Nonenzymatic free radical-catalyzed generation of thromboxane-like compounds (isothromboxanes) in vivo.

The isoprostanes (IsoPs) are novel bioactive prostaglandin-like compounds produced in vivo by free radical-catalyzed peroxidation of arachidonyl-containing lipids. Previously, we have identified IsoPs containing F-type and D- and E-type prostane rings that are formed by reduction and rearrangement of IsoP endoperoxide intermediates, respectively. We now explore whether thromboxane B2 (TxB2)-like compounds, termed B2-isothromboxanes (B2-IsoTxs), are formed by rearrangement of IsoP endoperoxides. Detection of these compounds was carried out using a stable isotope dilution mass spectrometric assay originally developed for quantification of cyclooxygenase-derived TxB2. Incubations of arachidonic acid with Fe/ADP/ascorbate for 30 min in vitro generated a series of peaks representing putative B2-IsoTx at levels of 62.4 +/- 21.0 ng/mg arachidonate. Using various chemical modification and derivatization approaches, it was determined that these compounds contained hemiacetal ring structures and two double bonds, as would be expected for B2-IsoTx. Analysis of the compounds by electron ionization mass spectrometry yielded multiple mass spectra similar to those of TxB2. B2-IsoTxs are also formed esterified to phospholipids; oxidation of arachidonyl-containing phosphatidylcholine in vitro followed by hydrolysis resulted in the release of large amounts of these compounds. To explore whether B2-IsoTxs are also formed in vivo, a well characterized animal model of lipid peroxidation consisting of orogastric administration of CCl4 to rats was used. Levels of B2-IsoTx esterified in lipids in the liver increased 41-fold from 2.5 +/- 0.5 to 102 +/- 30 ng/g of liver. In addition, circulating levels of free compounds increased from undetectable (<5 pg/ml) to 185 +/- 30 pg/ml after CCl4, a 37-fold increase. Thus, we have provided evidence that IsoTxs constitute another novel class of eicosanoids produced in vivo nonenzymatically by free radical-catalyzed lipid peroxidation. These studies thus expand our understanding of products of lipid peroxidation formed in vivo from the free radical-catalyzed peroxidation of arachidonic acid.

Free radical-catalyzed lipid peroxidation has been implicated in the pathogenesis of a wide variety of human disorders (1)(2)(3)(4). Nonetheless, much remains to be understood about the mechanisms of oxidant injury in vivo. Previously, we reported the discovery that a series of prostaglandin (PG) 1 F 2 -like compounds (F 2 -isoprostanes (F 2 -IsoPs)) capable of exerting potent biological activity are produced in vivo in humans as products of the free radical-catalyzed peroxidation of arachidonic acid (5). Formation of these compounds occurs independently of the cyclooxygenase enzyme, which had heretofore been considered obligatory for endogenous prostanoid biosynthesis. Circulating levels of these compounds increase dramatically in animal models of free radical injury, and quantification of F 2 -IsoPs has proven to be an important advance in our ability to assess oxidant stress in vivo (5,6). Formation of F 2 -IsoPs proceeds through intermediates comprising four positional peroxyl radical isomers, which undergo endocyclization to yield PGG 2 -like bicycloendoperoxides. These are then reduced to F-ring IsoPs. F 2 -IsoPs are initially formed in situ from arachidonic acid esterified in phospholipids and are subsequently released preformed by a phospholipase (6,7). This mechanism of formation is in contradistinction to the formation of cyclooxygenase-derived prostanoids in which arachidonic acid esterified in phospholipids must be released prior to oxygenation.
More recently, we reported that IsoPs that are PGD 2 -and PGE 2 -like compounds (D 2 /E 2 -IsoPs) also are produced in vivo from rearrangement of isoprostane endoperoxides (8). Like Fring compounds, they are formed in situ on phospholipids, their formation increases markedly in animal models of oxidant injury, and they exert potent bioactivity. Because Tx can also be formed by nonenzymatic rearrangement of cyclooxygenase-derived PGH 2 (9), we explored whether Tx-like compounds can also be generated as rearrangement products of the IsoP endoperoxide intermediates. We present evidence that TxB 2 -like compounds are, in fact, produced both in vitro and in vivo and that they are present both esterified to phospholipids and in the free form. Because these compounds are isomeric to cyclooxygenase-derived TxB 2 , they henceforth will be referred to as B 2 -IsoTx.

