Anti-Inflammatory Actions of Lipoxin A₄ Stable Analogs Are Demonstrable in Human Whole Blood: Modulation of Leukocyte Adhesion Molecules and Inhibition of Neutrophil-Endothelial Interactions

LIPOXINS (trihydroxytetraene-containing eicosanoids) formed by leukocytes during cell-cell interactions represent a unique class of lipid mediators with potent anti-inflammatory actions.1 In particular, lipoxin A₄ (LXA₄) was found to inhibit neutrophil and eosinophil chemotaxis in vitro,2,3 to block transmigration of neutrophil granulocytes (PMNL) across epithelial cells4and endothelial monolayers,5 and to reduce PMNL entry into inflamed renal tissues.6 Recent results showed that stable analogs of LXA₄ resist rapid inactivation and retain biological activities of native LXA₄, including inhibition of PMNL adhesion and transmigration across epithelial and endothelial monolayers,7 inhibition of interleukin-8 (IL-8) production with consequent impairment of the ability of bacteria-infected epithelia to direct PMNL movement,8 and attenuation of P-selectin–dependent PMNL adhesion and rolling on the mesenteric microvasculature.9 These studies described novel actions of LXA₄ and aspirin-triggered LXA₄ (15-epimer of LXA₄) on isolated cells, intestinal epithelial cells, or isolated vessels after their topical application, yet the question of whether these lipid mediators are active within the microenvironment of human whole blood that possess a wide array of components that can bind lipophilic compounds as well as the issue of the cellular mechanisms that account for their novel inhibitory actions in leukocyte trafficking have not been addressed.

Leukocyte extravasation into inflamed areas involves a multistep interaction of leukocytes and endothelial cells via regulated expression of surface adhesion molecules.10,11 The initial capture and tethering of circulating neutrophils to endothelium is mediated by L-selectin. L-selectin is constitutively expressed by most leukocytes and is rapidly shed after cell activation with a concomitant upregulation of Mac-1 (CD11b/CD18).12 The CD18 integrins, Mac-1 and LFA-1 (CD11a/CD18), are largely responsible for subsequent tightening of the adhesion and transendothelial migration of neutrophils via interactions with their endothelial counterreceptors, intercellular adhesion molecule-1 (ICAM-1) and ICAM-2.10,11 

In the present experiments, we studied the impact of stable LXA₄ analogs in human whole blood and addressed their cellular mechanisms of action with isolated cells. We examined the expression of adhesion molecules on human leukocytes and human coronary artery endothelial cells (HCAEC) and on adhesion of PMNL to HCAEC. We used 16-phenoxy-LXA₄, an analog of the native LXA₄,7 and 15-R/S-methyl-LXA₄, an analog of aspirin-triggered 15-epi-LXA₄.13Platelet-activating factor (PAF) and IL-8 were chosen to activate leukocytes, because these mediators can serve as signals for neutrophils to bind tightly to the endothelium.14,15 

In these studies, the monoclonal antibodies (MoAbs) used included fluorescein isothiocyanate (FITC)-conjugated mouse antihuman L-selectin MoAb DREG-56 (PharMingen, San Diego, CA), R-phycoerythrin–conjugated mouse antihuman CD18 MoAb MEM-48 (Monosan, Uden, The Netherlands), FITC-labeled mouse antihuman CD11a MoAb G-25.2 (Becton Dickinson Immunocytometry Systems, Mountain View, CA), FITC-labeled mouse antihuman E-selectin MoAb 1.2B6 (Serotec, Kidlington, UK), and R-phycoerythrin–conjugated mouse antihuman ICAM-1 MoAb HA58 (PharMingen). Appropriately labeled, class-matched irrelevant mouse IgG₁ was used as a negative control for each staining. The following murine MoAbs were used in neutrophil-endothelial cell adhesion assays: anti–L-selectin MoAb DREG-56 (IgG₁; PharMingen) at 20 μg/mL16; anti–E-selectin MoAb ENA-2 [IgG₁, purified F(ab′)₂ fragments; Monosan] at 10 μg/mL17; and anti-CD18 MoAb L130 (IgG₁; Becton Dickinson) at 10 μg/mL.18 The irrelevant MoAb MOPC-21 (IgG₁; PharMingen) at 20 μg/mL was used as a negative control.

