Aspirin-triggered 15-Epi-Lipoxin A₄ (LXA₄) and LXA₄ Stable Analogues Are Potent Inhibitors of Acute Inflammation: Evidence for Anti-inflammatory Receptors

Tritiated LXA₄ ([11,12-³H]LXA₄; ∼40 Ci/mmol) was obtained by custom catalytic hydrogenation of 11,12-acetylenicLXA₄ methyl ester that was supplied to and performed at Du Pont New England Nuclear (Boston, MA), and the labeled LXA₄ was purified as in Fiore et al. (18). α-[³²P]dCTP (3,000 Ci/mmol) and γ-[³²P] GTP (30 Ci/mmol) were purchased from Du Pont NEN. LXA₄ synthetic analogues, 15-epi-LXA₄-methyl ester, 5(S),6(R),15(R/S)- trihydroxy-15-methyl-7,9,13-trans-11-cis-eicosatetraenoic acid (15 [R/S]-methyl-LXA₄) methyl ester, and 16-phenoxy-17,18,19,20tetranor-LXA₄-methyl ester, were prepared, isolated and analyzed as described (23). 15(R)-16-phenoxy-17,18,19,20-tetranor-LXA₄ methyl ester (15-epi-16-phenoxy-LXA₄) and 15(R/S)-16-phenoxy-11,12-acetylenic-17,18,19,20-tetranor-LXA₄ methyl ester (15[R/S]-16-phenoxy-11,12-acetylenic-LXA₄) as methyl esters were designed from knowledge of 15-epi-LXA₄ structure and bioactivities (12) and were separated and isolated by RP-HPLC, and their identities were confirmed by NMR. Synthetic LXB₄, LTB₄, and LTD₄ were obtained from Cascade Biochem Ltd. (Reading, Berkshire, England). SKF-104353 was a gift from Smith Kline and French Laboratories (King of Prussia, PA). Dulbecco's phosphate-buffered saline (DPBS) and cell culture reagents were from Whittaker M.A. Bioproducts (Walkersville, MD). Casein was from Sigma Chemical Co. (St. Louis, MO), and silicon oil was from Hüls America (Bristol, PA). Balb/c mice were purchased from Jackson Laboratory (Bar Harbor, ME).

A mouse spleen cDNA library was purchased from Clontech (Palo Alto, CA), and 6 × 10⁵ clones were screened with the EcoRI fragment from the human LXA₄R cDNA employing high stringency. A positive clone (designated 15-2) was isolated. Phage DNA was amplified and purified, and the insert cDNA was excised by EcoRI digestion and subcloned into the EcoRI site of pBluescript II KS(+) (obtained from Stratagene, La Jolla, CA). Sequence analysis showed that this clone 15-2 was a partial clone lacking the amino terminal region (nucleotide 87, of full-length clone; see Fig. 1). To obtain the missing amino-terminal region, we used the rapid amplification of cDNA end or rapid amplification of cDNA end (RACE) technique. The 5′-RACE-Ready cDNA^™ from spleen was purchased from Clontech (Palo Alto, CA), and RACE was performed according to the manufacturer's instructions. The first round of PCR was done between the anchor primer provided by the manufacturer and synthetic primer 5′-GCCATTTCAACAAGAAGGAATGGTAGAG-3′ (antisense of nucleotide 229–257) for 30 cycles (94°C for 30 s, 60°C for 45 s, 72°C for 2 min). The first PCR product was diluted to 1:50, and a second round of PCR was carried out between the anchor primer and a synthetic primer 5′-GCTGTGAAAGAGAAGTCAGCCAATGCTA-3′ (antisense of nucleotide 199–227) using the same condition for 35 cycles. A PCR product of ∼300 bp was obtained and subcloned into pBluescript II KS(+) for sequencing. Overlapping regions of RACE product and clone 15-2 (nucleotide 87–198) were found to be identical. The RACE product was subcloned to the 5′ end of clone 15-2 to construct a full-length clone, using a SpeI site at nucleotide 136. Hydrophobicity analysis of amino acid sequence and homology comparison were performed using Lasergene (DNASTAR Inc., Madison WI).

