pharmacokinetics and pharmacodynamics of the endothelin-receptor antagonist bosentan in healthy...

l?HARMACOI~NETICS AND DRUG DISPOSITION

Pharmacokinetics and pharmacodynamics of the endothelin-receptor antagonist bosentan in healthy human subjects

Intvodmtion: Bosentan (Ro 47-0203) is a potent and mixed ET,- and ET,-receptor antagonist. Its activity has been studied in a variety of preclinical disease models. Methods: Two double-blind placebo-controlled studies were performed to investigate the pharmacokinetics and pharmacodynamics of bosentan after single oral and intravenous doses in healthy volunteers; doses of 3, 10,30,100,300,600,1200, and 2400 mg were given in a single ascending oral dose study, and doses of 10, 50, 250, 500, and 750 mg were given in a single ascending intravenous dose study (six subjects received active drug and two received placebo at each dose level). In an open-label crossover added to the second study, six subjects received a single oral dose of 600 mg and a single intravenous dose of 250 mg in randomized order. At regular intervals, blood pressure, pulse rate, and skin responses to intradermally injected endothelin-1 (ET-l) were recorded, and plasma levels of ET-l, proendothelin-1 (big ET-l), and ET-3, and drug and urinary levels of ET-l and drug were determined. Results: Systemic plasma clearance and volume of distribution decreased with increasing dose to limiting values of around 6 L/hr and 0.2 L/kg, respectively. The absolute bioavailability was 50% and appeared to decrease with doses above 600 mg. Plasma ET-l increased maximally twofold (oral) and threefold (intravenous), and this increase was directly related to bosentan plasma concentrations according to an E max model. Bosentan reversed the vasoconstrictor effect of ET-l measured in the skin microcirculation. There was a tendency toward decreased blood pressure (approximately 5 mm Hg) and increased pulse rate (approximately 5 beats/mm), neither was clearly dose dependent. Oral bosentan was well tolerated. Vomiting and local intolerability was observed at the higher intravenous doses. Con&&on: Bosentan is an orally bioavailable, well-tolerated, and active ET-l antagonist with a low clearance and a moderate volume of distribution. Its intravenous use is limited because of local intolerability. (Clm Pharmacol Ther 1996;60:124-37.)

Cornelia Weber, PhD, Rita Schmitt, PhD, Herbert Birnboeck, PhD, Gerard Hopfgartner, PhD, Sjoerd P. van Marle, MD, Pierre A. IM. Peeters, PhD, Jan H. G. Jo&man, PhD, and Charles-Richard Jones, MD Busel, Switzerland, and Zuidlaren, The Netherlands

From Clinical and Preclinical Research and Development, F. HoEmann-La Roche Ltd., Basel, and Pharma Bio-Research International B.V., Zuidlaren.

Received for publication Nov. 6, 1995; accepted March 3, 1996. Reprint requests: Cornelia Weber, PhD, F. Hoffmann-La Roche

Ltd., Department of Clinical Pharmacology, CH-4002 Basel, Switzerland.

Copyright 0 1996 by Mosby-Year Book, Inc. 0009-9236/96/$5.00 + 0 13/l/73206

Endothelin-1 (ET-l), a 21-amino acid peptide, has been the subject of intense research interest since its first description and isolation in 198K2 It is the most potent and longest-lasting vasoconstrictor known to date. A role for ET-l has been proposed in the pathogenesis of a variety of both cardiovascular and noncardiovascular diseases, including myocardial infarction, hypertension, heart failure,

124

CLINICAL PHARMACOLOGY & THERAPEUTICS VOLUME 60, NUMBER2 Webev et al. 125

renal failure, and respiratory diseases.3’” ET-l be- longs to a family of related peptides of which two others, ET-2 and ET-3, have been identified. It is generated from a precursor mo lecule called proendothel in-1 (or big ET-l) through an unusual enzymic cleavage by the action of an endothelin converting enzyme. So far, two receptor types with different sensitivities to the three endothelin iso- forms have been described. The ETA-type receptor is characterized by its very high affinity for ET-1 compared with a much lower affinity for ET-3. This type plays a dominant role in the vasoconstrictor effects of ET-l. Two subtypes of the ET,- type site are known. The ET,,-receptor, which is present on endothelial cells, is not selective for any isoform and med iates vasorelaxation through release of nitric oxide. The ET,,-receptor subtype is located on certain smooth muscle cells, has greater affinity for ET-3, and med iates direct vasoconstriction.

The availability of potent and selective endothelin-receptor antagonists has led to tremen- dous insight into the important physiologic and pathophysiologic roles of endothelins and their receptors within the short time frame from the first description of endothelin until today.‘z6 A few of these ET-l antagonists are now in clinical development and may elucidate the role of endothelin in certain cardiovascular and other disease states.

Bosentan (see Structure) is a potent nonpept ide m ixed ET,- and ET,-receptor antagonist with higher affinity to the ETA-receptor type.7 The drug has shown activities in a number of preclinical disease mode ls, including reversal and prevention of cerebral vasospasm in dogs and rabbitssJ9 and acute effects (decrease of mean arterial pressure) in rats with chronic heart failure.” This report describes the pharmacokinetics and pharmacodynamics after oral and intravenous dosing with bosentan over a wide dose range in healthy volunteers.

METHODS Subjects

Sixty-four healthy ma le subjects (mean age, 22 2 3 years; mean weight, 75 + 8 kg) participated in the first oral study. Forty-six healthy ma le subjects (mean age, 22 t 2 years; mean weight, 78 + 9 kg) participated in the second intravenous study, 40 in part I and six in part II. Both studies were approved by the independent Ethics Committee in Assen, The Netherlands. Written informed consent was obtained from each subject before enrollment. Physi-

Chemical structure of bosentan (Ro 47-0203).

cal examinations were performed and med ical his- tories, routine laboratory tests, ECGs, and vital signs were recorded before and after each drug administration. Subjects with clinically relevant de- viations from normal or any ma jor illness within 1 month before the screening examination were not included. No subject was included who smoked more than 10 cigarettes a day. No concomitant med- ication was allowed during the study, with the exception of med ications to treat adverse events.

