1356-1359.

Cardon, P. V., Donnell, F. W., Jr., & Trumble, R. R. (1976). Injuries to research subjects: a

survey ofinvestigators. Ne w Eng. 1. Med.. 295, 65D-<i54. Comford, F. M. (1908). Microcosmographia academ ica. Being a guide for the young academic

politician. Cambridge: Bowesand Bowes.

Epstein, L. C. & Lasagna, L. (1969). Obtaining informed consent. Form or substance. Arch. Im .

Med.. 123,682-688.

Joubert, P & Lasagna, L. (1975a). Patient package inserts. Part I. Clin. Pharmac. Ther., 18,

507-513.

Joubert, P. & Lasagna, L. (1975b). Patient package inserts. Part 11. Clin. Pharmac. Ther.. 18,

663-669 .

Zarafonetis, C. r. D., Riley, P. A., lr., Willis, P. W. III, Power, L. H., Werbelow,r., Farhat, L.,

Beckwith, W. & Marks, B. H. (1978). C1inically significant adverse elfects in a phase I

testing program. Clin. Pharma c. Ther.. 24, 127-132.

PREDICTIVE VALUE OF

PRE-CLINICAL DRUG

SAFETY EVALUATION

G.ZBINDEN

Institute ofToxicology,

Swiss Federallnstitute ofTechnology and University of Zurich ,

Schwerzenbach, Switzerland

Toxicological standard procedures

Pre-c1inical safety studies were originall y designed to imitate the future c1inical use of

new drugs in experimental animals. Many scientific committees, drug regulatory

agencies and individual scientists have feIt the need to standardize these experiments.

For this reason , they have proposed toxicological guidelines or even detailed

protocols. As a consequence, a certa in uniformity exists among the various

regulatory requirements for the registration of new drugs, This applies mostly to the

general format oftoxicity testing. But there remain still man y annoying discrepancies

between the exigencies ofvarious drug regulatory agencies (Griffm, 1979).

In order to fulfil regulatory requirements for world-wide sales of a new drug,

studies must, therefore , often be duplicated. For each test the most detailed and allinclusive protocol must be used, in the hope that it will be acceptable to all

government agencies. This greatl y increases the alread y exorbitant costs (Gelzer,

1979)ofroutine animal toxicity experiments.

In his famous 'Introduction into experimental medicine' Claude Bernard lists two

major reasons for embarking on a scientific experiment. These are I) the departure

point is an observation, and 2) the departure point is a hypothesis or a theory. The

father of modern experimental biology would certainly be surprised if he could see

the enormous number of toxicological experiments conducted today which have

neither a scientific observation as a point of departure nor the declared goal to test an

intelligent hypothesis. The current approach to toxicity testing is essentially designed

to complete a check list which includes, besides a variety of chemical and physical

constants, several biological characteristics of a new drug. These data must be

submitted to regulatory agencies before the compound is registered for marketing or,

in certain countries, permitted to be used for c1inical studies.

In order to be acceptable, pre-clinical safety evaluation studies must be done in

accordance with guidelines that are acknowledged in the particular country where

the new drug is registered. In general several types of experiments are required,

namely :

I) Single dose (acute) toxicity studies in two or more animal species with two or more

routes of administration.

10 G.ZBINDEN

2) Repeated dose (subacute, subchronic and chronic) toxicity studies in two animal

species and with all routes ofadministration expected to be used in man .

3) Reproduction studies including experiments on fertility of male and female

animals of one species, on teratogenicity in two species and on peri- and postnatal

development in one species.

4) Mutagenicity and life-time carcinogenicity studies are increasingly required,

particularly for drugs that are intended for continuous administration over an

extended part of the human life, and for agents that are chemically and pharmacologically related to known genotoxic substances.

5) A variety of acute experiments for the evaluation of miscellaneous functional and

biochemical effects (safety pharmacology).

As a general rule all the above experiments must be done with several dose levels,

and the requirement that the highest dose used must induce measurable pharmacological or toxic effects is rigidly enforced.

