Formation of Highly Reactive Cyclopentenone Isoprostane Compounds (A₃/J₃-Isoprostanes) in Vivo from Eicosapentaenoic Acid

Abstract

Omega-3 (ω-3) polyunsaturated fatty acids (PUFAs) found in marine fish oils are known to suppress inflammation associated with a wide variety of diseases. Eicosapentaenoic acid (EPA) is one of the most abundant ω-3 fatty acids in fish oil, but the mechanism(s) by which EPA exerts its beneficial effects is unknown. Recent studies, however, have demonstrated that oxidized EPA, rather than native EPA, possesses anti-atherosclerotic, anti-inflammatory, and anti-proliferative effects. Very few studies to date have investigated which EPA oxidation products are responsible for this bioactivity. Our research group has previously reported that anti-inflammatory prostaglandin A₂-like and prostaglandin J₂-like compounds, termed A₂/J₂-isoprostanes (IsoPs), are produced in vivo by the free radical-catalyzed peroxidation of arachidonic acid and represent one of the major products resulting from the oxidation of this PUFA. Based on these observations, we questioned whether cyclopentenone-IsoP compounds are formed from the oxidation of EPA in vivo. Herein, we report the formation of cyclopentenone-IsoP molecules, termed A₃/J₃-IsoPs, formed in abundance in vitro and in vivo from EPA peroxidation. Chemical approaches coupled with gas chromatography/mass spectrometry (GC/MS) and liquid chromatography/mass spectrometry (LC/MS) were used to structurally characterize these compounds as A₃/J₃-IsoPs. We found that levels of these molecules increase ∼200-fold with oxidation of EPA in vitro from a basal level of 0.8 ± 0.4 ng/mg EPA to 196 ± 23 ng/mg EPA after 36 h. We also detected these compounds in significant amounts in fresh liver tissue from EPA-fed rats at basal levels of 19 ± 2 ng/g tissue. Amounts increased to 102 ± 15 ng/g tissue in vivo in settings of oxidative stress. These studies have, for the first time, definitively characterized novel, highly reactive A/J-ring IsoP compounds that form in abundance from the oxidation of EPA in vivo.

Eicosapentaenoic acid (C20:5, ω-3; EPA)2 is one of the most abundant ω-3 polyunsaturated fatty acids (PUFAs) present in fish oils. Both animal and human epidemiological studies, and more recently, human clinical intervention trials, suggest that fish consumption, or dietary fish oil supplementation, reduces the incidence of important diseases including atherosclerosis and sudden cardiac death, neurodegeneration, asthma, and other inflammatory disorders (1–3). The mechanism(s) by which these effects occur is unknown, but it has been hypothesized that both various enzymatically and non-enzymatically derived EPA oxidation products are responsible for some of the benefits associated with EPA (4–7). Whereas specific biologically active, enzymatically derived EPA oxidation products such as polyhydroxylated EPA derivatives, termed the resolvins, have been reported by Serhan et al., (7), no studies to date have determined which non-enzymatic oxidation product or products of EPA might contribute to its bioactivity. A recent report from our laboratory, however, determined that EPA peroxidation products containing α,β-unsaturated carbonyl ring structures induced the NF-E2-related factor 2 (Nrf2)-based antioxidant response through inhibition of a negative regulator of Nrf2, but this study did not definitively characterize the molecular structure of these compounds (8).

Isoprostanes (IsoPs) are prostaglandin (PG)-like molecules that are formed non-enzymatically from the free radical-induced oxidation of arachidonic acid (C20:4, ω-6). These compounds are generated via bicyclic endoperoxide PGH₂-like intermediates that are reduced to PGF_2α-like compounds termed F₂-IsoPs, or the endoperoxides can undergo rearrangement to PGE₂ and PGD₂-like compounds (E₂/D₂-IsoPs) (9, 10). Unlike PGs, however, IsoPs have been shown to form in situ in phospholipids and are subsequently cleaved from phospholipid storage sites by phospholipases (11, 12). It is well known that eicosanoids containing E- and D-type prostane rings are unstable and readily dehydrate in aqueous solutions to form cyclopentenone-containing compounds (13). Previously we reported that, analogous to the formation of A-ring and J-ring PGs from the dehydration of the cyclooxygenase-generated PGE₂ and PGD₂, A₂/J₂-IsoPs as well as A₄/J₄-neuroprostanes (NPs), are generated in vitro and in vivo from the oxidation of arachidonic acid and docosohexaenoic acid (C22:6, ω-3; DHA), respectively (13–15). More recently, our laboratory has confirmed that F-ring IsoP-like compounds, termed F₃-IsoPs, are also formed from the free radical-catalyzed peroxidation of EPA (16). Based on these studies as well as our interest in the potential bioactivity of cyclopentenone-containing EPA oxidation products, we have hypothesized that, analogous to the formation of A₂/J₂-IsoPs from the oxidation of arachidonic acid, A- and J-ring IsoPs (A₃/J₃-IsoPs) can be generated from the oxidation of EPA in vitro and in vivo.

A proposed mechanism of formation of A₃/J₃-IsoPs is shown in Fig. 1. There are four bis-allylic carbons in EPA at carbons 7, 10, 13, and 16 as opposed to only three bis-allylic carbons in arachidonic acid. Hydrogen abstraction can occur at each of these bis-allylic carbons and is followed by oxygen insertion. Depending on the site of oxygen insertion, one of eight different hydroperoxides are generated; these eight hydroperoxides are then further oxidized to yield six bicyclic endoperoxides that can then undergo ring arrangement to form 6-series of E₃/D₃-IsoPs. These products subsequently dehydrate to generate one of six different series of A₃/J₃-IsoPs. Each regioisomer is theoretically comprised of eight racemic diastereomers for a total of 48 compounds. A nomenclature system for the IsoPs has been established and approved by the Eicosanoid Nomenclature Committee in which each different regioisomer class is designated by the carbon number on which the side chain hydroxyl is located with the carboxyl carbon designated as C-1 (17). Thus, in accordance with this nomenclature system, the A₃/J₃-IsoP regioisomers are designated as 5-, 8-, 11-, 12-, 15-, and 18-series A₃/J₃-IsoPs. Herein, we present evidence that A₃/J₃-IsoPs are, in fact, formed in significant amounts in vitro and in vivo from the free radical-catalyzed peroxidation of EPA.