The LXA₄ analogs, 15-R/S-methyl-LXA₄, 16-phenoxy-LXA₄ and 15-deoxy-LXA₄, were prepared by total organic synthesis7 and stored in ethanol at −70°C. An aliquot was removed and diluted in assay medium immediately before use. The highest concentration of ethanol (0.1%) had no detectable effects in any of the assays used.

Lipopolysaccharide (LPS; Escherichia coli O111:B4), C-reactive protein (CRP)-derived peptide 201-206 (Lys-Pro-Gln-Leu-Trp-Pro), dexamethasone 21-phosphate, wortmannin, and genistein were obtained from Sigma Chemical Co (St Louis, MO); PD98059, herbimycin A, and PAF were from Calbiochem (San Diego, CA); and human recombinant IL-8 and tumor necrosis factor-α (TNF-α) were purchased from R&D Systems (Minneapolis, MN).

Venous blood (anticoagulated with 50 U/mL sodium heparin) was obtained from nonsmoking healthy volunteers (male and female, 24 to 55 years of age) who had not taken any drugs for at least 10 days before the experiments. Informed consent was obtained from each volunteer, and the protocol was approved by the Clinical Research Committee. White blood cell counts were between 4,500 and 9,500 cells/μL. Whole blood aliquots were incubated with one of the LXA₄ analogs for 30 minutes at 37°C, 95% air/5% CO₂. Preliminary experiments showed that the maximum effects of LXA₄ analogs could be achieved after 30 minutes of preincubation. Some blood samples were preincubated for 30 minutes with LXA₄ analogs or treated for 120 minutes with dexamethasone (100 nmol/L) with or without LXA₄ analogs and then challenged with PAF (1 μmol/L), IL-8 (10 nmol/L), or CRP peptide 201-206 (100 μg/mL), which downregulates L-selectin expression without inducing cell activation.18 In some experiments, blood samples were challenged with Mg²⁺ (1 mmol/L) and EGTA (2 mmol/L) to induce a higher affinity form of β₂integrins.19 

Direct immunofluorescence labeling of resting and treated leukocytes in whole blood was performed as described.18,20 Leukocytes were stained with saturating concentration of fluorescence dye-conjugated antihuman L-selectin or antihuman CD18 MoAb. Nonspecific binding was evaluated by using appropriately labeled mouse IgG₁. Double- or single-color immunofluorescence staining was analyzed by a cytofluorometer (FACScan; Becton Dickinson) with Lysis II software. Antibody binding was determined as mean fluoresence intensity after gating for neutrophils, monocytes, and lymphocytes by their characteristic forward and side scatter properties. The results are presented as relative fluorescence units (RFU): RFU = (Fu_experimental − Fu_isotype) × 100/(FU_control − Fu_isotype), where FU_experimental and FU_control are the L-selectin and CD18 fluorescence intensity of treated cells and cells cultured in medium only, respectively, and FU_isotypeis the fluorescence intensity of class-matched irrelevant antibody.

PMNL were isolated from peripheral blood by centrifugation through Ficoll-Hypaque (Pharmacia Diagnostics, Uppsala, Sweden), sedimentation through dextran (3%, wt/vol), and hypotonic lysis of erythrocytes. Neutrophils (5 × 10⁶ cells/mL; purity, >97%) were suspended in a modified Hanks' balanced salt solution; incubated with the MAPK kinase (MEK) inhibitor PD98059, the phosphatidylinositol 3-kinase inhibitor wortmannin, or the protein tyrosine kinase inhibitors genistein or herbimycin A for 30 minutes at 37°C; and then challenged with 15-R/S-methyl-LXA₄ for 30 minutes. Surface expression of L-selectin and CD18 was analyzed as just described.

Normal HCAEC obtained from Clonetics Corp (San Diego, CA) were cultured as described.18 HCAEC (passages 3 to 6) seeded into 24-well or 96-well microplates and grown to confluence were used in the experiments.