The sequence of the murine LXA₄R clone from the spleen library was confirmed by RT-PCR of RNA obtained from isolated mouse neutrophils. In brief, for isolation of mouse neutrophils, 2 ml of 2% casein solution was injected into the peritoneal cavity of 8-wk-old female Balb/c mice. 4 h later, peritoneal lavages were performed with DPBS²⁺. Wright-Giemsa staining and light microscopy showed that ∼85% of the cells harvested were neutrophils, which were next taken for centrifugation and extraction of RNA using TriZol reagent (GIBCO BRL, Gaithersburg, MD) following manufacturer's instructions. 0.5 μg RNA was reverse-transcribed and used as the template for PCR, which was performed with sense primer 5′-CAGCTGGTTGTGCAGACAAAATG-3′ (corresponding to nucleotide −20 to 3) and antisense primer 5′-CATCCCACAGCCCCCTCCTCA-3′ (corresponding to nucleotide 1054– 1074), for 25 cycles (94°C for 30 s, 64°C for 45 s, 72°C for 80 s). The PCR product was subcloned into pBluescript II KS(+) and five isolated independent clones were sequenced for confirmation; each was found to be identical.

The coding region of mouse LXA₄R was subcloned into mammalian expression vector pcDNA3 (Invitrogen, San Diego, CA). This plasmid was transfected into Chinese hamster ovary (CHO) cells by the DEAE-dextran method, and after 48 h a binding experiment of [³H]LXA₄ was performed as in Fiore et al. (6). Cells were harvested in DPBS²⁻ (5 mM EDTA), washed twice in DPBS²⁺, and adjusted to 10⁶ cells/ml. For time course experiments, cells were incubated with 0.3 nM of [³H]LXA₄ in the presence or absence of 300 nM unlabeled LXA₄ at 4°C for the indicated times. For further analysis, cells were incubated with 0.3 nM of [³H]LXA₄ in the presence of increasing concentrations of the homoligand LXA₄ for 5 min. Reactions were terminated by rapid centrifugation (12,000 g, 60 s) through silicon oil (d = 1.028), and cell-associated radioactivity was determined by liquid scintillation counting. Specific binding was obtained by subtracting nonspecific binding in the presence of 3 log order excess of unlabeled LXA₄ from each count, and 10⁶ cells were used per each point. Data were analyzed with the Ligand program (Biosoft Elsevier, Cambridge, U.K.).

Mouse LXA₄R cDNA in mammalian expression vector pcDNA3 was also used to establish a CHO cell line stably expressing the mouse LXA₄R (i.e., mouse LXA₄R stable transformant). Mouse LXA₄R cDNA and vector alone, namely, pcDNA3 (used as a mock control), were transfected into CHO-K1 cells by electroporation, and transformants were selected under the existence of 1g/l G418 for ∼2 wk. Expression levels of mouse LXA₄R mRNA of individual colonies were examined by northern hybridization, using mouse LXA₄R cDNA as a probe.

Guanosine triphosphate activity was determined in mouse LXA₄R stable transformant as in Fiore et al. (6) with slight modification. Reactions were terminated at 3 min and the rate of GTP hydrolysis was calculated between time zero and 3 min.

Mouse Multiple Tissue Northern Blot^™ including 2 μg each poly(A)⁺ RNA of heart, brain, spleen, lung, liver, skeletal muscle, kidney, and testis was purchased from Clontech (Palo Alto, CA). Mouse neutrophils were isolated by casein-induced peritonitis, as described in “Confirmation of the mouse PMN receptor sequence” (above). Cells were then pelleted by centrifuge, and RNA was extracted using TriZol reagent (GIBCO BRL) following manufacturer's instructions. 10 μg total RNA was separated by gel electrophoresis in 1% agarose gel containing 1.9% formaldehyde and blotted to nylon membrane. BamHI–PstI fragment of mouse LXA₄ receptor (nucleotide 78–966) was labeled with α-[³²P]dCTP using an oligolabeling kit (Pharmacia, Piscataway, NJ), and hybridization was performed following the protocol of Church and Gilbert in 1% BSA, 7% SDS, 0.5 M phosphate buffer (pH 6.8), 1 mM EDTA (24), followed by washing in wash buffer A (0.5% BSA, 5% SDS, 40 mM phosphate buffer, pH 6.8, 1 mM EDTA) twice for 20 min at 65°C, and then in wash buffer B (1% SDS, 40 mM phosphate buffer, pH 6.8, 1 mM EDTA) four times for 20 min at 65°C. Filters were exposed to x-ray film with intensifying screen at −70°C for 72 h (neutrophil) or 120 h (multiple tissue blot). The human multiple tissue blot was also purchased from Clontech, and northern hybridization was performed as with mouse, using EcoRI fragment of human LXA₄R cDNA as a probe, and exposed to the x-ray film for 24 h.

