Resolvin D1 limits PMN recruitment to inflammatory loci: receptor dependent actions

Abstract

Objective

Resolvin D1 limits neutrophil recruitment during acute inflammation and is derived from omega-3 DHA to promote catabasis. The contribution of its specific receptors, the lipoxin A₄/Annexin-A1 receptor (FPR2/ALX) and the orphan receptor GPR32 are of considerable interest.

Methods and Results

RvD1 reduced human PMN recruitment to endothelial cells under shear conditions as quantified using a flow chamber system. Receptor specific antibodies blocked these anti-inflammatory actions of RvD1, with low (1nM) concentrations sensitive to GPR32 blockade, whilst the higher (10nM) concentration appeared FPR2/ALX-specific. Interestingly, PMN surface expression of FPR2/ALX but not GPR32 increased following activation with pro-inflammatory stimuli, corresponding with secretory vesicle mobilization. Lipid mediator metabololipidomics carried out with 24h exudates revealed that RvD1 in vivo gave a significant reduction in the levels of a number of pro-inflammatory mediators including prostaglandins and LTB₄. These actions of RvD1 were abolished in fpr2 null mice.

Conclusions

Pro-resolving lipid mediators and their receptors, such as RvD1 and the two GPCRs studied here regulate resolution and may provide new therapeutic strategies for diseases with a vascular inflammatory component.

Polymorphonuclear leukocytes (PMN) are of paramount importance for host innate immune defence, and are essential for ongoing health. However, excessive inflammation due to ungoverned leukocyte infiltration can become deleterious to the host and can progress to chronic disease. Indeed, uncontrolled, non-resolving inflammation is a hallmark of many prevalent pathologies including atherosclerosis and arthritis ¹. Contrary to initial belief, an acute inflammatory response does not merely subside due to dissipation of pro-inflammatory mediators. Accruing evidence now signifies that the resolving phase of inflammation is not a passive process, but actively ‘switches-off’ via the biosynthesis of endogenous anti-inflammatory mediators ^{2, 3}.

Omega-3 polyunsaturated fatty acids (PUFA) are known to bestow protective clinical effects in the cardiovascular system and inflammatory disorders including rheumatoid arthritis ^{4, 5}. Hence, the mechanisms by which omega-3 PUFAs exert their biological effects are of on-going interest. Lipidomics profiling of self-limited, inflammatory exudates led to the identification of unique omega-3 derived lipid mediators, designated resolvins and protectins, for their ability to actively accelerate inflammatory resolution ^{6, 7}. Resolvins exert potent anti-inflammatory and pro-resolving actions in experimental animal models and promote wound healing (reviewed in ⁸). Furthermore, resolvins also help maintain vascular homeostasis; RvD2 stimulates vasoprotective prostacyclin and nitric oxide release from the vascular endothelium ⁹, and EPA-derived RvE1 counter regulates platelet activation ¹⁰.

These anti-inflammatory and pro-resolving autacoids mediate their bioactions via specific G-protein coupled receptors (GPCRs). Recently, two GPCRs for Resolvin D1 (RvD1; 7S, 8R, 17S-trihydroxy-docosa-4Z,9E,11E,13Z,15E,19Z-hexaenoic acid¹¹) were identified, and validated using a GPCR/ -arrestin coupled system, namely FPR2/ALX and an orphan receptor GPR32 ¹². Initially characterized as a low affinity N-formyl-Met-Leu-Phe (fMLP) receptor and termed formyl-peptide receptor like 1, FPR2/ALX has since been reclassified due to its high affinity for the endogenous ligand lipoxin A₄ ¹³. Intriguingly, accumulating reports implicate FPR2/ALX as a multi-recognition receptor that recognises a repertoire of protein/peptide agonists as well as high-affinity endogenous lipid ligands, which modulate the host immune response (reviewed in ^{14, 15}). Comparatively, little is currently known about the orphan human receptor GPR32. High-affinity ligands for both GPCRs include RvD1 and LXA₄ as well as the small molecule FPR2/ALX agonist compound 43 ¹². Expression levels of GPR32 are increased following monocyte exposure to GM-CSF or zymosan A for 24-48h ¹², although its relevance in pathophysiological settings needs to be determined. Moreover, the functional relevance of these two receptors in the biological actions of RvD1 on human PMN remains unclear.

