Urinary metabolites of leukotriene B4 in the human subject.

Leukotriene B4 (LTB4) is a potent chemoattractant for neutrophils and is thought to play a role in a variety of inflammatory responses in humans. The metabolism of LTB4 in vitro is complex with several competing pathways of biotransformation, but metabolism in vivo, especially for normal human subjects, is poorly understood. As part of a Phase I Clinical Trial of human tolerance to LTB4, four human subjects were injected with 150 nmol/kg LTB4 with one additional subject as placebo control. The urine of the subjects was collected in two separate pools (0-6 and 7-24 h), and aliquots from these urine collections were analyzed using high performance liquid chromatography, UV spectroscopy, and negative ion electrospray ionization tandem mass spectrometry for metabolites of LTB4. In the current investigation, 11 different metabolites of LTB4 were identified in the urine from those subjects injected with LTB4, and none were present in the urine from the placebo-injected subject. The unconjugated LTB4 metabolites found in urine were structurally characterized as 18-carboxy-LTB4, 10,11-dihydro-18-carboxy-LTB4, 20-carboxy-LTB4, and 10,11-dihydro-20-carboxy-LTB4. Several glucuronide-conjugated metabolites of LTB4 were characterized including 17-, 18-, 19-, and 20-hydroxy-LTB4, 10-hydroxy-4,6,12-octadecatrienoic acid, LTB4, and 10,11-dihydro-LTB4. The amount of LTB4 glucuronide (16.7-29.4 pmol/ml) and 20-carboxy-LTB4 (18.9-30.6 pmol/ml) present in the urine of subjects injected with LTB4 was determined using an isotope dilution mass spectrometric assay before and after treatment of the urine samples with beta-glucuronidase. The urinary metabolites of LTB4 identified in this investigation were excreted in low amounts, yet it is possible that one or more of these metabolites could be used to assess LTB4 biosynthesis following activation of the 5-lipoxygenase pathway in vivo.

Leukotriene B 4 (LTB 4 , 1 (5S,12R)-dihydroxy-6,14Z-8,10E-eicosatetraenoic acid) is a biologically active metabolite of arachi-donic acid that is chemotactic for the human neutrophil and a potent lipid mediator of inflammation (1). LTB 4 is not stored within a cell but synthesized following activation of certain cells through release of arachidonic acid by the cytosolic phospholipase A 2 (2) and activation of the enzyme 5-lipoxygenase (3). The immediate product of the 5-lipoxygenase pathway is the reactive epoxide intermediate leukotriene A 4 that is transformed by leukotriene A 4 hydrolase into LTB 4 (4). A specific G-protein-coupled receptor for LTB 4 has now been cloned and expressed (5) and is known to be expressed by many cell types including the neutrophil (6). Through this receptor, LTB 4 is a potent stimulus of many functional responses of cells that are involved in immune responses (7) such as neutrophils, monocytes, macrophages, and various lymphocytes. LTB 4 is therefore thought to be an important component of the host defense mechanisms and is currently under investigation as a potent drug for the prophylaxis and/or treatment of infectious diseases.
The inactivation of endogenous LTB 4 takes place through metabolism so that little, if any, escapes into the circulation or appears in urine as an intact molecule. The metabolic transformation of LTB 4 has been studied in various cellular systems including the human neutrophil, which efficiently metabolizes LTB 4 by a specific cytochrome P-450-dependent pathway (CYP4F3) to produce 20-hydroxy-LTB 4 (20-OH-LTB 4 ) (8). Additionally, other cells carry out this same step of -oxidation but do so with different members of this unique family of P-450 enzymes including CYP4F2 and CYP4F4 (8). The initial -oxidized metabolite, 20-OH-LTB 4 , retains significant biological activity (9) but is further metabolized to biologically inactive 20-carboxy-LTB 4 (20-COOH-LTB 4 ) by both a P-450-dependent pathway (10,11) as well as an alcohol dehydrogenase-dependent pathway in certain cells (12). Formation of these -oxidation products has been observed after the incubation of LTB 4 with human hepatocytes (13). Once -oxidation has occurred, 20-COOH-LTB 4 can be further metabolized by ␤-oxidation into 18-carboxy-LTB 4 (18-COOH-LTB 4 ) and 16-COOH-LTB 3 (13,14) which requires formation of the CoA ester at the -terminal carboxyl moiety. These chain shortened products, as CoA esters, can undergo additional ␤-oxidation cycles from the -terminus to form substantially less lipophilic metabolites (14).
An additional pathway of LTB 4 metabolism identified in various human cells, including lung macrophages, monocytes, and keratinocytes, occurs via the 12-hydroxyeicosanoid dehydrogenase/⌬ 10 -reductase pathway (15,16). This pathway leads to oxidation of the hydroxyl group at C-12 with intermediate formation of the 12-oxo product, which is then reduced at the immediately adjacent 10,11 double bond by a ⌬ 10 -reductase. This leads to a series of 10,11-dihydro metabolites, which can be precursors foror ␤-oxidation. Metabolism by this pathway also results in a substantial reduction in biological activity (17).
LTB 4 can also participate in synthetic metabolic reactions by conjugation with polar molecules at either hydroxy or carboxyl substituents of LTB 4 and oxidized metabolites. The first LTB 4 conjugation metabolite observed was a taurine conjugate of 18-COOH-LTB 4 formed in the rat hepatocyte (14). Additionally, when LTB 4 was incubated with human keratinocytes, a glutathione adduct was observed as an intermediate metabolite (16). Finally, glucuronide conjugates of 20-COOH-LTB 4 , LTB 4 , and 10,11-dihydro-LTB 4 have been observed as products following incubation of LTB 4 with human hepatocytes (13).