FIGURE 1.

Pathway for the formation of the A₃/J₃-IsoPs by the nonenzymatic peroxidation of EPA. For simplicity, the stereochemistry of both the lipid peroxidation intermediates and final IsoPs is not shown, although it has been extensively characterized previously utilizing other fatty acids (32, 55).

EXPERIMENTAL PROCEDURES

Materials—EPA was purchased from Nu-Chek Prep (Elysian, MN) and [²H₄]PGA₂ was purchased from Cayman Chemical Co. (Ann Arbor, MI). Dimethylformamide, undecane, pentafluorobenzyl bromide, diisopropylethylamine, methoxylamine hydrochloride, and glutathione (GSH) were from Sigma-Aldrich. Bis(trimethylsilyl)trifluoroacetamide was purchased from Supelco (Bellefonte, PA). Bis-[²H₉]-(trimethylsilyl)acetamide was procured from CDN Isotopes (Pointe-Claire, PQ). [²H₃]Methoxylamine hydrochloride was purchased from Cambridge Isotope Laboratory (Andover, MA), and 2,2′-azobis(2-amidinopropane) hydrochloride (AAPH) was from Polysciences, Inc. (Warrington, PA). C-18 and silica Sep-Pak cartridges were from Waters Associates (Milford, MA). 60ALK6D TLC plates were from Whatman (Maidstone, UK). Equine liver glutathione S-transferase (GST) was purchased from BioPur (Bubendorf, Switzerland).

Oxidation of EPA—5 mg of fresh EPA was dissolved in 100 μl of ethanol and added immediately to 4.9 ml of phosphate-buffered saline solution (pH 7.4) containing 10 mm AAPH. The EPA oxidation reaction mixture was incubated at 37 °C for varying amounts of time, after which it was placed immediately at –80 °C until further processing.

Purification and Analysis of A₃/J₃-IsoPs—Free and esterified A₃/J₃-IsoPs were extracted using C-18 Sep-Pak cartridges, converted to methoxylamine (MOX) derivatives in pyridine, extracted, converted to pentafluorobenzyl ester derivatives, purified by thin layer chromatography (TLC), converted to trimethylsilyl (TMS) ether derivatives, and quantified by stable isotope dilution gas chromatography (GC)/negative ion chemical ionization (NICI) mass spectrometry (MS) with [²H₄]PGA₂ as an internal standard using a modification of the method described for the quantification of A₂/J₂-IsoPs (14). The only change to the procedure was the TLC purification. Instead of scraping from 1 to 3 cm above where a PGA₂ methyl ester standard migrates on TLC for analysis of A₂-IsoPs, the scraped area was extended to 5 cm above where PGA₂ methyl ester migrates. The [M–CH₂C₆F₅]^– ions were monitored for quantification (m/z 432 for A₃/J₃-IsoPs and m/z 438 for [²H₄]PGA₂). Quantification of the total amount of A₃/J₃-IsoPs was determined by integrating the peak area of material in the m/z 432 channel in comparison with the m/z 438 channel. GC/NICI/MS was carried out using an Agilent Technologies 6890N Network GC/MS system.

Conjugation of A₃/J₃-IsoPs with GSH in Vitro—A₃/J₃-IsoPs obtained from the in vitro oxidation of EPA were purified by Sep-Pak extraction and incubated in 0.1 m K₃PO₄ buffer (pH 6.5) in the presence of a 10-fold excess (∼1 μg) of GSH and 1 mg of equine liver GST containing a mixture of GSTs at 37 °C for 2 h (14). The incubation mixture was then acidified to pH 3 and extracted with 2 volumes of MeCl₂. Unconjugated A₃/J₃-IsoPs were measured in the organic fraction by GC/MS, and conjugated IsoPs represented the difference between the amount of A₃/J₃-IsoPs added to the incubation versus that amount present in the organic fraction (15).

Subsequently, A₃/J₃-IsoP-GSH adducts from a separate incubation were definitively identified by LC/MS. Adducts were purified by extraction using a C18 Sep-Pak cartridge pre-conditioned with acetonitrile and 0.1 m ammonium acetate (pH 3.4) (14) and eluted with 10 ml of 95% ethanol (10 ml), dried under a stream of N₂, and analyzed by LC/MS. Samples were suspended in a 1:3 mixture of Phase 1/Phase 2 (Phase 1: acetonitrile/methanol/acetic acid, 95:5:0.1; Phase 2: water/Phase 1/acetic acid, 95:5:0.1). LC was carried out using the Surveyor MS Pump equipped with a Phenomenex Luna C-18 column (50 × 2.00 mm, 3 μm) using Phase 1 and Phase 2 as mobile phases (0.3 ml/min) and an isocratic gradient of Phase 1 for 1 min and then a gradient of 20–70% Phase 1 in 26 min. Samples were analyzed using a ThermoFinnigan TSQ/Quantum Ultra mass spectrometer operating in the negative ion mode. The electrospray (ESI) source was fitted with a deactivated fused silica capillary (100-mm inner diameter). Nitrogen was used as both the sheath gas and the auxiliary gas, and all parameters were tuned in order to have maximum response. For MS/MS experiments, putative A₃/J₃-IsoPs were collisionally activated between energies of 15 and 30 eV under 1.5 mTorr of argon. Data acquisition and analysis were performed using Xcalibur software, version 2.0.