Herein, we employed an integrated approach with isolated human cells and a model of acute peritonitis, to provide direct evidence for GPCR-dependent actions of RvD1. On human PMN, low concentrations of RvD1 are sensitive to constitutive GPR32 expression, and higher concentrations elicit FPR2/ALX-mediated responses. We demonstrate unequivocally that the actions of RvD1 are dependent upon receptor mechanisms, which could inform novel anti-inflammatory drug discovery programmes.

Methods

Ethics

All animal studies were conducted with ethical approval from the Queen Mary University of London Local Ethical Review Committee and in accordance with the UK Home Office regulations (Guidance on the Operation of Animals, Scientific Procedures Act, 1986). Human cells were prepared according to a protocol approved by the East London & The City Local Research Ethics Committee (Ref. 05/Q0603/34 ELCHA, London, United Kingdom).

In vitro flow chamber assay

Primary human umbilical vein endothelial cells (HUVEC) were isolated in house by collagenase digestion and used at passage 2 for all experiments. To assess leukocyte-endothelial interactions, HUVEC were plated in μ-Slides VI^0.4 (Ibidi, München, Germany), and confluent monolayers were stimulated with TNF-α (10 ng/ml, 4h; R&D systems). Human polymorphonuclear cells (PMN) were freshly isolated from healthy volunteers using dextran sedimentation followed by gradient-centrifugation as described previously. Immediately prior to flow, PMN were suspended at 1×10⁶/ml in Dulbecco’s PBS supplemented with Ca²⁺ and Mg²⁺ containing 0.1% BSA, and pre-incubated with vehicle (0.1% EtOH) or RvD1 (0.01-100nM; Cayman Chemical, Michigan, USA) for 10 min at 37°C. PMN were perfused over HUVEC at 1 dyne/cm² using a programmable syringe pump (Stoelting, Germany) for 8 min, then 6 random fields/treatment were recorded for 10 sec each. The total number of interacting PMN was quantified as captured and further classified as rolling or adherent if stationary for the 10 sec period. In some experiments PMN were pre-incubated for 10 min with anti-FPR2/ALX (10 μg/ml, clone FN-1D6-A1, Genovac, Freiburg, Germany), anti-GPR32 (10 μg/ml; clone GTX71225, GeneTex, CA, USA), or isotype controls, mouse IgG1 (BD biosciences) or rabbit IgG (10 μg/ml; GeneTex) prior to incubation with vehicle or RvD1 (1 - 100 nM, 10 min; Cayman Chemical, Michigan, USA).

Flow cytometric analysis

To assess PMN degranulation, cells were suspended at 1×10⁶/ml (in DPBS supplemented with Ca²⁺ and Mg²⁺ containing 0.1% BSA) and stimulated with vehicle (0.1% EtOH), TNF-α (10ng/ml, R&D systems), platelet activating factor (PAF, C16 form: C₂₆H₅₄NO₇P; 1 nM, Sigma-Aldrich, Poole, UK), interleukin-8 (IL-8; 10 nM, R&D systems), N-formyl-Met-Leu-Phe (fMLP; 1 μM, Sigma), or fMLP plus cytochalasin B (5 μM, Sigma-Aldrich) for 15 min at 37°C. Surface expression of FPR2/ALX (10 μg/ml Genovac) and GPR32 (10 μg/ml Genetex) were determined following incubation with FITC-conjugated anti-mouse IgG (5 μg/ml, STAR9B, Serotec) or Alexa-488 goat anti-rabbit IgG (10 μg/ml, Invitrogen), respectively. Granule markers CD35 (20 μg/ml, clone E11, Serotec), CD63 (20 μg/ml, MEM-259, Serotec), CD63 (20 μg/ml, MEM-259, Serotec) and CD66b (1 μg/ml, clone G10F5, Biolegend, Cambridge, UK) were determined by flow cytometry. In addition CD62L (1 μg/ml, clone DREG56, ebioscence) and CD11b (1 μg/ml, clone ICRF44, ebioscience) were also monitored. In a separate set of experiments, surface expression of FPR2/ALX and GPR32 were determined on human PMN treated with or without TNFα (10 ng/ml; 5 min, 37°C), then vehicle (0.1% EtOH) or RvD1 0.01-100 nM for 10min at 37°C. In all cases, antibodies or relevant isotype controls were incubated at 4°C prior to flow cytometric analysis using a FACSCalibur flow cytometer using CellQuest II software (Becton Dickinson, Cowley, UK).