Human PMN were isolated by dextran sedimentation followed by Ficoll gradient separation, and resuspended in DPBS²⁺ (20 × 10⁶/ml) as in Fiore et al. (18). Human umbilical endothelial cells (HUVEC) were cultured in a gelatin-coated (1%) 12-well plate (3.5 × 10⁵/well). PMN and HUVEC were incubated with [³H]LXA₄ (0.3 nM, 5 min at 4°C) and [³H]LTD₄ (5 nM, 90 min at 4°C), respectively, in the presence or absence of the increasing concentrations of LXA₄ (3–300 nM), LTD₄ (5–500 nM) and the indicated LXA₄ analogues (20– 2,000 nM). After incubations, cell associated label was separated from free label by silicon oil method (PMN, see above) or washing (HUVEC) as in Fiore et al. (18).

Ear inflammation was induced as described by Ekerdt and Müller (25) with the following minor modifications. Balb/c mice (8–10 wk old, female) were anesthetized with intraperitoneal injection of pentobarbital (60 mg/kg). 20 μl of acetone was applied to the inner side of the right ear, and indicated amount of each test compound (i.e., LXA₄ analogues or dexamethasone) suspended in 20 μl acetone was applied to the inner side of the left ear. Approximately 7 min later, LTB₄ (5 μg in 20 μl of acetone) was applied to both ears. At the indicated times, mice were euthanized with an overdose of pentobarbital, and a 6-mm diam of ear sample was obtained from each ear using a skin biopsy punch. Ear skin samples were used for myeloperoxidase assay following the method of Bradley et al. (26). In brief, samples were homogenized in potassium phosphate buffer (pH 6.0) containing 0.5% hexadecyltrimethylammonium bromide, sonicated, and freeze-thawed three times, after which sonication was repeated. The suspension was centrifuged at 16,000 g for 20 min, and 100 μl of supernatant was added to 900 μl of potassium phosphate buffer (pH 6.0) containing 0.167 mg/ml o-dianisidine dihydrochloride (Sigma) and 0.0006% hydrogen peroxide (American Bioanalytical, Natick, MA). Changes in OD were monitored at 460 nm at 25°C, at 30- and 90-s intervals. Isolated mouse neutrophils obtained from casein-induced peritonitis (as in preceding sections) were processed in the same manner and used to obtain a calibration curve for neutrophils, which were enumerated by light microscopy.

Skin samples were obtained using a skin biopsy punch and fixed in 10% buffered formaldehyde. Samples were then paraffin-embedded, sliced, and stained with hematoxylin-eosin.

Statistical analysis was performed using the Student's t-test.

To evaluate the role and actions of LX and aspirin-triggered LX stable analogues in vivo, we first sought the identification and distribution of the murine LXA₄ receptor. The mouse LXA₄ receptor was obtained from a mouse spleen cDNA library. Since this initial clone lacked the amino terminus, a full-length clone was obtained using rapid amplification of cDNA end (RACE) technique (see Materials and Methods). The resulting fulllength clone had an open reading frame encoding 351 amino acids (Fig. 1). To confirm this clone obtained by the RACE technique, we amplified the coding region of mouse LXA₄R by RT-PCR from RNA obtained from isolated mouse neutrophils. When the PCR product was sequenced, five independent clones were obtained that were each identical to the clone isolated from spleen. Therefore, we concluded that this LXA₄R cDNA is indeed present not only in the mouse cDNA spleen library, but it is also in mouse neutrophils (see also Fig. 5 A).

As expected, hydrophobicity analysis showed the presence of a putative seven transmembrane domain characteristic of the G protein–coupled receptor superfamily (27). The mouse LXA₄R has two N-glycosylation sites in NH₂terminal extracellular domain, which are also conserved in human LXA₄R (6, 28). The carboxyl-terminal or cytoplasmic tail of the mouse LXA₄R had nine serine or threonine residues, among which six were also conserved in human LXA₄R (Figs. 1 and 2). The overall homology between human and mouse LXA₄Rs was 76% in nucleotide sequence and 73% in the deduced amino acid. An especially high homology was noted in the sixth transmembrane domain and second intracellular loop (Fig. 2), suggesting important roles for these regions in ligand recognition and signal transduction.