EXPERIMENTAL PROCEDURES
Reagents-Methoxyamine HCl, FeCl 3 , ascorbate, ADP, pentafluorobenzyl (PFB) bromide, diisopropylethylamine, and Apis mellifera venom phospholipase A 2 were obtained from Sigma. Dimethylformamide, undecane, and sodium borohydride were obtained from Aldrich. N, O-Bis(trimethylsilyl) Analysis of Isothromboxanes-IsoTxs were analyzed by gas chromatography (GC)/negative ion chemical ionization (NICI) mass spectrometry (MS) using a modification of methods described previously for the analysis of TxB 2 (10). Briefly, 1.5 ng of [ 2 H 3 ]TxB 2 internal standard was initially added to a biological fluid and adjusted to pH 3 with 1 M HCl. The sample was applied to a C-18 Sep-Pak cartridge that had been prewashed with 5 ml of methanol and 5 ml of H 2 0 (pH 3). The cartridge was then washed with 10 ml of H 2 O (pH 3) followed by 10 ml heptane, and compounds were eluted with 10 ml of ethyl acetate and evaporated to dryness under nitrogen. Compounds were subsequently methoximated by treatment with 250 l of a 2% solution of aqueous methoxyamine HCl for 30 min at room temperature. Compounds were extracted with 1 ml of ethyl acetate, and the organic layer was evaporated under nitrogen. Compounds were then converted to a PFB ester by addition of 40 l of a 10% solution of PFB bromide in acetonitrile and 20 l of 10% diisopropylethylamine in acetonitrile and incubated for 30 min at 37°C. Reagents were dried under nitrogen, and the residue was reconstituted in 30 l of chloroform and 20 l of methanol and chromatographed on a silica TLC plate to the top in a solvent system of ethyl acetate: methanol (98:2, v/v). The O-methyloxime and PFB ester derivative of TxB 2 (approximately 5 g), was chromatographed on a separate lane and visualized with 10% phosphomolybdic acid in ethanol by heating. The R F of the derivatized TxB 2 standard in this solvent system was ϳ0.46. Compounds migrating in the region 1.5 cm above and below the standard were scraped from the TLC plate, extracted with 1 ml of ethyl acetate, and dried under nitrogen.
Following TLC purification, compounds were converted to trimethylsilyl (TMS) ether derivatives by addition of 20 l of N,O-bis(trimethylsilyl)trifluoroacetamide and 10 l of dimethylformamide. The sample was incubated at 37°C for 10 min and then dried under nitrogen. The residue was redissolved for GC/MS analysis in 10 l of undecane, which had been stored over a bed of calcium hydride. GC/NICI MS was carried out on a Nermag R10-10C mass spectrometer interfaced with a Digital DEC-PDP computer. GC was performed using a 15-m, 0.25-m film thickness, DB-1701 fused silica capillary column (J & W Scientific, Folsom CA). The column temperature was programmed from 190°to 300°C at 20°C/min. The major ion generated in the NICI mass spectrum of the PFB ester, O-methyloxime, and TMS ether derivative of TxB 2 , which would be the same ion generated by IsoTx, was the m/z 614 carboxylate anion M Ϫ 181 (M Ϫ ⅐ CH 2 C 6 F 5 ). The corresponding ion generated by the [ 2 H 3 ]TxB 2 internal standard was m/z 617. Levels of endogenous B 2 -IsoTx in a biological sample were calculated from the ratio of the area under the m/z 614 chromatographic peaks to the m/z 617 chromatographic peak. In some experiments, compounds were reacted with trimethylsilyimidazole, subjected to catalytic hydrogenation, or reduced with sodium borohydride following TLC purification as described (7). IsoTxs were also analyzed by GC/electron ionization (EI) MS as methyl ester O-methyloxime and TMS ether derivatives. Purification and derivatization of compounds for analysis by GC/EI MS were as noted above, except the methyl ester derivatives were formed by treatment of compounds with excess ethereal diazomethane (7).
