Resolvin E1 Selectively Interacts with Leukotriene B₄ Receptor BLT1 and ChemR23 to Regulate Inflammation¹

Inflammation and resolution are major mechanisms involved in many human diseases, including cardiovascular disease, arthritis, diabetes, asthma, Alzheimer’s disease, and periodontitis (1). Most inflammatory challenges are self-limited in healthy subjects, implicating the existence of endogenous circuits for anti-inflammation and proresolution mediators that are operative during the temporal events of host defense and inflammation (2). Resolution of inflammation is an active process governed by timely and spatially regulated formation and actions of local mediators so that tissues can return to homeostasis (3, 4, 5).

Resolvins (Rv)⁴ and protectins are local mediators derived from omega-3 polyunsaturated fatty acids such as eicosapentaenoic acid (EPA) and docosahexaenoic acid that are generated during the spontaneous resolution phase and act locally at sites of inflammation. These recently uncovered pathways and mediators counterregulate polymorphonuclear leukocyte (PMN) infiltration and promote resolution (2, 3). They are generated during multicellular responses such as inflammation and microbial infections, a unique pathway that involves cell-cell interactions and transcellular biosynthetic routes. When aspirin is present, resolvin E1 (RvE1) is formed from EPA via cell-cell interactions involving cells bearing cyclooxygenase-2 that has been acetylated by aspirin treatment and cells that possess 5-lipoxygenase (5-LO) (5, 6). RvE1 biosynthesis can also be initiated by microbial cytochrome P450 monooxygenase in an aspirin-independent manner (7), which can contribute to its production in vivo. These newly produced resolvins may be responsible for some of the beneficial effects of taking omega-3 EPA that are enhanced with aspirin therapy (8, 9).

PMNs are one of the main cellular targets of RvE1’s anti-inflammatory actions. For example, RvE1 reduces PMN transendothelial migration and release of superoxide generation in the nanomolar concentration range (5, 10). These actions are also demonstrable in vivo, where RvE1 blocks PMN infiltration both in peritonitis (6) and in inflamed colon tissue during colitis (8). RvE1 also attenuates APC functions targeting dendritic cell migration and reducing IL-12 production via ChemR23 (6). ChemR23 was identified from a panel of G-protein-coupled receptors (GPCRs) using a counterregulatory screening system. We reasoned that proresolving ligand activation of GPCR would block proinflammatory signals such as TNF-α-driven NF-κB activation. From our GPCR panel described in Ref. 6 , only RvE1 selectively bound and activated ChemR23 to block TNF-α signaling in a stereoselective fashion.

In this study, we report that RvE1 specifically binds the leukotriene B₄ (LTB₄) receptor BLT1 on human PMN. RvE1 interacts with BLT1 as a partial agonist serving as a local damper of LTB₄-BLT1 signals on PMN. These RvE1-BLT1 interactions join RvE1-ChemR23-mediated counterregulatory actions to promote the resolution of acute inflammation.

RvE1 (5S,12R,18R-trihydroxy-eicosa-6Z,8E,10E,14Z,16E-pentaenoic acid) was prepared by total organic synthesis as described in Ref. 6 . LTB₄ and U-75302 (6-(6-(3R-hydroxy-1E,5Z-undecadien-1-yl)-2-pyridinyl)-1,5S-hexanediol) was obtained from Cayman Chemical. Forskolin, Ro-20-1724, and pertussis toxin (PTX) were purchased from Sigma-Aldrich. Fura 2-AM was obtained from Invitrogen Life Technologies. The expression vectors pcDNA3-hBLT1, hBLT2, and hChemR23 were constructed as described in a recent publication (6).

BLT1-deficient mice were described earlier (11). C57BL/6 strain from Charles River Laboratories was used as the wild-type (WT) control strain. Female 8- to 10-wk-old mice were used. All procedures were reviewed and approved by the Harvard Medical School Standing Committee on Animals (protocol 02570).

Human PMN and PBMC were freshly isolated from venous blood of healthy volunteers (that declined taking medication for 2 wk before donation; Brigham & Women’s Hospital protocol 88-02642) by Ficoll gradient as described in Ref. 5 .

