Journal of Applied Physiology
Vol. 82, No. 3, pp. 866-873, March 1997
CELLULAR ASPECTS OF LUNG FUNCTION

XO increases neutrophil adherence to endothelial cells by a dual ICAM-1 and P-selectin-mediated mechanism

Lance S. Terada, Brooks M. Hybertson, Kevin G. Connelly, David Weill, Dale Piermattei, and John E. Repine

Webb-Waring Institute for Biomedical Research and Department of Medicine, University of Colorado Health Sciences Center, Denver, Colorado 80262

ABSTRACT

INTRODUCTION

MATERIALS AND METHODS

RESULTS

DISCUSSION

ACKNOWLEDGEMENTS

FOOTNOTES

REFERENCES

ABSTRACT

Terada, Lance S., Brooks M. Hybertson, Kevin G. Connelly, David Weill, Dale Piermattei, and John E. Repine. XO increases neutrophil adherence to endothelial cells by a dual ICAM-1 and P-selectin-mediated mechanism. J. Appl. Physiol. 82(3): 866-873, 1997.Circulating xanthine oxidase (XO) can modify adhesive interactions between neutrophils and the vascular endothelium, although the mechanisms underlying this effect are not clear. We found that treatment with XO of bovine pulmonary artery endothelial cells (EC), but not neutrophils or plasma, increased adherence, suggesting that XO had its primary effect on EC. The mechanism by which XO increased neutrophil adherence to EC involved binding of XO to EC and production of H₂O₂. XO also increased platelet-activating factor production by EC by a H₂O₂-dependent mechanism. Similarly, the platelet-activating factor-receptor antagonist WEB-2086 completely blocked XO-mediated neutrophil EC adherence. In addition, neutrophil adherence was dependent on the ₂-integrin Mac-1 (CD11b/CD18) but not on leukocyte functional antigen-1 (CD11a/CD18). Treatment of EC with XO for 30 min did not alter intercellular adhesion molecule-1 surface expression but increased expression of P-selectin and release of von Willibrand factor. Antibodies against P-selectin (CD62) did not affect XO-mediated neutrophil adherence under static conditions but decreased both rolling and firm adhesive interactions under conditions of shear. We conclude that extracellular XO associates with the endothelium and promotes neutrophil-endothelial cell interactions through dual intercellular adhesion molecule-1 and P-selectin ligation, by a mechanism that involves platelet-activating factor and H₂O₂ as intermediates.

xanthine oxidase; platelet activating factor; CD18; CD11b; Mac-1; CD11a; leukocyte functional antigen-1; heparin; hydrogen peroxide; multiorgan failure; acute respiratory distress syndrome

INTRODUCTION

MULTIORGAN FAILURE IS A SYNDROME characterized by the dissemination of inflammation from one organ to another, prompting its more recent moniker, systemic inflammatory response syndrome. The hallmark of this disorder is the attachment of neutrophils to endothelium in a number of organs. Indeed, a localized primary injury causes neutrophil recruitment and inflammation in a secondary organ in a number of experimental models. For instance, reperfusion injury to hindlimbs (14) or intestines (33) produces an inflammatory lung injury, and focal injury to one lung causes inflammatory changes in the opposite lung (3).

Although the mechanism by which the secondarily affected organ initially retains neutrophils is not clearly understood, the prooxidant enzyme xanthine oxidase (XO) may be pivotal. Endogenous XO participates in postischemic neutrophil adherence in vivo (10) and reoxygenation-induced neutrophil adherence in vitro (34). In addition, circulating levels of XO are elevated in patients with acute lung injury (11) and after limb ischemia (8), and infused XO increases neutrophil adherence to rat mesenteric venules (9). We have recently shown that circulating XO increases and mediates neutrophil retention in lungs after intestinal ischemia (33), highlighting the potential role of XO in initiating distant organ inflammation. Accordingly, the goal of this study was to identify pathways involved in extracellular XO-mediated neutrophil adherence to endothelial cells (EC).

