Introduction

Top Abstract Introduction Methods Results Discussion References

Neutrophils (PMN) play important roles in host defense, inflammation and tissue injury (Baggiolini et al., 1993; Ben-Baruch et al., 1995; Weissmann, 1989). PMN activation and subsequent responses such as chemotaxis, adhesion, O₂ generation and enzyme granule release are regulated by a range of pro-inflammatory signals, which are believed to be potential targets for therapeutic interventions. Actions of many chemotactic peptides or lipid-derived mediators are transduced via seven-transmembrane receptors that are coupled to specific heterotrimeric G proteins. Biological activities, signal transduction and specific receptors are well characterized on these cells for the chemokine IL-8 (Knall et al., 1996), the eicosanoid LTB₄ (Sumimoto et al., 1988), chemotactic peptide fMLP (Bommakanti et al., 1995) and PAF (Nakamura et al., 1991). PMN express high-affinity receptors for these ligands which, despite distinct biological profiles of activity, share common modes of signal transduction involving activation of phospholipase C, phospholipase D, mitogen-activated protein kinase cascade, inositol 1,4,5-trisphosphate generation and Ca⁺⁺ mobilization (reviewed in Ben-Baruch et al., 1995).

Cytosolic pH (pH_i) changes are coupled tightly to functional responses during neutrophil activation. The well-characterized PMN agonists LTB₄ (Sumimoto et al., 1988) and fMLP (Weisman et al., 1987) each stimulate a biphasic change in pH_i that is characterized by rapid acidification and gradual Na⁺/H⁺ antiport-dependent alkalinization. These receptor-mediated changes in cytosolic pH apparently regulate certain PMN responses such as O₂ generation and chemotaxis (Simchowitz, 1985). Impaired PMN functions in diseases such as chronic granulomatous (Åhlin et al., 1995) and end-stage renal failure (Haynes et al., 1992) are associated with reduced electron proton shifts and inability of cytosolic acidification. The impact of other PMN ligands or those of vascular endothelial cells on changes in pH_i have not been established. Development of the microphysiometer (Cytosensor), a silicon-based potentiometric sensor that monitors rapid changes in EAR, has facilitated the study of plasma membrane receptors (McConnell et al., 1992). Ligand-operated changes in EAR are "read-outs" of metabolic changes within cells, such as ATP demand, and/or can reflect direct ion channel activation (e.g., Na⁺/H⁺ exchange). This system has been used to characterize both seven-transmembrane and tyrosine kinase receptors (McConnell et al., 1992; Owicki and Parce, 1992). Despite wide interest, inflammatory mediators that define surface receptors on either PMN or EC have not been characterized with respect to receptor-mediated changes in EAR.

Endogenous anti-inflammatory mediators, such as lipoxins (LXA₄ and LXB₄), can play counter-regulatory roles with PMN agonists (reviewed in Serhan, 1997) that may be relevant in inflammation and reperfusion injury. In this regard, LXA₄ in the nanomolar range inhibits PMN adhesion to and transmigration across both endothelial and epithelial cells. These actions of lipoxin, within this concentration range, are mediated via G-protein-coupled membrane receptors and sharply contrast the proadhesive and transmigratory actions of either the lipid ligand PAF and LTB₄ or the chemotactic peptide fMLP (reviewed in Serhan, 1997). The human PMN and monocyte (Maddox et al., 1997) seven-transmembrane receptor that is activated by LXA₄ and its analogs was cloned and sequenced. Interest in these lipid mediators has heightened with the discovery of an aspirin-triggered circuit that leads to the biosynthesis of a new class of mediators, namely 15-epi-lipoxins. 15-epi-LXA₄ is more potent than LXA₄ as an inhibitor of PMN adhesion (reviewed in Serhan, 1997). In view of the inactivation pathway for lipoxin, stable LXA₄ analogs were designed. LXA analogs were prepared by total organic synthesis to be used as both experimental tools and for potential therapeutic evaluations. These analogs resist rapid metabolism and retain the activity of native LXA₄ (i.e., inhibiting both PMN adhesion and transmigration) (Serhan et al., 1995).

Changes in oxygen levels within the microenvironment of tissues is associated with inflammatory processes, various tissue diseases and ischemia/reperfusion (Derevianko et al., 1996; Welbourn et al., 1991; West and Wilson, 1996). Change in oxygen levels, from normoxia to hypoxia, followed by reoxygenation can profoundly alter cell function and signal transduction events. The impact of these changes in oxygen levels in either PMN or vascular endothelial cells remains to be fully appreciated. Yet, ischemic events clearly lead to alterations in the pH microenvironment within tissues (Bak and Ingwall, 1994). The pivotal role of PMN in tissue injury observed after hypoxia/reoxygenation is emphasized by the protective effects observed in certain animal models with either depletion of circulating PMN or treatment with CD18/ICAM antibodies that prevent leukocyte accumulation in target tissues (reviewed in Welbourn et al., 1991; West and Wilson, 1996). The extent to which ligand-receptor interactions and subsequent functional responses in PMN are affected directly by hypoxia/reoxygenation remains to be determined. In this study, we characterized PMN and endothelial cell EAR with known ligands to evaluate the impact of hypoxia/reoxygenation using the microphysiometer and extended these observations to PMN-endothelial cell interactions.

