CHIRALITY 4484-487 (1992)

Enantioselective Pharmacodynamics of the Nonsteroidal Antiinflammatory Drug Ketoprof en: In Vitro Inhibition of Human Platelet Cyclooxygenase Activity PETER J. HAYBALL, ROGER L. NATION, AND FELIX BOCHNER Phamcy Department, Repatriation General Hospital, Daw Park, South Australia (PJfiJ, Department of Clinical and Experimental Pharmacology, University of Adelaide, South Australia, (P.J.H., F.B.), and School of Pharmacy, University of South Australia (RLN.), South Australia, Australia

ABSTRACT The pharmacological activity of ketoprofen enantiomers was investigated in humans by an in vitro method. The antiplatelet effect of ketoprofen was assessed by measuring the inhibition of platelet thromboxane Bz (TXB2) generation during the controlled clotting of whole blood obtained from each of four healthy volunteers. Ketoprofen was added separately to whole blood as a range of concentrations of (1)predominantly (S)-ketoprofen,(2) racemic ketoprofen, and (3)predominantly (R)-ketoprofen.(S)-Ketoprofen was found to be solely active at inhibiting human platelet TXBz production; (R)-ketoprofenwas devoid of such activity and did not modify the potency of its optical antipode. A relationship between the percentage inhibition of TXBz generation and the unbound concentration of (S)-ketoprofenin serum was modelled according to a sigmoidalEmaxequation. The mean ( fSD) serum unbound concentration of (S)-ketoprofenrequired to inhibit platelet TXBz generation by 50% (ECS0)was 0.320 ( f0.062) ng/ml. This value for ketoprofen is considerably lower than previously reported values for (S)-ibuprofenand (S)-naproxen. o 1992 Wiley-Liss, Inc.

KEY WORDS: ketoprofen enantiomers,thromboxane,in vitro activity, 2-arylpropanoicacids, sigmoidal &,, modelling INTRODUCTION

(RS)-ketoprofen [(RS)-2-(3'-benzoylphenyl)propanoicacid] is used clinically as an antiinflammatory and analgesic agent. In common with most congeners of the 2-arylpropanoicacid class of nonsteroidal antiinflammatory drugs (NSAIDs), ketoprofen is marketed as a racemic compound. While a number of mechanisms have been proposed to explain the pharmacological effects of NSAIDs, it is believed that inhibition of prostanoid biosynthesis is of the greatest importance. Specifically, NSAIDs act as competitive inhibitors of the cyclooxygenase subunit of prostaglandin synthetase. Moreover, in every case examined, inhibition of prostaglandin biosynthesis by chiral 2-arylpropanoicacids has been highly enantioselective with major or exclusive activity residing with enantiomers of (S)-config~ration.~-" To date, no detailed analysis of the potential enantioselective pharmacodynamics of ketoprofen has been performed. Moreno et aL5 have reported in abstract form the effect of ketoprofen enantiomers on arachidonic acid-induced rabbit platelet aggregation. These data were largely qualitative having been carried out at a single drug concentration(10 -5 M); the (S)-enantiomer inhibited platelet aggregation by 83% in contrast to 16% inhibition reported for its optical antipode. In addition, the optical purity of the ketoprofen enantiomers was not stated. The present study was designed to examine, for the first time in man, the effect of ketoprofen enantiomers on platelet cyclooxygenase.The amount of thromboxane A2 (TXAz)generated during the controlled clotting of whole blood was used o 1992 Wiley-Liss, Inc.

