Metabolism of Thromboxane Bz in Man IDENTIFICATION OF TWENTY URINARY METABOLITES*

[SHs]Thromboxane Bz (12.2 Ci/mol) was infused into a healthy adult male. Urinary metabolites of thromboxane Bz were isolated by reversed phase partition chro- matography and high performance liquid chromatography. Structural identification of metabolites was ac- complished by gas chromatography-mass spectrome-try. Twenty metabolites were identified. Three primary pathways of metabolism of thromboxane Bz were found. A small quantity of thromboxane Bz was excreted unchanged, representing 2.5% of total recovered radioactivity. Two additional metabolites retained the original thromboxane Bz hemiacetal ring; one of these metabolites, 2,3-dinor-thromboxane Bz, was the major urinary metabolite and represented 23.0% of total re- covered radioactivity. The other, 2,3,4,5-tetranor- thromboxane Bz, represented 5.3% of total recovered radioactivity. Two metabolites representing 1.1% of total recovered radioactivity had initially undergone reduction of the hemiacetal ring and indicated a second but relatively minor pathway of metabolism. A major pathway of metabolism was found to involve dehydro- genation of the hemiacetal alcohol group of thromboxane Bz resulting in a series of metabolites with a 6- lactone ring. Sixteen metabolites representing 29.3% of total recovered radioactivity were identified as prod- ucts of this pathway of metabolism.

$ The Joe and Morris Werthan Professor of Investigative Medicine.
It is well established that quantification of circulating or urinary prostaglandin metabolites represents a more reliable means of assessing endogenous prostaglandin synthesis in vivo than does quantification of the parent compound (3,4). Therefore, we initially investigated the metabolic fate of TxBz in the non-human primate. The major urinary metabolite was found to be 9cu,ll,l5(S)-trihydroxy-2,3-dinor-thromba-52,13Edienoic acid (2,3-dinor-TxBz) (5), as also was reported by Kindahl (6). We found another major pathway of metabolic transformation to involve dehydrogenation of the hemiacetal alcohol group at C-11, resulting in the formation of a series of metabolites with a &lactone ring structure (7).
Before quantitative studies of thromboxane synthesis in vivo in man are possible, however, pathways of human thromboxane metabolism must be defined. This present work describes our investigation into the metabolic fate of TxBz in man. A preliminary report of part of this work describing the identification of the major urinary metabolite as 2,3-dinor-TxBz has been published earlier (8). The description of identified metabolites in this paper adheres to the recently proposed nomenclature for thromboxanes (9). Although not confirmed, the stereochemistry of identified metabolites has been presumed to have been unaltered by metabolic transformation of TxB2.

RESULTS AND DISCUSSION
At no time during the infusion of TxB2 were there any significant changes in blood pressure or pulse rate and no clinically apparent adverse effects were observed. Seventyfour % of the infused radioactivity was recovered in the urine within 13 h.
The urine collected was extracted by Amberlite XAD-2 chromatography and the radioactivity was quantitatively eluted with 1600 ml of methanol. The residue obtained after evaporation of the methanol was dissolved in ethyl acetate and applied to a silicic acid column. Ninety-one % of the ' Portions of this paper (including all of "Experimental Procedures" and most of "Results" which describe chromatographic purification of metabolites and the mass spectral data pertaining to their structural elucidation) are presented in miniprint at the end of this paper. In the standard print section of "Results" are summarized the overall findings of this work. Miniprint is easily read with the aid of standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, Md. 20814. Request Document No. 80M-2569, cite author(s), and include a check or money order for $26.40 per set of photocopies. Full size photocopies are also included in the microfdm edition of the Journal that is available from Waverly Press. applied radioactivity was eluted with 1190 ml of ethyl acetate.
Further sample purification and initial compound isolation was effected by reversed phase partition chromatography on a support of 45 g of Hyflo Super-Cel using the solvent system water:n-butyl alcoho1:acetic acid (300:100:4) (v/v/v). Three polar peaks emerged ( Fig. 1) designated Peak A (130 to 160 ml of eluate, 17% of recovered radioactivity), Peak B (370 to 660 m l , 12%), and Peak C (670 to 890 ml, 5%). These were followed by the elution of material not resolved by this solvent system, designated Area D (900 to 2930 ml , 16%) and a large peak of relatively less polar material, designated Peak E (2940 to 4250 ml , 24%). Twenty-six % of the recovered radioactivity remained on the stationary phase and was eluted with 110 ml of methanol and was designated M. A flow diagram outlining the urine purification procedures that are described in the miniprint section and the isolation and letter-number designation of identified TxB2 metabolites are illustrated in Fig. 2. The letter-number designation of each identified metabolite along with its respective chemical name(s) and structure is shown in Fig. 3. Several metabolites were identified from two or more peaks that were completely resolved chromatographically. Explanations for this finding include the possibility of ion-pairing of metabolites with urine impurities and the existence of more than one structural form such as a &lactone and its acid-alcohol form. Table I lists the relative abundances of each metabolite and the total sum of recovered radioactivity identified.
