Anti-inflammatory effect of glucose—mannose binding lectins isolated from Brazilian beans

Selectins are essential for leukocyte recruitment in inflammation. Because of a lectin domain present in the selectin structure, we investigated the anti-inflammtory activity of six mannose–glucose binding lectins from brazilian beans: Dioclea guianensis-DguiL; D. grandiflora-DgL; Cratylia floribunda-CfL; D. violacea-D.vL; D. virgata-DvirL and Canavalia brasiliensis-ConBr. The lectins were injected intravenously (i.v.) into rats (0.1 and 1.0 mg/kg; 30 min before irritants) and its activities compared to E. coli endotoxin (LPS,30 μg/kg i.v.). Three lectins (DvL, CfL and DguiL), although less intense than LPS, inhibited the neutrophil migration induced by carrageenan (Cg, 300 μg) in a dose-dependent manner (0.1 and 1.0 mg/kg). DvL activity was reversed by 0.1 M α-D-methyl-mannoside (α-CH3), but not by 0.1 M α-D-galactose. The fMLP (44 ng)-induced neutrophil migration was also reduced by these lectins. Endotoxin contamination of lectin samples could be excluded since α-CH3 treatment reversed the DvL effect, but did not modify LPS inhibitory activity. Carrageenan (300 μg)-induced paw oedema was also reduced by LPS or lectin treatments. Conversely, none of the tested lectins inhibited dextran (Dex, 300 μg)-induced paw oedema, a classical leukocyte independent model, or zymosan (Zy, 1.0 mg)-induced peritonitis and paw oedema. LPS showed no effect upon Dex-induced paw oedema and barely reduced (25%) the oedematogenic effects of zymosan. As proposed for LPS, the lectin inhibitory activity was better observed on neutrophil-mediated inflammatory reactions. We speculate that the plant lectin antiinflammatory activity is probably due to a competitive blockage of a common leukocyte and/or endothelial selectin carbohydrate ligand.


Introduction
Neutrophil in ltration from the blood into the affected tissue is the hallmark of acute in ammatory reactions. The recruitment of these cells involves a complex and multi-mediated process which includes sequential interactions between these cells and endothelial cells and extracellular matrix components. 1 -3 The control of these mechanisms involves both the activation of adhesion receptors already present in endothelial and resting blood cells, and the expression of new adhesion molecules on cell membranes. The crucial initial step of neutrophil migration depends on binding of the neutrophils to selectins expressed by both the above mentioned cells and venular endothelial cells acti-vated by in ammatory mediators. 4 Three selectins, belonging to the C-type family of eukaryotic lectins, have been characterized. 5 The interactions involving selectins are dependent on the recognition of speci c cell surface glycoconjugates. 6,7 For example, human endothelial P-selectin is able to speci cally bind to sialylated, fucosylated polylactosamines on the surfaces of leukocytes. 8 A recent study has demonstrated that the lectin domain of two subunits of the B oligomer from pertussis toxin, S2 and S3, shares amino acid sequence similarity with the lectin domains of the eukaryotic selectin family. 9 Because selectins are essential for rolling of leukocytes on endothelial cells, in the in ammatory process, it was postulated that pertussis toxin could block leukocyte recruit-ment into in amed tissue by competitive blockage of selectin binding sites. In fact, it has been demonstrated that S2 and S3 lectin domains or peptides derived from them competitively inhibit adherence of neutrophils to selectin-coated surfaces and to endothelial cells in vitro 10 as well as leukocyte recruitment into cerebrospinal uid after pneumococcal challenge in vivo. 11 However, we have recently demonstrated that for the anti-in ammatory effect of pertussis toxin in vivo, the ADP-ribosilating activity of the A-protomer is more important than S2 and S3 lectin domains. 12,13 We have investigated the capacity of some plant lectins to antagonize, in vivo, the neutrophil migration induced by different in ammatory stimuli. Plant lectins can be de ned as all plant proteins that possess at least one non-catalytic domain that binds reversibly to a speci c mono-or oligosaccharide, 14 and are able to mediate several biological effects including histamine releasing properties, 15 human lymphocyte stimulation, 16 and, as the term lectin was initially referred, the ability to selectively agglutinate erythrocytes. 14 It is proposed that com-petitive activity between plant lectins and selectins by glycosylated binding sites present at the surface of the endothelium, leukocyte membranes and/or extracellular matrix components can expand the range of possibilities for development of selectin analogues with potential anti-in ammatory properties.

