Pharmacoakinetics in Early Drug Research

The Report and Recommendations of ECVAM Workshop 221,2

Reprinted with minor amendments from ATLA 25, 17-31

David E. Leahy,3 Ruth Duncan,4 Hans J. Ahr,5 Martin K. Bayliss,6 A. (Bert) G. de Boer,7 Ferenc Darvas,8 Julia H. Fentem,9 Jeffrey R. Fry,10 Robert Hopkins,11 J. Brian Houston,12 Johan Karlsson,13 Gregory L. Kedderis,14 Margaret K. Pratten,15 Pilar Prieto,9 Dennis A. Smith,16 and Donald W. Straughan17
3Lead Discovery Department, ZENECA Pharmaceuticals, Alderley Park, Macclesfield SK10 4TG, UK; 4Centre for Polymer Therapeutics, The School of Pharmacy, University of London, 29/39 Brunswick Square, London WC1N 1AX, UK; 5PH-PD-T Research Toxicology, Bayer AG, 42096 Wuppertal, Germany; 6Bioanalysis and Drug Metabolism, Glaxo Wellcome, Park Road, Ware, Herts SG12 0DP, UK; 7Division of Pharmacology, LACDR, Leiden University, Sylvius Laboratories, Wassenaarseweg 72, 2300 RA Leiden, The Netherlands; 8ComGenex Ltd, 1393 Budapest 62, Hungary; 9ECVAM, JRC Environment Institute, 21020 Ispra (VA), Italy; 10Department of Physiology & Pharmacology, Medical School, Queen's Medical Centre, Nottingham NG7 2UH, UK; 11Corning Hazleton, Otley Road, Harrogate, North Yorkshire HG3 1PY, UK; 12Department of Pharmacy, University of Manchester, Oxford Road, Manchester M13 9PL, UK; 13Elan Corporation Research Institute, Trinity College, Dublin 2, Ireland; 14CIIT, Research Triangle Park, NC 27709, USA; 15Department of Anatomy and Cell Biology, Medical School, Queen's Medical Centre, Nottingham NG7 2UH, UK; 16Department of Drug Metabolism, Pfizer Central Research, Sandwich CT13 9NJ, UK; 17FRAME, Russell & Burch House, 96-98 North Sherwood Street, Nottingham NG1 4EE, UK

1ECVAM - The European Centre for the Validation of Alternative Methods. 2This document represents the agreed report of the participants as individual scientists.

Address for correspondence: Dr D. Leahy, Lead Discovery Department, ZENECA Pharmaceuticals, Alderley Park, Macclesfield SK10 4TG, UK

Address for reprints: ECVAM, TP 580, JRC Environment Institute, 21020 Ispra (VA), Italy

Preface

This is the report of the twenty-second of a series of workshops organised by the European Centre for the Validation of Alternative Methods (ECVAM). ECVAM's main goal, as defined in 1993 by its Scientific Advisory Committee, is to promote the scientific and regulatory acceptance of alternative methods which are of importance to the biosciences and which reduce, refine or replace the use of laboratory animals. One of the first priorities set by ECVAM was the implementation of procedures which would enable it to become well-informed about the state-of-the-art of non-animal test development and validation, and the potential for the possible incorporation of alternative tests into regulatory procedures. It was decided that this would be best achieved by the organisation of ECVAM workshops on specific topics, at which small groups of invited experts would review the current status of various types of in vitro tests and their potential uses, and make recommendations about the best ways forward (1).

The workshop on Pharmacokinetics in Early Drug Research was held in Bath, UK, on 27-29 March 1996, under the co-chairmanship of David Leahy (ZENECA Pharmaceuticals, Macclesfield, UK) and Ruth Duncan (School of Pharmacy, London, UK). The aims of the workshop were to: a) review the current methodology for providing early pharmacokinetic and related information; b) define current best practice and identify those problem areas where scientific and technological advances could significantly improve the efficiency of drug discovery; and c) recommend further work which should be conducted in these high priority areas.

Introduction

The drug research and development process is evolving rapidly due to technological advances in the identification of biological targets, as well as in the automation of chemical synthesis and screening. Methodology for designing compounds with good activities in vivo, which is typically based on a combination of high throughput synthesis and screening, will need to be adapted in the light of these technological advances. One challenge that presents itself is how to improve our efficiency in the optimisation of activity in vivo, especially with respect to aspects of pharmacokinetics and safety assessments, without increasing our reliance upon animal testing.

The workshop was held to consider these changes and to develop recommendations for priority areas of research that could lead to a more effective use of animal studies while increasing the capacity for efficient optimisation of activity in vivo.

Lead Structure Identification

Drug discovery continues to undergo enormous changes, driven by technological innovation and the pressures of competition. Remarkable advances have been seen in the identification of new therapeutic targets, as a consequence of progress in molecular biology and of the Human Genome Project. New chemical lead identification is becoming much easier as a result of two developments: a) the revolution in synthetic chemistry due to the impact of robotics and combinational libraries; and b) the establishment of high throughput screening procedures. It is probable that these changes will continue to be of great significance, as the Human Genome Project continues to deliver massive amounts of information, which will be analysed by ever more sophisticated computer methods, and as the robotisation and miniaturisation of screening and synthesis become more commonplace.

High throughput primary screening can produce many thousands of compounds of different kinds which show the required effects at relatively high concentrations. The identification of lead structures is the process whereby these compounds are sifted (secondary screening) to provide congeneric series suitable for lead optimisation programmes. This requires a variety of rapid, relatively straightforward, relevant, and reliable systems which provide physicochemical and biological information.

Key physicochemical properties are solubility, stability, lipophilicity (as determined by measuring the octanol-water [P] or membrane-water partition coefficients [2]), and pKa (the negative log of the acid dissociation constant); these can be used to predict protein binding, tissue distribution, and gastrointestinal (GI) absorption (3).

There can be no standard approach to the use of biological models in the secondary screening stage, as this will depend upon the type of effect required, as well as on the number and variety of compounds identified in the primary screen. Estimations of the probability that compounds will reach adequate blood levels in vivo for them to be active can be obtained from physiologically-based pharmacokinetic (PBPK) models or from relatively simple in vivo experiments.