We report the first results with PMN and endothelial cells which indicate that lipid and peptide ligands each evoke unique profiles for EAR and demonstrate that responses to well established pro-inflammatory signals are not dramatically altered after intervals of hypoxia/reoxygenation. In contrast, sequential changes in oxygen levels enhanced the inhibitory actions of a LXA₄ stable analog on PMN transendothelial migration, a key event in both inflammation and reperfusion injury.

	Methods

Top Abstract Introduction Methods Results Discussion References

Materials. Synthetic LXA₄, LXB₄, LTB₄, LTB₄ dimethyl amide, PAF (C-16), hexanol amino PAF (C-16) and PGE₂ were purchased from Cascade Biochem Limited (Reading, Berkshire, England). 16-phenoxy-LXA₄-ME and 15(R/S)-methyl-LXA₄-ME prepared by total organic synthesis and characterized as in Serhan et al. (1995) were synthesized for the present experiments by Dr. Petasis and Dr. Fokin (University of Southern California, Los Angeles, CA). Diazomethane was prepared from N-methyl-N'-nitro-N-nitroguanidine (Aldrich Chemical Company, Milwaukee, WI) and LTB₄-treated (20 min, 25°C) to obtain LTB₄ carboxy methyl ester as used in Serhan et al. (1995). Recombinant human IL-8 and GM-CSF were purchased from R&D Systems (Minneapolis, MN) and pertussis toxin from ICN (Costa Mesa, CA). Transwell cell capsules (tissue culture treated and untreated, 12 mm, 0.3 mm) and tissue culture plates were obtained from Costar Corp. (Cambridge, MA), and MD-RPMI 1640 (1 mM phosphate) and low melting point agarose in MD-RPMI (entrapment media) were from Molecular Devices (Sunnyvale, CA). DPBS, RPMI 1640 and MD-Dulbecco's modified Eagle's medium (1 mM phosphate) were purchased from Bio Whittaker (Walkersville, MD). 5-(N,N-Hexamethylene)-amiloride, fMLP, calcium ionophore (A23187) and all other reagents were obtained from Sigma Chemical Co. (St. Louis, MO).

Cell isolation, culture and preparation for microphysiometer analysis. Human PMN from healthy donors who had not taken aspirin or other medication for at least 2 weeks were freshly obtained by venipuncture and isolated by the modified Böyum method as used in Serhan et al. (1995). Cells were suspended in modified MD-RPMI or RPMI (1.7 × 10⁷ cells/ml) adjusted to pH 7.4 and contained 96 ± 3% PMN, as numerated by light microscopy. After isolation, viability was evaluated by trypan blue exclusion and was >98%. HUVEC were studied at passages 1 and 2 and isolated according to Brezinski et al. (1989) and as used previously (Papayianni et al., 1996; Serhan et al., 1995). HUVEC were isolated by collagenase digestion and grown on gelatin-coated (1%) tissue culture plates in RPMI 1640 supplemented with 15% bovine calf serum (HyClone Laboratories, Logan, UT), 15% NU-serum (Collaborative Research Inc., Lexington, MA), 50 µg endothelial mitogen (Biomedical Technologies Inc., Stoughton, MA), 8 U/ml heparin, 50 U/ml penicillin and 50 µg/ml streptomycin.

Cell preparation for microphysiometer analysis. HUVEC were plated on gelatin-coated (1%) tissue culture-treated transwell capsules and propagated (as described above) at 5% CO₂ and 37°C until 75% confluence was achieved (3 days), after which transwell capsules were transferred into a microphysiometer (Cytosensor, Molecular Devices, Sunnyvale, CA). For microphysiometer analysis, nonadherent PMN were immobilized within 25% entrapment media, per manufacturer's instructions, always containing ~1.3 × 10⁵ PMN (10 µl), and were placed onto nontissue culture-treated transwell cell capsules and transferred into the microphysiometer. In indicated experiments, PMN were incubated at 37°C in an atmosphere of 5% CO₂ (cells were mixed gently every 15 min) before embedding in entrapment media in the following settings: 1) 1-hr incubation with 25 ng/ml GM-CSF or buffer alone; 2) 2-hr incubation with 5 µg/ml pertussis toxin or buffer alone; 3) in hypoxia/normoxia experiments, cells were suspended in conditioned RPMI and maintained (37°C/5% CO₂) either in normoxic conditions for 120 min in an incubator (pO₂ 147 mm Hg), or in hypoxic conditions in a chamber (Coy Laboratory Products, Ann Arbor, MI) (pO₂, 20 mm Hg). The hypoxic chamber consisted of an airtight glove box with the atmosphere continuously monitored by an oxygen analyzer interfaced with oxygen and nitrogen flow adapters. Oxygen concentration in the hypoxic chamber was 2% (pO₂, 20 mm Hg) with the balance containing nitrogen, carbon dioxide (constant 5%) and water vapor from the humidified chamber. Media were conditioned for 2 hr in the hypoxic chamber before use and achieved a pO₂ of 20 mm Hg within 90 min as monitored by a blood gas analyzer (Ciba-Corning, Essex, England) as described previously (Colgan et al., 1996; Zünd et al., 1996).

