Identification of Novel Urinary Metabolites of the Lipid Peroxidation

HNA, its mercapturic acid conjugate, and two diastereoisomers of the ... of 9-carboxy-HNA and of the corresponding lactone mercapturic acid conjugate...
3 downloads 0 Views 240KB Size
1368

Chem. Res. Toxicol. 1998, 11, 1368-1376

Identification of Novel Urinary Metabolites of the Lipid Peroxidation Product 4-Hydroxy-2-nonenal in Rats Jacques Alary,* Laurent Debrauwer, Yvette Fernandez, Alain Paris, Jean-Pierre Cravedi, Laurence Dolo, Dinesh Rao, and Georges Bories Laboratoire des Xe´ nobiotiques, INRA, BP3, 31931 Toulouse Cedex, France Received March 31, 1998

Following iv administration of 4-hydroxy-2-nonenal (HNE) and [4-3H]HNE to rats, 15 polar urinary metabolites accounting for about 50% of the urinary radioactivity were separated by HPLC. Among them, eight major compounds and tritiated water were quantified. The metabolites were unequivocally characterized using GC/MS and ESI/MS/MS/MS. Most of “HNE polar metabolites” originate from ω-oxidation of 4-hydroxy-2-nonenoic acid (HNA): 9-hydroxyHNA, its mercapturic acid conjugate, and two diastereoisomers of the corresponding lactone. The oxidation of 9-hydroxy-HNA by alcohol and aldehyde dehydrogenases leads to the excretion of 9-carboxy-HNA and of the corresponding lactone mercapturic acid conjugate. 1,4-Dihydroxy2-nonene (DHN) originating from the reduction of HNE by alcohol dehydrogenase was to a lesser extent ω-hydroxylated, leading to 9-hydroxy-DHN which was excreted as a mercapturic acid conjugate (two diastereoisomers).

Introduction Lipid peroxidation of ω-6 polyunsaturated fatty acids results in the formation of reactive aldehydes (1), of which 4-hydroxy-2-nonenal (HNE)1 is the major and the most cytotoxic product (2). Despite its biological relevance, the in vivo metabolism of HNE still remains partially unknown. In vitro, metabolic studies carried out with rat liver subcellular fractions showed that HNE was reduced to 1,4-dihydroxy-2-nonene (DHN) by cytosolic alcohol dehydrogenase (3) and oxidized by aldehyde dehydrogenase to 4-hydroxy-2-nonenoic acid (HNA) (4). In addition to these phase I products, the HNE-GSH conjugate was also formed after short time incubations of HNE with hepatocytes, enterocytes, and tumor cells (5). In a study carried out after iv injection of HNE into the rats, four mercapturic conjugates have been characterized in the 0-2 h urine (6), namely, 1,4-dihydroxynonene mercapturic acid (DHN-MA), 4-hydroxynonenal mercapturic acid (HNE-MA), 4-hydroxynonenoic mercapturic acid (HNAMA), and the corresponding lactone. The presence of these conjugates in rat urine was recently confirmed by De Zwart et al. (7). In our first study (6), the formally characterized metabolites accounted for only about one-half of the * To whom correspondence should be addressed. Telephone: 33 (0)5 61 28 53 83. Fax: 33 (0)5 61 28 52 44. 1 Abbreviations: HNE, 4-hydroxy-2-nonenal; HNA, 4-hydroxy-2nonenoic acid; 9-hydroxy-HNA, 4,9-dihydroxy-2-nonenoic acid; 8-hydroxy-HNA, 4,8-dihydroxy-2-nonenoic acid; 9-carboxy-HNA, 4-hydroxy9-carboxy-2-nonenoic acid; DHN, 1,4-dihydroxy-2-nonene; 9-hydroxyDHN, 1,4,9-trihydroxy-2-nonene; HNE-MA, 4-hydroxynonenal mercapturic acid; DHN-MA, 1,4-dihydroxynonene mercapturic acid; HNA-MA, 4-hydroxynonenoic mercapturic acid; 9-hydroxy-HNA-MA, 4,9-dihydroxynonenoic mercapturic acid; 3-hydroxy-HNA-MA, 3,4-dihydroxynonenoic mercapturic acid; 3-hydroxy-HNE-MA, 3,4-dihydroxynonenal mercapturic acid; 3-hydroxy-DHN-MA, 1,3,4-trihydroxynonene mercapturic acid; 9-hydroxy-HNA-lactone-MA, 9-hydroxynonenoic lactone mercapturic acid; NAC, N-acetylcysteine.