The human LXA₄R displays both specific and stereoselective binding and activation with LXA₄, which is ∼0.5 nM K_d in human PMN and ∼1.7 nM with transfected CHO cells (5, 6, 18). To test the ability of this mouse LXA₄R to bind LXA₄, CHO cells were transfected with mouse LXA₄R cDNA and tested for their ability to specifically bind [³H]LXA₄. Mouse LXA₄R showed specific binding to [³H]LXA₄. When performed at 4°C, the binding saturated after 5 min (data not shown). Scatchard analysis gave a K_d of 1.5 ± 0.6 nM, which was obtained with four separate concentrations of unlabeled LXA₄ (mean ± SEM, n = 4), a value similar to that of the human receptor (K_d 1.7 nM) (6), suggesting that this mouse clone encodes the high-affinity LXA₄ receptor (Fig. 3).

Next, we established a stable transformant of the mouse LXA₄R to further evaluate the ability of mouse LXA₄R to transmit signals. One clone (designated P4-5) had the highest mRNA expression of mouse LXA₄R when examined by northern hybridization (data not shown). To test whether this mouse LXA₄R transmits signal when exposed to LXA₄, we examined GTP hydrolysis activity using this stable cell line P4-5. Treatment of CHO cells expressing human LXA₄R with LXA₄ stimulates GTPase activation (see reference 6). Upon exposure to 10⁻⁸ M LXA₄, the mouse LXA₄R stable transformant activated GTPase approximately twofold when compared to vehicle alone (vehicle 0.07 ± 0.01 vs. LXA₄ 0.15 ± 0.03 pmol/min/10⁶ cells, mean ± SEM, n = 5, P <0.05). This stimulation was stereoselective because, when tested in parallel, equimolar of the positional isomer of LXA₄, LXB₄ did not elicit this activation (vehicle 0.07 ± 0.01 vs. 0.06 ± 0.01 pmol/10⁶ cells, mean ± SEM, n = 5 and 4, respectively). Mock transfected CHO cells did not respond to LXA₄ (Fig. 4). Thus, we concluded that we have identified a functional LXA₄R of mouse.

To obtain the tissue distribution of mouse LXA₄R mRNA, and compare it to that of human, northern hybridization of mouse and human multiple tissue blot was performed (Fig. 5). In mouse, there was one major band of ∼1.4 kb with high stringency conditions. LXA₄R mRNA was most abundant in neutrophils, followed by spleen and lung, organs known to carry contaminants for leukocytes (Fig. 5,A). When exposed longer (for 14 d), there were faint bands associated with heart and liver (data not shown). In humans, we have previously found that LXA₄R mRNA is abundant in lung tissue (6). In the present experiments, we extended these previous observations and observed that LXA₄R mRNA was also abundant in peripheral leukocytes, followed by spleen (Fig. 5 B). In these two organs, in addition to a major band of ∼1.4 kb, another band of ∼2.4 kb was also observed. In testis, there was a single band of ∼1.7 kb, which was not present in mouse. These bands of different sizes in human suggest the possibility of alternative splicing in human tissues, but by comparison they are lacking in mouse tissues.

Like other autacoids, LXA₄ is rapidly converted within seconds to minutes in the microenvironment to inactive LX metabolites, which involves dehydrogenation at the carbon C-15 position as the predominant route. To evaluate the action of lipoxins in vivo, LXA₄ stable analogues were designed and prepared by total organic synthesis, which resist dehydrogenation and ω-oxidation, and retain bioactivity (23) (Fig. 6). Both 15(R/S)- methyl-LXA₄ and 15(S)-16-phenoxy-17,18,19, 20-tetranor-LXA₄ methyl ester (16-phenoxy-LXA₄) were more resistant to the dehydrogenation by recombinant 15-hydroxy prostaglandin dehydrogenase (15-PGDH) than LXA₄, and also resistant to conversion by differentiated HL-60 cells, and were potent inhibitors of neutrophil transmigration and adhesion with IC50 ranging from 1–50 nM. 15-epi-LXA₄ (carrying a C-15 alcohol in the R configuration), one of the new eicosanoids triggered by aspirin treatment, also displays a slower rate of conversion by recombinant enzyme, suggesting a longer bio-half-life than native LXA₄ with its C-15 alcohol in the S configuration.

