The Role of Integrin α_Dβ₂ (CD11d/CD18) in Monocyte/Macrophage Migration

Abstract

Integrin α_Dβ₂ (CD11d/CD18) is a multiligand macrophage receptor with recognition specificity identical to that of the major myeloid cell-specific integrin α_Mβ₂ (CD11b/CD18, Mac-1). Despite its prominent upregulation on inflammatory macrophages the role of α_Dβ₂ in monocyte and macrophage migration is unknown. In this study, we have generated model and natural cell lines expressing different densities of α_Dβ₂ and examined their migration to various extracellular matrix proteins. When expressed at a low density, α_Dβ₂ on the surface of recombinant HEK293 cells and murine IC-21 macrophages cooperates with β₁/β₃ integrins to support cell migration. However, its increased expression on the α_Dβ₂-expressing HEK293 cells and its upregulation by PMA on the IC-21 macrophages results in increased cell adhesiveness and inhibition of cell migration. Furthermore, ligation of α_Dβ₂ with anti-α_D blocking antibodies restores β₁/β₃-driven cell migration by removing the excess α_Dβ₂-mediated adhesive bonds. Consistent with in vitro data, increased numbers of inflammatory macrophages were recovered from the inflamed peritoneum of mice after the administration of anti-α_D antibody. These results demonstrate that the density of α_Dβ₂ is critically involved in modulating macrophage adhesiveness and their migration, and suggest that low levels of α_Dβ₂ contribute to monocyte migration while α_Dβ₂ upregulation on differentiated macrophages may facilitate their retention at sites of inflammation.

Introduction

The β₂ integrins are heterodimeric adhesion receptors expressed on leukocytes which play an important role in the immune-inflammatory response. The group consists of four members, α_Lβ₂ (CD11a/CD18, LFA-1), α_Mβ₂ (CD11b/CD18, Mac-1), α_xβ₂ (CD11c/CD18, p150, 95) and α_Dβ₂ (CD11d/CD18), which have a common β₂ subunit and different, albeit homologous, α subunits. The two related members of the group, α_Mβ₂ and α_Dβ₂, are expressed predominantly on myeloid leukocytes and are strongly upregulated after agonist stimulation on neutrophils and monocytes, respectively. A large number of studies have been performed both in vitro and in vivo to understand the role of the major myeloid cell-specific integrin α_Mβ₂ in neutrophil migration [1]. In the current model of leukocyte trafficking during the inflammatory reaction, α_Mβ₂ does not appear to be important for adhesion to the endothelium but cooperates with α_Lβ₂ in leukocyte emigration from the vessel [2,3]. The same integrins, together with α₄β₁, have also been implicated in migration of monocytes although α_Mβ₂ and α_Lβ₂ appear to be less important than α₄β₁ [4,5]. While α_Mβ₂ is required for transendothelial extravasation, it does not support neutrophil migration through the interstitial extracellular matrix (ECM). Instead, neutrophils appear to utilize β₁ integrins for migration through tissues [6,2,7,8]. Nevertheless, α_Mβ₂ may have a unique role in neutrophils migration. As studies in different models of inflammation have shown, neutrophils from the α_Mβ₂-deficient mice had enhanced migratory properties since the leukocyte influx was increased ~2–3-fold in the α_Mβ₂-deficient mice compared with wild-type counterparts [9,10,3,11]. Furthermore, using model cells expressing different levels of α_Mβ₂, we have demonstrated that the progressive increase of α_Mβ₂ density inhibited cell migration mediated by β₁ integrins [12]. These observations suggest that the major function of α_Mβ₂ is not to support migration but rather to serve as a brake during neutrophil migration through the interstitial space. The basis for this function of α_Mβ₂ is thought to be its unusual stickiness and its ability to bind the same ECM proteins as the β₁ integrins. For example, both α_Mβ₂ and α₅β₁ can adhere to fibronectin [12]. Another important characteristic of α_Mβ₂ is that this receptor can be upregulated ~7–10-fold on the surface of neutrophils in a stepwise manner[1] while the levels of β₁ integrins increase modestly [7]. Accordingly, when the density of α_Mβ₂ exceeds that required for optimal β₁-driven cell locomotion, migration halts. These observations would be consistent with the idea that the major function of α_Mβ₂ is not migration but neutrophil adhesion and the control of adhesion-dependent leukocyte responses such as degranulation, oxidative burst and phagocytosis at sites of inflammation [10].