Study designs First oral study. This was a single ascending oral

dose, double-blind, placebo-controlled study. Doses of 3, 10, 30, 100, 300, 600, 1200, and 2400 mg were given as 100 m l aqueous suspension. Subjects were randomized to receive bosentan (n = 6 per dose level) or placebo (n = 2 per dose level). The match- ing placebo consisted of an aqueous suspension on the basis of sodium methylcellulose. The drug was administered at approximately 9 AM after subjects had fasted overnight. Blood was sampled from a forearm vein of each subject through an intravenous catheter before dosing and at i/4, % , 1, 2, 3, 4, 6, 8, 10, 12, 16 (at doses higher than 100 mg), and 24 hours after dosing for determination of plasma drug levels and predose and 1, 2, 4, and 8 hours after dosing for determination of plasma concentrations of ET-l, ET-3, and proendothelin-1. Total urine for measurement of drug concentrations was collected during the time intervals from 0 to 4,4 to 8,8 to 12, and 12 to 24 hours after dosing. Supine blood pressure and pulse rate were measured before administration and at 1, 2, 4, 8, 12, and 24 hours after administration, just after each blood sampling. An ECG (12-lead) was done before dosing and at 3 and 24 hours after administration. In addition, a telemetric ECG lead II was recorded during the first 6

126 Weber et al. CLINICAL PHARMACOLOGY & THERAPEUTICS

AUGUST 1996

hours after drug administration. Standardized light meals were supplied at approximately 4, 8, and 13 hours after dosing. The subjects remained supine for the first 4 hours after dosing, with the exception of the first half hour and for urine collections.

Second intravenous study. This study was divided in two parts. Part I was a single ascending intravenous dose, double-blind placebo-controlled study. Doses of 10, 50, 250, 500, and 750 mg were given by intravenous infusion over 5 minutes as either a 0.2% solution (10 and 50 mg), a 3% solution (750 mg), or a 6% solution (250 and 500 mg) through an indwell- ing catheter. Subjects at each dose level were randomized to receive bosentan (n = 6 per dose level) or placebo (n = 2 per dose level). Infusion solutions of bosentan were prepared for each subject sepa- rately by addition of water for injection to bosentan lyophilisate immediately before dosing by an independent pharmacist. The placebo consisted of 5% dextrose solution, and drug and placebo were always administered between 8 and 9 AM after subjects had fasted overnight. Blood was sampled through an intravenous catheter (distal from the infusion site) before dosing and at 5,10,20, and 35 minutes and at 1, 1X, 2Y2, 4, 6, 8, 10, 12, and 24 hours after dosing for determination of plasma drug levels; blood was sampled before dosing and at 35 minutes, 2X, 6, and 24 hours after dosing for determination of plasma concentrations of ET-l. Total urine for measurement of drug and ET-l concentrations was collected during the time intervals from 0 to 4,4 to 8, 8 to 12, and 12 to 24 hours after dosing. Skin responses (blood flow) after intradermal injections of 10 pmol ET-l and 10 pmol calcitonin gene-related peptide (CGRP) were measured predose and at 35 minutes, 2% hours, and 6 hours after start of the intravenous infusion of bosentan. Blood pressure and pulse rate were measured predose and at 10 and 35 minutes and at 1, 2, 4, 6, 12, and 24 hours after dosing, just after each blood sampling. An ECG (1Zlead) was done before dosing and at 10 minutes and 24 hours after dosing. In addition, a telemetric ECG lead II was recorded during the first 4 hours after dosing. Standardized light meals were supplied at approximately 5, 9, and 14 hours after drug administration. The subjects remained supine for the first 4 hours after dosing.

Part II was an open-label, randomized, two-way crossover study. Subjects received a single oral dose of 600 mg bosentan as 100 ml aqueous suspension and a single intravenous dose of 250 mg (n = 2 received a 0.8% solution over 5 minutes; y1 = 4

received a 0.2% solution over 15 minutes). The washout period between the treatments was at least 4 days. Blood for measurement of plasma drug levels after the intravenous dose was collected at the same time points as described for part I; after oral administration blood was collected before dosing and at 35 minutes and at 1, 1X, 2Y2, 4, 6, 8, 10, 12, 16, and 24 hours after dosing. Blood pressure and pulse rate were recorded at 35 minutes and at 1, 2, 4, 6, 12, and 24 hours after dosing. A 1Zlead ECG was done before administration and at 24 hours after administration.

Drug assay Blood samples for drug level measurements were

collected in Vacutainers (Becton Dickinson Vacu- tainer Systems, Rutherford, N.J.) that contained ethylenediaminetetraacetic acid (EDTA) as an an- ticoagulant and centrifuged at 4” C. Plasma was sep- arated and stored at -20” C until assayed. Plasma concentrations of bosentan were determined by specific high performance liquid chromatography with ultraviolet detection (HPLC-UV) and a more sensitive assay using narrow bore liquid chromatography and ion spray tandem mass spectrometry (LC-MS)?

The HPLC-UV method used liquid/liquid extraction of the drug and the internal standard (Ro 47- 4441) from buffered plasma (pH 9) with dichloromethane. The organic extract was evaporated to dryness at 40” C under a stream of nitrogen and redissolved in an aliquot of mobile phase. Analysis was done by reversed-phase chromatography on Su- pelcosil LC-ABZ (5 pm; Supelco, Bellefonte, Pa.) or NovaPak Phenyl (5 pm; Waters Chromatogra- phy, Millford, Mass.) with gradient elution and ultraviolet detection at 270 nm. The mobile phase consisted of an acetic acid (0.05 mol/L, pH 3.0)/ acetonitrile gradient (Supelcosil LC-ABZ) or a phosphoric acid (3 mmol/L)/acetonitrile gradient (NovaPak Phenyl). The limit of quantification was 20 rig/ml.