Murine peritonitis

Peritonitis was assessed in fpr2 null and littermate controls (male mice, 6-8 wks) generated and bred in house as detailed previously ¹⁶. Vehicle (1.0 % EtOH) or RvD1 (0.1-10 ng/mouse, Cayman Chemical) were administered i.v. followed by i.p. administration of zymosan A (0.2 mg; Sigma-Aldrich). Peritoneal lavages were collected after 4 or 24 h and leukocyte infiltration was assessed by light microscopy, followed by differential analysis using anti-Ly6G (clone 1A8, BD Pharmingen) and anti-F4/80 (clone BM8, BD Pharmingen) staining and flow cytometry analysis.

Sample extraction and mediator lipidomics

All samples for LC-MS/MS analysis were extracted with SPE columns as outlined in ¹⁷. Briefly, columns were equilibrated with 1X column volume of methanol and 2X volume ddH2O. Prior to sample extraction 500pg of deuterium labeled internal standards d₈5S-HETE, d₄LTB₄, d₅LXA₄ and d₄PGE₂ were added to facilitate quantitation of sample recovery. Sample supernatants were diluted with 10X volume of ddH2O, acidified (to pH ~3.5), and immediately loaded onto an SPE column. After loading, columns were washed with 1X volume of neutral ddH₂O and hexane. Samples were eluted with 1.5X volume methyl formate and dried by Speedvac or nitrogen stream. Dried samples were resuspended in methanol/water for LC-MS/MS analysis. Extracted samples were analyzed by a LC-UV-MS/MS system, QTrap 5500, (ABSiex) equipped with an Agilent HP1100 binary pump and diode-array detector (DAD). An Agilent Eclipse Plus C18 column (50 mm × 4.6 mm × 1.8 μm) was used with a gradient of methanol/water/acetic acid of 60:40:0.01 (v/v/v) to 100:0:0.01 at 0.4-ml/min flow rate. Identification was conducted using previously published criteria where a minimum of 6 diagnostic ions were employed ¹⁷. To monitor and quantify the levels of the various lipid mediators, a multiple reaction monitoring (MRM) method was developed with signature ion fragments for each molecule. Calibration curves were determined using synthetic lipid mediator (LM) mixture (d₈-5S-HETE, d₄-LTB₄, d₅LXA₄, d₄-PGE₂, LXA₄, LXB₄, PGE₂, PGD₂, PGF_2α, TXB₂, LTB₄, 15-HETE, 12-HETE, 5-HETE) at 12.5, 25, 50, 100 pg. Linear calibration curve for each compound was obtained with r² values ranging from 0.98-0.99. Quantification was carried out based on the peak area of the Multiple Reaction Monitoring (MRM) transition and the linear calibration curve for each compound.

Zymosan phagocytosis

Macrophages were harvested from Fpr2 null and WT mice 4 days after i.p administration of 1ml, 2% P-100 Bio-Gel in sterile PBS (Bio-Rad, Hemel Hempstead, UK). Peritoneal lavages were filtered through 70 μm cell strainers (BD Biosciences), and cells seeded in 96-well plates (1×10⁵/well). Non-adherent cells were removed and macrophages were incubated with RvD1 (0.001-100 nM, Cayman Chemical) for 30 min, 37°C in RPMI containing 0.1% FCS prior to the addition of FITC-labeled zymosan at a 1:20 ratio (2×10⁶/well; Invitrogen, Paisley, UK). After 20 min, cells were washed three times and fluorescence was determined using a plate reader (NOVOstar, BMG Labtech, Germany).

Statistics

Data are mean ± S.E.M. Multiple group comparisons were made using one-way ANOVA followed by Dunnett’s or Bonferroni’s post hoc analysis. P<0.05 was considered significant.

