Leukocyte Extravasation: An Immunoregulatory Role for α-l-Fucosidase?¹

Fucose can be linked through α (1, 3), α (1, 4), and α (1, 6) bonds to N-acetylglucosamine and through α (1, 2) to galactose in a range of glycoproteins and glycolipids. Fucosylated oligosaccharides and glycoconjugates play roles in several biological events, including the cell-to-cell adhesion processes that mediate inflammation, embryonic development, tumorigenesis/metastasis, and Ag recognition. The important role played by fucose residues has led to an increasing interest in α-l-fucosidase (ALF)³ (α-l-fucoside fucohydrolase; EC3.2.1.51), an exoglycosidase involved in the hydrolytic degradation of fucose-containing components of glycoproteins (1), glycolipids, and oligosaccharides (2).

Approximately 10–20% of total ALF activity occurs on the surface of a range of human cell types, including hemopoietic, epithelial, and mesenchymal cells (1). Mammalian ALF are multimeric, containing glycoprotein subunits of ∼53 kDa; they have a maximal activity between pH 4 and 6.5. At least five major isoforms of the ALF enzyme have been identified (3). There is a considerable degree of structural heterogeneity, both tissue specific and within tissues. This has been attributed to variations in the sialic acid content as well as the two different alleles and polymorphisms within the FUC1 gene. The serum ALF activity in normal individuals is reported to be 381 U/ml ± 10.86 (4).

The ALF enzyme is associated with a variety of biological functions. The significance of this enzyme in human catabolism is implied by the genetic neurovisceral storage disease, fucosidosis, which results from an absence or gross deficiency of ALF; this allows accumulation of fucoglycoconjugates, which result in mental and motor retardation (5, 6). A syndrome previously known as leukocyte adhesion deficiency II is now recognized as a generalized metabolic disease caused by severe hypofucosylation of glycoconjugates, including selectin ligands (7). Further to this, alterations in levels of ALF have been reported in the plasma and/or serum of patients with endometrial, hepatocellular, and colorectal cancer (4, 8, 9).

Several molecules involved in leukocyte transmigration, including selectin ligands, proteoglycans, integrins, and CD31, are glycosylated and contain fucosylated moieties. L-selectin-dependent interactions mediate lymphocyte tethering and rolling before chemokine-dependent cell activation and transmigration across high endothelial venules. On lymphocytes, L-selectin is a carbohydrate-binding protein that binds to ligands on high endothelial venules; the function of these ligands is dependent on the expression of carbohydrate 6-sulfo sialyl Lewis × (sialyl-(2–3)-galactopyranosyl-(1–4)-Nacetyl glucosamine-((1–3)-fucopyranosyl) groups termed sLe^X or CD15s. The pivotal function of fucosylation of L-selectin ligands has been established by genetic studies in mice. Mice lacking fucosyltransferase VII or both fucosyltransferase VII and fucosyltransferase IV show a reduction of 80% or more than 95%, respectively, in lymphocyte homing to lymph nodes (10).

P- and E-selectins are expressed on activated endothelial cells and cooperatively mediate leukocyte rolling. They are membrane-bound C-type lectins that bind to cell surface glycoconjugate ligands. P-selectin glycoprotein ligand-1 (PSGL-1), a mucin expressed on leukocytes, is the best-characterized selectin ligand. The receptor function of PSGL-1 depends on two posttranslational events. One consists of the generation of O-linked glycan and includes α (1, 3) fucosylation and α (2, 3) sialylation (such as sLe^X). The other involves sulfation of tyrosine residues. In vivo, PSGL-1 mediates leukocytes tethering to and rolling on P-selectin and supports tethering to E-selectin (11, 12). Fucosylated proteoglycans also play a key role in CD11d/CD18-mediated adhesive interactions and the migration of polymorphonuclear leukocytes across the intestinal epithelium (13). Additionally, systemic treatment with fucoidin, a sulfated polysaccharide containing l-fucose and l-fucose-4-sulfate, interferes with selectin receptor function and inhibits leukocyte rolling (14, 15, 16).

