Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Several chronic inflammatory lung diseases are characterized by the presence of leukocytic infiltrates of neutrophils, lymphocytes, macrophages, eosinophils, mast cells, alveolar damage, and edema, which all lead to decreased respiratory function (1). To infiltrate the peribronchial tissue and to promote inflammation, the leukocytes must extravasate from blood circulation through the vascular endothelial layer into the tissue (2). Extravasation of leukocytes is initiated by an interaction of selectins and their oligosaccharide-containing ligands, which leads to tethering and rolling of leukocytes on the inner surface of the vascular wall. L-selectin is constantly expressed on mature leukocyte surfaces and recognizes its endothelial counterreceptors, such as CD34, provided that they are decorated with 2,3sialylated, 1,3fucosylated, and sulfated lactosamines (2). Other L-selectin ligands (e.g., GlyCAM-1) are expressed only on high endothelial venules (HEVs) in lymph nodes. These ligands on HEVs constantly express glycans that fulfill the requirements listed earlier, the prototype decoration being sulfated sialyl Lewis x (sLex) for selectively binding L-selectin on lymphocytes (3, 4). On the contrary, endothelium in other sites does not express proper glycoforms of L-selectin ligands under normal conditions. Our previous work in both in vitro and in vivo models, as well as observations made with specimens obtained from heart transplant patients, showed that proinflammatory stimuli can induce endothelium to express these glycans de novo and thus promote lymphocyte extravasation (5, 6).

We now extend these observations into patients with chronic lung diseases and analyze whether the infiltrates of lymphocytes and other white blood cells observed here are due to de novo-induced endothelial sLex expression. The number of the vessels was determined with monoclonal antibody (mAb) anti-CD31 and the presence of functionally active L-selectin ligands was analyzed with a panel of mAbs recognizing sulfo-sLex glycosylations on, e.g., L-selectin. Our results show that first, the number of vessels present in the peribronchial area of lung specimens per unit area increased 2-fold as analyzed by the expression of CD31 in all inflammatory specimens. Second, endothelium in peribronchial tissue did not express the sulfo-sLex glycomodifications in specimens with normal histology. On the other hand, the expression of sulfo-/sLex-epitopes as detected with mAbs 2F3, HECA-452, and MECA-79 increased significantly in specimens obtained from patients with bronchial asthma. Further, these specimens obtained from asthma patients differed significantly from all other chronic inflammatory lung diseases, such as adult respiratory distress syndrome (ARDS), chronic bronchitis, fibrosing alveolitis, and granulomatous inflammation, as practically no or only very mild expression of sulfo-sLex was observed in them.

These data suggest that the upregulation of endothelial sulfo-sLex glycans acting as functionally active L-selectin ligands is connected to accumulation of leukocytes in the peribronchial tissue of patients with bronchial asthma. Thus, inhibition of the carbohydrate-dependent entry of leukocytes via L-selectin may provide a putative novel target for immunosuppressive therapies without targeting the activation and proliferation of T cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Specimens

The bronchial biopsies taken from patients with bronchial asthma, n = 26, 37 ± 3 yr (mean ± standard error of the mean [SEM]), 11:15 male-female ratio, were divided into five groups: Group 1 (n = 11, 37 ± 4 yr, 6:5) consisted of patients who used only short-acting ₂-agonist when needed. Group 2 consisted of patients who were treated with the sodium cromoglycate (Lomudal; Aventis, Strasbourg, France) (n = 5, 38 ± 3 yr, 1:4). Patients in Group 3 used flutikasone (Flixotide; Glaxo-Wellcome, Middlesex, UK) (n = 4, 37 ± 5 yr, 1:3). Patients in Group 4 had salmeterol treatment (Serevent; Glaxo-Wellcome) (n = 2, 39 ± 0 yr, 1:1), and Group 5 represented patients with occupational asthma (n = 4, 54 ± 2 yr, 2:2). The bronchoscopic examination and biopsy-taking conformed to international guidelines (7). Bronchial biopsies were performed according to our previously used methods on patients with suspected bronchial asthma (8).

