Fractional conversion of thromboxane A2 and B2 to urinary 2,3-dinor-thromboxane B2 and 11-dehydrothromboxane B2 in the cynomolgus monkey

Biochimica et Biophysica Acta. 992 (1989) 71-77

71

Elsevier BBAGEN23139

Fractional conversion of thromboxane A 2 and B 2 to urinary 2 , 3 - d i n o r - t h r o m b o x a n e B 2 and 11-dehydrothromboxane B 2 in the cynomolgus monkey P a o l a P a t r i g n a n i L . H o w a r d M o r t o n t, M a f i a C i r i n o ~, A n n e L o r d ~, L u c C h a r e t t e ~, J o h n G i l l a r d ~, J o s h u a R o k a c h ~ a n d C a r l o ? a t r o n o 2 t Mercg Frosst Research laboratories Pointe Claire-Dorval, Quebec (Canada) and " Department of Pharmacology, Catholic University School of Medicine, Rome (Italy)

(Received23 November1988) Key words: Reversedphase high performanceliquidchromatography; Radioimmunoassay;ThromboxaneA2; Thromboxane 132: 2.3-Dinorthromboxane~ ; ll-DehydrothromboxaneBe Following the intravenous administration of thromboxane (TX) i~, the stable hydration product of TXA z, to human and nonhuman prim_~.t~,~file mo~¢ abundant urinary metabolites are 2,3-dinor-TXl~ and ll-dehydro-TXl~. However, it is not known w£,.+~~',,+-.-::,~:~,::,~:.:+ ~,:,,~~,'-~+onof TXlgz to its enzymatic metabolites is an accurate representation of TXA z metabolism. Thus, we have compared the metabolic disposition of synthetic TXA z and TXB z via the fl-oxidation and ll-OH-dehydrogenase pathways in vivo in the monkey. T X A , or ~ (20 n g / k g ) was intravenously administered to four cynomolgus monkeys pretreated with aspirin in order to suppress endogenous TXA z production. Urinary TXI~, 2,3-diuor-TXl~ and ll-dehydro-TXl~ were measm'ed before, during and up to 24 h after thromboxane administration by means of reversed-phase high-pedormance liquid chromatography radioimmunoassay. Aspirin treatment suppressed urinary 2,3-dinor-TXl~ and ll-dehydro-TXl~ by approx. 75%. A similar fractional conversion of T X A , and ~ into 2,3-dinor-TXl~ and ll-dehydro-TXl~ was found. These results suggest that TXA 2 is hydrolyzod to ~ prior to enzymatic degradation and that metabolites of the latter represent reliable indices of TXA 2 biosynthesis. Due to t h e variability in the conversion of thromboxanes into 2,3-dinor-TXi~ and II-dehydro-TXBz, the measurement of both metabolites seems to represent a more reliable index of acute changes in TXA, production. Introduction Thromboxane (TX) A 2 is an unstable derivative of prostaglandin (PG) endoperoxide metabolism for which Hamberg et al. [1] have proposed a bicyclic oxetane structure. TXA 2 possesses a potent contractile activity on bronchial and vascular smooth muscle [2], giomerular mesangium [3] as well as a potent pro-aggregatory activity on platelets [1]. TXA 2 might play a role in the pathogenesis of acute coronary syndromes as well as in the progression of chronic glomerular disease. Thus, several investigators have developed methods for evaluating TXA 2 production in vivo [4,5].

* Present address: Departmentof Pharmacology,Cathofic University School of Medicine, Rome(ltaly) Abbreviations: TX, thromboxane; PG, prostaglandin; RP-HPLC, reversed-phase high performanceliquidchromatograpity; RIA, radioimmunoassay. Correspondence: C. Patrono, Department of Pharmacology,Catholic University School of Medicine, Largo F. Vito I, 00/68 Rome, Italy.

