The novel 13S,14S-epoxy-maresin is converted by human macrophages to maresin 1 (MaR1), inhibits leukotriene A₄ hydrolase (LTA₄H), and shifts macrophage phenotype

Abstract

Maresins are produced by macrophages from docosahexaenoic acid (DHA) and exert potent proresolving and tissue homeostatic actions. Maresin 1 (MaR1; 7R,14S-dihydroxy-docosa-4Z,8E,10E,12Z,16Z,19Z-hexaenoic acid) is the first identified maresin. Here, we investigate formation, stereochemistry, and precursor role of 13,14-epoxy-docosahexaenoic acid, an intermediate in MaR1 biosynthesis. The 14-lipoxygenation of DHA by human macrophage 12-lipoxygenase (hm12-LOX) gave 14-hydro(peroxy)-docosahexaenoic acid (14-HpDHA), as well as several dihydroxy-docosahexaenoic acids, implicating an epoxide intermediate formation by this enzyme. Using a stereo-controlled synthesis, enantiomerically pure 13S,14S-epoxy-docosa-4Z,7Z,9E,11E,16Z,19Z-hexaenoic acid (13S,14S-epoxy-DHA) was prepared, and its stereochemistry was confirmed by NMR spectroscopy. When this 13S,14S-epoxide was incubated with human macrophages, it was converted to MaR1. The synthetic 13S,14S-epoxide inhibited leukotriene B₄ (LTB₄) formation by human leukotriene A₄ hydrolase (LTA₄H) ∼40% (P<0.05) to a similar extent as LTA₄ (∼50%, P<0.05) but was not converted to MaR1 by this enzyme. 13S,14S-epoxy-DHA also reduced (∼60%; P<0.05) arachidonic acid conversion by hm12-LOX and promoted conversion of M1 macrophages to M2 phenotype, which produced more MaR1 from the epoxide than M1. Together, these findings establish the biosynthesis of the 13S,14S-epoxide, its absolute stereochemistry, its precursor role in MaR1 biosynthesis, and its own intrinsic bioactivity. Given its actions and role in MaR1 biosynthesis, this epoxide is now termed 13,14-epoxy-maresin (13,14-eMaR) and exhibits new mechanisms in resolution of inflammation in its ability to inhibit proinflammatory mediator production by LTA₄ hydrolase and to block arachidonate conversion by human 12-LOX rather than merely terminating phagocyte involvement.—Dalli, J., Zhu, M., Vlasenko, N. A., Deng, B., Haeggström, J. Z., Petasis, N. A., Serhan, C. N. The novel 13S,14S-epoxy-maresin is converted by human macrophages to maresin 1 (MaR1), inhibits leukotriene A₄ hydrolase (LTA₄H) and shifts macrophage phenotype.

Tissue injury gives rise to a local inflammatory response that can be amplified by the human leukocytes that may release substances that can inadvertently damage the host (1). When self-limited, inflammation leads to clearance of debris, wound repair, and return to homeostasis via leukocyte resolution traffic. Soluble signals, such as lipid mediators (LMs), primarily those derived from arachidonic acid (AA), eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA), actively participate in regulating cell trafficking and tissue repair at the site of injury during resolution (2). Leukotrienes and prostaglandins derived from AA possess potent chemoattractant and vasodilatory actions contributing to leukocyte recruitment and edema formation (3, 4). The EPA and DHA metabolomes include bioactive mediator E-series and D-series resolvins and protectins that exert potent anti-inflammatory, proresolving, and tissue protective actions (2). Macrophages play key roles in regulating the innate host response to local inflammation. These cells are also central in orchestrating other processes, including neovascularization and wound healing (5–7). A new family of DHA-derived proresolving mediators from macrophages was recently described and named macrophage mediators in resolving inflammation (maresins). Members of this family exert potent phagocyte-directed actions that include inhibition of neutrophil recruitment and stimulation of macrophage efferocytosis (8). The biosynthesis of maresins is initiated in macrophages by the 14-lipoxygenation of DHA producing the 14S-hydro(peroxy)-4Z,7Z,10Z,12E,16Z,19Z-DHA (14S-HpDHA). LM metabololipidomics of murine self-resolving exudates demonstrates that levels of 14S-hydroxy-4Z,7Z,10Z,12E,16Z,19Z-DHA (14S-HDHA), the maresin pathway marker, peak late into the resolution phase, suggesting a role for these mediators in reestablishing tissue homeostasis (8). The maresin pathway is also present in human macrophages (8, 9) and was recently identified in human synovial fluids from patients with arthritis (10).