Interest in LTB 4 as an endogenous inflammatory mediator has been somewhat hampered by an inability to assess in vivo production of LTB 4 either in normal individuals or individuals thought to have the leukotriene pathway of arachidonate metabolism activated as a result of pathologic events. One quite successful strategy in assessing in vivo production of prostaglandins and thromboxane A 2 has been the quantitative analysis of corresponding metabolites appearing in urine using sensitive and specific assays for each unique metabolite (18,19). This approach has also been used to monitor the production of leukotriene C 4 in human subjects through measurement of leukotriene E 4 excreted into urine (20). LTB 4 has been observed in urine only in individuals with a deficiency of fatty aldehyde dehydrogenase (Sjögren-Larsson Syndrome) and Zellweger patients who lack peroxisomes and have limited ␤-oxidation of lipids (21,22).
In the course of investigations into LTB 4 as a potential agent useful in the prophylaxis or treatment of infections, synthetic LTB 4 was administered into healthy human volunteers in studies of safety, tolerability, pharmacokinetics, and pharmacodynamics. These experiments afforded the possibility to examine the urine for excreted LTB 4 and LTB 4 metabolites when known quantities of LTB 4 were administered intravenously to human subjects. Detailed study of the urine was therefore undertaken to investigate whether or not previously identified metabolites of LTB 4 described above were present in human subjects.
Preparation of LTB 4 for Injection-In order to prepare LTB 4 for injection into human subjects, ethanolic LTB 4 was converted into the sodium salt by the addition of 1 eq of aqueous sodium hydroxide. This ethanolic solution of neutralized LTB 4 was evaporated to dryness under reduced pressure at 40°C. The clear yellowish oily residue was immediately redissolved in sodium phosphate (30 mM) buffered 0.7% sodium chloride containing 0.01% EDTA to a final concentration of 0.7 mg/ml LTB 4 . The pH of the solution was adjusted to 7.5 Ϯ 0.1, and the solution was sterilized by filtration through a 0.22-m microporous membrane filter. The sterile solution of LTB 4 was then transferred into 10-ml clear glass vials, which were sealed and stored at Ϫ80°C. Solutions of LTB 4 for injection were analyzed prior to use to assess sterility, endotoxin content, concentration, and impurity levels. The purity of LTB 4 for injection was also ϳ96% with a profile of impurities similar to that of the starting material with the exception that the ␦-lactone of LTB 4 was not present, but an impurity corresponding to the ethyl ester of LTB 4 (ϳ0.5%) was now detectable.
Injection of LTB 4 -Subjects enrolled in this study were members of the community at large (Quebec, Canada) who were screened according to inclusion and exclusion criteria. In addition, individuals were tested for the presence of alcohol as well as human immunodeficiency virus and hepatitis B and C. Subjects abstained from alcohol for 2 days prior to inclusion in the study. Subjects also provided written informed consent and agreed to abide with the study restrictions, including collection of urine samples. Study subjects excluded were those with any clinically significant abnormalities as assessed from an electrocardiogram, clinical laboratory tests, or had abused alcohol or used illicit drugs. Other exclusion criteria included positive hepatitis B, hepatitis C, or human immunodeficiency virus or an active infection caused by herpes simplex virus, or a virus of the orthomyxovirus or rhinovirus families, and any clinically significant surgery or illness within the previous month. None of the subjects had used antiviral, antibiotic, or corticosteroids within 14 days preceding the study or had used an investigational drug within 30 days preceding this study.
After overnight fasting for 10 h, subjects were injected with LTB 4 as a bolus intravenous dose of 50 g/kg (150 nmol/kg) or an identical placebo injection of sterile saline. Both the LTB 4 and the placebo were administered in a total volume of ϳ5 ml as a bolus injection completed within ϳ1 min. A total of four subjects received LTB 4 and one subject received the placebo injection.
The subjects fasted for at least 1 h after drug administration, after which time they were served a controlled breakfast, and standard meals at ϳ4, 9, and 13 h post-injection. Water was permitted ad libitum. Subjects remained lying down for 1 h after LTB 4 injection, and vigorous physical activity was prohibited at all times during the study. Subjects were asked to empty their bladder prior to administration, and urine was then collected at two interval pools of 0 -6 and 7-24 h postinjection. The urine samples were collected in polypropylene containers and kept in an ice water bath; at the end of both collection periods, the urine samples were quickly frozen and stored at Ϫ80°C until analyzed. Throughout the study, subjects were monitored for adverse events, and a qualified medical investigator was on-site during injection of LTB 4 and until 4 h after injection.
Urine Extraction-The two urine sample pools from each of the five subjects (representing 0 -6 and 7-24 h time periods after injection of LTB 4 or placebo) were thawed and kept at 4°C. The volume of urine in each sample was measured, then transferred into polypropylene 50-ml conical tubes, and centrifuged at 4°C for 10 min at 750 ϫ g. The urine samples were then acidified to pH 3.7-3.8 using formic acid and subjected to solid phase extraction using 1-g cartridges (Waters, Milford, MA). The cartridges were first conditioned with 2 volumes of methanol (40 ml) and then equilibrated with 2 volumes of water. The acidified urine samples were then loaded onto the cartridges and washed with 2 volumes of water. Solvent flow through the extraction cartridges was 20 Ϯ 5 ml/min. Lipophilic metabolites were eluted with 1 volume of methanol directly into 20-ml glass ampoules. A maximum of 500 ml of urine was loaded on each 20-ml cartridge, and when the urine sample volume exceeded 500 ml, the extraction cartridge was reconditioned, and the remaining volume of the urine sample was extracted using the same reconditioned solid phase extraction cartridge. Two extraction blanks (0.9% saline acidified with formic acid) were also prepared. The ampoules were flame-sealed and stored at Ϫ80°C until analysis. The urine sample volumes varied from 160 to 400 ml for the 0 -6-h samples and 625-1050 ml for the 7-24-h collections.