Microphysiometer evaluation of receptor ligand activation. Changes in PMN and HUVEC EAR were evaluated by use of a Cytosensor microphysiometer and computer workstation as configured in McConnell et al. (1992) and Owicki and Parce (1992), without modifications. Cells were equilibrated for 30 min (PMN in MD-RPMI and HUVEC in MD-Dulbecco's modified Eagle's medium). Extracellular acidification rates were determined by 30-s potentiometric rate measurements (µV/s; pump off cycle) after an 80-s pump cycle with a 10-s delay (120 s total cycle time). Cells were perfused with the indicated agonists for 30 s before the first rate measurements. Perfusion duration depended on the ligand being evaluated and ranged from 5 to 15 min. Acidification rates (µV/s) were normalized to basal rates (100%) three cycles before agonist or antagonist addition. In selected experiments, PMN were perfused with LXA₄ (100 nM, 15 min), hexamethylene amiloride (10 µM, 60 min) or vehicle before addition of agonists. Vehicle (EtOH) concentrations did not exceed 0.1% vol/vol and did not cause statistically significant changes in the basal acidification rate. Unless otherwise indicated each experiment represents the mean of three separate donors.

Superoxide anion generation by PMN. NADPH oxidase activity was measured as the superoxide dismutase inhibitable reduction of ferricytochrome C, as described in McCord and Fridovich (1969). PMN (4 × 10⁶ cells/ml) were incubated at pH 7.45, 37°C for 15 min with LXA₄, the selected LXA₄ analogs or vehicle alone at the indicated concentrations. The incubations were then exposed to fMLP (50 nM) for 10 min, and cytochrome C reduction was monitored. In parallel incubations, PMN were maintained in DPBS for 15 min and exposed to LXA₄, 16-phenoxy-LXA₄-ME or 15-R/S-methyl-LXA₄-ME for 10 min. Incubations were terminated by rapid pelleting of cells at 4°C and supernatants were taken and placed immediately on ice. Changes in absorbance at 550 nm were monitored and calculated as nanomoles of reduced cytochrome C.

PMN transendothelial (HUVEC) migration assay. HUVEC were grown to confluence on transwell polycarbonate filters (3 µm) placed in 24-well tissue culture plates which divided the monolayer into a 200-µl upper and 1-ml lower compartment. HUVEC were exposed (24 hr) to normoxia (i.e., pO₂, 147 mm Hg) or hypoxia (defined as pO₂, 20 mm Hg) for 24 hr (37°C, 5% CO₂) as described (Zünd et al., 1996; Colgan et al., 1996). Before the assessment of PMN transendothelial migration, conditioned monolayers of HUVEC were washed (2 times) with either normoxic or hypoxic DPBS (1 ml). Normoxic PMN (2.5 × 10⁷ cells/ml) were treated with indicated concentrations of the LXA₄ stable analog (16-phenoxy-LXA₄-ME) for 30 min before addition (40 µl, 1 × 10⁶ PMN) to the upper endothelial compartment containing 160 µl phosphate-buffered saline. PMN exposed to 16-phenoxy-LXA₄-ME were added directly to the monolayers; therefore, a 5-fold dilution of the indicated lipoxin analog concentration was present during the transmigration assay (90 min, 37°C). In selected experiments, PMN were maintained in a hypoxic environment for 2 hr, followed by a 30-min incubation with 16-phenoxy-LXA₄ before transfer to either normoxic or hypoxic conditioned monolayers of HUVEC. Transmigration was initiated by the addition of LTB₄ (10 nM), a concentration that evoked maximal PMN migration as reported in Papayianni et al. (1996). All experiments were performed in a 37°C incubator. After 90 min, stimulant-induced transmigration of PMN was quantitated by assaying the marker of azurophilic granules myeloperoxidase (MPO) as in (Parkos et al., 1991). Nonadherent PMN were washed extensively from the endothelial surfaces, and PMN that had traversed the monolayer completely (PMN in the lower reservoir) were quantitated with a standard calibration curve of the MPO activity in total PMN added to the upper compartment.

Statistical analysis. Unless otherwise indicated all values are represented as mean values ± S.E.M. Results were analyzed using Student's t test or by analysis of variance. Differences were considered significant at the P < .05 level.