It is well established that the responsiveness of migratory inflammatory cells to chemokines is governed by the expression of chemokine receptors, but the processes involved in the resolution of the inflammatory response are poorly defined. It is believed that chemokines are normally presented to intravascular leukocytes in the form of complexes bound to endothelial cell surface glycosaminoglycans. Significantly, it has been shown that the concentration of ALF and soluble glycosaminoglycans increases in Grave’s disease (17), suggesting the potential for cleavage of proinflammatory chemokine complexes and selectin ligands from the endothelial cell surface.

Based on these observations, we hypothesized that ALF plays an important regulatory role in the later stages of inflammation. A series of experiments was designed to examine the role of ALF in leukocyte migration both in vitro and in vivo. Transendothelial chemotaxis assays were used to determine the functional significance of ALF in regulating leukocyte trafficking in vitro. A murine experimental autoimmune uveoretinitis (EAU) model was then used to assess the role of ALF in leukocyte rolling and tissue infiltration in vivo.

THP-1 monocytic cells were supplied by European Cell Culture Collection and expanded in RPMI 1640 medium with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, and 1 μg/ml folic acid (all from Sigma-Aldrich). EAhy926, a human aortic endothelial cell hybridoma created by fusion of HUVECs with human lung carcinoma-derived cell line A549 (18), was grown in DMEM medium (Sigma-Aldrich) supplemented with FCS (10%), 2 mM glutamine, penicillin, and streptomycin, as above.

EAhy926 cells (5 × 10⁴) were seeded onto 3-μm-pore Falcon Transwell filters and allowed to grow to confluency; the endothelial monolayers were then treated for 1 h at 37°C with either no ALF, an optimal 0.08 U/ml ALF (Sigma-Aldrich), or 0.08 U/ml ALF, which had been denatured by heating to 100°C for 5 min.

THP-1 cells were assayed for their ability to migrate through a 3-μm-pore-diameter filter toward the basal compartment containing either no CCL5 or 100 ng/ml CCL5 (PeproTech). THP-1 cells were stimulated for 16 h using 300 U/ml IFN-γ, and then 5 × 10⁵ cells were added to each well and incubated for 2 h at 37°C with 5% CO₂. For some experiments, monocytic THP-1 cells and EAhy926 cells were also treated with 0.08 U/ml ALF for 1 h. Following the removal of excess cells from both the upper and lower chambers, the upper surface of each filter was carefully wiped with a clean swab. The filters were then fixed in 100% cold methanol for 1 h. After fixation, the filters were stained with Gill’s hematoxylin (Sigma-Aldrich) for 30 min, followed by a 5-min wash in Scott’s tap water (166 mM MgSO₄, 24 mM NaHCO₃). Sequential washing and dehydration using increasing concentrations of ethanol were then performed before mounting with DPX mounting medium. Migrant cells were counted using high power microscopy (×400) in four random fields on each filter.

THP-1 cells (200,000/ml) were washed twice with PBS containing 5% FCS. Cells were incubated with biotin-conjugated fucose-specific lectin from Tetragonolobus purpureas (50 μg/ml; Sigma-Aldrich) in HBSS buffer containing 1 mM CaCl₂, 1 mM MgCl₂, and 1% FBS for 1 h at 37°C. Cells were subsequently washed twice with HBSS buffer and resuspended in 50 μl of HBSS buffer containing streptavidin-FITC-conjugated Ab (BD Pharmingen) for 30 min at 4°C. Following washing, the stained cells were analyzed by flow cytometry (FACSort; BD Biosciences).

Cells (100,000/ml) resuspended in RPMI 1640 containing 0.1% BSA and 2% HEPES were treated with ALF (0.5-0.04 U/ml) for different time periods (1–24 h). These cells were then stained with the fucose-binding lectin, as described above. Trypan blue exclusion was assessed to ensure cell viability. Controls included cells stained with secondary Ab only. The stained cells were analyzed by flow cytometry (FACSort; BD Biosciences); data analysis was performed using Lysis II software (BD Biosciences).