It is well established that the measurement of urinary PG metabolites represents a reliable and noninvasive index of endogenous PG synthesis in vivo [6-8]. Since TXA 2 is rapidly transformed in vitro to its hydration product TXB 2 (half-life: 30 s at 3 7 ° C and pH 7.4) [1], which is chemically stable and biologically inactive, all metabolic studies have been performed using TXI~. Roberts et al. [9] identified 20 urinary metabolites following the administration of radiolabelled TXB 2 to a healthy volunteer. Two major pathways of metabolism were identified, one involving B-oxidation resulting in the formation of 2,3-dinor-TXl~, and the other involving dehydrogenation of the hemiacetal alcohol group at C-11, resulting in the formation of a series of metabolites with a 6-1actone ring structure. The most abundant urinary metabolites were 2,3-dinor-TXB 2 and ll-dehydro-TXB2. Measurements of these metabolites have been recently utilized to evaluate the endogenous production of TXA 2 in various pathophysiologic conditions in man [10-131. The same metabolic pathways have been characterized in the monkey [14]. However, no information is

0304-4165/89/$03.50 © 1989 ElsevierSciencePublishers B.V.(Biomedical Division)

available on the metabolic fate of TXA 2 in vivo. Due to its chemical instability, the natural "D~,A2 has no*, been isolated and characterized as a pure compound. Bhagwat et al. [15] have recently reported the synthesis of the proposed structure of TXA 2Binding to plasma proteins, mainly albumin, may protect TXA 2 from its rapid hydrolysis to TXB2 [16,17]. Thus, we set out to investigate comparatively the fractional conversion to and time-course of urinary excretion of 2,3-dinor-TXB2 and 11-dehydro-TXB2 following the administration of TXA 2 or TXB2 in the cynomolgus monkey. In order to measure TXB2 and its metabolites in urine, we have developed a reversed-phase/highperformance liquid chromatography (RP-HPLC) separation technique, allowing the specific recognition of the different immunoreactive eicosanoids by radioimmunoassay (RIA). Materials and Methods

Synthesis and biologicalactivity of TX_42. The synthesis of 1,15-anhydro-TXA 2 was carded out as described by Bhagwat et al. [15,18]. 5 0 / t g of the TXA2-1actone were hydrolyzed immediately prior to each experiment. Hydrolysis of the lactone was achieved at room temperature (1 h) using an excess of sodium hydroxide (0.15 M) in tetrahydrofuran/methanol/water (2 : 1 : 1) [19]. The hydrolyzed product was diluted 10 times with anhydrous methanol. This solution may be stored at - 7 8 ° C for up to 1 week without significant loss of biological activity. The capacity of synthetic TXA 2 to induce platelet aggregation in human platelet rich plasma (PRP) was tested both at the beginning and at the end of each experiment by using a Payton dualchannel aggregometer [19]. Platelet count in peripheral venous blood was monitored before, during and 1 h after TXA 2 administration (Coulter Thrombocounter, Coulter Electronics, Hialeah, FL, U.S.A.). Animal experiments. Four cynomolgns monkeys (three males, one female, body weight 3.8 + 0.5 kg) were anesthetized with ketamine hydrochloride (25 mg/kg i.m.) and supplemental doses of either ketamine hydrochloride or sodium pentobarbital (1 mg/kg i.v.) were given as required. A catheter was inserted into the bladder for urine collection during the 3 h of the experiment and a lead II electrocardiogram (ECG) was recorded and continuously displayed on a defibrillator screen (Marquette Electronics). The heart rate was obtained from the ECG signal. The animals were pretreated with aspirin (10 mg/kg per day for 3 days) in order to minimize the contribution of endogenous TXB2 production to the measurements of urinary TXB2, 2,3dinor-TXB2 and ll-dehydro-TXB2. Synthetic TXA 2 or TXB2 (Cayman Chemical Company, Ann Arbor, MI, U.S.A.) (20 ng/kg) was added to 100/tl of homologous plasma and immediately injected as a bolus every 2 rain