We have recently demonstrated that another member of the subfamily, the most recently discovered integrin α_Dβ₂, exhibits multiligand binding properties and has recognition specificity overlapping that of α_Mβ₂ [13]. Specifically, α_Dβ₂ is capable of supporting cell adhesion to various ECM proteins, including fibronectin, vitronectin, fibrinogen, CCN1 (Cyr61) and others. We have also shown that in α_Dβ₂, the α_D I-domain is responsible for the binding function and that the mechanism whereby α_D I-domain recognizes its ligands is similar to that utilized by α_Mβ₂ [13]. The finding that α_Dβ₂ and α_Mβ₂ have similar recognition specificity and bind a broad assortment of proteins in the ECM suggests that α_Dβ₂ might perform an analogous function in leukocyte migration. Like α_Mβ₂, integrin α_Dβ₂ is poorly expressed on peripheral blood leukocytes [14] and can be rapidly upregulated in response to chemotactic stimulation as a result of its transport to the cell surface from internal secretory granules [14]. However, its expression in neutrophils is much lower than that of α_Mβ₂ [14]. In contrast, α_Dβ₂ levels are gradually increased by de novo synthesis upon monocyte differentiation into macrophages, and oxidized LDL and AcLDL further upregulate its expression [15]. Furthermore, α_Dβ₂ (but not α_Mβ₂) is strongly upregulated on macrophages within atherosclerotic plaques [14]. Such a difference in upregulation of α_Mβ₂ and α_Dβ₂ on neutrophils and monocytes/macrophages, respectively, might indicate their distinct roles in acute and chronic inflammatory responses. The unique pattern of α_Dβ₂ up-regulation by atherogenic lipoproteins suggests a potential role for this integrin in the regulation of monocyte/macrophage migration to extravascular sites and, thus, in the development of atheroscerosis. However, exactly how α_Dβ₂ contributes to monocyte/macrophage migration is still unknown.

In the present study we have examined the migratory properties of α_Dβ₂. We have generated model and natural monocytic cell lines expressing different levels of α_Dβ₂ and tested their migration to various ECM proteins. In addition, a mouse model of inflammation has been used to further clarify the role of α_Dβ₂ in monocyte/macrophage migration in vivo.

Materials and Methods

Proteins and antibodies

Human fibrinogen was obtained from Enzyme Research Laboratories (South Bend, IN). Fibronectin was isolated from fresh human plasma by gelatin-agarose affinity chromatography [16]. Vitronectin was isolated from outdated plasma treated with 8M urea using heparin-agarose affinity chromatography [17]. CCN1 (connective tissue growth factor, Cyr61) was purified as described [18]. The mAb IB4 directed against the β₂ integrin subunit and mAb M1/70 which recognizes the mouse α_M integrin subunit were purified from the conditioned media of the hybridoma cell line obtained from American Type Culture Collection (Manassas, VA) using protein A agarose (GE Healthcare, Piscataway, NJ). F(ab’)₂ preparations of Mab M1/70 were isolated after digestion with pepsin (1:200 w/w). F(ab’)₂ retained its full immunoreactivity as assessed by FACS analyses and did not contain intact mAb molecules or its Fc fragments as assessed with goat anti-rat IgG Fc alkaline phosphatase-conjugated antibody (Pierce). Polyclonal antibodies 1950 and 1932 directed against the β₁ and β₃ integrin subunits, respectively, were purchased from Chemicon Int. (Temecula, CA). MAbs against murine CD14 (PE conjugated), F4/80 (PE conjugated) and α_X integrin (FITC conjugated) were from eBioscience (San Diego, CA). Monoclonal antibody against the mouse α_V integrin (clone RMV-7) was purchased from Fitzgerald Industries International Inc. (Concord, MA). Polyclonal antibody against the α_DI-domain was described previously [13]. The antibody recognizes both human and mouse α_DI-domains and has no cross-reactivity with recombinant human and mouse α_M, α_X and α_L I-domains. The antibody was isolated from rabbit serum by affinity chromatography using α_DI-domain-Sepharose. Fab preparations were isolated after enzymatic digestion with papain (Sigma-Aldrich, St. Louis, MO) followed by purification by protein A agarose.

Cells

The IC-21 macrophage cell line was obtained from the ATCC. Cells were cultured in RPMI-1640 medium (Invitrogen, Carlsbad, CA) supplemented with 10% FBS, 0.1 mg/ml streptomycin, 0.1 unit/ml penicillin, and 2 mM L-glutamine. The α_Dβ₂-expressing HEK 293 cells were generated as described previously [13] and maintained in DMEM/F-12 (Invitrogen) supplemented with 10% FBS, 2 mM glutamine, 15 mM HEPES, 0.1 mg/ml streptomycin, and 0.1 unit/ml penicillin. Populations of cells expressing low and high density of α_Dβ₂ integrin were selected by fluorescence activated cell sorting using a BD Biosciences FACS Vantage Instrument.