Analysis of F 2 -and D 2 /E 2 -IsoPs-Purification, derivatization, and analysis of F 2 -IsoPs and D 2 /E 2 -IsoPs by GC/NICI MS were performed as described (7,8) Quantification of either F 2 -IsoPs or D 2 /E 2 -IsoPs in the present studies differed from previous reports in that the amounts of endogenous IsoPs were determined by comparing the ratios of the area under the chromatographic peaks representing endogenous material to that of the respective standard.
Extraction, Purification, and Hydrolysis of Phospholipids-1-Palmitoyl-2-arachidonylphosphatidylcholine oxidized in vitro or lipids from livers of CCl 4 -treated rats were extracted as described (6,11) Depending on the experiment, 0.005% butylated hydroxytoluene was added to the lipid extracts during the extraction procedure. The lipid extracts (containing approximately 1 mol of phospholipid) were then hydrolyzed by chemical saponification or by reaction with A. mellifera venom phospholipase A 2 (approximately 200 g) as described (6,8) and subsequently analyzed for free B 2 -IsoTx. As a positive control for phospholipase A 2 activity, phosphatidylcholine containing [ 3 H]arachidonate in the sn-2 position was added to the incubation mixture, and the percent of radiolabeled arachidonate released was determined as described (6,8). In all experiments, Ͼ95% of esterified [ 3 H]arachidonate was released.
Oxidation of Arachidonic Acid and Arachidonyl Phosphatidylcholine in Vitro-Arachidonic acid and arachidonyl phosphatidylcholine were oxidized for 30 min using a Fe/ADP/ascorbate oxidizing system as described (12). Animal Model of Free Radical-induced Lipid Peroxidation-Free radical-catalyzed lipid peroxidation was induced in rats by intragastric administration of CCl 4 as described previously (13). At various time intervals, animals were sacrificed, and the livers were removed, snap frozen in liquid N 2 , and either processed immediately or stored at Ϫ70°C. In some experiments, animals were pretreated with 5 mg/kg indomethacin at 24, 12, and 2 h prior to receiving CCl 4 . This has been previously shown to inhibit the cyclooxygenase Ͼ90% (14).

Evidence of the Formation of Isothromboxanes in Vitro-
Previously, we had shown that oxidation of arachidonic acid in vitro results in the formation of large amounts of both F 2 -IsoPs and D 2 /E 2 -IsoPs (7,8). Thus, we initially explored whether IsoTxs are also formed in vitro by analyzing arachidonic acid that had been oxidized with Fe/ADP/ascorbic acid. As described above, IsoTxs were detected using an assay originally developed for cyclooxygenase-derived TxB 2 . The selected ion current chromatograms obtained from this analysis monitoring m/z 614 for B 2   The finding that large quantities of a series of compounds were formed during oxidation of arachidonic acid in vitro that had TLC and GC/MS properties similar to those of TxB 2 would be consistent with their being B 2 -IsoTx. However, additional experimental approaches were used to obtain further evidence that the compounds detected in oxidized arachidonic acid were B 2 -IsoTxs. First, no peaks were present when m/z 613 was monitored, indicating that the m/z 614 peaks were not natural isotope peaks of compounds generating an ion of less than 614 Da. When the compounds were analyzed as [ 2 H 9 ]TMS ether derivatives, the m/z 614 peaks all shifted upward 27 Da, indicating that the compounds have three hydroxyl groups. When the compounds were analyzed as [ 2 H 3 ]O-methyloxime derivatives, the m/z 614 peaks all shifted upward 3 Da, indicating that they contain one carbonyl group. When the compounds were analyzed following catalytic hydrogenation, there was a disappearance of the m/z 614 peaks and the appearance of new intense peaks 4 Da higher at m/z 618 (Fig. 2). No peaks were detected at m/z 616 or 620. This indicated that all of the compounds contained two double bonds. Collectively, these results indicated that the compounds represented by the m/z 614 peaks contain the same functional groups and the number of double bonds expected for the PFB ester, O-methyloxime, and TMS ester derivative of B 2 -IsoTx.