Results

Resolvin D1 attenuates PMN recruitment under flow

We first tested RvD1 actions on human PMN recruitment under shear conditions, at a physiological flow rate relevant to leukocyte recruitment in an inflamed post-capillary venule ¹⁸. RvD1 drastically reduced PMN-endothelial interactions; significantly blunting initial PMN capture, rolling and firm adhesion to TNF-stimulated endothelium in a concentration-dependent manner (Figure 1A-C). We initially selected 1nM and 10nM concentrations, giving near maximal inhibition, to identify which receptors were mediating the effect of RvD1.

Figure 1. RvD1 potently reduces neutrophil-endothelial interactions under flow.

Human PMN were incubated with vehicle (0.1% EtOH) or RvD1 (0.01 to 100 nM) for 10 min at 37°C. Cells were then flowed over TNF-stimulated endothelial monolayers at 1 dyne/cm² for 8 min, and the extent of cell capture (A), rolling (B) and adhesion (C) were quantified from 6 frames per treatment using Image Pro-plus software analysis. Results are mean ± SEM, n=7 donors, *P<0.05 vs. Vehicle (0.1% EtOH), one-way ANOVA followed by Dunnett’s post-hoc test.

PMN were pre-incubated with blocking antibodies for either FPR2/ALX or GPR32 prior to treatment with RvD1 and the extent of PMN accumulation was then assessed. At the low concentration of 1nM, RvD1 led to a significant attenuation in PMN capture, rolling and adhesion, which was reversed when cells were pre-incubated with anti-GPR32 (Figure 2A). Intriguingly, the neutralizing anti-FPR2/ALX antibody did not annul the actions of 1 nM RvD1, and in fact led to a further decrease in the number of rolling cells. Simultaneous pre-treatment with both antibodies led to a similar level of recruitment seen with vehicle control-treated PMN, highlighting that the actions of RvD1 were abrogated when both GPCRs were neutralized. Notably, pre-incubation of PMN with isotype control antibodies did not alter the anti-inflammatory effects of RvD1 (data not shown).

Figure 2. RvD1 reduces PMN recruitment: GPCR dependent actions.

Human PMN were pre-incubated with anti-FPR2/ALX or anti-GPR32 (10 μg/ml, 10 min, 37°C) prior to incubation for 10 min at 37°C with vehicle or RvD1 at 1 nM (A) or 10 nM (B) final concentration. Then, PMN were flowed over TNF-stimulated endothelial monolayers at 1 dyne/cm² for 8 min, and neutrophil capture, rolling and adhesion were quantified. Results are mean ± SEM, n=3-5 donors, *P<0.05 vs. Vehicle (0.1% EtOH), #P<0.05 vs. RvD1 treatment, one-way ANOVA followed by Bonferroni’s post-hoc test.

Using a higher concentration of 10 nM RvD1, a converse finding was observed, with the inhibitory actions now sensitive to FPR2/ALX blockade (Figure 2B). Also, preincubation with anti-GPR32 caused a further decrease in PMN-endothelial interactions by RvD1. Congruently, co-incubation with both blocking antibodies depleted the actions of 10 nM RvD1, similarly to low 1 nM concentrations, emphasizing GPCR dependency for RvD1 actions on human PMN. We next tested an intermediate concentration of 3 nM RvD1 that seemed to act via both GPCRs, as blocking either receptor alone did not revert its effects whereas cell co-incubation with both antibodies was required to abrogate the actions of RvD1 (Figure S1). Notably, pre-incubation of PMN with anti-FPR2/ALX or GPR32 alone did not alter PMN-endothelial interactions suggesting that these antibodies do not exert agonist actions (Figure S1).

Further support of receptor engagement was attained monitoring GPR32 and FPR2/ALX expression following incubation with the agonist RvD1 (0.01-100 nM). Internalization of GPR32 was observed following exposure to low concentrations of RvD1 (0.1-1 nM), whereas, FPR2/ALX was maximally internalized when PMN were first activated with a pro-inflammatory stimulus (Figure S2).