MATERIALS AND METHODS

Source of reagents. Platelet-activating factor (PAF; from bovine heart lecithin), superoxide dismutase (SOD; bovine erythrocyte, 3,000 U/mg) and XO (grade III from bovine milk, 1.2 U/mg) were obtained from Sigma Chemical (St. Louis, MO). Blocking monoclonal antibodies against CD11a and intercellular adhesion molecule-1 (ICAM-1; CD54) were from AMAC (Westbrook, ME). Blocking monoclonals against CD11b were from Sigma Chemical, and blocking monoclonals against P-selectin (CD62) were from Monosan. Each antibody was of subtype immunoglobulin G₁ (IgG₁), and, therefore, a mouse IgG₁ isotype control (derived from the MOPC-21 tumor line) was obtained from Sigma Chemical. Unconjugated and peroxidase-conjugated rabbit polyclonal antibodies to human von Willibrand factor (vWF) was from Dako (Carpinteria, CA). WEB-2086 was a kind gift of Boehringer Ingelheim (Ridgefield, CT). Heparin (porcine intestinal mucosa) was purchased from SoloPak Laboratories (Elk Grove Village, IL). 4-Hydroxy-2,2,6,6-tetramethylpiperidinyloxy (Tempol) was obtained from Aldrich (Milwaukee, WI). Rat interferon (IFN) -/ (3.5 × 10⁶ IU/mg) was obtained from Lee Bio Molecular (San Diego, CA). Catalase (bovine liver, 81,536 U/mg) was purchased from Worthington Biochemical (Freehold, NJ). [³H]acetate (>500 mCi/mmol) was from New England Nuclear (Boston, MA). Hypoxanthine (HX) and all other reagents were obtained from Sigma Chemical.

Endothelial cell culture. Bovine pulmonary artery EC were harvested by using collagenase digestion (34). After two passages in D-valine (Sigma Chemical) to minimize smooth muscle contamination, EC were cultured in Eagle's minimum essential medium with 10% fetal calf serum and studied after three to five passages. Cultures were found to be >99.5% EC by vWF and low-density lipoprotein receptor staining.

Static neutrophil adherence. Neutrophils were isolated from healthy human donors and labeled with ⁵¹Cr, as previously described (34). EC were plated in 24-well plates and uniformly used at 1-2 days after reaching confluence. EC were washed twice with Hanks' balanced salt solution (HBSS) and incubated with various combinations of XO (10 mU/ml), HX (200 µM), heparin (50 U/ml), catalase (815 U/ml), SOD (30 U/ml), Tempol (5 mM), WEB-2086 (1 mM), reagent PAF (1 µM), or monoclonal antibodies (4 µg/ml), in HBSS, for 30 min at 37°C. In some experiments, EC were incubated with various agents in 50% pooled normal human plasma. EC were then washed twice with HBSS, and unstimulated polymorphonuclear leukocytes (PMN; 10⁶/well) were gently added. In some experiments, WEB-2086 was added to EC and PMN. After a 30-min incubation at 37°C, nonadherent neutrophils were collected and pooled with the first wash. Residual ⁵¹Cr in adherent neutrophils was released with NaOH and quantified by liquid scintigraphy. Adherence was defined as the percentage of counts per minute (cpm) in the media and first wash, relative to total cpm in media per wash plus EC (cpm media + first wash/cpm media + first wash + EC releasate). Visual inspection under all conditions revealed persistence of EC monolayers before NaOH lysis. In some experiments, monolayers were first pretreated with XO and/or heparin for 60 min at 37°C then triply washed and incubated with neutrophils in the presence of HX. In some experiments, neutrophils were fixed with paraformaldehyde (1%) for 10 min at 25°C after ⁵¹Cr loading.

PAF assay. EC monolayers in 25-cm² flasks were triply washed to remove serum, then incubated with 25 µCi ³H-acetate in 1 ml HBSS with 10 mM N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid (21) in the presence or absence of XO (10 mU/ml), HX (200 µM), SOD (30 U/ml), or catalase (815 U/ml) for up to 30 min. Without aspirating the media, the reaction was stopped with 0.5 ml of 50 mM acetic acid in MeOH, and EC were scraped off, pooled with a subsequent 2-ml acetic acid/MeOH wash, and added to 1.25 ml CHCl₃. Labeled PAF was extracted (4) with 10 µg cold PAF as carrier by using 1.25 ml CHCl₃ and 1.25 ml of 0.1 M sodium acetate. The lower phase was dried under N₂ and resuspended in 100 µl CHCl₃/MeOH (9:1). An aliquot was removed for assessment of total [³H]acetate incorporation, and the remainder was separated by thin-layer chromatography (silica gel 60, Merck, Darmstadt, Germany) by using CHCl₃/MeOH/glacial acetic acid/H₂O (50:25:8:4). Lipids were visualized with iodine vapor, and lanes were scraped off in fractions by using the mobility of authentic PAF as a guide. After quantification of ³H in these fractions with liquid scintigraphy, the amount of [³H]acetate incorporated into PAF was calculated (36).

