Abstract
Resolvin E1 (RvE1; 5S,12R,18R-trihydroxy-6Z,8E,10E,14Z,16E-eicosapentaenoic acid) is a potent anti-inflammatory and proresolving mediator derived from the omega-3 eicosapentaenoic acid. In this study, we report the RvE1 metabolome, namely, the metabolic products derived from RvE1. RvE1 was converted to several novel products by human polymorphonuclear leukocytes and whole blood as well as in murine inflammatory exudates, spleen, kidney, and liver. The potential activity of each of the newly identified products was directly compared with that of RvE1. The new RvE1 products elucidated included 19-hydroxy-RvE1, 20-carboxy-RvE1, and 10,11-dihydro-RvE1. Metabolomic profiles of RvE1 were species-, tissue-, and cell type-specific. Direct comparisons of the bioactions between isolated RvE1 metabolic products indicated that 10,11-dihydro-RvE1, 18-oxo-RvE1, and 20-carboxy-RvE1 displayed reduced bioactivity in vivo. At concentrations as low as 1 nM, RvE1 enhanced macrophage phagocytosis, a proresolving activity that was reduced by metabolic inactivation. These results document novel metabolic products of RvE1 that impact its actions and that both omega-1 hydroxylation and reduction of conjugated double bonds in RvE1 are new pathways of four main routes of RvE1 metabolism in mammalian tissues. Together, these findings indicate that, during inflammation and its controlled resolution, specific tissues inactivate proresolving signals, i.e., RvE1, to permit the coordinated return to homeostasis. Moreover, the RvE1 metabolome may serve as a biomarker of these processes.
Resolution of inflammation is an active process (1) controlled in part by the temporal and spatially regulated formation and inactivation of endogenous mediators (2, 3, 4). The termination of proresolving signals is also required to bring the inflamed or injured system back to homeostasis. Recent results have revealed several new families of lipid mediators that are both anti-inflammatory and proresolving that are generated during the resolution phase of inflammation; these include lipoxins, resolvins, and protectins (2). For example, resolvin E1 (5S,12R,18R-trihydroxy-6Z,8E,10E,14Z,16E-eicosapentaenoic acid or RvE1)5 is a member of the resolvin family that is biosynthesized from the precursor eicosapentaenoic acid (EPA) and is a potent local mediator that i) reduces inflammation in skin and peritonitis (5), ii) is formed during colitis (6) and dramatically increases survival of mice with trinitrobenzene sulfonic acid colitis (7), and iii) reduces ocular neovascularization (8) and protects from periodontal inflammation and bone destruction in rabbits (9). RvE1 is a local acting autacoid that proved to display proresolving actions when treatments were given either by topical, i.v., or i.p. routes of administration (4, 10). Thus, it is critical to gain an appreciation of the routes and pathways involved in metabolic inactivation of local proresolving signals, such as RvE1, during homeostasis and inflammation-resolution.
Along these lines, the inactivation of lipoxin A4 (LXA4) (11) and RvE1 involves both carbon position and site-specific dehydrogenation as routes of metabolic inactivation (12). Protection of RvE1 from rapid dehydrogenation and/or metabolic inactivation stabilizes and prolongs its potent in vivo anti-inflammatory actions (12). RvE1 is endogenously formed from eicosapentaenoic acid via cell-cell interactions when aspirin is administered during inflammation and/or via cytochrome P450 conversion of its omega-3 fatty acid precursor EPA (5, 13). The omega-3 polyunsaturated fatty acids (PUFAs) have long been appreciated for their beneficial actions in many human systems, including the immune (14), neural (15, 16) and cardiovascular (17). The molecular mechanisms responsible, however, for these beneficial actions in human trials with supplementation of omega-3 PUFAs remained to be convincingly established (18). In this regard, RvE1 fulfills local mediator criteria in that it biosynthesized from precursor EPA and displays potent and stereoselective actions both in vitro and in vivo, where RvE1 is anti-inflammatory, proresolving, and tissue-protective. For example, RvE1 stops neutrophil infiltration in vivo and transendothelial migration (5), attenuates IL-12 production by dendritic cells and their mobility toward pathogens (13), and stimulates resolution programs in vivo (10).