Pro-inflammatory stimuli mobilize FPR2/ALX to the PMN surface

Given that PMN respond to physiological concentrations of RvD1 via two distinct receptors, we next addressed whether these receptors could be differentially modulated following cell activation. Stimulation of PMN with TNF-α, PAF and IL-8 caused a significant increment in FPR2/ALX expression (Figure 3A), whereas surface levels of GPR32 were not altered (Figure 3B). The status of PMN activation was monitored by CD11b upregulation (Figure 3C) and CD62L shedding (Figure 3D). PMN store a vast selection of functionally important molecules in cytoplasmic granules and vesicles, which can be promptly mobilized upon cell activation, allowing fast PMN responses ¹⁹. To decipher whether FPR2/ALX expression correlates with PMN degranulation, cell surface expression of granule markers was also quantified. Liberation of secretory vesicles was monitored using PE-conjugated anti-CD35; these vesicles have the highest propensity for mobilization, and surface levels were elevated upon PMN stimulation with PAF, IL-8 and TNF-α, correlating with enhanced FPR2/ALX mobilization. Stimulation with high concentrations of fMLP caused secretory vesicle release but only marginally increased FPR2/ALX levels. Following PMN incubation with fMLP and cytochalasin B, we did not observe a significant elevation in cell surface CD35 levels most likely due to a complete release of these vesicles from the PMN cell surface (Figure 3E). Specific granules contain a large number of essential adhesion molecules including the heterodimer CD11b/CD18, and indeed CD11b expression (Figure 3C) mirrored mobilization of the specific granule marker CD66b (Figure 3F). Although a prominent increase in CD66b was observed with fMLP alone or in combination with cytochalasin B, FPR2/ALX expression was not analogous, suggesting secretory vesicles store the major reserve of this receptor. Primary azurophil granule release was initiated by stimulating PMN with fMLP along with cytochalasin B and monitored using FITC-conjugated anti-CD63 (Figure 3G). Surface levels of RvD1 receptors were not increased, indicating that these GPCRs are not localized within primary granules. Notably, following exposure of PMN to RvD1 (0.01-100 nM), no change in CD11b or CD62L expression was observed, nor did RvD1 stimulate the release of granule-associated markers (data not shown). In contrast to the rapid modulation of FPR2/ALX observed in PMN, stimulating HUVEC along our experimental conditions (with or without TNF-α; 4 h) did not alter FPR2/ALX or GPR32 expression (Figure S3).

Figure 3. Neutrophil activation mobilizes FPR2/ALX, but not GPR32, to the cell surface.

Freshly isolated human PMN were stimulated for 15 min with various pro-inflammatory or chemoattractive substances and FPR2/ALX (A) and GPR32 (B) cell surface expression was assessed by flow cytometry. Neutrophil activation was monitored with antibodies recognizing adhesion molecules CD11b (C) and CD62L (D). PMN granule mobilization was assessed using specific surface markers for secretory vesicles; CD35 (E), specific granules; CD66b (F) and azurophilic granules; CD63 (G). Results are mean ± SEM, n=6 donors, *P<0.05 vs. Vehicle (0.1% EtOH), one-way ANOVA followed by Dunnett’s post-hoc test.

RvD1 counter-regulates leukocyte recruitment via Fpr2 in vivo

Intraperitoneal injection of zymosan provoked marked accumulation of leukocytes at 4 hours, with an average of 7.5×10⁶ leukocytes per mouse exudate, which was similar in both genotypes. Administration of RvD1 (at 1ng/mouse) significantly reduced total leukocytes and Ly6G⁺ neutrophil infiltration in WT mice (Figure 4B, D). These anti-inflammatory actions of RvD1 were abolished in the absence of Fpr2 signifying the essential role of FPR2/ALX in mediating RvD1 responses in the mouse.

Figure 4. FPR2/ALX mediates the RvD1 anti-inflammatory actions in vivo.

Total leukocyte and neutrophil infiltration to the peritoneum in wild-type (WT) and Fpr2 null mice was assessed 4 h (A, B) and 24 h (C, D) after injection of zymosan (0.2 mg, i.p.). Mice were treated with vehicle (0.1% EtOH) or RvD1 (1 or 10 ng, i.v.) immediately prior to zymosan administration. Results are mean ± SEM, n=3-6 mice per group. *P<0.05 vs. Vehicle (0.1% EtOH), one-way ANOVA followed by Dunnett’s post-hoc test.