Aliquots of the methanol eluate, typically 5%, were taken for subsequent extraction, purification, and treatment with ␤-glucuronidase. For some studies, solid phase extraction aliquots from the 0 -6-h collection of each subject in the study were evaporated to dryness under vacuum and then reconstituted in 5 ml of 100 mM phosphate buffer, pH 7. The pH of the mixture was adjusted to 3.8 using formic acid, and the leukotrienes present in the solid phase-extracted urine were extracted using 1:1 (v/v) hexane/ethyl acetate. After vortexing 4 times for 15 s, each sample was centrifuged at 135 ϫ g for 5 min to separate the two layers. This liquid-liquid extraction procedure was repeated 3 additional times, and the combined organic layers from each extraction were pooled and taken to dryness under vacuum.
Normal-phase HPLC and RP-HPLC-The 1:1 hexane/ethyl acetate extract was further purified by normal phase chromatography using an Ultremex 5-m silica 250 ϫ 4.6 mm column (Phenomenex, Torrance, CA). The normal phase solvents used were 90:10:0.1 hexane/isopropyl alcohol/acetic acid (solvent A) and 90:10:0.3 isopropyl alcohol, 20 mM ammonium acetate, acetic acid (solvent B). The initial mobile phase was 10% solvent B at a flow rate of 1 ml/min. This initial mobile phase was held for 3 min and then a linear gradient was started to 30% solvent B at 13 min. This was followed by a second linear gradient to 40% solvent B at 25 min and finally a third linear gradient to 85% solvent B at 43 min. The column effluent was monitored using UV detection at 270 and 235 nm, and one fraction was collected each minute for 43 min.
Due to the complexity of the urine matrix, it was necessary to perform reversed phase chromatography of the normal phase fractions of interest in order to increase the extent of purification. Normal phase fractions of interest were pooled, dried down under vacuum, and analyzed by reversed phase HPLC using a Synergi 10-m Hydro-RP 250 ϫ 4.6 mm column (Phenomenex, Torrance, CA). The reversed phase solvents used were 8.3 mM acetic acid adjusted to pH 5.7 with ammonium hydroxide (solvent A) and 65:35 acetonitrile/methanol (v/v) (solvent B). The initial mobile phase was 15% solvent B at a flow rate of 1 ml/min. This initial mobile phase was held for 3 min, and then a linear gradient was started to 70% solvent B at 43 min. This was followed by a second linear gradient to 100% solvent B at 55 min. The column effluent was monitored using UV detection at 270 and 235 nm, and one fraction was collected each minute for 55 min.
The final analysis was carried out using reversed phase fractions of interest, which had been dried under vacuum, and then subjected to a second reversed phase HPLC separation (subsequently termed the analytical RP-HPLC) using on-line RP-HPLC with electrospray mass spectrometry (LC/MS), which employed a 150 ϫ 1.0-mm Ultremex C 18 column (Phenomenex, Torrance, CA) at a flow rate of 50 l/min. The same reversed phase solvents were used that are described above. The initial mobile phase was 15% solvent B, which was held for 3 min, and then a linear gradient was started to 60% solvent B in 30 min. This was followed by a second linear gradient to 100% solvent B at 45 min.
Electrospray Mass Spectrometry-Mass spectrometry was performed on a Sciex API III ϩ triple quadrupole mass spectrometer (PE-Sciex, Thornhill, Ontario, Canada). The mass spectrometry experiments were carried out in the negative ion mode with a spray voltage of Ϫ2800 V and an orifice voltage of Ϫ65 V. Collisional activation and multiple reaction monitoring (MRM) data were obtained using an offset potential of 20 eV and argon as the collision gas at a thickness of 230 ϫ 10 13 molecules/cm 2 .
Gas Chromatography/Mass Spectrometry-Reversed phase fractions of interest were taken to dryness under vacuum and derivatized for GC/MS analysis by the addition of 10% N,N-diisopropylethylamine in acetonitrile (50 l) followed by the addition of 10% pentafluorobenzyl bromide in acetonitrile (50 l). These samples were kept at room temperature for 30 min and evaporated under a stream of dry nitrogen. The samples were then dissolved in 500 l of methylene chloride and introduced onto a silica solid phase extraction column (Supelco, Bellefonte, PA), which was conditioned with methanol and rinsed thoroughly with hexane prior to sample addition. After sample introduction, the silica SPE column was rinsed with 6 ml of hexane, and the PFB derivatives of interest were eluted with 3 ml of ethyl acetate. The ethyl acetate eluate was dried down under nitrogen and further derivatized with the addition of 50 l of acetonitrile and 50 l of bis(trimethylsilyl)trifluoroacetamide by incubating at 60°C for 20 min followed by evaporation under nitrogen. The derivatized sample was dissolved in 10 l of acetonitrile and subjected to GC/MS analysis. A gas chromatograph/mass spectrometer (Trace 2000, Thermo Finnigan, San Jose, CA) was employed for both electron ionization (EI) and negative ion chemical ionization (NCI) analysis. NCI spectra were obtained using methane as the moderating gas, and electron ionization (EI) spectra were obtained using an electron energy of 70 eV. The [M Ϫ PFB] Ϫ ions obtained from NCI (23) were used to determine the retention times of the -hydroxylated LTB 4 compounds of interest. EI was used to provide detailed structural information regarding the hydroxyl group position from fragmentations that occur adjacent to the trimethylsilyl ether positions (24,25).
Hydrolysis of Glucuronides-An aliquot of methanol eluate (5%) from the 0 -6-h pool of each subject injected with LTB 4 or the placebo was dried down under vacuum and brought back up in 4.5 ml of 100 mM phosphate buffer, pH 7. This solution was adjusted to pH 7, and 5,000 units of ␤-glucuronidase in 0.5 ml of 100 mM phosphate buffer (pH 7) was added (26). This mixture was incubated in a 37°C shaking water bath for 16 h for the quantitation experiments. For the time course experiments, the incubation times ranged from 1 to 24 h. After treatment, the liquid-liquid extraction procedure described in the initial metabolite extraction section was carried out followed by normal and reversed phase HPLC as described.