Analysis of A₃/J₃-IsoPs by LC/ESI-MS/MS—After extraction using the C-18 Sep-Pak methodology described above, samples were suspended in 2:1 methanol/water, and A₃/J₃-IsoPs were analyzed by LC. LC was performed with a Surveyor MS Pump equipped with a Phenomenex Luna C-18 column (50 × 2.00 mm, 3 μm) at a flow rate of 0.2 ml/min starting with 85% Phase 1 (95:5:0.1, acetonitrile/methanol/acetic acid) to 50% from 1 to 17.50 min, holding for 30 s, and returning to 85% at 19.00 min. Phase 2 consisted of water/Phase 1/acetic acid (95:5:0.1). LC/MS was carried out on a ThermoFinnigan TSQ/Quantum Ultra mass spectrometer. The ESI source was fitted with a deactivated fused silica capillary (100-μm inner diameter), and the mass spectrometer was operated in the negative ion mode. Nitrogen was used as both the sheath gas and the auxiliary gas, at 31 and 17 p.s.i., respectively. The capillary temperature was 300 °C. The spray voltage was 5.0 kV, and the tube lens voltage was 80 V. Collision-induced dissociation (CID) of the molecular ion of putative A₃/J₃-IsoPs was performed from 12 to 28 eV under 1.5 mTorr of argon. Spectra that are shown were obtained at either 18 or 21 eV. Spectra were displayed by averaging scans across chromatographic peaks. Selective reaction monitoring (SRM) was performed according to characteristic fragmentation patterns of A₃/J₃-IsoPs determined by CID. The collision energy for SRM was 18 eV. Data acquisition and analysis were performed using Xcaliber software, version 2.0.

Preparation of A₃/J₃-IsoPs from Rodent Tissue—Rats were fed with a rodent AIN-93 diet containing either 0 or 4% Menhaden fish oil (by weight). In some studies, animals were administered CCl₄ (1 ml/kg) orogastrically to induce oxidative stress 4 h before sacrifice. After 4 weeks of feeding, the rats were sacrificed. Liver and other tissues were removed and immediately flash-frozen in liquid nitrogen and stored at –80 °C. Analyzed tissue samples were homogenized in 5 ml of ice-cold chloroform/methanol (2:1, v/v) containing butylated hydroxytoluene (0.005%) to prevent ex vivo autoxidation. Esterified A₃/J₃-IsoPs in phospholipids were enzymatically hydrolyzed using bee venom PLA₂ to liberate free A₃/J₃-IsoPs. As noted previously (18), the addition of various PUFAs including arachidonic acid, EPA, or DHA to tissues during workup does not increase the levels of IsoP-like compounds in tissue extracts. A₃/J₃-IsoPs in samples were then purified and analyzed as described above.

Analysis of A₃/J₃-IsoP-PCs in Vivo by UPLC/ESI-Ion Trap-MS—Liver tissue samples were obtained from previously mentioned feeding studies, and tissue samples were homogenized in 5 ml of ice-cold chloroform/methanol (2:1, v/v) containing butylated hydroxytoluene (0.005%) to prevent ex vivo auto-oxidation. Phospholipids were separated from other lipids as follows. An inert II silica Sep-Pak (500 mg) was prewashed with 5 ml of hexanes followed by 5 ml of chloroform. Each sample was applied to a column and initially extracted with 12 ml of hexanes/methyl tert-butyl ether (MTBE) (200:3, v/v). Unoxidized cholesterol esters and triglycerides elute in this fraction. Subsequently, the column was extracted with 12 ml of methanol/MTBE (5:95, v/v) to remove non-polar lipids and cholesterol. A final extraction using 15 ml of MTBE/methanol/ammonium acetate (0.01 m) (5:8:2, v/v/v) eluted oxidized and unoxidized phospholipids. Samples were dried under N₂ and reconstituted in methanol/water (2:1). Products were analyzed by UPLC/ESI-Ion Trap-MS using a Waters Acquity UPLC (Waters Corp., Milford, MA) with a Phenomenex Luna 3μ C-8 column (150 × 4.6 mm) at a flow rate of 1.0 ml/min and a flow splitter, which sent ∼35% of the mobile phase to the mass spectrometer. A linear gradient was used starting with 75% solvent A (methanol/water 1:1, 20 mm ammonium acetate) to 0% in 1–10 min, held at 0% from 10- 20 min., and returned to 75% at 25 min. Mobile phase B consisted of 20 mm ammonium acetate in methanol. MS was carried out on a ThermoFinnigan LCQ Deca-XP Ion Trap mass spectrometer. The ESI source was fitted with a deactivated fused silica capillary (100-μm inner diameter), and the mass spectrometer was operated in the negative ion mode. Nitrogen was used as the sheath gas at 38 p.s.i. The capillary temperature was 350 °C. The spray voltage was 4.5 kV, and the tube lens voltage was 100 V. CID of the molecular ion of putative A₃/J₃-IsoPs esterified in phospholipids was performed at 25 or 35 eV under 1.5 mTorr of argon. Spectra were displayed by averaging scans across chromatographic peaks. Data acquisition and analysis were performed using Xcaliber software, version 2.0.

RESULTS

Formation and Characterization of A₃/J₃-IsoPs in Vitro—EPA was oxidized using the free radical azo initiator AAPH and then analyzed for A₃/J₃-IsoPs. A representative SIM chromatogram obtained from the GC/MS analysis of putative A₃/J₃-IsoPs is shown in Fig. 2. The two large peaks shown in the lower m/z 438 ion current chromatogram represent the syn- and anti-MOX isomers of the internal standard [²H₄]PGA₂. The peak marked with an asterisk (*) is used for calculations of IsoP formation, while the peak marked with a plus (+) represents the other internal standard isomer. In the upper m/z 432 ion current chromatogram are a series of chromatographic peaks eluting over a 1-min interval. These compounds possess the molecular mass predicted for derivatized A₃/J₃-IsoPs. Furthermore, these chromatographic peaks elute at a later retention time than the deuterated PGA₂ internal standard; it would be expected that A₃/J₃-IsoPs might elute at a different time than PGA₂ because of the presence of an additional carbon-carbon double bond in the molecule (16). Similar to A₂/J₂-IsoPs formed from arachidonic acid oxidation and A₄/J₄-IsoPs formed from DHA oxidation, it is not possible to differentiate between A-ring-containing molecules and J-ring-containing molecules because both chromatograph similarly and have an identical m/z ratio. It should also be noted that the retention times over which A₃/J₃-IsoPs elute may differ somewhat in the various figures in this report as analyses were performed on different days using different columns that varied in length.