The LC-MS method used narrow-bore chromatography with tandem mass spectrometric detection (API III, PE Sciex from Perkin-Elmer, Thornhill, Canada) in selected reaction monitoring mode. Af- ter protein precipitation with acetonitrile, bosentan and the internal standard were extracted with dichloromethane at pH 11. The compounds were chromatographed on a 2 mm internal diameter reversed-phase column and introduced into the mass spectrometer with an ion spray (pneumatically

CLINICAL PHARMACOLOGY & THERAPEUTICS VOLUME 60, NUMBER2 Weber et al. 127

assisted electrospray) interface at a flow rate of 170 kl/min without postcolumn splitting. Two different internal standards were used for the assay: either a structural analog or a deuterated analog. The lim it of quantification was 0.5 rig/ml with a 0.5 m l aliquot of plasma.

In general, drug plasma concentrations lower than 200 r&ml were determined by LC-MS, whereas concentrations above this lim it were determined by HPLC-UV. Both methods were successfully cross- validated.

Urine samples were stored at approximately 4” C until the end of the collection interval. Thereafter the portions were weighed in duplicate to derive the volume from the mean weight, and an aliquot of 10 m l was taken and stored at -20” C until analysis. Urinary concentrations of bosentan were deter- m ined with use of an HPLC method similar to the analysis of plasma samples, involving liquid/liquid extraction at pH 9 with dichloromethane and subsequent gradient reversed-phase chromatography with ultraviolet detection at 270 nm. The quantification lim it was 50 rig/ml.

The performance of all chromatographic methods was mon itored by simultaneous analysis of indepen- dently prepared quality control samples of various concentrations. The interassay coefficient of variation was always below 12.6% with the LC-MS method, below 10.2% with the plasma HPLC-UV method, and below 3.2% with the urine HPLC-UV method. The inaccuracy was between -0.6% and 7.5% for the LC-MS method, between -2.2% and 9.6% for the plasma HPLC-UV method, and between - 1.0% and -5.7% for the urine HPLC-UV method.

Pharmacokinetic evaluation Pharmacokinetic evaluation was performed with

mode l- independent methods. The pharmacokinetic parameters calculated were peak plasma concentrations (C,,), time to reach C,, (tm,), area under the curve [AUC(O-m)], apparent terminal half-life (t,,,), systemic plasma clearance (CL), volume of distribution at steady state (Vss), the fraction of dose excreted unchanged in the urine (F,) and the absolute bioavailability (F).

C max and t,, values after oral administration were taken directly from the observed plasma concentration-time data. The area under the curve was estimated with use of the linear trapezoidal rule” up to the last measured concentration value. Extrapolation to infinity was performed by division of that last measurable concentration by the appar-

ent terminal elimination constant p. The terminal elimination constant B was estimated by performing standard unweighted log-linear least-squares regression analysis of the terminal phase. The t,,z was calculated by division of ln2 by l3. Systemic plasma clearance (CL) was estimated by division of the dose by AUC(O-a). The volume of distribution (Vss) was calculated with use of the relationship13:

Vss = MRT . CL

in which MRT is the mean residence time. MRT was estimated as follows:

MRT = AUMC/AUC - T/2

in which T represents the duration of infusion, and AUMC is the area under the hrst moment curve. Oral bioavailability (F) was estimated as follows:

(Dose, . AUC,,)/(Dose,, * AUC,) * 100

Because the infusion solution concentrations for a number of subjects deviated by more than 5% from the foreseen one, the calculated dose (instead of the planned dose) was used for pharmacokinetic evaluations. F, is reported for only the highest dose groups. During sample analysis it became evident that bosentan in urine is adsorbed to the walls of plastic containers. However, all samples except those of the highest dose group in the intravenous study were collected in plastic containers.

Pharmacodynamic assessments Endothelin measurements. Blood samples for de-

termination of endothelins were collected in pre- chilled tubes that contained EDTA and were centrifuged at 4” C. The plasma samples were stored at -70” C until the assay. The full details of the assays are publ ished elsewhere. l4 In brief, extraction of ET-l, ET-3, or proendothel in-1 from human plasma or urine (ET-l only) was performed by solid-phase adsorption on Sep-Pak C,, cartridges (Waters Chro- matography, M ilford, Mass.). The cartridges were condit ioned with phosphate/citrate buffer, pH 7.0. After application of the plasma to the solid phase, the cartridge was washed with water. ET-l or proendothel in-1 was extracted from the cartridge with methanol/water (90/10, vol/vol). After evapora- tion the residue was dissolved in assay buffer (borate buffered, pH 7.4), and the corresponding antibody (anti-ET-l, anti-ET-3, or anti-proendothelin-1) was added. After preincubation at 4” C for 24 hours, the corresponding tracer, endothelin (ET-l, ET-3, or

128 Weber et al. CLINICAL PHARMA COLOGY & THERAPEUTICS

AUGUST 1996

proendothelin-1) labeled with 1251, was added. The tubes were incubated for 24 hours at 4” C. The separation of bound 1251-endothelin from free 1251- endothelin was performed by precipitation with a second antibody fixed on Amerlex-M magnetobeads (Amersham International plc, Aylesbury, Bucks, England). Radioimmunoassays were then performed as described previously.12 ET-l was measured with the specific rabbit anti ET-l antiserum RAS 6901 (Penninsula Laboratories, Merseyside, England), ET-3 (first study only) with the specific rabbit anti-ET-3 antiserum RAS 6911 (Penninsula Laboratories), and proendothelin-1 (first study only) with the specific rabbit anti-proendothelin-1 antiserum RAS 5313 (Hoffmann-La Roche Ltd, Basel, Switzerland). The cross-reactivity of anti-ET-l serum with ET-2, ET-3, and proendothelin-1 was 0.37%, 0.14%, and 6%, respectively. The cross- reactivity of anti-proendothelin-1 serum with ET-l and ET-3 was <O.OOl%. The cross-reactivity of anti- ET-3 serum with ET-l, ET-2, and proendothelin-1 was l.l%, l%, and l.l%, respectively.