Control samples from adult human lungs, n = 10, 55 ± 4 yr (mean ± SEM), 3:2 male-female ratio, were obtained from the clinically uninvolved areas of lungs removed for lung hamartoma. Open-lung biopsies were obtained to confirm the diagnosis of chronic bronchitis (n = 19, 48 ± 4 yr, 10:9), fibrosing alveolitis (n = 11, 56 ± 5 yr, 7:4), or granulomatous inflammation (n = 20, 50 ± 3 yr, 13:7). The ethiologic subgroups of granulomatous inflammation, i.e., sarcoidosis (n = 11) and tuberculosis (n = 9), were considered as one group because the results did not differ significantly from each other (P > 0.05, Mann-Whitney U test). Lung specimens were obtained from the autopsies of patients with ARDS (n = 12, 62 ± 4 yr, 2:1). The reactivity of mAb anti-CD31 was studied by reduced numbers of samples in the following patient groups because of the statistically significant results even with fever samples: chronic bronchitis (n = 12, 51 ± 6 yr, 2:1), granulomatous inflammation (n = 14, 46 ± 3 yr, 4:3), and control subjects (n = 7, 50 ± 4 yr, 4:3). The normal lymph nodes that had been removed for diagnostic purposes were studied as positive control samples. All specimens were formalin-fixed and paraffin-embedded, and processed for routine histologic diagnosis.

Antibodies

The glycan epitopes on L-selectin ligands identified by the mAbs used in the present study are described in Table 1. Both mAbs 2F3 (9) (5 µg/ml; kindly provided by R. Kannagi, Aichi Cancer Center, Nagoya, Japan) and HECA-452 (10) (2 µg/ml; kindly provided by S. Jalkanen, University of Turku, Turku, Finland) are anti-sLex mAbs, and they both require the presence of both 2,3 sialylation and 1,3 fucosylation of the lactosamine to recognize their antigens. mAb MECA-79 (11) (1:100 culture supernatant; also from S. Jalkanen) requires 6-sulfation of the GlcNAc-residue of N-acetyllactosamine on a family of endothelial proteins for proper recognition. CD31 is a member of the immunoglobulin (Ig) gene superfamily that is expressed among other things in the endothelial intercellular junction. Anti-CD31 (0.5 mg/ml, clone 1A10; Novocastra Laboratories Ltd., Newcastle upon Tyne, UK) was used as positive control for vascular endothelium. Isotype-matched mouse and rat IgMs (15 µg/ml; PharMingen, San Diego, CA) as well as mouse IgG (1:5 culture supernatant; kindly provided by O. Vainio, University of Turku) were used as negative control reagents.

Immunohistochemistry

Lung biopsy sections 5 µm thick were deparaffinized and rehydrated by rinsing them with xylene (five times for 5 min each time), 100% ethanol (twice for 5 min each time), 96 and 70% ethanol (for 5 min each), and aqua (twice for 5 min each time). One set of slides (mAbs 2F3, HECA-452, and MECA-79) was first incubated for 60 min at 37°C with 0.5% trypsin in phosphate-buffered saline (PBS), pH 7.5. The trypsin was then rinsed with PBS (three times for 10 min each time). The other set of slides (mAb anti-CD31) was first heated in a microwave oven with 10 mM citrate buffer (450 W, three times for 5 min each time), and then the sections were cooled at room temperature in the citrate buffer for 2 h. The peroxidase reaction in lung tissue was prevented by incubating both sets of slides first with 5% H₂O₂ in aqua (for 5 min); the H₂O₂ was then rinsed first with aqua and next with PBS (for 5 min each). A Histostain-Plus kit (Zymed Laboratories, Inc., San Fransisco, CA) was used for immunoperoxidase staining according to the instructions of the manufacturer.