as an index of cyclooxygenase activity. The study demonstrated that the ability to inhibit prostaglandin biosynthesis resides exclusively with the (S)-enantiomer of ketoprofen. Moreover, a relationship was established between the degree of inhibition of TXA2 formation (drug effect) and the unbound serum concentration of (S)-ketoprofen. MATERIALS AND METHODS Chemicals (RS)-Ketoprofen was purchased from Sigma Chemical Company (St Louis, MO). The (R)- and (S)-enantiomersof ketoprofen were gifts of Dr Kathy Knights (Flinders University, Bedford Park, South Australia) and Dr Ralph Massy-Westropp (University of Adelaide, Adelaide, South Australia), respectively. The optical purity of these ketoprofen enantiomers was determined by an indirect enantioselective high-performance liquid chromatographic (HPLC) method. l2 The optical purity of the HPLC chiral derivatizing reagent [(S)-1-phenylethylamine, Sigma Chem. Co., Lot: 10H3457l was determined by an extension of a lH NMR spectroscopy method l3(kindly performed by Dr David P.G. Hamon, unpublished data). The percentage optical purity of this chiral derivatizing amine used for assessing the optical purity of the enantiomeric ketoprofen substances was estimated to be in the order of 99.5%. Following replicate Received for publication February 7, 1992; accepted June 26, 1992. Address reprint requests to Peter J. Hayball, Pharmacy Department, Repahiation General Hospital, Daw Park, South Australia, 5041, Australia.