The radioactivity of the precipitate was measured in the y-counter for 5 minutes. The intraassay coefficient of variation for ET-l, ET-3, and proendothelin-1 was 8%, 4%, and 5%, respectively. The interassay coefficient of variation for ET-l, ET-3, and proendothelin-1 was 16%, 19%, and 16%, respectively.

Skin responses. ET-l, 10 pmol (Clinalfa, Laufelfin- gen, Switzerland), was injected intradermally to- gether with the vasodilator CGRP (Clinalfa, Laufelfingen, Switzerland) at a concentration of 10 pmol in the skin of the forearm of each subject before dosing and at 5 minutes, 2 hours, and 5% hours after the start of the bosentan administration. The visual skin response of CGRP when injected alone (done at baseline only) consists of a prolonged erythemic response. The coinjection with ET-l pro- duces an area of vasoconstriction (pallor-like) encir- cling the injection site that still is surrounded by an area of vasodilation (redness).” Skin blood flow measurements were made in triplicate 30 minutes after the injections in the area associated with vasoconstriction, approximately 5 mm away from the injection site, with use of a single channel Doppler blood flow monitor (Moor Instruments Ltd., Axminster, England) with a laser probe on the skin. The skin blood flow is expressed in ar- bitrary “perfusion units.” The injection sites for the skin responses were randomized to predeter- mined positions on the forearm sites to avoid

interference of the position on the forearm on the response. The stability of the basal flow during the study was confirmed by Doppler recording at a site where no injections were made during the study.

WuZ S&W. Supine systolic and diastolic blood pressure (Dinamap blood pressure monitor, Cri- tikon Inc., Tampa Fla.) and pulse rate (pulse fre- quency counting) were measured in the single mode.

Pharmacokinetic-pharmacodynamic modeling Plasma ET-l levels were related to the plasma

bosentan concentrations by use of the E,, model:

E=(E,; CYW,, + Cl

in which E,, is the maximal change of ET-l from baseline, EC,, is the concentration of bosentan required to produce 50% of the maximal effect, and C is the plasma concentration of bosentan. Estimates for E,, and EC,, (fixed effects) and for the random effects (intersubject variability), linked to the fixed effects as constant coefficients of variance, were obtained with double-precision NONMEM, ver- sion 4.0 level 2.r6,17 Data input to and data re- trieval from NONMEM and graphic display were facilitated by an RS/l shell developed in house. The placebo observations were included in this analysis.

Statistics Pharmacokinetic parameters were descriptively

analyzed, calculating mean values and their coefficients of variation (CV). The dose dependency of CL and Vss values after intravenous administration was assessed by standard unweighted linear least- squares regression analysis. As the data revealed, a linear relationship between CL and log-dose and between log-Vss and log-dose could be assumed. Dependency on doses was tested by comparison of the slopes of the regression lines to zero. The dose proportionality after oral administration was assessed by ANOVA of dose-normalized and log- transformed C,, and AUC values, including dose level as factor. Pharmacodynamic data are expressed as mean values t SEM. Vital signs (blood pressure and pulse rate) were descriptively analyzed. For the ET-l plasma concentrations and the laser Doppler blood flow data, an ANOVA for repeated measures with Bonferroni adjustment was done. This was followed by a Student t test for unpaired

CLINICAL P HARMACOLOGY &THERAPEUTICS VOLUME 60, NUMBER2 Weber et al. 129

4 8 12 16 20 24

Time (h)

u - -A-

1J 0 4 8 12 16 20 24

B Time (h)

Fig. 1. Time profiles of the mean plasma concentrations of bosentan after single intravenous and oral administration (n = 6 per dose level) A, Oral administration of 3, 10,30,100,300,600, 1200, and 2400 mg. B, Intravenous administration of 10,50,250,500, and 750 mg as a Sminute infusion.

data at each time point to establish whether signif- RESULTS

icant differences were present between the values of Pharmacokinetics placebo and the different doses of bosentan. A p The time course of bosentan in plasma after in- value of ~0.05 was considered to be statistically travenous and oral doses is given in F ig. 1 (mean significant. plots). The estimated pharmacokinetic parameters

130 Weber et al.

Table I. Pharmacokinetic parameters of bosentan after single ascending oral doses

C AUC (0-a) CL/F Dose (mg) “i;lT ($r) & (NIL * hr) (Llhr)

3 Mean cv (%)

10 Mean cv (%)

30 Mean cv (%)

100 Mean cv (%)

300 Mean cv (%)

600 Mean cv (%)

1200 Mean cv (%)

2400 Mean cv (%)

267.8 29.7

2.8 38.68 3.7 12.2 46.9 51.0 22.0 32.0

2.7 139.1 4.0 1,009 11.0 45.4 53.0 9.5 34.5 34.9

2.5 477.7 4.5 2,840 10.8 41.8 30.7 29.2 16.0 16.6

2.2 1,786 5.0 8,180 14.9 18.8 56.8 32.3 51.8 44.0

2.3 5,000 4.8 18,450 18.1 22.1 48.7 14.3 38.9 32.3

2.3 9,987 5.3 41,480 15.3 35.0 33.0 29.7 25.3 26.6

2.8 14,830 7.5 61,420 23.3 26.7 54.1 33.7 44.8 44.0

2.0 17,220 7.1 79,810 32.1 0.4 31.3 32.9 27.6 27.8

C,,,,, Peak plasma concentration; t,,, time to reach C,,; t1,2, apparent terminal half-life; AUC(O-m), area under the plasma concentration-time curve from time zero to infinity; CL/F, oral plasma clearance; CV, coefficient of variation.

are summarized in Tables I and II for the oral and intravenous studies, respectively.