The reactivities of mAbs anti-CD31, 2F3, HECA-452, and MECA-79 were evaluated by two blinded observers who had no knowledge of the pathologic diagnosis of the specimens. The number of positive vessels was determined per square millimeter, and the capillaries, arterioles, and venules were distinguished on the basis of histology. The numbers of positive venules and capillaries were combined because their numbers were approximately equal in every specimen. Arterioles did not react and thus were not included. Both the number of reactive vessels and the staining intensity of each positive vessel were determined (0 = no reactive vessels, and 1 = mild, 2 = moderate, and 3 = strong reaction) and the mean value for each patient sample was calculated. The whole specimen was analyzed but only peribronchial regions were included in the analysis. The mean surface area of peribronchial regions per specimen was 68 ± 7.9 mm² (± SEM). One or two independent sections were evaluated for each patient. All fields with peribronchial regions of the sample were included in the analysis, thus the areas of leukocyte infiltration were not favored. The region of caseating necrosis of specimens from patients with tuberculous granulomatous inflammation was not analyzed, nor were regions with bronchial-associated lymphatic tissue (BALT).

Statistical Analysis

The results were analyzed by the Kruskal-Wallis one-way analysis of variance by ranks. This test is basically a comparison of the medians of several unpaired groups and the null hypothesis is that the medians are equal. P < 0.05 was considered significant. We used Dunn's test for multiple comparisons of medians in control and lung inflammation groups. The null hypothesis is that the medians are equal. Where appropriate, Dunn's test was used for comparing medians of pooled groups. Thus, the reactivity of mAb anti-CD31 was analyzed in the control group and the pooled group of all inflammatory diseases and the reactivities of mAbs 2F3, HECA-452, and MECA-79 were assessed in bronchial asthma and the pooled group of the other inflammatory diseases. In Dunn's test, differences were considered significant at P < 0.01.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

The expression of an endothelial intercellular junction molecule, also known as platelet endothelial cell adhesion molecule-1, was analyzed in the specimens with mAb anti-CD31. This was used as a marker to determine the number of vessels per square millimeter because this molecule has been shown to be expressed on all vascular endothelium (12). In control specimens representing normal lung histology 41.9 ± 6.5 vessels/mm² (mean ± SEM) were reactive with mAb anti-CD31. The numbers of mAb anti-CD31- positive vessels increased in the peribronchial area of specimens with lung inflammation, and the mean value for all pooled specimens with inflammatory conditions was significantly elevated (80.9 ± 2.4 vessels/mm², P < 0.001 with Dunn's test) (Figure 1).

View larger version (33K): [in this window] [in a new window]   Figure 1.   Distribution of the number of positively stained vessels in controls (Co) and five groups of inflammatory lung diseases. Compared with the controls, the numbers of positively reacted vessels with mAbs 2F3, HECA-452, and MECA-79 were significantly increased in bronchial asthma (Ba) (P < 0.003, Dunn's test), whereas in the other groups of inflammatory diseases the increase was not significant. Cb, chronic bronchitis; Fa, fibrosing alveolitis; Gi, granulomatous inflammation. Horizontal lines denote mean values; a circle symbolizes one sample and a box denotes five samples.

High endothelium in lymph nodes expressed specific glycoforms of L-selectin ligands, fulfilling most or all of the reqirements (i.e., a prototype decoration being a sulfated sLex) for L-selectin recognition. However, the same protein or lipid backbones of these L-selectin ligands are not properly glycosylated in other organs under normal conditions (13). Because there is not a single antibody available to recognize all known decorations of the L-selectin ligands, we needed to probe the differently decorated L-selectin ligands with a panel of mAbs. The reacting epitopes of antibodies used in this study are shown in Table 1. mAbs 2F3 and HECA-452 detect 2,3 sialylation and 1,3 fucosylation of lactosamine and do not distinguish between sulfation and nonsulfation (10). On the other hand, mAb MECA-79 needs sulfation of its epitope for proper recognition (11). The ligands with highest relative binding affinity toward L-selectin would possess all these decorations (2). In addition to L-selectin ligands, these mAbs also recognize sLex-related (2F3, HECA-452) or sulfated (MECA-79) moieties on other molecules.