for 20 min into the saphenous vein of the monkey. Urine was collected during the 20 h preceding the administration of thromboxanes using a metabolic cage. Beginning with the administration of thromboxanes, 1 h urine collections were obtained for 3 h via the catheter. During the following 21 h urine was collected in a metabolic cage. Urine samples were immediately frozen and kept at - 2 0 ° C until assayed. Urinary measurements of TXB2 metabotites. 12000 dpm of radiolabelled TXB2 ([3H]TXB2, 142 Ci/mmtA, New England Nuclear, Boston, MA, U.S.A.) was added to 1-5 ml of urine ~ q u o t s for recovery evaluation. After adjusting the urine pH to 4-4.5 with formic acid, TXB2, 2,3-dinor-TXB2 and ll-dehydro-TXB2 were extracted from urine on SEP-PAK Cls cartridges (Waters Associate,,~, Milford, MA, U.S.A.) and eluted with ethyl acetate. Eluted TXB2, 2,3-dinor-TXB2 and ll-dehydroTXB2 were separated on RP-HPLC with the solvent system acetonitrile/water/acetic acid (27: 73 : 0.18) at a flow rate of 0.5 ml/min, l-rain fractions were collected, dried under a stream of N 2 and reconstituted with phosphate (0.02 M, pH 7.4) or Tris-phosphate (0.02 M, pH 10) buffer and analyzed for TXB2, 2,3-dinor-TXB2 and ll-dehydro-TXB2 by RIA techniques. Radioimmunoassays. Twenty 250/tl HPLC-fractions were assayed in a 1.5 ml volume of 0.02 M phosphate buffer (pH 7.4) for TXB 2 and 2,3-dinor-TXB2 measurements or 0.02 M Tris-phosphate buffer (pH 10) for 11-dehydro-TXB2 measurements. Both buffers contained 0.5% of charcoal-adsorbed human plasma. The RIA for TXB~. and 2,3-dinor-TXB2 used 4000 dpm [3H]TXB2 and an anti-TXB2 serum diluted 1:850000 which showed 45-50% crossreaction with authentic 2,3-dinor-TXB2 [20]. Radiolabelled ll-dehydro-TXB2 was prepared by incubating [3H]TXB2 with the highspeed supernatant of gninea-pig liver homogenate in the presence of 4 mM N A D (Sigma, St. Louis, MO, U.S.A.). The ll-dehydro[3H]TXB2 formed was separated from unreacted [3 H]TXB2 and other enzymatic derivatives by RP-HPLC [21]. Approx. 4000 dpm of 11-dehydro[3H]TXB 2 and a very specific anti-ll-dehydro-TXB2 serum [22] diluted 1 : 300000 was used. ll-Dehydro-TXB2 occurs in two forms in a pH-dependent equilibrium [9]. At acidic pH it exists in the 8-1actone form with an intact thromboxane ring, while at basic pH it is characterized by an open ring structure w;,th a second carboxyl group at C-11. The anti-ll-dehydro-TXB2 serum has been shown to recognize preferentially the open ring form [22]. Thus, before performing the RIA both the radiolabelled and the unlabelled ll-dehydro-TXB2, as well as the HPLC-fractions corresponding to the retention time (RT) of ll-dehydro-TXB2 (open and lactone forms), were kept at pH 10 (0.02 M Tris-phosphate buffer) at room temperature for 2 h in order to allow the complete conversion of 11-dehydro-TXB2 into the open ring form (Fig. 1).

0-02fA l I-DEHYDRO-TXB2

6-LACTONE FORM

I

I

20 30 40 RETENTION TIME (MIN) 2h, pill0, 2oC

10

o

B

0.02-

50

11-DEHYDRO-TXB2 OPEN FORM H0 0 ~ " ~

+- ' "

tn

0

1=0

I

20

--1--~ 30

t 40

50

RETENTION TIME (MIN) Fig. 1. (A) RP-HPLCof 11-dehydro-TX~&lactoneform. Moving phase: acetonitrile/water/aceticacid (30:"/0:0.18) (B) RP-HFLCof 11-dehydro-TXBz open form, obtainedby incubatingthe 8-1actone format pH 10, at 22°C for 2 h.