urinary radioactivity. Unidentified compounds corresponded to more polar metabolites remaining mostly unresolved under our chromatographic conditions. This study was designed to characterize and quantify the major “polar metabolites” of HNE to complete an overview of the pathways involved in the metabolic fate of HNE in rats.

Materials and Methods Chemicals. HNE and [4-3H]HNE were obtained by acid hydrolysis of the corresponding synthesized diethylacetal derivatives (8, 9) and purified by HPLC. DHN and HNA were prepared from HNE as previously described (6). All solvents and reagents used for the preparation of buffers and HPLC eluents were of the highest commercial grade available from Merck (Nogent-sur-Marne, France) or Carlo Erba (Rueil Malmaison, France). Ultrapure water from Milli-Q system (Millipore, Saint Quentin en Yvelines, France) was used for HPLC eluent preparation. Radioactivity Determination. The samples were directly counted in a model 4330 Packard Tricarb scintillation counter (Packard Instrument Co., Downers Grove, IL) with Ultimagold (Packard) as the scintillation cocktail. Animal Treatment. Two male Wistar rats (250 g) were weakly anesthetized with diethyl ether and injected in the penis vein with 500 µL of Ringer’s solution containing a mixture of 1.2 mg of HNE and 0.6 MBq (550 ng) of [4-3H]HNE. The rats were housed in individual metabolism cages, and water was provided ad libitum. Twenty-four hour urine samples were collected. The urine samples were filtered through a Millex-HA 0.45 µm filter (Millipore) and stored at -20 °C until analysis. Extraction of Urinary HNE Metabolites. Filtered urine samples were diluted 1:1 with distilled water and adjusted to pH 2.5 with 1 M H3PO4. Each urine sample was loaded onto a 1 g LC-C18 Supelco cartridge (Saint Quentin Fallavier, France) preconditioned with 5 mL of methanol and 10 mL of dilute H3PO4 (pH 2.5). The cartridge was washed with 20 mL of dilute H3PO4 (pH 2.5). The aqueous phases from the column and washing were recovered and combined. The cartridge was dried

10.1021/tx980068g CCC: $15.00 © 1998 American Chemical Society Published on Web 10/24/1998