Different decorations of L-selectin ligands were detected by immunohistochemical analysis of the various lung biopsy specimens with all three mAbs. The intensity of endothelial staining was rated in a semiquantitative manner from 0 up to 3, and the number of reactive vessels per square millimeter in each sample was determined as well. The first antibody used in this analysis was anti-sLex mAb 2F3, which did not react with endothelium in control specimens with normal histology (mean 0). Instead, in specimens from patients diagnosed with bronchial asthma the number of mAb 2F3 reactive vessels was significantly increased (3.7 ± 0.92 vessels/mm²) compared both with control specimens (P = 0.0015, Dunn's test) and with the pooled specimens with the other inflammatory diseases (P < 0.0001). Because the peribronchial tissue had 80 vessels/mm², the proportion of mAb 2F3-reactive vessels represented approximately 5% of the total number of vessels. It should be noted that only capillaries and venules expressed sulfo-sLex-decorated L-selectin ligands, whereas the arteries or arterioles did not express these glycodecorations. We also analyzed the grade of endothelial reactivity with mAb 2F3 in sections from patients with bronchial asthma. The proportion of specimens with negative, mild, or moderate staining intensity with mAb 2F3 were 30, 30, and 33%, respectively (Figure 2). However, strong staining intensity with mAb 2F3 was observed only in the 7% of samples. No significant differences were observed between the patients of the five subgroups of bronchial asthma (data not shown). The number of mAb 2F3 reactive vessels was not significantly increased in specimens from patients with ARDS (mean 0.6 ± 0.37 vessels/mm²) and granulomatous inflammations (1.2 ± 0.83/mm²) compared with control values. The endothelial reactivity of mAb 2F3 in specimens from patients with chronic bronchitis and fibrosing alveolitis was practically negative (Figures 1, 3, and 4).

View larger version (40K): [in this window] [in a new window]   Figure 2.   Distribution of the staining intensity of anti-sulfo-sLex mAbs in bronchial asthma specimens. The staining intensity of each positive vessel was determined and the mean value of a sample was counted. 0 = no positive vessels, 1 = weak, 2 = moderate, and 3 = strong staining.

View larger version (18K): [in this window] [in a new window]   Figure 3.   Expression of L-selectin ligands on capillary and venular endothelium in controls (Co) and different inflammatory lung diseases. The y-axis represents the mean value of positively stained vessels per sample. The reactivities of anti-sLex mAbs (2F3, HECA-452) and antisulfated mAb (MECA-79) were greater in all groups of inflammatory diseases. Compared with the control samples, the number of positive venules stained with all tested mAbs increased significantly only in bronchial asthma (Ba) (*P < 0.003, **P < 0.0003, Dunn's test). Further, compared with the pooled group of the other inflammatory diseases, the expression of L-selectin ligands detected by these mAbs increased most in bronchial asthma (***P < 0.0001, Dunn's test). The minor increase of positive vessels in the other groups of inflammatory lung diseases was not significant. Cb, chronic bronchitis; Fa, fibrosing alveolitis; Gi, granulomatous inflammation. The y-error bars represent SEM.