Flow cytometry

FACS analyses were performed to assess the expression of integrins on the surface of the α_Dβ₂ transfected HEK 293 cells and the mouse IC-21 macrophage cell line. The cells were incubated with integrin-specific antibodies and analyzed using a FACScan™ (Beckton Dickinson) as described previously [19]. To determine integrin expression on murine peripheral blood monocytes, a population of cells isolated from whole blood by Ficoll fractionation which contained both monocytes and lymphocytes was incubated with a mixture of phycoerythrin-conjugated anti-CD14 (5 µg/ml) and selected FITC-conjugated anti-integrin antibodies (10 µg/ml). Likewise, cells isolated from the mouse peritoneal cavity were incubated with a mixture of the phycoerythrin-conjugated macrophage-specific marker F4/80 (5 µg/ml) and FITC-conjugated anti-integrin antibodies. The levels of integrins were determined using back-gaiting with anti-CD14 and F4/80 mAbs for peripheral blood monocytes and peritoneal macrophages, respectively.

Adhesion and migration assays

For adhesion assays, 96-well tissue culture plates (Costar, Cambridge, MA) were coated with different concentrations of vitronectin for 3 h at 37°C. The wells were post-coated with 0.5% PVA for 1 h at 22°C. The α_Dβ₂-expressing cells were labeled with 10 µM Calcein AM (Molecular Probes, Eugene, OR) and assays were performed as described previously. [12]

Cell migration assays with calcein-labeled cells were performed under sterile conditions using 24-well transwell plates (Costar, Corning, NY) with 8-µM pore size uncoated polycarbonate filters as previously described [20,12]. Lower wells contained 600 µl of vitronectin or other ECM proteins in DMEM/F-12 which during the assay coat both sides of the filter. Upper wells, inserted into the lower wells, contained a final volume of 250 µl after the addition of cells. To begin the assay, 150 µl of cells (2.5 × 10⁶/ml) in medium, with or without mAbs, was added to upper wells and cells were allowed to migrate for 18 h at 37°C in a 5% CO₂ humidified atmosphere. In this format, migration is largely directional and cells migrate from the upper chamber to the underside of the filter coated with a ligand [20]. Two hours prior to the completion of the migration assay, cells were labeled with calcein AM which was added to the lower chamber at 10 µM final concentration. Assays were stopped by removing cells from the upper surface of the polycarbonate membrane by wiping the surface twice with a cotton-tipped applicator. Cells migrating to the underside of the filter were detected using a CytoFluorII fluorescence plate reader (Applied Biosystems, Farmington, MA).

Peritoneal Inflammation

Peritonitis was induced by intraperitoneal injection of 0.5 ml of 3% Brewer thioglycollate medium (Sigma-Aldrich, St. Louis, MO) in C57BL/6 mice (Jackson Laboratories, Bar Harbor, ME). After 72 hours the mice were euthanized by isoflurane inhalation and the peritoneal cavities were lavaged with 5 ml PBS. In experiments with antibodies, 100 µg of purified anti-α_D polyclonal antibody, Mab M1/70 or normal mouse IgG diluted in PBS were injected intraperitoneally 4 hours before thioglycollate injection. In selected experiments, mice were injected with F(ab’)₂ preparations of Mab M1/70 (100 µg). Macrophages were counted by nonspecific esterase assay as described [21]. In selected experiments, the number of cells in total lavage fluid was counted in a hemacytometer and macrophage number was determined by performing differential counts after cytospin and Wright stain. In each experiment, 7 mice were used for a control (non-treated) and 7 mice were used for the treated group.

Statistical analysis

Results are presented as the means ± SD. Statistical analyses were performed using Student’s t test. The value of P<0.05 was considered significant.

Results

Generation of the α_Dβ₂ expressing cells with different densities of receptors and characterization of their adhesive and migratory properties