vWF release and P-selectin and ICAM-1 surface expression. EC were grown to confluence in 96-well enzyme-linked immunosorbent assay (ELISA) plates and exposed to XO (10 mU/ml) and HX (200 µM) in HBSS for 30 min. Control EC were exposed to IFN-/ (24,000 U/ml) for 24 h to increase ICAM-1 expression. EC were then washed twice with HBSS and blocked with 2% bovine serum albumin (BSA) in HBSS at 25°C for 30 min. After two subsequent washes, EC were incubated with either anti-P-selectin or anti-ICAM-1 (1:500 in 0.1% BSA) for 30 min at 37°C, washed twice, then incubated with rabbit anti-mouse IgG-peroxidase (1:1,000 in 0.1% BSA) for 20 min at 25°C. After two subsequent washes, the plates were developed with o-phenylenediamine dihydrochloride (OPD). Baseline values (EC treated with OPD in the absence of antibodies) were subtracted from sample values to account for the low levels of endogenous EC peroxidases. vWF released into the media was assessed by ELISA as previously described (13).

Dynamic neutrophil adherence. EC were passaged into gelatin-coated glass capillary tubes (1.1-mm ID, Scientific Manufacturing Industries, Emeryville, CA). After attachment of EC, medium was changed once, and EC were allowed to grow overnight to form confluent monolayers on approximately one-half of the internal surface of the tube. After treatment of EC with various agents, the capillary tubes were secured on the stage of an inverted microscope with EC in the dependent position. Neutrophils (10⁶/ml) were perfused through the tube with a syringe pump at a constant rate (0.173 ml/min). A second syringe pump infused HBSS + 5% fetal calf serum at a variable rate. From the measured total flow rate and the cylindrical geometry of the tube, the shear at the surface of the tube was calculated by using the Hagen-Poiseuille equation (22). The interaction of neutrophils with EC was recorded on videocassette recorder (Javelin recorder) for later playback analysis. The number of rolling neutrophils crossing a standardized 250-µm bar, neutrophil rolling velocity (average of 6-12 neutrophils), and number of firmly adherent neutrophils (no movement for at least 30 s) per 0.0625 mm² were determined for three random fields per tube.

RESULTS

Effect of XO on neutrophil adherence. Exposure of EC and neutrophils to XO (5-30 mU/ml) and substrate (HX) for 10-30 min increased (P < 0.01) neutrophil adherence to EC monolayers (Figs. 1 and 2). Addition of substrate alone did not affect (P > 0.05) neutrophil adherence. Pretreatment of EC, but not neutrophils, with XO and HX increased (P < 0.001) neutrophil adherence to EC monolayers (Fig. 3). Addition of plasma pretreated with XO and HX had no effect on neutrophil adherence (not shown). In another experiment, conditioned media from HX/XO-treated EC were added to untreated EC monolayers. Adherence of neutrophils was higher (P < 0.01) to HX/XO-treated EC than to EC exposed to HX/XO-EC conditioned media (control 3.09 ± 0.25%; HX/XO 8.13 ± 0.25%; conditioned media 4.27 ± 0.25%).

Effect of heparin on neutrophil adherence. The ability of EC-bound XO to modify neutrophil adherence was assessed by first preincubating monolayers with XO in the absence of substrate. After removal of unassociated XO by extensive washing, HX was then added and resulted in increased (P < 0.001) neutrophil adherence compared with EC not preincubated with XO (Fig. 4). Heparin, which interferes with binding of XO to cell surfaces (1, 31), decreased (P < 0.001) neutrophil adherence when added with XO to EC in the preincubation phase (Fig. 4). In contrast, heparin did not affect (P > 0.05) neutrophil adherence when added concurrently with both XO and HX in HBSS. Heparin alone had no effect (P > 0.05) on neutrophil adherence, compared with baseline. In addition, heparin had no effect on phorbol myristate acetate-mediated neutrophil adherence (PMA, 46 ± 3%; PMA + heparin, 41 ± 4%, P > 0.05). To investigate the importance of XO binding to EC in a more relevant antioxidant-replete milieu, we exposed EC to HX and XO in 50% human plasma. HX and XO increased neutrophil adherence (P < 0.001) in 50% plasma (Fig. 5). In contrast to the experiment performed in HBSS alone, heparin decreased neutrophil adherence when added concurrently with HX and XO in 50% plasma (P < 0.001). Again, heparin had no effect (P > 0.05) on baseline adherence.