View larger version (127K): [in this window] [in a new window]   Figure 4.   Expression of endothelial L-selectin ligands on human lung biopsies detected by immunohistochemistry. (A) A biopsy of a normal lung tissue close to hamartoma, which represents a control sample, showed no reaction with an isotype-matched murine IgM antibody. The arrow points to a capillary. Scale bar, 100 µm. (B) This biopsy represents bronchial asthma with recruitment of peribronchial leukocytes, yet the control murine IgG antibody showed no endothelial reactivity. The arrow points to a venule. (C) mAb HECA-452 identifies sLex modifications on endothelial cells during inflammation and therefore did not react with the control sample. (D) mAb anti-CD31 detects a member of the Ig superfamily on endothelial cells in a control sample. The arrow points to a venule. (E) mAb anti-CD31 shows a strong reaction in peribronchial venules and capillaries in a bronchial biopsy of asthma. (F ) mAb 2F3 detects sialylation and fucosylation modifications relevant to L-selectin-mediated adhesion of a venule (arrow) in bronchial asthma tissue. Moreover, mAb 2F3 reacts with mucin-secreting goblet cells of bronchial epithelium (not presented here). (G) In a biopsy of bronchial asthma two venules are partly positive with mAb HECA-452, i.e., only one or two of endothelial cells are stained positively (arrows). This antibody also reacts with mucin-secreting goblet cells of bronchial epithelium (not presented here). In addition, the very slight background reaction is probably due to the unspecific reaction of mucous material. (H ) mAb MECA-79 stained strongly sulfated lactosamine structures as L-selectin ligands both on capillaries and venules in bronchial asthma. (I ) This implicates ARDS with diffuse damage of the alveolar wall as well as the positive reaction of vessels with mAb HECA-452. The arrows point to venules. (J ) Bronchiolar injury (arrow) and increase in inflammatory cells characterize chronic bronchitis. The peribronchial capillaries and venules reacted with mAb MECA-79. (K ) The accumulation of leukocytes and the alveolar damage and fibrosis are seen in a sample of fibrosing alveolitis. Several strongly mAb 2F3-positive vessels can also be detected. (L) Reaction of venules and capillaries with mAb MECA-79. The arrows point to giant cells that are characteristic of granulomatous inflammation.

The other anti-sLex antibody, mAb HECA-452, was also nonreactive with specimens with normal histology. Here again, in contrast to controls, a strong induction in the endothelial mAb HECA-452 reactivity was found in specimens from bronchial asthma patients (3.8 ± 0.72 vessels/mm²). This level of mAb HECA-452 expression was significantly elevated in specimens from patients with bronchial asthma (P = 0.001, Dunn's test) compared with the normal control specimens and with the pooled specimens with all other inflammatory diseases (P < 0.0001). The mAb HECA-452-expressing vessels represented some 5% of the total number of vessels (Figure 3). As seen in Figure 3, all the inflammatory specimens other than bronchial asthma were essentially negative for mAb HECA-452. We also analyzed the grade of endothelial reactivity with mAb HECA-452 in sections with bronchial asthma. The distribution of specimens with negative reaction or mild or moderate staining intensity with mAb HECA-452 were essentially similar to the distribution of staining intensities with mAb 2F3 (Figures 2 and 4).

To analyze sulfated variants of functionally active L-selectin ligands on endothelium the anti-sulfo-sLex antibody mAb MECA-79 was used. mAb MECA-79 did not react positively with endothelium in specimens with normal histology (mean 0), and the number of mAb MECA-79-positive vessels remained low in specimens obtained from the inflammatory conditions other than bronchial asthma (Figure 3). On the contrary, strong endothelial reactivity was observed in specimens from patients with bronchial asthma (5.2 ± 1.3 vessels/mm²); the number was significantly higher compared with controls (P = 0.0003, Dunn's test) as well as compared with the pooled specimens with the other inflammatory diseases (P < 0.0001). This endothelial mAb MECA-79 reactivity in the asthma specimens represented roughly 7% of the total number of vessels. The staining intensity of mAb MECA-79 was mild in 30%, moderate in 19%, and strong in 4% of the asthma samples (Figure 2).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The present study with 98 histologic lung samples shows that normal control specimens did not express sLex and/or sulfated glycoforms detected by mAbs 2F3, HECA-452, and MECA-79. Negative or very low levels of these glycans on L-selectin ligands were expressed on venules and capillaries in the following groups of chronic lung inflammation: ARDS, chronic bronchitis, fibrosing alveolitis, and granulomatous inflammation. In contrast, specimens obtained from patients with bronchial asthma showed significant increase of endothelial expression of mAbs 2F3, HECA-452, and MECA-79, i.e., putative functionally active L-selectin ligands, in comparison with controls and other inflammatory diseases. No significant differences in the endothelial expression of L-selectin ligands were observed in specimens from patients with nontreated versus treated bronchial asthma, most probably due to the small n-values of the five groups (data not shown). There was a clear qualitative correlation between endothelial reactivity detected with all three mAbs and the lymphocyte infiltrates in the vicinity of reactive vessels. The regions defined histologically as BALT were not included in the analysis because BALT is lymphatic tissue. However, BALT as well as normal peripheral lymph nodes used as positive control samples had very strong endothelial reactivity in venules and capillaries and a high number of positive vessels per square millimeter detected by all three mAbs. The increase of CD31-positive vessels in peribronchial tissue taken from patients with inflammatory diseases might be due to the formation of new vasculature during inflammation (14), or it might be due to the fact that mucosal areas are more vascularized than submucosal areas (15), the latter being putatively more frequent in the control specimens.