Maresin 1 (MaR1) is the first member identified from this family of macrophage-derived proresolving mediators. As demonstrated by the identification of alcohol-trapping products, the biosynthesis of MaR1 is proposed to involve a 13,14-epoxide intermediate as the MaR1 precursor. This epoxide intermediate is then proposed to be enzymatically hydrolyzed via an acid-catalyzed nucleophilic attack by water at carbon 7, resulting in the introduction of a hydroxyl group at that position and double-bond rearrangement to form the stereochemistry of bioactive MaR1 (8). The complete stereochemistry of MaR1 proved to be 7R,14S-dihydroxydocosa-4Z,8E,10E,12Z,16Z,19Z-hexaenoic acid, which was shown to possess potent bioactions (11). MaR1 exerts stereospecific leukocyte-directed actions, characteristic of the maresin family. This proresolving mediator also exerts potent cancer-induced antinociceptive and tissue regenerative actions in wound healing in planaria (8). These findings and the biosynthesis of MaR1 in flatworms suggest that maresin structure and function are conserved in evolution. Given the potent actions of MaR1, it was deemed critical to establish the complete stereochemistry and stereospecific conversion of the proposed epoxide intermediate to MaR1. Also, using stereochemically pure material obtained by total organic synthesis, we uncovered novel bioactions of this key epoxy-containing intermediate that regulates leukotriene B₄ production via directly inhibiting leukotriene A₄ (LTA₄) hydrolase and provides novel mechanisms for phagocytes to stimulate the tissues' return to homeostasis.

MATERIALS AND METHODS

RPMI 1640 and DPBS (with and without calcium and magnesium), Ficoll-Histopaque 1077-1, and zymosan A were purchased from Sigma-Aldrich (St. Louis, MO, USA). Human recombinant granulocyte-monocyte colony stimulating factor (GM-CSF) was obtained from R&D Systems (Minneapolis, MN, USA); FCS was obtained from Invitrogen (Grand Island, NY, USA); Liquid chromatography (LC)-grade solvents were purchased from Fisher Scientific (Pittsburgh, PA, USA); Eclipse Plus C18 column (100×4.6 mm×1.8 μm) was obtained from Agilent (Santa Clara, CA, USA); C18 SPE columns were from Waters (Milford, MA, USA); AA and DHA were purchased from Cayman Chemicals (Ann Arbor, MI, USA) along with synthetic standards for LC-tandem mass spectrometry (MS-MS) quantitation and deuterated internal standards [d₈-5S-hydroxy-6E,8Z,11Z,14Z-eicosatetraenoic acid (5S-HETE), and d₄-5S,12R-dihydroxy-6Z,8E,10E,14Z-eicosatetraenoic acid (LTB₄)].

Extraction and LM lipidomics

All samples for LC-MS-MS-based metabololipidomics were extracted with solid-phase extraction columns as previously reported (9). Prior to sample extraction, d₈-5S-HETE and d₄-LTB₄ representing each region in the chromatographic analysis (500 pg each) were added to facilitate quantification. Extracted samples were analyzed by a LC-MS-MS system, QTrap 5500 (AB Sciex, Framingham, MA, USA) equipped with a Shimadzu SIL-20AC autoinjector and LC-20AD binary pump (Shimadzu Corp., Kyoto, Japan). An Agilent Eclipse Plus C18 column (100×4.6 mm×1.8 μm) was used with a gradient of methanol/water/acetic acid of 75:25:0.01 (v/v/v) to 98:2:0.01 at 0.7-ml/min flow rate or a gradient of methanol/water/acetic acid 55:45:0.01 for 2 min, ramped to 85:15:0.01 over 10 min, and to 98:2:0.01 over the next 8 min at a flow rate of 0.4 ml/min. To monitor and quantify the levels of 12-HETE, 14-HDHA, LTB₄, and MaR1, a multiple reaction monitoring (MRM) method was developed with the following signature ion fragments (m/z) for each molecule: 12-HETE, 319-179; 14-HDHA, 343-205; LTB₄, 335-195; MaR1 and related products, 359-221, 359-250, and 359-141. Identification was conducted using published criteria where a minimum of 6 diagnostic ions were employed (12). Calibration curves were determined using synthetic LM mixture containing d₈-5S-HETE, d₄-LTB₄, 12-hydroxy-5Z,8Z,10E,14Z-eicosatetraenoic acid (12-HETE), 14-HDHA, and MaR1. Linear calibration curve for each compound was obtained with r² values ranging from 0.98 to 0.99, and detection limit was ∼1 pg. Quantification was carried out on the basis of the peak areas obtained with multiple reaction monitoring (MRM) transition and the linear calibration curve for each compound as reported previously (9).

Recombinant enzyme incubations

Human recombinant LTA₄ hydrolase (LTA₄H) was prepared (13); this was then incubated (∼10 μg) with LTA₄ or 13S,14S-epoxide (10 μM, 30 min, 37°C, pH 8) in Tris buffer. The incubation was stopped with 2 vol of ice-cold methanol, d₄-LTB₄ was added, and LMs were extracted over C18 columns as detailed above. In designated experiments, LTA₄H (∼10 μg) was incubated with LTA₄, synthetic 13S,14S-epoxy-docosa-4Z,7Z,9E,11E,16Z,19Z-hexaenoic acid [13S,14S-epoxy-DHA; the Z or E stereochemistry for each C=C bond was assigned using 2-dimensional NMR spectroscopy (Varian Medical Systems Inc., Palo Alto, CA, USA), and data were processed and analyzed using MestReNova 7.1.1 software (Mestrelab Research, Santiago de Compostela, Spain); detailed synthesis will be reported elsewhere; unpublished results], or the aqueous hydrolysis products of 13S,14S-epoxy-DHA for 30 min (10 μM, 37°C, pH 8) prior to the addition of 10 μM of LTA₄ and incubation for 30 min (37°C, pH 8). The reaction was stopped by ice-cold methanol, d₄-LTB₄ was added, and products were quantified by LC-MS-MS metabololipidomics following solid-phase extraction.