Lipid mediator metabololipidomics were performed with exudates obtained 24h post zymosan as cell free supernatants to elucidate the potential mechanism(s) underlying RvD1 actions in WT mice. We identified the following eicosanoids; lipoxin (LX) B₄, 5,15-dihydroxyeicosatetraenoic acid (HETE) and leukotriene (LT) B₄ from the lipoxygensae pathway and prostaglandin (PG) E₂, PGD₂, PGF_2α and thromboxane (TX) B₂ from cyclooxygenase pathways (Figure 5A). Identification was conducted using published criteria ²⁰ for example LXB₄ and 5,15-diHETE see Figure 5B. Using multiple reaction monitoring (MRM) we quantified the individual mediator amounts and found that RvD1 activated lipoxin biosynthesis stimulating the production of the anti-inflammatory mediator LXB₄ ²¹ and LX-pathway marker 5,15-diHETE (Figure 5C,D). RvD1 also stimulated the biosynthesis of the cyclooxygenase derived PGE₂ (Figure 5E) whilst down regulating production of the pro-inflammatory lipoxygenase derived LTB₄ (Figure 5F). We also found that RvD1 reduced the biosynthesis of the cyclooxygenase-derived prostanoids PGD₂, PGF_2α (~25-50%; Figure 5G,H) and TXB₂ (Figure 5I). This exquisite lipid mediator regulation by RvD1 is lost in the fpr2 null mice, where a significant reduction in the baseline levels for some of the lipoxygenase products such as LXB₄ and LTB₄ was obtained (Figure 5C and F).

Figure 5. RvD1 regulates exudate lipid mediators via fpr2.

Mice were treated with vehicle (0.1% EtOH) or RvD1 (10 ng, i.v.) immediately prior to zymosan administration and exudates collected after 24 h. Lipid mediators in peritoneal in cell free exudates were assessed using LC-MS/MS techniques following solid phase extraction. (A) Representative schedule reaction monitoring chromatograms for selected ion pairs for arachidonic acid derived lipoxins and eicosanoids, Compound I = 12-epi-6-trans-LTB₄ and compound II = 6-trans-LTB₄ (B) Representative MS/MS spectra detailing diagnostic ions employed for the identification of LXB₄ and 5,15-diHETE based on published criteria ²⁰. Quantification of peritoneal lipid mediators (C-I). Results are mean ± SEM, 4 mice per group. *P<0.05 vs. WT + Vehicle (0.1% EtOH), one-way ANOVA followed by Dunnett’s post-hoc test.

RvD1 also exhibits pro-resolving actions, enhancing clearance of microbial particles and apoptotic cells to promote catabasis ^{12, 22}. Herein, we investigated whether the pro-resolving actions of RvD1 on murine macrophage phagocytosis was mediated via FPR2/ALX. To test this, we harvested Bio-gel elicited macrophages from either WT or Fpr2 null mice, and pre-incubated them with vehicle or RvD1 (0.01-100 nM) for 30 min prior to addition of fluorescently labelled zymosan. Vehicle MFI values for macrophage zymosan phagocytosis were 1885±380 WT vs. 1840±396 fpr2^−/− thus no difference was observed between baseline phagocytosis levels of macrophages isolated from either WT or fpr2^−/− mice after 20 min phagocytosis. As predicted, RvD1 significantly enhanced zymosan phagocytosis in WT macrophages (Figure 6A), but failed to increment phagocytosis in fpr2 null macrophages (Figure 6B).

Figure 6. Pro-resolving actions of RvD1 are FPR2/ALX dependent.

Bio-gel elicited macrophages were harvested from WT or Fpr2 null mice and pretreated with vehicle or RvD1 (0.01-100 nM, 30 min) prior to addition of fluorescent zymosan (1:20 ratio, 20 min). Zymosan phagocytosis was evaluated following vigorous washing, using a fluorescent plate reader (NOVOstar). Results are mean ± SEM, n=6 experiments, **P<0.01 vs. Veh (0.1% EtOH), one-way ANOVA followed by Dunnett’s post-hoc test.

Discussion

Polymorphonuclear cell (PMN) recruitment to inflammatory sites comprises a cascade of events including capture, rolling and adhesion to the vascular endothelium, and diapedesis into the underlying tissue ²³. Although these processes are well characterized, the counter-regulation of PMN recruitment by endogenous antiinflammatory and pro-resolving mediators is less renowned. Indeed, control of PMN influx is necessary to prevent unwanted tissue damage, thus the timely resolution of an inflammatory reaction is pertinent to maintain tissue homeostasis.