Chinese hamster ovary (CHO), HEK293, and HeLa cells were cultured in Ham’s F-12 and DMEM, respectively, supplemented with 10% FBS. HEK293 cells stably expressing human BLT1 or BLT2 were established by transfecting pcDNA3-hBLT1 or pcDNA3-hBLT2, selected and maintained with 500 μg/ml G418. CHO cells stably expressing human ChemR23 (CHO-hChemR23) were prepared (6) and maintained in the presence of 500 μg/ml G418.

Binding studies were conducted with tritiated RvE1 (6,7,14,15-[³H]RvE1; 100 Ci/mmol) synthesized as described in Ref. 6 using custom tritiation (American Radiolabeled Chemicals) of acetylenic RvE1 followed by HPLC isolation. The binding mixture (100 μl) contained isolated membrane fractions (10 μg protein) and indicated concentrations of [³H]RvE1 with or without unlabeled competitors in binding buffer (50 mM HEPES, 1 mM CaCl₂, and 5 mM MgCl₂) for 1 h at 4°C. For CHO-hChemR23, cells (1 × 10⁶) were incubated in Dulbecco’s PBS with CaCl₂ and MgCl₂ for 1 h at 4°C. For determination of nonspecific binding, at least 1,000 × concentration of unlabeled RvE1 was used. The bound and unbound radioligands were separated by filtration through Whatman GF/C glass microfiber filters and radioactivity was determined. Scatchard plot was obtained and the K_d value was calculated using Prism (GraphPad).

Cells were seeded on 24-well plates (7.5 × 10⁵ cells/well) and cultured for another 48 h. The medium was replaced with 400 μl of DMEM containing 5 μM forskolin, 100 μM Ro-20-1724 (4-(3-butoxy-4-methoxybenzyl)-2-imidazolidinone), and test compounds, and after 15 min of incubation, the reaction was terminated by replacing media with 100 μl of 0.1 N HCl. The cAMP contents in 20 μl of the lysates were determined using the cAMP enzyme immunoassay kit (Cayman Chemical).

Mobilization of calcium was measured in fura 2-AM-loaded cells. Human PBMC (1.0 × 10⁷ cells/ml) were incubated in HBSS with fura 2-AM (5 μM; <0.05% v/v DMSO) for 30 min at 37°C. Cells were washed twice and suspended (12.5–15 × 10⁶ cells/ml) in HBSS supplemented with Ca²⁺ (1.6 mM). The cell suspension was maintained in a stirred thermostatted cuvette and fluorescence was monitored using 335-nm excitation and 505-nm emission in a spectrofluorometer (PerkinElmer; LS-3B). RvE1 and LTB₄ were added in a volume of 20–40 μl to cuvettes to give the indicated concentrations.

HeLa cells were maintained in DMEM supplemented with 10% heat-inactivated FBS. Cells (5 × 10⁵ cells/well) in 24-well plates were transfected with 100 ng of pNF-κB luciferase (Stratagene), 800 ng of either pcDNA3 or pcDNA3-BLT1, and the internal standard pRL-TK (Promega) using Superfect transfection reagent (Qiagen). After 24 h, cells were exposed to test compounds for 6 h in serum-free DMEM. Luciferase activity was measured by the Dual-Luciferase reporter system (Promega).

Murine peritonitis was conducted as described in Ref. 4 , and 100 ng or 1 μg/mouse RvE1 was injected into the tail vein and followed by 1 ml of zymosan A (1 mg/ml) into the peritoneum. Peritoneal lavages were collected at 2 h and cells were enumerated. For differential leukocyte counts, 100 μl of the lavaged cells was added to 100 μl of 30% BSA and centrifuged onto microscope slides at 2200 rpm for 4 min using a Cytofuge (StatSpin). The slides were air dried, and cells were visualized using Wright-Giemsa stain and light microscopy.