Quantitation of LTB 4 Glucuronide and 20-COOH-LTB 4 -
The quantity of LTB 4 glucuronide and 20-COOH-LTB 4 present in the urine of subjects injected with LTB 4 and the placebo was determined using a stable isotope dilution LC/MS/MS protocol, essentially as described previously (27). The amount of LTB 4 glucuronide present in urine was determined by adding 150 pmol of d 4 -LTB 4 as internal standard to an aliquot of the solid phase extracted urine in 5 ml of 100 mM phosphate buffer, pH 7. Hydrolysis of glucuronides was then carried out using the procedure described above followed by extraction and HPLC separation.  a Reversed phase retention time using the reversed phase solvent system. b Identification of -oxidized metabolite elution order required GC/MS analysis and detection as the pentafluorobenzyl ester, trimethylsilyl ether derivatives (Fig. 3B).
c Ion transitions reported previously (13,26) for each indicated m/z pair in a tandem quadrupole mass spectrometer. d The reversed phase HPLC retention time and MRM transitions monitored are for the aglycone formed after hydrolysis by ␤-glucuronidase. e Chromophore identified as characteristic for a conjugated triene with evidence of vibronic shoulders (Ϯ10 nm) on either side of the reported max . f Chromophore identified as characteristic for a conjugated diene at the reported max .

RESULTS
The administration of LTB 4 to four healthy male subjects at the dose of 150 nmol/kg was well tolerated with no report of severe adverse effects. A full report of the safety and tolerability of the intravenous administration of LTB 4 to healthy human subjects as well as pharmacokinetic and pharmacodynamic data will be published separately.
The urine of subjects treated with LTB 4 , as well as the one placebo-treated subject, was collected into two separate pools. The first pool was 0 -6 h after administration of LTB 4 or placebo, and the second collection was 7-24 h after the procedure. Due to the complexity of the urine matrix, a total of five purification steps were performed in order to prepare the sample for mass spectrometric analysis. The first step in purification involved a solid phase extraction using a reversed phase protocol to remove lipophilic components from the urine matrix. The pH of the urine was adjusted to pH 3.8 prior to solid phase extraction in order to maximize recovery of LTB 4 metabolites (28). The next purification step was a liquid-liquid extraction carried out on separate aliquots of the methanol eluate using a hexane/ethyl acetate extraction procedure that had been shown to result in excellent recovery of LTB 4 as well as several of the target metabolites (data not shown). The solventextracted components were then subjected to normal phase HPLC using a silica column with gradient elution maximized for the separation of LTB 4 , 20-OH-LTB 4 , and 20-COOH-LTB 4 (data not shown). The retention times of these leukotrienes were 7.6, 14.9, and 16.8 min, respectively, and these specific fractions (Ϯ1 min) were collected. The normal phase fractions of interest were introduced onto a 4.6-mm reversed phase HPLC column where the retention time of LTB 4 , 20-OH-LTB 4 , and 20-COOH-LTB 4 standards were 36.1, 21.8, and 17.3 min, respectively. Finally, the reversed phase fractions that eluted at the expected retention times (Ϯ1 min) were introduced onto a 1-mm analytical RP-HPLC column and analyzed by on-line mass spectrometry (LC/MS). The retention times for standards and known UV absorption characteristics of leukotrienes were used as a guide to screening additional fractions than those noted for metabolites of LTB 4 . For example, fractions that eluted from the normal phase column after 16.8 min were analyzed for metabolites more polar than 20-COOH-LTB 4 . Another experimental strategy used to verify the identity of LTB 4 metabolites was to monitor two MRM transitions for each known LTB 4 metabolite. This strategy was critical to identify unambiguously the metabolites listed below. The MRM transitions used in this study were obtained from a previous study of the collision-induced dissociation mass spectra derived from the corresponding metabolites (13,29).
Collisional activation of intact LTB 4 glucuronide-conjugated metabolites had been shown previously to result in a neutral loss of 176 daltons to yield the aglycone (parent) anion as well as a glucuronic acid anion at m/z 193 (13). These ions and the neutral loss event did not provide structural details concerning the aglycone portion of the metabolite but did provide relevant information as to the existence of glucuronides eluting from the HPLC column. Because glucuronide conjugates of 20-COOH-LTB 4 , LTB 4 , and 10,11-dihydro-LTB 4 had been observed previously (13) following incubation of LTB 4 with human hepatocytes, a constant neutral loss scan was carried out for the LC/MS analysis of the solid phase extracted urine. However, a complex mixture of components was clearly evident in the placebo control as well as in the urine from the individual receiving the LTB 4 injection. Despite the solid phase extraction of this sample, it was clear that the use of a single ion transition, in this case a constant neutral loss of 176 daltons, as well as a precursor mass specific for LTB 4 were not sufficiently unique criteria to identify specific glucuronides present in the complex urine mixture. The most likely reason for this was that multiple LTB 4 glucuronide conjugates (see below) were present and each one at sufficiently low levels and that naturally occurring glucuronide conjugates from unrelated molecules produced signals that were greater in magnitude compared with any single LTB 4 glucuronide. Thus it was necessary to carry out the second stage of reversed phase HPLC as well as hydrolyze the glucuronides prior to analysis. A summary of the metabolites that were found in urine of all the subjects injected with LTB 4 is reported in Table I. The urine from the subject that received the placebo injection was also analyzed, but none of the metabolites reported (Table I) satisfied all of the criteria in chromatography as well as tandem mass spectrometry to suggest the presence of the metabolite. Specific details are provided below.