Therefore, it is essential to chart the metabolic conversion of RvE1 and its potential inactivation pathways as well as the bioactivities of RvE1-derived metabolic products. These are needed to appreciate their potential contributions(s) in the local control of inflammation and its resolution. These are also critically required components to fully appreciate the relationship(s) between dietary intake of omega-3 PUFAs and the local biosynthesis of resolvins and protectins in vivo as well as to design RvE1-related therapeutic interventions. In the present report, we investigated the metabolic profiles and pathways for RvE1 in several mammalian tissues and isolated cells. We also determined the anti-inflammatory and proresolving properties of these newly isolated RvE1 metabolic products that together demonstrate the specificity and selectivity of the RvE1 inactivation process.
Materials and Methods
Materials
RvE1 was prepared by total organic synthesis according to published matching criteria (5, 13) and obtained from the Organic Synthesis Core of the National Institutes of Health Center P50-DE016191 (Core Leader Prof. Nicos A. Petasis, University of Southern California). Calcium ionophore (A23187) was from Sigma-Aldrich. Radiolabeling of RvE1 was purchased as a custom tritiation from American Radiolabeled Chemicals and obtained from catalytic hydrogenation of the provided diacetylenic RvE1 in [3H]gas and isolated in this lab before use via reversed phase-HPLC (RP-HPLC) (as in Ref. 13).
Acute inflammation
Peritonitis was conducted using 6–8-wk-old male FVB mice (Charles River Laboratories) that were anesthetized with isoflurane. Mice were fed Laboratory Rodent diet 5001 (LabDiet) containing 0.19% omega-3 fatty acids without additional fatty acid supplementation. RvE1, RvE1-metabolic products, or vehicle was injected into the tail vein (20 ng/mouse). Zymosan A in 1 ml of saline (1 mg/ml) was injected ∼2 min later into the peritoneum to induce acute inflammation or peritonitis. At 2 h after zymosan administration, the peritoneal lavages were collected and total leukocytes, polymorphonuclear leukocyte (PMN), monocyte, and lymphocyte numbers were enumerated (19), and the lungs, livers, kidneys, and spleens were harvested. Mice were euthanized with isoflurane in accordance with the Harvard Medical Area Standing Committee on Animals (Protocol Number 02570).
RvE1 metabolism/conversion by cells and tissues
Human whole (venous) blood was collected with heparin from deidentified healthy volunteers (who denied taking medication 2 wk before donation; Brigham and Women’s Hospital Protocol No. 88-02642). Following collection, whole blood (5.0 ml) was immediately incubated with RvE1 (500 ng; 60 min, 37°C). In the meantime, human PMNs (50 × 106 cells) were isolated from the whole blood by Ficoll gradient, then immediately incubated with 500 ng RvE1 and in 0.5 ml PBS +/+ with calcium ionophore (A23187, 5 μM (pH 7.45), 30 min, 37°C). For incubations with murine resident macrophages, male FVB mice (6–8-wk old) were euthanized with isoflurane and the peritonea were lavaged with PBS −/−. Harvested cells (7 × 106 cells) were plated on 6-well plates with RPMI 1640 with 10% FBS and incubated at 37°C. After 2 h, the supernatants were removed and adhered cells were washed with PBS +/+ two times (70–80% of harvested cells adhered to the plate). Cells were exposed to A23187 (5 μM) for 1 min, then 200 ng of RvE1 was added and cells were incubated ((pH 7.45) 37°C). After 2 h, cells were scraped and 2 vol of ice-cold methanol were added; samples were stored at −80°C until C18 solid phase extraction.