According to our knowledge this is the first demonstration of increased expression of sulfated and/or sLex-decorated ligands in peribronchial endothelium of asthma. This is in accordance with our previously published results of another lymphocyte-mediated inflammation, i.e., rodent models of heart and kidney allograft rejection (5, 16), as well as a heart transplant patient study with endomyocardial biopsies (6). The de novo induction and upregulation of endothelial sLex glycans at the onset of both animal and human allograft rejection suggest that this pattern of regulation is crucial to the site-directed lymphocyte extravasation into the allograft and thus to the generation of local inflammation. Further, the downregulation of the expression of the bioactive glycoforms of selectin ligands at the time the rejection resolved is linked with the decrease of lymphocytic infiltrate in the allograft, which indicates that the graft endothelium has the machinery to induce and decrease the expression of these glycans (6).

The reason for using a panel of antibodies to detect various parts of decoration of L-selectin ligands was that no single antibody has been shown solely to detect all these glycoforms. We selected our anti-sLex antibody panel according to our previous experience with endomyocardial biopsies (6). The original antigens in mAb 2F3 and mAb HECA-452 preparation were an sLex containing glycolipid and stromal components of human lymph nodes. MAb 2F3 is able to inhibit in vitro the L-selectin-dependent adhesion of lymphocytes to human lymph node HEV, whereas mAb HECA-452 is not capable of blocking this (10). The mAb HECA-452 epitope, which has also been termed as cutaneous lymphocyte antigen, is an sLex-decorated variant of P-selectin ligand glycoprotein-1 on the skin-homing lymphocytes (17). Both mAbs 2F3 and HECA-452 detect sLex-related structures also, for example, on P- and E-selectin ligands, however in most cases these are not expressed on endothelium. mAb MECA-79 was originally defined to recognize an antigen entitled peripheral lymph node adressin (18). This is actually a family of sulfated glycoproteins, many of which are able to bind L-selectin (19). Currently, there is not a single antibody that can recognize glycans on endothelial L-selectin ligands.

