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

Citation preview

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 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 (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 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=OCH31 >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 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, Ki100 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, IC5090%) 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 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