FIGURE 2.

SIM chromatogram obtained from the analysis of the MOX-TMS ether derivative of A₃/J₃-IsoPs generated during a 24-h in vitro oxidation of EPA. The series of chromatographic peaks in the upper m/z 432 chromatogram represent putative A₃/J₃-IsoPs. The two large peaks in the lower m/z 438 chromatogram represent the syn- and anti-MOX isomers of the [²H₄]PGA₂ internal standard. The amount of A₃/J₃-IsoPs in this sample was calculated to be 180 ng/mg EPA by comparing the area of the peaks in the m/z 432 chromatogram to the area of the m/z 438 peak indicated by the asterisk (*). The other MOX isomer of the deuterated internal standard is represented with a plus (+) sign.

Additional experimental approaches were used to provide further evidence that the compounds represented by the chromatographic peaks in the m/z 432 channel were A₃/J₃-IsoPs. First, the m/z 431 ion current chromatogram contained no chromatographic peaks, indicating that the peaks in the m/z 432 channel are not natural isotope peaks of compounds generating an ion of less than m/z 432 (data not shown). Analysis of A₃/J₃-IsoPs as [²H₃]MOX derivatives resulted in a shift of the m/z 432 chromatographic peaks up 3 Da to m/z 435, indicating the presence of one ketone (Fig. 3). As well, analysis of putative A₃/J₃-IsoPs as [²H₉]TMS ether derivatives resulted in a shift of the m/z 432 chromatographic peaks up 9 Da to m/z 441, indicating the presence of one hydroxyl group (Fig. 4). Finally, A₃/J₃-IsoPs were analyzed following catalytic hydrogenation as MOX-TMS ether derivatives. Prior to hydrogenation, there were only a few chromatographic peaks present 8 Da above m/z 432 in the m/z 440 chromatogram that did not elute with putative A₃/J₃-IsoPs. Following hydrogenation, however, intense chromatographic peaks appeared at m/z 440 with a loss of chromatographic peaks in the m/z 432 chromatogram, indicating the presence of four double bonds (Fig. 5). Collectively these data indicate that the compounds represented by the chromatograms shown in Fig. 2 have the functional groups and the double bonds predicted for A₃/J₃-IsoPs.

FIGURE 3.

Analysis of putative A₃/J₃-IsoPs as [²H₃]MOX derivatives. A, analysis of compounds prior to [²H₃]MOX derivatization. The peaks in the m/z 432 chromatogram represent the MOX-TMS ether-derivatized A₃/J₃-IsoPs shown in Fig. 2. As seen in the lower chromatogram, only small amounts of compounds were detected 3 Da above m/z 432. B, analysis after [²H₃]MOX derivatization. No peaks were detected at m/z 432. A series of peaks were detected at m/z 435, which correspond to the m/z of A₃/J₃-IsoPs after [²H₃]MOX derivatization. This increase of 3 Da indicates the presence of one ketone group in these molecules.

FIGURE 4.

Analysis of putative A₃/J₃-IsoPs as [²H₉]TMS ether derivatives. A, analysis of compounds prior to [²H₉]TMS ether derivatization. The peaks in the m/z 432 chromatogram represent the MOX-TMS ether-derivatized A₃/J₃-IsoPs shown in Fig. 2. As shown in the lower chromatogram, no compounds were detected 9 Da above m/z 432. B, analysis after [²H₉]trimethylsilyl ether derivatization. Only a small amount of material was detected at m/z 432. A series of peaks were detected at m/z 441, which correspond to A₃/J₃-IsoPs after [²H₉]TMS ether derivatization. The increase of 9 Da indicates the presence of one hydroxyl group in these molecules.

FIGURE 5.

Analysis of the putative A₃/J₃-IsoPs generated in vitro as MOX-TMS ether derivatives prior to and after catalytic hydrogenation. A, analysis of compounds prior to hydrogenation. The peaks in the m/z 432 chromatogram represent the MOX-TMS ether-derivatized A₃/J₃-IsoPs shown in Fig. 2. In the lower chromatogram, only small amounts of non-coeluting compounds were detected 8 Da above m/z 432. B, analysis after catalytic hydrogenation. No peaks were detected at m/z 432. A series of peaks were detected at m/z 440, which correspond to the m/z of the EPA-derived cyclopentenone IsoPs after catalytic hydrogenation. This increase of 8 Da indicates the presence of four double bonds in these molecules.

Adduction of A₃/J₃-IsoPs with Glutathione in Vitro—One of our primary interests in the formation of cyclopentenone IsoPs in vivo is related to the fact that A₃/J₃-IsoPs contain α,β-unsaturated carbonyl moieties, which render these molecules electrophilic (14). Thus, as a further characterization step, we sought to determine whether purified A₃/J₃-IsoPs obtained from the oxidation of EPA in vitro conjugate with GSH in the presence of GST. We have previously shown that cyclopentenone eicosanoids including 15-A_2t-IsoP and A₄/J₄-NPs readily adduct GSH in the presence of GST in vitro (14, 15). Formation of GSH conjugates of IsoPs was assessed first by determining the percent of A₃/J₃-IsoPs that did not extract into an organic solvent, methylene chloride, at pH 3. We found that 62 ± 5% of the A₃/J₃-IsoPs were in the form of a polar conjugate that would not extract after 2 h of incubation with GSH and GST (data not shown). These results were analogous to those found for 15-A_2t-IsoP and for A₄/J₄-NPs (14, 15). These data suggest that, like other cyclopentenone IsoPs, A₃/J₃-IsoPs are reactive molecules capable of undergoing nucleophilic addition.