After intravenous administration, CL varied widely, between 4.7 and 21.3 Whr (individual data) and was apparently dose-dependent (see Fig. 2). It decreased with increasing dose from a mean of 11 to 12 L/hr (dose levels 10 and 50 mg) to a mean of 6 to 7 L/hr (750 mg dose level; p < 0.05). Vss also decreased with increasing dose from a mean of 0.6 to 0.7 L/kg (10 and 50 mg dose level) to a mean of approximately 0.2 L/kg (750 mg dose level; p < 0.05). Both parameters apparently approached a limiting value at higher doses. The terminal t,,, was 3 to 5 hours on average after both intravenous and oral administration but may have been underestimated in most cases because no samples were collected between 12 and 24 hours after dosing, except for the higher oral doses. There the average t,,, was indeed higher, in the magnitude of 5 to 8 hours.

After oral administration, peak plasma levels were reached within an average of 2 to 3 hours. These peak levels and the area under the curve

CLINICAL P HARMACOLOGY & THERAPEUTICS AUGUST 1996

Table II. Pharmacokinetic parameters of bosentan after single ascending intravenous doses

Planned Actual dose dose Concentration t,,, CL V,, 6%) (mg) of solution (%) (hr) (Llhr) (L/kg)

Part I 10

Mean 10 0.2 4.3 10.8 0.67 cv (%) 4 23.3 24.5 21.1

50 Mean 68 0.2 3.9 12.3 0.49 CV(%) 10 17.9 40.2 38.0

250 Mean 308 6 3.3 8.2 0.24 cv (%) 20 12.1 25.8 20.5

500 Mean 500 6 3.1 6.6 0.17 CV(%) 0 12.9 27.0 8.2

750 Mean 904 3 2.8 5.7 0.16 CV (%) 24 14.3 21.7 38.9

Part II 250

Mean 292 610.2 3.4 10.4 0.28 CV (%) 16 13.6 36.7 24.4

t,,, Apparent terminal half-life; CL, systemic plasma clearance; Vs,, volume of distribution at steady state; CV, coefficient of variation.

increased proportionally with dose, up to and including 600 mg (Fig. 3). At this dose level the bioavailability was estimated as 49.8%, with a rather high interindividual variability (CV was 43.8%). Above 600 mg, the increase in Cm, and AUC was much less than dose proportional.

Less than 1.3% of the dose was excreted unchanged in the urine in each individual. After oral dosing, a mean of 0.7% (CV, 52.9%) and after intravenous dosing a mean of 0.8% (CV, 54.2%) of the dose were found in the 24-hour urine portions. The value for the top oral dose may be underestimated by maximally 20% because these samples were collected in plastic containers, where adsorption of bosentan has been observed.

Pharmacodynamics Bosentan significantly increased plasma ET-l lev-

els at doses ~300 mg orally and at doses 2250 mg intravenously. ET-l levels did not change in the placebo group (Fig. 4). After the highest oral dose of 2400 mg, 1.8-fold higher plasma levels (p < 0.001) were observed, whereas 3.1-fold higher levels (p < 0.001) were measured after the highest intravenous dose of 750 mg. At 24 hours after dosing ET-1 levels

CLINICAL PHARMACOLOGY &THERAPEUTICS VOLUME 60, NUMBER2 Weber et al. 13 1

0'

A

1.00 .-

0.50--

G 22 > 0.25~-

: ?

0.10 --

10 2i 50 100 250 560 1000

Dose (mg)

0.05' 10 25 50 100 250 500 1000

B Dose (mg)

Fig. 2. A, Plot of systemic plasma clearance (CL) versus log-dose. B, Plot of volume of distribution at steady state (log-V,,) versus log-dose. Regression lines are included (p < 0.05 and n = 29 in each case).

were back to baseline levels. Bosentan did not virtually unchanged compared with placebo (data change the excretion of ET-l in the urine compared not shown). with placebo (data not shown). Plasma ET-3 and CGRP when injected alone produced a prolonged proendothel in-1 were measured only at the higher erythemic response and a marked vasodilation. doses (600 and 1200 mg) in the oral study and were Blood flow was increased compared with basal

132 Weber et al. CLINICAL P HARMACOLOGY &THERAPEUTICS

AUGUST 1996

A

Dose (mg)

Fig. 3. Plot of area under the plasma concentration-time curve from time zero to infinity [AUC(O-M)] versus dose for oral administration. Open trim&s represent individual values; solid triangles represent the mean values.

blood flow (Fig. 5). The coinjection of endothelin with CGRP caused a complete inhibition of the dilator response of CGRP. Bosentan reversed this inhibitory response of endothelin. The ET-l induced vasoconstriction was completely suppressed 35 minutes after the 750 mg dose, but 6 hours after dosing this effect had disappeared.

There was a tendency toward lower blood pressure after oral and intravenous bosentan in the magnitude of 5 mm Hg compared with placebo, but no clear dose-dependent effect was observed (Fig. 4). At the highest oral dose of 2400 mg, the maximum placebo-adjusted effect in systolic blood pressure was a decrease of 9 mm Hg, 2 hours after dosing. Pulse rate tended to be slightly increased at all dose levels. The increase was in the magnitude of 5 beats/ min compared with placebo, but no dose dependency was observed (Fig. 4).