In the present study, we demonstrate that the anti-inflammatory and pro-resolving lipid mediator, Resolvin D1 (RvD1) limits human PMN recruitment under shear conditions in a mechanism dependent upon its receptors. Recently, two independent GPCRs were identified that RvD1 specifically binds on human leukocytes, namely FPR2/ALX and GPR32 ¹². As yet, a thorough investigation into which of these receptors RvD1 signals to limit human PMN recruitment remained elusive. An apparent dichotomy in receptor involvement was evident depending on the amounts of RvD1, with low concentrations appearing to be GPR32 specific and increased concentrations eliciting FPR2/ALX dependent responses (illustrated in Figure S4). Accordingly, incubation of unstimulated PMN with low concentrations of RvD1 (0.1-10 nM) led to internalization of GPR32 from the plasma membrane, whereas FPR2/ALX was maximally internalized on TNF-α activated PMN using 10-100nM RvD1, implicating receptor engagement. Recently, FPR2/ALX internalization was reported as a prerequisite for the pro-phagocytic actions of LXA₄ and Annexin-A1 derived peptides ²⁴. The anti-inflammatory actions of RvD1 on human PMN are in concordance with recent results, whereby RvD1 blocks actin polymerization and CD11b upregulation induced by the pro-inflammatory lipid mediator LTB₄ ¹², decreases transendothelial migration under static conditions ¹¹, and reduces chemotaxis towards an IL-8 gradient ²⁵.

An intriguing observation was noted when PMN were pre-incubated with anti-GPR32 prior to incubation with 10 nM RvD1, which resulted in a further decrease in PMN-endothelial interactions. We postulate that in inflammatory settings, when RvD1 concentrations are raised to orchestrate catabasis ²⁶, RvD1 signals via FPR2/ALX, which is upregulated following PMN activation. Recent studies have examined FPR2/ALX promoter activity and found that LXA₄ enhances the transcription of this receptor, in a pro-resolving feedback circuit ²⁷. Noteworthy, increased levels of pro-resolving mediators and FPR2/ALX are detected in human pathologies including rheumatoid arthritis ^{28, 29} and acute post-streptococcal glomerulonephritis ³⁰, suggesting that protective mediators and their receptors are both operative within inflammatory settings to aid resolution. Of interest, increasing evidence suggests that several pathologies perpetuate due to the failure of inflammatory resolution, including atherosclerosis ²² and periodontitis ³¹. Thus, persistent/chronic inflammation may not merely be due to excessive production of pro-inflammatory mediators but also attributed to a defect in endogenous anti-inflammatory and pro-resolving pathways. Recently, lipid mediator metabolomics has proved a useful measure of resolution outcomes, with protective mediators associated with early rather than late responders following post-operative abdominal aortic aneurysm repair ³².

PMN are equipped with multiple types of intracellular granules allowing efficient innate effector functions. Upon stimulation, PMN mobilize vesicles and granules to the plasma membrane, which contain receptors, cell adhesion molecules and proteases to promptly facilitate PMN adhesion, diapedesis and killing of bacteria amongst other roles. Although these processes are necessary for host defence, granule release is tightly controlled as their contents can cause extensive tissue-damage. There are four main types of PMN vesicles/granules that are categorised based on their content and propensity for liberation. In sequential order, secretory vesicles are first mobilized and provide adhesion molecules for PMN recruitment and trans-endothelial migration, while gelatinase granules contain metalloproteases required for matrix degradation and hence PMN migration to the inflammatory milieu. Specific granules comprise anti-microbial peptides, whereas the final type to be rallied, the azurophil granules permit proteolytic enzyme release into the phagosome to aid bacterial killing ¹⁹.

Using flow cytometric analysis, we investigated the surface expression of FPR2/ALX and GPR32 following stimulation with an array of pro-inflammatory mediators and in parallel monitored PMN granule mobilization. FPR2/ALX surface levels were increased with pro-inflammatory stimuli, which coincided with a similar increase in CD35 expression, signifying secretory vesicle release. Notably, FPR2/ALX surface expression is elevated on human blister exudate PMN ³³, further supporting the notion that this receptor can be rapidly mobilized on the cell surface as part of a counter regulatory mechanism. Surface expression of CD66b was also enhanced following PMN activation; therefore it is also possible that FPR2/ALX is harboured within the specific granules.