1

1

1

1

1

DISCUSSION

We recently found that after mesenteric ischemia plasma levels of XO, an efficient enzymatic source of reactive oxygen intermediates, increased and mediated retention of neutrophils in the lung parenchyma (33). The ability of circulating XO to increase neutrophil-EC interactions provides an interesting mechanism by which a primarily injured organ may initiate inflammation in other vascular beds. However, the literature provides conflicting paradigms of the mechanism by which prooxidants may promote retention of neutrophils in inflammed tissues. For instance, high levels (10 mM) of H₂O₂ were found to rapidly incite PAF production by EC (21), whereas PAF did not appear to be responsible for neutrophil adherence when lower, more physiological doses of H₂O₂ were employed (27). In addition, conclusions diverge regarding whether H₂O₂ mediates neutrophil adherence exclusively by P-selectin- (27) or ICAM-1-related mechanisms (23). Similarly, conflicting data exist regarding the mechanism by which XO alters neutrophil adherence (9, 29).

We found evidence to suggest that extracellular XO promotes neutrophil adherence to EC by a mechanism involving the production of PAF by EC and the binding of both ICAM-1 and P-selectin to neutrophil ligands. We found first that XO increased neutrophil adherence in concentrations we have previously found in plasma to affect lung neutrophil retention in vivo (33). The effect was rapid, occurring within 10 min, and did not require de novo protein synthesis, precluding involvement of nonconstitutive EC ligands such as E-selectin. Curiously, neutrophil adherence decreased to baseline levels by 40 min. Although the reason for this time-dependent decrease in adherence is not clear, it may relate to a decrease in EC PAF levels, as occurs after stimulation of EC with bradykinin (35) or thrombin (28). Alternatively, ₂-integrin expression may also decrease despite continued neutrophil stimulation (24).

The effect of XO-derived oxidants appeared to be primarily on EC, since treatment of EC, but not neutrophils, increased adherence. We did not find evidence for formation of plasma-derived factors to mediate this effect. The direct effect of XO on EC is particularly relevant, since XO binds to anionic EC surface moieties both in vitro and in vivo in a heparin-reversible manner (1, 31). Our data are consistent with these prior studies and further suggest that EC-bound XO can activate neutrophil adherence mechanisms. This close approximation of XO to EC may be particularly germane in vivo, as a means of bypassing the antioxidant-rich milieu of blood. We found evidence for this effect when EC were exposed to XO in 50% plasma. Human plasma is rich in oxidant-scavenging activity (20) and could potentially dampen the effect of XO on EC. In 50% plasma, however, XO continued to mediate an increase in neutrophil adherence, consistent with a site-specific mode of action of XO at the cell surface. In this environment, as opposed to HBSS, heparin decreased neutrophil adherence when added concurrently with XO and HX, providing further evidence that displacement of XO from the surface of EC allows plasma scavengers to intervene and diminish the effects of delocalized XO.

H₂O₂ was primarily involved in mediating static neutrophil adherence, since catalase, but not SOD or the membrane-permeable SOD mimic Tempol, decreased adherence. It is not clear why some studies of reperfusion implicate XO-derived superoxide anion radical rather than H₂O₂ in mediating adherence (10, 34), although in such studies endogenous, presumably intracellular, XO is likely the source of oxygen metabolites. In the present situation with exogenous XO, the diffusion of superoxide anion radical into EC from external sites through anion channels may limit its participation, or the specific targets may differ.