Each type of organ (∼200 mg) harvested from mice with peritonitis (2 h) was gently homogenized on ice (∼4°C) then incubated with RvE1 (500 ng) in PBS +/+ (pH 7.45) at 37°C for 60 min. Each incubation was stopped with the addition of 2 vol of ice-cold methanol (19, 20), centrifuged, and the supernatants were extracted using C18 solid phase extraction.
LC-MS/MS-based mediator lipidomics
Following extraction, isolated materials were taken for analysis of RvE1 and the potential RvE1-derived products using liquid chromatography-UV-tandem mass spectrometry (LC-UV-MS/MS) equipped with a HPLC (P4000) coupled to a photo-diode-array UV detector and an ion trap (LCQ) MS/MS (Thermo Electron) (for further details, see Refs. 5, 21). The mobile phase flowed at 0.2 ml/min using a C18 LC column (Phenomenex Luna 2.1 mm × 150 mm × 5 μm) for these profiles.
To monitor endogenous RvE1-related compounds and RvE1-derived metabolites from the time course of acute murine peritonitis, the samples were extracted as in Ref. 21 and injected into an HPLC-UV (HP1100; Agilent) coupled to an ion trap-tandem-mass spectrometer (Q Trap 3200; Applied Biosystems/Sciex) equipped with a C18 LC column (Agilent Eclipse Plus, 4.6 mm × 50 mm × 1.8 μm) and a mobile phase flow rate of 0.4 ml/min. The chromatography cycle was 15 min using a mobile phase of methanol/water/acetic acid (60/40/0.01;v/v/v) that was changed to 80/20/0.01 (v/v/v) after 5 min, 95/5/0.01 (v/v/v) after 8 min, and 100/0/0.01 (v/v/v) to wash the column after 14 min.
Radioactive tracing of RvE1 metabolic profile
Isolated human PMNs (50 × 106 cells) in 1 ml PBS +/+ were exposed to zymosan A (1 mg) for 3–5 s and then incubated with 1 μg unlabeled RvE1 along with 104 cpm/μg tritium-labeled RvE1 (6,7,13,14-tetra-tritiated-RvE1) for 40 min ((pH 7.45) 37°C). Extracted materials were subject to RP-HPLC equipped with a Luna C18 column (2.1 mm × 150 mm × 5 μm; Phenomenex). The flow rate was set at 0.2 ml/min and collected at 30-s intervals. Each fraction was mixed with scintillation fluid, and radioactivity was counted with a scintillation counter (13).
Macrophage phagocytosis with RvE1 and its related metabolic products
The proresolving actions of RvE1 and related products were assessed as in Godson et al. (22). Briefly, exudates from the peritonea of naive mice that were euthanized with isoflurane were collected, and resident macrophages were plated on a 24-well plate (105 cells/well) in PBS +/+ and incubated for 30 min at 37°C. The compounds to be tested were added to the wells at indicated concentrations, and cells were incubated in the dark for 15 min at 37°C. FITC-labeled zymosan was then added to the wells, which were incubated again in the dark for 30 min at 37°C. The wells were subsequently aspirated, and extracellular fluorescence was quenched by brief addition of trypan blue, followed by aspiration and suspension again in PBS +/+ (pH 7.45). Plates were read using a PerkinElmer Victor (3) plate reader.
Statistical analysis
All results are expressed as mean ± SEM. Statistical significance for differences between groups was determined using Student’s t test and Fisher’s protected least significant difference.