In vitro cytotoxicity and genotoxicity tests are suitable for use with large numbers of compounds, and the use of freshly isolated hepatocytes in suspension (4), or microsomes, can provide important information on metabolism. Where extrahepatic metabolism could be important, the incubation of compounds with blood or tissue homogenates may be appropriate.

At this stage, the various methods do not need to provide precise information on pharmacokinetic parameters or toxicological effects, but, ideally, should give relative values which can be used as a basis for lead identification or further structural searching.

Lead Optimisation

The aim of lead optimisation is to obtain good biological activity which is likely to translate to humans. Its success is increasingly likely to be based on the rational modelling of in vivo behaviour, as a result of the provision of information on metabolism, pharmacokinetics, and toxicology much earlier in drug discovery than at present.

In early lead optimisation the focus should be on characterising metabolism and pharmacokinetics. The number of compounds under consideration is now considerably reduced relative to lead identification, but the systems employed still need a capacity of tens of compounds per week. Intestinal absorption, tissue penetration, metabolism, stability, and elimination are pharmacokinetic parameters which must be considered. Ideally, assessment of metabolism should include the determination of metabolic rates, since these can show great species variation. Human hepatocytes are especially relevant here, and other in vitro systems, such as the Caco-2 cell line for evaluating transcellular absorption, can also be helpful.

Knowledge of aqueous solubility, lipophilicity, pKa, protein binding, and other physicochemical parameters is as essential in lead optimisation as it is in lead identification. The types of in vivo studies and PBPK modelling employed previously can also be applied to any new chemical derivatives.

Early indications of toxic potential and of target organs are very useful at this stage. Cytotoxicity data obtained with organ-specific cells can be employed and, as they become available, mechanistically-based screening methods (for example, for immunotoxicity, nephrotoxicity, neurotoxicity, and phototoxicity) will also be of value for lead optimisation purposes. The identification of toxic effects which are dependent on the particular chemical series or are pharmacologically-related is of particular importance.

Compound Selection

Compound selection implies that drug discovery has been successful, and that a batch of compounds have emerged which are likely to be sufficiently safe and efficacious. Further selection will take place on the basis of other considerations, such as ease of synthesis and stability. Another important factor is the absence of significant toxicity in more than one species.

The use of experimentally-determined in vitro metabolism kinetics in estimating the likely kinetic behaviour of a compound is now an exciting development (5, 6). Estimates of in vivo metabolic clearance can be integrated with data from PBPK models for different species to predict likely human pharmacokinetics. If this type of methodology is shown to be reliable, it will permit a more rapid selection of compounds for further development (7).

The Integrated Use of Measurements and Models

To achieve rapid drug development, active candidate compounds can now be appraised via an integrated programme of in vitro and in vivo studies, ideally initiated after a computer-based screening of their likely physicochemical and toxicological properties (8). The goal is not merely to identify which of a series of lead compounds has the most appropriate safety profile, but also to permit the early commissioning of Phase I clinical trials. The evaluation of a compound in healthy human volunteers will provide important data on tolerance, pharmacokinetics and possible pharmacodynamics, which can rapidly be fed back to the discovery team to help its search for lead compounds and/or aid in the design of the subsequent development programme. This approach is only feasible because of our increased understanding of biological mechanisms, and advances in computerisation and analytical instrumentation. It also means that compounds may now be initially deselected without resorting to animal testing. For lead candidates, some animal testing will be necessary, at least for the foreseeable future.

Physicochemical Properties

Consideration should be given to molecular weight (MW), physical properties (appearance, salts, etc.), lipophilicity (log P; log D - the log of the distribution coefficient at pH 7.4), pKa, solubility, chirality, and the likely formulation and route of administration. Quantitative structure-activity relationship (QSAR) and computer modelling studies can be of value here, for example, in the assessment of absorption and/or barrier penetration.

Computer-based Systems/Molecular Modelling

By using rule-based programmes or neural networks, it is possible to predict, with increasing success, the potential toxic effects (relating to specific endpoints) associated with a particular chemical structure (for example, by using DEREK, "Deductive Estimation of Risk from Existing Knowledge" [9]), or the likely cytochrome P450 isoform that mediates metabolism.

Molecular modelling has developed slowly, since it is generally based on the crystal structure of bacterial cytochrome P450s (10), the only cytochromes crystallised to date. These computational programmes are being developed to provide either models of human enzyme substrate binding sites (11), or sites of metabolic instability based on frontier orbital electron densities (12). This approach has also been applied to predicting various forms of toxicity. The COMPACT ("Computer Optimised Molecular Parametric Analysis of Chemical Toxicity") system employs modelling software to calculate molecular and electronic structures of molecules, such as frontier orbital energies, which are of fundamental importance in describing the potential reactivities of molecules (13).

The current limitations of such systems should not be overlooked (for example, the cytochrome P450 models are based on bacterial enzymes). As yet, most computer models for predicting absorption, metabolism and toxicity have not been validated to any great extent (14).

In Vitro Studies

The in vitro studies conducted typically include metabolic and barrier penetration screens; these are primarily low throughput systems. The metabolic screens used have mainly been hepatic microsomal, hepatocyte or slice incubations, primarily derived from rat and dog. Compounds can be ranked quickly in terms of their metabolic stabilities by using these methods, especially with respect to those that are rapidly turned over; the sites at which metabolism occurs can be identified with the use of qualitative mass spectrometry. This information can be fed back rapidly to the project team, so that effort can be directed toward the introduction of functional groups which will alter the physical properties and/or make the molecules more metabolically stable.