Human macrophage 12-lipoxygenase (hm12-LOX) was cloned, expressed, and prepared for the present experiments (unpublished results). Briefly, human macrophages were obtained as reported previously (9). 12-LOX cDNA was cloned from these cells and inserted into pET20b vector. This was then subcloned into pFastBac vector, and the recombinant Bacmid containing 12-LOX for insect expression was obtained with Bac-to-Bac Baculovirus Expression System (Life Technologies, Carlsbad, CA, USA). Next, the enzyme was isolated by fast protein LC (FPLC) for incubation with the epoxide. The RNA sequence (Supplemental Table S1) and deduced protein matched that reported previously (14). To determine the production of 13S,14S-epoxy-DHA by hm12-LOX, the enzyme (0.2μM) was incubated with DHA (5 μM, 37°C, pH 8.0) in Tris buffer for 10 min. The reaction was stopped with 2 vol of ice-cold methanol, d₄-LTB₄ was added, and LMs were extracted. To determine whether 13S,14S-epoxy-DHA inactivates 12-LOX, hm12-LOX (1.0 μM) was incubated with 13S,14S-epoxy-DHA (2.5 μM, 30 min, 37°C, pH 8) prior to addition of 50 μM AA or DHA (room temperature, pH 8). After 5 min, the reaction was stopped with ice-cold methanol, d₄-LTB₄ was added, and products were extracted and assessed by LC-MS-MS. For acid-catalyzed methanol trapping of 13S,14S-epoxy-DHA intermediates, hm12-LOX (10 μM) was incubated with DHA (10 μM, 37°C, pH 8, 0.03% Tween 20) for 20 s, and the reaction was stopped with 10 vol of acidified methanol as reported previously (8). Products were extracted and assessed by LC-MS-MS metabololipidomics.

Cell incubations

Human peripheral blood mononuclear cells (PBMCs) were isolated by density-gradient Ficoll-Histopaque isolation (9). Isolated PBMCs were washed 3 times to remove platelets, and monocytes were purified by adhesion to tissue culture plates for 30 min (37°C, DPBS^+/+, pH 7.45). These cells were then cultured for 7 d in RPMI 1640 medium supplemented with 10% FCS and 20 ng/ml GM-CSF. Macrophages (∼10×10⁶ cells) were suspended in PBS^+/+ containing 10% BSA and incubated for 10 min (37°C, pH 7.45) prior to the addition of 13S,14S-epoxy-DHA (∼1.0 μM), serum-treated zymosan (STZ; 0.1 mg), or apoptotic neutrophils (5×10⁷ cells/ml). After 30 min (37°C, pH 7.45), the incubation was stopped with 2 vol ice-cold methanol, deuterated internal standards were added to facilitate quantification and identification, and products were extracted using C18 columns and quantified by LC-MS-MS metabololipidomics, as outlined above. M1 and M2 macrophages were prepared and characterized as reported previously (9). These were then suspended in DPBS (pH 7.45, with calcium and magnesium) containing 100 mg/ml BSA. 13S,14S-epoxy-DHA (2 μM) was then added, and the cells were incubated for 30 min at 37°C. The incubations were stopped with 2 vol of ice-cold methanol, and d₄-LTB₄ was added to facilitate quantification. The products were extracted using C18 columns and MaR1 levels in these incubations as above.

In designated experiments, M1 macrophages were plated in 24-well plates, ∼1 × 10⁵ cells/well, in RPMI supplemented with 0.1% FCS. These were then incubated with 10 nM of 13S,14S-epoxy-DHA, MaR1, resolvin D1 (RvD1), or vehicle [DPBS containing 0.1% ethanol (EtOH)] for 6 h at 37°C. At the end of the incubation, the cells were detached, and surface expression of CD54 (clone HCD54; BD Biosciences, San Jose, CA, USA), CD80 (clone 2D10; BD Biosciences), CD163 (clone GHI/61; BD Biosciences), and CD206 (clone 19.2; BD Biosciences) was determined by flow cytometry as described previously (9).

Statistics

All data are expressed as means ± se. Differences between groups were compared using Student's t test (2 groups) or 1-way ANOVA (multiple groups) followed by post hoc Bonferroni test. The criterion for statistical significance was P < 0.05.

RESULTS

Macrophage LOX produces 13S,14S-epoxy-DHA from DHA

We assessed whether this human 12-LOX in its recombinant form is responsible for the conversion of the 14S-hydro(peroxy) precursor to the 13S,14S-epoxy-DHA intermediate. To this end, we incubated human recombinant hm12-LOX with DHA and assessed the product profile using LC-MS-MS-based metabololipidomics. Chromatographic analysis using MRM demonstrated the presence of dihydroxy products in these incubations (Fig. 1A). We next determined the identity of these DHA-derived products by assessing their respective MS-MS spectra and matching with synthetic and/or authentic standards (11). Several of these products were found to be positional or double-bond isomers of MaR1. These products were denoted as compound I–V, where compound I and II were found to be 7R/S,14S-dihydroxydocosa-4Z,8E,10E,12E,16Z,19Z-hexaenoic acid, compound III was identified as 7S,14S-dihydroxydocosa-4Z,8E,10Z,12E,16Z,19Z-hexaenoic acid (7S,14S-diHDHA; ref. 8), and compound IV and V were identified as 13R/S,14S-dihydroxydocosa-4Z,7Z,9E,11E,16Z,19Z-hexaenoic acid (Fig. 1B). Of note, the major product identified following sodium borohydride reduction of hm12-LOX incubation with DHA was the double oxygenation product 7S,14S-diHDHA, identified earlier (8). These results suggest that hm12-LOX can catalyze the conversion of 14S-hydro(peroxy)-DHA to an epoxide intermediate. Interestingly, this enzyme can also conduct a second oxygenation at carbon position 7 to produce the double oxygenation product 7S,14S-diHDHA.