The adhesion assay was performed as in Zouki et al.18 In brief, monolayers of HCAEC in 96-well microplates were stimulated with LPS (1 μg/mL) with or without LXA₄ analogs or dexamethasone (100 nmol/L) for 6 hours at 37°C in a 5% CO₂ atmosphere. The cells were then washed 3 times, and 2 × 10⁵⁵¹Cr-labeled neutrophils in 100 μL were added. In some experiments, neutrophils were preincubated with one of the LXA₄ analogs for 30 minutes or with dexamethasone (100 nmol/L) for 120 minutes with or without LXA₄ analogs for 30 minutes before the addition to HCAEC. In another set of experiments, LPS-activated HCAEC were incubated for 15 minutes with ENA-2 or MOPC-21 MoAb before the addition of neutrophils. Radiolabeled neutrophils were incubated with DREG-56, L130, or MOPC-21 MoAb for 15 minutes before the addition to HCAEC. After incubation of HCAEC with neutrophils for 30 minutes at 37°C on an orbital shaker at 90 rpm, loosely adherent or unattached neutrophils were washed 3 times and the endothelial monolayer plus the adherent neutrophils were lysed in 200 μL of 0.1% Triton X-100. The number of adhered neutrophils in each experiment was estimated from the radioactivity of a control sample. Treatment of HCAEC with any of the LXA₄ analogs did not affect the integrity of viable endothelial monolayers.

After incubation for 4 hours at 37°C in a 5% CO₂atmosphere with LPS (1 μg/mL) or TNF-α (2.5 ng/mL) in the absence or presence of various concentrations of 15-R/S-methyl-LXA₄, HCAEC were removed from the 24-well microplates by exposure to EDTA (0.01%) in phosphate-buffered saline (PBS) for 10 minutes at 37°C, followed by gentle trituration. Cells were resuspended in ice-cold saline containing sodium azide (0.02%), incubated with a saturating concentration of fluorescein dye-conjugated anti–E-selectin or anti–ICAM-1 MoAb for 30 minutes at 4°C, washed, and fixed in formaldehyde (3.9% in PBS). Nonspecific binding was evaluated by using appropriately labeled mouse IgG₁. Immunofluorescence of HCAEC was then analyzed with a FACScan.

Results are expressed as the means ± SEM. Statistical comparisons were made by ANOVA using ranks (Kruskal-Wallis test), followed by Dunn's multiple contrast hypothesis test to identify differences between various treatments, or by the Wilcoxon signed rank test for paired observations. P values less than .05 were considered significant for all tests.

Incubation of heparinized whole blood with either 15-R/S-methyl-LXA₄ or 16-phenoxy-LXA₄ resulted in higher L-selectin expression on treated than control (untreated) cells and gave concentration-dependent downregulation of CD18 expression on PMNL. Figure 1 reports representative results illustrating the impact of 15-R/S-methyl-LXA₄ added to whole blood. Incubation of blood samples for 30 minutes at 37°C with 5 μmol/L of 15-R/S-methyl-LXA₄ or 16-phenoxy-LXA₄ resulted in 31% ± 8% and 32% ± 8% higher levels of expression for L-selectin than on cells from untreated whole blood and gave 26% ± 5% and 27% ± 5% decreases in CD18 expression, respectively (Fig 2). Similar changes were observed with monocytes and lymphocytes (Fig 2). Neither 15-deoxy-LXA₄(Fig 2) nor structurally degraded (by thermal degradation) 15-R/S-methyl-LXA₄ or 16-phenoxy-LXA₄ affected expression of adhesion molecules (data not shown). When blood samples were incubated at 4°C, none of the LXA₄ analogs studied affected L-selectin expression. For instance, mean fluorescence intensity for L-selectin on PMNL was 129 ± 8 and 120 ± 11 (n = 3, P > .5) in control samples and in the presence of 15-R/S-methyl-LXA₄, respectively.