Flat-profile 96-well ELISA plates (Linbro; MP Biomedicals) were coated with goat anti-human IgG Fc Ab (Sigma-Aldrich) at 1 μg/well in PBS at 4°C for 18 h. Excess Ab was removed by washing twice with PBS, and human P-selectin.Fc fusion protein (1 μg/well; R&D Systems) was added in HBSS buffer (Sigma-Aldrich) for an additional 18 h at 4°C. Excess protein was washed twice with HBSS buffer, and nonspecific binding was blocked for 2 h at 37°C with HBSS containing 2% FBS.

THP-1 cells (100,000/ml) resuspended in RPMI 1640 containing 0.1% BSA and 2% HEPES were treated with ALF (0.08 U/ml; 8 h), and untreated cells were incubated in the same medium for 8 h. Viability of treated cells was 95%; equal number of viable cells from control and treated group was used in subsequent experiment. THP-1 cells (2 × 10⁶/ml) were labeled with 2′,7′-bis(2-carboxyethyl)-5-(6)-carboxyfluorescein tetrakis (acetoxymethyl) ester (Sigma-Aldrich) at 20 μg/ml for 15 min at 37°C. After washing by centrifugation, the cells were resuspended in HBSS, 5 × 10⁵ cells were added to each assay well, and the plates were centrifuged at 60 × g for 2 min at 37°C.

Following incubation of the assay plates at 37°C for 1 h, the nonadherent cells were removed by mechanical oscillation and gentle washing with warm HBSS, and the remaining, adherent T cells were lysed by the addition of 2% Triton X-100 (Sigma-Aldrich). The quantity of released fluorochrome was measured using a plate-format fluorometer (Fluostar Optima; BMG Labtech).

Female B10R.III mice, 12–20 wk old (Medical Research Facility, University of Aberdeen), were treated humanely according to the Animals (Scientific Procedures) Act (United Kingdom). EAU was induced with a s.c. 50-μl injection into each thigh of 25 μg of peptide 161–180 (SGIPYIISYLHPGNTILHVD; purity >85%; Sigma Genosys) of the human interphotoreceptor retinoid-binding protein emulsified 1:1 with CFA (H37Ra; Difco Laboratories) (19, 20). Animals were observed using a slit-lamp for clinical evaluation (grading) of the eye at day 11 postimmunization.

A single-cell suspension of splenocytes was prepared 11 days after sensitization for EAU. Erythrocytes were lysed with 0.75% (w/v) NH₄Cl in 17 mM Tris buffer (pH 7.2), and the remaining leukocytes were washed twice in RPMI 1640 medium. The cells were then resuspended in 10 ml of medium supplemented with 0.1% (w/v) BSA at 1 × 10⁶ cells/ml, and some samples were incubated at 37°C with 0.08 U/ml ALF per 5 × 10⁵ cells for optimal time. The cells were then pelleted, resuspended in 5 ml of RPMI 1640 plus 0.1% BSA, and incubated with 40 μg/ml calcein-AM (Molecular Probes) at 37°C for 30 min with gentle mixing. Cells were spun down and resuspended in 150 μl of medium.

Eleven days after sensitization, mice of equivalent EAU severity were anesthetized and fitted with contact lenses, as described (21). A total of 100 μl of 0.05% (v/v) sodium fluorescein (Sigma-Aldrich) was injected via the tail vein, followed by 1 × 10⁷ calcein-AM-labeled cells in 150 μl of medium. For each eye, three regions of interest containing one to three veins/venules were observed by SLO, and images were recorded for at least 15 min after cell injection (21). Video analysis was conducted off-line, as described elsewhere (21). Rolling leukocytes and those not interacting with the endothelium were counted in each venule (22, 23, 24). The rolling efficiency was calculated as the percentage of labeled rolling cells among the total number of labeled cells that entered a venule. The sticking efficiency was determined as the percentage of labeled cells becoming firmly adherent for at least 20 s compared with the total number of labeled cells that rolled within the venule during the same time interval.

Fifty minutes after the injection of labeled resting cells and SLO, the anesthetized mice were injected via the tail vein with 100 μl of Evans blue solution (2% (w/v) in PBS; Sigma-Aldrich). After 10 min, animals were killed by terminal anesthesia, and the eyes were immersed in 2% (w/v) paraformaldehyde for 1 h. The retinas were removed, washed twice in PBS for 15 min, spread on clean glass slides, and mounted with the vitreous side upward (25). The mounts were examined using a confocal laser scanning microscope, LSM 510 (Carl Zeiss).

