Inhibition of Selectin-mediated Cell Adhesion and Prevention of Acute Inflammation by Nonanticoagulant Sulfated Saccharides

Selectins play a major role in the inflammatory reaction by initiating neutrophil attachment to activated vascular endothelium. Some heparin preparations can interact with L- and P-selectin; however, the determinants required for inhibiting selectin-mediated cell adhesion have not yet been characterized. We now report that carboxyl-reduced and sulfated heparin (prepared by chemical modifications of porcine intestinal mucosal heparin leading to the replacement of carboxylates by O-sulfate groups) and trestatin A sulfate (obtained by sulfation of trestatin A, a non-uronic pseudo-nonasaccharide extracted from Streptomyces dimorphogenes) exhibit strong anti-P-selectin and anti-L-selectin activity while lacking antithrombin-mediated anticoagulant activity.In vitro experiments revealed that both compounds inhibited P-selectin- and L-selectin-mediated cell adhesion under laminar flow conditions. Moreover, carboxyl-reduced and sulfated heparin and trestatin A sulfate were also active in vivo, as assessed by experiments showing 1) that microinfusion of trestatin A sulfate reduced by 96% leukocyte rolling along rat mesenteric postcapillary venules and 2) that both compounds inhibited (by 58–81%) neutrophil migration into thioglycollate-inflamed peritoneum of BALB/c mice. These results indicate that nonanticoagulant sulfated saccharides targeted at P-selectin and L-selectin may have therapeutic potential in inflammatory disorders.

Leukocyte migration in inflammatory lesions is a reaction that is sequentially regulated by adhesion receptors and inflammation products. Selectins play a major role in initiating neutrophil attachment to cytokine-activated endothelium. Lselectin is expressed by most circulating leukocytes; E-selectin expression is induced after several hours of endothelial cell activation by interleukin-1, tumor necrosis factor-␣, or endotoxin; P-selectin is rapidly expressed by endothelial cells or platelets exposed to thrombin or histamine (1)(2)(3)(4)(5)(6).
Several studies have indicated that heparan sulfate and heparin are ligands for L-selectin and P-selectin. Heparan sulfate proteoglycans isolated from calf pulmonary endothelial cells or kidney can interact with L-selectin (40 -42, 45, 46). Moreover, heparin or heparin-like oligosaccharides can inhibit L-selectin or P-selectin binding to sLex-related compounds or HL-60 cells (46,47). Heparin is a glycosaminoglycan composed of alternating D-glucosamine and uronic acid (L-iduronic or D-glucuronic acid) residues that are heterogeneous in size and degree of sulfation. The heparin anticoagulant effect is caused by antithrombin activation, a reaction that follows a conformational change that takes place when the serpin binds to a specific pentasaccharide sequence. In contrast, little information is available on the structural determinants required for L-selectin and P-selectin binding. The beneficial effects of selectin inhibitors in animal models of ischemia/reperfusion injury (for example in models of myocardial infarction, stroke, traumatic shock or solid organ transplantation) suggest that heparin or heparin-like compounds could be helpful in inhibiting selectin function and preventing tissue damage (48 -54). However, heparin anticoagulant properties and the potential of bleeding complications may contraindicate its use as an antiadhesive compound. In this study, we have examined whether the antithrombin-mediated anticoagulant activity of heparin could be separated from its anti-selectin activity. Two compounds with high anti-L-selectin and anti-P-selectin activity but negligible anticoagulant activity were identified including a chemically modified (carboxyl-reduced and sulfated) form of heparin and the non-uronic pseudo-nonasaccharide trestatin A sulfate. Furthermore, the anti-selectin activities of these compounds were compared with those of three unfractionated (UFH) and three low molecular weight heparin (LMWH) preparations approved for anticoagulant therapy in clinical practice.