[³H]RvE1 specifically binds to the isolated membrane fraction of human PMN (Fig. 1). At 4°C, RvE1 binding to human PMN membrane fraction was specific with a K_d of 48.3 nM. We next examined the competition of [³H]RvE1 binding with several related eicosanoids. Among the compounds tested, [³H]RvE1-specific binding to human PMN was displaced by the homoligand RvE1 (K_i = 34.3 nM), LTB₄ (K_i = 0.08 nM), and LTB₄ receptor 1 (BLT1) selective antagonist U-75302 (K_i = 1.5 nM) (12), but not by the chemerin peptide (13); a ligand specific for another RvE1 receptor denoted ChemR23 (Fig. 2,A). For direct comparison, [³H]RvE1 to human ChemR23 was competed with RvE1 (K_i = 330 nM) or chemerin peptide (K_i = 429 nM), but not with LTB₄ (Fig. 2 B). These results demonstrate that the RvE1 binding site on human PMN is pharmacologically distinct from ChemR23.

We next determined whether RvE1 binds to recombinant BLT1. HEK293 cells stably expressing human BLT1 (HEK-hBLT1) were prepared to determine RvE1-specific binding. As shown in Fig. 3,A, RvE1 specifically bound to HEK-hBLT1 with a K_d of ∼45 nM. RvE1-specific binding was competed with LTB₄ (K_i = 3 nM) or RvE1 (K_i = 70 nM) (Fig. 3 B). Recombinant human LTB₄ receptor type 2 (BLT2) did not display specific binding for [³H]RvE1 at concentrations up to 10 nM (data not shown). These results clearly demonstrated that RvE1 binds to BLT1 on human PMN.

To examine whether the binding of RvE1 to BLT1 transduces functional responses, such as intracellular signaling, we determined adenylyl cyclase activity by measuring the cAMP accumulation. As shown in Fig. 4,A, LTB₄ inhibited 5 μM forskolin-activated adenylate cyclase activities in HEK-hBLT1 cells with EC₅₀ of 0.015 nM, a value consistent with that reported previously (14). In these experiments, RvE1 inhibited adenylate cyclase activities with EC₅₀ of 3.2 nM. On the other hand, RvE1 gave no response with HEK-hBLT2 cells where LTB₄ gave signals (EC₅₀ = 11.5 nM), a value consistent with the previous report (12) (Fig. 4 B). These results indicate that RvE1 selectively binds and activates BLT1 to transmit intracellular signals.

Fig. 5 reports the increase in intracellular calcium mobilization stimulated by LTB₄ or RvE1 in human PBMC. RvE1 at 100 nM increased intracellular calcium, but the maximum intensity of signal was only about one-third that of LTB₄ (Fig. 5, A and B). Importantly, prior exposure to RvE1 completely blocked LTB₄-induced calcium response in a concentration-dependent manner (Fig. 5 C). These results indicate that RvE1 is a partial agonist to attenuate LTB₄-induced calcium responses in leukocytes.

Because LTB₄ induces proinflammatory cytokine and chemokine expression by activating NF-κB (15), we next determined whether RvE1 could modulate LTB₄-BLT1-induced NF-κB activation by using a luciferase reporter gene system. LTB₄ induced NF-κB activation in HeLa cells transfected with human BLT1 with EC₅₀ of 0.36 nM (Fig. 6,A). This induction was almost completely abolished by 100 ng/ml PTX, suggesting that BLT1 coupled to PTX-sensitive G proteins such as Gi/o to activate NF-κB transcription factor (Fig. 6,A, inset). RvE1 concentrations above 100 nM gave partial activations of NF-κB that were in a range comparable to that of the BLT1 antagonist U-75302. Importantly, LTB₄-dependent NF-κB activation was blocked by ∼40–50% with RvE1 as low as 1 nM, which is an ∼10-fold molar excess of LTB₄ (Fig. 6 B). These results clearly demonstrated that RvE1 binds to BLT1 and attenuates LTB₄-induced proinflammatory signals as a nonphlogistic ligand.