Polar Urinary Metabolites of 4-Hydroxy-2-nonenal under a nitrogen stream and eluted with 10 mL of methanol. The respective aqueous phases and methanolic eluates from each urine sample were combined and the samples stored at -20 °C until analysis. HPLC System 1. The system consisted of a Philips 4100 apparatus (Pye Unicam, Cambridge, U.K.) equipped with a 2 mL loop and a semipreparative Ultrabase ODS column (5 µm, 250 mm × 7.5 mm, SFCC Eragny, France) thermostated at 35 °C and connected to a Radiomatic Flo-one beta model 515 instrument (Packard, La-Queue-lez-Yvelines, France) or to a Gilson model 202 fraction collector (Gilson France, Villiers-leBel, France). Two mobile phases were used. Mobile phase A contained 2.5% acetonitrile and 97.5% (v/v) ammonium acetate (20 mM) adjusted to pH 4.5 with acetic acid. Mobile phase B contained 70% acetonitrile and 30% ammonium acetate (20 mM, pH 4.5). Solvents were delivered at a flow rate of 2 mL/min as follows: 100% A from 0 to 4 min, a linear gradient from 4 to 5 min from 0 to 14.3% B, a linear gradient from 5 to 30 min from 14.3 to 21% B, 21% B from 30 to 40 min, a linear gradient from 40 to 45 min from 21 to 100% B, and 100% B from 45 to 60 min. HPLC Systems 2-6. The system consisted of the apparatus described above equipped with a Supelcosil LC-18-DB 5 µm, 250 mm × 4.6 mm column (Supelco France, Saint-Germain-en-Laye, France) thermostated to 35 °C and connected to a Gilson fraction collector. Two mobile phases were used. Mobile phase A consisted of 2.5% acetonitrile and 97.5% acetic acid (1% v/v). Mobile phase B contained 20% acetonitrile and 80% acetic acid (1% v/v). Solvents were delivered at 1 mL/min as described below. System 2: 100% A from 0 to 5 min, a linear gradient from 5 to 30 min from 0 to 14.3% B, a linear gradient from 30 to 40 min from 14.3 to 20% B, a linear gradient from 40 to 46 min from 20 to 50% B, and 50% B from 46 to 60 min. System 3: 100% A from 0 to 5 min, a linear gradient from 5 to 30 min from 0 to 14.3% B, a linear gradient from 30 to 40 min from 14.3 to 20% B, 20% B from 40 to 45 min, a linear gradient from 45 to 46 min from 20 to 50% B, and a linear gradient from 46 to 60 min from 50 to 100% B. System 4: 100% A from 0 to 5 min, a linear gradient from 5 to 10 min from 0 to 14.3% B, a linear gradient from 10 to 25 min from 14.3 to 20% B, 20% B from 25 to 40 min, a linear gradient from 40 to 45 min from 20 to 50% B, and a linear gradient from 45 to 60 min from 50 to 100% B. System 5: 80% A and 20% B from 0 to 5 min, a linear gradient from 5 to 10 min from 20 to 40% B, 40% B from 10 to 15 min, a linear gradient from 15 to 45 min from 40 to 60% B, 60% B from 45 to 50 min, and a linear gradient from 50 to 60 min from 60 to 100% B. System 6: 80% A and 20% B from 0 to 4 min, a linear gradient from 4 to 5 min from 20 to 40% B, a linear gradient from 5 to 45 min from 40 to 60% B, 60% B from 45 to 50 min, and a linear gradient from 50 to 60 min from 60 to 100% B. Derivatization for GC/MS Analysis. The dry compounds were methylated by addition of 50 µL of etheral diazomethane. After 30 min at room temperature, the solvent was removed under a nitrogen stream. The dry residue was then dissolved in a mixture of N,O-bis (TMS) trifluoroacetamide/trimethylchlorosilane (Pierce Chemical, Rockford, IL) in a 99:1 ratio and heated at 60 °C for 1 h. The solvent was evaporated at 30 °C under a stream of nitrogen, and hexane (20 µL) was added to the dry residue. One microliter of the solution was injected into the gas chromatograph. Mass Spectrometry. (1) FAB/MS. FAB mass spectra were obtained on a Nermag R-10-10H (Delsi Nermag Instruments, Argenteuil, France) single quadrupole mass spectrometer fitted with a M-Scan FAB gun (M-Scan Ltd., Ascot, U.K.). Xenon gas was used for bombardment at an accelerating voltage of 8 kV, with a discharge current of 1-2 mA. Samples were prepared by mixing 1 µL of the sample solution (1 µg/µL in methanol) with 1 µL of the matrix [Magic Bullet (5:1 dithiothreitol/ dithioerythritol) or thioglycerol].