Results
Human PMN and whole blood generate novel RvE1 metabolic products
Because neutrophils play a key role in inflammation and RvE1 is biosynthesized via interactions of human endothelial cells and PMNs (5), we first studied RvE1 conversion with isolated human PMN (Fig. 1). Mediator lipidomic analysis using LC-UV-MS/MS revealed that ionophore A28137-stimulated PMNs converted RvE1 to novel metabolic products 20-carboxy-RvE1 and 19-hydroxy-RvE1, shown in the selected ion chromatograms in Fig. 1,A. The metabolic product 20-carboxy-RvE1 was identified based on its MS/MS spectrum at molecular anion m/z 379 (M-H), which displayed diagnostic fragment ions at m/z 361 (M-H-H2O), 343 (M-H-2H2O), 325 (M-H-3H2O), 317 (M-H-H2O-CO2), 297, 275 (293-H2O), 263, 255 (291-2H2O), 237 (293-2H2O), 226 (263-2H2O-H), 167 (185-H2O), 143 (185 plus 2H-CO2), and 115 (Fig. 1 B; inset illustrates the main ions). In addition, its UV spectrum displayed λmax = 271 nm, which was indicative of the presence of a conjugated triene chromatophore, and its chromatographic retention time was shorter than that of native RvE1 or other RvE1 metabolites. These properties were consistent with its higher polarity resulting from the addition of a carboxyl group to the ω end (carbon 20 position) of RvE1.
The metabolic product 19-hydroxy-RvE1 was identified based on its physical properties, which included a chromatographic behavior with a C18 column of eluting after 20-hydroxy-RvE1 (Fig. 1,A). The UV spectrum with the λmax = 271 nm obtained for this RvE1-derived product displayed a conjugated triene (Fig. 1,C, inset), and the presence of diagnostic ions in its MS/MS spectrum was consistent with the structure shown in the inset (Fig. 1,C). These ions of diagnostic value present in the MS/MS spectrum (see Fig. 1 C, inset) included m/z 365 (M-H), 347 (M-H-H2O), 321 (320 plus H), 311 (M-H-3H2O), 303 (320-H2O plus H), 291, 283 (320-2H2O-H), 273 (291-H2O), 267 (320-3H2O plus H), 257 (320-H2O-CO2-H), 251, 239 (320-2H2O-CO2-H), 221 (320-3H2O-CO2-H), 205 (223-H2O), 195, 179 (223-CO2), 170, 161 (223-H2O-CO2), 123 (141-H2O), and 115. The UV and MS/MS spectrum of 20-hydroxy-RvE1 were consistent with the recently reported physical properties of this new metabolic product of RvE1 (cf. Ref. 12).
To determine whether RvE1 is converted to metabolic products in human blood, we conducted mediator lipidomic LC-UV-MS/MS analysis with incubations of RvE1 and human blood. One of the new products identified in these incubations was 10,11-dihydro-RvE1 (Fig. 2), which eluted slightly later than RvE1 as shown in the selected ion monitoring chromatogram at m/z 351 (M-H) (Fig. 2,A). This chromatographic behavior was consistent with its structure having one fewer double bond than RvE1. Its structure was further supported on the basis of the MS/MS spectrum, which contained diagnostic ions at m/z 351 (M-H), 333 (M-H-H2O), 321, 315 (M-H-2H20), 307 (M-H-CO2), 297 (M-H-3H2O), 293, 289 (M-H-H2O-CO2), 275 (293-H2O), 271 (M-H-2H2O-CO2), 257 (293-2H2O), 253 (M-H-3H2O-CO2), 247 (267-H2O-2H), 235, 231 (293-H2O-CO2), 225, 217 (235-H2O), 213 (231-H2O), 207 (225-H2O), 197, 189 (225-2H2O), 179 (197-H2O), 163 (225-H2O-CO2), 145 (225-2H2O-CO2), 135 (197-H2O-CO2), 123 (125-2H), and 107 (125-H2O) (Fig. 2 B).
