metastases, 334 Metered-dose inhaler, 39–40 (S)-β-Methacholine, 121 Methadone, 168 Methaqualone, 101 Methacycline, 469 Methane sulphonates, 352 Methicillin, 453–4, 462 Methiothepin, 400 Methisazone, 484, 529

671

Index

Methotrexate, 123, 268, 341–4, 474 7α-Memoxycepham, 290 Methoxyflurane, 15 8-Methoxypsoralen, 378 Methsuximide, 238 α-Methyldopa, 116, 327 (+)-(S)-α-Methyldopamine, 116 (R)- and (S)-Methylflumequine, 158 α-Methylhistamine, 417 (R) and (S)-α-Methylhistamine, 120, 418 O-Methyl 5-HT, 400 2-Methyl 5-HT, 403 (−)-(1R, 2S)-α-Methylnoradrenaline, 116 Methylphenobarbitone, 238 α-Methylserotonin, 400 Metoclopramide, 385 Metronidazole, 242, 478 Mezlocillin, 456 Mianserin, 407 Microsomal oxidations, 14 Miocomycin, 472 Miconazole, 478 Microtubule assembly, 366 Midecamycin, 472 Migraine, 407 Minimal steric difference, 183 Minocycline, 469 Misonazole, 255 Mithramycin, 368 Mitomycin, 377, 439 Mitotane, 368 Mitroxantrone, 342, 360 Mitozolomide, 354 Mixed function oxidases, 13 Molecular connectivities, 185 interactions, 53 modelling, 51, 68, 275 refractivity, 182 similarity, 184 volume, 183 Monoamine oxidase, 272 Monobactams, 466 Monoclonal antibodies, 491–507 α-Monofluoromethyldopa, 273, 278, 328 Monokines, 499 Monphytol, 523 Morphine, 9 Morpholine antifungals, 480 mRNA, 493 MSC30, 45 Mucus, 40 Multidrug resistance, 383

672

Index

Multiple fragment probes, 74 Multivariate analysis, 198 Mustine, 348, 350 Mycomycin, 168 Mycocrisin, 523 Myoglobin, 521 Myeloma cells, 502 Na+, K+-ATP’ase, 267, 274 Nalidixic acid, 443–4, 474 NAN-190, 396 NAPAP, 297 Naproxen, 118, 153 (S)-Naproxen, 151 Nebulizers, 41 Neopolyoxins, 48 Neostigmine, 260, 317 Netilmicin, 467 Neural networks, 204 Neuropeptides, 389 Neuronal signal effects, 380 Neurotransmitters, 387 Nicotine, 7 Nicotinic receptor, 7 Nikkomycins, 481 Nilutamide, 374 Nilvadipine, 132 Nitracine, 255 Nitrogen mustards, 340, 348 Nitrosourea, 354 NMDA, 390 NMR spectroscopy, 74, 77 Nocarcidins, 466 Nolvadex, 210 Non-cell-cycle-specific agents, 342 Noradrenaline, 108, 389 biosynthesis, 273, 327 (R)-Noradrenaline, 115 Norethindrone, 210 Norethisterone, 210 Norfloxacin, 443–4, 474 Nortryptyline, 9 NOT, 398 Novobiocin, 20, 33, 439, 448 Noyes-Whitney equation, 32 NSAIDs, 110, 150 Nucleoside analogues, 484 Nystatin, 476 Octreotide, 377 ODC, 274, 329