Heparin and Oligosaccharide Preparations-The modified heparins and sulfated oligosaccharides used in this study showed nonparallel dose-response curves with the International Standard for Heparin. Therefore, the anticoagulant activities of these compounds were not expressed in IU of heparin, but characterized by the IC 50 indicating the concentration of the compound leading to a clotting time of twice the control (56). Three heparin preparations were obtained from Hoffmann-La Roche. 1) The heparin (Fig. 1a) used as "control heparin" was heparin lot 33 88 12 (Hoffmann-La Roche) and had a mean molecular weight of 12,500, an sulfur content of 10.7%, an anti-IIa anticoagulant activity of 2.2 mg/ml, an anti-Xa activity of 2.7 g/ml, and an IC 50 on smooth muscle cell proliferation of 100 g/ml (56). 2) CR-heparin (CR for carboxyl-reduced, Fig. 1a) was prepared by reduction of control heparin (56). During the reaction no depolymerization had occurred, as evidenced by the high performance liquid chromatography profiles on an Ultrapac TSK G 3000 gel permeation column in comparison with the starting material, as described previously (56,57). CR-heparin had no detectable anti-IIa and anti-Xa anticoagulant activity and no inhibitory effect on smooth muscle cell proliferation. The sulfur content of CRheparin was 12.2% (56). 3) CRS-heparin (CRS for carboxyl-reduced and sulfated; Fig. 1a) was obtained by sulfation of the primary hydroxyl groups of CR-heparin; the experimental procedure and characterization have been reported previously (56). CRS-heparin had a sulfur content of 15.7% and a low anticoagulant activity (anti-IIa activity: 170 mg/ml; anti-Xa activity: 680 g/ml) (56). The selective sulfation of the primary hydroxyl groups was supported by characteristic peaks observed in the 13 C NMR spectrum. The inhibitory effect of CRS-heparin on smooth muscle cell proliferation was higher than for the control heparin (relative inhibition (r i ) ϭ 1.2) (58).
The following oligosaccharides were also obtained from Hoffmann-La Roche. Trestatin A, a non-uronic pseudo-nonasaccharide (Fig. 1b) extracted from Streptomyces dimorphogenes, had neither anticoagulant nor antiproliferative activity (56). Trestatin A sulfate (M r ϭ 3550 Ϯ 300) was prepared by extensive sulfation of trestatin A (M r ϭ 1435), as previously reported (56). Briefly, trestatin A was dissolved in absolute N,N-dimethylformamide and sulfated in the presence of sulfur trioxide pyridine complex using excess of reagent (2.3 eq/hydroxyl group in trestatin A). After precipitation of the sulfated compound, the crude material was treated with aqueous sodium acetate solution to obtain the trestatin A sulfate sodium salt. A mixture of differently sulfated pseudo-nonasaccharides was obtained with an average number of sulfate groups per monosaccharide unit (degree of sulfation ϭ 2.3 Ϯ 0.3). The degree of sulfation for trestatin A sulfate was calculated from the integrals in the NMR spectrum of the pyridinium salt (56). Trestatin A sulfate had no anticoagulant activity as assessed by the activated partial thromboplastin time assay, and anti-IIa and anti-Xa activity in chromogenic assays, but a high inhibitory activity on smooth muscle cell proliferation (r i ϭ 1.2) (58).
Immunofluorescence Studies-Immunostaining was carried out by incubating cells for 20 min at 4°C with appropriate fluorescein isothiocyanate/phycoerythrin-conjugated mAbs or soluble adhesion receptors (L-selectin/, P-selectin/, E-selectin/, or CD4/ chimeric proteins) (14). mAbs and chimeric proteins were used at optimal concentrations in phosphate-buffered saline supplemented with 1% albumin and 1 mM CaCl 2 . Cell surface binding of chimeric proteins was detected using a polyclonal fluorescein isothiocyanate-conjugated rabbit anti-human IgM heavy chain antibody (Dako, Glostrup, Denmark). Flow cytometry was performed with a EPICS Profile cytofluorimeter (Coulter Electronics, Hialeah, FL). Mononuclear cells were gated by forward and side scatter signals. A total of 5000 cells was analyzed in each experiment.