To examine the role of α_Dβ₂ in migration, we initially examined HEK 293 cells stably transfected with α_D and β₂ integrin subunits [13]. To generate cells with different levels of α_Dβ₂, the initial heterogeneous population of the α_Dβ₂-expressing cells was sorted into two fractions. Surface expression of α_Dβ₂ was determined by flow cytometry using mAbs directed against the α_D and β₂ subunits (Fig.1). The mean fluorescence intensities of the two cell lines differed by ~10-fold, indicating that α_Dβ₂ was expressed at different densities. Moderate levels of endogenous integrins α₅β₁, α₂β₁, α₃β₁, α₄β₁, α₆β₁ and α_Vβ₁ were detected on the surface of transfected cells in agreement with previous data [12] (not shown). Because adhesion is necessary for cells to obtain traction upon migration, we examined the adhesive capacity of the generated cell lines. In these experiments, vitronectin was selected as a representative ECM ligand and the maximal levels of adhesion were determined. Consistent with the α_Dβ₂ density differential, adhesion of cells expressing low levels of integrin was ~ 2-fold lower than that of high expressors (Fig.2). In previous studies we demonstrated that in the α_Dβ₂-expressing cells both α_Dβ₂ and β₁ integrins (apparently α_Vβ₁, a vitronectin receptor in HEK 293 cells) contribute to adhesion to vitronectin [13]. In agreement with these observations, anti-α_D function blocking antibody partially inhibited cell adhesion of both cell lines and this effect correlated with the level of α_Dβ₂ expression. Accordingly, anti-α_D inhibited adhesion of cells expressing high and low densities of α_Dβ₂ by ~80% and 50%, respectively (Fig.2). Furthermore, mock-transfected HEK 293 cells adhered to vitronectin; however, the adhesion was low because these cells rely only on endogenous α_Vβ₁ for binding to this ligand.

Fig.1. Characterization of HEK 293 cells expressing different levels of α_Dβ₂.

Binding of anti-α_D polyclonal antibody (10 µg/ml) and anti-β₂ mAb IB4 (10 µg/ml) to mock-transfected and the α_Dβ₂-expressing cells was analyzed by flow cytometry. Results are presented as histograms with the logarithm of fluorescence intensity on the abscissa and the cell number on the ordinate. The level of integrin expression is shown as filled histograms. Control cells incubated with the Alexa Fluor-488 conjugated secondary antibody are shown as open histograms.

Fig.2. Adhesion of HEK 293 cells expressing different densities of α_Dβ₂ and mock-transfected HEK 293 cells to vitronectin.

Aliquots of calcein-labeled cells (5×10⁴) were allowed to adhere to wells of 96-well plates coated with 5 µg/ml vitronectin in the absence or presence of blocking anti-α_Dβ₂ antibody (10 µg/ml). After incubation for 30 min at 37°C in 5% CO₂, the nonadherent cells were removed by two washes with PBS and fluorescence was measured. The number of adherent cells was calculated by using the fluorescence of aliquots with a known number of labeled cells. Data are expressed as a percentage of added cells and are the mean ± SD of three individual experiments (* denotes P<0.05; ** denotes P<0.01).

To investigate the relationship between the α_Dβ₂ density and cell migration, we compared the ability of cells expressing different levels of integrin to migrate toward vitronectin in a Transwell system. In these experiments, vitronectin (50 µg/ml) was placed in the lower chamber and migration was initiated by adding cells to the upper chamber. Migration was allowed to proceed for 18 h at 37°C, at which time the number of cells adherent to the underside of the filter coated with vitronectin was determined. Migration of the two cell lines to vitronectin was very different (Fig.3A). Both mock-transfected cells and the cell line expressing a low density of α_Dβ₂ migrated efficiently, with migration of the latter cells being 1.7-fold higher. However, migration of cells expressing a high density of α_Dβ₂ was suppressed (by ~3-fold) compared to mock-transfected cells. To examine the contribution of α_Dβ₂ and α_vβ₁ to migration, cells were preincubated with 20 µg/ml of either anti-α_D or anti-β₁ function-blocking antibodies before their addition to the upper chamber of the transwell system. Migration of cells with a low density of α_Dβ₂ was inhibited by ~50% and 85%, respectively (Fig.3B). Thus, both α_Dβ₂ and α_vβ₁ support migration of these cells. However, preincubation of cells expressing a high-density of α_Dβ₂ with anti-α_D produced the opposite effect (Fig.3C), i.e., the antibody did not inhibit migration but instead enhanced migration by ~4-fold. At the same time, anti-β₁ still blocked migration. Hence, ligation of α_Dβ₂ with anti-α_D in these cells could restore migration to the level observed with mock-transfected cells. Since the level of β₁ integrins in the two cell lines expressing different densities of α_Dβ₂ is the same, the difference in their migratory ability appears to arise from the dissimilarity in the density of α_Dβ₂. Thus, these results suggest that in both cell lines migration is mediated by α_vβ₁, and that α_Dβ₂ expressed at a low density contributes to migration. However, when the density of α_Dβ₂ increases it suppresses migration.

Fig.3. Migration of the α_Dβ₂-expresing HEK 293 cells to vitronectin.