Subsequently, putative A₃/J₃-IsoP-GSH adducts were definitively identified by LC/MS. Adducts were partially purified from an incubation mixture and subjected to LC as described above. The predicted [parent molecule–H⁺]^– ion, hereafter referred to as “M^–,” for A₃/J₃-IsoP-GSH adducts is m/z 638. SIM analysis showed multiple chromatographic peaks with this m/z eluting from 10.5 to 11.5 min, presumably representing GSH adducts of different A₃/J₃-IsoP stereoisomers. CID experiments confirmed the molecules at m/z 638 to be A₃/J₃-IsoP-GSH conjugates; a composite CID spectrum was obtained by summing scans over these peaks and is shown in Fig. 6. As is apparent, the molecular ion M^– is present at m/z 638. Relevant product ions include m/z 620 [M–H₂O]^–, m/z 306 [M–A₃/J₃-IsoP]^–, m/z 272, and m/z 254. The structures of these latter two product ions are shown in Fig. 6. Similar CID spectra were observed when a synthesized standard, 15-A_3t-IsoP, was incubated with GST and GSH (data not shown). Together, these data confirm that the molecules putatively identified as A₃/J₃-IsoPs are reactive and capable of undergoing nucleophilic addition reactions with GSH.

FIGURE 6.

LC/ESI-MS/MS analysis of the GSH conjugate of A₃/J₃-IsoPs. The parent ion is m/z 638. Material was subjected to CID at 18 eV, and the product ions scanned from m/z 30–700. Spectra were obtained by averaging the scans across the eluting LC chromatographic peaks from 10.5 to 11.5 min.

Time Course of A₃/J₃-IsoP in Vitro—Having provided significant evidence for the formation of the A- and J-ring IsoPs derived from EPA in vitro, we next examined the time course for their formation. For these studies, EPA was oxidized in the presence of AAPH for varying times from 0–36 h. The results demonstrated that A₃/J₃-IsoP levels increase dramatically in a time-dependent manner to a maximum of 195 ± 33 ng/mg EPA when quantified by GC/MS and to a maximum of 196 ± 23 ng/mg EPA when quantified by LC/MS methods (Fig. 7). As with GC/MS quantification, LC/MS quantification was performed by measuring the relative abundance of material representing A₃/J₃-IsoPs compared with that of a deuterated PGA₂ internal standard. Despite the fact that the quantification of these compounds using both GC/MS and LC/MS gives similar results, it should be kept in mind that our methodology relies upon the quantification of A₃/J₃-IsoPs in total based upon a single ion as opposed to other studies we have undertaken herein utilizing various tandem MS approaches to structurally identify different stereo- and regioisomers.

FIGURE 7.

Time course of formation of A₃/J₃-IsoPs during oxidation of EPA in vitro by incubation with AAPH using LC/MS and GC/MS quantification methods. A, LC/MS quantification. B, GC/MS quantification. Data are expressed as means of experiments (n = 6).

Analysis of A₃/J₃-IsoPs by LC/ESI-MS/MS—To provide direct evidence that the compounds analyzed by GC/NICI-MS are A₃/J₃-IsoPs, LC/ESI-MS/MS in the negative ion mode was employed. The predicted M^– for A₃/J₃-IsoPs is m/z 331. The SIM chromatogram of this ion is shown in Fig. 8A. As is evident, multiple chromatographic peaks are present that presumably represent different A₃/J₃-IsoPs stereoisomers. All of the chromatographic peaks in Fig. 8A were next analyzed using CID to obtain structural information about these molecules. The observed product ions were consistent with various series of A₃/J₃-IsoP regioisomers. As a representative sample, the composite CID spectra at three retention times, including RT 14.91, 17.33, and 19.51 min, are shown in Fig. 8, B–D. CID of the parent ion at m/z 331 resulted in the formation of a number of relevant product ions common to all of the A₃/J₃-IsoP regioisomers including m/z 313 [M–H₂O]^–, m/z 295 [M–2H₂O]^–, m/z 287 [M–CO₂]^–, and m/z 269 [M–H₂O-CO₂]^–. Other prominent product ions were present that resulted from the fragmentation of each of the six specific A₃/J₃-IsoP regioisomers. On the basis of our previous work and studies by other groups (16, 19–21), these ions can be explained as follows. In Fig. 8B (RT 14.91), they include m/z 229 ([M–CHOCH₂CH₃-H₂O-CO₂]^–) (18-series) and m/z 215 ([M–CHOCH₂CH=CHCH₂CH₃-H₂O-CO₂]^–)(15-series). These data suggest that the mass spectrum shown in Fig. 8B represents a mixture of two predominant A₃/J₃-IsoP regioisomers (15- and 18-series) that would be predicted to form. In Fig. 8C (RT 17.33), the prominent product ions are m/z 215 (15-series), and m/z 175 ([M–CHOCH₂CH=CHCH₂CH=CHCH₂CH₃-H₂O-CO₂)]^–) (12-series). In Fig. 8D (RT 19.51), the major product ions are m/z 149, the characteristic fragment for the 11-series (-CHOCH₂CH=CHCH₂CH= CH(CH₂)₃COO^–-CO₂), m/z 109, the characteristic fragment of 8-series (-CHOCH₂CH=CH(CH₂)₃COO^–-CO₂), and m/z 97, the characteristic fragment of the 5-series (-CHO(CH₂)₃COO^–-H₂O). Taken together, these data provide direct evidence for the formation of a series of A₃/J₃-IsoPs generated from EPA peroxidation.

FIGURE 8.