Cell Adhesion Assays-A well defined laminar flow was produced over confluent CHO cells stably expressing PSGL-1/C2GnT and Fuc-T VII (CHO-PSGL-1/C2GnT/Fuc-T VII cells), E-selectin (CHO-E-selectin cells), or P-selectin (CHO-P-selectin cells). Cells were grown on 25-mm glass circular coverslips introduced in a parallel plate flow chamber (40). Peripheral blood neutrophils or U937 cells, suspended at 0.5 ϫ 10 6 /ml in RPMI medium plus 1% FCS, were perfused through the chamber via a syringe pump (Harvard Apparatus, Indulab AG, Switzerland) for 4 min at room temperature and at a constant shear stress of 1.5 dyn/cm 2 . Neutrophil interactions with transfected CHO cells were visualized using an inverted phase contrast videomicroscope (Leica, Lausanne, Switzerland) and Sony CCD-IRIS videocamera) and videotaped (Panasonic s-VHS recorder, TSA Telecom, Switzerland). Sequential images of neutrophil or U937 cell interactions with transfected CHO cells were digitalized and analyzed using a software developed for use in the public domain (National Institutes of Health Image software, version 1.57). Images were analyzed on a Power-Macintosh 8600/200 equipped with a Scion LG-3 board (Scion, Frederick, MD). Cell-cell interactions were analyzed from videotapes at 2-4 min of perfusion. Most U937 cells interacting with CHO-P-selectin cells and most neutrophils interacting with CHO-PSGL-1/C2GnT/Fuc-T VII cells were rolling cells. Heparin preparations and oligosaccharides were diluted in cell suspensions at appropriate concentrations. The LAM 1-3, WASP12.2, and H18/7 or 7A9 mAbs were used as anti-L-selectin, anti-P-selectin, and anti-E-selectin function blocking mAbs. Isotype-matched mAbs were used as controls. Experiments were performed in quadruplicate. Results were expressed as percentage of the control obtained from experiments using cells in medium with vehicle alone. These control studies employing cells in medium with vehicle alone were conducted at the beginning and the end of each experimental condition.
Intravital Microscopy-Harlan Sprague-Dawley rats (250 -300 g) were anesthetized with ketamine (75 mg/kg administered intramuscularly; Parke Davis, Berlin, Germany) after premedication with pentobarbital (Nembutal, Sanofi, Hannover, Germany; 20 mg/kg administered intramuscularly). Anesthesia was maintained by a continuous infusion of pentobarbital (0.2 mg/ml in physiologic saline) at 40 ml/kg/h. Leukocyte concentration was determined at 45-min intervals using a Coulter D N cell counter (Coulter, Herts, United Kingdom). After opening of the peritoneal cavity, a few loops of ileum proximal to the appendix were exteriorated onto a limited stage and were superfused with bicarbonate-buffered isotonic saline at 37°C (61). Oligosaccharides or phosphate-buffered saline were infused through a micropipette in an upstream side branch of the venule to be investigated. Mesenteric microcirculation was observed using a Leitz (Wetzlar, Germany) intravital microscope and recorded on a Sony U-matic videotape (Sony, Berlin, Germany) via a video camera (RCA, Lancaster, PA). Rolling leukocyte flux was determined by counting the number of rolling leukocytes passing a line perpendicular to the vessel axis (61). For comparison between observations made in different venules, flux values were normalized to the average rolling leukocyte flux during the control period that followed each micro-infusion.
Thioglycollate-induced Peritonitis-BALB/c mice (6 weeks old, ϳ30 g) were anesthetized with metofane (Arovet AG, Switzerland) and treated with a subcutaneous injection of 3.0 mg of trestatin A sulfate or CRS-heparin. Doses of CR-heparin, CRS-heparin, and trestatin A sulfate were in the same range as those chosen for thrombosis prophylaxis or for inhibition of L-selectin or P-selectin activity in mice (47). After 30 min, a time point that corresponds to the peaks of blood levels of the injected drugs, the animals received an intraperitoneal injection of 1 ml of 3% thioglycollate broth (T9032, Sigma) or sterile pyrogen-free saline. The mice were sacrificed 3 h later, at a time point where both P-selectin and L-selectin are involved in regulating neutrophil migration into the peritoneal cavity (62,63). Peritoneal leukocytes were harvested by peritoneal lavage with 5 ml of saline containing 2 mM EDTA. After red blood cell lysis, leukocytes were counted in a hemocytometer. Neutrophils were counted after staining with Tü rck or by counting cytospin preparations stained with Giemsa (Fluka, Switzerland).
Statistical Analysis-The Mann-Whitney test was used to compare medians of unpaired groups while medians of paired groups were compared with the Wilcoxon signed rank test. When three or more groups were compared, differences between treatments were evaluated by analysis of variance and Bonferroni multiple comparison tests. p values Ͻ 0.05 were considered as significant. Data are shown as means Ϯ 1 S.E.
These results show that the presence of carboxyl group in heparin is important for anti-L-selectin or anti-P-selectin activity. However, it can be artificially replaced by an O-sulfate ester, giving a compound with a more potent anti-L-selectin and anti-P-selectin activity without antithrombin-mediated anticoagulant activity.