The assay for TXIh, 2,3-dinor-TXBz and 11-dehydro-TXEh were incubated at 4 ° C for 24 h, and separation of the antibody-bound ligand from free ligand was achieved by rapidly adding 0.1 ml of human plasma and 50 /~1 of a charcoal suspension (50 mg/ml) and subsequent eentrifugation at 3000 rpm for 10 min at 4 ° C. The ICs0 values (concentration of unlabelled antigen which displaces by 50% the binding of the homologou~ tracer) were 9 pg/nd, 20 pg/nd and 9 pg/ml for the RIA of TXB2, 2,3-dinor-TXBz and 11-dehydro-TXB2 respectively. The least detectable concentrations that could be measured with 95~ confidence (i.e., 2 S.D. at zero) were 2 pg/ml of incubation mixture for the RIA of 2,3-dinor-TXB2 and 1 pg/mi for the RIA of both 11-dehydro-TX~ and TXl~.

Validation of the urinary measurements was obtained by different independent criteria: (1) dilution and recovery studies and (2) detection of homologous immunoreactivity only in the HPLC fractions corresponding to the retention times of the authentic standards. These RIA techniques have been previously validated for the measurements of TXB~ [2% 2,3-dinor-TXB2 [23] and 11-dehydro-TXB2 [22] in human urine samples by comparison with capillary gas chromatography negative ion chemical ionization-mass spectrometry. Statistical analysis. Results were analyzed for statistical significance using the Student's t-test. Results

RP-HPLC-RIA technique A very sensitive and specific HPLC-RIA technique has been developed for the simultaneous analysis of TXB2, 2,3-dinor-TXB 2 and ll-dehydro-TXB2 in urine. 2,3-Dinor-TXB2, the open form of ll-dehydro-TXB2, TXB2 and the lactone from of ll-dehydro-TXB2 were separated by RP-HPLC with retention times of 13.2, 23.7, 39.7 and 69.2 min, respectively (Fig. 2, upper panel). 2,3-dinor-TXBz, TXB2 and ll-dehydro-TXB2 concentrations in HPLC fractions were measured by previously characterized RIA techniques [20,22,23]. The detection limit of these methods was 6-12 pg/ml of HPLC-fractions. Fig. 2 (lower panel) depicts the distribution of TXB2-, 2,3-dinor-TXB2-, and ll-dehydroTXBe-like inununoreactivities, as detected by RIA in the HPLC fractions of two urine samples obtained before and during TXA 2 infusion. Homologous immunoreactivity was detected only in the fractions corresponding to the retention times of authentic 2,3-dinorTXB2, TXBz and ll-dehydro-TXB2. This establishes identical chromatographic as well as immunologic behaviour of endogenously released and exogenously induced TXBz metabolites with the corresponding standards. Biological properties of synthetic TXA , The effect of synthetic TXA2 on human platelet aggregation is reported in Fig. 3. TXA 2 at concentrations of 0.075-0.75 I*M caused a concentration-dependent aggregation of human PRP. Irreversible platelet aggregation occured at 0.15/tM TXA 2. TXA2 was 7and 9.2-times more potent than the stable PGH~/TXA 2 ~,',dog~s, U46619 and U40,069, re~po.ctively[24]. Synthetic TXA2 incubated in 0.02 M phosphate buffer (pH 7.4) for 2 rain was quantitatively recovered as a substance that was immunologically and chromatographically indisting~,.ishable from authentic TXI~. Since plasma has been demonstrated to stabilize TXA2 (tl/2: approx. 3 rain) [16,17] both TXA2 and

80 70 60

f"

50 40

11-DEHYDRO-TX~ OPEN FORM 2 •3 -TX~ D NOR -"l

.

il

/

30 11 - DE HYDRO-TXE2

TX~

20

6-LACTONE FORM

10

t

t

0 2o

3[0

410

;o

6/

71o

81

RETENTION TIME (MIN)