Figure 1.

Human macrophage LOX produces a novel 13S,14S-epoxide from DHA. A) Selected ion chromatogram (m/z 359-221) obtained from incubation of hm12-LOX (0.2 μM, 37°C, 10 min, pH 8) with DHA (5 μM). B) Incubations were stopped with ice-cold methanol, and products were extracted and assessed by metabololipidomics; MS-MS spectra were employed for the identification of compounds I–V. C) Acid alcohol trapping: hm12-LOX was incubated with DHA (20 s, 10 μM, 37°C) and stopped by addition of 10 vol of acidified (HCl) ice-cold methanol. MS-MS spectrum of the methoxy trapping products. Inset: selected ion chromatogram (m/z 373-297). Results are representative of n = 3 for each incubation condition.

To further assess production of MaR1 epoxide intermediate by human recombinant macrophage LOX, we next sought evidence for the formation of methoxy-trapping products following acid methanol trapping. Metabololipidomics of these incubations gave a single peak at retention time (T_R) 14.1 min. Assessment of the fragmentation pattern of material under this peak in the MS-MS gave characteristic fragmentation of 7R/S-methoxy,14S-hydroxy-containing products that are likely two products that coelute and display m/z 373 = M-H, m/z 355 = M-H-H₂O, m/z 341 = M-H-CH₃OH, m/z 329 = M-H-CO₂, m/z 323 = M-H-H₂O-CH₃OH, m/z 311 = M-H-H₂O-CO₂, m/z 297 = M-H-CH₃OH-CO₂, m/z 297 = M-H-CH₃OH-CO₂-H₂O, m/z 245 = 263-H₂O, m/z 201 = 263-H₂O-CO₂ and m/z 159 = 235-CH₃OH-CO₂ (see Fig. 1C). Hence, this recombinant macrophage LOX can directly convert DHA to the novel epoxide intermediate.

Total organic synthesis of stereochemically pure 13S,14S-epoxy-DHA

To confirm the assigned structure and stereochemistry of this epoxide intermediate and precursor to MaR1, we developed a stereo-controlled synthetic strategy to the 13S,14S-epoxide of DHA. The approach that we used for the synthesis of this chemically labile epoxide carboxylic acid involved a final hydrolysis of the corresponding methyl ester, which was constructed in a stereochemically pure form via 3 key precursors (Fig. 2A). The 13S,14S chirality of the epoxide was obtained via a highly enantioselective epoxidation reaction, while the Z/E geometry for each of the 6 C=C bonds was secured by employing highly stereocontrolled processes. The stereochemical purity of the 13S,14S-epoxide methyl ester was validated via the unambiguous assignment of the Z/E geometry of all C=C bonds by using 2-dimensional correlation spectroscopy (COSY) NMR (Fig. 2B). Following its synthesis and structural confirmation, this methyl ester was hydrolyzed immediately just prior to its addition to cells and enzyme incubations. Note that MaR1 was not obtained in appreciable quantities from any of the aqueous hydrolysis conditions in vitro, namely without cells or isolated enzymes (n=6).

Figure 2.

Total organic synthesis strategy for the production of 13S,14S-epoxy-DHA. A) Synthetic strategy and key precursors used for the preparation of 13S,14S-epoxy-DHA. B) Assignment of the Z or E stereochemistry for each C=C bond using 2-dimensional NMR spectroscopy. The ¹H-¹H gCOSY spectrum of a solution of the epoxide in C₆D₆ (c=9.6×10⁻³ M) shown was acquired on a Varian VNMRS 600 MHz NMR spectrometer at 25°C on a 5-mm triple resonance PFG ¹H, ¹³C, ¹⁵N probe. The data were processed and analyzed on MestReNova 7.1.1 software and referenced to the C₆D₆ as an internal standard. This spectrum depicts all of the connectivities between adjacent alkenyl hydrogens (H₄-H₅, H₇-H₁₂, H₁₆-H₁₇, H₁₉-H₂₀). Colors denote a bitmap plotting method using a rainbow palette that gives depth to the positive and negative contours. The complete identification of each H atom using this method, in combination with its corresponding coupling constants (J values) enabled the unambiguous e/z assignment of all alkenyl hydrogens. The following chemical shifts and coupling constants were recorded: H₉: 6.53 ppm, J = 11.2, 14.8 Hz; H₁₁: 6.38 ppm, J = 11.2, 15.6 Hz; H₁₀: 6.10 ppm, J = 10.8, 14.8 Hz; H₈: 6.01, J = 11.2, 11.2 Hz; H_4, H_5, H_7, H_12, H_16, H_17, H_19, H₂₀: 5.55–5.25 ppm.