Analysis of putative A₃/J₃-IsoPs as free acids by LC/ESI-MS/MS from in vitro oxidation of EPA. A, SIM chromatogram of the [parent molecule–H⁺]^– (M^–) ion at m/z 331 from LC/ESI-MS analysis of putative A₃/J₃-IsoPs obtained from AAPH-induced oxidation of EPA in vitro. Peaks that were subjected to CID are denoted by arrows. The m/z 331 ion was subjected to CID at 21eV, and product ions were scanned from 30–400. B, CID spectrum by obtaining scans over the chromatographic peak at RT: 14.91 min. C, CID spectrum obtained by summing scans over the chromatographic peak at RT: 17.33 min. D, CID spectrum obtained by summing scans over the chromatographic peak at RT: 19.51 min.

The above results were confirmed using SRM. The major unique identifying fragments of the six EPA-derived A₃/J₃-IsoP regioisomers identified in our CID experiments are shown in Fig. 9. They result from cleavage of the molecule α to hydroxyl groups on the carbon backbone. In the SRM experiments, the transition of the precursor ion (m/z 331) to the product ions in Fig. 9 was monitored, and the resulting chromatograms are shown in Fig. 10. Fig. 10A shows the chromatograms obtained from the analysis of EPA oxidized in vitro. All 6-series of A₃/J₃-IsoP regioisomeric characteristic fragments are detected eluting between 14 and 21 min, suggesting that all of these regioisomers are formed from in vitro oxidation of EPA.

FIGURE 9.

Structurally specific fragmentation of the different A₃/J₃-IsoP regioisomers. The major fragmentation of these molecules occurs α to their hydroxyl group.

FIGURE 10.

SRM analysis of putative A₃/J₃-IsoPs obtained from (A) in vitro oxidation of EPA and (B) rat livers after 4 weeks of fish oil supplementation. A, in vitro. B, from rat livers after 4 weeks of EPA supplementation (4% by weight). Please refer to Fig. 9 for specific ions monitored for the different A₃/J₃-IsoP regioisomers.

Formation of A₃/J₃-IsoPs in Vivo—We then undertook experiments to determine whether A₃/J₃-IsoPs are formed esterified in phospholipids in vivo. Baseline levels of EPA in animals and humans are extremely low; thus, A₃/J₃-IsoPs are below the levels of detection (<25 pg/g of tissue). To determine if these compounds are generated in vivo, we supplemented rats with a Mendehen fish oil-containing diet (4% by volume) for 4 weeks. Subsequently, animals were sacrificed, and tissue lipids were extracted and analyzed for oxidation products. Putative A₃/J₃-IsoPs were analyzed as free compounds following enzymatic hydrolysis of oxidized compounds from phospholipids. Livers were chosen for these studies as it has been shown that fish oil supplementation markedly increases the level of EPA found in this organ in animals and humans. A representative GC/MS ion current chromatogram obtained from these analyses is shown in Fig. 11. The chromatographic peaks in the lower m/z 438 chromatogram represent the syn- and anti-MOX isomers of the internal standard [²H₄]PGA₂ whereas the upper m/z 432 chromatogram represents endogenous A₃/J₃-IsoPs. The peak marked with an asterisk was used for quantification.

FIGURE 11.

SIM chromatogram obtained from the GC/MS analysis for A₃/J₃-IsoPs esterified in liver tissue from fish oil-fed rats. The series of peaks shown in the m/z 432 chromatogram represent putative A₃/J₃-IsoPs. The two large peaks in the lower m/z 438 chromatogram represent the syn- and anti-O-N-methyoxime isomers of the [²H₄]PGA₂ internal standard. The amount of A₃/J₃-IsoPs present in this sample was 125.5 ng/g of liver tissue by comparing the area of the peaks in the m/z 432 chromatogram to the area of the m/z 438 peak indicated with the asterisk (*). The other N-methyloxime isomer is represented with a plus (+) sign.

Analogous to studies performed with EPA oxidized in vitro, LC/MS experiments were carried out to obtain further evidence that the chromatographic peaks in the m/z 432 ion current chromatogram of Fig. 11 were A₃/J₃-IsoPs. For these studies, liver tissue from the previously mentioned feeding study was utilized. Employing CID, all of the product ions predicted to be common to all A₃/J₃-IsoP regioisomers, including m/z 313 [M–H₂O]^–, m/z 295 [M–2H₂O]^–, m/z 287 [M–CO₂]^–, and m/z 269 [M–H₂O-CO₂]^–, were detected (Fig. 12, B and C). In addition, product ions representing the six different A₃/J₃-IsoP regioisomers (m/z 229, m/z 215, m/z 175, m/z 149, m/z 109, and m/z 97) were identified (Fig. 12, B–D). These six regioisomers were also detected in SRM experiments (Fig. 10B). Taken together, these experiments provide evidence that A₃/J₃-IsoPs are formed in abundance in vivo.

FIGURE 12.

Analysis of putative A₃/J₃-IsoPs as free acids by LC/ESI-MS/MS from rat livers after 4 weeks of fish oil supplementation. A, SIM chromatogram of the [parent molecule–H⁺]^– (M^–) ion at m/z 331 from LC/ESI-MS analysis of putative A₃/J₃-IsoPs obtained from liver tissue from rats supplemented with 4% EPA for 4 weeks. Peaks that were subjected to CID are denoted by arrows. The m/z 351 ion was subjected to CID at 21eV, and product ions were scanned from 30 to 400. B, CID spectrum by obtaining scans over the chromatographic peak at RT: 12.60 min. C, CID spectrum obtained by summing scans over the chromatographic peak at RT: 17.01 min. D, CID spectrum obtained by summing scans over the chromatographic peak at RT: 19.16 min.

Finally, we determined the effect of enhanced oxidative stress on endogenous A₃/J₃-IsoPs in rats fed a 4% fish oil-supplemented diet. Levels of these compounds increased from a baseline level of 19 ± 2 ng A₃/J₃-IsoPs/g liver to 102 ± 15 ng/g liver after administration of CCl₄, a potent inducer of oxidant stress (Fig. 13). The effect of fish oil supplementation on levels of A₂-/J₂-IsoPs derived from arachidonic acid oxidation was also studied, and we found that EPA supplementation significantly reduced their level of formation (p < 0.05, data not shown). In addition, it should be noted that we have previously undertaken studies to confirm the stability of EPA in the fish oil diet fed to rodents to ensure that the source of oxidized EPA products was not the diet and that they were generated in vivo (22).