Inhibition of L-and P-selectin-mediated Leukocyte Adhesion by Control Heparin, CR-heparin, CRS-heparin, and Trestatin A Sulfate: Studies under Shear Flow Conditions-The ability of control heparin, CR-heparin, CRS-heparin, or trestatin A sul-fate to inhibit L-selectin-and P-selectin-mediated cell adhesion was then studied under flow conditions. Adhesion assays were performed in a parallel plate flow chamber, at a constant shear stress of 1.5 dyn/cm 2 , on CHO cell monolayers stably expressing E-selectin or P-selectin at levels similar to those expressed by activated human endothelial cells. For studying interactions between P-selectin and PSGL-1, we used U937 cells expressing high levels of PSGL-1, C2GnT, and Fuc-T VII and CHO-Pselectin cell monolayers. For studying interactions between L-selectin and PSGL-1, we used peripheral blood neutrophils and CHO cell monolayers expressing high levels of PSGL-1, C2GnT, and Fuc-T VII. It was observed that U937 cells efficiently interacted with CHO-P-selectin cells (57 Ϯ 9 rolling cells/min/mm 2 , n ϭ 19). The specificity of this interaction was demonstrated by blocking studies with specific mAbs. WASP 12.2, an anti-P-selectin mAb, abolished the rolling of U937 cells on CHO-P-selectin cells (99 Ϯ 1% of inhibition, n ϭ 8). Conversely, anti-PL1 mAb eliminated PSGL-1-dependent interactions of U937 cells with CHO-P-selectin cells (99 Ϯ 1% of inhibition, n ϭ 4).
The anti-P-selectin activity of control heparin, CR-heparin, CRS-heparin, trestatin A, and trestatin A sulfate were initially determined at a saccharide concentration of 1.0 mg/ml. Control heparin strongly inhibited U937 cell rolling on CHO-P-selectin cells (96 Ϯ 2% of inhibition, p Ͻ 0.001, n ϭ 4; Fig. 4). Carboxyl reduction of control heparin caused a complete loss of this inhibitory activity. Indeed, rolling of U937 cells on P-selectin was not significantly inhibited by 1.0 mg/ml CR-heparin (21 Ϯ 12% of inhibition, n ϭ 4). However, sulfation of CR-heparin reestablished anti-P-selectin activity. For example, the rolling of U937 cells on CHO-P-selectin cells was almost abolished in presence of 1.0 mg/ml CRS-heparin (98 Ϯ 1% of inhibition, p Ͻ 0.001, n ϭ 8; Fig. 4). A similar inhibition of U937 cell rolling cells on CHO-P-selectin cells was seen with trestatin A sulfate (99 Ϯ 1%, p Ͻ 0.001, n ϭ 8; Fig. 4). No anti-P-selectin activity was detectable with 1.0 mg/ml trestatin A (84 Ϯ 10%, n ϭ 4; Fig. 4), emphasizing the importance of sulfate groups for observing antiadhesive activity. The anti-P-selectin activities of control heparin, CRS-heparin, and trestatin A sufate were then quantified by estimating IC 50 values for the inhibition of U937 cell rolling on CHO-P-selectin cells (Fig. 5). The IC 50 value for control heparin was 0.38 mg/ml; it was 0.10 mg/ml for CRSheparin and 0.17 mg/ml for trestatin A sulfate (Fig. 5).

CRS-heparin or Trestatin A Sulfate Inhibit Neutrophil Migration into Thioglycollate-inflamed Peritoneum-
The ability of CRS-heparin and trestatin A sulfate to reduce neutrophil migration at sites of acute inflammation was evaluated in BALB/c mice using a thioglycollate-induced model of peritonitis. Three hours after thioglycollate injection, neutrophils were collected from the inflamed peritoneal cavity and counted. Thioglycollate injection induced a 4-fold increase in neutrophil accumulation. Neutrophil migration into the peritoneal cavity was efficiently inhibited by a subcutaneous injection of trestatin A sulfate (3.0 mg) or CRS-heparin (3.0 mg). In two experiments, neutrophil accumulation was reduced by 81 Ϯ 11% (p Ͻ 0.001, n ϭ 5) and 73 Ϯ 10% (p Ͻ 0.001, n ϭ 7, Fig. 8) when mice were treated with trestatin A sulfate, whereas the unsulfated form of this nonasaccharide had no effect. Subcutaneous injection of CRS-heparin inhibited neutrophil migration by 58 Ϯ 10% (p Ͻ 0.01, n ϭ 5) and 73 Ϯ 10% (p Ͻ 0.01, n ϭ 5, Fig. 8). As expected, CR-heparin (3.0 mg subcutaneous) did not influence neutrophil migration (Fig. 8).