	Results

Top Abstract Introduction Methods Results Discussion References

Distinct EAR profiles for PMN exposed to inflammatory mediators. Extracellular acidification rates obtained with a representative panel of known pro- as well as potential anti-inflammatory mediators were evaluated by use of a Cytosensor microphysiometer with isolated PMN embedded in soft agar matrix (see "Methods"). Each mediator gave a characteristic and reproducible EAR profile. Among the group of mediators evaluated, LTB₄ and fMLP were the most potent agonists, increasing EAR by as much as 112 ± 15% and 278 ± 25%, respectively (fig. 1A). Their individual profiles were characterized by a sharp peak response within 38 s of agonist addition followed by a rapid return toward baseline with continuous agonist perfusion. Distinct increases in basal acidification rates were detected with these agonists with concentrations as low as ~100 pM (data not shown). In contrast, the potent and well-documented PMN agonists IL-8 (Grob et al., 1990; Thelen et al., 1988) and PAF (Shaw et al., 1981), which typically evoke strong in vitro PMN responses within 1 to 10 nM for IL-8 and 10 to 100 nM for PAF, proved to be only weak stimuli for inducing changes in EAR, as monitored by this methodology.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Characterization of ligand-operated EAR in human PMN. PMN extracellular acidification rates were measured by microphysiometry as described under "Methods." Rates (µV/s) were normalized to base line at t = 0 (100%), and traces represent separate experiments with mean values obtained from four independent chambers (~1.3 × 105 PMN/chamber) from three separate donors. All compounds were added at t = 4 (indicated by arrow). Chambers containing PMN were challenged with PMN ligands fMLP (50 nM) () or LTB4 (100 nM) () for 5 min (A), PAF (1 µM) () or IL-8 (100 nM) () for 15 min (B). PMN responses to these agonists were compared with other ligands; cells were challenged with LXA4 (100 nM) (), 16-phenoxy-LXA4 (100 nM) () or PGE2 (1 µM) () for 15 min (C).

LTB₄- and fMLP-stimulated EAR exhibits rapid desensitization and is Na⁺/H⁺ antiport dependent. Rapid homologous desensitization is a component of LTB₄ and fMLP interaction with plasma membrane receptors on PMN (Baggiolini et al., 1993).To determine whether ligand-operated EAR can also reflect homologous receptor desensitization, PMN were challenged in tandem with agonist and elicited EAR was quantified by microphysiometer. PMN were perfused for 5 min with LTB₄ (100 nM) or fMLP (50 nM) after EAR returned to base line. The representative on-line microphysiometer trace (fig. 2) demonstrates desensitization to a second or third challenge with agonist. These patterns were observed consistently with cells from three separate donors. The second agonist challenge resulted in a 69 ± 3% (fMLP) and 71 ± 3% (LTB₄) decrement, whereas a third challenge did not decrease this response further.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   fMLP- and LTB4-stimulated PMN EAR exhibits rapid desensitization and is Na+/H+ antiport-dependent. Representative on-line microphysiometer tracings of four independent chambers (~1.3 × 105 PMN/chamber) on exposure to agonists. PMN were treated with either vehicle alone (, ) or hexamethylene amiloride (HMA; 10 µM) (, ) for 60 min before agonist addition. Acidification rates (µV/s) were normalized to base line at t = 0. Chambers containing PMN were perfused with fMLP (50 nM) (, ) or LTB4 (100 nM) (, ) at t = 4 for 5 min. Homologous receptor desensitization was evaluated by repeated challenge of PMN with either LTB4 (100 nM, 5 min) or fMLP (50 nM, 5 min). Agonist addition at t = 4, t = 28 and t = 52 is indicated by arrows. These responses were observed with cells from three separate donors.

Inhibition of ligand-operated EAR by pertussis toxin. LTB₄ and fMLP mediate and respond through PTX-sensitive G-protein-coupled receptors (Bommakanti et al., 1995; Sumimoto et al., 1988). PMN were incubated with PTX (5 µg/ml) to determine whether agonist-induced EAR was mediated through ligand-specific PTX-sensitive receptors. Pertussis toxin inhibited the PMN response to LTB₄ (86 ± 4%) and fMLP (99 ± 1%) (fig. 3), thus establishing that the observed ligand-operated changes in EAR were indeed G-protein receptor-coupled.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Pertussis toxin inhibits LTB4- and fMLP-induced EAR in PMN. PMN were treated (120 min, 37°C) with PTX (5 µg/ml) (, ) or vehicle alone (, ), embedded in agar and placed in a microphysiometer. Extracellular acidification (µV/s) was normalized to base line at t = 0. Traces represent the mean of four independent chambers (~1.3 × 105 PMN/chamber) with PMN isolated from three separate donors. PMN were challenged with fMLP (50 nM) (, ) or LTB4 (100 nM) (, ) for 5 min at t = 4 (indicated by arrow) in the presence of PTX or vehicle.