With these LXA₄ stable analogues in hand, we next examined whether they compete at the same site as native LXA₄ (Fig. 7,A). [³H]LXA₄ binds to its specific receptor on human neutrophils (18), and therefore in the present study we used human cells for the purpose of direct comparison with previous findings (Fig. 7 A). Moreover, it was not possible to obtain mouse peripheral blood neutrophils in the amounts that would permit specific binding experiments and parallel evaluation of each of the synthetic LX analogues. Each of these bioactive LXA₄ analogues competitively displaced [³H]LXA₄ binding in the following rank order: homoligand LXA₄ ≈ 15-epi-LXA₄ >15(R/S)-methyl-LXA₄ >16-phenoxy-LXA₄. The K_i values for LXA₄ in the present experiments (∼2.0 nM) were comparable to the previously reported values for human neutrophils (18) and CHO cells expressing human LXA₄R (6). Thus, LXA₄ stable analogues compete at the same myeloid receptor as native LXA₄ on human PMN.

LXA₄ also carries a vasodilatory action, and is known to modulate the specific binding of LTD₄ on endothelial cells (21) and glomerular mesangial cells (20). Thus, in addition to acting at the myeloid LXA₄R, LXA₄ also competes at a receptor subtype that also recognizes and responds to the peptidoleukotriene LTD₄. LXA₄ also inhibits peptidoleukotriene (LTC₄ and LTD₄)-induced upregulation of P-selectin in endothelial cells (4). In view of these findings, we examined whether these LXA₄ stable analogues could compete for [³H]LTD₄–specific binding to endothelial cells. As seen in Fig. 7 B, both native LXA₄ and 15-epi-LXA₄ competed with [³H]LTD₄ as effectively as the homoligand LTD₄ and the well-characterized LTD₄ receptor antagonist SKF-104353. The LX stable analogues 15-(R/S)-methyl-LXA₄ and 16phenoxy-LXA₄ each gave K_i values within the same range as the LTD₄ antagonist SKF-104353. These results suggest that these LXA₄ stable analogues and 15-epi-LXA₄ recognize the same receptors present on both neutrophils and endothelial cells as LXA₄.

LXA₄ inhibits neutrophil adhesion and transmigration in vitro (4), and causes downregulation of CD11b/ CD18 on human neutrophils (5). To evaluate this potential anti-inflammatory action of LXA₄ and test whether these compounds do indeed carry this action in established in vivo models, we induced skin inflammation in mouse ears and examined the impact of several new LXA₄ stable analogues. After topical application of LTB₄ on mouse ear, time-dependent increase of myeloperoxidase (MPO) activity in ear skin was observed (Fig. 8, A–C). Since MPO is well established as a marker of neutrophil infiltration (26), we calibrated the degree of MPO activity and the number of neutrophils using standard curves obtained from neutrophils isolated after peritoneal lavage (Fig. 8,A). At equimolar concentrations neither FMLP nor platelet-activating factor induced PMN infiltration when applied topically in acetone (Fig. 8,A). In contrast to the chemotactic ability of LTB₄, the same amounts of LTD₄ did not stimulate neutrophil infiltration into this tissue (Fig. 8 B).

Applied alone to the ear, 16-phenoxy-LXA₄ had no direct effect on neutrophil infiltration (0–8 h; Fig. 8,B). Also, no neutrophil influx was noted at intervals up to 48 h (not shown). When mouse ears were exposed to 16-phenoxyLXA₄ just before the application of LTB₄, neutrophil infiltration was markedly attenuated at each time point and the percent inhibition observed at 24 h was ∼85% (Fig. 8,C). Hematoxylin-eosin staining of ear biopsies confirmed these results and established that 16-phenoxy-LXA₄ applied alone did not alter the tissue architecture (Fig. 9). Ears exposed to LTB₄ exhibited prominent PMN infiltration in perivascular regions (B), and this PMN infiltration was markedly attenuated when the ear was exposed to topical 16-phenoxyLXA₄ (Fig. 9 C).