FIGURE 13.

Levels of A₃/J₃-IsoP levels in liver tissue from rats fed a fish oil-supplemented diet (4% by weight). A, at baseline without orogastric administration of CCl₄ (1 ml/kg). B, with administration of CCl₄ (1 ml/kg). Data are expressed as means ± S.D. n = 4 experiments. **, p < 0.005; –CCl₄ to +CCl₄.

Formation of A₃/J₃-IsoPs in Vivo in EPA-containing Phospholipids—Because IsoPs have been shown to form in situ on phospholipids, we sought to characterize the presence of A₃/J₃-IsoPs esterified in phospholipids in vivo from liver tissue from rats fed a fish oil-supplemented diet. A number of LC/MS techniques have been developed to analyze intact phospholipids as well as oxidized phospholipids (23–28). To provide direct evidence that these compounds exist esterified in phospholipids, we utilized UPLC/ESI-ion trap-MS⁴. Oxidized phospholipids were purified using the solid phase extraction method mentioned above. Structures monitored based on their m/z ratio in these experiments are shown in Fig. 14. Using a mobile phase containing acetate, phospholipid species were detected adducted with an acetate moiety. Thus, the predicted m/z for A₃/J₃-IsoPs esterified in the sn-2 position of a phospholipid with a stearoyl (C18:0) in the sn-1 position and a phosphatidylcholine headgroup (SIsoPPC) adducted to an acetate ion is 896 [SIsoPPC + acetate ion]^–. The CID spectrum of m/z 896 is shown in Fig. 14A. The most abundant product ion detected was m/z 822, which represented [SIsoPPC–CH₃]^–. We next utilized MS³ to obtain structural data on m/z 822. As can be seen in the spectrum in Fig. 14B, it was possible to detect a product ion at m/z 331, which corresponds to the expected m/z for A₃/J₃-IsoP fatty acids, and its dehydration product m/z 303. It was also possible to detect stearic acid from the sn-1 position at m/z 283.

FIGURE 14.

UPLC/ESI-Ion Trap-MS analysis of putative A₃/J₃-IsoPs formed in situ on phospholipids from rat livers after 4 weeks of fish oil supplementation (4% by weight). A, MS² CID mass spectrum of m/z 896, the parent molecule SIsoPPC adducted to an acetate ion [SIsoPPC+acetate ion]^–. B, MS³ CID spectrum of m/z 822 that represents loss of a methyl group [SIsoPPC–CH₃]^–. C, MS⁴ CID spectrum of m/z 331, which represents A₃/J₃-IsoP fatty acids.

Consecutive CID (MS⁴) was carried out again to obtain structural data on m/z 331. The product ions of m/z 331 [M^–] were monitored between m/z 90 and m/z 340. As seen in the MS⁴ spectrum of m/z 331 (Fig. 14C), the product ions observed include m/z 313 [M–H₂O]^–, m/z 287 [M–CO₂]^–, m/z 295 [M–2H₂O]^–, and m/z 233 [M–C₆H₁₀O]^–, an ion that indicates fragmentation at carbon C-5. (Previous reports have noted that this loss of 98 Da is common to many different fatty acids and IsoPs (15, 19, 20, 21, 29–31)). In addition to these product ions, the 15-series A₃/J₃-IsoP regioisomer specific α-fragment (m/z 215) was also detected. Similar in vivo studies were undertaken to show the presence of A₃/J₃-IsoPs esterified in phospholipids containing a palmitoyl (C16:0) group in the sn-1 position (data not shown). Similar data were also obtained from experiments in which EPA-containing phospholipids were oxidized in vitro. Taken together, these findings, for the first time, directly demonstrate the formation of cyclopentenone eicosanoids esterified in phospholipids in vivo.

DISCUSSION

These studies have elucidated a novel class of cyclopentenone eicosanoids, A₃/J₃-IsoPs, formed in vitro and in vivo from the free radical-initiated peroxidation of EPA. The mechanism of formation of A₃/J₃-IsoPs is outlined in Fig. 1. Like other classes of IsoPs, the A₃/J₃-IsoPs are formed in situ esterified in phospholipids and are then presumably released in the free form by a phospholipase(s). Recently, it has been shown that platelet-activating factor acetyl hydrolases are capable of hydrolyzing F₂-IsoPs from phospholipids, and, thus, it is conceivable that these enzymes also hydrolyze phospholipids containing A₃/J₃-IsoPs (12). Previously, we have demonstrated that another class of IsoP-like molecules, F₃-IsoPs, is also formed from the oxidation of EPA in vivo (16). From our studies with arachidonic acid, it is known that F-ring compounds are formed as a result of reduction of the bicyclic endoperoxide intermediates. On the other hand, A₃/J₃-IsoPs are generated after rearrangement of their endoperoxides precursors to molecules with E- and D-prostane rings that subsequently dehydrate (10, 14, 15, 29). The identification and characterization of D₃-/E₃-IsoPs, however, remains the subject of future work. Nonetheless, it should be noted that A₃/J₃-IsoPs are formed in significant amounts comparable to F₃-IsoPs.