MaR1 production from synthetic 13S,14S-epoxy-DHA

We next investigated the conversion of synthetic 13S,14S-epoxy-DHA to MaR1. To this end, this synthetic material was incubated with human macrophages, and products were assessed by metabololipidomics. Chromatographic analysis using MRM gave a major sharp peak with T_R = 5.1 min (Fig. 3A) that demonstrated a characteristic fragmentation pattern: m/z 359 = M-H, m/z 341 = M-H-H₂O, m/z 323 = M-H-2H₂O, m/z 315 = M-H-CO₂, m/z 297 = M-H-H₂O-CO₂, m/z 279 = M-H-2H₂O-CO₂, m/z 161 = 221-H₂O, and m/z 123 = 141-H₂O, corresponding to MaR1 (Fig. 3B). Of note, when the epoxide was incubated with macrophages that had been heated (100°C for 1 h), the preparation no longer converted the epoxide to MaR1. Instead, the product profile corresponded to the nonenzymatic hydrolysis products of the epoxide where compounds I and II were identified as 7R/S,14S-dihydroxydocosa-4Z,8E,10E,12E,16Z,19Z-hexaenoic acid, while compounds IV and V were found to be 13R/S,14S-dihydroxydocosa-4Z,7Z,9E,11E,16Z,19Z-hexaenoic acid (Fig. 3C). These results indicate that the stereochemically pure epoxide obtained by total organic synthesis, 13S,14S-epoxy-DHA, was converted to MaR1 by human macrophages, thus identifying the structure and stereochemistry of the endogenously formed epoxide intermediate. In addition to phagocytizing zymosan, macrophages phagocytizing apoptotic human polymorphonuclear neutrophils (PMNs) also converted this epoxide to MaR1 (n=3). Moreover, this process requires the enzymatic conversion of this intermediate to produce MaR1, which possesses the stereochemistry of 7R,14S-dihydroxydocosa-4Z,8E,10E,12Z,16Z,19Z-hexaenoic acid.

Figure 3.

Human macrophages convert synthetic 13S,14S-epoxy-DHA to MaR1. Human macrophages (1×10⁷ cells) obtained from peripheral blood mononuclear cells were incubated with 13S,14S-epoxy-DHA (1.0 μM) intermediately prior to addition of STZ (0.1 mg; DPBS containing 10% BSA) and incubation for 60 min (37°C, pH 7.45). The incubations were stopped, and LM was assessed by metabololipidomics (see Materials and Methods). A) Selected ion chromatogram (m/z 359-141) depicting MaR1. B) MS-MS spectrum employed for identification of MaR1. C) Selected ion chromatogram (m/z 359-221) obtained following incubation of 13S,14S-epoxy-DHA with cells that were incubated at 100°C (60 min). Results are representative of n = 4 separate cell preparations.

13S,14S-epoxy-DHA inhibits LTB₄ biosynthesis by LTA₄H

We next investigated whether 13S,14S-epoxy-DHA could be a potential substrate for conversion to MaR1 by LTA₄H, since this enzyme is responsible for the stereoselective conversion of the AA-derived epoxide LTA₄ to LTB₄ (3, 13). Metabololipidomic assessment of the product profile obtained following incubation of LTA₄ with LTA₄H gave 2 minor sharp peaks with T_R = 4.1 and 4.3 min that corresponded to the known nonenzymatic hydrolysis products of LTA₄ (3, 4) with the major peak at T_R = 5.0 min corresponding to LTB₄ (Fig. 4A). Chromatographic analysis using MRM of the products produced on incubation of 13S,14S-epoxy-DHA with LTA₄H gave 4 products: peaks I and II were identified as 7R/S,14S-dihydroxydocosa-4Z,8E,10E,12E,16Z,19Z-hexaenoic acid, while peaks III and IV were identified as 13R/S,14S-dihydroxydocosa-4Z,7Z,9E,11E,16Z,19Z-hexaenoic acid (Fig. 4B). These compounds correspond to the nonenzymatic hydrolysis products of the epoxide, thus indicating that LTA₄H did not convert synthetic 13S,14S-epoxy-DHA to MaR1 and is not responsible for MaR1 biosynthesis.

Figure 4.

13S,14S-epoxy-MaR inhibits LTB₄ production by human recombinant LTA₄H. Human LTA₄H was incubated with LTA₄ (10 μM), 13S,14S-epoxy-DHA (10 μM), or the aqueous hydrolysis products of 13S,14S-epoxy-DHA (10 μM; 30 min, 37°C, pH 8.0). The incubation was either stopped by addition of 2 vol ice-cold methanol or 10 μM of LTA₄ was added, and the enzymes were incubated for 30 min (37°C, pH 8.0). The incubations were stopped, and LMs were obtained by solid-phase extraction (see Materials and Methods). A, B) Selected ion chromatograms for m/z 335-195 (A) and m/z 359-221 (B). Results are representative of n = 4 separate incubations. C, D) Quantification (C) and percentage inhibition (D) of LTB₄ formation in the incubations by metabololipidomics. Values are expressed as means ± se; n = 4. **P < 0.01 vs. LTA₄H + LTA₄.

