Research Paper Mediators of Inflammation, 8, 107–113 (1999)

The natural physiological ligands for selectins are oligosaccharides found in glycoprotein or glycolipid molecules in cell membranes. In order to study the role of sugar residues in the in vivo lectin anti-inflammatory effect, we tested three leguminous lectins with different carbohydrate binding affinities in the peritonitis and paw oedema models induced by carrageenin in rats. L. sericeus lectin was more anti-inflammatory than D. virgata lectin, the effects being reversed by their specific binding sugars (N-acetylglucosamine and alpha-methylmannoside, respectively). However, V. macrocarpa, a galactose-specific lectin, was not anti-inflammatory. The proposed anti-inflammatory activity of lectins could be due to a blockage of neutrophil-selectin carbohydrate ligands. Thus, according to the present data, we suggest an important role for N-acetylglucosamine residue as the major ligand for selectins on rat neutrophil membranes.


Introduction
Neutrophil infiltration into inflamed tissue is a complex sequential chain of events comprising spatial and temporal expression of adhesion molecules, both in leukocytes and endothelial cell membranes. [1][2][3] The earliest adhesion molecules involved in neutrophilendothelium interactions are selectins. These transmembrane adhesion molecules, leukocyte L-selectin, endothelial P-selectin and E-selectin, have similar structural organization and mediate leukocyte rolling. The molecular structure of selectins is composed by an N-terminal lectin domain, an epidermal growth factor-like domain, multiple consensus repeats (CR), a transmembrane, and a cytoplasmatic domain. These selectin domains can functionally be involved in the neutrophil recruitment in vivo. 4 Recently, it was demonstrated that a specific monoclonal antibody, that recognizes an epitope in the CR domain of E-selectin, was in the C57BL/6 strain, and which had no effect on leukocyte arrest and rolling. 5 Thus, these data raise the possibility for an additional function for E-selectin, downstream from leukocyte rolling, and stress that considerable species differences may be found in respect to the role of selectin domains. It is well recognized, however, that the lectin domain has a pivotal role on neutrophil arrest and rolling. 3,4 The adhesive interactions of this domain involve a lowaffinity, reversible binding to complex carbohydrates present in the cytoplasmatic membranes of neutrophils and endothelial cells. 6 This adhesive interaction between neutrophils and endothelium can be reduced by a variety of simple and complex carbohydrates, most of which are sialylated, fucosylated, or both. 4 It has been proposed that sugar-based inhibitors directed against adhesive activities of selectins might provide for new and more effective antiinflammatory drugs. [7][8][9] Recently, we have demonstrated that glucose-mannose binding plant lectins were able to inhibit the neutrophil migration into rat peritoneal cavities induced by carrageenin and fMLP, and paw swelling induced by carrageenin. This oedema is considered as a classical neutrophilmediated inflammatory model. These effects were not due to endotoxin contamination or lectin-induced haemaglutination and leukopaenia, and could be reversed by a -methyl-mannoside (a -MM) but not but a -D-galactose (a -D-gal). We propose that the inhibitory effect upon neutrophil recruitment was due to a competitive blockage of glycosylated selectin binding sites by plant lectins. 10 In humans, the most important selectin binding sialylated glycoconjugate is the sialosyl-Lewis-X (sLe x ). However, since this epitope is highly species specific and is absent in non-human mammalian species, 11 it cannot be considered as a general binding site for interaction with selectins in non-human leukocytes. 11,12 Thus, sugar residues other than sialylated glycoconjugates must be involved in selectin-mediated neutrophil recruitment process in non-human species. The objective of the present work was to investigate the relative involvement of galactose, N-acetylglucosamine and glucose-mannose residues in the neutrophil-mediated rat peritonitis and paw oedema induced by carrageenin, using lectins obtained from leguminous seeds.

Animals
Male Wistar rats (150-200 g) were housed in a temperature-controlled room with access to water and food ad libitum, until used.

Neutrophil migration into peritoneal cavity of rats
As the inflammatory stimuli, carrageenin (Cg; 300 m g) was injected i.p. in 1 ml of sterile NaCl 0.15 M solution. After 4 h, animals were sacrificed and peritoneal lavage was performed with 10 ml of sterile phosphatebuffered saline (PBS) containing 5 U/ml heparin. The fluid was removed for total and differential cell counts and the results are reported as mean±S.E.M. of the number of cells per microliter of peritoneal wash. 16 Lectins (0.01-1 mg/kg), LPS (30 m g/kg) or saline (0.1 ml/100 g body weight) were injected i.v. 30 min before carrageenin. Lectins were administered alone or in solution with 0.1 M a -D-methyl mannoside (a -MM), 0.2 M N-acetylglucosamine (GlcNAc) or 0.1 M of a -D-galactose (a -D-gal). LPS was used as positive control for the inhibitory effect upon neutrophil migration 17 . All drugs were dissolved in sterile NaCl 0.15 M.