PHARMACODYNAMICSOF KETOPROFEN ENANTIOMERS

485

methodology in detail elsewherel4 where we had also established that the plasma protein binding of both stereoisomers was independent of concentration over the range used in these experiments and nonenantioselective.The blood to serum concentration ratio of (S)-ketoprofen([B], /[S],) was calculated as the ratio of the measured concentration of (S)-ketoprofenin Assessment o f Platelet CyclooxygenaseInhibition serum derived from spiking harvested serum with 2.0 pg/ml of To monitor the effects of ketoprofen enantiomers on platelet (RS)-ketoprofen to the serum concentration of (S)-ketoprofen cyclooxygenase, whole blood was allowed to clot under con- subsequent to spiking whole blood with 2.0 pg/ml of (RS)trolled conditions. As an index of cyclooxygenase activity, the ketoprofen. Pilot studies revealed that the partitioning of amount of TXA, generated was assessed by measuring the (S)-ketoprofenbetween blood and serum was independent of concentration of its stable breakdown product, thromboxane drug concentration and that equilibration occurred within 15 Bz (TXB,), in the harvested serum as described by Patron0 et min. Consequently, individual subject estimates of PI, /[S], al. l5 Blood (20 ml) was collected by venepuncture from each of were obtained in duplicate following incubations for 1 h at four healthy volunteers, none of whom was taking any medica- 37°C. Concentrations of total (bound plus unbound) (S)-ketotion. The mean ( f SD) age of the subjects (2 female and 2 male) profen were determined by an enantioselective HPLC assay was 34.5 (f13.7) years. For a given human subject, 1 ml of detailed previously.l2 The concentration of unbound (S)-ketoblood was immediately transferred (needleremoved) to each of profen in serum following spiking of whole blood with drug was calculated [Eq. (l)]. a series of pre-calibrated PyrexR medium-walled borosilicate (Conc(s).blood) culture tubes (10 x 75 mm) containing a range of quantities (vide znfra) of (R)-ketoprofen and/or (S)-ketoprofen.Each tube, which also contained 50 1 of sodium chloride 0.9% injection B.P., had been prewarmed to 37°C in a dry heat bath. The tube Pharmacodynamic Analysis o f T m z Data contents was gently mixed and returned to the heat bath and For each subject, the control serum concentration of TXB, maintained at 37°C for 1 h. After the 1h incubation, the tube was rimmed with a glass rod to detach any unretracted por- was determined by averaging the TXB, concentration in the tions of the clot and centrifuged at 2000g (10 min). The har- two blood samples containing no ketoprofen. For samples convested serum was stored at - 20°C until analysed for TXB, taining ketoprofen, inhibition of TXB, production during whole blood clotting was calculated as the percentage decrease using a radioimmunoassay method. l6 The final blood concentrations of (R)-ketoprofen,(S)-keto- in the serum concentration of TXB,, relative to the control profen, and (RS)-ketoprofen for each subject were as follows: concentration. For each subject, data from each of three apcontrol samples (n = 2), containing no ketoprofen; (R)-keto- proaches for spiking various amounts of (S)-ketoprofeninto profen at 0.01, 0.1, 0.5, 1.0, 5, 10, and 20 pg/ml; (S)-ketoprofen blood [as predominantly (S)-ketoprofen,as (RS)-ketoprofen,and at 0.005, 0.01, 0.05, 0.1, 0.5, 1.0, and 5.0 pg/ml; and (RS)-keto- as predominantly (R)-ketoprofen]were pooled. This was done profen at 0.01, 0.02, 0.1, 0.2, 1.0, and 2.0 pg/ml. The blood since it was evident that the logarithmic concentration/reconcentration of (S)-ketoprofenin each tube containing racemic sponse data for (S)-ketoprofen,for each of the three spiking drug was 50% (by definition) of the stated (RS)-ketoprofen approaches, were superimposed. Subsequently, the relationconcentration value. In those tubes containing (R)-ketoprofen ship between the unbound serum concentration of (S)-ketoalone, the blood concentration of (S)-ketoprofen (present as profen and the percentage inhibition of TXB, generation was equation to the optical impurity) was derived by transforming the (R)-keto- examined by fitting a standard sigmoidal ha, profen value by the fractional impurity value (4.30%, vide data with an extended least-squares modelling computer prosupra). Thus (S)-ketoprofenwas spiked into blood in three gram (Multifit 2.0, Day Computing, Cambridge, U.K.). The sigfashions: (1) as the predominant enantiomer, (2) in equal pro- moidal Emax equation is portion to its optical antipode, and (3) as the minor enantiomer. ha,x Conc" E = +Eb Derivation of Unbound (S)-Ketoprofen Concentration ECN " + Conc" in Serum where E is the measured effect at drug concentration Conc; The unbound (S)-ketoprofenserum concentration (Conqs). is the basal effect in the absence of drug; ha,is the maximal was derived for each of the spiked blood samples by effect; ECW represents the concentration of drug required to performing separate in vitro drug binding and blood/serum cause 50% of Emax;and n is the steepness factor for the log partitioning experiments with each study subject's blood and concentration-effect relationship. l7 In the present case E was serum. The COnc(S).unbound values could not be measured di- expressed as a percentage of the maximal possible effect of the rectly in the serum samples used for TXB, analysis as these drug, and hence 4 was set at zero and ha, at 100, thereby fell below the assay limits of quantification of the unbound simplifying Eq. (2) to species.l 4 The fraction unbound of (S)-ketoprofen (fu,) was determined as the arithmetic average of four values over a YO Inhibition = loo (ConC(S)-unbound ) ' . (3) spiked serum concentration range of (RS)-ketoprofen of Ec50 " + (ConC(S)-unbound " 2.0-12.0 pg/ml for each individual subject. In addition, we confirmed that the fraction unbound of (S)-ketoprofenwas con- Equation (3) was applied individually for each subject after stant over this concentration range (analysis of variance, pooling data derived from the three approaches to spiking P< 0.05). We have described the enantioselective binding (S)-ketoprofen into blood. Several weighting schemes were ex-

analysis (n = 3), the mean ( f SD) optical purity of (R)-ketoprofen was 95.7 f 0.06% and of (S)-ketoprofen99.0 f 0.06%. Suppliers of chemicals used in the experimentswhen the serum protein binding of (S)-ketoprofenwas determined have been cited previously. l4

486

HAYBALL ET AL.