300

ADMINISTRATION

250 l

/I

'°°l

+',iit

TxAzS

Fig. 3. Dose-dependent platelet aggregation induced by synthetic TXA2 in human platelet rich plasma. The intravenous administration of both TXI~z and TXA 2 caused a time-dependent increase in the urinary excretion of T X Bz, 2,3-dinor-TXBz and ll-dehydroT X Bz (Fig. 4) that was maximal during the 1st h beginning with the injections and returned to baseline values between 3 and 24 h. A statistically significant increase in the urinary excretion of "IXBz, 2,3-dinor-TXl~ and ll-dehydro-TXB 2 was reached during the 1st and the 2nd h of urine collection ( P < 0.01). During the 3rd h, 11-dehydro-TXB2 (in both T X A , and TXB2 experiTXA2 ADMINISTRATION(20 nglkg, i.v.)

6°°°T 4,~

"io ~o

~.zW5000 T

RETENTION TIME (MIN) Fig. 2. (A) Separationof TXB2,2,3-dinor-TXB2 and 11-dehydro-TXB2 (open and laetone forms) by RP-HPLC. (B) Distribution of TXB2-, 2,3-dinor-TXB2-,and ll-dehydro-TXB2 (open and lactone forms)-llke immunoreactivity,as detected by RIA in the HPLC fractions of two urine samplesobtained before and during TXA2 infusion.

~, 3°°°T

o6

!

0.7'5 uM

0.075 uM 0.15 uM 11o

;o

io

~o

~o

,~o

6o

T X Bz were diluted in homologous plasma just before being administered intravenously to the monkey. The time required to dilute and inject the thromboxanes was always less than 20 s. T X A 2 and T X B z fractional conversion in vivo

Four cynomolgus monkeys were pretreated with oral aspirin (10 m g / k g ) for 3 days in order to minimize the contribution of the endogenous production of TXB2, 2,3-dinor-TXB2 and ll-dehydro-TXB2 to their urinary measurem~cnts following thromboxane administration. Aspirin caused only a 36% reduction in urinary TXB 2 excretion, while suppressing by approx. 74% both 2,3dinor-TXB2 and ll-dehydro-TXB 2.

OTX~

• 2.3-DINOR-TXB2

4oo0T

Ca''-DE"VDRO-TX~

o=,MEA'--+SE

2ooo T

E

,oooT o--.~

~

BEFORE

B°°°T

0-1

1-2

2-3

--~ 3-24

HOURS

TXB2 ADMINISTRATION(20 ng/kg, i.v.)

4,

~.'~ sooo

i ooo T ~ 2000

..... T E o_.__m E looo T

1

~1~ _.~.~

0-1 1-2 2-3 3-24 HOURS Fig. 4. Urinary excretion of TXB2, 2,3-dinor-TXB2and 11- dehydxoTXB2 measuredbefore, during and after TXA2 administration(upper panel) and TXB2 administration (lower ":;~el) to four cynomolgus monkeys. BEFORE

75 TABLE ! Urinary excretion rates (pg/mg creatinine) of TXB~, 2,3-dinor.TXB z and l l-dehydro-TXB 2 measured in four cynomo/gus monkeys before, during the 3 h beginning with the injection of TXA 2 or TXB z and during the subsequent 21 h.

All valuesare mean+S.D. Time Before 0-3h 3-24h

TXA2 (20 ng/kg i.v.) TXB2 2,3-dinor-TXB 2 118+ 133 200+ 194 3303+1666 5148+3307 974- 61 169± 136