18-COOH-LTB 4 and 10,11-dihydro-18-COOH-LTB 4 -The pooled normal phase fractions eluting between 19 and 20 min were injected onto a 4.6-mm reversed phase HPLC column and a component eluted that produced a signal at 270 nm at 12 min with an UV absorption spectrum consistent with that of a conjugated triene (data not shown). Subsequent LC/MS/MS analysis of the components eluting between 11 and 13 min from the 4.6-mm reversed phase column was then carried out for this region where components somewhat less lipophilic than 20-COOH-LTB 4 eluted. The specific transitions m/z 337 3 141 and 337 3 195 were investigated since they correspond to 18-COOH-LTB 4 , a known metabolite in cell studies (13). Both transitions were clearly present in the unhydrolyzed urine (not treated with ␤-glucuronidase) at a retention time of 13.1 min (Fig. 1, a and b), which had a retention time shorter than that expected for 20-COOH-LTB 4 (17.1 min). The unique UV absorption spectrum as well as the coelution of each MRM transition confirmed the presence of 18-COOH-LTB 4 in the aliquot of urine taken from the subject injected with LTB 4 . This experiment was repeated for all the subjects (data not shown), but the placebo-treated individual did not generate a component eluting at the appropriate retention time. Additionally, the MRM transitions m/z 339 3 115 and m/z 339 3 141 were monitored to determine whether or not 10,11-dihydro-18-COOH-LTB 4 was present in these HPLC fractions. Both of these MRM transitions were present in the unhydrolyzed urine (Fig. 1, c and d) at a slightly more lipophilic retention time (13.6 min) compared with 18-COOH-LTB 4 , which suggested that 10,11-dihydro-18-COOH-LTB 4 was present in the urine of subjects injected with LTB 4 .

20-COOH-LTB 4 and 10,11-Dihydro-20-COOH-LTB 4 -
The normal phase and reversed phase retention times of 20-COOH-LTB 4 were determined to be 16.8 and 17.1 min, respectively. Therefore, the pooled normal phase fractions 17-18 from each subject treated with LTB 4 were subjected to RP-HPLC on a 4.6-mm column, and a conjugated triene chromophore was observed for a component eluting at 17.3 min (Fig. 2a, inset).  (Fig. 2, a and b). In addition to monitoring the MRM transitions for 20-COOH-LTB 4 , the MRM transitions at m/z 367 3 115 and 367 3 169 were also monitored in order to assess whether or not 10,11-dihydro-20-COOH-LTB 4 was present in these HPLC fractions. Both of these MRM transitions were significant in the unhydrolyzed urine at a slightly more lipophilic retention time (17.6 min, Fig. 2, c and  d), which suggested that indeed 10,11-dihydro-20-COOH-LTB 4 was present in urine samples of subjects injected with LTB 4 . The collisional activation of m/z 367 yielded a product ion mass spectrum (Fig. 2c, inset) that was virtually identical to that published previously (13) for 10,11-dihydro-20-COOH-LTB 4 . The abundant product ions observed at m/z 115 and 169 were consistent with cleavage adjacent to the hydroxyl substituents with charge retention on the corresponding moiety as reported previously (29).

Glucuronide Conjugates of 20-, 19-, 18-, and 17-OH-LTB 4 -
The normal phase retention time of 20-OH-LTB 4 standard was 14.9 min, and the reversed phase retention time was 21.8 min using a 4.6-mm column. The pooled normal phase fractions 15-16 were separated on the 4.6-mm RP column, and fractions 21-23 of the ␤-glucuronidase-treated and untreated urine aliquots were analyzed by LC/MS/MS. Any metabolites of LTB 4 that were hydroxylated near the methyl terminus were detected using the MRM transition m/z 351 3 195 (29). This transition was clearly observed in urine that had been treated with ␤-glucuronidase (Fig. 3a) but was not observed in the urine not treated with ␤-glucuronidase. Furthermore, four separate HPLC components eluted within a 3-min time window at 21.4, 22.4, 22.9, and 23.6 min in the ␤-glucuronidase-treated urine. In order to establish the identity of the four components that share the MRM transition m/z 351 3 195, reversed phase fractions 21-23 were derivatized for GC/MS analysis as the pentafluorobenzyl ester (PFB) and trimethylsilyl ether (TMS) derivative. The NCI of these derivatives had the most abundant ion at m/z 567 (Fig. 3b), which corresponded to a trihydroxy-TMS-derivatized carboxylate anion. Negative ion chemical ionization was used to determine the retention times (9.98, 10.10, 10.17, and 10.42 min) for the -hydroxylated LTB 4 compounds of interest. After the retention times were established, EI was used to provide detailed structural information regarding the hydroxyl group position from bond cleavage that occurs adjacent to the trimethylsilyl ether positions. The EI mass spectrum of the derivatized metabolite with a retention time of 9.98 min (Fig. 3c) indicated that there are hydroxyl groups at C5 from the m/z 369 fragment (TMS-O ϩ ϭCH-(CH 2 ) 3 CO 2 PFB) and C12 from the m/z 549 (TMS-O ϩ ϭCH-(CH) 6 -CH(TMSO)-(CH 2 ) 3 CO 2 PFB) (24). Additionally the ions at m/z 73 (TMS ϩ ) and m/z 181 (C 7 H 2 F 5 ϩ ) were from the derivative groups. Finally, the major fragment ion at m/z 145 (TMS-O ϩ ϭCH(CH 2 ) 2 CH 3 ) was indicative of a 17-hydroxylated metabolite due to ␣-cleavage of the C16 -C17 bond adjacent to the trimethylsilyl ether (25). From the EI spectrum shown in Fig.  3c, it was determined that the LTB 4 metabolite with a retention time of 9.98 min was 17-OH-LTB 4 . The EI spectra of the hydroxylated LTB 4 metabolites with retention times of 10.10 and 10.18 min (Fig. 3b) were obtained and showed major ␣-cleavage ions at m/z 131 (TMS-O ϩ ϭ CHCH 2 CH 3 ) and m/z 117 (TMS-O ϩ ϭ CHCH 3 ), respectively, as expected for -1and -2-hydroxylated metabolites (25). In addition the relative abundance of other ions in the electron ionization mass spectra (data not shown) was identical to spectra published previously (30,31) of the -1and -2-hydroxylated metabolites of LTB 4 . These data were consistent with the identification of the metabolites with retention times of 10.10 and 10.18 min as 18-OH-LTB 4 and 19-OH-LTB 4 . Finally, the EI spectrum of the derivatized LTB 4 metabolite with a retention time of 10.42 min, identical to the gas chromatographic retention time of the 20-OH-LTB 4 standard, was identical to the EI mass spectrum of 20-OH-LTB 4 (24). The GC/MS data indicated that glucuronide conjugates of 20-, 19-, 18-, and 17-OH-LTB 4 were present in urine of subjects injected with LTB 4 and that the most abundant -oxidized metabolite was 17-OH-LTB 4 .