Total RNA from control and stimulated cells was extracted using RNAzol B (Ambion), according to the manufacturer’s protocol. Poly(A)⁺ RNA was purified using Oligotex mRNA kits (Qiagen), as recommended by the manufacturer. RNA samples were electrophoresed in formaldehyde-containing, denaturing agarose gels and blotted onto Hybond XL nitrocellulose (Amersham Biosciences). A quantity amounting to 25 μg of total RNA was loaded onto the gel. Probes were prepared using 25 ng of PCR product as template and [α³²P]dCTP (Amersham Biosciences) as label. Excess nucleotides were removed from the probes by purification using Sephadex G-50 nick columns (Amersham Biosciences). Blots were prehybridized for 4 h at 50°C in high SDS Church buffer (7% SDS, 50% formamide, 25% 20× SSC, 5% 1 M sodium phosphate (pH 7), 10% 50× Denhardt’s solution, and 1% 10× N-lauroyl sarcosine (Sigma-Aldrich)) containing 100 μg/ml salmon sperm DNA for blocking. Hybridization was performed for 16 h at 50°C, and detection was performed by x-ray film exposure.

THP-1 cells were washed in sterile 0.01% PBS, and then serum starved for 24 h. The cells were washed and supplemented with 2 mg/ml BSA or 2 mg/ml BSA and 100 ng/ml CCL5 or CCL3 (PeproTech). Cells were incubated at 37°C for 24 h or in subsequent experiments for 0, 2, 4, 16, or 24 h, respectively. Following incubation, cells were resuspended in lysis buffer (dH₂O, 150 mM NaCl, 1% Nonidet P-40, 50 mM Tris, 0.2 mM NaVO₄, 1 mM DTT, and 25 μg/ml each of aprotinin, leupeptin, and pepstatin). Lysates were incubated at 4°C with agitation for 30 min, followed by centrifugation at 12,000 × g for 2 min. An equal volume of glycerol (Sigma-Aldrich) was added to each sample. Protein estimation was performed on samples using the bicinchoninic acid protein assay kit (Pierce).

An optimum concentration of protein (20 μg) was loaded onto 10% SDS-PAGE and transferred to a Hybond P nitrocellulose membrane (Amersham Biosciences). Western blot analysis was performed using 5% nonfat milk in TBS with 0.5% Tween 20 for blocking. The membrane was incubated with primary Ab Mab72 specific for ALF (University of Iowa) at 4°C overnight at a concentration of 1:50. This was followed by incubation with goat anti-rat HRP-conjugated Ab (Santa Cruz Biotechnology) for 1 h at room temperature with shaking. Reactions were developed using the Pierce ECL chemiluminescence substrate kit.

To quantify the membrane-bound fraction of ALF, flow cytometry was performed using polyclonal rabbit anti-ALF, provided by M. Páez de la Cadena (Universidad de Vigo, Vigo, Spain) (1). THP-1 cells were labeled with an optimal concentration of primary Ab at 4°C for 30 min, washed, and counterstained with a FITC-conjugated goat anti-rabbit Ig (Sigma-Aldrich) for an additional 20 min. CD31 was labeled using a murine mAb (clone JC/70a; Abcam) for 30 min before washing and counterstaining with an appropriate fluorochrome-conjugated goat anti-mouse IgG reagent (Sigma-Aldrich). Ab specificity was controlled with preimmune rabbit serum or an isotype-matched, irrelevant mAb (DakoCytomation). In both cases, the stained cells were analyzed by flow cytometry (FACSort; BD Biosciences); data analysis was performed using Lysis II software (BD Biosciences).

To determine whether ALF has the potential to modify the extravasation process, a series of chemotaxis experiments was performed. In these experiments, a monolayer of the endothelial fusion cell line EAhy926 was grown on the upper surface of the membrane, which contained 3-μm pores. Fig. 1 shows a visual representation of cell morphology following treatment with ALF. Fig. 1,a shows the normal chemotactic response toward 100 ng/ml CCL5 in the lower chamber; treatment of the monocytic cells with 0.08 U/ml ALF markedly decreased migration (Fig. 1 b), but no decrease in cell viability was observed. These findings suggest that the cell morphology following treatment with ALF remains generally unchanged, whereas migration is reduced to basal levels.