Predictability ofthe toxicological standard procedures

A unique feature of toxicological experiments is the fact that positive and negative

results are of equal importance. This is quite different from the work ofthe pharmacologists who can discard negative test results and pay attention mainly to potentially

useful effects, Toxicologists, on the other hand , must not only detect adverse

properties of the test compounds, but must, if no positive effects are evident, prove

the absence of adverse qualities of new drugs . The latter is, of course, much more

difficult to establish.

The overall performance ofthe current pre-clinical safety evaluation procedures is

very difficult to assess. There are no published data on the number of compounds

that were found to be too toxic to be released for clinical trials. Of the compounds

that were withheld from clinical use because of adverse effects in animals it is not

possible to determine the percentage that would have been tolerated by humans

('false positive' animal toxicity), and that which would have induced comparable

toxic reactions in man ('true positive' animal toxicity).

Some useful information can be gathered from a retrospective evaluation of drugs

that have been used extensively in humans and whose toxicological spectrum can be

compared with the effects observed in animal experiments (Fletcher, 1978).

Experience has shown that the adverse effects commonly referred to as 'dose- and

time-dependent toxic reactions' are read ily predicted by animal experiments. A

selection of such reactions is listed in Table 1. In the light of present criticisms of

animal experimentation even by certain academics, it must be stressed that a large

number ofserious toxic reactions ofpotential new drugs can be detected reliably with

toxicological tests. To negate publicly the scientific basis and validity of experimental toxicology is an irresponsible act that could hurt the development ofessential

new drugs and could cause serious harm to mankind.

A much more difficult problem for toxicologists and clinicians alike are the

unpredictable adverse effects referred to as 'not dose- and tirne-related toxic

reactions'. The fact that they occur usually only in a small segment of patients (Towincidence responses', Plaa, 1978) proves that factors other than the administered

drugs play an essential role in their development. Ifthese factors are not operational

in the animal experiments, such toxic reactions cannot be pred icted. Even grotesque

overdosage, an experimental trick proposed to overcome these shortcomings of the

experimental models , can often not help to avoid such 'false negative' toxicity

experiments.

The recognition ofthis important problem represents already a decisive step in the

direction of its solution. If the toxicologist can defme the type oftoxic reactions that

PREDICTIVE VALUE OF PRE-CLINICAL DR UG SAFETY EVALUATION 11

Table 1 Examples for time- and dose-related toxic reactions that can be demonstrated in

animal experiments

Drug

Isoniazid

Aminoglycoside antibiotics

Non-steroidalanti-inflammatory drugs

Anthracycline antibiotics

Cytostaticdrugs

Hexachlorophene

Alkylatingdrugs

Chloroquine

Thiouracil

Corticosteroids

Acetaminophene

Oxygen

Reserpine, neuroleptic drugs

Emetine

Sulphonamides

Taxie reaction

Peripheral neuropathy

Tubular damageofkidneys and ototoxicity

Gastrointestinal ulceration

Cardiomyopathy

Bonemarrow depression,gastrointestinal

ulceration, testiculardamage

Spongiformencephalopathy

Carcinogenicity

Retinopathy

Thyroid hyperplasia

Adrenal atrophy

Hepatotoxicity

Lungdamage

Mammary hyperplasia

ECGchanges

Crystalluria

are difficult or impossible to detect in animals, the clinician is warned and can take

appropriate measures to look for such effects in his patients. Moreover, the recognition ofthe reasons why animal experiments sometimes fail to predict relevant toxic

effects will often lead to arevision and improvement oftoxicological methods.

The most important reasons for false negative toxicity tests (Table 2) shall brietly

be discussed.