Chem. Res. Toxicol., Vol. 11, No. 11, 1998 1369 (2) GC/MS. For GC/MS analyses, a Nermag R-10-10-T single quadrupole instrument was coupled to a Delsi DI 200 (Delsi Nermag Instruments) gas chromatograph fitted with a BPX5 (25 m × 0.22 mm × 0.25 µm) capillary column (SGE, Villeneuve St. Georges, France). The samples were injected in the splitless mode. Helium was used as the carrier gas at a flow rate of 1 mL/min with a back pressure of 0.8 bar. The oven temperature was programmed as follows: 50 °C for 50 s, then from 50 to 230 °C at 25 °C/min, and from 230 to 270 °C at 5 °C/min. The injector temperature was 270 °C, and the interface temperature was 270 °C. EI mass spectra were generated at 70 eV with an emission current of 200 µA, at a source temperature of 220 °C. (3) ESI/MS. ESI/MS experiments were carried out either on a Nermag R-10-10-H single quadrupole instrument fitted with an Analytica of Branford (Branford, CT) ESI source or on a Finnigan LCQ quadrupole ion trap mass spectrometer equipped with the Finnigan ESI source. In both cases, the sample solutions (typically 10 ng/µL in 50:50 methanol/water) were infused at 1-3 µL/min into the electrospray interface. The potentials applied to the ESI sources were as follows: needle, ground potential; surrounding electrode, -2500 V; end plate/ nozzle, -3500 V; metallized intet end of the glass transfer capillary, -4500 V, for the Analytica source, and needle, 5200 V; and heated transfer capillary, 5-20 V, for the Finnigan source. Structural information was obtained by carrying out MSn experiments achieved on the LCQ instrument under automatic gain control conditions, and using helium as the collision gas. Biosynthesis of 4,9-Dihydroxy-2-nonenoic Acid. An HPLC-purified sample of [4-3H]HNA (0.3 µmol, 1 MBq) was incubated with 3 mg (200 µL) of liver microsomal protein from a clofibrate-treated rat in the presence of 5 mM glucose 6-phosphate, glucose-6-phosphate dehydrogenase (2 IU), and 1 mM NADPH dissolved in 500 µL of 0.1 M phosphate buffer (pH 7.4). The mixture was stirred for 2 h at 37 °C. The reaction was stopped by addition of 3 mL of methanol, and proteins were removed by centrifugation (10 min at 10000g). The supernatant was recovered, and methanol was evaporated under vacuum. The resulting aqueous phase was diluted with distilled water, adjusted to pH 2.5 (1 M H3PO4), and applied to a 360 mg SepPak Plus cartridge (Waters). The cartridge was washed with dilute H3PO4 (pH 2.5), dried under a nitrogen stream, and eluted with 3 mL of methanol. HPLC radiochromatography (system 1) revealed the presence of two peaks accounting for about 35 and 65% of the radioactivity, respectively. The polar fraction (tR ) 15 min) was tentatively assigned to the expected mixture of 4,9-dihydroxy-2-nonenoic acid (9-hydroxy-HNA) and 4,8-dihydroxy-2-nonenoic acid (8hydroxy-HNA), whereas the major peak (tR ) 40 min) corresponded to unchanged HNA. After methylation and silylation, the polar fraction was analyzed by GC/MS using electron impact ionization. The resulting gas chromatogram exhibited two peaks. From the major one where tR ) 9.4 min (95%), the molecular ion of the bis-TMS methyl ester derivative of the expected 9-hydroxy-HNA (m/z 346) was not detected but several fragments provided evidence for this structure: m/z 331 [M - CH3]+, m/z 314 [M - CH3OH]+•, m/z 299 [M - CH3 - CH3OH]+, m/z 256 [M - (CH3)3Si - OH]+•, m/z 241 [M - CH3 - (CH3)3Si OH]+, m/z 187 [[(CH3)3Si - OdCH - CHdCH - COOCH3]+, R-cleavage of the 4-OTMS group], m/z 147 [[(CH3)2SidO Si(CH3)3]+, presence of two hydroxy groups], m/z 103 [[(CH3)3Si - OdCH2]+, presence of the 9-hydroxy group], m/z 75 [(CH3)2Si - OH]+, and m/z 73 [(CH3)3Si]+. All these fragments were in total agreement with the fragmentation pattern published by Eglinton et al. (10) for ω-hydroxy TMS ethers of fatty acid methyl esters and allowed characterization of the major compound as 9-hydroxy-HNA. The presence of trace amounts of 8-hydroxy-HNA could be expected due to the type of biological oxidation used. Indeed, the gas chromatogram showed the presence of a minor compound ( 334