Animals
Male Wistar rats (150 ±250 g body weight) were housed in a temperature-controlled room with free access to water and food.

Lectins
Lectins from six leguminous seeds from the Dioclin ae subtribe (family Leguminos ae, tribe Ph as eol ae) were studied. Lectin isolation by af nity chromatography on Sephadex G-50 has been previously described. Diocle a gui anensis, 17 Diocle a gr andi or a, 18 Cr atyli a oribund a, 19 Diocle a viol ace a, 20 , Diocle a virg at a, 21 and Can av ali a br as iliensis. 22 All lectins used in this study are of the glucose ±mannose type.
Neutrophil migration into peritoneal cavity of rats Carrageenan (300 mg), zymosan (1 mg) or fMLP (44 ng) were injected intraperitoneally (i.p.) in 1 ml of sterile 0.15 MNaCl. After 2 ±8 h, animals were sacri ced and peritoneal lavage was performed with 10 ml of sterile phosphate buffered saline (PBS) containing 5 U/ ml heparin. The uid was removed for total and differential cell counts and results were reported as mean SEM of the number of cells per microlitre of peritoneal wash. 23 Lectins (0.1 and 1·0 mg/ kg), alone or in combination with 0.1 M of a-Dmethyl-mannoside or a-D-galactose were injected i.v. into rats 30 min before the in ammatory stimuli (0·1 ml/ 100 g body weight). LPS (30 mg/ kg, i.v.) was used as positive control for the inhibitory effect on neutrophil migration. All drugs were dissolved in sterile 0.15 M NaCl. Control animals received sterile saline i.v. (0·1 mg/ 100 g body weight).

In¯ammatory paw oedema
Paw oedema was induced by subplantar injection of carrageenan (300 mg/ paw), zymosan (1 mg/ paw) or dextran (300 mg/ paw) in a nal volume of 0.1 ml into the right hind paw of rats under light ether anaesthesia. All drugs were dissolved in sterile 0.15 M NaCl. Paw volume was measured immediately before the irritant injections, and at selected time intervals thereafter with a hydroplethysmometer. 24 LPS (30 mg/ kg) or lectins (1 mg/ kg) were injected i.v. 30 min before injection of the irritants. Results were expressed as the increase in paw volume (ml) calculated by subtracting the basal volume. The area under the time-course curve was also calculated using a trapezoidal rule and results expressed in arbitrary units. 24

Haemagglutinating activity
Clumping by the six puri ed lectins of a 2%rat erythrocyte suspension in 0.15 M NaCl containing 5 mM CaCl 2 and MnCl 2 was estimated as described elsewhere. 25 Haemagglutinating activity was expressed as haemagglutinating units (HU)/ ml. One HU was de ned as the inverse of the end point of a serial dilution of an initial lectin solution (2 mg/ ml) required to induce visible agglutination. Determinations were made in duplicate. We also compared this activity using erythrocytes from other rodent species, e.g. rabbit and mouse.

Statistical analysis
All results were expressed as the mean SEM for n experiments with the exception of haemaglutinating activity which was expressed in median values. Statistical evaluation was undertaken by analysis of variance (ANOVA) and Duncan's test for multiple comparison. A P-value of less than 0.05 was considered signi cant.