LTB₄-elicited EAR reflects changes in ligand receptor interaction. PMN were perfused with LTB₄ analogs to determine whether elicited EAR profiles correlated with reported impact of these analogs on PMN responses. Carboxy terminus modification has a severe impact on the biological activity of LTB₄, namely, LTB₄-methyl ester reportedly is inactive as a chemotactic agent (Clancy et al., 1987), and LTB₄-dimethyl amide is a potent antagonist and a partial agonist at high concentrations (Falcone and Aharony, 1990; Showell et al., 1982). Both of these LTB₄ derivatives elicited rapid changes in PMN extracellular acidification that were distinct in both magnitude and profile from native LTB₄ (fig. 4). Maximal EAR induced by these LTB₄ derivatives were 66 ± 13% (LTB₄-dimethyl amide; fig. 4A) and 68 ± 12% (LTB₄-methyl ester; fig. 4B) less than that of LTB₄ (fig. 4C) at equimolar concentrations. Thus, these results indicate that changes in receptor ligand interaction (Clancy et al., 1987; Falcone and Aharony, 1990; Showell et al., 1982) brought about by alterations at the carboxylic end of LTB₄ also have an impact on ligand activation of EAR (fig. 4).

View larger version (17K): [in this window] [in a new window]   Fig. 4.   LTB4-elicited EAR profiles reflect changes in ligand receptor interaction. PMN (~1.3 × 105/chamber) were equilibrated in a microphysiometer, acidification rates (µV/s) normalized to base line at t = 0 and potential ligands were added at t = 4 (indicated by arrow). PMN were challenged with LTB4 analogs. (A) LTB4 dimethyl amide panel and (B) LTB4 methyl ester panel for 5 min. In parallel experiments, PMN were treated with GM-CSF (25 ng/ml) () or vehicle () for 60 min before LTB4 (100 nM) stimulation. Traces represent the mean of four independent chambers with cells isolated from three different donors.

Does LXA₄ inhibit PAF, fMLP and LTB₄ microphysiometer profiles or PMN O₂^-generation? LXA₄ and its analogs are potent modulators of PMN function (Serhan et al., 1995) and inhibit both LTB₄- and fMLP-elicited transmigration and chemotaxis in vitro (reviewed in Serhan, 1997) and in vivo (Takano et al., 1997). To determine whether PMN exposure to LXA₄ altered microphysiometer profiles obtained for pro-inflammatory ligands (fig. 1), cells were incubated (15 min) with LXA₄ (100 nM) and subsequently challenged with either PAF (1 µM), fMLP (50 nM) or LTB₄ (100 nM). Exposure to LXA₄ did not alter the ligand-operated EAR (n = 3, data not shown). This pattern also was seen when LXA₄ exposure was extended to 60 min or when concentrations were increased to 500 nM. Together these results demonstrate that pro-inflammatory ligands elicit changes in EAR that represent signal transduction events that are distinct from LXA₄ receptor-activated pathways.

Impact of mediators and eicosanoid analogs on HUVEC EAR. PMN adhesion and transmigration across endothelial cells are obligatory components of host defense as well as leukocyte-mediated tissue injury (Baggiolini et al., 1993; Ben-Baruch et al., 1995; Weissmann, 1989). HUVEC were used as a model to evaluate the impact of inflammatory mediators on endothelial cell EAR (fig. 5). Here too, endothelial cell responses demonstrate that EAR is mediated via ligand-specific signal transduction pathways. Histamine, a well-characterized endothelial cell agonist (Bull et al., 1992), elicited a small increase in EAR (30% ± 9) that was characterized by a slow and gradual return to base line during the 15-min agonist perfusion. On the other hand, 16-phenoxy-LXA₄ (100 nM), LXA₄ (100 nM), LTB₄ (100 nM), fMLP (100 nM) or iloprost (1 µM) [the prostaglandin I₂ stable analog and potent endothelial cell agonist (reviewed in Coleman et al., 1994)] did not stimulate EAR after 15 min perfusion, even though specific receptors that evoke functional responses on endothelial cells were characterized for both iloprost and LXA₄ on endothelial cells (reviewed in Serhan, 1997).

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Impact of mediators and analogs on HUVEC extracellular acidification. HUVEC were equilibrated in a microphysiometer, acidification rates (µV/s) normalized to base line at t = 0 and ligands were added at t = 4 (indicated by arrow). Chambers containing HUVEC were perfused with either histamine (1 µM) (), fMLP (100 nM) (), LTB4 (100 nM) (), iloprost (1 µM) (), 16-phenoxy-LXA4 (100 nM) () or LXA4 (100 nM) () for 15 min. Traces represents the mean of four independent chambers.