1l-dehydro-TXB2 349+350 5325:1:701 326:1:177

ments) and 2,3-dinor-TXB2 (in the TXB2 experiment) were numerically higher than control values; however, this difference did not reach statistical sig~fificance. The excretion rates of TXB2, 2,3-dinor-TXEh and 11-dehydro-TXBz measured before, during the 3 h beginning with the injection of TXA 2 or TXB2 and during the subsequent 3-24 h are detailed in Table I. Approx. 2% of the injected thromboxanes was recovered as unmetabolized TXI~. The fractional conversion of injected TXA2 to urinary 2,3-dinor-TXB2 and 11-dehydro-TXlh averaged 3.3~ and 3.6~, respectively. The corresponding figures for the fractional conversion of injected TXB2 were 3.7~ and 2~. No statistically significant difference was reached between the two sets of measurements. The administration of both TXA2 and TXB2 did not produce any significant change in heart rate and platelet count in peripheral venous blood. Discussion

Various methods have been developed for the measurement of TXB2, and its enzymatic metabolit?s in urine. G C / M S combines very high specificity with sensitivity in the low picogram range, particularl',' in the negative ion chemical ionization mode (NICI) [10,21]. Eicosanoid measurements l~erformed in highly purified urine samples by very sensitive and specific RIA techniques were shown to correlate with those obtained by N I C I - G C / M S [22,23]. Here, we describe a very sensitive and specific HPLC-RIA technique which allows the simultaneous measurement of TXlh, 2,3-dinor-TXB2 and 11-dehydro-TXlh in the same urine sample. One difficulty in developing a RIA technique for 11-dehydro-TXB2 is its existence in two forms (open and lac~one form~) in ~. pH-dependem cq~.;L;hrh_~_~[.9]. Because of the preferential recognition hy our antiserum of the open ring form of 11-dehydro-TXB2 [22], we allowed the full conversion of 11-dehydro-TXlh into its open ring form by incubating the HPLC fractions, the radiolabelled and unlabelled 11-dehydro-TXB2 for 2 h at pH 10 at room temperature prior to the assay. Furthermore, in order to avoid the interconversion of the two forms

TXB2(20 ng/kg i.v.) TXB2 2,3-dinor-TXB2 11-dehydro-TXB2 137+ 74 130+ 51 437+ 231 3065±1898 6958±3055 4569+2207 834- 50 171+ 100 5084- 135

during the overnight incubation at 4 ° C the assay was performed in a pH 10 buffer. The differential suppression by aspirin of the urinary excretion of TXB e (36%) vis41~vis that of its enzymatic metabolites (74%) is consistent with a predominant renal origin of the former and largely platelet origin of the latter, as characterized in human beings [4,8]. Due to the great instability of the oxetane ring, T X A : is very difficult to handle. In order to avoid its rapid hydrolysis to TXB2, TXA2 was diluted in plasma that had been demonstrated to prolong its biological half-life in vitro ( t l / 2 : 3 min) [16,17]. In our experiments, the time required to dilute and inject TXA, in the monkey was always less than the chemical half-life of TXA: in aqueous solution, i.e., 30 s. By using the same procedure of dilution in plasma, we have found that the injection of TXA2 (20 ng/kg) into the left renal artery of the pig caused a significant reduction of renal blood flow that was reversed by a selective TXA 2 receptor antagonist (Ciri~o e~ al., unpublished data). Thus, it is reasonable to bfer that our experimental conditions did allow biologically active TXA 2 to reach the bloodstream. However, the intravenous administration of 20 ng/kg of TXA z to the monkey did not affect the heart rate or peripheral platelet count to any statistically significant extent. The failure of administered TXA z to cause systemic biological effects is likely to be due to its low peripheral plasma concentrations. In fact, a bolus injection of 20 ng/kg of TXA 2 should give a peripheral blood concentration of approx. 300-400 pg/ml. Due to its chemical instability and rapid metabolism in vivo repeated bolus injections of TXA: every 2 mln (as done in this study) should not cause its accumulation in the circulation. On the other hand, as repgr,'e~ here, irreversible platelet aggregation occurred in vitro at approx. 50 n g / m l of TXA:. In our metabolic studies we have chosen to administer to the monkeys a dose of TXA 2 which would have no systemic haemodynamic effects, in order not to alter the hepatic handling of TXA2/TXB2 and the renal clearance of its metabolites. The dose of thrombox:-mes that we administered is comparable to that of TXB,