Metabolomic profiles of RvE1 with human and mouse tissues
To determine the capacity of inflammatory exudates to convert RvE1, we incubated murine inflammatory exudates collected at different time intervals during peritonitis with RvE1. The LC-UV-MS/MS analysis indicated that exudates collected at later points during peritonitis converted less RvE1 (Fig. 3,A), i.e., the capacity of inflammatory exudates to metabolize RvE1 apparently declined as peritonitis continued from 2–24 h into the resolution phase. During this phase, the total number of cells in exudates, especially PMN, decreases during resolution and PMN begin to undergo apoptosis (Fig. 3,A, left panel). These factors may account, in part, for the decreased metabolic capacity to convert RvE1, for example at 24 h, in resolving exudates. For the metabolism of tritium-labeled RvE1 by human PMN, the profile of the radioactive chromatogram is plotted as an overlay with the UV chromatogram (Fig. 3 B), where 20-hydroxy-RvE1 was a major component, thus demonstrating that 20-hydroxy-RvE1 is the major metabolic product in human leukocytes. The radioactive chromatogram represents the complete profile of metabolites from tritium-labeled RvE1 because the 3H in labeled RvE1 (6,7,13,14-tetra-tritiated-RvE1) was retained in these products. Also, the use of labeled RvE1 as tracer in these studies provides additional evidence to confirm the conversion of RvE1 to novel compounds.
To address the pathways involved in metabolic conversion of RvE1 in different tissues, we determined the quantitative metabolomic profiles of RvE1 in human blood, human PMN, and murine tissues (lung, liver, kidney, or spleen) (Fig. 3,C). The major RvE1 products from human blood were 10,11-dihydro-RvE1, 18-oxo-RvE1, and 20-hydroxy-RvE1 (Fig. 3,C, right panel). However, with isolated human PMN, more 20-hydroxy-RvE1 was produced, followed by 19-hydroxy-RvE1 and 18-oxo-RvE1 (Fig. 3). The metabolic products obtained were the same for both murine spleen and lung (Fig. 3,C, left panel). It is of interest that murine kidney and liver generated more 18-oxo-RvE1 than other RvE1-derived metabolic products (Fig. 3 C). Along these lines, RvE1 is converted to 18-oxo-RvE1 in vitro by human recombinant 15-prostaglandin dehydrogenase (12).
To determine the profile and appearance of endogenous RvE1 and its metabolic products in inflammatory exudates during the established time course of murine peritonitis and its resolution (4), we obtained peritoneal exudates at specific time intervals and monitored levels of endogenous RvE1 and its related precursor and metabolic products identified herein using LC-UV-MS/MS (Fig. 3 D). The biosynthetic precursor to RvE1, 18-HEPE, was identified in exudates in amounts ranging from ∼50 to 325 pg per mouse in both mice with peritonitis and those not exposed to zymosan. The peak level of 18-HEPE was determined to be 2 h after inducing inflammation and rapidly declined by 12 h. This time interval coincides with the period of rapid leukocytic infiltration in to the peritoneum (4). These findings suggest that leukocytes appearing in exudates in the acute inflammatory phase can rapidly convert 18-HEPE in situ. Endogenous RvE1 accumulated at later time points, identifiable in trace amounts at 48 h and at ∼27 pg at 72 h, and other metabolic products were likely below limits of detection in individual exudates. This accumulation of RvE1 at 48 and 72 h may reflect differences in the cellular exudate composition, as PMN are the predominant cell type during the initial acute inflammatory phase of murine peritonitis whereas mononuclear cells are the predominant cell type during the later resolution phase. Endogenous accumulation of RvE1 at earlier time points in the murine skin air pouch (5), and the temporal differences in the peritoneal exudates, likely reflect permeability in the peritoneum vs the skin (cf. Ref. 4, 5).