Table 2 Reasonsforfalse negativetoxicitytests

Pharmacokinetic and metabolieditTerences Absenceof pre-existingpathologicalconditions

Anatomical and physiological ditTerences

Tolerance - enzyme induction

Lack ofappropriate assay methods

Failure to useappropriate assaymethods

Pharmaeokinetie and metabolie differences

Species differences for plasma half-lives and metabolic disposition of drugs are weil

recognized (Baeder, 1980). Rapid detoxification and excretion in animals can account

for the failure to recognize a relevant toxic effect. The often heard desideratum that

toxicity experiments should be conducted in animals that handle the test drug like

man, is therefore understandable, though rather naive, if one considers that hardly

more than half a dozen an imaI species are available for toxicity testing and that

qualitative and quantitatives differences in drug metabolism between species are the

rule rather than the exception. It is more reasonable to demand that the extent of

these differences be determined and be brought to the attention ofthe clinicians to be

included in the evaluation process ofthe toxicological data. Moreover, the pharmacokinetic and metabolic da ta should be used to optimize the toxicological

experiment, for example in the selection of the application schedules and other

modifications ofthe procedures, such as, the use ofanimals with induced or inhibited

drug metabolizing enzymes (Mitchell & Jollow, 1974).

In the case oflow-incidence responses in man the existence ofsubpopulations with

special sensitivities against certain drugs has been suggested (Plaa, 1978). Typical

examples are sporadically occurring reactions such as cholestatic jaundice, hepatitis,

12 G.ZBINDEN

agranulocytosis and collagen diseases. Since the exact mechanisms of such

idiosyncrasies are not known , it is impossible to develop rational toxicological

strategies. Sometimes the use of several inbred strains of animals permits the

demonstration of a rare toxic efTect that was missed in routine experiments, for

example, induction of cholestatsis in certain strains of inbred mice by anabolic

steroids (Imai & Hayashi, 1970). In other cases animals with chemically induced

organ lesions may be helpful , for example, demonstration of hepatotoxicity of

isoniazid, iproniazid and derivatives in rats with CC14 - induced hepatic lesions

(Zbinden & Studer, 1959).

Absence 0/pre-existing pathological conditions

Drugs may exacerbate pre-existing diseases (for example, diabetes, gastric ulcer ,

mental depression, latent infections), and diseases can increase the toxicity of

therapeutic agents . Since pre-clinical safety studies are done in healthy animals, such

interactions can often not be demonstrated. From the pharmacodynamic and

pharmacokinetic proftle and the toxicological data obtained in healthy animals, it is,

however, usually possible to predict the particularly unfavourable conditions that

would make the drug especially hazardous. In high risk patients special precautions

can then be taken.

Anatomical and physiological differences

A review of adverse reactions to drugs in man ind icates that they are most frequentl y

connected with functional disturbances ofthe central and autonomic nervous system

(Zbinden, 1966). It is understandable, therefore, that their symptomatology difTers

considerably between experimental animals and man . For example, a frequent side

efTect of man y drugs in man is hypotension, and this efTect is often not present in

quadruped animals (Fletcher, 1978).

Tolerance and enzyme induction

In recent years drug regulatory agencies have tended to demand more and more longterm toxicity experiments. Itshould be recognized , however, that the predictability of

such studies is not alwa ys better than that of the short-term tests. Man y drugs are

able to stimulate their own metabolism. For example, Poilmann (1980) reported

fmdings with a coronary dilator Me 621, (N-[3-phenyl-3-hydroxypropyl]-N-[2-

(p-hydroxyphenyl)-2-hydroxyethyl]-am ine). After only seven doses of 16 kg-I per

os in rats , peak blood levels were markedly decreased , and excretion was greatly

accelerated. With such drugs, toxic tissue concentrations may never be reached, and

tissue lesions that are initially present may disappear, although treatment is

continued for a long time. Development of tolerance is, therefore, a frequent

occurrence in chronic toxicity experiments. Among many examples reviewed by

Balazs (1974), the disappearance of kidney lesions during chronic administration of

several nephrotoxic substances, (for example, mercuric chloride, salicylates,

nephrotoxic antibiotics) is particularly instructive.