The addition of PAF (1 μmol/L) to whole blood gave a significant decrease in L-selectin and a marked upregulation of CD11/CD18 on leukocytes (Figs 3 and4). Figure 3 shows a representative experiment on the effect of 15-R/S-methyl-LXA₄ on PAF-induced changes in the expression of these adhesion molecules. Preincubation of blood with either 15-R/S-methyl-LXA₄ or 16-phenoxy-LXA₄ attenuated PAF-induced upregulation of CD18 expression on PMNL, monocytes, and lymphocytes in a concentration-dependent fashion, with IC₅₀ values of approximately 1.5 μmol/L (Fig 4). Essentially complete inhibition was achieved with 5 μmol/L of the LXA₄ analogs. PAF-induced decreases in L-selectin expression on PMNL and monocytes were also markedly, although not completely, inhibited by these LXA₄analogs (Fig 4). Incubation of blood with IL-8 (10 nmol/L) downregulated L-selectin and upregulated CD18 expression on PMNL (Fig 5). As with PAF, both 15-R/S-methyl-LXA₄ and 16-phenoxy-LXA₄ at 5 μmol/L completely inhibited IL-8–induced upregulation of CD18, whereas they partially inhibited changes in L-selectin expression (Fig5). Immunostaining of leukocytes with an anti-CD11b MoAb showed similar changes as those observed with the anti-CD18 MoAb (data not shown).

We also investigated whether LXA₄ analogs are able to modify the affinity of β₂ integrins. Formation of a higher affinity form of LFA-1 and Mac-1 is thought to involve conformational changes and association of the I domain with the β-propeller domain.21,22 The MoAb G-25.2 recognizes this latter domain in LFA-1.22 Incubation of whole blood with PAF or IL-8 in the absence or presence of 15-R/S-methyl-LXA₄ was not associated with a detectable increase in expression of MoAb G-25.2 (Fig6A and B). On the other hand, activation of cells with Mg²⁺and EGTA, which is known to induce the formation of a higher affinity form of LFA-1 without inducing clustering,19 resulted in increases in MoAb G-25.2 expression (Fig 6C) that was not affected by 15-R/S-methyl-LXA₄ (Fig 6C and D) or 16-phenoxy-LXA₄ (data not shown).

Next, we examined whether LXA₄ analogs when added to whole blood could also affect the cell activation-independent downregulation of L-selectin expression. Incubation of blood with CRP peptide 201-206 (100 μg/mL) resulted in, on average, 49%, 42%, and 17% decreases in L-selectin expression on PMNL, monocytes, and lymphocytes, respectively (Fig 7). These changes were inhibited by 15-R/S-methyl-LXA₄ in a concentration-dependent manner, with apparent IC₅₀ values of approximately 250 nmol/L (Fig 6). Similar inhibition was observed with 16-phenoxy-LXA₄ (Fig 7).

Treatment of neutrophils with the MAPK kinase inhibitor PD98059 increased by approximately 27% L-selectin expression without altering CD11/CD18 expression (Fig 8). PD98059 prevented 15-R/S-methyl-LXA₄–induced downregulation of CD11/CD18 expression in a concentration-dependent fashion, whereas changes in L-selectin expression were not affected (Fig 8). Wortmannin at 0.2 μmol/L had no effect, whereas at 2 μmol/L, it prevented 15-R/S-methyl-LXA₄–induced downregulation of CD18 (Fig 8). Furthermore, wortmannin (2 μmol/L) alone downregulated L-selectin expression and inhibited the action of 15-R/S-methyl-LXA₄(Fig 8). 15-R/S-methyl-LXA₄ induction was not affected by the tyrosine kinase inhibitor genistein (Fig 8) or herbimycin-A (data not shown).

After stimulation by LPS, HCAEC increased 27- and 2.6-fold the expression of E-selectin and ICAM-1, respectively (n = 4, bothP < .05; Table 1). 15-R/S-methyl-LXA₄ did not affect basal expression of these adhesion molecules (data not shown) and produced only a slight inhibition of LPS-induced changes (Table 1). The maximum inhibition did not exceed 10%. Similarly, 15-R/S-methyl-LXA₄ failed to significantly inhibit TNF-α–induced E-selectin and ICAM-1 expression (Table 1).