plored with the least-squares analysis, the most appropriate profen leading to a 50% inhibition of TxB2 generation was being the reciprocal of YOTxB2 inhibition (i.e., 1/% inhibition). found to be 0.320 ng/ml (Table 1). The corresponding slope factor of the logarithmicconcentration-effectcurve (n)was 1.52. RESULTS AND DISCUSSION The optical purity data for the ketoprofen enantiomers were The effect of (S)-ketoprofen(expressed as pharmacologically obtained by an indirect WLC method l2 following chromatogactive unbound drug) on in vitro TXB2 generation during the raphy of the (S)-1-phenylethylamidesof (R)- and (S)-ketoprofen. controlled clotting of whole blood for each subject is depicted These data were based on the assumption that the derivatizain Figure 1. When (RS)-ketoprofen and (R)-ketoprofen were tion reagent was optically pure. We estimated (S)-1-phenyleadded to blood to assess their effect on platelet TXB2 produc- thylamine to contain approximately 0.5% of its optical antion, no effect was attributed to the presence of (R)-ketoprofen. tipode. Thus derived serum concentrations of (S)-ketoprofen, In contrast, (S)-ketoprofen present either alone, as racemic when spiked into blood as the 4.3%impurity in (R)-ketoprofen, drug, or as a minor enantiomeric impurity elicited a concentra- may be subject to a small error (approximately 10%).The net tion-dependent inhibition of TXB2 production by platelets. result on the sigmoidal &,, model parameter estimates (obThis effect was not modified by the presence of (R)-ketoprofen tained by combining data from all three spiking approaches) (Fig. 1). It was found, for a given subject, that the concentra- would be expected to be minor. These data clearly demonstrate the enantiospecificity of intion-response curves [expressed in terms of unbound (S)-ketoprofen in serum] constructed for each of the three ketoprofen hibition of platelet cyclooxygenase-mediatedTXB2 generation, spiking approaches were identical. Subsequently, to improve elicited by ketoprofen in vitro. Further, (R)-ketoprofendoes not the precision of each subject's model parameter estimates, data modify the activity of the active (S)-enantiomer, even when derived from adding (S)-ketoprofento blood as the predomi- present as the predominant enantiomeric species. A similar nant enantiomer,as racemic drug and as the minor enantiomer, study conducted with ibuprofen3 demonstrated that (R)-ibuwere pooled. When the effect of (S)-ketoprofenon TXB2 genera- profen did not influencethe in vitro activity of its phannacologtion was modelled, according to a sigmoidal &,, equation, the ically active optical antipode when both enantiomers were prearithmetic mean unbound serum concentration of (S)-keto- sent as racemic drug. We have previously shown that

Subject 1

z 9 I-

2

P I-

dW

d W

ma

Z

W

a N

p1

m

m X I-

100

-

80

-

60-

X

t-

z 2 t I

Subject 2

1

40

-

20-

I

s *

.01 .1 1 10 100 CONC UNBOUND S-KETOPROFEN (nglml)

0,-

.01

z n

1

dW

z W

a

- ....

-7

3

.l

..

1

-

,

. .-.--I

10

100 1000 CONC UNBOUND S-KETOPROFEN (ng/ml)

Subject 4

7 80

er

3

0 . .01

......... ........ ....... . . . .1

I

1

1

1

10

100

CONC UNBOUND S-KETOPROFEN (nglml)

Fig. 1. In vitro relationships between the percentage inhibition of TXBz generation and logarithmic serum concentration of unbound (S)-ketoprofen for study subjects 1 to 4. The symbols are actual data points when (S)-ketoprofen was added to blood as 99.0% optically pure drug (0). as racemic drug (O),and

.01

.1

1

10

100

CONC UNBOUND S-KETOPROFEN (nglml)

as 4.3%optical impurity together with its antipode (A). The line represents the predicted relationship,according to a sigmoidal Em,, model, from least squares regression analysis.