Targeted metabolomic analyses using LC-UV-MS/MS-based mediator lipidomics demonstrated that the key products of RvE1 metabolism in cells and tissues are 20-hydroxy-RvE1, 19-hydroxy-RvE1, 18-oxo-RvE1, 10,11-dihydro-RvE1, and 20-carboxy-RvE1. These quantitative results for each indicated that omega-1 hydroxylation and enzymatic reduction of one conjugated carbon-carbon double-bond in RvE1 are two novel major metabolic pathways that coexist in parallel to the recently delineated pathway, namely dehydrogenation of the 18-hydroxy-position of RvE1 (12). Specific cytochrome P450 enzymes are known to oxygenate arachidonic acid, leukotriene B4 (LTB4), and prostaglandins to their respective omega-hydroxy and omega-1-hydroxy metabolites (23, 24). Similarly, 20-hydroxy-RvE1 and 19-hydroxy-RvE1 are likely to be P450 products of RvE1, and 20-carboxy-RvE1 is likely generated from the further oxidation of the intermediate 20-hydroxy-RvE1.
Our earlier results demonstrated that 18-oxo-RvE1 is generated from RvE1 via enzymatic dehydrogenation (12). In the present report, we identified 10,11-dihydro-RvE1 as a key metabolite in tissues in vivo. It is likely that RvE1 is converted enzymatically first to an intermediate 12-oxo-RvE1, then rapidly reduced to 10,11-dihydro-12-oxo-RvE1 that can then be further converted to 10,11-dihydro-RvE1. This may also explain the finding that 10,11-dihydro-RvE1 is a major product in several tissues (Fig. 3). In support of this route of RvE1 metabolism, 10,11-dihydro-12-oxo-RvE1 (Fig. 4) was also identified from incubations of RvE1 with the human THP-1 cell line. To accumulate the transient intermediate, freeze-thaw lysates of THP-1 cells were incubated with RvE1 and the cofactor NAD (500 μM, 37°C, 30 min) that enhanced the yield and mass spectral identification (not shown) of 10,11-dihydro-12-oxo-RvE1. However, we were unable to isolate quantities of this intermediate that would permit assessment of the intermediate’s potential biological activity. This pathway of RvE1 further metabolism appears to be similar to the metabolism proposed for LTB4 conversion to 10,11-dihydro-LTB4, first identified in kidney (25). Similar reduction products were identified earlier for prostaglandins (reviewed in Ref. 26) and lipoxins with human monocytes and macrophages (27). Also, when RvE1 was incubated with isolated murine resident peritoneal macrophages, the major product identified using LC-MS/MS proved to be 10,11-dihydro-RvE1, indicating that both human and mouse isolated macrophages use this pathway for RvE1 (not shown). The proposed metabolic routes and pathways in the RvE1 metabolome are depicted in the scheme shown in Fig. 4.
Anti-inflammatory activities of RvE1 and related metabolic products
To establish whether the newly isolated and identified RvE1 metabolites retain the anti-inflammatory actions of RvE1, each of the new RvE1-derived products was isolated and we compared their actions directly in vivo using acute inflammation, i.e., zymosan A-induced peritonitis (where RvE1 has been established to have anti-inflammatory and proresolving actions). When administered i.v., 20-hydroxy-RvE1 proved to be essentially as potent as RvE1 in reducing infiltration of leukocytes into inflamed peritonea (Fig. 5,A) and particularly at stopping PMN infiltration (Fig. 5 B). The product 19-hydroxy-RvE1 was significantly less potent than its precursor RvE1. Furthermore, 20-carboxy-RvE1, 18-oxo-RvE1, and 10,11-dihydro-RvE1 were each essentially inactive and did not prevent PMN infiltration. Thus, further metabolism of RvE1 toward the last three metabolites belongs to pathways inactivating RvE1.
RvE1 enhances macrophage phagocytosis: a proresolving action
A key and important step in the resolution of acute inflammation is the uptake and clearance of apoptotic PMN by macrophages (1, 2). A recently appreciated proresolving action of LXA4 is its potent ability to stimulate the uptake of apoptotic PMN by macrophages (22, 28) in addition to the now widely documented anti-inflammatory actions of LXA4 (reviewed in Refs. 1, 29, 30, 31, 32). Recently, LXA4 and RvE1 were each shown to stimulate resolution in mice as well as stimulate the uptake and clearance of zymosan by macrophages in vivo (10).