3a

338 MH+ 320 [MH - H2O]+

338

320 302

[MH - H2O]+ [MH - 2H2O]+

338 > 320

3b

338 MH+ 320 [MH - H2O]+

338

320 302

[MH - H2O]+ [MH - 2H2O]+

338 > 320

4

348 MH+ 330 [MH - H2O]+

348

330 312 302 185 164

[MH - H2O]+ 348 > 330 [MH - 2H2O]+ [MH - H2O - CO] A+ BH+

242

5b

334 MH+ 316 [MH - H2O]+

6

334 MH+

334

316 298 164 153

-

-

[MH - H2O]+ [MH - 2H2O]+ BH+ [A - H2O]+

-

Scheme 1. General Fragmentation Pattern and Fragment Ion Notation of Mercapturic Acid Conjugates of HNE Polar Metabolites

of 16. The major water loss during MS collisional activation strongly suggests metabolite 2b should be an oxidized derivative of HNA-MA. Theoretically, biological oxidation could take place (i) on the sulfur atom of NAC

334 > 316

-

185 167 146 298 288 270 164 153 146 135 130 -

intrepretation [MH - 2H2O]+ [MH - 3H2O]+ [A - H2O]+ BH+ [A - 2H2O]+ [BH - H2S]+ [A -2H2O - CO2]+ [MH - 2H2O]+ [ASH - H2O]+ BH+ [A - H2O]+ [A - 2H2O]+ [BH - H2S]+ [MH - 2H2O]+ [MH - 3H2O]+ [ASH - H2O]+ [ASH - 2H2O]+ BH+ [A - 2H2O]+ [BH - H2S]+ [MH - 2H2O]+ [MH - H2O - CO]+ [MH - 2H2O - CO]+ [MH - 3H2O - CO]+ [MH - H2O - CO CH2CO]+ [MH - 2H2O - CO CH2CO]+ A+ [A - H2O]+ [BH - H2O]+ [MH - 2H2O]+ [MH - H2O - CO]+ [MH - 2H2O - CO]+ BH+ [A - H2O]+ [BH - H2O]+ [A - 2H2O]+ [BH - H2S]+ -

proposed structure 9-hydroxy-HNA-MA

9-hydroxy-DHN-MA

9-hydroxy-DHN-MA

9-carboxy-HNA-lactone-MA

9-hydroxy-HNA-lactone-MA

9-hydroxy-HNA-lactone-MA

to give a sulfoxide, (ii) on the C2 or C3 atoms following the previous epoxidation of the double bond, or (iii) at the ω or ω - 1 positions of the alkyl chain. These different possibilities have been investigated. The search for an eventual mercapturic acid sulfoxide was based upon the fact that the desulfurization of the sulfoxide and that of the thioether by Raney nickel lead to the same compound, namely, 4-hydroxynonanoic acid. Following such a treatment (6), metabolite 2b did not give rise to 4-hydroxynonanoic acid. Epoxidation of HNE was performed using H2O2 as previously described (11), giving a racemic mixture of HNE-epoxide which was subsequently reacted at pH 7.8 with a stoichiometric amount of NAC. The reaction mixture chromatographed by HPLC (system 1) showed the presence of two major compounds with retention times of 31 and 32 min, respectively. The two compounds analyzed by FAB/MS (negative mode) gave a unique and intense [M - H]- ion at m/z 334 consistent with the opening of the HNE-epoxide ring by the mercapturate. Periodic oxidation of the mixture of the two mercapturic conjugates isomers afforded [3H]hexanal (yield of 70%) which was characterized as its dinitrophenylhydrazone. This result provides conclusive evidence of the presence