DISCUSSION
Several studies have indicated that L-selectin and P-selectin can interact with heparin and heparan sulfate glycosaminoglycan chains. Fragments of heparin that bind to L-selectin include the more heavily sulfated and epimerized regions (42,46). P-selectin binding fragments contain these regions but also less modified fragments. The present study has examined the structural requirements for L-selectin or P-selectin binding to heparin. Our main observations are that: 1) the heparin derivative CRS-heparin, devoid of antithrombin-mediated anticoagulant activity, exhibits high anti-L-selectin and anti-P-selectin activity; 2) trestatin A sulfate, a non-uronic pseudo-nonasaccharide obtained by sulfation of trestatin A, also possesses high anti-L-selectin and anti-P-selectin activity; 3) when tested in two animal models of experimental inflammation, both CRSheparin and trestatin A sulfate exhibit strong anti-inflammatory activity as a result of inhibition of L-selectin-and Pselectin-mediated cell adhesion. From this finding it can be concluded that 1) the carboxylates in heparin can be replaced by a primary O-sulfate group to obtain heparin derivatives that inhibit L-selectin or P-selectin activity, and 2) antithrombindependent anticoagulant activity of heparin is not required for anti-selectin activity.
Control heparin partially inhibited binding of recombinant L-selectin/ and P-selectin/ chimeric proteins to KG-1 cells (Fig. 2), a reaction almost completely dependent on the interaction of L-selectin or P-selectin with PSGL-1 (14). As control heparin inhibited both selectins, it is possible that L-selectin and P-selectin have similar heparin binding domains. It has been suggested that appropriate sulfation and carboxylation of heparin play a critical role in selectin recognition (42,46). In cell rolling assays, control heparin efficiently inhibited PSGL-1-mediated U937 cell rolling on P-selectin (Fig. 4), an interaction dependent on post-translational modifications of PSGL-1 including tyrosine sulfation and expression of sialylated and fucosylated O-glycans on threonine 57 (22). In contrast, control heparin did not inhibit L-selectin-mediated rolling on PSGL-1 (Fig. 6). This difference may result from differences in L-selectin and P-selectin binding sites for heparin or PSGL-1. As expected, control heparin did not affect E-selectin-mediated interactions.
Earlier studies have shown that inhibition of arterial smooth muscle proliferation by heparin is independent from its anticoagulant activity and that the carboxyl group of the uronic acid unit is not essential for antiproliferative activity (56,64). To determine if the carboxyl group of the uronic acid units were required for anti-selectin activity, we compared the inhibitory activities of control heparin and CR-heparin. The anti-P-selectin and anti-L-selectin activity of control heparin was lost after carboxyl reduction (Figs. 2 and 4) indicating that the uronic acid unit 2-O-sulfate groups and the D-glucosamine unit 2-N-,6-O-sulfate groups are not sufficient for this activity. Another possible interpretation is that, by removal of the carboxyl group, just one of the several charges required for efficient P-selectin and L-selectin binding, have been removed. This would be in agreement with stronger binding to L-selectin and P-selectin observed after introduction of a sulfate group in CRS-heparin, prepared by sulfation of the carboxyl-reduced group of CR-heparin (Figs. 2, 4, and 6). Importantly and in contrast to control heparin, CRS-heparin had no antithrombindependent anticoagulant activity (56). Thus, requirements for antithrombin binding are different from those necessary for L-selectin or P-selectin binding (65,66).
The key role played by sulfate groups in the anti-selectin Each experimental group contained 5-7 BALB/c mice. Control mice were injected with sterile, pyrogen-free saline alone (white column); neutrophil migration 3 h after the injection of 3% thioglycollate broth (black or hatched columns) is shown. Saline and oligosaccharides were injected subcutaneously 30 min before thioglycollate. Results, indicated as mean Ϯ S.E., are representative of two separate experiments (**, p Ͻ 0.01; ***, p Ͻ 0.001). activity of CRS-heparin and trestatin A sulfate is a common feature of biological selectin ligands which regulate leukocyte migration (67,68). Specific sulfotransferases are involved in the sulfation of sialyl-6-sulfo-Le x and sialyl-6Ј-sulfo-Le x , two capping structures of GlyCAM-1 and CD34 that interact with L-selectin (32,33,68). Sulfation of PSGL-1 is regulated by two tyrosylprotein sulfotransferases and is required for L-selectin and P-selectin binding (14,17,18,22,70,71). Sulfation of heparan sulfate proteoglycans may also be important to support L-selectin binding (40 -42).