A, HEK 293 cells expressing different densities of α_Dβ₂ were analyzed for their ability to migrate through the Transwell inserts having a pore size of 8 µM and 6.5 mm in diameter. Cells were allowed to migrate toward 50 µg/ml of vitronectin placed in the lower chamber of the Transwell system for 18 hours at 37°C in 5% CO₂. Two hours prior to the completion of the migration assay, cells were labeled with calcein AM and fluorescence of cells that migrated to the undersurface of the filter was determined. The cells expressing low (B) and high (C) densities of α_Dβ₂ were analyzed for their ability to migrate to 50 µg/ml of vitronectin in the presence of function blocking polyclonal anti-α_D antibody or blocking anti-β₁ mAb. Cells were preincubated with 20 µg/ml of each antibody for 20 min before their addition to the upper chamber of Transwell plates. Migration data are expressed as a percentage of the migration of mock-transfected cells (A) or control α_Dβ₂ transfected cells in the absence of antibodies (B, C). Results are the mean ± SD from three to five separate experiments, each with duplicate measurements. (*, P<0.05; **, P<0.01 compared to mock transfected cells in (A) or control (B,C)).

To confirm that the rate of migration in this system depends on the density of α_Dβ₂ but not on the nature of the adhesive ligand, we examined cell migration to different ECM proteins. Previous studies have demonstrated that fibronectin, fibrinogen and CCN1 are ligands for α_Dβ₂ [13]. Furthermore, fibronectin and fibrinogen can interact with the β₁ integrin α₅β₁ [22,23], and CCN1 can engage α₆β₁ [24]. As shown in Fig.4, pretreatment of cells expressing a high density of α_Dβ₂ with anti-α_D antibody significantly enhanced migration to all proteins while the anti-β₁ specific antibody inhibited migration. These results indicate that as with vitronectin, the increased density of α_Dβ₂ results in the decline of β₁-driven migration to other ligands and that the retarded migration could be restored by the removal of the restraint imposed by the excess of the α_Dβ₂ - ligand bonds. However, the extent to which ligation of α_Dβ₂ by function-blocking anti-α_D increased migration was different, which may reflect the relative contribution of α_Dβ₂ and β₁ integrins to adhesion. For instance, cell adhesion to CCN1 is mediated predominantly by α_Dβ₂ with a small contribution of β₁ integrins [13], and, accordingly, anti-α_D exerted a strong stimulating effect on cell migration. In contrast, anti-α_D caused a moderate effect on cell migration to fibronectin because both α₅β₁ and α_Dβ₂ mediate cell adhesion to this ligand [13]. Nevertheless, regardless of these differences the stimulating effect of anti-α_D was detected with all ligands suggesting that the increased density of α_Dβ₂ inhibits cell migration to various ECM proteins.

Fig.4. Migration of HEK 293 cells expressing a high density of α_Dβ₂ to various ECM proteins.

HEK 293 cells expressing a high density of α_Dβ₂ were analyzed for their ability to migrate to 50 µg/ml fibrinogen, fibronectin and CCN1 in the absence or in the presence of blocking polyclonal anti-α_D antibody or blocking anti-β₁ mAb. Cells were preincubated with 20 µg/ml antibodies for 20 min before their addition to the upper chamber of Transwell plates. Cells were allowed to migrate toward ligands for 18 hours at 37°C in 5% CO₂ and the extent of cell migration was assessed as described in Materials and Methods. Data are expressed as a percentage of migration; cell migration in the absence of blocking antibodies was assigned a value of 100%. Values represent the means ± SD with duplicate wells for each experiment from three or more individual experiments (*, P<0.05; **, P<0.01 compared to control).

Migratory properties of α_Dβ₂ expressed on the macrophage cell line IC-21

Because macrophage-derived foam cells that are present in atherosclerotic lesions express high levels of α_Dβ₂ [14], we examined the role of α_Dβ₂ in monocyte/macrophage migration. In previous studies, we have demonstrated that the mouse macrophage cell line IC-21 expresses similar levels of α_Mβ₂ and α_Dβ₂ [13]. After stimulation of these cells with 100 ng/ml PMA for 48 h, the level of α_Dβ₂ was upregulated by more than 5-fold while expression of α_Mβ₂ did not change (Fig.5). After 72 h, expression of α_Dβ₂ remained high and gradually returned to the basal levels after 7 days in culture.

Fig.5. Analyses of integrin α_Dβ₂ and α_Mβ₂ expression on mouse IC-21 macrophages after PMA stimulation.

IC-21 cells were stimulated with 100ng/ml PMA for different periods of time and the levels of α_Mβ₂ (open bars) and α_Dβ₂ (black bars) integrins were assessed by flow cytometry using anti-α_M mAb M1/70 and polyclonal anti-α_D antibody. Results are the mean ± SD values of mean fluorescence from three independent experiments (*, P<0.05; **, P<0.01 compared to α_D expression before PMA treatment.)