Dose-response assays using 0–0.1 U/ml ALF revealed that the prior incubation of THPs with 0.08 U/ml produced a maximal inhibition of the transmonolayer migration of monocytic THP-1 cells (Fig. 1 c). The addition of 100 ng/ml CCL5 and 0.08 U/ml ALF decreased migration to background levels.

To examine whether both the THP-1 migratory cells and the endothelial monolayers were sensitive to treatment with the acid hydrolase, a parallel experiment was conducted in which either the model endothelium or the migratory monocytic cells were treated individually or together with 0.08 U/ml ALF. As with previous experiments, an optimal 100 ng/ml CCL5 was added to the lower chamber during the chemotaxis assay. The results demonstrated that compared with the controls, migration was markedly inhibited in all cases in which ALF was added (Fig. 1 d). Furthermore, compared with the basal migration levels, pretreatment of the monocytic cells with ALF produced a greater decrease in cell migration (p = 0.0004) than pretreatment of the endothelial monolayer (p = 0.0374). Furthermore, combined pretreatment of monocytes and endothelial cells seems to have an additive effect on transmigration. The decrease in migration compared with individual treatment of monocytes and endothelial cells was significant (p = 0.0255 and p = 0.0195, respectively). Controls for basal migration in ALF-pretreated monocytes and endothelial cells in absence of CCL5 were not significantly different from untreated cells (data not shown).

To control for the possibility that the migratory process was affected by agents present in the ALF suspension buffer, further chemotaxis experiments were performed. Fig. 1 e demonstrates that compared with samples containing CCL5 alone, the additional presence of denatured ALF did not significantly decrease cell migration (p > 0.4).

Labeling experiments were performed to examine the effect of fucosidase treatment on binding of fucose-specific lectin (T. purpureas) to THP-1 cells. Suspension of THP-1 cells was treated with fucosidase (0.08 U/ml) for 1, 2, 8, and 16 h, respectively, followed by binding with lectin. Untreated cells strongly labeled with the lectin (Fig. 2). In contrast, ALF treatment for 8 and 16 h demonstrated a significant reduction in labeling of THP-1. However, incubation for 1 and 2 h with 0.08 U/ml ALF did not show significant reduction in binding of lectin (data not shown). The enzyme-treated cells were 95 and 85% viable following 8- and 16-h incubation based on trypan blue exclusion assay.

To determine whether the decrease in transmigration following ALF treatment is due to the effect of enzyme on selectin ligands, we conducted adhesion experiments for ALF-treated THP-1 cells to P-selectin.Fc. For these experiments, THP-1 cells were treated with 0.08 U/ml ALF, followed by labeling with 2′,7′-bis(2-carboxyethyl)-5-(6)-carboxyfluorescein tetrakis. As a control, a parallel experiment was performed in which cells were untreated (Fig. 3). ALF treatment significantly reduced THP-1 cell adhesion to P-selectin.Fc (p = 0.0004).

To examine the functional consequences of ALF treatment under shear flow, an in vivo study was conducted using a single-cell splenocyte suspension from mice immunized (11 days postimmunization) with interphotoreceptor retinoid-binding protein peptide to induce EAU. Cells were divided into two groups, and one group was treated with ALF. To ensure the cleavage of fucose residues present on the surface of splenocytes, the cells were stained with a biotinylated fucose-specific lectin (T. purpureas agglutinin) following treatment with ALF to validate removal from the cell surface (data not shown).

Results from SLO (Fig. 4,a) revealed that the treatment of splenocytes with ALF had a highly significant effect on both the rolling and retinal infiltration of cells. Following treatment with ALF (0.08 U/5 × 10⁵ cells), an 8.6-fold decrease (p < 0.0001) in rolling compared with untreated cells was observed (Fig. 4,b). Furthermore, the ability of the splenocytes to migrate through retinal vasculature to penetrate the retinal tissue was also inhibited following treatment of migratory cells with the fucosidase, resulting in a 3.5-fold decrease in infiltration (Fig. 4 c; p = 0.0119).