We also found evidence for PAF as an EC-derived intermediate in neutrophil adherence. First, incorporation of [³H]acetate into PAF increased after stimulation with XO, and this effect was also diminished by catalase but not SOD. Both PAF synthesis and neutrophil adherence increased within 10 min. In other systems, oxidative stress such as ischemia-reperfusion (15) or high concentrations of H₂O₂ (21) have been shown to increase PAF levels. Although the mechanism is not clear, it is noteworthy that treatment of some cells with XO increases cytosolic free calcium (7) and phospholipase A₂ activity (37), which may facilitate PAF production. Second, the PAF-receptor antagonist WEB-2086 completely blocked XO-mediated adherence and, as expected, addition of reagent PAF increased neutrophil adherence. It is likely that EC-derived PAF activates neutrophils directly, since WEB-2086 blocked adherence only when added with neutrophils and not when used to pretreat EC; however, the additional role of PAF in transducing EC signals in an autocrine fashion has not been completely excluded. In addition, PAF may increase neutrophil ₂-integrin expression by undetermined EC or neutrophil intermediates. Our results differ from those of Sellak et al. (29), who found that PAF-receptor antagonism did not diminish XO-stimulated neutrophil adherence to EC. It is unclear what accounts for this discrepancy, although this latter group examined a somewhat shorter treatment interval and employed a different PAF antagonist (BN-52021). PAF levels were not measured in this latter study, so it is unclear whether PAF production by EC increased. Finally, it is likely that PAF remained largely associated with EC, since conditioned media from HX/XO-treated EC increased neutrophil adherence to naive EC monolayers only slightly.

Our data also implicate Mac-1 rather than LFA-1 in mediating XO-dependent adherence to endothelial ICAM-1. This pattern of ₂-integrin involvement is consistent with the proposed role of PAF in activating neutrophils, since PAF increases the surface expression of Mac-1 in neutrophils (5, 26). It is possible that stimulation of neutrophils with chemotaxins like PAF promotes Mac-1- rather than LFA-1-dependent adherence (2, 30). Importantly, Mac-1 expression on neutrophils is increased in acute respiratory distress syndrome (18), a condition associated with high circulating levels of XO (11). Furthermore, monoclonals directed against the common ₂-integrin subunit CD18 decrease lung injury after intestinal ischemia-reperfusion (12), another condition associated with elevated levels of circulating XO (33), and also decrease neutrophil adherence after intravascular infusion of purified XO (9).

An important endothelial ligand for Mac-1 in the present investigation appears to be constitutive ICAM-1, since inhibition of protein synthesis did not affect adherence, and surface expression of ICAM-1 did not increase after XO treatment. Although ICAM-1 expression has been reported to increase as early as 30 min after exposure of EC to high levels of H₂O₂ (23), most reports suggest a time frame of 6-24 h after oxidative stress for induction of surface ICAM-1 expression to occur (6, 17).

Besides ICAM-1, an additional endothelial ligand responsible for XO-mediated neutrophil adherence appears to be P-selectin, whose expression increased after exposure to XO. A dynamic system was required to demonstrate an effect of the monoclonal antibody against P-selectin, suggesting that the integrin-mediated adherence dominated the effect seen in the static assay. This may explain the apparent discrepancy between studies that have investigated the role of P-selectin using static (29) vs. dynamic (9) systems. The cooperativity between the selectin and integrin adherence systems was more evident at higher shear rates, where interference with P-selectin markedly suppressed firm adherence. This is consistent with the notion that, under higher levels of shear, rolling is a prerequisite for firm adherence (19). In contrast, XO-mediated firm adherence was not significantly reduced by the antibody to P-selectin at the lowest shear rate tested. Indeed, observations in vivo suggest that ICAM-1-mediated firm adherence can occur at low shear rates even with effective blockade of P-selectin sites (16). A potential limitation of this study is the use of bovine rather than human EC. However, our data suggest that human neutrophils appear capable of interacting with bovine EC via both integrin and selectin ligand pairs.

In summary, extracellular XO can associate closely with EC and stimulate adherence of neutrophils through interactions with endothelial ICAM-1 and P-selectin. We suggest that circulating XO may contribute to the initial recruitment of neutrophils to secondarily affected vascular beds.

ACKNOWLEDGEMENTS

This work was supported by the American Heart Association and National Heart, Lung, and Blood Institute Grants R29-HL-52591, R01-HL-5582, and P50-HL-0784. L. S. Terada is an Established Investigator of the American Heart Association, and B. M. Hybertson is a fellow of the Parker B. Francis Foundation.

FOOTNOTES

Address for reprint requests: L. S. Terada, Univ. of Colorado Health Sciences Center, Box C322, 4200 E. Ninth Ave., Denver, CO 80262.

Received 2 April 1996; accepted in final form 20 September 1996.

REFERENCES