Polar Urinary Metabolites of 4-Hydroxy-2-nonenal

of a vic-diol group at C3-C4 and implies that most of the mercapturic groups are attached at C2. 1,3,4-Trihydroxynonene mercapturic acid (3-hydroxyDHN-MA) was synthesized from 3-hydroxy-HNE-MA by reduction with sodium borohydride. Two isomers were obtained with the same m/z 336 [M - H]- ion (negative FAB/MS). HPLC analysis (system 1) of the mixture reveals two peaks with retention times of 27 and 28 min, respectively. 3-Hydroxy-HNA-MA was synthesized by oxidation of 3-hydroxy-HNE-MA with sodium chlorite in the presence of sulfamic acid (6). The reaction mixture was composed of two isomers exhibiting the same [M - H]- m/z 350 quasi-molecular ion (negative FAB/MS). Chromatographed by HPLC (system 1), the mixture of 3-hydroxyHNA-MA isomers gave two peaks with retention times of 25 and 26 min, respectively. The retention times of the three mercapturic acid conjugates synthesized from HNE-epoxide and that of metabolite 2b in HPLC (system 1) were quite different. This demonstrates that metabolite 2b is not hydroxylated at C3. The unambiguous characterization of metabolite 2b as the mercapturic acid conjugate of 9-hydroxy-HNA (9-hydroxy-HNA-MA) was fully confirmed by the comparison of its fragmentation pattern with that of the synthesized standard. Peak 3. Analyzed by HPLC (system 3), peak 3 is composed of a mixture of three metabolites (3a-c) with tR values of 41.4, 42.7, and 51.4 min, respectively. Metabolite 3a accounted for 2% of the urinary radioactivity and was analyzed by ion trap ESI/MS (positive mode). The quasi-molecular ion observed at m/z 338 and the fragmentation pattern of this metabolite (Table 1) are consistent with the assigned structure, 9-hydroxy-DHNMA. Metabolite 3b accounted for 1.9% of the urinary radioactivity. ESI/MS analysis gave a fragmentation pattern similar to that described for metabolite 3a, suggesting that metabolite 3b is a diastereoisomer of metabolite 3a. Metabolite 3c accounted for only 1% of the urinary radioactivity. This compound gave no signal when analyzed by ESI/MS and no interpretable spectrum after analysis by GC/MS. Metabolite 3c was not further investigated due to the low amount of available material. Peak 4. HPLC system 4 analysis of peak 4 exhibited a unique compound (tR ) 52.50 min) which accounted for 3.8% of the urinary radioactivity. Analysis of metabolite 4 by ESI/MS (positive mode) showed an intense MH+ ion at m/z 348 and relevant fragment ions obtained by MS/ MS performed on the MH+ parent ion (Table 1). This fragmentation pattern is consistent with the proposed structure, 9-carboxy-HNA-lactone-MA. In this case, owing to the presence of a carboxy group at C9, the formation of the -lactone cannot be totally ruled out. However, spontaneous lactonization is easier for γ- than for -hydroxy acids, and the γ-lactone shows a greater stability to hydrolysis. These two observations support the proposed structure. Peak 5. Analysis of peak 5 by HPLC (system 5) revealed the presence of two compounds (5a and 5b) with retention times of 21.6 and 27.3 min, respectively. Metabolite 5a accounted for 1.2% of the urinary radioactivity. This compound gave no signal in ESI/MS and was not further investigated. Metabolite 5b accounted for 3.5% of the urinary radioactivity and was analyzed