To determine whether ALF plays a natural role in regulation of inflammation, the human monocytic cell line THP-1 was stimulated for various time periods with a prototypical inflammatory chemokine CCL5 at 100 ng/ml.

Northern blot analysis was performed to examine regulation of the gene encoding ALF. Fig. 5 shows the densitometric analysis of the Northern blots and demonstrates that there was a 2.2-fold increase in the expression of ALF following 24-h stimulation with CCL5 and a 1.8-fold increase in expression following 24-h stimulation with CCL3. The Northern membranes were also subsequently probed for GAPDH, to ensure equal loading (data not shown).

To further examine the regulation of ALF at the protein level, Western blot analysis was performed; this revealed two bands at 56 and 51 kDa, which showed increased expression after 4 h (data not shown) and 24 h (Fig. 6). Densitometric analysis (normalized to an α-tubulin loading control) was performed, which yielded values for OD; this showed that after 24-h stimulation with CCL5, the 51- and 56-kDa bands were up-regulated 2.4- and 3.2-fold, respectively.

ALF is usually found as a soluble component of the lysosomes and antiserum. However, it has also been reported that it can exist associated with plasma membrane (10–20% of the total cellular fucosidase activity) in a range of cell types, including monocytes (1). Therefore, to determine whether stimulation with chemokines altered the distribution of membrane-bound and cytoplasmic ALF, a flow cytometric analysis was performed. Although ALF was detected on the cell surface, no significant change in cell surface expression was observed following stimulation with CCL5 for 30 min, 4 h, or 24 h (data not shown).

The process of transendothelial migration involves a series of adhesion molecules together with their respective receptors; ALF is known to degrade fucosylated moieties. Because the fucosylated molecule CD31 (PECAM-1) is central to the process of leukocyte migration, immunofluorescence flow cytometry was used to measure the surface expression of CD31 on monocytic THP-1 and EaHY.926 cells up to 24 h after the addition of 0.08 U/ml ALF. CD31 expression was examined at each time point using flow cytometric analysis.

EAhy926 cells demonstrated no significant difference in CD31 expression following ALF treatment for 2 h; however, at 5 h, there was a 7% increase in cell surface expression (p = 0.0001). The expression of CD31 declined dramatically between 5 and 16 h, and gradually increased by 24 h such that there was no significant difference between 0 and 24 h. Following treatment with ALF, THP-1 cells demonstrated decreased expression of CD31 between 0 and 5 h (p = 0.002), but Ab binding then gradually increased (Fig. 7). The level of CD31 expression was significantly lower at 24 h compared with 0 h with p value of 0.006.

ALF is a ubiquitous acid hydrolase that is normally found as a soluble component of lysosomes and blood plasma, but up to 20% of ALF activity is associated with the cell surface (1). This enzyme can process fucosylated residues that normally play a role in stabilizing intercellular adhesion, potentially regulating cell trafficking in health and disease. Furthermore, ALF from some sources has also been shown to have unique property to perform the synthesis of complex oligosaccharides by transglycosylation (26). This might provide a valuable approach for synthesis of fucose containing oligosaccharides because α-glycosynthases are difficult to obtain. The current study was designed to assess the possibility that ALF can modulate inflammation by reducing the interaction between fucoslyated adhesion molecules that normally support leukocyte extravasation.

The ability of the monocytic cell line, THP-1, to migrate through an endothelial cell monolayer in response to the chemokine CCL5 was tested in a model system. Dose-response studies showed that the presence of 0.08 U/ml ALF inhibited leukocyte migration to a background level. A comparison of treatment of either the endothelial cells or the monocytes showed that both cell types express migration-critical residues that were inhibited by ALF. However, the slightly greater inhibition produced by treatment of the monocytes suggests that fucosylated ligands may play a dominant role on these cells. Furthermore, combined pretreatment of endothelial and THP-1 cells had an additive effect on inhibition of transmigration.