used in previous human studies by Patrono et al. [23] (0.1-5 ng. kg - t . min - t ) but much lower than that employed by Roberts et al., both in the monkey [14] (2 p g . kg - t - rain - t ) and in man [9] (65 n g . kg - t - iron-t). This study has demonstrated that the metabolic handling of synthetic TXA 2 is indistinguishable from that of TXI~ in non-human primates. Thus, a similar fractional conversion of TXA 2 and TXB 2 into 2,3-dinorTXB 2 and ll-dehydro-TXB~ was found. The urinary excretion of 2,3-dinor-TXB, and ll-dehydro-T~CB2 was characterized by different kinetics, likely ieflecting differences in polarity of the two metabolites. While the vast majority of 2,3-dinor-TXBz was excreted during the first hour, ll-dehydro-TXE h, a less polar substance, showed a slower urinary elimination that was not completed until after 3 h. The different time-courses of urinary excretion of 2,3-dinor-TXBz and l l - d e h y d r o TXBz is well related to the different rates of plasma clearance of the two metabolites in man. As reported by Lawson et al. [21], following the intravenous administration of TXB2 to healthy volunteers the post-infusion half-life of ll-dehydro-TXI~z approximated 1 h, while that of 2,3-dinor-TX~ was 15-17 rain. Although characterized by different kinetics, the urinary excretion of 2,3-dinor-TXIh and li-dehydro-TXB2 was not significantly different during the 3 h beginning with the administration of either TXA 2 or TXBz (Table I). The low fractional conversion of thromboxanes into the two major urinary metabolites (2,3-dinor-TXIgz: 3.5 + 1.7%; ll-dehydro-TXBz: 2.8 + 1.5%; mean + S.D., n = 8) is at variance with previous observations of Roberts et al. [14]. In one cynomolgus monkey 14.5% and 6% of [3H]TXI~ (2 p g . kg - t . m;a - t ) was recovered as urinary 2,3-dinor-TXB2 and l l - d e h y d r o - T X Bz, respectively. The use of anesthesia and the more 'physiologic' dose of thromboxanes administered in our experiments might be responsible for the lower fractional conversion. Variable results were also reported on the metabolism of TXI~ in man. Roberts et al. [9] showed that 23% and 7.6% of infused [3H]TXBz (65 ng. kg -1 • rain -1) was recovered in the urine of one healthy volunteer as 2,3-dinor-TXB 2 and ll-dehydro-TXB2, respectively. On the other hand, Ciabattoni et at. [25] have recently reported that the fractional conversion of unlabeUed T X B 2 infused at more physiologic rates (0.1-5 n g . kg - t • m i n - 1) to four healthy volunteers was 6.4 + 1.2% for 2.3-dinor-TXB2 and 6.8 + 0.7% for ll-dehydro-TXBz. Several factors might explain these apparent discrepancies: (1) different metabolic handling of radiolabelled and unlabelled T X B 2, (2) inter-individual variability in thromboxane metabolism, (3) faster saturability of the enzyme involved in ll-OH-dehydrogenation thus resulting in overestimation of the biotransformation to t -

oxidation products following the high rates of infusion used by Roberts et at. [9,14]. In conclusion, our results suggest that TXA 2 is hydrolyzed to TXB 2 prior to enzymatic degradation and that metabolites of the latter may represent reliable indices of TXA2 biosynthesis. Because of the low and variable fractional conversion of thromhnxanes into 2,3-dinor-TXB2 and ll-dehydro-TXB2, the measurement of both metabolites seems to represent a more reliable index of acute changes in TXA 2 production. Acknowledgements We ~dsh to thank A.W. Ford-Hutchinson and R. Zamboni for helpful discussions and advice, D. Singh for expert technical assistance, and M. Redaelli and M.L. Bonanomi for expert editorial assistance.