Results
Screening for inhibitory activity of six glucose± mannose type lectins on carrageenan-induced neutrophil migration Carrageenan (Cg, 300 mg/ cavity) caused a signi cant (P , 0·05) neutrophil migration when injected into the peritoneal cavities of rats ( Fig.  1 and Fig. 3). Among the lectins tested, three inhibited the neutrophil migration induced by Cg in a dose-dependent manner: D. viol ace a lectin caused reductions of 29 and 70%, Cr atyli a oribund a lectin caused 29 and 62% reductions and D. gui anensis lectin caused 43 and 63% reductions in inhibition at 0.1 and 1·0 mg/ kg, respectively, measured 4 h after the in ammatory challenge (Fig. 1A ±C). D. virg at a lectin also decreased the Cg-induced neutrophil in ltration (54 and 40% reductions), but not in a dose-dependent manner (Fig. 1D). On the other hand, D. gr andi or a and Can av ali a br as iliensis lectins, although belonging to the same group of lectins, did not alter the number of in ltrating neutrophils, when compared to the saline-injected (i.v.) animals (SAL group), at either dose tested ( Fig. 1E and 1F). However, the effect of those inhibitory lectins was less intense than that promoted by LPS (30 mg/ kg), which reached 93% of inhibition, and was used in this work as positive control. At this dose LPS produces maximal inhibitory effect in this model, 26 and this activity could not be accounted for by hypotension. 27 Effect of lectins from Cratyliā oribunda, D. guianensis and D. violacea on neutrophil migration induced by fMLP and zymosan fMLP (44 ng, i.p.) and zymosan (1 mg, i.p.) caused a signi cant (P , 0·05) neutrophil migra-tion into the peritoneal cavity when measured 4 h after the in ammatory challenge (Fig. 2). Treatment of animals, 30 min before the in ammatory stimuli, with 1 mg/ kg (i.v.) of Cr atyli a oribund a, D. gui anensis or D. viol ace a lectins strongly reduced (P , 0·05) the fMLPinduced neutrophil migration, by 73, 84 and 100%, respectively ( Fig. 2A). Conversely, none of these lectins inhibited zymosan-induced neutrophil migration (Fig. 2B). The neutrophil migration induced by fMLP or zymosan was strongly inhibited by 30 mg/ kg of LPS (P , 0·05). D. violacea seed lectin effect upon carrageenan(Cg)-and fMLP-induced neutrophil migration and the involvement of carbohydrate residues on its inhibitory activity D. viol ace a lectin was used to investigate the involvement of mannose and glucose residues as a site for the inhibitory effect of lectins. This lectin was chosen because it showed the highest activity at 1·0 mg/ kg. At this dose, the lectin from D. viol ace a, when injected 30 min before Cg inhibited the neutrophil migration seen 4 and 8 h following i.p. administered Cg (Fig. 3A).  3B). Co-injection of LPS (30 mg/ kg) with 0.1 M of a-methyl-mannoside did not change the LPS inhibitory effect on the fMLP-induced neutrophil migration model (Fig. 4). Thus, we can rule out the possibility that the D. viol ace a lectin activity was caused by endotoxin contamination.
Effect of LPS and plant lectins on rat paw oedema induced by carrageenan (Cg), zymosan and dextran Subplantar injection of Cg (300 mg/ paw) induced a progressive and intense paw oedema that reached a maximal value by 3 h (Fig. 5A). Lectins from D. viol ace a, Cr atyli a oribund a or D. gui anensis seeds, injected i.v. at 1 mg/ kg, 30 min before the chemotactic agent, reduced the Cg-induced paw oedema by 32, 30 and 17%, respectively. LPS (30 mg/ kg) treatment caused a 50% reduction in the paw oedema induced by Cg (Fig. 5B). The inhibitory effect of the D.  (Table 1) or zymosan ( Table  2). LPS (30 mg/ kg) showed no effect upon dextran-induced paw oedema (Table 1) and reduced the zymosan oedematogenic effect by only 25% (Table 2).