Since LTA₄ is known to inactivate LTA₄H, preventing the conversion of LTA₄ to LTB₄ (15), we next investigated whether 13S,14S-epoxy-DHA also exerted inhibitory actions on LTA₄H. Incubation of LTA₄ with LTA₄H led to an ∼50% inhibition in LTB₄ formation (Fig. 4C, D), as determined by LC-MS-MS-based metabololipidomics. Incubation of the 13S,14S-epoxide with LTA₄H also led to a significant reduction in LTB₄ formation (∼40%), while incubation of equimolar amounts of the aqueous hydrolysis products of 13S,14S-epoxide did not significantly alter LTB₄ formation by LTA₄H (Fig. 4C, D). Incubation of LTA₄H with LTA₄ and 13S,14S-epoxy-DHA did not give a statistically significant increase in LTB₄ formation by this enzyme. These results indicate that 13S,14S-epoxy-DHA inhibits LTA₄H to a similar extent as LTA₄ itself and that the actions of each of these products on LTA₄H were not synergistic.

13S,14S-epoxy-DHA inhibits lipoxygenation of AA by human 12-LOX

We next assessed whether this 13S,14S-epoxide could also regulate AA conversion by hm12-LOX. Incubation of hm12-LOX with 13S,14S-epoxy-DHA prior to addition of AA led to a significant reduction (∼60%) in 12-HETE formation, as determined by LC-MS-MS-based metabololipidomics (Fig. 5). In sharp contrast, when the 12-LOX substrate was DHA, prior exposure of the enzyme to the 13S,14S-epoxide did not lead to a significant reduction in DHA conversion to 14-HDHA (Fig. 5A). Together, these results indicate that the 13S,14S-epoxy-DHA selectively inhibits conversion of AA, but not that of DHA, by hm12-LOX.

Figure 5.

13S,14S-epoxy-DHA selectively inhibits hm12-LOX AA lipoxygenation. 13S,14S-epoxy-DHA (2.5 μM) or vehicle (4% EtOH) was added to recombinant 12-LOX (1 μM) and incubated for 30 min (37°C, pH 8). AA or DHA (50 μM) was then added to the incubation mixture, the reaction was stopped after 5 min (room temperature), and products were assessed by metabololipidomics. Results are expressed as means ± se of 3 separate incubations. *P < 0.05 vs. corresponding control incubations.

Macrophage phenotypes

We next investigated whether incubation of 13S,14S-epoxy-DHA with M1 macrophage led to a change in macrophage phenotype toward an M2 profile. Incubation of M1 macrophages with either 13S,14S-epoxy-DHA (10 nM) or MaR1 (10 nM) led to significant reductions in CD54 and CD80 expression and a concomitant up-regulation of CD163 and CD206 (Fig. 6A--D). For comparison, this regulation was also found with RvD1 (10 nM).

Figure 6.

13S,14S-epoxy DHA displays proresolving properties and is more readily converted by M2 macrophages. A–D) M1 macrophages were incubated with vehicle (Veh; PBS containing 0.1% EtOH), 13S,14S-epoxy-DHA, MaR1, or RvD1 (10 nM, pH 7.45, RPMI 0.1% FCS, 37°C) for 6 h. Expression of CD54 (A), CD80 (B), CD163 (C), and CD 206 (D) on the surface of these cells was assessed by flow cytometry using fluorescently conjugated antibodies. E) 13S,14S-epoxy-DHA (2 μM) was incubated with M1 or M2 macrophages (40×10⁶ cells/ml, DPBS^+/+, pH 7.45, 100 mg/ml BSA, 30 min, 37°C). Incubation was stopped with ice-cold methanol, and products were assessed by LM metabololipidomics. Results for A are means ± se; n = 3 separate cell preparations. Results for B–E are representative of n = 3 for each incubation condition.

We also investigated the conversion of 13S,14S-epoxy-DHA to MaR1 by different human macrophage subtypes. Macrophage phenotype was determined using flow cytometry, where M1 macrophages expressed higher CD80 and CD54 levels, while M2 macrophages displayed higher CD163 and CD206 levels. Incubation of 13S,14S-epoxy-DHA (2 μM) with M2 macrophages gave higher MaR1 levels then when the 13S,14S-epoxy-DHA was incubated with M1 macrophages, as determined by LM metabololipidomics (Fig. 6).

Hence, these findings identify novel regulatory actions of DHA-derived products within the eicosanoid cascade and macrophage phenotypes (see Fig. 7).

Figure 7.

Biosynthesis and actions of 13S,14S-epoxy-DHA and its role in MaR1 production. See text for details.

DISCUSSION

In the present report, we demonstrate production of a novel 13S,14S-epoxy-DHA by recombinant human macrophage LOX. Using material obtained by stereocontrolled organic synthesis in combination with NMR spectroscopy and LM metabololipidomics, we establish the complete stereochemistry of this 13S,14S–epoxy-DHA and demonstrate its precursor role in conversion to MaR1 by human macrophages. Also, incubation of this epoxide intermediate with isolated human recombinant LTA₄H inhibited LTB₄ production from LTA₄. Moreover, incubation of human macrophage LOX with 13S,14S-epoxy-DHA selectively inhibited AA conversion to 12-HpETE. Together, these findings demonstrate novel bioactions within the maresin pathway in controlling proinflammatory LM (i.e., LTB₄), as well as establish the stereochemistry of the epoxide intermediate in MaR1 biosynthesis.