We selected five different chronic lung inflammations and compared the expression of L-selectin ligands in the inflammed peribronchial area of bronchial or lung biopsies. Although ARDS is histologically characterized by diffuse epithelial and endothelial injury in alveolar wall, which leads to increased capillary permeability and infiltration of leukocytes (principally neutrophils), no significant increase of expression of endothelial L-selectin ligands was detected in our study. On the other hand, neutrophil-mediated lung injury of animal models that resemble acute ARDS have been demonstrated to depend on L-, P-, and E-selectins (20). In addition, prophylactic treatment with sLex-oligosaccharides as well as with anti-P- and anti-L-selectin mAbs prevented alveolar accumulation of leukocytes and acute lung injury resembling ARDS (23, 24). Inhaled heparin, which has been shown to inhibit L- and P-selectin-mediated leukocyte binding (25), has been reported to be advantageous in ARDS (26). Fibrosing alveolitis is characterized by the presence of principally mononuclear cells within the interstitium and alveolar spaces (1). Soluble E-selectin increased and was associated with an increased number of lymphocytes in bronchoalveolar lavage fluid of patients with fibrosing alveolitis, however, the expression of endothelial E-selectin did not differ from controls (27). The hallmarks of chronic bronchitis are hypersecretion of mucus, increased numbers of goblet cells, and inflammatory infiltration with pigmented alveolar macrophages. In the bronchial mucosa of patients with chronic bronchitis or sarcoidosis, the increased expression of endothelial E-selectin has been documented (28, 29). On the other hand, we were not able to detect significant increase of endothelial L-selectin ligands. Granulomatous inflammation is characterized by caseating (tuberculosis) or noncaseating (sarcoidosis) granulomas consisting of mononuclear epitheloid cells surrounded by and interspersed with lymphocytes (1). Collet and Munro found endothelial MECA-79 reactivity in specimens from patients with tuberculosis and in other lung conditions (bronchiectasis, interstitial pneumonia), but not in specimens from patients with sarcoidosis and some other lung conditions (30). The positive reaction was detected in only a minority of specimens from patients with tuberculosis and on average only a relatively small proportion of microvessels was positive, which is in accordance with our results of small but not significant increase in the reactivity of mAb MECA-79 in specimens from patients with granulomatous inflammation. Thus, the lymphocyte infiltration, which is characteristic of both subgroups of granulomatous inflammation, may be due to the increased expression of other adhesion molecules.

A consistent feature of the asthmatic response is produced by a combination of mucosal edema, smooth-muscle contraction, epithelial damage, thickening of basement membrane, and excessive production of bronchial secretions as a result of the selective recruitment of eosinophils, mast cells, and lymphocytes, and the release of chemical mediators (31, 32). The greatly increased peribronchial lymphocyte population may control the recruitment and activation of eosinophils by producing several cytokines (32). Allergen challenge or cytokines interleukin-1, tumor necrosis factor-, and interferon- released by local macrophages upregulate the expression of E-selectin and result in an influx of inflammatory cells (33). In addition, increased expression of L-selectin ligands in peribronchial vessels may influence mediator release by enhanced L-selectin binding during asthmatic response (34). The treatment with inhaled heparin, which is an L-selectin ligand, has been shown to inhibit eosinophil recruitment and to reduce the early or late asthmatic response after allergen inhalation or exercise provocation in several studies of patients with asthma (26). Further, treatment with an L-selectin-specific mAb, DUI-29, as well as with selectin-binding inhibitors (TBC-1269, MBPA) reduced early and/or late airway response in allergic sheep after antigen challenge (34). In addition, it has been shown that the eosinophil infiltration and the late asthmatic response in guinea pigs can be inhibited by intravenous pretreatment with sLex analog before antigen challenge (35). This would indicate that the recruitment of lymphocytes and eosinophils to the site of inflammation in asthma is L-selectin- and sulfo-sLex-dependent. Further, this is in accordance with our previous findings that multivalent sLex analogs inhibit in vitro lymphocyte adhesion on rejecting rat heart allograft vessels (5).

It can be concluded that selectin-dependent leukocyte extravasation could be a crucial step in the initiation of leukocyte inflammatory reactions, as our previous work on acute allograft rejection has indicated. The data presented in this paper broaden our previous observations as well as those of others and demonstrate that a similar strong induction of endothelial sulfo-sLex glycans, i.e., putative decorations on ligands for L-selectin, can be found in lung specimens obtained from patients with bronchial asthma. This induction is specific to asthma inasmuch as no other chronic inflammatory process in lungs enhanced this expression. Not only the inducible expression of endothelial E- and P-selectin but also the inducible endothelial glycoforms of L-selectin ligands seem to guide lymphocyte and eosinophil traffic to the sites of bronchial inflammation in asthma patients, thus providing a novel target to glycan-based immunomodulatory interventions.