Qualitative information on routes of metabolism can be obtained routinely from in vitro incubations. Furthermore, there is growing interest in the combined use of in vitro systems and biokinetic models to predict metabolic clearances in vivo (5). Conventional kinetic studies with in vitro systems can reveal the apparent maximum rate of metabolism (Vmax) and the substrate (drug) concentration required to give half the apparent maximum rate (Km). Under first-order conditions, Vmax and Km can be related to intrinsic clearance (Clint) by the relationship: Clint = Vmax/Km. By using this relationship, Michaelis-Menten parameters can be determined and the intrinsic clearance can be calculated. With the use of scaling factors, the hepatic clearance can be determined (15, 16), though with varying degrees of success depending upon the particular in vitro system used. To use in vitro data to give quantitative information, PBPK models can be developed for use in predicting the kinetics and toxicities of compounds (17).

Barrier penetration has been studied and assessed in vitro during the lead optimisation stage, by using either Caco-2 cells as a model of GI absorption (18, 19), or cerebral endothelial cells as a model of the blood-brain barrier (BBB; 20-23). These systems have not yet been fully validated for these purposes, and limited data exist for comparing in vitro and in vivo models of absorption (24) or brain penetration (20). However, cell-based systems have been used successfully to rank the ability of compounds to penetrate a membrane and have provided a valuable insight into the behaviour of compounds. Furthermore, studies with artificial membranes for measuring penetration could lead to the development of a high throughput system for use in the future.

A number of in vitro models have been used for the detection and characterisation of target organ toxicity (4, 25-27), of which isolated hepatocyte cultures have been the most extensively studied. To date, hepatocyte cultures have not been used in primary screening, but have proved to be valuable for characterising mechanisms of toxicity at later stages during drug development. In addition, various non-cellular in vitro methods have been proposed as screens for metabolically-activated toxicants (28, 29) and for compounds that induce pro-oxidant states (30); these may provide starting points for the development of mechanistically-based high throughput screens for such compounds. Thus, it seems likely that the development of in vitro screens for toxicity will proceed on two fronts: a) the use of cultures of relevant target cells; and b) the use of mechanistically-based non-cellular systems.

Human material has not been used routinely in discovery screening in vitro because of the amount of material which would be needed to satisfy all of the drug discovery projects. In addition, supply has been a problem in the past, as has confidence in the quality of the material used and, consequently, of the data produced. The issue of supply has begun to be resolved now that material is available commercially. The availability of a battery of human enzymes heterologously expressed in bacteria, yeast, insect cells, and mammalian cells (31) provides the opportunity to produce large quantities of protein which can be used in high throughput metabolic screens. However, given the significant requirement for throughput of compounds in the discovery mode, cost is still an important consideration.

The in vitro methods used during lead optimisation are intended to support internal decisions. It is therefore not necessary for them to be subjected to formal, interlaboratory validation. It is, of course, necessary to have standardised procedures and to validate the biological significance as well as the reliability of the individual tests, for instance, by using known reference compounds. The main issue, however, is the relevance of the individual test and the specific test battery to the intended target of the compound which is to be optimised.

In Vivo Studies

Generally, the most promising compounds from in vitro screens have been taken and their pharmacokinetic parameters evaluated in the rat or dog, in order to assess and optimise their kinetic profiles. Attempts should also be made to gain information about structure-toxicity or structure-pharmacokinetic relationships from these data. Although this is not possible in every case, such relationships, for example, those between physicochemical parameters such as partition coefficients, and metabolic clearance, or volume of distribution, may help in defining how chemical structures should be modified (3).

If the plasma clearance is high and approaching liver blood flow, in any species, then the assumption is generally made that hepatic clearance is likely to be high. Excreta and plasma are examined for metabolites, and the structural information is passed back to the project chemists. If metabolic clearance is not a problem in a structural series, then limited oral studies are conducted in a single species, usually in the rat. These studies consist of giving a single dose, with limited sampling points, to a very few animals. This reduces the numbers of animals required and increases efficiency and compound throughput.

Advances in mass spectrometry, particularly in the atmospheric pressure machine, have enabled a number of compounds to be co-administered to either the rat or dog, simultaneously and at very low doses. This approach has been used to good effect in producing pharmacokinetic data with particular emphasis on absorption. However, the numbers of compounds studied and the analytical limits of this approach have not yet been fully explored. There has been an increase in the use of liquid chromatography mass spectrometry for plasma analysis, rather than high performance liquid chromatography. The former has the advantages of automation, short analytical times, and multiple analysis after pooling samples.

When undertaking in vivo studies, it is best practice to initially conduct small "pilot" investigations with only a few animals, and to base any decisions about the most appropriate way to proceed on the results obtained in these pilot studies. In general, the workshop participants supported the view that many current in vivo pharmacokinetic studies used more animals than the minimum which would be required if physicochemical and in vitro data were used in a more-structured manner. The use of such an approach is important, since it offers the possibility of reducing the numbers of animals used in pharmaceutical research and development without prejudicing safety.

Specific Examples

Selection of Macromolecular Products for Chemical Development

Soluble synthetic polymers can be covalently bound to therapeutic agents via linkers which are designed for site-specific cleavage within the body. The resultant macromolecular products are suitable for intravenous administration, and can be directed to specific organs or tissues such as tumours. As the polymer-bound drug is inert, this reduces considerably any toxicity associated with the drug (32).

A polymer conjugate was developed, comprising the water-soluble copolymer, N-(2-hydroxylpropyl)methacrylamide (HPMA), linked to doxorubicin via the tetrapeptide spacer, Gly-Phe-Leu-Gly. This compound is called PK1 and has a MW of approximately 30,000 Daltons and a doxorubicin content of approximately 10% (by weight). The conjugate is only taken into cells by endocytosis, providing the opportunity for lyposomotropic drug delivery and thus by-passing the mechanisms of resistance associated with the plasma membrane efflux pump, P-glycoprotein. Doxorubicin must be liberated from the conjugate by lysosomal thiol-dependent proteases before it is able to exert its cytotoxic effect; it is believed that this happens intracellularly. A complex mechanism of action is involved, which results from pharmacokinetically-guided design (33).

The results of the Phase I clinical evaluation of PK1 (FCE 28068) have recently been reported, and show that the behaviour of the conjugate in humans (reduced toxicity, suitable pharmacokinetics) was similar to that observed preclinically, and that polymer-doxorubicin displays anti-tumour activity in man (34). This supports the approach taken for the rational design of polymeric anti-tumour agents (32, 33).