Impact of hypoxia/reoxygenation on PMN response to pro-inflammatory ligands. Changes in oxygen tension dramatically alter PMN functions, such as enhancing phagocytosis, increasing tyrosine kinase activity and altering expression of adhesion molecules (reviewed in West and Wilson, 1996). To determine whether these alterations extend to receptor-mediated signal transduction of selected pro-inflammatory mediators, PMN were equilibrated in hypoxic conditioned buffer (pO₂, 20 mm Hg) and exposed to hypoxia (2 hr) followed by reoxygenation (30 min). PMN responses to PAF, LTB₄ and fMLP were quantitated by monitoring changes in EAR (fig. 6). Reoxygenation did not alter basal PMN EAR (e.g., statistically significant differences were not observed between hypoxic and normoxic conditioned PMN) during the 30-min microphysiometer equilibration period. This finding agrees with the observation that PMN obtain all required ATP via anaerobic glycolysis (Karnovsky, 1968). Thus, hypoxia should have no impact on the metabolic state of PMN or EAR. Also, PMN exposed to normoxic or hypoxic conditions and challenged with either PAF (1 µM), LTB₄ (100 nM) or fMLP (50 nM) did not demonstrate statistically significant differences between the agonist profiles obtained for EAR both with respect to shape or magnitude (fig. 6). Thus, local changes in oxygen tension apparently do not have an impact on PMN EAR with these inflammatory mediators.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Hypoxia/reoxygenation does not alter PMN responses to pro-inflammatory ligands. Normoxic (, , ) or hypoxic (, , ) conditioned PMN (1.3 × 105/chamber) were equilibrated in a microphysiometer and acidification rates (µV/s) normalized to base line at t = 0. PMN were challenged with either LTB4 (100 nM, 5 min) (, ), fMLP (50 nM, 5 min) (, ) or PAF (1 µM, 10 min) (, ) at t = 4 (indicated by arrow). Traces represent the mean of four independent chambers with cells from three separate donors.

Inhibition of PMN transendothelial migration by 16-phenoxy-LXA₄ is enhanced after hypoxia/reoxygenation. Because the LXA₄ receptor activation was silent via microphysiometry in both PMN and EC, we sought evidence that these ligands were functionally active in parallel experiments with cells from the same donors. To this end, we determined whether changes in oxygen tension altered LTB₄-driven PMN transmigration, a potential functional endpoint for LTB₄, because we found no impact with PMN alone (fig. 6) and LTB₄ was indeed generated after ischemia/reperfusion (Goldman et al., 1992; Welbourn et al., 1991). PMN and recruitment and ensuing tissue injury partly depend on the enhanced biosynthesis of the potent chemotactic factor LTB₄ that follows hypoxia/reoxygenation (Goldman et al., 1992; Welbourn et al., 1991). Transendothelial migration was analyzed by exposing HUVEC and/or PMN to hypoxia for 24 hr and 2 hr, respectively. Of particular interest, PMN transmigration in response to a gradient of LTB₄ (10 nM) was also not significantly altered here following exposure to a hypoxic environment for either PMN or HUVEC (fig. 7, panel A).

View larger version (32K): [in this window] [in a new window]   Fig. 7.   Hypoxia/reoxygenation enhances LXA4 inhibition of PMN transendothelial migration. PMN-endothelial interaction under hypoxia/reoxygenation conditions was evaluated in a PMN transendothelial migration assay. PMN (2 hr) and monolayers of HUVEC (24 hr) were normoxic or hypoxic conditioned (see "Methods"). Normoxic () or hypoxic () conditioned PMN were treated with indicated concentrations of 16-phenoxy-LXA4 (B, C) or vehicle alone (A) for 30 min. Treated PMN (1 × 106 cells/40 µl) were added to the upper compartment of hypoxic or normoxic conditioned endothelial monolayers and transmigration was initiated by the addition of LTB4 (10 nM) to the lower compartment. After 90 min of transmigration, PMN MPO activity in the lower reservoir was assayed and is expressed as percent of total PMN that traversed the endothelial monolayer (A) or as percent inhibition compared to vehicle alone (B, C). The values are the means of four experiments with PMN from four separate donors. Statistically significant differences (P < .05) between treatments are indicated by symbols.

	Discussion

Top Abstract Introduction Methods Results Discussion References

These are the first results documenting ligand-operated EAR for lipid mediators in PMN and endothelial cells (HUVEC) with microphysiometry. Despite some common components in their respective signal transduction pathways, both PMN and HUVEC specific ligand-activated seven-transmembrane receptors evoked highly unique profiles of EAR (figs. 1-5) that do not necessarily correlate with their known and previously reported specific cellular functions. In addition, ligand-specific responses to LTB₄, fMLP and PAF were not altered when PMN were subject to hypoxia/reoxygenation designed to mimic the in vivo scenario. In contrast, hypoxia/reoxygenation enhanced the inhibitory responsiveness of PMN and endothelial cell coincubations to a metabolically stable LXA₄ analog (16-phenoxy-LXA₄).