This study has demonstrated that TxB2 is transformed by humans into a variety of metabolites which are excreted into the urine. We have previously identified the major urinary metabolite as 2,3-dinor-TxB2 (8). The present study has described the isolation and structural identification of 19 additional urinary metabolites as well as the excretion of a small quantity of unchanged TxB2.
There are several distinctive features in the metabolism of TxBz in man which are outlined in Table 11. Three separate series of metabolites were categorized, based on ring structure. One series retained the original TxBz hemiacetal ring. A major pathway of metabolism involved dehydrogenation of the hemiacetal alcohol group at C-11. The proposed pathways of TxB2 metabolism in man are iUustrated in Fig. 4. Additional detailed studies of sequences and mechanisms of these biochemical transformations are required to define precisely actual metabolic pathways. Compounds in brackets were not isolated but are proposed intermediates in the formation of identified metabolites.
Only two metabolites in addition to a small amount of unchanged TxBz were identified with an intact hemiacetal ring, even though these metabolites represented a major por- [SSD-RPPC] = Hytlo Super-Cel reversed phase partition chromatography using solvent system D (Table I) tion of the total radioactivity in the urine. Both metabolites, 2,3-dinor-TxB2 and 2,3,4,5-tetranor-TxB~, were products of p oxidation.
Sensitized guinea pig lungs have been described as capable of converting TxBz to 15-keto-13,14-dihydro-TxBz (20). The efficiency of this conversion also apparently increases with successive antigenic challenges (21). We have incubated TXBZ with the 100, OOO X g supernatant of guinea pig liver with added NAD+ and have not found any conversion of TxBz to 15-keto-13,14-dihydro-TxBz. PGEz incubated with the same 100, OOO X g liver supernatnat was essentially quantitatively converted to 15-ket0-13,14-dihydro-PGE~.~ These data and L. J. Roberts, 11, B. J. Sweetman, and J. A. Oates, unpublished observations. the present study suggest that in man and in the absence of immunologic sensitization in the guinea pig that TxBz is not a good substrate for the 15-hydroxy-prostaglandin dehydrogenase enzyme. We cannot exclude the possibility, however, that 15-keto-13,14-dihydro-TxBz was formed in the present study and merely escaped detection since all of the urinary radioactivity was not identified. Alternatively, any 15-keto-13,14-dihydro-TxBZ formed may not be excreted into the urine due to possible conversion to ll-dehydro-15-keto-13,14-dihydro-TxBz. We are presently investigating this latter possibility.
The second major series of compounds, formed as a result of dehydrogenation of the hemiacetal alcohol group of T x B z , was comprised of 16 identified metabolites. The second most   with NAD' results in efficient dehydrogenation of the hemiacetal alcohol group. The dehydrogenation was very inefficient in the absence of NAD+? Therefore, the enzyme appears to be a soluble, NAD-dependent enzyme.
The several compounds identified that had undergone dehydrogenation of the C-15 alcohol group suggests that the 11dehydro derivatives are better substrates for the 15-hydroxyprostaglandin dehydrogenase than compounds with an intact hemiacetal ring. It is assumed that the 15-hydroxy-prostagladin dehydrogenase enzyme is responsible for the dehydrogenation of the C-15 alcohol group although a different enzyme cannot be excluded.