Chem. Res. Toxicol., Vol. 11, No. 11, 1998 1373

by positive ESI/MS (Table 1). By using sequential MS2 and MS3 experiments, metabolite 5b was identified as 9-hydroxy-HNA-lactone-MA. Peak 6. HPLC of peak 6 (system 5) showed a unique compound with a tR of 31 min (metabolite 6) which accounted for 2.6% of the urinary radioactivity. The ESI/ MS analysis of this metabolite showed an MH+ ion at m/z 334. Owing to the low quantities of metabolite 6 that were available, MS2 and MS3 experiments could not be successfully carried out. On the basis of its molecular weight (MW ) 333), metabolite 6 was thus tentatively assigned to a diastereoisomer of metabolite 5b. The proposed structures for metabolites 5b and 6 were unequivocally confirmed by both an identical chromatographic behavior and fragmentation pattern of the synthesized 9-hydroxy-HNA-lactone-MA isomers. Peak 7. Analysis of peak 7 by HPLC (system 6) showed the presence of two unresolved compounds with tR values of 25 and 25.5 min, respectively. Peak 7 accounted for 1.5% of the urinary radioactivity. Analysis by ESI/MS gave no interpretable signal. Therefore, due to the low amount of available material, further investigations have not been carried out. The fragmentation spectra of HNE polar metabolites 2b-6 showed a common fragment ion at m/z 164 characteristic of a protonated NAC moiety which unequivocally showed that these urinary metabolites corresponded to mercapturic conjugates. Also, this m/z value confirmed the absence of a sulfoxide group on the mercapturic acid moiety. ESI/MSn spectra of metabolites 2b-6 all gave fragment ions which showed a major water loss, suggesting that oxidation takes place on the alkyl chain of HNA. Following periodic acid treatment, none of the HNE polar metabolites gave [3H]hexanal contrary to synthesized HNE-epoxide conjugates which all afforded the labeled aldehyde. This result supports the assumption that the former compounds do not bear a vic-diol group at C3C4. Moreover, the chromatographic behavior of metabolites 2b-6 showed that these compounds were more polar than the synthesized HNE-epoxide mercapturates. This higher polarity is likely due to a lower capability of intramolecular hydrogen bonding in metabolites 2b-6 when compared to that of HNE-epoxide mercapturates. This suggests that the oxidation of the alkyl chain would occur beyond C5. This assumption was confirmed by the analysis of three standards bearing a hydroxy group at C9 which showed both a chromatographic behavior and a mass spectrum identical with those of metabolites 2b, 5b, and 6, respectively. Although the authentic standards of all the metabolites were not synthesized, the observed similarities in the fragmentation pattern of metabolites 2b-6 suggest that all resulted from the initial hydroxylation at C9. Metabolites Unretained by the C18 Cartridge at pH 2.5. This group of metabolites accounted for 18% of the urinary radioactivity. A fraction of these metabolites corresponding to 3.5% of the urinary radioactivity was evolved during the distillation step and was quantitatively recovered in the condensate. Experiments were carried out on the condensate to tentatively characterize this fraction. The radioactivity recovered in the condensate was totally unextracted by ethyl acetate in acid, neutral, and alkaline medium. It was completely unretained on SAX, SCX, C4, C8, and HLB (Waters) cartridges. Treatment of the

1374 Chem. Res. Toxicol., Vol. 11, No. 11, 1998

condensate with dinitrophenylhydrazine at pH 1 showed that the radioactivity does not correspond to aldehydic compound(s). All these data strongly suggest that the evolved fraction would correspond to tritiated water. Additional experiments carried out to investigate the time course excretion of tritiated water showed that [3H]H2O could not be detected in 0-2 h urine, but the amount increased thereafter, accounting for 3.5% of the radioactivity in the 2-24 h urine and for 28% of the radioactivity in the 24-48 h urine sample. Following the distillation step, highly polar products (14.5% of the urinary radioactivity) remained in the residue. These compounds were retained by a SAX cartridge (500 mg of Chromabond SB, Macherey Nagel Hoerdt, France) and were methylated with a good yield (80%). These results showed that this group of compounds could correspond to a mixture of four short chain hydroxy acids likely generated by β-oxidation of HNA and 9-carboxy-HNA. However, at this time, we are unable to obtain their respective methyl ester with a sufficient purity to allow their unambiguous characterization by GC/MS.