Polymeric anti-cancer agents are complex macromolecular drugs comprising the polymer backbone, the polymer drug-linker, and a bioactive anti-tumour agent. As such, they can only be captured by cells by endocytosis, so an appreciation of their pharmacokinetics at the cellular and the whole organism level is essential. To achieve an improved therapeutic index in vivo, it has been shown that it is essential to study the following factors systematically and feed the parameters back into the molecular design (Figure 1):

  1. biocompatibility (or toxicology) of the proposed carrier - this can largely be undertaken in vitro and is suitable for automation by using high throughput screens;
  2. the stability of the polymer-drug linker in the environments it will encounter, extracellularly and intracellularly (plasma, lysosomal) - this is suitable for automation by using model enzyme systems in vitro;
  3. rates of endocytosis - in vitro analysis is possible; and d) biodistribution and pharmacology in vivo.

Figure 1: Rational Design of Polymeric Anti-Cancer Agents

Experience gained during the transfer of this novel approach from the laboratory to the clinic allowed certain general comments to be made with regard to the importance of early pharmacokinetic evaluation in the selection and optimisation of lead candidates for in vivo and/or clinical study. In addition, early pharmacokinetic data are essential, to allow effective in vitro/in vivo human correlations to be made.

Although this research has focused on the design and development of polymer therapeutics, the main issues are directly relevant to any macromolecular drug. The last decade has seen a steady stream of protein and peptide drugs coming to the market, and there are currently great hopes for compounds which might be called "gene therapeutics".

Caco-2 Monolayers in Experimental and Theoretical Predictions of Drug Transport

During the last few years, the use of intestinal epithelial cell lines such as Caco-2 and HT29-H has increased dramatically in many research fields, including the pharmaceutical sciences. The cell lines are now routinely cultured as monolayers on permeable filters for studies of the transepithelial transport of drugs (18, 19, 35, 36).

Comparison of drug transport in Caco-2 monolayers with intestinal drug transport in vivo indicates that the monolayers can be used to predict drug transport by different pathways across the intestinal epithelium (24, 37-41). The best correlation to the in vivo situation is obtained for drugs which are transported by the passive transcellular route (42). The passive paracellular route is less permeable in the cell monolayers than in vivo, but the data so far indicate that the selectivity of this pathway is comparable to the in vivo situation (43). From these results, it can be concluded that Caco-2 monolayers can be used to identify drugs with potential absorption problems, and possibly also to select drugs with optimal passive absorption characteristics from series of pharmacologically-active molecules generated in drug discovery programmes. The absorption of drugs transported via carrier-mediated mechanisms can probably also be predicted in some, but not in all, cases. However, to confirm this, a more extensive characterisation of each active transport mechanism needs to be performed. Moreover, it is evident that there is a need for standardisation of Caco-2 cultures. Direct comparison of drug permeabilities obtained in different laboratories will only be possible if the same Caco-2 cell population, cell culture, and experimental conditions are used (44, 45).

Many attempts have been made to explain and predict passive drug absorption directly from the properties of a particular drug molecule. In these studies, single physicochemical properties, such as log P (46), hydrogen bonding capacity (47), or desolvation energy (48), have been correlated with the intestinal absorption rate or the cell membrane permeability. One advantage of using single parameters for this purpose is that they are relatively easy to determine experimentally or to derive from theoretical models. However, the relative importance of these properties will vary from one chemical type to another, so only approximate correlations can be obtained with single physicochemical properties.

Initial studies with a new theoretical method based on molecular surface properties suggest that the dynamic polar surface area, calculated by using molecular modelling, could be an interesting alternative for the prediction of drug absorption (49). The dynamic polar surface area correlates well with drug permeability in Caco-2 monolayers and excised intestinal segments for the limited series of compounds investigated to date, suggesting that Caco-2 monolayers could be used as a convenient reference model for theoretical predictions of drug absorption.

Very powerful methods have recently been developed for the combinatorial synthesis of large libraries of peptides and organic compounds, as have methods for the high throughput screening of pharmacological activity. As a result, large numbers of compounds with promising pharmacological activities are being obtained. This has increased the demand for screening methods for oral drug absorption (50), which suggests that the interest in cell culture models for experimental and theoretical predictions of drug absorption will continue to increase.

The In Vitro Blood-brain Barrier model for Investigating Drug Transport to the Brain

The isolation and culture of brain microvessel endothelial cells (BMEC) has had a major impact on research on the transport of drugs across the BBB into the brain, in particular at the cellular and sub-cellular levels (21, 51-53). BMEC can be isolated from human, bovine, porcine, and rat brains by using a completely non-enzymatic procedure (22), a combined mechanical and enzymatic method (23), or a solely enzymatic procedure (21). The non-enzymatic procedure may reduce the loss of surface molecules which can occur during treatment with enzymes, particularly with non-specific proteases. Cells are grown to confluency on cell culture filters in the presence of astrocyte-conditioned medium or astrocytes. The use of astrocyte-conditioned medium (22, 23) or co-culture with astrocytes (22) improves the maintenance in vitro of properties associated with the BBB in vivo (54, 55).

It is important that the in vitro systems to be used are well-characterised morphologically, biochemically and functionally. Cultured BMEC monolayers are useful for studying a wide range of important processes, for example: a) drug transport across the BBB, including passive hydrophilic (paracellular) transport, and passive lipophilic and carrier-/receptor-mediated (transcellular) transport; b) drug metabolism; c) the effects of specific compounds and conditions on the functionality of the BBB, both in normal situations and in disease states; d) visualisation of drug transport routes through confluent BMEC monolayers by using confocal laser scanning microscopy; and e) measurement of the trans-endothelial electrical resistance. It is important to investigate whether disease states change the properties of the BBB and, as a result, affect the transport of drugs to the brain. Inflammatory conditions, which may occur, for example, in multiple sclerosis, Alzheimer's disease, AIDS-related dementia, meningitis, and encephalitis, can be induced in vitro by the addition of lipopolysaccharide. Under such disease conditions, BBB permeability, functionality, and drug transport can be studied (56, 57).