The third minor series of metabolites were acyclic compounds with alcohol groups at C-11 and C-12. This acyclic structure is the same as the sodium borohydride-reduced product of TxB2 (13). This biochemical transformation is envisioned to occur by a process of reduction of the C-11 aldehyde group of the aldehyde-alcohol form of the original hemiacetal ring. The enzyme responsible for this conversion is unknown. However, TxB2 is converted in part to this acyclic derivative when incubated with 100, OOO X g supernatant of guinea pig liver with added NADPH. The conversion is less efficient in the absence of NADPH or in the presence of NADH.3 It, therefore, appears that this enzyme is a NADPHdependent soluble enzyme. Two minor metabolites (E2b and M5dJ were identified in which the A5 double bond has been reduced. Although reduction of the d5 double bond has been described in the metabolism of PGF2, in the rat (15), this has not been found to occur in the metabolism of PGEz and PGF2, in man, although 5,6dihydro metabolites may have escaped detection in previous metabolism studies since all urinary radioactivity was not identified (22,23). The mass spectral data on metabolite A3al suggests the presence of an oxo group attached on one of the carbon atoms from C16-20 and A3al was tentatively identified as either 9a,l5(S)-dihydroxy-ll,19-dioxo-2,3-dinorthromba-52,13E-dienoic acid or 9~~,15(S)-dihydroxy-11,20-dioxo-2,3-dinor-5ZJ3E-dienoic acid. Analogous metabolites have not been previously described in the metabolism of prostaglandins. Insufficient material was present to pennit further structural analysis. The biochemical mechanisms leading to the formation of this oxo group are unknown. The two most likely sites for the location of the oxo group would seem to be at (3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19) or C-20 since metabolites of prostaglandins and TxB2 with hydroxyl groups at the w-1 and w-2 positions have been described (7, [16][17][18][19]. The formation of 11-0-methyl-TxB2 (M2), 11-0-methyl-2,3dinor-TxBz (M5b), ll-O-butyl-2,3-dinor-TxBz (M7), and 2,3,4,5-tetranor-TxB2 butyl ester (M5d3) are all considered to have formed artifactually from TxB2, 2,3-dinor-TxBZ, and 2,3,4,5-tetranor-TxBz during chromatography. We had previously found that the hemiacetal alcohol group of the thromboxane ring is highly reactive with an alcohol in the presence of acid and identified several 11-0-ethyl derivatives in our study of the metabolism of TxBz in the monkey (7). In that study, the original XAD-2 column was eluted with ethanol and ethanol was frequently used during evaporation procedures to form an azeotrope with water. In the present study, ethanol was purposely avoided throughout urine processing and the XAD-2 column was eluted with methanol. This explains the presence in the present study of 11-0-methyl derivatives and the absence of ll-0-ethyl derivatives. 11-0butyl-2,3-dinor-TxB~ undoubtedly formed during the initial reversed phase partition chromatography with the solvent system of butanol/water/acetic acid. The finding of a small quantity of 2,3,4,5-tetranor-TxB2 butyl ester may be indicative that tetranor-TxBn can exist as S-lactone, although this form has not been isolated and identifed. We have found that the S-lactone ring of the 11-dehydro metabolites stored in ethanol at -30 "C will esterify with the ethanol whereas no ester& cation occurs with the upper side chain carboxyl We have also observed that 2,3,4,5-tetranor-TxBz stored in ethanol will form 2,3,4,5-tetranor-TxB2 ethyl ester.3 In contrast, 2,J-dinor-TxBz and TxB2, which would not be expected to lactonize, when stored in ethanol do not form ethyl esters. This is, therefore, suggestive that the 2,3,4,5-tetranor-TxB2 butyl ester formed as a result of reaction of its S-lactone form with butanol rather than direct esterification of the upper side chain carboxyl group by butanol.
Of interest was the identification of 2,3-dinor-PGFZ, in the urine during the course of analysis of material in peak E2b. This compound was not identified in previous studies of the metabolism of PGFz, in man (22), although we have recently reported 2,3-dinor-PGFzU as the major urinary metabolite of PGD, in the non-human primate (16). The fact that a sufficient amount of 2,3-dinor-PGF& was present in the 13-h urine collection after chromatographic losses to obtain a complete mass spectrum would suggest that several micrograms of the endogenous compound are excreted during a 24-h period.
This study has provided the background biochemical information necessary to begin quantification of urinary metabolites of TxBz as a means to assess in vivo production of TxBz in man. In this regard, we have now developed a stable isotope dilution assay for 2,3-dinor-TxBz using combined gas chromatography-mass spectrometry. Six normal adult males have been found to excrete a few hundred picograms/mg of creatinine of 2,3-dinor-TxBz4 These initial studies document the excretion of endogenous 2,3-dinor-TxB~ in human urine and indicate the possibility that an assay for urinary 2,3-dinor-TxB2 may prove to be a useful index of in uiuo TxBz production in man.