Discussion These results show that 24 h after HNE iv injection, 50% of the urinary radioactivity correspond to 15 HNE polar metabolites, eight of which and tritiated water have been unequivocally characterized. These HNE polar metabolites do not originate directly from HNE but mainly from HNA formed by oxidation of HNE by cytosolic and/or membrane-bound aldehyde dehydrogenases and to a lower extent from DHN formed by reduction of HNE by cytosolic alcohol dehydrogenase. The relevant role of HNA in the whole metabolism of HNE is not surprising and can be attributed to two factors. (i) HNA is both an early and a major metabolite of HNE in hepatocytes (12). (ii) HNA and DHN show a low reactivity with GSH for Michael addition. Therefore, free HNA and DHN may remain available for metabolic pathways other than GSH conjugation for a longer period of time. On the basis of the results of this study, the hypothetical metabolic pathway of HNE in rats is proposed in Scheme 2. The characterization of the major HNE polar metabolites indicates that for HNA, ω-hydroxylation is especially relevant. Indeed, the results show that 9-hydroxy-HNA accounts for the major HNE polar metabolite. This supports the assumption that ω-hydroxylation of HNA corresponds to an early metabolic step. In fact, the involvement of the microsomal P450 4A isozymes in the metabolization of HNE was recently demonstrated in mice in our laboratory.2 This finding was obtained from in vivo studies carried out in both conventional mice and peroxisome proliferator-activated receptor R-deficient mice, and confirmed in vitro with hepatic microsomes from clofibrate-induced and noninduced animals. It thus justified our methodological approach in the biochemical synthesis of the ω-hydroxylated standard. Assays carried out with liver microsomes from rats induced with clofibrate and incubated with HNA and DHN show that 35% of the HNA dose and 4% of the DHN dose are ω-hydroxylated, respectively. Under the same 2 F. Gue ´ raud, J. Alary, P. Costet, L. Debrauwer, L. Dolo, T. Pineau, and A. Paris, manuscript in preparation.

Alary et al.

conditions, no ω-hydroxylation was detected with HNAMA and DHN-MA as substrates (data not shown). These findings indicate that ω-hydroxylation of HNA occurs before GSH conjugation in the presence of liver glutathione transferases. The resulting glutathione conjugates are then exported from the liver to the kidney and excreted in the urine as mercapturic acid conjugates. Whether a similar sequence of biochemical reactions can be involved in the formation of 9-hydroxy-DHN-MA isomers remains unkown. The results show that 60% of 9-hydroxy-HNA-MA is characterized as a mixture of two diastereoisomers of the corresponding lactone. No significant amount of lactone was observed during the syntheses of HNA and of 9-hydroxy-HNA. On the other hand, high levels of HNA-lactone-MA were also excreted in rat urine (6), suggesting that the saturation of the C2C3 double bond of HNA is a prerequisite for spontaneous γ-lactonization. Therefore, lactonization of the hydroxy acid occurred after GSH conjugation. The occurrence of such a sequence of reactions is in agreement with the fact that compounds with a R,β-unsaturated lactone ring do not react with GSH (13). It is noteworthy that lactonization of γ-hydroxy acids occurs spontaneously. In our previous study (6), following direct urine injection into HPLC, the HNA-MA:HNAlactone-MA ratio was found to be 0.54. In this study, the ratio was 0.22. This decrease of about 2.5 times was due to the use of strong acidic conditions (pH 2.5) required for both the extraction and the HPLC purification of HNE polar metabolites which artificially favor lactone formation. HNA is a medium-chain fatty acid, a class of compounds which are usually excreted in urine as the corresponding dicarboxylic acids (14). The presence of 9-carboxy-HNA in the urine is in agreement with this statement. This compound originates from 9-hydroxyHNA by oxidation catalyzed by alcohol dehydrogenase and subsequent oxidation by aldehyde dehydrogenase located mainly in the cytosolic fraction of hepatocytes and in the endoplasmic reticulum (15). As observed for 9-hydroxy-HNA, the dicarboxylic derivative is then partially conjugated with GSH and lactonized. Tritiated water in 24 h urine accounted for about 3.5% of the urinary radioactivity. It originates from the HNE metabolic pathway and not from a label exchange since previous attempts carried out with HNE and DHN-MA (as a model of mercapturic acid conjugates) showed that no significant amount (