The inhibitory action of 16-phenoxy-LXA₄ was concentration-dependent (Fig. 8,D), and an IC50 was estimated to be ∼120 nmol/cm². When its potency was directly compared in the same model to a known anti-inflammatory agent, dexamethasone, 16-phenoxy-LXA₄ was as potent as dexamethasone at the equivalent concentrations (Fig. 8 D).

In the case of native LXA₄, its stereoisomer 15-epiLXA₄, which is generated with aspirin treatment, carries its C-15 alcohol in the R configuration and is known to be more potent than native LXA₄ in vitro (12). To address the importance of chirality at C-15 position in vivo, we prepared and examined the actions of 15-epi-16-phenoxy-LXA₄ that carries its C-15 alcohol in the R configuration. This analogue, like 16-phenoxy-LXA₄, was designed to resist inactivation (23) and is an analogue of the aspirin-triggered 15-epi-LXA₄ from this series of LXA₄ mimetics (see Fig. 6 for structure). 15-epi-phenoxy-LXA₄ was as potent as, or more potent than, 16-phenoxy-LXA₄ (Fig. 8,E), a finding that supports the notion that the R configuration at the C-15 position augments the inhibitory actions of LXA₄ (23). When the acetylenic racemate 15(R/S)-16-phenoxy-11,12acetylenic-LXA₄ (see Fig. 6) was examined in this model, it showed much less anti-inflammatory action than did either of the tetraene-containing compounds (i.e., 16-phenoxyLXA₄ or 15-epi-16-phenoxy-LXA₄), suggesting that the loss of tetraene configuration abrogates bioactivity of these analogues (Fig. 8 E). Thus, LXA₄ and aspirin-triggered 15-epiLXA₄ stable analogues have anti-inflammatory actions in the mouse ear inflammation model as exemplified by inhibition of neutrophil infiltration, and this bioaction was stereoselective.

The present results demonstrate, for the first time, cloning of the mouse myeloid LXA₄R and an anti-inflammatory action for both LXA₄ and novel aspirin-triggered LXA₄ stable analogues in vivo, namely, that topical application of these analogues and association with the murine receptor inhibits neutrophil infiltration in skin. The mouse LXA₄R isolated from a spleen cDNA library had a characteristic sequence of seven transmembrane spanning G protein–coupled receptors (27), and its homology to the human LXA₄R (6) was 73% in amino acid (Figs. 1 and 2). Mouse LXA₄R gave high-affinity binding to [³H]LXA₄ (K_d 1.5 nM), with values similar to those obtained with the human LXA₄R (1.7 nM) expressed in CHO cells (6) (Fig. 3). CHO cells stably transfected with mouse LXA₄R and exposed to LXA₄ selectively hydrolyzed GTP, indicating that LXA₄ stimulates functional coupling of LXA₄R and G protein (Fig. 4). Tissue distribution of mouse LXA₄R mRNA paralleled the appearance of human LXA₄R mRNA, and this mRNA was most abundant in mouse neutrophils, followed by spleen and lung (Fig. 5). Bioactive LXA₄ stable analogues effectively displaced both [³H]LXA₄ and [³H]LTD₄ binding to human neutrophils and endothelial cells, respectively (Fig. 7), results that are consistent with those obtained with native LXA₄ (18, 21). These LXA₄ analogues inhibited neutrophil infiltration in the mouse ear inflammation model, and the level of inhibition was as potent as the well established anti-inflammatory steroid, dexamethasone (Fig. 8). The inhibitory actions of LXA₄ stable analogues were stereoselective, with the R epimer (i.e., 15-epi-16-phenoxy-LXA₄: aspirin-triggered 15-epi-LXA₄ analogue) showing a trend for greater potency in this model (Fig. 8).

Of interest, in the phylogenetic tree, mouse and human LXA₄R belong to the chemokine receptor family, rather than having an association with other eicosanoids such as the family of prostanoid receptors (29). The leukotriene receptors remain to be identified at the molecular level. The chemokine receptor group also includes both the Fusin receptor and RANTES receptor, which have recently been shown to serve as cofactors for HIV-I infection (30, 31). LXA₄ is generated during HIV infection at least in vitro (32), but whether there is a connection between these and the present results with LXA₄ actions remains to be determined.