Lack 0/ appropriate assay methods

Many adverse reactions to drugs cannot be measured object ively but have to be

communicated verbally by the patient. Such efTects are, of course, very difficult to

recognize in animal experiments. However, behavioural toxicity is an emergent

discipline, and it is hoped that such subjecti ve efTects as hallucinations, dizziness,

difficult y in concentrating and disturbance of memory will soon also be measurable

in animal experiments. A much more difficult problem are allergic reactions that are

PREDICTIVE VALUE OF PRE-CLINICAL DRUG SAFETY EVALUATION 13

among the frequent drug side elTects in man. In this area there are still few promising

research leads.

Failure to use appropriate research methods

In the beginn ing of this review it was pointed out that toxicologists are forced to

conduct their studies according to official guidelines. This has led to a deplorable

monotony in toxicological practice and a dearth of innovative approaches. In recent

years, however, toxicologists have looked beyond the limits of their usual cruising

radius, and have tried to apply scientific methods of other disciplines. This has been

particularly fruitful in the area of short-term mutagenicity and carcinogenicity

testing (Zbinden & Schlatter, 1978; Zbinden, 1979), and in the incorporation of

pharmacological methods into conventional drug safety assessment (Zbinden &

Gross, 1979).

The problem of false positive results

In routine animal toxicity experiments, many organ lesions are induced that are ofno

relevance to the safety ofthe drug in man. The most important reason why these tests

often overpredict adverse drug elTects is the regulatory requirement to conduct all

experiments with doses that must induce a measurable toxic elTect. This problem is

particularly evident in the acute toxicity experiments where the doses must be

increased until the animals die. This then leads to often dramatic agonal symptoms

such as ataxia and convulsions that are not encountered during clinical application of

the drug in man (Fletcher, 1978). In subacute, subchronic and chronic experiments

the symptoms caused by overdosage are usually not.as dramatic, but they may still be

quite serious. This is particularly the case when overloading with the chemical

substances leads to chronic mal nutrition of the animals. Organ changes , such as

single cell necrosis of the liver, gastrointestinal inflammation and ulceration, bone

marrow .depression, hyperplasia and neoplasia of endocrine-dependent organs and

the liver, may occur. These lesions are often difficult to separate from those that are

specifically related to drug treatment.

The best way to deal with the problem offalse positive results is to apply pharmacokinetic concepts also to experimental toxicology. It isjust not acceptable anymore

to conduct a safety test according to standard protocols, to register and to report the

many serious lesions that have occurred and then to put olT the whole thing by

pointing out that the experiment was done with a highly exaggerated dose. In modern

toxicology, all lesions , regardless at what dose they have occurred, must be judged

against the background of sound pharmacokinetic data. Only when blood and tissue

concentrations, measured in the poisoned animals, are significantly higher than those

found after therapeutic application in humans, can the observations in the animal

model be attributed to 'intentional overdosage'.

Conclusions

Modern pre-clinical drug safety studies generate a fantastic amount of'scientific data,

and most of it is negative . Uninformed critics have pointed to this mass of negative

results and have used it to ridicule the elTorts of industrial toxicologists and govemmental regulatory agencies. They have forgotten, however, that the large majority of

drug exposures of human subjects, are, from the toxicological point of view, totally

uneventful encounters and in perfect accordance with the results of the animal

experiment.

14 G.ZBINDEN

Pre-clinical safety evaluation studies are today the only means to conduct clinical

research ethically and safely. For the patient volunteer and his physician it is

essential to know that a substantial number of animals have received the new drug

without serious ill effects, that a large group oftrue and serious toxic reactions have

reliably been excluded by well validated methods, and that many of the potential

target organs of toxicity or adverse symptoms have been identified and can now be

especially monitored.

Clinicians are also well aware that the pre-clinical safety evaluation cannot

guarantee total absence of risk. But the areas of imperfection and insufficient

predictability of experimental toxicology are much better defmed now than 10 or 20

years ago. Drug safety in clinical trials has become a shared responsibility of toxicologists and clinical pharmacologists. Thanks to their collaboration and their

continued endeavour to sharpen their tools, clinical trials ofnew drugs have become

a responsible and justifiable medical activity.