A unique feature of TxB 2 is that it contains a hemiacetal ring, which exists in aqueous solution in an equilibrium between open and closed forms. Thus, powerful evidence that the compounds detected were B 2 -IsoTxs would be to demonstrate that these compounds contain a hemiacetal ring. Such evidence can be obtained using different derivatization and chemical modification approaches (15). As shown in Fig. 3A, if putative B 2 -IsoTxs are first reacted with methoxyamine, derivatives will be formed in which the hemiacetal ring is open. Subsequent conversion to PFB ester and TMS ether derivatives would be expected to result in a series of compounds with a major fragment ion of 614 Da (M Ϫ 181 and M Ϫ ⅐ CH 2 C 6 F 5 ) when analyzed by GC/NICI MS. The selected ion monitoring analysis of presumed IsoTx derivatized in this manner has been previously discussed and is shown in Fig. 1. If, on the other hand, as shown in Fig. 3B, the treatment with methoxyamine is omitted, and the compounds are converted to PFB ester and TMS ether derivatives, the hemiacetal ring will remain closed. Derivatives of these compounds would be expected to generate major fragment ions of 585 Da (M Ϫ 181) when analyzed by NICI MS. Results using this derivatization approach are shown in Fig. 4A. As is evident, in the upper m/z 585 chromatogram, a series of chromatographic peaks are present that elute at a similar retention time to the [ 2 H 3 ]TxB 2 internal standard represented in the lower m/z 588 chromatogram. Finally, as shown in Fig. 4C, if the carbonyl at C-11 in the open ring form is first reduced with NaBH 4 followed by conversion to PFB ester and TMS ether derivatives, the major M Ϫ 181 fragment ion would be generated at 659 Da. Results of this analysis are shown in Fig. 4B. Again, a series of m/z 659 peaks elute from the GC at a retention time similar to this derivative of the TxB 2 internal standard. Collectively, the results of these studies provide additional significant evidence that these compounds contain a hemiacetal ring as does TxB 2 .
Analysis of B 2 -IsoTx by EI MS-To obtain more direct evidence that the compounds detected by NICI MS were B 2 -IsoTx, the compounds were analyzed as methyl ester, O-methyloxime, and TMS ether derivatives by EI MS. The results of this analysis yielded a series of compounds eluting over approximately a 20-s period from the capillary GC column, which yielded mass spectra with characteristics of the EI mass spectrum of TxB 2 . One of the mass spectra obtained from a major peak is shown in Fig. 5. Other mass spectra obtained from the analysis of the other peaks were similar to that shown in Fig. 5, except that either the relative abundance of some of the fragment ions varied or some of the lower molecular weight fragment ions were different. In the mass spectrum shown, there is a prominent ion at m/z 629, representing the molecular ion.  3 SiO ϩ ϭCHCH 2 CHϭNOCH 3 ). Of particular interest is the major fragment ion of 243 Da. This ion is not present in the mass spectrum of this derivative of cyclooxygenase-derived TxB 2 . However, this is an expected ion resulting from ␣ cleavage of the trimethoxysiloxy substitutent at C-8, as depicted in the regioisomer shown in Fig. 5 (7, 16). Thus, this EI mass spectral data provide additional confirmatory evidence for the formation of IsoTxs by nonenzymatic peroxidation of arachidonic acid.