Conclusions and Recommendations

Physicochemical Properties and Structure-activity Relationships

Significant problems remain in the calculation of log P, pKa, and log D, particularly when the very large numbers of compounds likely to be of interest in the early screening stage as a consequence of high throughput chemistry are taken into account. Comparative studies on the algorithms available for log P, pKa, and log D estimation are required, and validation of these methods is also needed. The current methods for estimating solubility are inadequate, and they are labour-intensive and have not been standardised.

If log P, pKa and log D, and other QSAR parameters are known, some important distribution parameters, such as GI absorption, protein binding, and BBB permeability, can be estimated adequately for use at the early compound screening stage. However, further validation of this approach with a wider range of chemical structures is required.

Methods that are less well developed are predictive models for toxicity and metabolism. Further methodological developments for the validation of knowledge bases for expert system approaches are necessary (14), while in the case of numerical calculations, such as modern pattern recognition methods, comparative studies for establishing their predictive performance are essential.

It will also be necessary to develop models which integrate and evaluate information about the component pharmacokinetic processes. Such overview models could be developed from data which are available on existing compounds but, at present, information is readily available only for a relatively small number of compounds. The question of which physical properties determine the distribution of macromolecules, and how they might be estimated, is still poorly understood, and there is a need for basic modelling to address this question.

The advantages and disadvantages of the methods currently used for determing physicochemical properties are outlined in Table I. The following recommendations are made:

  1. Expert systems or other models that can generate toxic or metabolic alerts based on historical precedent would be very valuable in supporting early decisions. The further development and validation of these methods should be supported.
  2. Physical properties, such as log P, pKa, and log D, provide essential information, and comparative studies of the predictive methods available should be carried out to identify the best methods for supporting early pharmacokinetic studies.
  3. Physical property modelling of macromolecular distribution should be undertaken.

Table I: Methods used for determining physicochemical properties

Methods Use Advantages Disadvantages Future Needs
clog P, prolog P log P estimation fast; computer-based unreliable for some structures mature area; little development needed
PKALC pKa estimation fast; computer-based unreliable for some structures urther development possible
Solubility and physical property measurements well-established, reliable experimental methods available costly and time-consuming automation, miniaturisation and improved analytical methods

Assessing Toxicity

The key issues in assessing toxicity are: a) target organ identification; b) the identification of general (that is, not organ-specific) targets; c) species differences; d) better understanding of mechanisms of toxicity; e) improving the availability of human material for in vitro studies; and f) improving the general accessibility of databases.

The advantages and disadvantages of the methods currently used for assessing toxicity are outlined in Table II. The future requirements for toxicological assessments at an early stage in drug discovery are given in Table III. It is recommended that:

  1. Support should be increased for the development of toxicology databases. This would be greatly facilitated if industry could be persuaded to release more information for use by others.
  2. There should be greater emphasis on the identification of common and relevant target cells and the establishment of suitable cell-based toxicity screens. Relevant target organs should be identified in collaboration with industry.
  3. Basic research into general mechanisms of toxicity, markers for specific toxic effects (for example, stress responses, free radical damage, covalent binding), and their relevance for predicting toxicity, should be encouraged. Mechanistically-relevant cellular systems and markers should then be integrated into cell-based toxicity screens.

Table II: Methods used for assessing toxicity

Methods Use Advantages Disadvantages Future Needs
Databases prediction alert uses prior knowledge limited accessibility to historical data improve accessibility
Expert systems prediction alert analyses SAR from a number of perspectives accuracy; limited expertise; limited endpoints improve and refine
Cellular toxicity models cell-based toxicology; target cell identification amenable to greater efficiency; fairly rapid; does not require much compound need to prove relevance maintenance of organotypic function; definition of a panel of meaningfulcells
Mechanistic models screens for common toxic effects high throughput; gives scope for chemical modification; species differences knowing relevance of individual mechanisms more basic research
In vivo identification of targets; integrated whole-body definitive assessment response; relevance; accepted species extrapolation; time-consuming better markers; transgenic animals; integration of data from non-animal methods in design

Table III: Future requirements for undertaking toxicological assessments at an early stage of drug discovery

Phase Needs Uses Future Trends
Lead structure identification
- high throughput screening increase probability for benign toxicological profile; exclude toxic structures new predictive markers for toxicity
  1. stress genes
  2. cell cycle markers

expert systems
- secondary screens limited cytotoxicity testing mechanistically-based screens
Lead optimisation identify critical targets; reduce toxicity on these targets exploratory toxicity for target class-/structure-related toxicity; in vitro screening on selected targets better in vitro screens

  1. recombinant cell lines
  2. mechanistic markers
  3. new techniques (PCR, mRNA mapping)
Candidate selection predict all relevant targets;
assess human risk;
select the best candidate		 comprehensive in vitro and in vivo test battery; mechanistic studies;
definitive in vivo studies		 transgenic animals;
better ex vivo markers

Assessing Metabolism

A key issue is whether metabolism screening needs to be conducted with biological systems, or whether it can be accomplished by using expert systems which would identify and rank possibilities. The current situation requires metabolism properties to be optimised by using a biological system, and a decision is taken about whether this should incorporate human tissue or tissues from the species to be used in any subsequent in vivo studies. Both metabolic route and rate should be determined.

It is likely that some combination of biological and expert systems would be the most productive approach, in combination with the use of PBPK models. One problem is that the design of studies and interpretation of the data they provide are currently dependent upon relatively few individual experts and the confidential databases to which they have access.

The advantages and disadvantages of the systems currently used for investigating metabolism are outlined in Table IV. It is recommended that:

  1. Research aimed at the provision of more recombinant systems, with a greater range of diversity, should be encouraged. These systems must be fully characterised.
  2. Attention should be focused on the circumstances in which the use of microsomal and/or whole cell systems is appropriate.
  3. An objective evaluation should be conducted on the comparative utility of hepatocytes and liver slice preparations.
  4. Procedures should be established at the European level for the safe, regular, and ethical supply of human livers of high quality.
  5. More effort should be invested in the development and optimisation of various types of techniques which could improve the preservation of metabolic functions in in vitro preparations, including cryopreservation and the use of culture matrices and various culture media.
  6. Effort should be invested in the further development of PBPK modelling as a means of improving the usefulness of data provided by in vitro test systems.