Haemagglutinating activity of lectins in rodent erythrocytes
Haemmagglutinating activity of lectins from D. viol ace a, Cr atyli a oribund a, D. gui anensis, D. virg at a, D. br as iliensis and D. gr andi or a seeds was estimated on a 2% suspension of rat, rabbit or mouse erythocytes; haemagglutinating activity differed greatly among animal species. In general, the highest susceptibility to lectin haemagglutinating activity was shown by rabbit erythrocytes. For example, in these cells the lectin haemagglutinating activity of Cr atyli a oribund a was 64-fold higher than that observed for rat erythrocytes. Furthermore, Cr atyli a oribund a and D. viol ace a lectins which showed the highest anti-in ammatory activity (Fig. 1), showed the lowest HU in rat erythrocytes (Table 3).

Discussion
We have demonstrated that some mannose ± glucose binding lectins from seeds of the Dioclin ae show anti-in ammatory activity. Similar to that observed with LPS, the i.v. administration of lectins from D. viol ace a, D. gui anensis, Cr atyli a oribund a and D. virg at a seeds inhibited both the neutrophil migration into rat peritoneal cavities induced by fMLP or carrageenan, and rat hind paw oedema induced by carrageenan. Among the lectins tested the highest anti-in ammatory activity was shown by D. viol ace a lectin (DvL), the anti-in ammatory effect of which could be blocked by a-D-methylmannoside (a speci c lectin binding sugar) but not by a-D-galactose. When administered alone these sugars showed no anti-in ammatory activity (data not shown); this is in accordance with the demonstration that D-mannose had no signi cant effect on the neutrophil migration induced in vivo or in vitro by fMLP. 28 Furthermore, a-D-methyl-mannoside did not modify the LPS inhibitory effect on fMLP-induced neutrophil migration showing that DvL activity was not due to endotoxin contamination. These data are in favour of a speci c lectin activity. However, D. gr andi or a and Can av ali a br asiliensis seed lectins, although belonging to the same mannose ±glucose binding group, were devoid of any inhibitory activity. In addition, we also observed that concanacalin A, a well-known member of this group of plant lectins, did not inhibit fMLP-induced neutrophil migration (data not shown), in spite of the demonstration of its immunosupressive effect, 29 and inhibitory activity on experimental allergic encephalomyelitis