Oral bosentan was well tolerated up to the pro- spectively defined highest dose of 2400 mg, and there were no serious adverse events. After intravenous bosentan, nausea and vomiting were observed at the 500 mg dose (two of six subjects) and the 750 mg dose (two of six subjects). The vomiting was observed between 4 and 6 hours after dosing and did not require treatment or any action that could in- terfere with the normal cause of the study. Local

tolerability problems (thrombophlebitis or partially occluded vein in the infusion arm) were observed in a few subjects at intravenous doses of 250 mg at concentrations of 6% (one of six subjects), 250 mg at concentrations of 0.8% (one of two subjects), 500 mg at concentrations of 6% (two of six subjects) and 750 mg at concentrations of 3% (four of six subjects) but not after infusion of doses less than 500 mg and at concentrations of 0.2%. These adverse events were of mild intensity and developed within a day after the infusion. No treatment was required in any of the subjects. All events resolved within a few weeks without sequelae. Mild headache was the most prominent adverse event, occurring at oral doses 2300 mg and at intravenous doses 2250 mg. ECG and routine laboratory tests were essentially unaltered by bosentan.

Pharmacokinetic-pharmacodynamic relationship The relationship between the increase in plasma

ET-l concentrations and plasma bosentan concentrations is shown in Fig. 6. The E,, model was found to best represent this relationship. No lag- time (hysteresis) was detected. E,, was estimated as 23.0 pg/ml (CV, 10.7%). EC,, was 8.7 t&ml (16 pmol/L; CV, 27.5%) with a relatively high intersubject variability of 96%.

CLINICAL PHAl iMACOLOGY &THERAPEUTICS VOLUME 60. NUMBER2 Webev et al. 133

A Diastolic Blood Pressure Systolic Blood Pressure I ‘“1 ‘“1

loom&-A 300 mg --o 5-

L-l;

5-

600mg-¤ IZOOmg-0 2400mp-•* I” O r”

o-

j-j 5-5s

- IO- - IO-

,124 8 12

Time(h)

Pulse Rate

B

Time(h)

Diastolic Blood Pressure

Time(h)

Systolic Blood Pressure IO

-‘jb -15i II246 12 24 -I 1246 12 24

Time(h) Tme (h)

Pulse Rate ET-1

15-

IO-

i 5- F E

d

O-

-5-

Time(h) Time(h)

Fig. 4. Time course of mean changes (+-SEM) in blood pressure and pulse rate and of mean plasma endothelin-1 (ET-l) concentrations (?SEM) after oral (A) and intravenous (B) administration of bosentan. ET-l plasma concentrations were not measured for oral doses ~100 mg. Blood samples at 24 hours after dosing were taken only for oral doses >300 mg. ANOVA for repeated measures: *p < 0.05; **p < 0.01; ““p < 0.001 versus placebo.

134 Weber et al. CLINICAL P HARMACOLOGY &THERAPEUTICS

AUGUST 1996

250 t

0-- 1 1 1 J I

-2 -1 0 1 2 3 4 5 6

Time (h)

Fig. 5. Time course of the mean (-SEM) laser Doppler blood flow in the skin after intradermal injections of ET-l. Units are perfusion units (PU). ANOVA for repeated measures: *p < 0.0s; **p < 0.01 versus placebo.

DISCUSSION The most striking pharmacokinetic properties

were the dose-dependent CL and Vss of bosentan. The individual plasma concentration-time profiles did not show any signs of a typical Michaelis- Menten-type elimination, and therefore saturable elimination seems to be a rather unlikely explanation for the observed nonlinearity. The change in Vss with dose could be the consequence of nonlinear tissue binding. Binding of bosentan to plasma proteins is concentration independent, at least in the concentration range covered in the intravenous study (R. Brandt, personal communication, data on file at Hoffmann-La Roche Ltd.). One explanation could be high-affinity binding of bosentan to pharmacologic targets in the tissue compartments, as was recently discussed by Levy!’ The striking difference in the shape of the plasma concentration-time curves between the two lowest intravenous dosages (faster apparent first phase of decline) and all higher intravenous dosages (slower phase) would also sup- port this hypothesis. On the other hand, the estima- tion of V,, might not be valid if the elimination of bosentan is in fact nonlinear, because the method used13 does not apply for such a case. Therefore the change in Vss might be artefactual only. The t,,, could well be underestimated, especially in the high dose range where the terminal phase was not yet

reached during the sampling interval. However, even if a much longer t,,, is assumed, the overall result of decreasing CL and V,, values would still be apparent. To answer the questions whether and why CL and Vss decrease with increasing dose, addi- tional experiments will be necessary.

Because only very low amounts of bosentan were found in urine, hepatic clearance by metabolism and biliary excretion of bosentan or its metabolites can be assumed as the major pathway(s) for the elimination of bosentan. This finding is consistent with the observations made in various animal species, including rats and dogs, in which hepatic metabolism and subsequent biliary excretion of metabolites and, to a lower extent, unchanged bosentan was identified as the major route of elimination (P. Coassolo, A. Viger-Chougnet, personal communication, data on file at F. Hoffmann-La Roche Ltd.).

The oral bioavailability varied extensively between subjects and was lower than could be expected on the basis of a low hepatic first-pass loss of bosentan. This loss can be assumed to be 520% based on a hepatic clearance of maximally 15 L/hr and a blood/plasma ratio of 0.6 (R. Brandt, personal communication, data on file at Hoffmann-La Roche Ltd.). The drug is poorly soluble in aqueous solutions, especially at low pH (D. Grab, personal communication, data on file at Hoffmann-La Roche

CLINICAL PHABMA COLOGY & THERAPEUTICS VOLUME 60, NUMBER 2 Webev et al. 135

t+ -10

0 10 20 30 40 50 60 70

Plasma Bosentan Cont. (pg/mL)

Fig. 6. Correlation between change of plasma ET-l and plasma concentrations of bosentan after intravenous and oral administration. Smoothed curves through the observed (single line) and the model-predicted (broken line) data points are shown.