673

Index

inhibitors, 329 Oestradiol, 9, 212 Oestrogen, 320, 322, 368 receptors, 320 sulphatase, 322 therapy, 37 Oestrone, 212 Ofloxacin, 443–4 (R)- and (S)-Ofloxacin, 158 4-OHA, 213, 322, 371 7-OH DPAT, 429 8-OH DPAT, 395, 407–8 Oligonucleotides, 381 Olivanic acids, 463 OM 805, 295 Omeprazole, 274 Ondansetron, 385, 403 Optimal partition coefficient, 191 Oral dosing, 5, 36 Orbital interactions, 61 Organoarsenicals, 274 Organophosphorus compounds, 318 Ormetroprim, 474 L-Ornithine decarboxylase, 274 inhibitors, 329 Osalazine, 244 Ovarian cancer, 367 Oxacephems, 465 Oxapenicillins, 457 Oxazepam, 12, 125, 130 Oxine, 528 Oxophenarsine, 237 Oxyphenbutazone, 239 Oxytetracycline, 469 P, 169 P-170, 141 P450, 208–9, 274, 282, 320 P450arom, 216 P450scc, 261, 282, 320, 322 Paclitaxel, 366 Paludrine, 237 Pamaquin, 237 PAPS, 16 Paracetamol, 18, 239 Paraquat, 42 Parkinson’s disease, 327 Particle size, 39 Partition coefficient, 169 Partitioning, 11 Passive reabsorption, 24

674

Index

Pattern recognition methods, 203 PBDs, 357 PBPs, 440 Penbutolol, 143 Penems, 465 Penciclovir, 247, 279, 487 D-Pencillamine, 161, 527, 530, 532 L-Penicillamine, 123 Penicillanic acid sulphone, 464 Penicillin G, 23, 34, 454 Penicillin V, 454 Penicillins, 246, 274, 440 Penicillin binding protiens, 440 Pentostam, 523 Peperacillin, 456 Pepleomycin, 364 Pepsin, 86 Pepstatin, 86, 310 Peptic ulcer, 522 Peptide chloromethyl ketones, 293 Peptidyl aldehydes, 281 Pergolide, 426 Perindoprilat, 303 Perphenazine, 258 Pethidine, 20 Pharmacodynamics, 135 Pharmacokinetics, 1, 134 Phase I and II, 12–3, 16 (R) and (S)-Pheniramine, 149 Phenacetin, 239 Phenobarbitone, 19–20 Phenelzine, 19, 272, 274 Phenothiazine, 429 Phenoxazines, 361 Phenoxymethyl pencillin, 454 (R)- and (S)-Phenprocoumon, 148 Phenylalanine mustard, 348, 350 Phenylbutazone, 20, 239, 253 Phenytoin, 15–7, 20, 37, 127, 243 D-Phe-Pro-Arg-H, 293 D-Phe-Pro-Arg-CH2Cl, 293 PHNO, 427–8 Phospholipase C, 391 Phosphonoformate, 487 Phosphoramidon, 307 Phosphoryl enzyme, 318 Phosphotransferases, 467 Photoactivated prodrugs, 378 Photodynamic therapy, 378 Picenadol, 137 Pilocarpine, 249 Pindolol, 396

675

Index

Pipemidic acid, 444 Pipothiazine, 258 Pipomidic acid, 444 Pirprofen, 153 Pirampicillin, 246, 456 Placental barrier, 11 Plasma albumin, 10 Plasma binding, 10 Plasmids, 453, 492 Plomestane, 322, 371 Podophyllotoxin, 366 Polyene antibiotics, 439, 442, 472 Polymixins, 439, 442 Polymorphism, 33 Polyoxins, 481 PPACK, 293 3-PPP, 137, 427 Physostigmine, 317 Practolol, 162 Pralidoxime, 253, 319 Prednisolone, 242 Prenalterol, 89 Presystemic metabolism, 18 Primidone, 238 Procainamide, 15 Procarbazine, 252, 358 Prochirality, 112 Prochiral to chiral, 127 Prodrugs, 42, 236, 377 Promethazine, 136, 413 Prontosil, 237 Propanolol, 17, 59, 144–5, 396 Propantheline, 6 Prostaglandin synthetase, 267 Prostatic cancer, 371–4 Protease inhibitors, 283–99 Protein binding, 124 double strand DNA recognition, 98 kinase C, 383, 391 ligand binding, 77 protein recognition, 87 single strand DNA recognition, 94 structure, 78 α1-Proteinase inhibitor, 285 Pulmonary absorption, 38 Puromycin, 439 Putrescine, 329 Pyridine-2-aldoxime mesylate, 319 Pyridoxal phosphate-dependent enzymes, 322 Pyridoglutethimide, 282 Pyrilamine, 412 Pyrimethamine, 237, 446, 474