We examined the role of BLT1 in the in vivo actions of RvE1 in regulating leukocyte infiltration in zymosan A-induced peritonitis. In this model, zymosan activates resident macrophages and/or early accumulating neutrophils to produce LTB₄ from arachidonic acid, resulting in PMN influx in an autocrine loop of PMN accumulation during acute inflammation (11). Intravenous administration of RvE1 at 100 ng dramatically blocked PMN infiltration by 38% at 2 h after zymosan injection in WT mice (Fig. 7,A). In contrast, RvE1 at 100 ng did not give statistically significant reduction in mice with targeted disruption of BLT1 (BLT1^−/−) (Fig. 7,A). The early reduction of PMN influx in BLT1^−/− mice, which represents BLT1-dependent PMN infiltrations, was noted as reported previously (11). RvE1 reduced the numbers of infiltrating PMN in WT mice to the levels of that obtained in BLT1^−/− mice (Fig. 7,A), indicating that i.v. administration of RvE1 at 100 ng blocked PMN migration by dampening BLT1 functions in vivo. Of interest, RvE1 at higher doses (1.0 μg i.v.) retained most of its antimigratory actions in BLT1^−/− mice, giving around 35% inhibition of PMN infiltration in both WT and BLT1^−/− mice (Fig. 7 B). Together, these results demonstrate that i.v.-administered RvE1 acts via BLT1, potentially serving as a local damper of BLT1 signals on leukocytes in addition to BLT1-independent mechanisms involving additional receptor(s) in vivo.

The omega-3 EPA-derived RvE1 is an endogenous mediator that protects tissues from leukocyte-mediated injuries (5, 6, 7 ; for recent review see Ref. 2). In the present study, we characterized RvE1 binding to human PMNs and identified these sites as the LTB₄ receptor BLT1. RvE1 selectively interacts with BLT1, but not with its closely related receptor, termed BLT2. BLT2 is structurally similar to BLT1 with ∼45% identity in deduced amino acid sequence (12). RvE1 effectively blocked LTB₄-induced calcium mobilization in human leukocytes, and also attenuated LTB₄-induced NF-κB activation in BLT1-transfected cells. In vivo anti-inflammatory action of RvE1 was sharply reduced in BLT1 knockout mice when given at low doses in zymosan-induced peritonitis, but RvE1 at high doses gave significant inhibition of PMN infiltration in a BLT1-independent manner. Therefore, RvE1 functions as a local damper of LTB₄-BLT1 signals on leukocytes in addition to other receptor(s)-mediated actions.

LTB₄ is an arachidonic acid-derived mediator produced mainly by activated leukocytes. LTB₄ is a potent chemoattractant for PMNs, eosinophils, and macrophages and also activates the respiratory burst and granule release from PMNs (16, 17). BLT1 is a high-affinity LTB₄ receptor responsible for the chemotactic actions of LTB₄ (14). In zymosan peritonitis, zymosan activates resident macrophages and/or early accumulating neutrophils that produce LTB₄ from arachidonic acid, resulting in BLT1-dependent PMN influx and PMN accumulation during acute inflammation (11). Along these lines, when we monitored exudate eicosanoids in zymosan-initiated peritonitis during a 72 h time course after zymosan injection, the maximal levels of LTB₄ were at 2 h, and subsequently subsided during the first 24 h (see Ref. 4). LTB₄ is also a strong chemoattractant for T cells, creating a functional link between innate and adaptive immune responses (18). RvE1 within a low nanomolar range was shown to block PMN transendothelial migration initiated by LTB₄ or fMLP (5), and in this study we report that RvE1 directly interacts with BLT1 attenuating the propagation of proinflammatory signals by LTB₄ with similar potencies to the BLT1 antagonist U-75302. This site of action may be particularly important in self-limiting PMN diapedesis, since RvE1 is locally generated from omega-3 EPA when activated PMN interacts with vascular endothelial cells (5, 6). ChemR23 was identified as an RvE1 receptor displaying specific binding with a K_d value of 11.3 nM, phosphorylation signals, G protein activation, and attenuation of TNF-α activated NF-κB (6). ChemR23 is abundantly expressed in macrophages and dendritic cells (DC), but apparently not as much in PMN (13). RvE1 attenuates APC functions targeting DC migration and reducing IL-12 production, and down-regulation of ChemR23 expression by small interference RNA abolished the effect of RvE1 on DCs (6). Taken together, the potent anti-inflammatory and/or proresolution actions of RvE1 in vivo could be mediated by two different sites of actions, namely attenuating LTB₄-BLT1 signals to stop leukocyte infiltration and activation as well as stimulating ChemR23 to regulate migration and cytokine production of macrophages and/or DC. These multiple sites of action are also the case with lipoxin (LX)A₄, serving as a local damper of both vascular LTD₄ receptor (CysLT1) signals as well as LXA₄ receptor (ALX)-regulated PMN traffic (19). LXA₄ was as potent as RvE1 in terms of magnitude of inhibition (40–50%) and effective concentration ranges (10–100 ng/mouse) in zymosan peritonitis (20). LXA₄ shares the same receptor denoted ALX with anti-inflammatory peptide annexin 1 (21), and annexin 1 gives the same magnitude of reduction of PMN migration (40–50%) in zymosan peritonitis at higher concentration ranges (10–100 μg) (22). Also, overexpression of human ALX in mouse leukocytes dramatically reduced acute PMN infiltration in ALX transgenic mice (20), suggesting ALX as a counterregulatory receptor whose ligands work as agonists of endogenous anti-inflammation.