In this study, we determined whether RvE1 can also stimulate isolated macrophages to phagocytize zymosan A. RvE1 proved to be a potent agonist of macrophage phagocytosis (Fig. 6). At concentrations as low as 0.1 nM, RvE1 enhanced the phagocytic uptake of zymosan A (Fig. 6,A). We also investigated the time course of exposure to RvE1 on phagocytosis from 0 to 90 min before zymosan A (Fig. 6,A). The increase in RvE1-stimulated phagocytosis was at 15 min. Consistent with our results from the time course of endogenous production in murine peritonitis (Fig. 3 D), we found that RvE1’s actions diminished with increasing time of exposure, suggesting that RvE1 is rapidly converted to metabolic products by resident peritoneal macrophages during phagocytosis. For comparison, the levels obtained with LXA4 were also examined. RvE1 was approximately twice as active in this system as LXA4 when compared at equal molar concentrations.
Next, we directly compared RvE1 and its metabolic products isolated herein for their ability to stimulate macrophage phagocytosis. As documented in Fig. 6,B (right panel), the 18-oxo-RvE1 metabolite was essentially inactive, whereas a stable analog of RvE1, namely 19-para-fluoro-phenoxy-RvE1, which was designed to resist metabolic inactivation (12), gave potent activity above that obtained with RvE1 with these cells (Fig. 6,B). Of interest, 20-carboxy-RvE1 and 20-hydroxy-RvE1 at 1 nM retained the same ability as RvE1 to stimulate the phagocytosis of zymosan by murine macrophages (results not shown, n = 3). Taken together, these findings indicate RvE1 is subject to several routes of metabolism (Fig. 4) that yield different compounds, some of which are biologically less active than RvE1 whereas others retain function. Thus, selective blockade of specific RvE1 metabolism pathways can enhance RvE1 actions. In this regard, the 19-para-fluoro-phenoxy-RvE1 analog reduces the endogenous conversion by both omega-1 metabolism as well as the dehydrogenation pathways initiated at the carbon 18 position.
Discussion
In the present report, we document the RvE1 metabolome and bioactivities of the newly identified RvE1-derived metabolic products. The omega-3 polyunsaturated fatty acids have emerged as an important approach to reducing the risk of cardiovascular disease (33, 34). Although it is now clear that omega-3 PUFAs have beneficial actions in human clinical trials, particularly those emphasizing cardiovascular risk, the mechanisms by which these beneficial actions occur remained to be established. It is increasingly apparent that inflammation plays a key role in the progression of cardiovascular diseases and many other diseases widely occurring in the Western population. Resolution of acute inflammation and its potential failure may underlie the mechanisms for presenting chronic unresolved inflammation in many diseases (1, 35).
RvE1 is a potent bioactive product generated from EPA and originally identified in resolving murine exudates. RvE1 proved to carry potent stereospecific biological actions in vitro and in complex animal disease models (7, 9). The biosynthesis of RvE1 is initiated by both P450- and aspirin-dependent COX-2-triggered mechanisms. To recapitulate one of several potential routes involved in the in vivo formation of RvE1 in humans, isolated human microvascular endothelial cells treated with aspirin in a hypoxic environment convert EPA to intermediates that are further transformed by human PMN in coincubations to RvE1 (5). In humans, aspirin increases the plasma levels of RvE1 in healthy volunteers also taking EPA supplements (13). Hence, in humans there are both aspirin-dependent and aspirin-independent routes of RvE1 biosynthesis. Along these lines, RvE1 is generated de novo in ocular tissue (8, 36). In addition to aspirin-triggered biosynthesis of 15-epi-LXA4 as a local endogenous anti-inflammatory mediator, recent studies by Birnbaum et al. (37, 38, 39) demonstrate that statins can also initiate the formation of 15-epi-LXA4. Interestingly, combination of EPA and statin in a study of >18,000 patients with a 5-year follow-up demonstrates a significant reduction in coronary events (34). Given the potent anti-inflammatory and proresolving actions documented in the present studies together with earlier in vitro and in vivo results (7, 9), it is likely that RvE1 plays a role in mediating some of the coronary-protective actions noted for EPA supplementation.