Since resting and PMA-stimulated IC-21 cells can model the two cell lines with low and high densities of α_Dβ₂, respectively, these cells were tested for their ability to support migration to vitronectin. At least three other integrins on the surface of IC-21 cells, including α_Vβ₁, α_Vβ₃, and α_Mβ₂, can potentially mediate migration to vitronectin. Expression of α_M, α_V, β₁ and β₃ integrins did not change significantly after 72 hours of stimulation with PMA (not shown) suggesting that only up-regulation of α_Dβ₂ can modulate cell adhesiveness. Both cell lines were able to migrate to vitronectin but migration of resting IC-21 was reproducibly higher than that of PMA-stimulated cells (Fig. 6A,B) and when compared in the same experiment (not shown). Preincubation of cells with the Fab preparation of the anti-α_D polyclonal antibody inhibited migration of resting macrophages indicating that α_Dβ₂ support migration (Fig.6A). No inhibition of migration by control Fab of rabbit IgG was detected. In these studies, we have utilized Fab preparations of anti-α_D because preliminary experiments demonstrated that intact rabbit IgG at the concentration 20 µg/ml inhibited migration of IC-21 cells, an effect often encountered in such assays that is attributed to the Fc portion of the antibody molecule. Pretreatment of resting macrophages with either intact anti-α_M mAb M1/70 or its F(ab’)₂ inhibited migration by ~45%, indicative of the contribution of α_Mβ₂ to cell migration. As a control, rat IgG did not inhibit cell migration. Function blocking antibodies against murine α_V, the common subunit for integrins α_Vβ₁ and α_Vβ₃, also inhibited cell migration indicating that either α_Vβ₁ or α_Vβ₃ or both can support migration (Fig. 6A).

Fig.6. Migration of IC-21 cells expressing different levels of α_Dβ₂ to vitronectin.

Unstimulated (A) and PMA-stimulated for 72 h (B) IC-21 cells were analyzed for their ability to migrate through the Transwell inserts to 50 µg/ml of vitronectin either in the absence or presence of either Fab preparations of anti-α_D or mAb M1/70. Cells were preincubated with 20 µg/ml of each antibody for 20 min before their addition to the upper chambers of Transwell plates. Cells were allowed to migrate toward vitronectin coated on the undersurface of the filter for 18 hours at 37°C in 5% CO₂. Data are expressed as a percentage of migrated cells in the absence of blocking antibodies. Results are the mean ± SD values from three to five individual experiments with duplicate wells in each experiment for each reagent tested (*, P<0.04; **, P<0.01 compared to control).

In contrast to resting cells, preincubation of PMA-stimulated macrophages with Fab preparations of anti-α_D resulted in an increase of their migration which was significant (Fig.6B). These results are consistent with the effect of anti-α_D on migration of HEK 293 cells expressing a high density of α_Dβ₂ and suggest that this integrin can be responsible for the increased adhesiveness of stimulated macrophages as well. By contrast, anti-α_M mAb M1/70 produced the opposite effect, i.e., it inhibited macrophage migration. Hence, like in resting macrophages, α_Mβ₂ contributes to migration. Collectively, these data demonstrate that upregulation of α_Dβ₂ in a cell line which naturally expresses this receptor leads to decreased cell migration. In addition, these results distinguish the opposite roles of two integrins in PMA-stimulated macrophages: while α_Mβ₂ promotes migration, α_Dβ₂ inhibits it.

Effect of anti-α_D antibody on migration of macrophages to mice peritoneum

To assess the role of α_Dβ₂ in monocyte migration in vivo, we examined the effect of anti-α_D antibody on emigration of monocytes into the peritoneal cavity of mice with thioglycollate-induced inflammation. Previous studies demonstrated that significant numbers of monocytes arrive into the peritoneum by 72 h after thioglycollate injection [25], where they apparently differentiate into macrophages. To determine whether α_Dβ₂ upregulation occurs in an environment such as that in inflamed peritoneum, we characterized expression of the integrin on monocytes at 72 h after thioglycollate injection and on monocytes in the circulation. Expression of several other macrophage integrins, including α_M, α_D, α_X, β₁ and β₃ have also been determined. Fig.7 shows that the levels of α_Dβ₂ and α_Mβ₂ were increased by ~4- and 1.5-fold, respectively, compared to basal levels on peripheral blood monocytes. No increase in the expression of other integrin subunits was detected. Resident macrophages, which constitute approximately 5% the total number of recruited macrophages, contained a low level of α_Dβ₂ and elevated levels of α_Mβ₂ (data not shown). Thus, α_Dβ₂ is the major integrin on the surface of stimulated macrophages elicited in response to inflammatory stimulation.