Activation of HCAEC with LPS resulted in a 4.1-fold increase in the number of adherent neutrophils (Fig 9). When HCAEC were challenged with LPS in the presence of one of the stable LXA₄ analogs, only slight decreases in adhesion could be detected (Fig 9A). By contrast, addition of neutrophils preincubated with either 15-R/S-methyl-LXA₄ or 16-phenoxy-LXA₄ to LPS-activated HCAEC attenuated neutrophil attachment in a concentration-dependent fashion (Fig 9A). Significant inhibition of adhesion was detected with 15-R/S-methyl-LXA₄ at a concentration as low as 50 nmol/L. At 5 μmol/L, 15-R/S-methyl-LXA₄ and 16-phenoxy-LXA₄ inhibited neutrophil adhesion by 38% ± 4% and 36% ± 3%, respectively (n = 5, both P < .05; Fig 9A). Neither 15-deoxy-LXA₄ nor decomposed 15-R/S-methyl-LXA₄ or 16-phenoxy-LXA₄ affected neutrophil adhesion (data not shown).

Because multiple receptors are involved in neutrophil adhesion to LPS-stimulated HCAEC18 and LXA₄ analogs affected expression of both L-selectin and CD11/CD18 on PMNL, we assayed the contribution of L-selectin, E-selectin, and CD18 to the binding interaction. A significant proportion of neutrophil-HCAEC attachment was blocked by MoAbs binding to L-selectin (44% ± 4%, n = 5), CD18 (31% ± 4%), or E-selectin (38% ± 5%; Fig 9B). The combination of these MoAbs inhibited neutrophil adhesion by 87% to 98%. Treatment of neutrophils with 15-R/S-methyl-LXA₄ and anti-CD18 MoAb resulted in only a slightly greater inhibition of adhesion than those observed with neutrophils treated with either 15-R/S-methyl-LXA₄ or anti-CD18 MoAb (Fig 9B). The combination of 15-R/S-methyl-LXA₄ with either anti–L-selectin MoAb or anti–E-selectin MoAb resulted in additive inhibition, and the degree of inhibition was similar to those observed when anti–L-selectin MoAb or anti–E-selectin MoAb was combined with anti-CD18 MoAb, respectively (Fig 9B). Combining 15-R/S-methyl-LXA₄, anti–L-selectin MoAb, and anti–E-selectin MoAb blocked approximately 80% of adhesion. Similar results were obtained with 16-phenoxy-LXA₄ (data not shown).

Dexamethasone was used at a concentration of 100 nmol/L, because our previous results showed that the maximum inhibitory effect of dexamethasone on neutrophil adhesion molecule expression can be achieved with this concentration.20 As expected, dexamethasone added alone attenuated by 35% to 45% PAF-induced changes in L-selectin and CD11/CD18 expression on PMNL (Fig 10A). Combining dexamethasone with 15-R/S-methyl-LXA₄ resulted in an almost complete inhibition (Fig 10A). Similar inhibition was observed with both monocytes and lymphocytes (data not shown). Culture of HCAEC with dexamethasone diminished the number of adherent neutrophils (Fig 10B). However, no further decrements in neutrophil adhesion were detected in the presence of LXA₄ analogs. On the other hand, the inhibitory actions of dexamethasone and LXA₄ analogs were additive when neutrophils were preincubated with dexamethasone and LXA₄ analogs before exposure to LPS-activated HCAEC (Fig10B).

In this report, we describe a novel mechanism(s) by which stable LXA₄ analogs can affect the inflammatory response, namely via the modulation of surface expression of adhesion molecules on resting and immunostimulated leukocytes, and inhibition of neutrophil adhesion to the activated endothelium. This inhibition is predominantly mediated via actions of LXA₄ analogs on leukocytes and is distinct from those of glucocorticoids. Moreover and of particular interest, these actions of stable LXA₄ analogs were monitored for the first time within the whole blood environment, where they clearly remain bioactive.

Because dehydrogenation of native lipoxins by human leukocytes results in their rapid inactivation,23 we used LXA₄stable analogs that carry substituents at the carbon 15 through ω 20 end of the native LXA₄ structure, resist conversion,7,24 and retain the biological activities of native LXA₄ in leukocyte adhesion and migration assays.5,24 15-R/S-methyl-LXA₄ is an analog of aspirin-triggered 15-epi-LXA₄ and 16-phenoxy-LXA₄ is an analog of native LXA₄.7,13 The effects of 15-R/S-methyl-LXA₄ and 16-phenoxy-LXA₄ observed in this study are attributed to the molecules themselves, because neither degraded LXA₄ analogs nor 15-deoxy-LXA₄caused detectable changes in the assays used.