Macrophages are key regulators of the inflammatory response with distinct macrophage subtypes linked with the propagation vs. resolution of inflammation (16). In this context, there are two broad categories: M1 macrophages or classical macrophages are considered as proinflammatory, while M2 macrophages are linked with the reestablishment of homeostasis, wound healing, and tissue regeneration (5, 7, 16). A third macrophage subtype, referred to as resolution-phase macrophages, was identified in resolving exudates, with these macrophages possessing characteristics of both M1 and M2 cells (5, 6), and each human subpopulation expresses 12-LOX. Recent results indicate that DHA and DHA-derived proresolving mediators, including RvD1, can stimulate a switch in macrophage phenotype from proinflammatory to a proresolving M2-like phenotype (17, 18). Using LM metabololipidomics, we recently found that human M2 macrophages are associated with higher MaR1 levels (9), a finding that is in line with the homeostatic and tissue regenerative actions of this proresolving mediator (8, 11). In the present study, we demonstrate that this is at least in part a result of enhanced ability of this macrophage subtype to convert the 13S,14S-epoxide intermediate to MaR1. Also, LM-metabololipidomics of human synovial fluids from patients with arthritis demonstrated that MaR1, along with other proresolving mediators, including LXA₄ and RvD5, are found within these synovial fluids at levels within their bioactive ranges (10). Together, these findings underscore a role for macrophage-derived proresolving mediators in potentially preventing further tissue damage and/or organ fibrosis in disease (10, 19).

LM biosynthesis involves multiple enzyme-regulated steps that determine the specific stereochemistry of each member of these local mediator families (2, 3). The absolute stereochemistry of proresolving LM, in turn, dictates their bioactivity and, hence, structure-activity relations. In this context, subtle differences in the stereochemistry of either resolvins or protectins, for example, result in changes in their potency (2). To establish the biosynthetic pathways for LM biosynthesis, it was deemed important to determine the presence and stereochemistry of their intermediates, a goal of the present experiments. In the biosynthesis of MaR1, DHA undergoes 14S-lipoxygenation to give 14-HpDHA that is converted to 13S,14S-epoxy-DHA, a proposed intermediate that is then further transformed to MaR1 (8). To this end, in the present report, we found that this 13,14-epoxide is produced by isolated hm12-LOX, as demonstrated by the identification of the aqueous hydrolysis products of 13S,14S-epoxy-DHA along with the corresponding acid methanol-trapping products from DHA (Fig. 1). Of note, in these incubations, we also identified the double dioxygenation isomer of MaR1, 7S,14S-diHDHA (cf. refs. 8, 11), suggesting that this macrophage 12-LOX also oxygenates DHA at C7. Although this double dioxygenation product, 7S,14S-diHDHA, retains some bioactivity (8, 11), it displays significantly lower anti-inflammatory and proresolving actions than MaR1. Along these lines, the 14-HpDHA intermediate can also undergo another double dioxygenation at the penultimate carbon (ω-1) to give the 14S,21R-dihydroxydocosa-4Z,7Z,10Z,12E,16Z,19Z-hexaenoic acid (20). This maresin-related pathway product displays tissue-protective actions promoting wound healing in diabetes models and following renal ischemia reperfusion injury by restoring mesenchymal stem cell function (21, 22).

LM biosynthesis occurs at sites of inflammation and tissue injury, whereby the intermediates are either rapidly transformed to bioactive mediators via stereospecific enzyme-mediated conversion or nonenzymatically hydrolyzed to virtually inactive products (2–4). Therefore, tissue levels of these intermediates are transient and, hence, do not reach quantities that can be isolated for direct stereochemical determination by NMR spectroscopy (2). Hence, in the present studies we have used direct matching of material obtained by total organic synthesis (which allows for the generation of compounds with known absolute stereochemistry) with biological materials. This approach permitted both the scale-up and confirmation of the potent anti-inflammatory and proresolving actions of resolvins, protectins, and MaR1 (reviewed in refs. 2, 11), as well as their unambiguous identification in biological tissues via targeted lipidomics (LC-MS-MS-based). The ability to identify resolvins revealed their remarkable protective actions, including their ability to enhance immune vigilance (23) and to stimulate host immune responses, lowering antibiotic requirements during infection (24).