Glucuronide Conjugates of LTB 4 and 10,11-Dihydro-LTB 4 -The normal phase and reversed phase retention times of LTB 4 were determined to be 7.6 and 36.1 min, respectively. Therefore, pooled normal phase fractions 7-9 were subjected to RP-HPLC. Only the ␤-glucuronidase-treated urine revealed a conjugated triene chromophore with a max of 270 nm at 35.8 min (Fig. 4a, inset). The pooled reversed phase fractions 35-37 of the ␤-glucuronidase-treated and untreated samples were then subjected to LC/MS/MS analysis. In order to ascertain if LTB 4 was present in the urine, the MRM transitions m/z 335 3 195 and m/z 335 3 129 were monitored, and neither of these transitions were present in the unhydrolyzed urine (Fig. 5, b  and c). However, the MRM transitions for LTB 4 were detected at 36.9 min when the ␤-glucuronidase-treated urine was analyzed (Fig. 4, a and b). In addition to monitoring the MRM transitions for LTB 4 , the MRM transitions m/z 337 3 115 and m/z 337 3 225 for 10,11-dihydro-LTB 4 were also monitored in the unhydrolyzed and ␤-glucuronidase-treated urine. The most abundant of the MRM transitions for 10,11-dihydro-LTB 4 was not present in the unhydrolyzed urine (Fig. 5a); however, both of these MRM transitions were present at the correct relative intensity ratio at the slightly more lipophilic retention time compared with LTB 4 of 37.9 min (Fig. 6, a and b). This mass spectrometric data indicated that glucuronide conjugates of LTB 4 as well as 10,11-dihydro-LTB 4 were present in urine of the subjects injected with LTB 4 .
Glucuronide Conjugate of 10-Hydroxy-4,6,12-octadecatrienoic Acid (10-HOTrE)-Normal phase fractions eluting between 4 and 6 min were also subjected to RP-HPLC on a 4.6-mm column, and reversed phase fractions eluting between 39 and 41 min were subject to LC/MS/MS. The specific MRM transitions for 10-HOTrE that were monitored during this LC/ MS/MS assay were m/z 293 3 137 and 293 3 153 (29). Neither of these transitions were present in the specific fractions of the unhydrolyzed urine, but the transitions were quite abundant at 39.4 min in the ␤-glucuronidase-treated urine (Fig. 7, a and b). These data were consistent with a glucuronide conjugate of 10-HOTrE present in urine of subjects injected with LTB 4 .

Quantitation of LTB 4 Glucuronide and 20-COOH-LTB 4 -
The quantity of LTB 4 glucuronic acid conjugates excreted into urine of subjects injected with LTB 4 was determined using a stable isotope dilution mass spectrometric assay (Table II). Enzymatic hydrolysis of the LTB 4 glucuronide conjugates was achieved by incubating the solid phase extracted urine with ␤-glucuronidase for 16 h at 37°C. After hydrolysis and purification using RP-HPLC, the MRM transitions for LTB 4 and the internal standard d 4 -LTB 4 were monitored during LC/MS/MS analysis. A time course for the hydrolysis of urine was also carried out (Fig. 8) on an aliquot of the solid phase extracted urine to follow the yield of released LTB 4 . The yield of free LTB 4 was found to maximize between 15 and 24 h of treatment, and therefore a hydrolysis time of 16 h was routinely used in subsequent studies. The quantity of 20-COOH-LTB 4 was also determined in each urine sample using the d 4 -LTB 4 as an analog internal standard. The standard curve for this latter assay was linear over the range of 9 -900 pmol with an excellent linearity and correlation coefficient (data not shown). The  quantity of 20-COOH-LTB 4 present in each urine pool (0 -6 h) was similar to that of LTB 4 glucuronide (Table II). DISCUSSION Previous studies of the metabolism of LTB 4 in vitro revealed the existence of three major metabolic pathways. A major pathway of LTB 4 metabolism involved a specific cytochrome P-450 family of enzymes termed CYP4F (8) responsible for -oxidation (Fig. 9) and formation of 20-OH-LTB 4 . Further oxidation of the terminal carbon atom, mediated by alcohol dehydrogenase and aldehyde dehydrogenase in most cells, results in 20-COOH-LTB 4 which was one of the human urinary metabolites observed. Once the -carboxyl moiety was present, conversion of the 20-carboxyl group to the CoA ester is possible, which is required for peroxisomal or mitochondrial ␤-oxidation and the formation of chained shortened metabolites, specifically 18-COOH-LTB 4 and 16-COOH-LTB 3 (14).
A separate pathway of LTB 4 metabolism found in specific cells, including the lung macrophage, monocytes, and keratinocytes, involved reduction of the conjugated triene to a conjugated diene and has been termed the 12-hydroxyeicosanoid dehydrogenase/⌬ 10 -reductase pathway (16). This pathway is initiated through oxidation of the 12-hydroxy group at carbon 12 by 12-hydroxyeicosanoid dehydrogenase and the formation of 12-oxo-LTB 4 . This conjugated ketone is then reduced by ⌬ 10 -reductase and further reduction of the ketone moiety at carbon-12 by a keto-reductase to the hydroxyl moiety. This pathway results in a series of metabolites termed 10,11-dihydro-LTB 4 metabolites, which can be further transformed by /␤-oxidation from either the methyl or carboxyl terminus (16).