Microphysiometer analysis of ligand-operated EAR demonstrated unique profiles for each of the pro-inflammatory mediators examined with PMN, namely, LTB₄, fMLP, IL-8 and PAF. These results suggest that, upon activation, each respective seven-transmembrane receptor for these ligands evokes specific signal transduction pathways that are reflected in the individual character of the EAR response. Evidence that the observed changes in basal EAR are mediated by G-protein-coupled receptors is two-fold (figs. 1-3). First, fMLP and LTB₄ responses demonstrate rapid homologous desensitization and second, ligand-stimulated EAR proved to be pertussis toxin-sensitive.

Common intracellular components of LTB₄, fMLP, IL-8 or PAF individual receptor signal transduction pathways do not appear to mediate ligand-operated EAR directly (fig. 1). This conclusion is supported by the finding that ligand-stimulated EAR showed dramatic differences among the mediators (ligands) both in terms of magnitude and profile (fig. 1). Of interest, LTB₄, fMLP, PAF and IL-8 are each well-characterized and potent activators of intracellular signal transduction pathways (reviewed in Ben-Baruch et al., 1995). Yet, differences between ligand responses for EAR become most apparent when we consider the results obtained with the chemokine IL-8 (fig. 1B). For instance, IL-8 is 10 to 100 times more potent than fMLP in mobilizing Ca⁺⁺ (Grob et al., 1990; Thelen et al., 1988), but it was only a weak agonist for evoking EAR. Along these lines, comparing the relatively weak EAR triggered by the calcium ionophore A23187 (144% increase) to another potent PMN agonist, fMLP (278% increase), provides further evidence that intracellular Ca⁺⁺ mobilization and ligand-stimulated EAR do not appear to be linked events in PMN. However, ligand-operated EAR clearly do reflect specific changes in certain receptor activation cascades as demonstrated by the impact of modifying the carboxy terminus of LTB₄ as well as results obtained with the cytokine GM-CSF (fig. 4). Each LTB₄ analog tested evoked a unique EAR profile, whereas cytokine exposure and subsequent LTB₄ challenge gave an attenuated but characteristic LTB₄-induced EAR response.

PMN receptors are cloned and sequenced for PGE₂ (Armstrong, 1995; Coleman et al., 1994) and LXA₄ (reviewed in Serhan, 1997) and, like the chemoattractant receptors including fMLP, PAF, LTB₄ or IL-8, they belong to the seven-transmembrane receptor superfamily. Both of these eicosanoids, LXA₄ and PGE₂, are potent in vitro inhibitors of fMLP-stimulated chemotaxis (Lee et al., 1991). It is of particular interest that these ligands, as well as the stable LXA₄ analog 16-phenoxy-LXA₄, did not evoke detectable changes in EAR (fig. 1C). PGE₂ and LXA₄ act via different receptors, but PGE₂ enhances LTB₄ actions in vivo whereas LXA₄ and its analogs inhibit PMN infiltration in vivo (Takano et al., 1997) in keeping with in vitro findings. LXA₄ does not compete with or alter either fMLP or LTB₄ binding at 4°C and binds to its own specific receptor to inhibit PMN functional responses (reviewed in Serhan, 1997). LXA₄ suppresses Ca⁺⁺ mobilization (Lee et al., 1989) mediated by both LTB₄ and fMLP. In addition, LXA₄ inhibits PAF-stimulated chemotaxis in eosinophils (Soyombo et al., 1994), another cell type with the well-characterized serpentine receptor specific for PAF (reviewed in Izumi and Shimizu, 1995). Thus, our findings with LXA₄ provide evidence that its inhibitory actions apparently are confined to motility in leukocytes that is mediated via a Cytosensor silent receptor, a property also shared by PGE₂ receptor on these cells.

When viewed together these results indicate that serpentine receptors of inflammatory ligands activate unique signaling pathways that are paralleled in their specific EAR profile and do not correlate directly to previously identified signal transduction pathways. This conclusion is supported with the observation that protons that are generated in response to signal transduction pathways (i.e., cAMP, phosphatidyl inositol or diacylglycerol generation and the activity of protein kinases) represents only a small component of the typically large change in EAR observed after receptor activation (reviewed in Owicki and Parce, 1992). Thus, ion channel opening (i.e., Na⁺/H⁺ exchange) after receptor activation most likely represents the largest component of the observed rapid and transient change in EAR. It is therefore likely that, upon activating receptors for LXA₄ and PGE₂, intracellular signals are generated that do not activate ion channels and therefore do not induce detectable changes in EAR, yet clearly inhibit PMN transmigration in vitro (fig. 7). This also proved to be the case for LTB₄, LXA₄ and PGI₂ with endothelial cells. LTB₄ recently has been found to evoke receptor-dependent actions on EC, as do both LXA₄ and PGI₂ (Nohgawa et al., 1997).