References

Baeder, C. (1980). Non-human primates und ihre Bedeutung für die chronischen

Toxizitätsstudien . In Zur Problematik von chronischen Toxizit ätspriifungen, ed.

Schnieders, B. & Grosdanoff, P., pp. 29-32 . Berlin: AMI Berichte. Dietrich Reimer Verlag.

Balazs, T. (1974). Development oftissue resistance to toxic effectsofchemicals. Toxicology, 2,

247-255.

Fletcher, A. P. (1978). Drug safety tests and subsequent clinical experience. J. Roy. Soc. Med. ,

71,693-687.

Gelzer, J. (1979). Governmental toxicology regulations: an encumbrance to drug research?Arch.

Toxicol., 43, 19-26.

Griffm, J. P. (1979). Animal toxicological studies in the safety evaluation of new drug

substances. In Drug Assessment. Criteria and Methods, ed. Bowers, J. Z. & Velo, G. P., pp.

107-119. Amsterdam, New York, Oxford: Elsevier/North Holland Biomedical Press.

Imai, K. & Hayashi, Y. (\970). Steroid-induced intrahepatic cho1estasis in mice. Japan. J.

Pharmac., 20,473-481.

MitchelI, J. R. & Jollow, D. J. (\974). Metabolie activation of acetaminophen , furosemide, and

activation of acetaminophen, furosemide, and isoniazid to hepatotoxic substances. In Drug

lnteractions, ed. Morselli, P. L., Garattini, S. &Cohen, S. N., pp. 65-79 . New York: Raven

Press.

Plaa, G. L. (19 78). The problems oflow-incidence response. In Proc. First ln t. Congr. Toxicol.

ed. Plaa, G. L. & Duncan, W. A. M., pp. 207-219. New York, San Francisco, London:

Academic Press.

Poilmann, W. (1980). Pharmakokinetik und Toxizitätversuche. In Zur Problematik von

chronischen Toxizlidtspr üfungen. ed. Schnieders, B. & Grosdanoff, P., pp. 49-51. Berlin:

AMI Berichte. Dietrich Reimer Verlag.

Zbinden, G. (1966). Toxicology of new drugs. In The Handbook 01 Biochem istry and

Biophysics, ed. Damm , H. c.,Besch, P. K. & Goldwyn, A. J., pp. 459-511.Cleveland and

New York: The World Publishing Company.

Zbinden, G. (1979). Application ofbasic concepts to research in toxicology. Pharmac. Rev., 30,

605-616.

Zbinden, G. & Gross, F. (1979). Pharmacological Methods in Toxicology. Oxford, New York,

Toronto, Sydney, Paris, Frankfurt: Pergamon Press.

Zbinden, G. & Schlatter, C. (1978). New approaches to mutagenicity and carcinogenicity testing

in in vivo mammalian systems. In Chemical Toxicology 01Food, ed. Galli, C. L., Paoletti,

R. & Vetorazzi, G., pp. 153-166.

Zbinden, G. & Studer, A. (1959). Experimental pathology of iproniazid and related compounds.

Ann. N. Y. Acad. Sei., 80, 873-884 .

THEPHARMACOLOGY

AND CLINICAL POTENTIAL

OF PROSTACYCLIN

S.MONCADA& J. R. VANE

Wellcome Research Laboratories

Langley Court. Beckenham, Kent BR3 3BS, England

Smith & Willis (1970) showed that aggregating platelets release, among other products, prostaglandins of the E and F type. Shortly afterwards it was shown that

aspirin-like drugs inhibit prostagiandin biosynthesis (Vane, 1971; Ferreira, Moncada

& Vane, 1971; Smith & Willis, 1971) and as a result the general theory was proposed

(Vane, 1971) that this enzyme inhibition accounts for the anti-inflammatory effects

(and perhaps the side effects)of aspirin-like drugs. Since then, substantial evidence to

support this theory has accumulated and it is now widely accepted that the analgesic,

anti-pyretic and anti-inflammatory effects of aspirin-like drugs are mediated via

inhibition ofprostaglandin biosynthesis (Vane, 1976; Moncada & Vane, 1979a).