Table IV: Biological systems used for investigating metabolism

System Status Use Advantages Dis- advantages Future Needs
Recombinant validation lead optimisation pure system single system integration with other enzyme systems
Microsomes established lead optimisation proven utility;
ready storage;
well-characterised		 artifactual; limited enzymes; cofactor requirement
Isolated cells (e.g. hepatocytes) established/ evolving lead optimisation integrated cellular system short survival;
limited enzyme stability;
technically demanding		 greater availability and use of human cells;
better preservation
Slices evolving lead optimisation ease of preparation;
intact architecture		 diffusion rate limitation;
short-term viability		 improved availability of human tissues;
cryo- preservation

Assessing Permeability

A numbers of key factors must be considered when permeability is to be assessed, including the proposed route of administration of the drug, the target site, primary and secondary barriers, and potential sites of toxicity. The nature of the compounds under investigation is also very important; for example, whether they are low MW or macromolecular (peptides, proteins, oligonucleotides or polymers) should be taken into account.

The advantages and disadvantages of the model systems currently used for assessing permeability are outlined in Table V. It is recommended that:

  1. Standard procedures should be developed and validated for the model systems used to monitor penetration/metabolism at barrier sites and at the target cell level.
  2. In vitro models should be used for studying GI and BBB permeabilities, especially for compounds with values of log P and log D at the extremes of their distributions.
  3. Closer collaboration, including the sharing of databases at the European level, would lead to improvements in the biological systems and other models needed.
  4. A workshop should be organised on the use of Caco-2 cells and other artificial and cell-based systems for assessing permeability.
  5. A workshop should be held on the early use of cell models and in vivo pharmacokinetic models for the rational design and selection of macro-molecular drugs.

Table V: Biological systems used for assessing permeability

Systems Use Advantages Disadvantages Future Needs
Membrane vesicles membrane transport;
qualitative interactions		 simple poor correlation metabolism
Isolated cell systems quantification of intracellular distribution;
active transportsystems		 rational design for (macromolecular) drugs time-consuming standardisation;
improved analytical methods
Cell monolayers, tissue sections and loops penetration;
transport		 controlled system;
more physiological		 variability;
time-consuming		 validation
Blood-brain barrier ± astrocytes transport;
metabolism		 routine cellular system simplified system;
not quantitative;
extrapolation		 include pericytes;
standard reference compounds