From the results of the present studies, at least four separate pathways for further metabolism of RvE1 are present in mammalian tissues (Fig. 4). These pathways appear to be species-, organ-, and cell type-specific. Although 10,11-dihydro-12-oxo-RvE1 could not be directly isolated and identified in the tissues studied to date, it is likely to be an intermediate in a 12-oxo-dehydrogenation-initiated route of RvE1 further metabolism. Given that the product 10,11-dihydro-RvE1 was essentially biologically inactive compared with RvE1, it is possible that this metabolite of RvE1 may serve as an inactive biomarker of RvE1 transient formation in vivo. Along these lines, it is of interest to point out that the 20-hydroxy-RvE1 product of RvE1 made via omega carbon 20 oxidation retains some of the activity of RvE1, namely in vivo anti-inflammatory actions and proresolving actions-accelerating the phagocytic uptake of zymosan. This point demonstrates that not all further metabolites of RvE1 or potentially other resolvins can be assumed to be biologically inactive. Hence, it is possible that further metabolites can retain activity and/or possibly possess new actions, whereas others are clearly pathways of RvE1 inactivation.
Omega-1 hydroxylation to 19-hydroxy RvE1 and reduction of a conjugated double bond to 10,11-dihydro-RvE1 are novel metabolic pathways identified in the present studies that inactivate RvE1. Adding a p-fluorophenoxy to the 19 position of RvE1, as in 19-p-fluorophenoxy-RvE1, blocks this route of RvE1 inactivation. From our earlier report, it is already known that 19-p-fluorophenoxy also blocks another metabolic inactivation route, i.e., dehydrogenation of RvE1 to inactive 18-oxo-RvE1 (12). In summation, the present results demonstrate and provide additional evidence that there is efficient endogenous “machinery” that can quench proresolving signals, such as RvE1, so that the exudates and tissues can return to homeostasis (1, 2). These results also indicate that blocking dehydrogenation of RvE1 and preventing the reduction of its conjugated double bond by modifying the RvE1 structure without attenuating its anti-inflammatory and proresolving activities could be one means to develop RvE1-based therapeutics that can serve as agonists of resolution. Moreover, identification of these further metabolic products in the RvE1 metabolome may be useful in qualifying suitable biomarkers relevant in omega-3 fatty acid supplementation studies as well as monitoring their relation to the biosynthesis and actions of the E-series resolvins.
Acknowledgments
We thank Mary H. Small for assistance with manuscript preparation and Katherine Gotlinger for expert technical assistance.
Disclosures
The lipoxins and resolvins as biotemplates for stable analogs are U.S. patents assigned to Brigham and Women’s Hospital, and Charles N. Serhan is the inventor. These analog patents are licensed for clinical development and are the subject of consultant agreements for Charles N. Serhan.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported in part by National Institutes of Health Grants DK074448 and P-50-DE-016191 (to C.N.S.).
Abbreviations used in this paper: RvE1, resolvin E1, 5S,12R,18R-trihydroxy-6Z,8E,10E,14Z,16E-eicosapentaenoic acid; LC-UV-MS, liquid chromatography-ultraviolet-tandem mass spectrometry; LXA4, lipoxin A4; PMN, polymorphonuclear leukocyte; PUFA, polyunsaturated fatty acid; Resolvin, resolution phase interaction product; 18-oxo-RvE1; 18-oxo-5S,12R-dihydroxy-6Z,8E,10E,14Z,16E-eicosapentaenoic acid; EPA, eicosapentaenoic acid; RP-HPLC, reversed phase-HPLC; LTB4, leukotriene B4.