Evaluation of endothelial cells (HUVEC) also revealed a similar pattern of ligand-selective EAR for lipid-derived mediators in that endothelial receptors also have been identified and biological activity characterized for LXA₄ (Papayianni et al., 1996), prostacyclin (reviewed in Coleman et al., 1994), LTB₄ (Nohgawa et al., 1997) and histamine (Bull et al., 1992). However, neither the potent prostacyclin analog iloprost, LTB₄, LXA₄, nor its stable analog elicited demonstrable changes in EAR. Histamine, on the other hand, stimulated EAR (fig. 5).

PMN play an important role in reperfusion injury, marked by recruitment, adhesion of PMN to endothelial cells and the generation of PMN-derived cytotoxic agents and lipid mediators. In this context, it is of interest that nitric oxide is a signal that is generated in lung tissues when exposed to LXA₄ (Tamaoki et al., 1995). Changes in oxygen levels profoundly alter cell function and certain signal transduction pathways (Welbourn et al., 1991; West and Wilson, 1996). In the present experiments, PMN that were exposed to intervals of hypoxia/reoxygenation in an in vitro environment gave essentially identical EAR to LTB₄, fMLP and PAF (fig. 6), which suggests that their individual pro-inflammatory signal transduction pathways remained responsive. These findings were reinforced by evaluating PMN functional responses, namely PMN transendothelial migration. Results from these experiments clearly demonstrated that LTB₄-stimulated PMN migration was neither enhanced nor diminished by hypoxic treatment of PMN and/or endothelial cells (fig. 7B). Periods of hypoxia/reoxygenation can activate HUVEC (Arnould et al., 1994), and via enhanced expression of selectins and intracellular adhesion molecule-1, they become an adhesive surface for PMN (Arnould et al., 1995; Shreeniwas et al., 1992). Hypoxia/reoxygenation, however, did not alter PMN transmigration (fig. 7A). It appears that the impact of hypoxia/reoxygenation on PMN adhesion and transendothelial migration is likely to be at independent loci that modulate PMN trafficking (see above). Taken together, these findings demonstrate that hypoxia/reoxygenation does not enhance or diminish pro-inflammatory signal transduction-activated pathways as monitored by PMN functional responses, namely EAR and transendothelial migration.

It was of interest in view of these results to determine the impact of hypoxia/reoxygenation on actions associated with lipoxin A₄ that represent counter-regulatory or migration inhibitory actions (reviewed in Serhan, 1997). 16-phenoxy-LXA₄ inhibition of PMN transendothelial migration was significantly augmented by hypoxia/reoxygenation. The enhanced inhibition of transmigration was most pronounced when PMN were first exposed to hypoxia/reoxygenation before coincubation with endothelial cells, which gave a shift in the IC₅₀ of ~ 2 log orders of magnitude (fig. 7, B and C). Thus, hypoxia/reoxygenation enhances protective actions of LXA₄, possibly as a potential means to dampen the pro-inflammatory milieu that is initiated in response to hypoxia/reoxygenation.

Lipoxin inhibitory action on PMN function encompasses blocking of -integrin-mediated adhesion and an implication for tyrosine kinase (Papayianni et al., 1996). In addition, LXA₄ exhibits potent actions on HUVEC (Papayianni et al., 1996) at concentrations of 1 nM to 1 µM. These bioactions with vascular endothelial cells include induction of nitric oxide formation (Tamaoki et al., 1995) and inhibition of P-selectin up-regulation mediated in part by prostacyclin release (Brezinski et al., 1989; Papayianni et al., 1996). In this context, changes in oxygen tension increase tyrosine kinase activity in PMN (reviewed in West and Wilson, 1996) and stimulate stress-activated kinases (JNKs/SAPKs) as well as mitogen-activated kinases in cardiac myocytes (Laderoute and Webster, 1997). Also, recent studies report increased expression of cyclooxygenase II in HUVEC after hypoxia/reoxygenation in vitro (Schmedtje et al., 1997). Thus, the impact of hypoxia/reoxygenation on certain signal transduction pathways, such as altered redox potentials (Welbourn et al., 1991; West and Wilson, 1996), and more specifically on increased cyclooxygenase activity and nitric oxide formation, might provide a framework for a cellular basis underlying the protective actions of 16-phenoxy-LXA₄ (fig. 7). Since neutrophil-mediated injury is an essential component in the sequelae of ischemia-reperfusion, the present in vitro findings suggest that LXA₄ stable analogs may be useful in limiting PMN-mediated damage in this arena since hypoxia followed by an interval of reoxygenation enhanced LXA₄ activity.

In summary, this report demonstrates that ligand-operated EAR reflects unique and receptor-specific signal transduction pathways in both PMN and endothelial cells. The Cytosensor is a useful tool to evaluate potential antagonist and receptor candidates for human neutrophil lipid ligands such as LTB₄ (figs. 1-3), whose receptor identification at the molecular level recently has been reported (Yokomizo et al., 1997). Moreover, these in vitro findings suggest a potential beneficial role for both endogenous LXA₄ and its synthetic stable LXA₄ analogs in attenuating PMN-mediated events in reperfusion injury.