PHARMACODYNAMICS OF KETOPROFEN ENANTIOMERS

TABLE 1. Fraction unbound in serum (fus)and the blood to serum concentration ratio (p],/[S],) of (S)-ketoprofenin each subject together with computer generated sigmoidal Em, model parameters describing the relationship between unbound serum concentration of (S)-ketoprofenand percentage inhibition of TXB, generation (gender)

of (S)-1-phenylethylamine. In addition, the authors wish to thank Dr John V. Lloyd and Ms Elizabeth Duncan (Department of Haematology, Institute of Medical and Veterinary Science, Adelaide, South Australia) for advice and technical assistance with the serum TXB2 assays and Miss Donna Lapins for secretarial assistance.

ECJSE)"

Age

Subject

487

fu,

Pk/M

(ng/ml)

n (SE)

0.00745 0.00770 0.00993 0.00782

0.578 0.578 0.625

0.588

0.00823 0.00115

0.592 0.022

0.331 (0.032) 0.402 (0.059) 0.286 (0.029) 0.262 (0.038) 0.320 0.062

1.93 (0.18) 1.25 (0.18) 1.48 (0.16) 1.41 (0.19) 1.52 0.29

"SE, standard error of the model parameter estimate.

(R)-ketoprofen does not displace (S)-ketoprofen from plasma protein binding sites at the drug concentrations used in the present study. l 4 Thus, it is unlikely that indirect dispositional effects can be attributed to the (R)-enantiomerin this in vitro pharmacological test system. A number of studies have reported some minor activity of (R)-enantiomers of 2-arylpropanoate NSAIDs (see Table 3 in Evans l) in in vitro systems examining inhibition of prostaglandin synthesis and platelet aggregation. However, it is unclear in the majority of these cases as to whether such effects were due to the presence of some (S)-enantiomerimpurity in the respective (R)-enantiomer test compound. These data demonstrate the sensitivity of human platelet cyclooxygenase to (S)-ketoprofen.By comparison with recent studies with (S)-ibuprofen3and (S)-naproxen,l8 where pharmacologically active unbound drug was examined with an identical in vitro test system, ketoprofen [based on EC50 of unbound molar concentration of (S)-enantiomer]was approximately 40-fold more potent than (S)-ibuprofen and 100-fold more potent than (S)-naproxen. Caution must be taken when extrapolating these data for ketoprofen into the clinical setting. Significantpharmacological activity of the (R)-enantiomerof 2-arylpropanoateNSAIDs can arise indirectly from metabolic chiral inversion. l9 In addition, antiinflammatory effects may, in part, be mediated by processes independent of cyclooxygenase inhibition with potentially different stereochemical determinants.20 However, based on estimates of (S)-ketoprofenconcentrations achieved with chronic dosing of racemic drug in humans21 and extrapolating unbound drug concentrations from in vitro binding studies, l4 it would appear that the thromboxane-related antiplatelet effects of ketoprofen would persist throughout the dosing interval. ACKNOWLEDGMENTS

This study was supported by a research grant from the Central Health and Medical Research Council of the Department of Veteran Affairs of Australia. The authors are indebted to Dr. David P.G. Hamon (Department of Organic Chemistry, University of Adelaide, Adelaide, South Australia) for perfonning the 'H NMR spectroscopicassessment of the optical purity