References

  1. Anon. (1994). ECVAM News & Views. ATLA 22: 7-11.
  2. Gobas, F.A.P.C., Lahittete, J.M., Gorofalo, G., Shiu, W.Y. & MacKay, D. (1988). A novel method for measuring membrane-water partition coefficients of hydrophobic organic chemicals: comparison with 1-octanol-water partitioning. Journal of Pharmaceutical Science 77: 265-272.
  3. Watari, N., Sugiyama, Y., Kaneniwa, N. & Hiura, M. (1988). Prediction of hepatic first-pass metabolism and plasma levels following intravenous and oral administration of barbiturates in the rabbit based on quantitative structure-pharmacokinetic relationships. Journal of Pharmacokinetics and Biopharmaceutics 16: 279-301.
  4. Blaauboer, B.J., Boobis, A.R., Castell, J.V., Coecke, S., Groothuis, G.M.M., Guillouzo, A., Hall, T.J., Hawksworth, G.M., Lorenzon, G., Miltenburger, H.G., Rogiers, V., Skett, P., Villa, P. & Wiebel, F.J. (1994). The practical applicability of hepatocyte cultures in routine testing. The report and recommendations of ECVAM workshop 1. ATLA 22: 231-241.
  5. Blaauboer, B.J., Bayliss, M.K., Castell, J.V., Evelo, C.T.A., Frazier, J.M., Groen, K., Gülden, M., Guillouzo, A., Hissink, A.M., Houston, J.B., Johanson, G., de Jongh, J., Kedderis, G.L., Reinhardt, C.A., van de Sandt, J.J.M. & Semino, G. (1996). The use of biokinetics and in vitro methods in toxicological risk evaluation. The report and recommendations of ECVAM workshop 15. ATLA 24: 473-497.
  6. Iwatsube, T., Hirota, N., Ooie, T., Suzuki, H. & Sugiyama, Y. (1996). Prediction of in vivo drug disposition from in vitro data based on physiological pharmacokinetics. Biopharmaceutics and Drug Disposition 17: 273-319.
  7. Charnick, S.B., Kawai, R., Nedelman, J.R., Lemaire, M., Niederberger, W. & Sato, H. (1995). Physiologically-based pharmacokinetic modelling as a tool for drug development. Journal of Pharmacokinetics and Biopharmaceutics 23: 217-235.
  8. Barratt, M.D., Castell, J.V., Combes, R.D., Dearden, J.C., Fentem, J.H., Gerner, I., Giuliani, A., Gray, T.J.B., Livingstone, D.J., Provan, W.Mc., Rutten, A.A.J.J.L., Verhaar, H.J.M. & Zbinden, P. (1995). The integrated use of alternative approaches for predicting toxic hazard. The report and recommendations of ECVAM workshop 8. ATLA 23: 410-429.
  9. Sanderson, D.M. & Earnshaw, C.G. (1991). Computer prediction of possible toxic action from chemical structure, the DEREK system. Human and Experimental Toxicology 106: 267-279.
  10. Lewis, D.F.V. (1995). Three dimensional models of human and other mammalian microsomal P450s constructed from an alignment with P450102 (P450bm3). Xenobiotica 25: 333-366.
  11. Mancy, A., Broto, P., Dijols, S., Dansette, P.M. & Mansuy, D. (1995). The substrate binding site of human liver cytochrome P450 2C9: an approach using designed tienilic acid derivatives and molecular modelling. Biochemistry 34: 10365-10375.
  12. Rietjens, I.M.C.M., Soffers, E.M.F., Veeger, C. & Vervoort, J. (1993). Regioselectivity of cytochrome P450 catalysed hydroxylations of fluorobenzenes predicted by calculated frontier orbital substrate characteristics. Biochemistry 32: 4801-4812.
  13. Lewis, D.F.V. (1995). COMPACT and the importance of frontier orbitals in toxicity mediated by the cytochrome P450 mono-oxygenase system. Toxicology Modelling 1: 85-97.
  14. Dearden, J.C., Barratt, M.D., Benigni, R., Bristol, D.W., Combes, R.D., Cronin, M.T.D., Judson, P.N., Payne, M.P., Richard, A.M., Tichy, M. & Yourick, J.J. The development and validation of expert systems for predicting toxicity. The report and recommendations of ECVAM workshop. ATLA 25: in press.
  15. Houston, J.B. (1994). Utility of in vitro drug metabolic clearance data in predicting in vivo metabolic clearance. Biochemical Pharmacology 47: 1469-1479.
  16. Bayliss, M.K., Bell, J.A., Jenner, W.N. & Wilson, K. (1990). The prediction of intrinsic clearance of loxtidine from kinetic studies in rat, dog, and human hepatocytes. Biochemical Society Transactions 18: 1198-1199.
  17. Frazier, J.M. (1995). Biokinetic modelling and in vitro assays. Application of in vitro systems to the prediction of in vivo biokinetics. Toxicology In Vitro 9: 527-536.
  18. Artursson, P. (1990). Epithelial transport of drugs in cell culture. A model for studying the passive diffusion of drugs over intestinal absorptive (Caco-2) cells. Journal of Pharmaceutical Science 79: 476-482.
  19. Hidalgo, I.J., Raub, T.J. & Borchardt, R.T. (1989). Characterisation of the human carcinoma cell line (Caco-2) as a model system for intestinal epithelial permeability. Gastroenterology 96: 736-749.
  20. Dehouck, M.P., Jolliet-Riant, P., Bree, F., Fruchart, J.C., Cecchelli, R. & Tillement, J.P. (1992). Drug transfer across the blood-brain barrier: correlation between in vitro and in vivo models. Journal of Neurochemistry 58: 1790-1797.
  21. Audus, K.L. & Borchardt, R.T. (1986). Characterization of an in vitro blood-brain barrier model system for studying drug transport and metabolism. Pharmaceutical Research 3: 81-87.
  22. Dehouck, M.P., Mresse, S., Delorme, P., Fruchart, J.C. & Cecchelli, R. (1990). An easier, reproducible, and mass-production method to study the blood-brain barrier in vitro. Journal of Neurochemistry 54: 1798-1801.
  23. Rubin, L.L., Hall, D.E., Parter, S., Barbu, K., Cannon, C., Horner, H.C., Janatpour, M., Liaw, C.W., Manning, K., Morales, J., Tanner, L.I., Tomaselli, K.J. & Bard, F. (1991). A cell culture model of the blood-brain barrier. Journal of Cell Biology 115: 1725-1735.
  24. Artursson, P. & Karlsson, J. (1991). Correlation between oral drug absorption in humans and apparent drug permeability coefficients in human intestinal epithelial (Caco-2) cells. Biochemical and Biophysical Research Communications 175: 880-885.
  25. Atterwill, C.K., Bruinink, A., Drejer, J., Duarte, E., McFarlane Abdulla, E., Meredith, C., Nicotera, P., Regan, C., Rodr'guez-Farré, E., Simpson, M.G., Smith, R., Veronesi, B., Vijverberg, H., Walum, E. & Williams, D.C. (1994). In vitro neurotoxicity testing. The report and recommendations of ECVAM workshop 3. ATLA 22: 350-362.
  26. Hawksworth, G.M., Bach, P.H., Nagelkerke, J.F., Dekant, W., Diezi, J.E., Harpur, E., Lock, E.A., MacDonald, C., Morin, J-P., Pfaller, W., Rutten, A.A.J.J.L., Ryan, M.P., Toutain, H.J. & Trevisan, A. (1995). Nephrotoxicity testing in vitro. The report and recommendations of ECVAM workshop 10. ATLA 23: 713-727.