Using total organic synthesis in the present report, we established the absolute stereochemistry of the novel 13S,14S-epoxy-containing intermediate, which proved to be 13S,14S-epoxy-DHA (Fig. 2). When incubated with activated human macrophages, this synthetic compound was readily converted to MaR1 (Fig. 3). In contrast, incubation of this synthetic epoxide with heat-inactivated cells gave no conversion to MaR1, since the epoxide was instead converted to its corresponding nonenzymatic hydrolysis products (Fig. 3). Hence, in the proposed biosynthesis of MaR1, the LOX first abstracts a hydrogen from the methylene group at C12 and inserts molecular oxygen at C14 to yield 14S-HpDHA. In macrophages lacking the murine homologue of this enzyme (12/15-LOX), the conversion of DHA to 14-HpDHA was greatly reduced (>95%), indicating the role of the murine ortholog of 12-LOX in initiating MaR1 biosynthesis in macrophages (8). This HpDHA is then converted to the 13S,14S-epoxide by the same enzyme following a second hydrogen abstraction from the methylene group at C9. Our present results indicate that in addition to catalyzing the 14-lipoxygenation of DHA, this same enzyme is also an epoxidase producing the novel maresin epoxide intermediate (see scheme in Fig. 6). This type of reaction has been demonstrated for 5-LOX, which converts AA to 5-HpETE and then to LTA₄ (25). The conversion of the 13S,14S-epoxide to MaR1 then proceeds via an enzyme-mediated hydrolysis that converts the double-bond geometry at the C8 position from Z to E and at the C12 position from Z to E, giving a final geometry around the triene region of MaR1 corresponding to 8E,10E,12Z (see Fig. 6), as well as directs the insertion of a hydroxyl group (from H₂O) at C7 (8). Thus, the epoxide intermediate to MaR1 requires an enzyme-mediated hydrolysis and Z/E reconfiguration in order to give the conversion of the double-bond geometry at Δ¹² from trans to cis orientation, giving a final geometry of MaR1. In the absence of this enzyme, the 13S,14S-epoxide intermediate decays to aqueous hydrolysis products 7R/S,14S-dihydroxydocosa-4Z,8E,10E,12E,16Z,19Z-hexaenoic acid and 13R/S,14S-dihydroxydocosa-4Z,7Z,9E,11E,16Z,19Z-hexaenoic acid (see Fig. 6).

Since LTA₄H catalyzes LTA₄ conversion to LTB₄ (13) and is involved in the conversion of EPA to RvE1 in the E-series resolvins (26), we assessed conversion of 13S,14S-epoxy-DHA with LTA₄H, which did not result in MaR1 production (Fig. 4). These results rule out a role for LTA₄H in converting the 13S,14S-epoxide intermediate to MaR1 (Fig. 3). Leukocytes and, in particular, phagocytes are key cellular players regulating the onset and resolution of inflammation (27). MaR1 is produced by macrophages during inflammation resolution and exerts phagocyte-directed actions (8, 11). When administered in vivo, MaR1 at concentrations as low as 0.1 ng/mouse significantly reduced PMN accumulation in response to sterile injury (8, 11). MaR1 is also a potent prophagocytic stimulus enhancing human macrophage phagocytosis, as well as efferocytosis of apoptotic cells (8, 11). Apart from leukocyte-directed actions, this proresolving mediator also exerts tissue-directed actions, whereby administration of MaR1 to surgically injured planaria accelerates tissue regeneration at the site of injury (11). These findings emphasize the role of MaR1 in regulating both local inflammatory responses and potentially stem cell functions. One mechanism by which MaR1 might regulate stem cell responses is by promoting stem cell differentiation, as recently described for another proresolving mediator, protectin D1 (28). MaR1 also displays potent analgesic actions controlling local inflammation resolution and associated inflammatory pain, as well as neuropathic pain, via blocking TRPV1-mediated responses (11). Of interest, our present results indicate that the maresin pathway can also affect inflammation resolution by selectively reducing LTB₄ production via direct inhibition of LTA₄H (Fig. 4). This is a regulatory mechanism that the 13S,14S-epoxide intermediate shares with LTA₄ (29) and that would be operative in the conversion of DHA to MaR1. The ability of the 13S,14S-epoxy-DHA to directly inactivate LTA₄H is not shared with MaR1; therefore, conversion of the epoxide to MaR1 could result in a reduction of its LTA₄H inhibition. Because MaR1 possesses potent anti-inflammatory properties, this proresolving mediator may still regulate LTA₄H activity by other means yet to be identified. We also found a novel regulatory mechanism in eicosanoid biosynthesis, whereby this 13S,14S-epoxy-maresin selectively inhibited AA conversion by 12-LOX (Fig. 5). Of interest, both MaR1 and 13S-14S-epoxy-DHA regulate macrophage phenotype, skewing the macrophage profile toward a homeostatic and tissue-protective phenotype (Fig. 6). Together, these findings point to novel anti-inflammatory mechanisms by the MaR1 biosynthetic pathway, and utilization of DHA can regulate proinflammatory eicosanoids via direct enzyme interactions with the potent autacoid actions of MaR1 and related products themselves, as well as changing the macrophage phenotype.

In summary, the present findings demonstrate that hm12-LOX serves as a DHA 14-lipoxygenase to catalyze the formation of 13S,14S-epoxy-DHA. We also established the complete stereochemistry of this novel 13S,14S-epoxy-DHA and its conversion to MaR1 by human macrophages. This epoxide displayed potent direct enzyme regulatory actions inhibiting proinflammatory LTB₄ biosynthesis and shifting enzyme substrate utilization toward 14-lipoxygenation of DHA. It is conceivable that these reactions can provide tissues with a feed-forward loop in the biosynthesis of maresins. It is now well appreciated that ω-3 fatty acids exert a positive bearing on several important aspects of cardiovascular disease (30, 31). Thus, these findings offer novel direct and stereoselective mechanisms for action of the MaR pathway in the DHA metabolome and specifically the maresin pathway in resolution, tissue homeostasis, and regeneration.