Since an intact uronic acid unit is not required for inhibiting L-selectin or P-selectin binding to PSGL-1, we examined the anti-selectin activity of trestatin A sulfate, a highly sulfated pseudo-nonasaccharide with antiproliferative activity on arterial smooth muscle cells (56). Trestatin A sulfate exhibited high anti-L-selectin and anti-P-selectin activities (Figs. 4 -6). Trestatin A sulfate was very efficient at inhibiting L-selectin-mediated rolling on PSGL-1, whereas CRS-heparin had a much weaker effect on this reaction and control heparin was completely inactive (Fig. 6). In vivo, the inhibitory activity of trestatin A sulfate on neutrophil rolling along postcapillary venules disappeared quickly upon termination of microinfusion. Transient inhibition of selectin-mediated rolling might be advantageous, as it will probably avoid prolonged suppression of the host cellular defense system. Importantly, sulfation of trestatin A was required for anti-L-selectin and anti-P-selectin activity, further emphasizing the role of sulfate residues for Lselectin and P-selectin binding to carbohydrate ligands (Fig. 7).
Three UFH and three LMWH had highly variable anti-Pselectin activity. UFH A, a preparation extracted from porcine intestinal mucosa, had no detectable anti-L-selectin and anti-P-selectin activity, whereas UFH C had a strong anti-P-selectin activity and UFH B a moderate inhibitory effect on P-selectinmediated rolling. These observations indicate that the exhibition of an anti-selectin activity is not a general property of UFH, a notion consistent with the data available in the literature. In some studies, certain heparin preparations were found to have significant inhibitory activity against L-selectin or P-selectin; in other studies, selectin inhibition was absent and no reaction was observed with heparin or heparan sulfate proteoglycans (41,46,47,72,73). These differences suggest that certain heparin could express unique structural features that support L-selectin and P-selectin binding. Differences in molecular weight, degree of sulfation, as well as by the expression of specific binding sequences for L-or P-selectin may be important.
Carboxyl reduction and sulfation of heparin or sulfation of non-uronic oligosaccharides such as trestatin A generates nonanticoagulant compounds with high anti-L-selectin and anti-Pselectin activities. These compounds have the potential of preventing leukocyte migration with reduced bleeding risk. The anti-inflammatory activity of CRS-heparin and trestatin A sulfate was evaluated in a thioglycollate-induced model of peritonitis. Subcutaneous injection of trestatin A sulfate and CRS, before thioglycollate intraperitoneal injection, inhibited neutrophil migration by 60 -80%, confirming the potent inhibitory activity of these compounds on neutrophil adhesion and migration (Fig. 8). Neutrophil migration in this peritonitis model is dependent on E-selectin, P-selectin, and L-selectin function (74 -76). Neutrophil migration was not completely inhibited despite blockade of L-selectin and P-selectin by CRS-heparin or trestatin A sulfate. Residual migration may result from Eselectin-dependent cell adhesion, a reaction not affected by CRS-heparin or trestatin A sulfate. Inhibition of cell migration could be enhanced by combining E-selectin inhibitors to CRSheparin or trestatin A sulfate (77,78).
In clinical practice, inhibitors of L-selectin and P-selectin could be very useful to prevent ischemia/reperfusion injury observed in various conditions such as myocardial infarction, stroke, or solid organ transplantation. The role of L-selectinand P-selectin-dependent cell adhesion in contributing to ischemia/reperfusion injury has been well established in animal models (48, 50 -54, 79 -82). Inhibition of L-selectin or P-selectin with mAbs or sLex-related compounds attenuates neutrophil accumulation in reperfused organ, reduces the area at risk of infarction or acute organ dysfunction, and results in better recovery. Some heparin preparations can preserve myocardial contractility after ischemia/reperfusion injury or reduce brain injury (83)(84)(85)(86). The anti-L-selectin and anti-P-selectin activity of heparin could constitute a major mechanism by which heparin derivatives may prevent reperfusion injury. Additional properties of heparin could also contribute to reduce reperfusion injury such as inhibition of complement activation or CD11b function (69,73,87). The anti-L-selectin and anti-Pselectin activity of CRS-heparin and trestatin A sulfate could be particularly useful for the prevention of reperfusion injury when the risk of hemorrhagic complications is increased, for example in the case of thrombolytic therapy, traumatic shock, or solid organ transplantation. Additional in vivo studies are now required to further assess the therapeutic potential of these compounds.