676

Index

Pyrrolo [2,1] [1,4-c] benzodiazepines, 357 Pyruvate dehydrogenase, 274 Quantitative structure-activity relationships, 168 QSAR, 116, 120, 167 Quinaprilat, 302 Quinidine, 133 Quinolones, 158, 439, 443, 453, 474 Quinpirole, 423, 427–9 r, 176 R 76713, 234 Racemates, 98 Raclopride, 429 Ramiprilat, 302 Ranitidine, 416 ras farnesyltransferase, 383 Receptors, 390 Recombinant DNA technology, 491–500 Rectal dosing, 7 Reductases, 15 5α-Reductase, 267, 371 Δ14-Reductase, 481 Regression coefficient, 189 Remoxipride, 429 Renal clearance, 21 elimination, 21 Renzapride, 403 Resistance mechanisms, 488 Restriction enzymes, 494 Reverse transcriptase, 248 Reversible inhibitors, 264 Rheumatoid arthritis, 525 Ribavirin, 484 Ribonucleotide reductase, 367 Riboxyl amidotransferase, 267 Ribozymes, 382 Rifampicin, 20, 439, 444–5, 453 RNA polymerase, 493 RNA retroviruses, 337 Ro 46–6240, 298 Rogletimide, 282 Routes of administration, 35 Roxatidine, 416 Roxithromycin, 471 RS-23597–190, 405 S-2581, 299 (R)-Salbutamol, 164 Salicylic acid, 16 Saliva elimination, 25

677

Index

Saquinavir, 276, 279, 312 Sarin, 319 SB 200646A, 401 SB 204070, 405 SB 204741, 402 SB 206553, 402 SB 207710, 405 SC 52151, 312 SC 53116, 405 SCH 23390, 423, 429, 431 SCH 32615, 278, 306 SCH 34826, 278, 308 SCH 39370, 306 SDAT, 317, 319 SDZ 205557, 405 SDZSER-082, 402 Selective toxicity, 341 antivirals, 484 Selegiline, 272, 274, 283 Selsun, 523 Sequene Rule system, 107 Serazide, 327 Serine proteases, 83, 284 Serotonin, 389 receptors, 391 Serum albumin, 521, 526 Shingles, 485 Single fragment probes, 76 Sinorphan, 308 Sisomicin, 467 Site specific, carrier, 47 drug delivery, 47, 253 SKF 38393, 423, 428–9 Soft acid, 518–9 base, 518–9 Soft gelatin capsules, 37 Soft quaternary salt, 249 Somatostatin, 492, 496 Sorivudine, 486 Sotalol, 146 Sparfloxacin, 443, 474 Sparteine, 19 Spermidine, 329 Spermine, 42, 329 Spiperone, 401 Spiraprilat, 302 SQ 20881, 301 Squalene epoxidase, 267, 480 Stereochemistry, 95, 99 Stereoisomerism, 18 Steric parameters, 179

678

Index

Sterimol parameters, 182 Steroidogenesis pathway, 321 Streptomycin, 437, 439, 448, 453, 467–8 Succinic semi-aldehyde dehydrogenase, 267 Suicide substrates, 268 Sulbactam, 274, 464 Sulindac, 101, 154, 239 Sulphadiazine, 523 Sulphamethoxazole, 9, 263, 274, 446 Sulphanilamide, 272 Sulphasalazine, 244 Suphatase, 322 Sulphathiazole, 17 Sulphonamides, 274, 439, 445, 448, 453 (−)-Sulpiride, 423, 429, 431 Sumatriptan, 398 Suramin, 377 Suspensions, 37 Sustained release dosage forms, 44 Suxamethonium, 19 Tablets, 38 Tabun, 319 Tacrine, 319 Tafts substituent constants, 178 Tafts steric substituent constants, 179 TAK-147, 319 Talampicillin, 246, 456 Tallimustine, 343, 352 Tamoxifen, 20, 210, 320, 342–3, 352 3-TAPAP, 296, 298 4-TAPAP, 291 Taxanes, 366 Taxol, 257, 340, 342, 366 Tazobactam, 464 3 TC, 486 T cells, 500 TCR, 45 Teicoplanin, 441 Telomerase, 384 inhibitors, 384 Temocillin, 455 Temafloxacin, 424, 443 Termazepam, 12 Temozolamide, 340, 343, 354 Terbinafine, 480 Terbutaline, 42, 124 Terfenadine, 121, 414 Terodiline, 162 Testololactone, 213 Testosterone, 370, 372