References 1 Hamberg, M., Svensson,J. and ;Samuelsson,B. (1975) Prec. Natl. Acad. Sci. USA 72, 2994-2998. 2 Ellis, E.F., Oelz, O., Roberts, LJ., !I, Payne,N.A., Sweetman,B.J., Nies, A.S. and Oates, J.A. (1976) Science193,1135-1137. 3 Scharshmidt, L.A., Llanos, E. and Dunn, M.I. (1983) Fed. Prec. 42, 3058-3063. 4 FitzGerald, G.A., Pedersen,A.K. and Patrono, C. (1983) Circulation 67,1174-1177. 5 Westlund, P., Granstri~m, E., Kumlin, M. and Nordenstri~rn,A. (1986) Prostaglandins 31, 929-960. 6 Samuelsson,B. and Green, K. (1974) Biochem.Med. 11, 298-303. 7 Samuelsson, B., Granstrom, E., Green, L., Hamberg, M. and Hammarstrom, S. (1975) Annu. Rev. Biochem.44, 669-695. 8 Catella, F., Nowak, J. and FitzGerald, G.A. (1986)Am. J. Med. 81 (Suppl. 2B) 23-29. 9 Roberts, LJ., II, Sweetman, B.J. and Oates, J.A. (1981) J. Biol. Chem. 256, 8384-8393. 10 Catena, F. and FitzGerald, G.A. (1987)Thromb. Res. 47, 647-656. 11 Vesterqvist, O., Edhag, O., Green, K. and Henriksson, P. (1985) Thromb. Res. 37, 459-464. 12 Fitzgerald, DJ., Roy, L., Catella, F. and FitzGerald, G.A. (1986) N. Engl. J. Med. 315, 983-989. 13 Vejar, M., Lipkin, D.P., Maseri, A., Born, G.V.R., Ciabauoni, G. and Patrono, C. (1988) Clin. Res. 36, 326A. 14 Roberts II, LJ., Sweetman, B.J. and Oates, J.A. (1978) J. Biol. Chem. 253, 5305-5318. 15 Bhagwat, S.S., Hamann, P.R., Still, W.C., Bunting, S. and Fitzpatrick, F.A. (1985) Nature 315, 511-513. 16 Smith, J.B., lngerraan, C. and Silver, M.J. (1976) J. Clin. Invest. 58,1119-1122. 17 Folco, G.C., Gransl~,:..c-'E.and Kindhal, H. (1977)FEBS Lett. 82, 321-324. 1 18 Bhagwat, S.S., Hamanii, P.R. and Still, W.C. (1985)J. Am. Chem. Sec. 107, 6372-6376. 19 Hall, R.A., Gillard, J.W., Gnindon, Y., Letts, ~.. Champion, E., Ethier, D., Evans, J., Ford-Hutchinson, A.W., Forttn, R., Jones, T.R. and Lord, ~.. (1987) Eur. J. Pharmaeol. 135,193-201. 20 Patrono, C., Ciabattoni, G., Remuzzi0G., Gotti, E., Bombardieri~ S., Di Manno, O., Tartarelli, G., CinoUi, G.A., Simonetti, B.M. and Pierac¢i, A. (1985) J. Clin. Invest. 76,1011-1018.

77 21 Lawson, J.A., Patrono, C., Ciabattoni, G. and FitzGerald, G.A. (1986) A~al. Biochem. 155,198-205. 22 Ciabattoni, G., Mae~ouf, J., Catella, F., FitzGerald, G.A. and C. Patrono (1987) Bi~chim. Biophys. Acta 918, 293-297. 23 Patrono, C., Clabattoni, G., Pugliese, F., Pierucci, A., Blair, I.A. and FitzGerald, G.A. (1986) J. Clin. Invest. 77, 590-594.

24 Letts, G., Lord, A. and Morton, H. (1986) Pharmacologist28,186. 25 Ciabattoni. G., Pugiiese. F., Dart. G., Pierucci, A.. Simonetti. B.M. and Pa:rono. C. (1989) Biochim. Biophys, Acta 991, 66-70.