Collectively, these results demonstrate that FPR2/ALX surface expression is rapidly increased upon activation with stimuli that PMN sense during an inflammatory response to help counter-regulate exuberant PMN recruitment. On the other hand, GPR32 in human PMN is constitutively and stably expressed, and can respond to low concentrations of endogenous anti-inflammatory mediators with the aim of conveying normal physiological functions.

In murine physiology, fpr2 represents the human orthologue for FPR2/ALX ¹⁴, whereas there is no known murine counterpart of human GPR32 ³⁴. Thus, we investigated the GPCR dependency of RvD1 in an acute inflammatory model by utilising fpr2 deficient mice. RvD1 significantly reduced peritoneal PMN infiltration in WT mice; of interest, the actions of RvD1 were abolished in the fpr2 deficient mice, signifying the essential role of the RvD1-FPR2/ALX axis in the mouse. No differences in cell recruitment between WT and fpr2 deficient mice were observed at this time-point. A finding that was inline with those published, with no differences in cell recruitment seen in the acute inflammatory models of peritonitis and IL-1β air pouch between genotypes, whereas enhanced responses emerged in fpr2 null mice when tested in the carrageenan paw oedema and inflammatory arthritis models ¹⁶.

Treatment with RvD1 modulated peritoneal lipid mediators, enhancing levels of anti-inflammatory and pro-resolving lipid mediators including LXB₄ and the LX-pathway marker 5,15-diHETE, whilst decreasing pro-inflammatory lipid mediators including TXB₂, PGF_2α and LTB₄ in WT mice. Importantly, PMN directed regulatory actions of RvD1 were not observed in fpr2 null mice demonstrating the pivotal function of this receptor in mediating these important and complex responses. RvD1 treatment also enhanced production of PGE₂ in WT mice which is known to promote vascular responses and lipid mediator class switching that enhances the resolution of acute inflammation ³⁵. Since RvD1 also stimulated endogenous LXB₄ (Figure 5C) we cannot rule out that the anti-inflammatory actions of RvD1 in vivo were mediated, at least in part, by LXB₄ which possesses potent anti-inflammatory actions ²¹. Relevantly, RvD1 was recently shown to modulate a repertoire of mediators and pathways via miRNAs, including the enzyme 5-lipoxygenase ³⁶. Thus, perhaps in the absence of the RvD1 receptor, baseline levels of 5-lipoxygenase products including LXB₄ and LTB₄ are differential in fpr2 null mice. The pro-resolving phagocytic actions of RvD1 were also abrogated in fpr2 null macrophages, further re-enforcing the crucial role of FPR2/ALX in conveying pro-resolving effects of RvD1 in murine tissue. Corroborative results were obtained for LXA₄, which enhanced phagocytosis by bone marrow-derived macrophages from WT but not fpr2 null cells, implicating FPR2/ALX as a multi-recognition yet pro-resolving receptor ²⁴. In the absence of exogenous agonists, transgenic mice overexpressing human FPR2/ALX display reduced PMN infiltration in zymosan peritonitis ³⁷, and mice lacking the murine homologue receptor display an exacerbated response to ischemia-reperfusion injury and arthritogenic serum ¹⁶, further supporting a protective role of this receptor in inflammation. Indeed, human monocyte-derived macrophages treated with RvD1 display enhanced zymosan phagocytosis, a response further enhanced when these cells are transfected with either FPR2/ALX or GPR32 ¹².

Our findings reinforce the central role of FPR2/ALX to transmit regulatory circuits of interest in controlling vascular inflammation, that highlight a functional role for GPR32 in mediating homeostasis. Importantly, we report the receptor-mediated actions of RvD1 with potent regulation of human PMN recruitment and regulation of lipid mediator biosynthesis (without completely blocking PMN-endothelial interactions) as well as enhanced clearance of zymosan. This is in line with other specialized pro-resolving lipid mediators that enhance anti-microbial actions of phagocytes without causing immunosuppression ^{9, 38, 39}. Hence these results establish the RvD1-GPCR interactions in vivo that govern pro-resolving mechanisms. Further study of these can provide a new appreciation of the underlying mechanisms in inflammation-associated diseases and may impact on drug discovery programmes aiming at developing novel therapeutics for vascular pathologies.