Analysis for the Presence of B 2 -IsoTx Esterified to Phospholipids in Vivo-Since the above results suggested that IsoTxs could be formed in vitro, we investigated whether these compounds may also be formed in vivo. Previously, we had shown that F 2 -IsoPs and D 2 /E 2 -IsoPs are initially formed in situ from arachidonic acid esterified in tissue phospholipids and subsequently released preformed (6,8). Therefore, we examined whether IsoTxs are also formed esterified in phospholipids in livers of rats that had been treated with CCl 4 to induce lipid peroxidation. To investigate this, lipids were extracted from the livers, subjected to hydrolysis using methanolic potassium hydroxide, and analyzed as free compounds. The results of this analysis are shown in Fig. 6. A series of m/z 614 peaks was present in a pattern very similar to that obtained from analysis of arachidonic acid oxidized in vitro, although the relative abundances of the various peaks differ slightly (cf. Fig. 1). Essentially identical results were obtained when phospholipids were hydrolyzed enzymatically with phospholipase A 2 from A. mellifera (data not shown). Table I compares the amounts of the B 2 -IsoTxs with D 2 /E 2 -IsoPs and F 2 -IsoPs measured following hydrolysis of lipids from the same livers of both untreated and CCl 4 -treated rats. The quantities of B 2 -IsoTx measured following hydrolysis of lipids from livers of CCl 4 -treated rats were 41-fold higher than those in untreated rats. Levels of free B 2 -IsoTx measured in lipid extracts that were not subjected to hydrolysis were Ͻ1% of the levels measured following hydrolysis (n ϭ 4), suggesting that the compounds detected following saponification were released from an acyl linkage on phospholipids. Pretreatment of animals with indomethacin prior to CCl 4 administration with a dosage regimen previously shown to inhibit cyclooxygenase activity by Ͼ90% (14) did not affect levels of the compounds measured (p Ͼ 0.7, Student's t test; n ϭ 4), indicating that the cyclooxygenase enzyme is not involved in their formation. Previously we had shown that butylated hydroxytoluene markedly suppresses the formation of F 2 -IsoPs by autoxidation in vitro (7). The presence of butylated hydroxytoluene (0.005%) in the extraction solution, however, did not affect levels of IsoTx measured (p Ͼ 0.6; n ϭ 4), arguing that these compounds are not formed ex vivo by autoxidation during sample processing.
Experiments were then carried out to obtain further evidence of the identity of the compounds represented by the m/z 614 peaks in Fig. 6  pearance of the m/z 614 peaks and the appearance of new intense peaks 4 Da higher at m/z 618. No peaks were detected at m/z 616 or 620. This indicated that all of the compounds contained two double bonds. Furthermore, analysis of the compounds using the different derivatization approaches and chemical modification as outlined in Fig. 3 indicated the presence of a hemiacetal ring, as was found in the compounds generated in vitro (Fig. 4). Collectively, these results indicated that the compounds generated in vivo represented by the m/z 614 peaks contain a hemiacetal ring and the same functional groups and number of double bonds as would be expected for B 2 -IsoTx.
Analysis for the Presence of Free B 2 -IsoTx in the Circulation in CCl 4 -treated Rats-We have previously demonstrated that F 2 -and D 2 /E 2 -IsoPs are initially formed esterified to tissue phospholipids in CCl 4 -treated rats and subsequently released into the circulation preformed (6,8,17). Thus, we examined whether increased concentrations of these IsoTxs could also be detected free in the circulation of rats 4 h following administration of CCl 4 to induce lipid peroxidation. In these studies, animals were pretreated with indomethacin to inhibit cyclooxygenase-derived TxB 2 from blood elements. IsoTxs could not be detected in plasma from normal rats that had not been treated with CCl 4 (lower limit of detection, 5 pg/ml; n ϭ 4). However, following treatment of rats with CCl 4 , B 2 -IsoTx were detected in plasma at concentrations of 185 Ϯ 30 pg/ml (n ϭ 4), representing an increase of 37-fold. The pattern of peaks was essentially identical to that shown in Fig. 6. We have previously shown that increases in levels of F 2 -IsoPs esterified to circulating lipids parallel increases in free levels after administration of CCl 4 (17). Therefore, we also quantified levels of B 2 -IsoTxs esterified in circulating plasma lipids and found that CCl 4 administration increased levels from 13 Ϯ 12 to 237 Ϯ 63 pg/ml (n ϭ 4). In summary, these studies support the concept that B 2 -IsoTxs are formed in situ on phospholipids and subsequently released preformed into the circulation.

DISCUSSION
These studies report the discovery that thromboxane-like compounds, termed IsoTxs, are formed both in vitro and in vivo by nonenzymatic free radical-catalyzed peroxidation of arachidonic acid. Analogous to the formation of F 2 -IsoPs and D 2 /E 2 -IsoPs, B 2 -IsoTxs are also formed in situ esterified to phospholipids and released in the free form, presumably by a phospholipase. A pathway for the formation of these compounds is outlined in Fig. 7. It is identical to that outlined previously for the formation of F 2 -IsoPs and D 2 /E 2 -IsoPs involving the formation of the bicyclic endoperoxide intermediates (6,8). In the formation of IsoTx, however, the endoperoxides undergo rearrangement to form TxA 2 -like compounds, termed A 2 -IsoTxs, which then rapidly decompose to more stable TxB 2 -like molecules, termed B 2 -IsoTxs. Analogous to the formation of F 2 -IsoPs, four regioisomers of B 2 -IsoTxs are formed, each of which can theoretically comprise eight racemic pairs of diastereomers.