Ltd.). Therefore incomplete and variable absorption due to lim ited solubility in the gastrointestinal tract may be the ma in reason for our observations. It is not known whether first-pass elimination in the gastrointestinal tract contributes to the lower-than- expected bioavailability.

Bosentan significantly increased plasma ET-l levels at oral doses 2300 mg and at intravenous doses ~2.50 mg. ET-l is regarded as a paracrine rather than an endocrine hormone because most of the ET-l secretion is toward the smooth muscle ce11.19 Plasma ET-l measurements are nevertheless useful because plasma concentrations have been found to correlate well with the severity of certain diseases such as chronic heart failure.” The observation that circulating concentrations of ET-l are increased by the m ixed ET,-/ET,-receptor selective antagonist bosentan in this study or by an ET,-receptor selective antagonist21 but not by ETA-selective antagonists in another studf2 is suggestive of a role for ET,-type receptors in the clearance of ET-1.23’24 The EC,, value of 16 pmo l/L is consistent with an effect at ET,-receptors. This is approximately equal to a free concentration of 300 nmo l/L, which is in the range of the inhibitory concentration (kJ values from radioligand binding studies for bosentan inhib- iting ET-l binding to ET,-receptors (100 to 300

nmo l/L) in contrast to a ki value of approximately 5 nmo l/L for ETA-receptors7 Another explanation for the increase in plasma ET-l would be the displace- ment of ET-l from the receptor, al though binding studies have shown that ET-l dissociates extremely slowly from its receptors.25 The unchanged proendothel in-1 concentrations and the rapid increase in plasma ET-l concentrations shortly after dosing exclude the possibility of de lzovo synthesis. Whether the increase in plasma ET-l can be used as a surrogate marker for drug action will be known only after therapeutic trials in patients have been performed.

Bosentan antagonized endothel in-induced vasoconstriction in the forearm skin. Endothelins inter- act with specific receptors that are activated with different potencies. The ETA-type site is characterized by its very high affinity for ET-l and ET-2 compared with a much lower affinity for ET-3, whereas the ET,-type site has a high and equal affinity for all three isopeptides.26 The receptors responsible for the vasoconstrictor effect of ET-l in the skin are probably ETA-receptors because ET-3 did not cause vasoconstriction in the same setting.27 Laser Doppler flow measurements have been used to assess vascular responses with the advantages of being noninvasive, sensitive enough to record small

136 Weber et al.

changes in blood flow, and reproducible. However, the injection of fluid per se (5% dextrose) caused a slight vasodilation. Certainly, this study assessed only responses of skin microvessels; the response of other vessels cannot be predicted.

However, there is increasing evidence from studies with ET-receptor antagonists that generation of ET-l contributes to the maintenance of basal vascular tone and the regulation of blood pressure.28 The widespread expression of messenger ribonu- cleic acid for the ET isoform and the distribution of ET,- and ET,-receptors in cardiovascular tissues suggest that the ET system may play an important role in cardiovascular control.28

The blood pressure-lowering effect of bosentan in the normotensive healthy subjects included in the present two protocols was moderate. In contrast, a recent study with bosentan in patients with chronic heart failure, a condition with elevated plasma endothelin levels, showed pronounced systemic, pulmonary, and venous vasodilation accompanied by an improved cardiac performance without reflex tachy- cardia.29 Likewise, animal experiments with bosentan have shown that blood pressure was unchanged in spontaneously hypertensive rats and control Wistar-Kyoto rats.30 In contrast, bosentan clearly reduced blood pressure in DOCA-salt hypertensive rats, a rat model with known impaired endothelium- dependent vasorelaxation.31

In summary, bosentan exhibited nonlinear pharmacokinetics after both oral and intravenous administration. CL and Vss decreased with increasing doses to limiting values of around 6 L/hr and 0.2 L/kg, respectively. The oral bioavailability was approximately 50% and decreased above oral doses of 600 mg. Hepatic clearance is assumed to be the major elimination pathway for this drug. Bosentan increased plasma ET-l levels after both oral and intravenous administration and reversed the vasoconstrictor effect of ET-l in the skin mi- crovasculature after intravenous doses of 250 mg and higher. The drug was very well tolerated after oral dosing. Its intravenous use is limited because of local intolerability.

We thank Dr. H. Eggers, Mr. A. Gotschi, and Mrs. G. Eckert for the measurements of urinary drug concentrations. We thank Mr. W. Lausecker and Mrs. R. Defoin for their skillful determinations of bosentan concentrations in plasma. We thank Mrs. A. Karwoth and Dr. S. Halbeisen for their assistance in the pharmacokinetic and pharmacodynamic evaluations.

CLINICAL P HARMACOLOGY & THERAPEUTICS AUGUST 1996

References 1.

2.

3.