Our study documents the complex nature of actions of stable LXA₄ analogs on expression of leukocyte adhesion molecules. These analogs upregulated L-selectin and downregulated CD11/CD18 expression on human resting neutrophils, monocytes, and, to a lesser extent, lymphocytes in whole blood and markedly attenuated changes evoked by immunostimulation. These results are consistent with earlier reports, which showed that native LXA₄ inhibits formyl-Met-Leu-Phe–stimulated upregulation of CD11b/CD18 on human isolated neutrophils.25 The present results indicated that this action is also demonstrable within whole blood with nanomolar to micromolar concentrations of LXA₄ analogs. As expected, higher concentrations were required in whole blood to inhibit the upregulation of CD18 expression than in isolated cells. This might reflect interactions of LXA₄ analogs with serum components such as albumin. Nevertheless, it is impressive that these lipophilic compounds are active in the microenvironment of whole blood and overcome interactions with blood components to specifically regulate leukocytes. Downregulation of CD11/CD18 expression on resting neutrophils by LXA₄ analogs does not involve phosphatidylinositol 3-kinase or tyrosine kinases; rather, it appears to be mediated via activation of MAPK kinase, as suggested by the experiments with PD98059 and wortmannin, which, at a concentration of 2 μmol/L, also inhibits MAPK kinase.26 The LXA₄analogs did not initiate detectable changes in tyrosin phosphorylation of either Erk-2 or p38 MAPK in resting human neutrophils as evaluated by immunoblotting using specific antibodies to the phosphorylated forms of these enzymes (Hachicha et al, unpublished observations). Recent results suggest that intracellular arachidonic acid induces integrin-dependent homotypic adhesion of neutrophils via the Raf-1/MAPK kinase/Erk pathway.27 With respect to LXA₄ receptor activation, it is possible that other Erk or p38 MAPK-independent pathways may be involved in signaling neutrophil adhesion molecule expression.

Physiological stimuli regulate adhesion by either altering the affinity of the individual integrin molecule or by inducing clustering of β₂ integrins, thereby increasing the overall strength of binding.19,28,29 Increasing intracellular Ca²⁺concentration does not induce detectable changes in the affinity of LFA-1.22 Because leukocyte activation evoked by PAF or IL-8 is mediated through increases in intracellular Ca²⁺concentration, these agonists would not increase integrin affinity. Indeed, neither PAF nor IL-8 affected MoAb G-25.2 expression. Activation of leukocytes with Mg²⁺ and EGTA, which results in the formation of a higher affinity form of LFA-1,19increased the expression of MoAb G-25.2, which did not appear to be affected by the LXA₄ analogs studied. These results suggest that LXA₄ analogs do not alter the affinity of either LFA-1 and probably of Mac-1, but rather they interfere with activation-induced clustering of β₂ integrins. However, our results do not preclude the possibility that, in the presence of integrin ligands, LXA₄ analogs might affect a ligand-induced affinity increase secondary to integrin clustering.

Comparison of L-selectin expression at 4°C and at 37°C indicated that, within whole blood, these LXA₄ analogs prevented temperature-induced L-selectin shedding on unstimulated leukocytes. For minutes of activation by either PAF or IL-8 or challenge with CRP peptide 201-206, which does not activate cells, leukocytes release L-selectin from their surface by a proteolytic enzyme. Because this enzyme appears to be constitutively active, formation of an appropriate 3-dimensional structure of L-selectin near the membrane is thought to regulate this proteolytic process.30,31 Whereas it remains to be established whether the conformational changes induced by cell activation dependent or independent stimuli are identical, our results demonstrate that these events are effectively blocked by LXA₄ analogs. A recent study has reported that the intracellular tail of L-selectin is phosphorylated on serine upon activation with chemoattractants.32 It is of interest that calmodulin inhibitors directly induce proteolytic shedding of L-selectin.33 Whether this action of LXA₄involves calmodulin or interference with phosphorylation step(s) remains a key area of inquiry.