Rat paw oedema
Paw oedema was induced in the left hind paw of rats under light ether anaesthesia by a single subcutaneous intraplantar injection of carrageenin (300 m g/paw) or dextran (300 m g/paw), diluted both in 0.1 ml of sterile saline 0.15 M. Oedema was measured plethysmographically 18

Statistical analysis
All results were expressed as mean±S.E.M. for n experiments. Statistical evaluation was undertaken by analysis of variance (ANOVA) and Ducan's test for multiple comparison. All i.v.-treated groups were compared to animals which received injection of irritants, and the positive controls of inflammation were compared to animals which received only saline from both routs. A P value of less than 0.05 was considered to indicate significance.

Results
Inhibitory activity of lectins on carrageenin-induced neutrophil migration Figure 1 shows the inhibitory effect of D. virgata and L. sericeus lectins, at doses varying from 0.01 to 1 mg/kg i.v., on neutrophil migration induced by intraperitoneal injection of carrageenin (Cg; 300 m g) in rats. L. sericeus, a specific N-acetylglucosamine lectin caused, respectively, 15, 74 and 92% of inhibition, measured 4 h after Cg injection (Fig. 1C). The effect of this lectin at the highest dose used (1 mg/kg) was similar to that obtained with the i.v. injection of LPS (30 m g/kg). D. virgata, a lectin that binds specifically to mannose/ glucose residues, also decreased, by 38%, the Cginduced neutrophil infiltration, but only at the highest dose (Fig. 1B). On the other hand, V. macrocarpa, a galactose-binding lectin, did not alter the number of neutrophils, when compared to the saline-injected animals (SAL group), at any dose tested (Fig. 1A).
Involvement of carbohydrate residues on the anti-inflammatory activity of D. virgata and L. sericeus lectins at the peritonitis model When D. virgata (DvigL) and L. sericeus (LserL), at 1 mg/kg, were conjointly injected i.v. with their specific binding sugars, a -MM (DvigL) and GlcNAc (LserL) at 0.1 and 0.2 M, respectively, the inhibitory effect on neutrophil migration was reverted (P<0.05). On the contrary, 0.1 M of a -D-gal did not modify these lectin-inhibitory activities ( Fig. 2A,B). The simultaneous i.v. injection of LPS (30 m g/kg) with a -MM or GlcNAc did not change the LPS inhibitory effect on the Cg-induced neutrophil migration model (data not  shown). Thus, we can rule out the possibility that the lectin activity would be due to endotoxin contamination.

Effect of plant lectins on rat paw oedema induced by carrageenin and dextran
The subcutaneous intraplantar injection of Cg (300 m g/paw) induced a progressive and intense paw oedema that reached a maximal value by the third hour (Fig. 3). Lectins from D. virgata and L. sericeus seeds injected i.v. at 1 mg/kg, 30 min before the flogistic agent, reduced by 24 and 49%, respectively, the Cg-induced paw oedema. The L. sericeus lectin inhibitory effect was similar to that obtained with LPS (30 g/kg) which caused a 51% of reduction in the paw oedema induced by Cg (Fig. 3). The co-injection of D. virgata (1 mg/kg) and L. sericeus (1 mg/kg) with their specific binding sugars reversed (P<0.05) the inhibitory effect of these lectins. However, 0.1 M of a -D-gal did not alter the lectin's anti-inflammatory effect (Fig.  4). Dextran (300 m g; s.c.), induced a paw oedema peaking 1 h after injection. From Fig. 5 it can be concluded that none of the tested lectins inhibited the development of the oedema induced by dextran, a classical leukocyte-independent inflammatory agent. 19 The i.v. injection of carbohydrates (a -MM and GlcNAc) alone, did not cause any reduction on the paw oedema induced by Cg (data not show).