A major impetus driving the present studies is the fact that fish oil supplementation has been shown to be beneficial in the prevention of important human diseases such as atherosclerosis and sudden cardiac death, neurodegeneration, and various inflammatory disorders (1, 3, 33–36). Although the mechanism(s) by which the ω-3 fatty acids exert their biological activity is unknown, an important anti-atherogenic and anti-inflammatory mechanism of EPA action is its interference with the arachidonic acid cascade that generates pro-inflammatory eicosanoids via the cyclooxygenases and lipoxygenases (1, 37). A recent Japanese study examining the effects of EPA on major coronary events in hypercholesterolemia (JELIS) demonstrated a 19% reduction in coronary events after 4.6 years of EPA intervention in hypercholesterolemic patients (33). EPA has also been shown not only to replace arachidonic acid in phospholipid bilayers but is also a competitive inhibitor of cyclooxygenases, reducing the production of 2-series PGs and thromboxanes as well as 4-series leukotrienes (38–40). The 3-series and 5-series PGs and leukotrienes derived from EPA have also been shown to be either less biologically active or are inactive in comparison to the products formed from arachidonic acid (39, 41, 42). Thus, compounds derived from EPA are thought to exert less inflammatory activity. Further, Serhan and co-workers have described a specific group of polyoxygenated EPA derivatives termed resolvins that are produced enzymatically in various tissues; these molecules have been shown to inhibit cytokine expression and other inflammatory responses (7, 43–47).

In addition to identifying enzymatically generated oxygenated derivatives of EPA, there has been significant interest in studying the beneficial effects of non-enzymatically derived EPA peroxidation products (8, 16, 48). Sethi and co-workers (4–6) reported Cu²⁺-oxidized EPA, but not native EPA, potently inhibits human neutrophil and monocyte adhesion to endothelial cells, a process linked to atherosclerosis. This effect was induced via inhibition of endothelial adhesion receptor expression and was modulated by the activation of the peroxisome proliferator-activated receptor-α (PPAR-α) by ω-3 oxidation products. Further studies by this group demonstrated that oxidized EPA markedly reduced leukocyte adhesion to venular endothelium of lipopolysaccharide-treated mice in vivo, and the effect was not observed in PPAR-α-deficient mice, suggesting that oxidized EPA products are partially responsible for the anti-inflammatory effects ascribed to EPA consumption. Similarly, Vallve et al. (49) have shown that various non-enzymatic aldehyde oxidation products of EPA decrease the expression of the CD36 receptor, a receptor has been linked to atherosclerosis, in macrophages. It has also been demonstrated that EPA peroxidation products modulate the generation of other endothelial inflammatory molecules (5, 50). Further, Arita et al. (51) have reported that non-enzymatically oxidized EPA enhances apoptosis in HL-60 leukemia cells showing that oxidized ω-3 PUFAs are anti-proliferative. Similar findings have been reported in HepG2 (human hepatoma) cells, AH109A (rat liver cancer) cells, and glomerular endothelial cells (52–54). In none of these studies, however, have the specific peroxidation products responsible for these effects been identified. Thus, the identification of specific oxidation products at a molecular level of EPA has been largely lacking.

Recently, we reported that oxidation products of EPA activate the Nrf2 transcription factor that mediates important anti-oxidant responses in response to reactive electrophiles via induction of gene transcription. These nonenzymatically derived products were postulated to interact with Keap1, the direct inhibitor of Nrf2, initiating Keap1 dissociation from Cullin3 to yield Nrf2 translocation and activation. It was hypothesized therein that molecules containing an α,β-unsaturated carbonyl moiety, specifically J₃-IsoPs, were responsible for eliciting this biological activity (8). The previously mentioned work, however, did not definitively characterize the structure of EPA-derived cyclopentenone IsoPs. Nonetheless, this previous work was the first to ascribe specific nonenzymatic EPA peroxidation products with biological activity. Concentrations of oxidized EPA products necessary to exert this biological activity were in the micromolar range, and, interestingly, in the studies reported herein, we measured levels of A₃/J₃-IsoPs in the range of 0.3 μm in liver tissue from rats after induction of an oxidant stress. These finding suggest that A₃/J₃-IsoPs can be generated in vivo at biologically relevant concentrations. Of note, these levels of A₃/J₃-IsoPs represent esterified compounds. Previously, however, we have shown that IsoPs are rapidly hydrolyzed from tissue lipids to the free form (11, 12). In addition, it is important to note that since, as a class, the biological activities of A₃/J₃-IsoPs are likely primarily related to the highly reactive cyclopentenone moiety. We would therefore postulate that all of the different stereo- and regioisomers of the A₃/J₃-IsoPs generated in vivo contribute to their biological activities.

The studies reported herein have systematically characterized the formation of A₃/J₃-IsoPs from the oxidation of EPA in vitro and in vivo. These molecules were identified using a variety of complementary chemical and MS approaches. As predicted, 6-series of A₃/J₃-IsoPs were identified. The fragmentation patterns of A₃/J₃-IsoPs are similar to F₃-IsoPs, and, indeed, previously acquired information regarding the F₃-IsoPs was extremely useful for the present work (16). In addition, based on previous studies with F₃-IsoPs, we would hypothesize that the 5-series and 18-series A₃/J₃-IsoPs should predominate, given the ability of the other 4-series of A₃/J₃-IsoP regioisomers to further oxidize to yield novel dioxolane-IsoPs derived from EPA (22). But, as is evident from data presented in Fig. 10, all 6-series of A₃/J₃-IsoPs appear to be formed in fairly equal abundance (16, 20). The reason for this discrepancy is unclear although we might propose this is due to preferential metabolism in vivo of particular A₃/J₃-IsoP regioisomers. In this regard, further experimentation will need to be undertaken to understand the mechanisms of formation and metabolism of A₃/J₃-IsoPs.

In summary, we report the discovery that A/J-ring IsoPs are formed in vitro and in vivo as products of the non-enzymatic free radical-catalyzed peroxidation of EPA, a major ω-3 fatty acid found in fish oil. The fact that A₃/J₃-IsoPs are readily detectable in liver tissues from animals that have been supplemented with fish oil suggests the compounds represent an important pathway for the oxidative metabolism of EPA. Further understanding of the biological consequences of the formation of these novel compounds and the factors influencing their formation and metabolism are underway and will likely provide valuable insights into the beneficial effects associated with fish oil consumption on human health.