The third major pathway of LTB 4 metabolism has been characterized by conjugation with polar intermediates by specific enzymes that convert hydroxyl or carboxyl substituents of LTB 4 into ether or ester conjugates. Conjugation with glucuronic acid was particularly prominent in the metabolism of LTB 4 with human hepatocytes (13). It is important to note that not a single glucuronide of LTB 4 existed, but rather a mixture of numerous glucuronide conjugate positional isomers, perhaps greater than five, was possible when considering acyl glucuronides (32). Additionally, glucuronide conjugates of LTB 4 metabolites derived from the pathways described above were also possible that could result in an exceedingly complex mixture of metabolites, making identification of prominent metabolites of LTB 4 excreted into the urine difficult.
Previous studies of LTB 4 metabolism in intact animals had involved the use of tritium-labeled LTB 4 to assist in isolation and purification of metabolites. When rats were injected with radiolabeled LTB 4 , both bile and urine were found to contain metabolites. The bile contained 20 -25% of the injected radioactivity, whereas the urine contained less than 10% of the injected dose (33). Reversed phase HPLC analysis of the bile and urine revealed that most metabolites were less lipophilic than LTB 4 ; however, structural identification of these metabolites was not reported. In another study, the in vivo metabolism of LTB 4 in the monkey was reported (34) after infusion with radiolabeled LTB 4 (8 mol/kg). In this investigation, 25% of the injected radioactivity was recovered in urine after 24 h. More than 70% of the infused radioactivity was reported as volatile metabolites, and no intact LTB 4 was detected in urine. Purification of nonvolatile radiolabeled metabolites was carried out, and one metabolite was found to coelute with 20hydroxy-LTB 4 . The structures of the additional urinary metabolites were not reported, although several metabolites were separated.
In the present study where LTB 4 (150 nmol/kg) was injected into human subjects in a Phase I Clinical Trial, an alternative strategy was employed to identify LTB 4 metabolites based upon the previous known metabolites of LTB 4 using sensitive and specific LC/MS/MS assays. Because the retention times and relative intensities of ion transitions were known for several LTB 4 metabolites, it was possible to take advantage of the chromatographic behavior of metabolites under both reversed and normal phase conditions. For example, it was possible to obtain data consistent with the existence of 20-COOH-LTB 4 , a component present in the urine of individuals treated with LTB 4 , but not present in the urine of the untreated control subject. The criteria for metabolite identification was based upon HPLC retention times as well as the relative abundance of several specific ion transitions. The relative intensity of m/z 365 3 169 compared with that of m/z 365 3 195 was 0.70 ( Fig.  2, a and b), which was similar to that published previously (29). If a single ion transition was used, multiple components could often be detected (Figs. 1d and 7a). This was mostly the result of the presence of unrelated compounds in urine that generated signals with the same ion transition channel, most likely due to the large number of molecules present in urine at levels much higher than those observed for the true LTB 4 metabolites. Therefore, criteria were established where two MRM transitions for each of the identified metabolites had to coincide at the correct retention time and, if possible, UV absorption spectra in order to report the presence of a specific metabolite. By using this strategy, it was also possible to detect the presence of 18-COOH-LTB 4 , 10,11-dihydro-18-COOH-LTB 4 , and 10,11dihydro-20-COOH-LTB 4 excreted into the urine in each of the subjects, even though authentic standards were not available to unambiguously establish the ratio of MRM transitions. In the case of 18-COOH-LTB 4 and 10,11-dihydro-18-COOH-LTB 4 , additional information was employed in that the compound with the correct MRM transitions eluted as a less lipophilic component compared with 20-COOH-LTB 4 . Additionally, 10,11-dihydro-20-COOH-LTB 4 had the correct MRM transitions and eluted at a slightly more lipophilic retention time than 20-COOH-LTB 4 . No data could be obtained consistent with the presence of 16-COOH-LTB 3 as urinary metabolite of LTB 4 . It was very clear, however, that neither 20-OH-LTB 4 nor LTB 4 itself were present as excreted products in urine of subjects treated with LTB 4 . The fact that LTB 4 was not present in urine of these subjects suggested that LTB 4 was rapidly metabolized following injection and was consistent with previous primate studies where ϳ50 times more LTB 4 was injected (34).
One of the major pathways of LTB 4 metabolism in human subjects in this study was glucuronic acid conjugation, which is a common means that the intact organism can enhance the excretion of lipophilic compounds (35,36). In order to establish unambiguously that glucuronide conjugates of LTB 4 were present in the urine of subjects injected with LTB 4 , a strategy that employed analysis of identical aliquots of ␤-glucuronidasetreated and untreated samples were analyzed by LC/MS/MS. A further complication of multiple glucuronides for any single LTB 4 metabolite reduced the overall sensitivity of the analysis of each individual metabolite as an intact glucuronide. Because ␤-glucuronidase can cleave ether-linked as well as 1-O-acyllinked glucuronide conjugates, it was employed as the most applicable method to remove the glucuronic acid conjugate. Studies employing acid or based catalyzed hydrolysis led to extensive degradation of LTB 4 metabolites (data not shown). The LC/MS/MS data clearly indicated the existence of several excreted glucuronide metabolites of the injected LTB 4 . These included 20-, 19-, 18-, and 17-OH-LTB 4 , 10-HOTrE, 10,11dihydro-LTB 4 , and even intact LTB 4 glucuronides. None of these components were present in the nonhydrolyzed urine samples nor were they present in the untreated placebo urine. The identity of each of these metabolites (after hydrolysis) could be clearly ascertained using the criteria described above.