Using flow cytometry, we identified a significant decrease in the binding of fucose-specific lectin following ALF treatment. Interestingly, treatment of both the monocytic and endothelial cells for 1 h with 0.8 U/ml ALF was sufficient to reduce cell migration to a basal level. However, a significant reduction in the binding of the lectin was only evident after treatment for 8 and 16 h. It is difficult to compare the results with previous observations due to variability in the source of enzyme and lectins/Abs used (27).

PSGL-1, originally identified as a ligand for P-selectin, can bind all three selectins and has a proadhesive function in many inflammatory settings (28). L-selectin is constitutively expressed by most leukocytes, whereas the other members of selectin family, P- and E-selectin, are expressed by activated endothelium. Acting in cooperation with L-selectin, P-selectin mediates the initial interaction between circulating leukocytes and endothelial cells that produces characteristic rolling of leukocytes on endothelium. To assess the contribution of fucose residues on selectin binding, solid-phase adhesion experiments were conducted. The current study was consistent with previous observations that fucosylation of appropriate carbohydrate determinants is critical for selectin ligand generation (13). Furthermore, treatment of THP-1 cells with ALF resulted in 53% decrease in the adhesion of THP-1 cells to P-selectin.Fc.

To further examine the role of ALF in vivo, we used an EAU model, induced by immunization of mice with retinal Ags (20). The blood-retina barrier is breached in EAU, allowing both T lymphocytes and monocytes to move into the retina (29). The pretreatment of splenocytes from sensitized animals with ALF resulted in a profound decrease in the ability of these cells to roll along the surface of the retinal endothelium, and also reduced the number of cells that undergo diapedesis and penetrate the retinal tissues.

Homophilic interactions between CD31 (PECAM-1) on the surface of inflammatory leukocytes and CD31 within the junctions between endothelial cells allow the passage of leukocytes across confluent endothelium (30), with Ab blockade of CD31 reducing transendothelial migration by up to 90% (31). To identify a further potential mechanism by which ALF can mediate its effect, a series of studies was performed to measure the potential of this enzyme to degrade CD31. Investigation of the nature of the N-linked glycans in Chinese hamster ovary cell-expressed CD31 has already revealed the presence of sialylated bi-, tri-, and tetra-antennary compounds, of which 62% are fucosylated at the core GlcNAc residue (32). The current study demonstrated CD31 expression by both endothelial and monocytic cell lines. ALF does have an effect on CD31 expression on EAhy926 endothelial cells at several of the time points that were observed; however, after 24-h exposure, the CD31 levels return to the basal level. THP-1 cells treated under the same conditions show a decline in CD31 expression between 0 and 5 h. The difference in the level of CD31 expression between the two cell types might be attributed to the location of CD31 molecule. It is possible that the cleavage of fucose residues, contained within the CD31 molecules, within the junction of an endothelial monolayer, would be more difficult to access than those that are expressed on cell surface of migratory cells. Significantly, the potential of a CD31 mAb to label these cells was reduced by pretreatment of the cells with ALF, suggesting degradation of a fucose-containing epitope that may be critical for transendothelial migration.

The possibility of regulation of ALF expression at both the mRNA and protein levels was investigated. Interestingly, following treatment with CCL5 (100 ng/ml), the gene product that coded for ALF was up-regulated by more than 2-fold. A similar increase in expression (1.8-fold) was produced by treatment with 100 ng/ml CCL3. To examine the protein expression of ALF, THP-1 cells were stimulated with 100 ng/ml CCL5 for 4 and 24 h; ALF bands were identified at 51 and 56 kDa by Western blotting, which is consistent with previous reports (1, 3, 5). Treatment with CCL5 increased protein expression after 4 h (data not shown), with stimulation for 24 h increasing expression of both products by between 2- and 3-fold.

In conclusion, this study shows that the acid hydrolase, ALF, is potentially up-regulated by chemokines in the later stages of inflammation. The function of this enzyme includes cleavage of fucosylated residues from adhesion molecules such as sLe^X and CD31, which play an important role in intravascular leukocyte rolling and subsequent extravasation. The potential for inflammatory chemokines to up-regulate the expression of ALF is consistent with activation of a natural regulatory loop, resulting in a gradual diminution of the potential of blood-borne leukocytes to penetrate the endothelium at sites of existing inflammation.

The authors have no financial conflict of interest.