Fig.7. Analyses of integrin expression on mouse monocytes isolated from peripheral blood and on peritoneal macrophages.

Peripheral blood monocytes were isolated from mouse blood collected by cardiac puncture. Monocytes were obtained as a monocyte-lymphocyte phase after separation using a Ficoll-Paque gradient. Peritoneal macrophages were collected at 72 hours after injection of 0.5 ml 3% thioglycollate by washing peritoneum cavity with 5 ml of PBS. The integrin densities on monocytes (open bars) and macrophages (black bars) were determined by flow cytometry using selected FITC-conjugated mAbs against specific integrin subunits. Monocytes isolated from the peripheral blood and macrophages in the peritoneal lavage were detected using back-gaiting with phycoerythrin-conjugated mAbs against CD14 and F4/80, respectively. Results are the mean ± SD values of mean fluorescence in three or more independent experiments (**, P<0.01 compared to α_D and α_M expression on monocytes).

To analyze the role of α_Dβ₂ in accumulation of macrophages in the peritoneum, mice were injected with 100 µg of anti-α_D antibody 4 hours before injection of thioglycollate. After 72 hours, macrophages in the peritoneal exudate were counted and compared with macrophage levels in a control group injected with rabbit IgG or PBS. The amount of macrophages in the mice treated with anti-α_D antibody was increased by 44 ± 8% compared to the control group. Fig.8 shows a representative of four independent experiments. In each experiment, 7 mice were used for the control (non-treated) and 7 mice were used for the treated group. The difference between the groups in each experiment was statistically significant. The results indicate that the blockade of α_Dβ₂ with anti-α_Dβ₂ antibody resulted in the increased numbers of monocyte/macrophage recovered from the peritoneum (Fig.8). In contrast, both mAb M1/70 or its F(ab’) ₂ preparation inhibited monocyte migration slightly and this effect was not significant (data not shown). These data suggest that up-regulation of α_Dβ₂ on activated macrophages might either suppress their migration into the peritoneum or enhance their adhesion to the peritoneal wall. Furthermore, α_Dβ₂ and α_Mβ₂ may have different roles in monocyte migration and macrophage accumulation in the peritoneum.

Fig.8. Effect of anti-α_D antibodies on macrophage accumulation in the peritoneum.

Mice were treated with affinity purified anti-α_D polyclonal antibody (100 µg) or rabbit IgG 4 hours before the development of peritonitis. Peritonitis was induced by intraperitoneal injection of 3% Brewer thioglycollate medium (0.5 ml). 72 hours later, the mice were euthanized by isoflurane inhalation, and their peritoneal cavities were lavaged with 5 ml PBS. The numbers of migrated macrophages in the peritoneum were determined by nonspecific esterase assay. The data shown are representative of four separate experiments. The data represent the means ± SD for the control and the treated groups with seven mice in each group. Asterisk denotes the statistically significant difference between the control and treated groups (P < 0.02). This experiment was replicated 4 times with the same results.

Discussion

The β₂ integrins play a central role in the inflammatory response by mediating adhesive reactions of leukocytes during their migration from the peripheral blood to extravascular sites of injury. In a previous study we have demonstrated that two members of the group, α_Mβ₂ and the most recently discovered integrin α_Dβ₂, have similar recognition specificity in that they are capable of supporting adhesion to numerous ECM proteins [13]. In addition to their salient multiligand binding properties, another distinctive feature of both integrins is that they are upregulated to high densities on the surface of leukocytes. When coming into play, these qualities may influence adhesive andmigratory properties of leukocytes in unusual ways. Indeed, previous studies proposed that a gradual increase of α_Mβ₂ density on the surface of stimulated neutrophils can serve as a “brake” in the overall process of neutrophil migration [3,12]. The commonality in properties between α_Mβ₂ and its sister integrin α_Dβ₂ suggests that upregulation of α_Dβ₂ may perform a similar function in monocyte/macrophage migration. To define the specific contribution of α_Dβ₂ to the process of macrophage migration, we have utilized α_Dβ₂-expressing HEK 293 cells as well as a murine IC-21 macrophage cell line expressing different levels of α_Dβ₂ and we also compared α_Dβ₂ function with that of α_Mβ₂. In addition, using anti-α_D blocking antibodies we have examined the role of α_Dβ₂ in a mouse model of peritonitis. The main findings of our study are that: 1) when expressed at a low density, integrin α_Dβ₂ cooperates with β₁/β₃ integrins in supporting cell migration; 2) when expressed at a high density, α_Dβ₂ inhibits β₁/β₃-governed cell migration; and 3) activation of macrophages in vitro and in vivo is accompanied by strong upregulation of α_Dβ₂.