Despite inhibition of L-selectin shedding from neutrophils within whole blood, which is thought to promote neutrophil-HCAEC attachment, LXA₄ analogs actually inhibit isolated neutrophil adhesion to HCAEC. This inhibition can primarily be attributed to their actions on neutrophils rather than on HCAEC in this interaction, because only slight decreases in the number of adherent neutrophils were observed after culture of HCAEC with LPS in the presence of LXA₄analogs. Consistently, LXA₄ analogs have little effect on LPS- or TNF-α–stimulated expression of E-selectin and ICAM-1 on HCAEC. Our results indicate that inhibition of neutrophil-HCAEC adherence by LXA₄ analogs is predominantly attributable to inhibition of CD11/CD18 expression on neutrophils. The LXA₄analogs or a function-blocking anti-CD18 MoAb resulted in similar decreases in neutrophil adhesion to HCAEC. The action of LXA₄ analogs and anti-CD18 MoAb was not additive, whereas the inhibition with LXA₄ analogs was additive with anti–E-selectin and anti–L-selectin MoAbs. These results suggest that LXA₄ analogs had little, if any, effect on E-selectin or L-selectin function and did not interfere with E-selectin ligand. Taken together, our data indicate that, in addition to inhibition of P-selectin expression on rat intestinal venular endothelial cells,8 lipoxins are potent regulators of both leukocyte L-selectin and CD11/CD18 expression. Whereas inhibition of P-selectin–dependent capture of neutrophils may be a key mechanism by which LXA₄ analogs inhibit neutrophil-endothelial interactions in the mesenteric circulation,8 functions of adhesion molecules may overlap, and even neutrophil recruitment into other vascular beds may not rely on a P-selectin–dependent adhesion.34 

The present study and previous studies20,35 indicate that the mechanisms of action of LXA₄ analogs differ from those of glucocorticoids. Adhesion molecule expression on resting neutrophils can be modulated by LXA₄ analogs, but not by dexamethasone.20 Both dexamethasone and LXA₄analogs alone inhibited by approximately 20% to 45% PAF-induced changes in L-selectin and CD11/CD18 expression on neutrophils, representing the near maximum inhibition that can be achieved with these compounds (Filep et al20 and the present study). Interestingly, the inhibitory actions of dexamethasone and LXA₄ analogs were additive, resulting in almost complete inhibition of PAF-stimulated changes. The inhibitory actions of dexamethasone on PMNL requires de novo protein synthesis.20Preincubation of neutrophils with either dexamethasone or LXA₄ analogs before addition to LPS-activated HCAEC resulted in partial inhibition of adhesion, and the inhibitory actions of dexamethasone and LXA₄ analogs were additive. Culture of HCAEC with LPS and dexamethasone, but not with LXA₄analogs, resulted in marked decreases in the number of adherent neutrophils. These findings are consistent with the glucocorticoid inhibition of LPS-induced expression of ICAM-1 and E-selectin on endothelial cells.35 By contrast, LXA₄ analogs had little, if any, effect on expression of these adhesion molecules. Interestingly, the mechanism by which LXA₄ analogs inhibit neutrophil-endothelial adhesion also differs from that of several nonsteroidal anti-inflammatory drugs, which induce shedding of L-selectin from neutrophils and inhibit L-selectin–mediated attachment.36 The actions of LXA₄ analogs on β₂ integrins appear to be similar to those reported for piroxicam,37 tepoxalin, a dual cyclooxygenase/lipoxygenase inhibitor,38 and the anti-inflammatory agent leumedin NPC 15669.39 Piroxicam prevents expression of a CD11b activation neo-epitope,37 whereas tepoxalin and leumedin NPC 15669 inhibit upregulation of Mac-1.38,39 However, unlike the LXA₄ analogs, none of these agents affected β₂ integrin expression on resting leukocytes.

In conclusion, this study demonstrates that stable LXA₄analogs modulate adhesion molecule expression on leukocytes monitored within the microenvironment of human whole blood and inhibit neutrophil adhesion to HCAEC via downregulation of CD11/CD18 expression. These effects are distinct from those of either glucocorticoids or nonsteroidal anti-inflammatory drugs. Therefore, stable analogs of native LXA₄ and aspirin-triggered 15-epi-LXA₄may represent a novel therapeutic approach for selective regulation of leukocyte trafficking in host defense, inflammation, and reperfusion injury.

The authors thank Valery V. Fokin for expert assistance in the synthesis of the lipoxin analogs.