679

Index

Tetracyclines, 37, 437, 439, 448, 453, 469, 529 Tetrahydrofolic acid, 344 Tetroxoprim, 446, 473 Thalidomide, 532 Thallassaemia, 530 Therapeutic concentration ratio, 45 Thermolysin, 304, 307 Theophylline, 19–20 Thiacetazone, 528 2-(2-Thiazolyl)ethanamine, 411 Thicarbamates, 480 Thienamycin, 463, 465 6-Thioguanine, 342, 346–7 Thiopentone, 11, 20 Thioperamide, 419 Thioridazine, 429 Thiorphan, 278, 305–6 Thiotepa, 352 α-Thrombin, 68, 84, 292 inhibitors, 292 Thymidine kinase, 267, 485, 488 Thymidylate kinase, 267 Thymidylate synthetase, 263, 274, 480 Ticarcillin, 455, 462 Timolol, 143 Tissue distribution (chiral compounds), 126 T lymphocytes, 500 TNF, 376 Tobramycin, 467 (R)- and (S)-Tocainide, 142 Tolbutamide, 19 Tolnaftate, 480 Topical application, 7 Topoisomerases, 159, 267, 359, 362, 444, 483 inhibition, 362 Topological parameters, 185 Toxicology, chiral compounds, 160 Trace elements, 512 supplementation, 534 Transfer of drugs, 3 Transition state analogues, 266 Transpeptidase, 274 Tranylcypromine, 272, 274 Trazodone, 407 TREN, 530 Triamterene, 474 Trichloroethanol, 15, 238 Triethylenetetramine, 530 Trimelamol, 358 Trimethoprim, 263, 439, 446, 448, 453, 473 Tropisetron, 403 Trypanosomiasis, 329

680

Index

Trypsin, 68, 83 Tumour formation mechanisms, 336 Tumour necrosis factor, 376 Tyrosine kinases, 383 U-75875, 311 U-77779, 358 U-80244, 358 U-81749, 311 UDP-Glucuronyl transferases, 19–20 UK-46245, 280 UK 69578, 307 UK 79300, 308 VD, 9 Vaccines, 385, 507 Valaciclovir, 246, 279, 487 Van der Waals dimensions, 181 interactions, 56 Vancomycin, 441 Varicella zoster, 485, 487 Vasopressin, 258 VDEPT, 257 (R)- and (S)-Verapamil, 140 Vidaribine, 484 Vigabatrin, 13, 274, 283, 325 Vinblastine, 340–2, 356, 364 Vinca alkaloids, 364 Vincristine, 341, 343, 361, 364 Vindesine, 364 α-Vinyl GABA, 13, 325 Viral DNA polymerase, 267 Viral reverse transcriptase, 486, 489 Virus-directed enzyme prodrug therapy, 257 Vitamin B12, 512 Vokitamycin, 472 Volume of distribution, 9 (+)-Vorozole, 322, 371 Warfarin, 10, 20, 129, 148 WAY-100135, 396 WIN 63759, 292 WIN 64733, 292 Xanthine oxidase, 262, 369 X-ray crystallography, 74, 77 Zacopride, 403 Zatosetron, 403 Zidovudine, 247, 486

681

Index

Zinc finger domain, 94 Zolantidine, 416

682