LITERATURE CITED 1. Evans, A.M. Enantioselective pharmacodynamics and pharmacokinetics of chiral non-steroidal anti-inflammatory drugs. Eur. J. Clin. Pharmacol. 4 2 237-256,1992. 2. Vane, J.R. Inhibition of prostaglandin synthesis as a mechanism of action for aspirin-like drugs. Nature (London) 231:232-235, 1971. 3. Evans, A.M., Nation, R.L., Sansom, L.N., Bochner, F., Somogyi, A.A. Effect of racemic ibuprofen dose on the magnitude and duration of platelet cyclooxygenase inhibition: Relationship between inhibition of thromboxane synthesis and the plasma unbound concentration of 3 +)-ibuprofen. Br. J. Clin. Pharmacol. 31: 131-138, 1991. 4. Cerletti, C., Manarini, S., Colombo, M., Tavani, A. The ( + )-enantiomer is responsible for the antiplatelet and anti-inflammatory activity of ( *)-indobufen. J. Pharm. Pharmacol. 42:885-887, 1990. 5. Moreno, JJ., Calvo, L., Femandez, F., Carganico, G., Bastida, F., Bujons, J., Messeguer, A. Biological activity of ketoprofen and its optical isomers. Eur. J. Pharmacol. 183:2263-2264,1990 (abstract). 6. Rubin, A., Knadler, M.P., Ho, P.P.K., Bechtol, L.D., Wolen, R.L. Stereoselective inversion of (R)-fenoprofen to (S)-fenoprofen in humans. J. Pharm. Sci. 7482-85.1985, 7. Buttinoni, A,, Ferrari, M., Colombo, M., Ceserani, R. Biological activity of indoprofen and its optical isomers. J. Pharm. Pharmacol. 35:603-604,1983. 8. Adams, S.S., Bresloff, P., Mason, C.G. Pharmacological differences between the optical isomers of ibuprofen: evidence for metabolic inversion of the (-)-isomer. J. Pharm. Pharmacol. 28256-257, 1976. 9. Ku, E.C., Wasvary, J.M. Inhibition of prostaglandin synthase by pirprofen. Studies with sheep seminal vesicle enzyme. Biochim. Biophys. Acta 384: 360-368, 1975. 10. Gaut, Z.N., Bamth, H.. Randall, L.O., Ashley, C., Paulsrud, J.R. Stereoisomeric relationships amoung anti-inflammatory activity, inhibition of platelet aggregation, and inhibition of prostaglandin synthetase. Prostaglandins 1059-66, 1975. 11. Nishizawa, E.E., Wynalda, DJ. Suydam, D.E., Molony, B.A. Flurbiprofen, a new potent inhibitor of platelet aggregation. Throm. Res. 3577-588, 1973. 12. Hayball, P.J., Nation, R.L., Bochner, F., Le Leu, R.K. Enantiospecific analysis of ketoprofen in plasma by high-performance liquid chromatography. J. Chromatogr. 570 446-452, 1991. 13. Whitesides, G.M., Lewis, D.W. Tris [3-(ferf-butylhydroxymethylene)-d-camphorato]europium (111). A reagent for determining enantiomeric purity. J. Am. Chem. Soc. 926979-6980,1970. 14. Hayball, PJ., Nation, R.L., Bochner, F., Newton, J.L., Massy-Westropp, R.A., Hamon, D.P.G. Plasma protein binding of ketoprofen enantiomers in man: Method development and its application. Chirality 3460-466, 1991. 15. Patrono, C., Ciabattoni, G., Pinca, E., Pugliese, F., Castrucci, G., DeSalvo, A,, Satta, M.A., Peskar, B.A. Low dose aspirin and inhibition of thromboxane B, production in healthy subjects. Thromb. Res. 17317-327, 1980. 16. Fitzpatrick, F.A. A radioimmunoassay for thromboxane B,. Methods Enzymol. 86:286-296,1982. 17. Holford, N.H.G., Sheiner, L.B. Understanding the dose-effect relationship: clinical application of pharmacokinetic-pharmacodynamicmodels. Clin. Pharmacokinet. 6429-453, 1981. 18. Williams, K.M. Personal communication, 1992. 19. Hutt, A.J., Caldwell, J. The metabolic chiral inversion of 2-arylpropionic acids-a novel route with pharmacological consequences.J. Pharm. Pharmacol. 35:693-704, 1983. 20. Abrarnson, S.B., Weissmann, G. The mechanism of action of nonsteroidal antiinflammatory drugs. Arthr. Rheum. 32 1-9, 1989. 21. Jamali, F., Brocks, D.R. Clinical pharmacokinetics of ketoprofen and its enantiomers. Clin. Pharmacokinet. 19 197-217, 1990.