Discussion
It was demonstrated that D. virgata (DvigL) and L. sericeus (LserL) lectins, when administered i.v., inhibited the carrageenin-induced neutrophil migration to peritoneal cavities of rats. These lectins could also inhibit the rat paw oedema induced by carrageenin. In both models L. sericeus lectin showed the most important inhibitory activity. DvigL and LserL inhibitory activities were reversed by the co-injection of their specific sugars (a -methyl mannoside; a -MM and N-acetylglucosamine; GlcNAc, respectively). On the contrary, when the lectin association was made with a -D-galactose (a -D-gal), a non-specific binding carbohydrate, the inhibitory effect was not reversed. This result suggests the involvement of a lectin domain of these proteins in this effect. On the other hand, V. macrocarpa (VmacL), a galactose-specific lectin, when injected i.v., did not affect the neutrophil migration, in spite of its pro-inflammatory activity by the intraperitoneal route (data not shown). Is known that dextran induces a type of oedema without involvement of polymorphonuclear leukocytes at the inflamed tissue, but via resident cell degranulation. 19 Since neither LserL nor DvigL inhibited the dextraninduced paw oedema, it was suggested that the lectin activity would be related to inflammatory leukocytemediated reactions. The leukocyte traffic into inflam-matory sites is an important event of the effective host response against infection. The early step involved is the interaction between selectins and leukocytes, which facilitates the rolling of these cells along the vascular endothelium wall. 4  receptors, belonging to the selectin family (E, P and L-selectins) participate in this adhesive interaction. The presence of a lectin domain in the selectin structure enables the interaction between these glycoproteins and carbohydrates of the cell mem-branes. 6 The physiological ligands for the selectins are subject to ongoing investigations. It is evident that selectins share affinity with a common carbohydrate core structure, but the detailed binding specificity of each is distinct. 20 In general, the natural ligands are generally found at the termini of O-linked glycoprotein oligosaccharides or of glycolipids. These ligands contain fucosylated core trisaccharides designated Lewis x (Le x ), Lewis a (Le a ) or related structures, which are elaborated in different ways. Some of the most effective ligands for selectins contain 39 -sialyl-Lewis x (sLe x ) with additional sulphate residues attached to the intervening galactose and the N-acetylglucosamine. 21 Several studies have suggested that the sialyl-Lewis x may function as a selectin ligand in humans. 4 However, it was demonstrated that this epitope is absent in non-human mammalian species. 11 Based on these data the adhesive interaction mediated by these proteins in the present rodent model would involve other mechanisms and/or polysaccharides related to sLe x . As LserL showed the most important inhibitory effect upon neutrophil migration and N-acetylglucosamine is present at the sLe + structure, we can suggest a possible similarity between human and rodent ligands. This raises the question of the bases for specificity, as well as avidity in selectincarbohydrate interaction. Several possible mechanisms could be envisioned to answer this question, including a role for protein-protein interactions between selectins and their in vivo ligands, multivalence of oligosaccharide ligands, or specific modifications of sLe x "backbone". In this article the differences in potency of the anti-inflammatory effect observed for L. sericeus (92%) and D. virgata (38%) lectins, in both models used, suggest an essential role for N-acetylglucosamine-containing glycoconjugates in leukocyte-mediated inflammatory models. Moreover, the anti-inflammatory effect of L. sericeus lectin was similar to the maximal inhibition obtained with LPS. 17 Another hypothesis can be raised based on the intermediary potency in inhibiting the neutrophil migration observed for D. virgata lectin. According to this, Assreuy and co-workers 10 could also demonstrate a less intense blockade (not superior to 70%) of the Cg-induced neutrophil migration by glucose-mannose type lectins compared to the effect of L. sericeus lectin. Thus, it is possible to postulate a partial involvement of glucose-mannose residues in the present model. On the other hand, the lack of inhibitory effect demonstrated for V. macrocarpa lectin is controversial. One could suggest that residues of galactose are not involved in this effect. However, we cannot rule out a role for these sugars. Some studies using plant lectins with similar physicochemical properties have demonstrated positive (at different potencies) or negative responses toward the same biological system. 10,22 -24 These responses would probably be due to small differences in the amino acid sequence of lectins in the region involved with sugar interaction. This may affect the fine specificity for sugars in the cell membrane. In this line, Sanz-Aparicio and co-workers 25 demonstrated that a change in a single amino acid residue of C. brasiliensis and C. ensiformis lectins causes a different geometrical arrangement in its molecular structure. Moreover, PSP-II, 26 an animal lectin belonging to the spermadhesin family, seems to have a cryptic binding site for mannose-6-phosphate in the native heterodimer (PSP-I/PSP-II). Thus, based on this, we suggest that the lectin domain of V. macrocarpa lectin would be located in an unfavourable region of the molecule, or in a cryptic manner preventing the binding to its specific sugar on leukocyte membranes. In summary, based on plant lectin inhibitory activity in the carrageenin-induced rat peritonitis model, the results show that: (a) galactose residues play no role in this model; (b) glucose-mannose residues are partially involved; and (c) N-acetylglucosamine-containing glycoconjugates are essential for neutrophil recruitment. This ranking of order explains the lack of activity of VmacL, the intermediary potency of DvigL, and the high anti-inflammatory activity of LserL. These data together open perspectives for the use of lectins, as sugar-based inhibitors, acting directly on specific carbohydrate targets in inflammatory cells, for better understanding the mechanisms involved in cellular inflammatory reactions.