Mediators of In¯ammation´Vol 6´1997 207
in guinea pigs and rats. 30 Discrepancies in the capacity of mannose binding plant lectins to induce human lymphocyte stimulation, interferon-gamma release by human peripheral mononuclear leukocytes, 16 and rat mast cell degranulation, 15 were also observed. Interest-ingly, the lectins from the Dioclin ae subtribe plants show very high homology (80 ±90%) with respect to their primary structures. 31 It has been suggested that despite very similar physicochemical properties, differences in the biological activities of lectins might be due to   differences in its ne sugar af nity sites that bind to target effectors. 15 Very recently, it was demonstrated that the crystal structure of Can av ali a br as iliensis lectin suggests a correlation between its quaternary conformation and its biological properties distinct from those of concanavalin A. 32 Our results suggest that the lectin inhibitory activity could be better observed on neutrophil-mediated in ammatory reactions, since all lectins tested, like LPS, had no effect on dextran-induced paw oedema, a classical neutrophil-independent in ammatory agent. 33 The exact mechanism involved in the anti-in amamtory effect of the studied plant lectins is currently unknown. The amount of lectin injected was in excess of the amount required for haemagglutination. However, no clear relationship could be demonstrated between lectin haemaglutinating and anti-in ammatory activities in the rat. DvL and Cr atyli a oribund a lectin showed the lowest haemagglutination activity and the highest anti-in ammatory activities. Beside this, D. gr andi or a and D. gui anensis lectins exhibited similar haemaglutinating activities but only the latter showed anti-in ammatory activity. Moreover, no alterations could be detected in the haematocrit of rats treated i.v. with Dvl compared to normal values (data not shown); no signs of acute toxicity were observed after i.v. administration of lectins, even in doses up to 5-to 10-fold higher than the maximal dose used in this study (data not shown), a nding which parallels previous results showing that concanavalin A, at similar dose range, did not cause acute toxicity. 29,30 Furthermore, the lectin inhibitory activity could not be accounted for by a possible leukopaenic or neutropenic effect, since none of the lectins tested changed, during the rst 4 h, the numbers of circulating leukocytes when compared to saline-treated rats (data not shown). We speculate that these plant lectins are operating against the adhesive activities or selectins, competing with these glycoproteins by carbohydrate ligands on endothelial or leukocyte cell membranes. Which selectin is competitively blocked by these mannose binding lectins is unclear. The leukocyte adhesion mediated by selectins can be reduced by a variety of simple and complex carbohydrates, most of which are sialylated, fucosylated, or both. 5 In humans, the most important sialylated glycoconjugate is the sialosyl-Lewis-X which is expressed in large amounts by neutrophils and is considered the major binding site for selectins. 5 However, since this epitope is highly species speci c and is absent in non-human mammalian species it cannot be considered to be a general binding site for interaction with selectins in non-human leukocytes. 34,13 As the lectin domain of E-selectin aligns closely with that of the mannose binding C-type lectins, 5 we postulated that it would be considered as the rst candidate to compete with mannose binding plant lectins by the same carbohydrate ligands on neutrophil membranes. Corroborating this suggestion, we showed that plant lectin inhibitory activity was not detectable at 2 h after i.p. carrageenan injection, and it has been demonstrated that E-selectin plays essentially no role in the earliest phases (up to 2 h) of leukocyte migration during acute in ammation, because it requires de novo gene transcription. 5 In addition, E-selectin expression by cultured human endothelial cell peaks between 3 and 6 h after TNF-a stimulation, 35 a similar time interval for maximal neutrophil recruitment response following carrageenan injection into the rat peritoneal cavity. The ability of carrageenan to induce TNF-a release, in rats, has also been suggested. 36 However, it must be pointed out that not all neutrophil-mediated in ammatory reactions could be inhibited by these plant lectins. It was observed that lectins were ineffective in inhibiting both neutrophil migration and paw oedema induced by zymosan. These contradictory ndings are dif cult to explain. These data, however, ruled out the possibility that the observed anti-in ammatory activity of the mannose binding plant lectins could be due to a non-speci c effect, such as hypotension or leukopaenia. Furthermore, as zymosan-induced neutrophil migration was unchanged or potentiated by plant lectins, a direct inhibitory effect upon neutrophils is probably not occurring. The possibility that the time interval we choose to test the lectin inhibitory activity was different from that reported for maximal zymosan-induced intraperitoneal neutrophil migration in mice 37 is unlikely, since, in rats, the maximal response occurred 4 h after peritoneal zymosan challenge (data not shown). It is possible, however, that zymosan triggers different neutrophil recruitment mechanisms, not sensitive to plant lectin treatment; zymosan is a well-known powerful agent in inducing the release of oxygen free radicals by macrophages, 37 the predominant peritoneal resident cell in rats, and it has been shown that hydrogen peroxide is able to induce long-lasting P-selectin expression 38 when compared to that induced by thrombin, histamine or complement fragments. 5 Thus we speculate that in zymosan-in amed rat peritoneum Pselectin could play a pivotal role, and that Pselectin ligands, such as P-selectin glycoprotein ligand-1, 39 are probably not recognized by these mannose binding plant lectins.
In conclusion, the present work shows that mannose binding plant lectins can reduce neutrophil-mediated in ammatory reactions induced by different stimuli, such as carrageenan and fMLP. We suggest that lectin anti-in ammatory activity involves competition between plant lectins and selectins for a common carbohydrate ligand present on leukocyte and/or endothelial cell membranes. This hypothesis reinforces the proposition by others that sugarbased inhibitors directed against adhesive activities of selectins might provide for new and more effective anti-in ammatory drugs. 40,10,11 Thus, plant lectins could be used as important and speci c tools for better understanding the role of sugar residues in leukocyte recruitment, as well as a template in the study of molecular interactions between selectins and their carbohydrate ligands.