Results from studies using mouse models and antagonists of LTB₄ suggested a role for BLT1 in chronic diseases such as rheumatoid arthritis, asthma, and cardiovascular disease (18, 23, 24, 25). Increased expression of the genes of the leukotriene pathway such as 5-LO, 5-LO-activating protein, and LTA₄ hydrolase in atherosclerotic plaques are observed and draw particular attentions (26). Results of genetic studies revealed variants of the 5-LO gene promoter, 5-LO-activating protein, and LTA₄ hydrolase genes as risk factors in human atherosclerosis and myocardial infarction (27, 28, 29). LTB₄ is produced in human atherosclerotic lesion (26), activates NF-κB transcription factor through BLT1, and induces atherogenic chemokine MCP-1 in human monocytes (15), smooth muscle cells (30), and endothelial cells (31). Of special interest, omega-3 polyunsaturated fatty acids such as EPA and docosahexaenoic acid are widely thought to be cardioprotective (9, 32, 33). The present results demonstrated that RvE1 derived from omega-3 EPA-attenuated LTB₄ induced NF-κB activation via BLT1, suggesting the therapeutic potential of RvE1 in inflammatory diseases where LTB₄ plays a major role in disease progression, including cardiovascular diseases. Moreover, the present findings could explain, at least in part, the molecular mechanism that can underlie the beneficial actions of omega-3 EPA observed in many clinical disorders where LTB₄ is believed to be a proinflammatory signal. Therefore, we reasoned that RvE1 acts as a proresolving ligand and would block proinflammatory signals.

In summation, RvE1 specifically interacts with the LTB₄ receptor BLT1 in addition to ChemR23. BLT1 is expressed abundantly in PMNs, and RvE1-BLT1 interactions regulate migration of leukocytes in acute inflammation. ChemR23 does not appear to be highly expressed in PMNs, but present in APCs such as macrophages and DCs, where it plays a regulatory role to control migration and cytokine production. Also, RvE1 attenuates LTB₄- dependent proinflammatory signals such as mobilization of intracellular calcium and NF-κB activation. These results provide a molecular basis not only for the involvement of omega-3 EPA-derived lipid mediators in controlling inflammatory responses, but also for a potential therapeutic utility of RvE1 and its mimetics for a wide range of inflammatory disorders wherein specifically regulating PMNs may be beneficial.

We thank M. H. Small for manuscript preparation; Drs. Andrew M. Tager and Andrew D. Luster for providing BLT1 knockout mice; Drs. Francesca Bianchini and Maria Stan for expert technical assistance; and the Organic Synthesis Core of P50-DE016191 (C.N.S.) led by Dr. Nicos Petasis for preparing synthetic RvE1 and its acetylenic precursor for labeling.

C. N. Serhan is the inventor of several U.S. patents on the structural elucidation of resolvins and their use in controlling inflammatory diseases. Brigham and Women’s Hospital is the assignee for these patents, and they have been licensed by Resolvyx Pharmaceuticals, with whom Dr. Serhan is a consultant.