It has been known for many years that aspirin inhibits the platelet release reaction

induced by ADP and collagen in vitro (O'Brien, 1968; Wciss, Aledort & Kochwa,

1968). Additionally, aspirin prolongs the bleeding time (Quick, 1966) and this effect

occurs at low concentrations that are achieved in plasma after an oral therapeutic

dose. Aspirin also inhibits prostagiandin production in platelets after oral administration (Smith & Willis, 1971). Because ofthis Smith & Willis (1971) suggested that the

inhibition of prostagiandin biosynthesis could account for the inhibitory effect of

aspirin on platelet aggregation. At that time, however, the proposition was difficult to

support for the known prostaglandins did not induce platelet aggregation. A further

important step in the knowledge of arachidonic acid metabolism was achieved with

the isolation of the prostagiandin endoperoxide intermediates and thromboxane A2

(TXA2) (Hamberg & Samuelsson, 1973; Nugteren & Hazelhof, 1973; Hamberg

Svensson, Wakabayashi & Samuelsson, 1974; Hamberg, Svensson & Samuelsson,

1975).

The release of prostagiandin endoperoxides during platelet aggregation was

demonstrated by Hamberg et al. (1974) and Smith, Ingerman, Kocsis & Silver (1974).

Prostagiandin endoperoxides (PGG2 and PGH2) are chemically unstable in aqueous

solution (half-life approximately 5 min at 37'C) and isomerize to stable prostaglandins.

In platelets, the prostagiandin endoperoxides are further metabolized into an even

more unstable compound, thromboxane A2 , by an enzyme which has been

designated thromboxane synthetase (Needleman, Moncada , Bunting, Vane,

Hamberg & Samuelsson, 1976). Thromboxane A2 is more potent than the parent

endoperoxides as an inducer ofplatelet aggregation (Hamberg et al., 1975; Moncada

16 S. MON CADA & J. R. VANE

& Vane, 1977), and as a constrictor of arterial smooth muscle in vitro (Bunting,

Moncada & Vane, 1976b; Hamberg et al., 1975), and in vivo in the hind limb and

mesenteric vascular beds of the dog (Dusting, Moncada & Vane, 1978b). Thromboxane A2 degrades chemically (half-life 30s) (Hamberg et al.. 1975)to the relatively

inert substance thromboxane B2. Thromboxane A2is, in fact, responsible for most of

the activity of a substance discovered by Piper & Vane (1969) to be released during

anaphylactic shock in guinea-pig isolated lungs (Bunting et al., 1976b;Hamberg et al.,

1975). Because ofthe activity by which they detected it, they called it 'rabbit aorta

contracting substance' or 'RCS'. Its release was inhibited by aspirin-like drugs (Piper

& Vane, 1969). In platelets the prostagiandin endoperoxides generated during aggregation most probably exert their aggregating effect through conversion to thromboxane A2 (Hamberg et al., 1975; Moncada & Vane, 1977).

In contrast to cyclo-oxygenase, thromboxane synthetase is poorly inhibited by

indomethacin (Moncada, NeedlemanBunting & Vane, 1976c). Selective inhibition

of thromboxane synthetase has been attempted using benzydamine (Moncada et al.,

1976c), a phenyl phosphonate derivative of phloretin phosphate designated as

N-0164 (Kulkarni & Eakins, 1976);the compound 1'-(isopropyl-2-indolyl)-3-pyridyl3-ketone, known as L 8027 (Gryglewski, Zmuda, Korbut, Krecioch & Bieron, 1977);

imidazole (Moncada , Bunting, Mullane, Thorogood, Vane, Raz & Needleman,

1977a; Needleman, Bryan, Wyche, Bronson, Eakins, Ferrendelli & Minkes, 1977)

and more potently by some of its analogues (Blackwell, Flower, Russell-Smith,

Salmon, Thorogood & Vane, 1978; Tai & Yuan , 1978). In addition, endoperoxide

analogues are inhibitors ofthromboxane synthetase (Gorman, Bundy, Peterson, Sun,

Miller & Fitzpatrick, 1977a).