Although B 2 -IsoTx can be generated both in vitro and in vivo from arachidonic acid, the chemistry involved in the conversion of isoprostane endoperoxides to IsoTx is not entirely clear. It may, however, be similar to that proposed for the conversion of the cyclooxygenase-derived endoperoxide PGH 2 to TxA 2 by the enzyme thromboxane synthase (9). Thromboxane synthase is a cytochrome P450 enzyme containing a catalytic iron moeity at its active site. Hecker and Ullrich (9) have proposed that the formation of TxA 2 initially involves complexing of the Fe 3ϩ in the enzyme active site with the oxygen at C-9 on the endoperoxide PGH 2 . This is followed by homolytic scission of the endoperoxide bond, leading to formation of an alkoxyl radical. Subsequently, ␤ scission of the C-11-C-12 bond occurs, followed by rearrangement of the molecule to form TxA 2 , which then rapidly decomposes to TxB 2 . A similar mechanism might also explain the nonenzymatic formation of B 2 -IsoTx from the iron-catalyzed peroxidation of arachidonic acid in vitro. Arguing against this mechanism, however, is the observation by Hecker and Ullrich (9) that reaction of inorganic Fe ϩ3 with PGH 2 does not result in the formation of significant quantities of TxB 2 . In those studies, however, Hecker and Ullrich (9) did find that large amounts of TxB 2 were generated from PGH 2 if iron was present complexed in a porphyrin such as hemin. Thus, it is possible that the formation of IsoTx in vivo might be catalyzed by porphyrin-containing compounds or Fe-containing enzymes, including Tx synthase. On the other hand, the fact that large amounts of B 2 -IsoTx can be formed in vitro when  arachidonic acid is oxidized with Fe/ADP/ascorbate would suggest that complexed iron or Fe-containing enzymes are not necessary for the formation of IsoTx.
It should be noted that the quantities of B 2 -IsoTxs that are formed in vivo are only slightly less than the amounts of D 2 / E 2 -IsoPs generated. Since the levels of many of the individual IsoPs in normal human biological fluids are at least an order of magnitude higher than cyclooxygenase-derived prostaglandins, the amounts of IsoTx that are produced in vivo are not trivial. We have previously reported that both F-and D/E-ring IsoPs possess potent biological activities (5, 8, 18 -20). Whether IsoTx possess biological activity will be difficult to ascertain. It is reasonable to assume that if IsoTx possess bioactivity, such activity would reside with the TxA-ring compounds rather than the TxB-ring compounds, analagous to cyclooxygenase-derived Txs, of which TxA 2 is bioactive but TxB 2 is devoid of biological activity. However, because the TxA-ring is extremely unstable, undergoing rapid hydrolysis to form the TxB-ring, isolation of A 2 -IsoTx for biological testing would be difficult, if not impossible.
It should also be mentioned that there are potentially important biological ramifications associated with the formation of IsoTx esterified in phospholipids. We previously reported that molecular modeling of phospholipids with F 2 -IsoPs esterified at the sn-2 position revealed them to be extremely distorted mol-ecules (6). Thus, the formation of isoprostane-containing phospholipids in settings of oxidant stress may have deleterious effects on membrane fluidity and integrity, well recognized sequelae of oxidant injury (21). Since we have now discovered that, in addition to F 2 -IP and D 2 /E 2 -IsoPs, IsoTxs are also formed esterified to phospholipids in large quantities, the total quantities of phospholipids containing products of the isoprostane pathway that may be formed in settings of free radical injury are substantially greater than previously thought.
In summary, we report the discovery that IsoTxs are formed in vivo as products of nonenzymatic free radical-catalyzed lipid peroxidation. Analogous to the formation of F 2 -IsoPs and D 2 / E 2 -IsoPs, IsoTxs are formed in situ esterified to phospholipids and subsequently released in free form. Further understanding the biological consequences of the formation of these novel compounds and mechanisms by which they are formed may provide valuable insights into the pathophysiology of oxidant injury.