The results presented here are also the first direct demonstration of the ability of both LXA₄ and novel aspirintriggered LXA₄ stable analogues to inhibit neutrophil migration in vivo. 16-phenoxy-LXA₄, which resists enzymatic degradation and is a potent inhibitor of neutrophil adhesion and transmigration in vitro (23), applied topically to mouse ear skin just before induction of inflammation clearly blocks neutrophil infiltration in a concentration-dependent fashion. LXA₄ analogue-induced inhibition of neutrophil infiltration into the ear skin was as potent as topical applications of dexamethasone and required the tetraene structure, since the 11,12-acetylenic-containing analogue was essentially not effective within this concentration range (Fig. 8). It is of particular interest to point out that other eicosanoid stable analogues such as the analogue of PGI₂, Iloprost, significantly enhance LTB₄-induced leukocyte infiltration in this mouse ear model (25). Thus the inhibitory actions of LXA₄ and 15-epi-LXA₄ stable analogues established here for the first time in an in vivo model provide evidence for the ability of LXA₄ to block neutrophil migration, which appears to be a unique action of lipoxins when compared to other eicosanoids that either initiate (i.e., LTB₄) or enhance (PGE₂, PGI₂) this response in vivo (9, 25).

These inhibitory actions of LXA₄ and 15-epi-LXA₄ analogues are not likely to be the result of blocking LTB₄ binding to its receptor on neutrophils, because LXA₄ does not compete for the specific binding of [³H]LTB₄ (4°C) to neutrophils or differentiated HL-60 cells expressing LTB₄ receptors (21). One possible cellular basis for this in vivo inhibitory action of these analogues is the downregulation of neutrophil CD11b/CD18 as recently demonstrated for native LXA₄ in vitro, which requires the engagement of LXA₄R (5). Since these LXA₄ and 15-epi-LXA₄ analogues, which stereoselectively inhibit neutrophil migration, also effectively displace specific binding of [³H]LXA₄ to its myeloid receptor on neutrophils (Fig. 7,A), it is highly likely that these LXA₄ analogues interact with the LXA₄R in vivo. This notion is supported by the finding that the mouse receptor is primarily associated with mouse neutrophils, which appear to be the target for their in vivo actions (see Results). To address a possible systemic action of the LX analogues applied topically, we added 15-epi-16-phenoxy-LXA₄ to one ear and applied LTB₄ to the other (as in Fig. 8). In this setting the LXA₄ analogue did not inhibit LTB₄-induced PMN infiltration (data not shown), suggesting that the action of these analogues was local. Along these lines, pharmacokinetic studies are in progress in this laboratory to design LX analogues with increased systemic actions. Together these findings indicate that activation of the LXA₄R in vivo results in an anti-inflammatory outcome counteracting the actions of proinflammatory signals such as LTB₄ in vivo. Moreover, they provide the first evidence for anti-inflammatory seven transmembrane spanning receptors and pathways.

Aspirin treatment of various cell types in vitro enhances native lipoxin production during cell–cell interactions and triggers the generation of 15-epi-lipoxins by a separate biosynthetic pathway (reviewed in reference 8). These novel lipoxin R epimers appear to mediate some of the beneficial actions of aspirin, in particular, 15-epi-LXA₄, which blocks neutrophil adhesion to endothelial cells (12). Therefore, it is of interest that 15-epi-16-phenoxy-LXA₄, an analogue of 15-epi-LXA₄, was the most potent inhibitor of neutrophil infiltration in vivo of this series of LXA₄ and 15-epiLXA₄ analogues (Figs. 6 and 8 E). Aspirin has both beneficial and deleterious actions in humans. It is possible that certain of aspirin's beneficial actions might now be open to further experimentation using analogues of 15-epi-lipoxins. Taken together, the present findings provide direct further evidence that LXA₄ and novel aspirin-triggered 15-epiLXA₄ stable analogues, as well as mouse LXA₄R cDNA, serve as useful tools to investigate the actions and role of LXA₄ and aspirin-triggered 15-epi-LXA₄ in vivo.

We thank Mary Halm Small for expert assistance in preparation of this manuscript and Dr. B. Schmidt of the Dermatopathology Division, Department of Pathology, Brigham and Women's Hospital for expert preparation and histological evaluation of the skin biopsies.

These studies were supported in part by National Institutes of Health grant nos. GM-38765 and P01DK50305 (C.N. Serhan) and a research grant from Schering AG (C.N. Serhan and N.A. Petasis). T. Takano was also supported in part by a fellowship from the Ministry of Education, Science, and Culture of Japan.