Of some interest was the unexpected observation of 17-OH-LTB 4 , 18-OH-LTB 4 , and 19-OH-LTB 4 glucuronide. Powell and Gravelle (30) found that rat polymorphonuclear leukocytes could form both 19-OH-LTB 4 and 18-OH-LTB 4 upon incubation with arachidonic acid and stimulation by A23187. Additionally, rat liver microsomes were found to convert LTB 4 into 20-OH-LTB 4 along with small amounts of 19-OH-LTB 4 (31). Recently, recombinant CYP4F5 and CYP4F6 were found to convert LTB 4 into 17-OH-, 18-OH-, 19-OH-, and 20-OH-LTB 4 during in vitro incubation (37). These P450 isozymes were found expressed in hepatocytes of rats (37). During the LC/MS/MS assay, detection of four separate components with the ion transition of m/z 351 3 195 was observed after treatment with ␤-glucuronidase (Fig.  3a). In order to determine the position of hydroxylation of the four components, reversed phase fractions 21-23 were pooled and analyzed by GC/MS as the PFB ester and TMS ether derivative. The four trihydroxy-TMS derivatized carboxylate anions (m/z 567) are shown in the NCI-GC/MS chromatogram in Fig. 3b at retention times of 9.98, 10.10, 10.18, and 10.42 min. Electron ionization was then used to provide structural information concerning the hydroxyl group position from cleavage that occurs adjacent to the TMS ether position. The EI mass spectra at retention times of 9.98, 10.10, and 10.18 min showed a major ␣-cleavage ion at m/z 145 (Fig. 3c), m/z 131, and m/z 117, respectively. From these data it was concluded that 17-OH-, 18-OH-, and 19-OH-LTB 4 were present as glucuronide conjugates in the urine of subjects injected with LTB 4 . Additionally, the metabolite with a retention time of 10.42 min was identified as 20-OH-LTB 4 . The main -hydroxylated LTB 4 metabolite found in the ␤-glucuronidase-treated urine of subjects injected with LTB 4 was found to be 17-OH-LTB 4 , whereas 20-OH-LTB 4 was a minor peak. This unexpected finding might reflect a situation where 20-OH-LTB 4 can undergo ␤-oxidation to form less lipophilic metabolites rather than forming a glucuronide (Fig. 9). Thus, the glucuronidation pathway was a more predominant fate for -1-, -2-, and -3-hydroxylation products of LTB 4 because subsequent ␤-oxidation was not possible.
The most abundant metabolites identified were LTB 4 glucuronide and 20-COOH-LTB 4 . Therefore, quantitative assays based on stable isotope dilution or analog dilution using stable isotope-labeled LTB 4 was employed to measure specifically these two metabolites of LTB 4 . The quantity of the LTB 4 glucuronide (after hydrolysis) was found to range between 15 and 30 pmol/ml of urine. The enzymatic hydrolysis experiment was performed five separate times for one subject, and the error was determined to be Ϯ1.7 pmol/ml of urine. The amount of 20-COOH-LTB 4 was also determined in these urine samples, and it ranged from 18 to 31 pmol/ml of urine. This quantitation was found to have an analytical error of Ϯ0.86 pmol/ml in five replicate analyses. From these results, it appeared that approximately the same amount of 20-COOH-LTB 4 and LTB 4 glucuronide were excreted into the urine of each subject. It is not known if a single glucuronide isomer predominated for LTB 4 glucuronide; however, based on previous studies it was likely that a number of glucuronides existed. The total amount of 20-COOH-LTB 4 and LTB 4 glucuronide represented only 0.2% recovery of the injected LTB 4 for each of the subjects. This rather low recovery was expected based on the previously described studies of LTB 4 metabolism as well as the metabolism of prostanoids into individual metabolites (18 -20).
It is important to note that the metabolic products of LTB 4 in vivo reported above were only found in the urine sample from a 0-to 6-h period after injection of LTB 4 . The 7-24-h urine sample was also analyzed, and none of these metabolites were observed. This was consistent with the time course described previously (33,34) for appearance of LTB 4 metabolites in urine in various animal experiments. These studies indicated that LTB 4 underwent degradation via multiple pathways leading to extensive chain shortening typical of fatty acids as well as glucuronidation (Fig. 9). Nevertheless, some caution must be exercised in extrapolation of the observations reported here to the metabolic fate of LTB 4 synthesized from endogenous arachidonic acid in vivo. This study involved administration of high doses of LTB 4 that could experience a different metabolic fate in quantitative terms from LTB 4 synthesized within a cell.
In summary, the metabolism of LTB 4 in the human subjects is quite complex with multiple pathways responsible for the degradation of LTB 4 prior to elimination into urine. Some of the most abundant products were those derived from conjugation with glucuronic acid of either intact LTB 4 or other highly lipophilic metabolites. In this investigation, 11 different metabolites of LTB 4 were identified in urine subjects injected with 150 nmol/kg LTB 4 . Four unconjugated metabolites were excreted into urine and identified by mass spectrometry as 18-COOH-LTB 4 , 10,11-dihydro-18-COOH-LTB 4 , 20-COOH-LTB 4 , and 10 -11-dihydro-20-COOH-LTB 4 . The glucuronide conjugates of 20-, 19-, 18-, and 17-OH-LTB 4 , 10-HOTrE, 10,11-dihydro-LTB 4 , and intact LTB 4 itself were also observed. None of these metabolites exceeded 1% of the injected dose; however, it was possible to detect the glucuronide conjugates by mass spectrometry after hydrolysis to the aglycone metabolite. It is possible that these LTB 4 metabolites may be relevant targets to assess the in vivo production of LTB 4 , but additional experiments will be required to define which of these observed me-tabolites from the exogenously administered LTB 4 best reflects endogenous LTB 4 production.