O’Brien RF, Robbins RJ, McMurtry IF. Endothelial cells in culture produce a vasoconstrictor substance. J Cell Physiol 1987;132:263-70. Yanagisawa M, Kurihara H, Kimura S, Tomobe Y, Kobayashi M, Mitsui Y, et al. A novel potent vasoconstrictor peptide produced by vascular endothelial cells. Nature 1988;332:411-5. Rubanyi GM. Potential physiological and pathologi- cal significance of endothelins. Drugs Future 1992;17: 915-36.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

Haynes WG, Webb DJ. The endothelin family of peptides: local hormones with diverse roles in health and disease? Clin Sci 1993;84:485-500. Douglas SA, Meek TD, Ohlstein EH. Novel receptor antagonists welcome a new era in endothelin biology. Trends Pharmacol Sci 1994;15:313-6. Warner TD. Endothelin receptor antagonists. Cardio- vast Drug Rev 1994;12:110-22. Clozel M, Breu V, Gray GA, Kalina B, Loffler BM, Burri K, et al. Pharmacological characterization of bosentan, a new potent orally active nonpeptide endothelin receptor antagonist. Pharmacol Exp Ther 1994;270:228-35. Shigeno T, Clozel M, Sakai S, Saito A, Goto K. The effect of bosentan, a new potent endothelin receptor antagonist, on the pathogenesis of cerebral vasospasm. Neurosurgery 1995;37:87-91. Rotrx S, Li%ler BM, Gray GA, Sprecher U, Clozel M, Clozel JP. The role of endothelin in experimental cerebral vasospasm. Neurosurgery 1995;37:78-86. Teerlink JR, Liiffler BM, Hess P, Maire JP, Clozel M, Clozel JP. Role of endothelin in the maintenance of blood pressure in conscious rats with chronic heart failure: acute effects of the endothelin receptor antagonist Ro 47-0203 (bosentan). Circulation [in press]. Lausecker B, Hopfgartner G. Quantitative analysis of a new endothelin receptor antagonist in human plasma by narrow-bore liquid chromatography and ion spray mass spectrometry. J Chromatogr 1995;712:75-83. Gibaldi M, Perrier D. Pharmacokinetics. 2nd ed. New York: Marcel Dekker, 1982. Benet LZ, Galeazzi RL. Noncompartmental determination of the steady-state volume of distribution. J Pharm Sci 1979;68:1071-4. Loeffler BM, Maire JP. Radioimmunological determination of endothelin peptides in human plasma: a methodological approach. Endothelium 1994;1:273- 86.

15.

16.

Crossman DC, Brain SD, Fuller RW. Potent vasoac- tive properties of endothelin 1 in human skin. J Appl Physiol 1991;70:260-6. Beal SL, Sheiner LB. The NONMEM system. Am Stat 1980;34:118-9.

17. Boeckmann AJ, Beal SL, Sheiner LB. NONMEM

CLINICAL P HARMACOLOGY & THERAPEUTICS VOLUME 60, NUMBER 2 Weber et al. 137

users guide. San Francisco: NONMEM Project Group, University of California, 1990

18. Levy G. Pharmacologic target-mediated drug disposition. Clin Pharmacol Ther 1994;56:248-52.

19. Yoshimoto S, Ishizaki Y, Sasaki T, Murota SI. Effect of carbon dioxide and oxygen on endothelin produc- tion by cultured porcine cerebral endothelial cells. Stroke 1991;22:378-83.

20. Wei CM, Lerman A, Rodeheffer RJ, McGregor CGA, Brandt RR, Wright S, et al. Endothelin in human congestive heart failure. Circulation 1994;89:1580-6.

21. Fukuroda T, Fujikawa T, Ozaki S, Ishikawa K, Yano M, Nishikibe M. Clearance of circulating endothelin-1 by ETB receptors in rats. Biochem Biophys Res Com- mun 1994;199:1461-5.

22. Loeffler BM, Breu V, Clozel M. Effect of endothelin receptor antagonists and of the novel non-peptide antagonist Ro 46-2005 on endothelin levels in rat plasma. FEBS Lett 1993;333:108-10.

23. MacCumber MW, Ross CR, Glaser BM, Snyder SH. Endothelin: visualization of mRNAs by in situ hybrid- ization provides evidence for local action. Proc Nat1 Acad Sci USA 1989;86:7285-9.

24. De Nucci G, Thomas R, D’Orleans-Juste P, Antunes E, Walder C, Warner TD, et al. Pressor effects of circulating endothelin are limited by its removal in the pulmonary circulation and by the release of prostacy-

clin and endothelium-derived relaxing factor. Proc Nat1 Acad Sci USA 1988;85:9797-800.

25. Waggoner WG, Genova SL, Rash VA. Kinetic anal- yses demonstrate that the equilibrium assumption does not apply to [‘251]endothelin-l binding data. Life Sci 1992;51:1869-76.

26. Maggi CA, Giuliani S, Patacchini R, Rover0 P, Giachetti A, Meli A. The activity of peptides of the endothelin family in various mammalian smooth muscle preparations. Eur J Pharmacol 1989;174:23-31.

27. Wenzel RR, No11 G, Luscher TF. Endothelin receptor antagonists inhibit endothelin in human skin microcirculation. Hypertension 1994;23:581-6.

28. Haynes W G Webb DJ. Contribution of endogenous generation of endothelin-1 to basal vascular tone. Lancet 1994;344:852-4.

29. Kiowski W , Sutsch G, Hunziker P, Muller P, Kim J, Oechslin E, et al. Evidence of endothelin-1 mediated vasoconstriction in severe chronic heart failure. Lan- cet 1995;346:732-6.

30. Li JS, Schiffrin EL. Effect of chronic treatment of adult spontaneously hypertensive rats with an endothelin receptor antagonist. Hypertension 1995;25:494-500.

31. Li JS, Lariviere R, Schiffrin EL. Effect of a nonselec- tive endothelin antagonist on vascular remodeling in deoxycorticosterone acetate-salt hypertensive rats. Hypertension 1994;24:183-8.

CORRECTION An error occurred in the article, “A case of bleeding requiring hospitalization that was likely caused by an

interaction between warfarin and levamisole” (Wehbe TW, Warth JA. Clin Pharmacol Ther 1996;59:360-2). In Table I, on page 362, levamisole was given orally and not intravenously as mentioned in the table on two occasions. The dose was 50 mg every 8 hours for 3 days.

pharmacokinetics and pharmacodynamics of the endothelin-receptor antagonist bosentan in healthy...

Documents