  27. Seibert, H., Balls, M., Fentem, J.H., Bianchi, V., Clothier, R.H., Dierickx, P.J., Ekwall, B., Garle, M.J., Gómez-Lechón, M.J., Gribaldo, L., Gülden, M., Liebsch, M., Rasmussen, E., Roguet, R., Shrivastava, R. & Walum, E. (1996). Acute toxicity testing in vitro and the classification and labelling of chemicals. The report and recommendations of ECVAM workshop 16. ATLA 24: 499-510.
  28. Garle, M.J. & Fry, J.R. (1989). Detection of reactive metabolites in vitro. Toxicology 54: 101-110.
  29. Mulder, G.J. & Le, C.T. (1988). A rapid, simple in vitro screening test, using [3H]-glutathione and L-[35S]-cysteine as trapping agents, to detect reactive intermediates of xenobiotics. Toxicology In Vitro 2: 225-230.
  30. Knobeloch, L.M., Blondin, G.A., Lyford, S.B. & Harkin, J.M. (1990). A rapid bioassay for chemicals that induce pro-oxidant states. Journal of Applied Toxicology 10: 1-5.
  31. Gonzalez, F.J. & Korezekwa, K.R. (1995). Cytochrome P450 expression systems. Annual Reviews of Pharmacology and Toxicology 35: 369-390.
  32. Duncan, R. (1992). Drug-polymer conjugates: potential for improved chemotherapy. Anti-Cancer Drugs 3: 175-210.
  33. Duncan, R. & Spreafico, F. (1994). Polymer conjugates: pharmacokinetic considerations for design and development. Clinical Pharmacokinetics 27: 290-306.
  34. Vasey, P.A., Duncan, R., Kaye, S.B. & Cassidy, J. (1995). Clinical Phase I trial (HPMA copolymer doxorubicin). European Journal of Cancer 31A: S193.
  35. Artursson, P. (1991). Cell cultures as models for drug absorption across the intestinal mucosa. Critical Reviews of Therapeutic Drug Carrier Systems 8: 305-330.
  36. Hillgren, K.M., Kato, A. & Borchardt, R.T. (1995). In vitro systems for studying intestinal drug absorption. Medical Research Reviews 15: 83-109.
  37. Cogburn, J.N., Donovan, M.G. & Schasten, C.S. (1991). A model of human small intestinal absorptive cells. 1. Transport barrier. Pharmaceutical Research 8: 210-216.
  38. Rubas, W., Jezyk, N. & Grass, G.M. (1993). Comparison of the permeability characteristics of a human colonic epithelial (Caco-2) cell line to colon of rabbit, monkey, and dog intestine and human drug absorption. Pharmaceutical Research 10: 113-118.
  39. Wils, A., Warney, A., Phung-Ba, V. & Scherman, D. (1994). Differentiated intestinal epithelial cell lines as in vitro models for predicting the intestinal absorption of drugs. Cell Biology and Toxicology 10: 393-397.
  40. Rubas, W., Villagran, J., Cromwell, M., McLeod, A., Wassenberg, J. & Mesny, R. (1995). Correlation of solute flux across Caco-2 monolayers and colonic tissue in vitro. S.T.P. Pharmaceutical Sciences 1: 93-97.
  41. Stewart, B.H., Chan, H., Lu, R.H., Reyner, E.L., Schmid, H.L., Hamilton, H.W., Steinerg, B.A. & Taylor, M.D. (1995). Comparison of intestinal permeabilities determined in multiple in vitro and in situ models: relationship to absorption in humans. Pharmaceutical Research 12: 693-699.
  42. Lennernas, H., Palm, K., Fagerholm, V. & Artursson, P. (1996). Correlation between paracellular and transcellular drug permeability in the human jejunum and Caco-2 monolayers. International Journal of Pharmaceutics 127: 103-107.
  43. Artursson, P., Ungell, A.L. & Lofroth, J.E. (1993). Selective paracellular permeability in two models of intestinal absorption. Cultured monolayers of human intestinal epithelial cells and rat intestinal segments. Pharmaceutical Research 10: 1123-1129.
  44. Artursson, P., Karlsson, J., Ocklind, G. & Schipper, N. (1996). Studying transport processes in absorptive epithelia. In Cell Culture Models of Epithelial Tissues - A Practical Approach (ed. A. Shaw), pp. 111-133. New York: Oxford University Press.
  45. Walter, E. & Kissel, T. (1994). Heterogeneity in the human intestinal cell line Caco-2 leads to differences in transepithelial transport. European Journal of Pharmaceutical Science 3: 656-671.
  46. Martin, Y.C. (1981). A practioner's perspective of the role of quantitative structure-activity analysis in medicinal chemistry. Journal of Medicinal Chemistry 24: 229-237.
  47. Young, R.C., Mitchell, R.C., Brown, T.H., Ganellin, R.C., Griffiths, R., Jones, M., Rana, K.K., Saunders, D., Smith, I.R., Sore, N.E. & Wilks, T.J. (1988). Development of a new physicochemical model for brain penetration and its application to the design of centrally acting H2 receptor histamine antagonists. Journal of Medicinal Chemistry 31: 656-671.
  48. Wright, L.L. & Painter, G.R. (1992). Role of desolvation energy in the nonfacilitated membrane permeability of dideoxyribose analogs of thymidine. Molecular Pharmacology 41: 957-962.
  49. Palm, K., Luthman, K., Ungell, A.L., Strandlund, G. & Artursson, P. (1996). Correlation of drug absorption with molecular surface properties. Journal of Pharmaceutical Science 85: 32-39.
  50. Stevenson, C.L., Augustijns, P.F. & Hendren, R.W. (1995). Permeability screen for synthetic peptide combinatorial libraries using Caco-2 cell monolayers and LC/MS/MS. Pharmaceutical Research 12: S94.
  51. van Bree, J.B.M.M., de Boer, A.G., Danhof, M., Ginsel, L.A. & Breimer, D.D. (1988). Characterization of in vitro blood-brain barrier: effects of molecular size and lipophilicity on cerebrovascular endothelial transport rates of drugs. Journal of Pharmacology and Experimental Therapeutics 247: 1233-1239.
  52. van Bree, J.B.M.M., de Boer, A.G., Danhof, M. & Breimer, D.D. (1992). Drug transport across the blood-brain barrier. II. Experimental techniques to study drug transport. Pharmaceutica Weekblad (Sci) 14: 338-348.
  53. Joo, F. (1993). The blood-brain barrier in vitro: the second decade. Neurochemistry International 23: 499-521.
  54. Meyer, J., Rauh, J. & Galla, H-J. (1991). The susceptibility of cerebral endothelial cells to astroglial induction of blood-brain barrier enzymes depends on their proliferative state. Journal of Neurochemistry 57: 1971-1977.
  55. Mresse, S., Dehouck, M-P., Delorme, P., Bensad, M., Tauber, J.P., Delbart, D., Fruchart, J-C. & Cecchelli, R. (1989). Bovine brain endothelial cells express tight junctions and monoamine oxidase activity in long-term culture. Journal of Neurochemistry 53: 1363-1371.
  56. de Vries, H.E., Blom-Roosemalen, M.C.M., de Boer, A.G., van Berkel, T.J.C., Breimer, D.D. & Kuiper, J. (1996). Effect of endotoxin on the permeability of bovine cerebral endothelial cell layers in vitro. Journal of Pharmacology and Experimental Therapeutics 277: 1418-1423.
  57. de Vries, H.E., Blom-Roosemalen, M.C.M., de Boer, A.G., van Berkel, T.J.C., Breimer, D.D. & Kuiper, J. (1996). The influence of cytokines on the integrity of the blood-brain barrier in vitro. Journal of Neuroimmunology 64: 37-43.