The involvement of endogenous cyclic endoperoxides (and thromboxane A2) in

platelet aggregation and the prevention oftheir formation by inhibition ofthe cyclooxygenase by aspirin provides a satisfactory explanation for the inhibitory effects of

aspirin-like drugs on platelet aggregation and the release reaction (Hamberg &

Samuelsson , 1974).

In 1976, whilst the conversion of cyclic endoperoxides by different tissues was

being studied (Moncada, Gryglewski, Bunting & Vane, 1976a), it was found that a

microsomal preparation ofblood vessels converted prostagiandin endoperoxides into

an unstable product which, in contrast to thromboxane A2,was a potent inhibitor of

platelet aggregation. It was called PGX. The chemica l structure ofthis substance was

elucidated (Johnson, Morton, Kinner, Gorman, McGuire, Sun, Whittaker, Bunting,

Salmon , Moncada & Vane, 1976)and PGX was then renamed prostacyclin and given

the abbreviation ofPGh.

Prostacyclin is the most potent endogenous inhibitor of platelet aggregation yet

discovered, being 30-40 times more potent than PGEI (Moncada & Vane, 1977). In

vivo, in the microcirculation of the hamster cheek pouch (Higgs, Higgs, Moncada &

Vane, 1978), prostacyclin applied locally in low concentrations inhibits thrombus

formation due to ADP and given systemically to the rabbit it prevents e1ectricallyinduced thrombus formation in the carotid artery and increases bleeding time

(Ubatuba, Moncada & Vane, 1979). These effects in vivo disappear within 30 min of

dosing. A property ofprostacyclin which has potential therapeutic importance is that

it disaggregates platelets in vitro (Moncada et al., 1976b), in vivo (Ubatuba et al., 1979)

and in extracorporeal circulations where platelet clumps have formed on collagen

strips (Gryglewski, Korbut & Ocetkiewicz, 1978a; Gryglewski, Korbut, Ocetkiewicz

&Stachwa, 1978c).

Prostacyclin is unstable at neutral pH, breaking down to the relatively inert

substance 6-oxo-PGFlu. Its activity disappears within 15s on boiling or within 10

min at 22'C. In blood at 37"C prostacyclin has a half-life of 2-3 min (Dusting,

Moncada & Vane, 1977a). Alkaline pH increases the stability ofPGh (Johnson et al.,

1976;Cho & Allen, 1978)so that at pH 1O-11, it is virtually stable. .

PROSTACYCLlN, PHARMACOLOGY ANDCLlNICAL POTENTIAL 17

Generation of prostacyclin is an active mechanism by which the vessel wall could

be protected from deposition of platelet aggregates. Thus, prostacyclin formation

provides a comprehensive explanation of the long recognised fact that contact with

healthy vascular endothelium is not a stimulus for platelet clumping, An imbalance

between formation of prostacyclin and TXA2 could be of dramatic consequence.

Vascular damage leads to platelet adhesion but not necessarily to thrombus formation. When the injury is minor, platelet thrombi are formed which break away from

the vessel wall and are washed away by the circulation. The degree of injury is an

important determinant, and there is general agreement that for the development of

thrombosis, severe damage or physical detachment of the endothelium must occur .

All these observations are in accord with the differential distribution of prostacyclin

synthetase across the vessel wall, decreasing in concentration from the intima to the

adventitia. Moreover, the pro-aggregating elements increase from the subendothelium to the adventitia. These two opposing tendencies render the endothelial

lining anti-aggregatory and the outer layers of the vessel wall thrombogenic

(Moncada, Herman, Higgs & Vane, 1977b).

The ability ofthe vascular wall actively to prevent aggregation has been postulated

before (Saba & Mason, 1974). For instance, the presence ofan ADP-ase in the vessel

wall has led to the suggestion that this enzyme, by breaking down ADP, limits platelet aggregation (Heyns, van den Berg, Potgieter & Retief, 1974; Lieberman, Lewis &

Peters, 1977). The presence of an ADP-ase in the vessel wall has been confirmed in

this laboratory. However, the anti-aggregating activity of the vessel wall is mainly

related to the release of prostacyclin, for 15-HPAA or 13-hydroperoxylinoleic acid

(l3-HPLA), two inhibitors ofprostacyclin formation which have no activity on ADPase, abolish most if not all of the anti-aggregatory activity of vascular endothelial

cells (Bunting, Moncada & Vane, 1977). Similar results have been obtained using an

antiserum which cross-reacts with and neutralises prostacyclin in vitro (Bunting,

Moncada, Reed, Salmon & Vane, 1978). Endothelial cells pretreated with this

antiserum lose the ability to inhibit ADP-induced aggregation (Bunting et al.. 1978;

Christofinis, Moncada, Bunting &Vane, 1979).

It is not yet clear whether or not prostacyclin is responsible for all the thromboresistant properties of the vascular endothelium. However, Czervionke, Smith, Fry

& Hoak (1979) using endothelial cell cultures have demonstrated that platelet

adherence in the presence ofthrombin increases from 4% to 44% after treatment with

1mM aspirin . This increase was parallelIed by a decrease in 6-oxo-PGF, a formation

from 107 nM to < 3nM and could be reversed by addition of 25 nM of exogenous

PGI2. This work clearly supports the thesis that prostacyclin plays an important role

in the prevention of deposition of platelet aggregates but may not be responsible for

all the thromboresistant properties ofvascular endothelium.

Prostacyclin inhibits platelet aggregation (platelet-platelet interaction) at much

lower concentrations than those needed to inhibit adhesion (platelet-collagen

interaction) (Higgs, Moncada, Vane, Caen, Michel & Tobelern, 1978). This suggests

that prostacyclin does not prevent platelets from sticking to vascular tissue and

interacting with it, but does limit or prevent thrombus forrnation . Certainly, platelets

adhering to a site where prostacyclin synthetase is present could weil feed the enzyme

with endoperoxide, thereby producing prostacyclin and preventing other platelets

from clumping onto the adhering platelets, limiting the cells to a monolayer.

Recently, Weiss &Turitto (1979)have observed some degree ofinhibition ofplateletsubendothelium interactions with low concentrations of prostacyclin at high shear

rates, but at none of the concentrations used could they observe total inhibition of

platelet adhesion.

It is also possible that formed elements of blood such as the white cells, which

produce endoperoxides and TXA2 (Davison, Ford-Hutchinson, Smith & Walker,

1978;Goldstein, Malmsten, Kaplan, Kindahl, Samuelsson & Weissman , 1977; Higgs,

18 s.MONCADA & J. R. VANE

Bunting, Moncada & Vane, 1976) interact with th e vessel wall and thereby promote

the formation ofprostacyclin, as do the platelets.

Prostacyc1in, thromboxane A2- thrombosis and haemostasis

Pro stagiandin endoperoxides are at the crossroads of arachidonic acid metabolism,

for they are precursors ofsubstances with opposing biological properties. On the one

hand, TXA2produced by the platelets is a strong contractor oflarge blood vessels and

induces platelet aggregation, On the other hand, prostacyclin produced by the vessel

wall is a strong vasodilator and the most potent inhibitor of platelet aggregation

known . Each substance has opposing elTects on cyclic AM P conce ntrations in

platelets (Moncada & Vane, 1979b), thereby giving a bala nced contro l mechanism

which will, therefore, alTect thrombus and haemostat ic plug formation. Selective

inh ibition of the formation of TXA2 should lead to an increased bleeding tim e and

inhibition of thrombus formation , whereas inhibition of prostacycl in formation

should be propitious for a ' pro-thrombotic state'. Th e amount of contro l exerted by

th is system can be tested, for selective inhibitors of each pathway have been described

(Moncada & Vane , 1977; Nijkamp, Moncada, Wh ite & Vane, 1977).