Inhibition of cell migration by integrins expressed at high density is well established. As shown experimentally in studies with transfected cells and based on generated mathematical models, cell migration exhibits a bell-shaped dependence on the cell-substratum adhesiveness with the maximal speed occurring at its intermediate value [26,27,28]. This is an adhesiveness at which the cell can form new attachments at the cell front but break bonds at the rear. The total cell-substratum adhesiveness is a resultant of three variables, including ligand concentration, integrin activation and integrin density [28]. Under conditions when two other parameters are constant, a moderate increase in integrin density enhances the strength of cell adhesion and promotes migration. However, as the receptor density continues to raise, this leads to an increased number of adhesive bonds between integrins and the substratum, which inhibits migration. The behavior of α_Dβ₂ observed in our experiments fits this general model well, i.e. α_Dβ₂ expressed at a low level promotes migration while high density α_Dβ₂ inhibits migration. The conclusion that the increase in the α_Dβ₂-mediated adhesiveness is responsible for inhibition of migration is based on the effect of anti-α_D function-blocking antibodies which enhance cell migration to multiple α_Dβ₂ ligands. Thus, disruption the excessive integrin-substrate bonds diminish cell adhesion and restores cell migration.

Although the relationships between integrin density and the rate of migration are evident, they have been established for recombinant cells transfected with only one integrin [28]. A unique feature of myeloid leukocytes (which is also modeled in our experiments with the α_Dβ₂-expressing HEK 293 cells) is that they express different types of integrins, β₁ and β₂, both of which can engage the same proteins in the ECM. Furthermore, these receptors are expressed differently on various leukocyte populations. While circulating neutrophils express low levels of β₁ integrins and their levels are only modestly induced on stimulated cells [29,7], α_Mβ₂ is strongly upregulated on neutrophils migrated into inflamed tissues [10]. Furthermore, as we demonstrate here, β₁ and β₃ integrins are not induced on migrating monocytes whereas the levels of α_Dβ₂ are increased several fold. Therefore, although both β₁/β₃ and β₂ integrins can support adhesion to the same ligands their function in migration appears to be different: β₁/β₃ integrins drive migration whereas β₂ integrins, through alterations in their density, modulate the β₁/β₃-driven migration. The role of β₁ integrins as primary migratory receptors is also supported by our data with anti-β₁ function-blocking mAbs which invariably inhibited migration of the α_Dβ₂-transfectd HEK 293 cells to various ECM proteins. In addition, blocking the common α_V subunit of α_Vβ₁ and α_Vβ₃ blocked migration of IC-21 macrophages. However, α_Dβ₂, when expressed at a low density, can contribute to migration as evidenced by the following data: 1) anti-α_D antibodies partially inhibited cell migration; and 2) migration of the α_Dβ₂-transfected cells expressing a low density of integrin was enhanced compared to mock-transfected cells (Fig.3A). In this regard, previous in vivo studies demonstrated that anti-α_Dβ₂ antibody inhibited the accumulation of monocytes and neutrophils at sites of spinal cord injury [30]. Since infiltration of leukocytes was determined 48 h after injury, the period when α_Dβ₂ seems not to be significantly upregulated, the anti-α_Dβ₂ antibody might exert the inhibitory effect on leukocyte extravasation. Thus, depending on its density, α_Dβ₂ may have a dual role in monocyte/macrophage migration, both pro-migratory and anti-migratory.

Previous studies demonstrated that expression of α_Dβ₂ requires de novo synthesis and occurs during in vitro differentiation of human peripheral blood monocytes to macrophages in the presence of selected cytokines [15]. Moreover, oxLDL and AcLDL have been identified as stimuli that augment α_Dβ₂ gene expression [15]. Also, activation with PMA can induce α_D mRNA in HL60 monocytic cells [15] and upregulate α_Dβ₂ surface expression in isolated eosinophils [31]. Consistent with these findings, culturing murine macrophages with PMA led to a 5-fold increase in α_Dβ₂ levels after 48–72 h whereas the levels of α_Mβ₂ remained unaltered (Fig.5). Our data that both anti-α_D and anti-α_M antibodies inhibited cell migration of non-stimulated cells in vitro suggest that both α_Mβ₂ and α_Dβ₂ support migration. However, on PMA-stimulated cells, over-expressed α_Dβ₂ suppressed migration while α_Mβ₂ still supported migration. Thus, in these cells α_Mβ₂ appears to provide no added contribution to the total cell adhesiveness which arises mainly from the excess of α_Dβ₂.

An important finding of the present study is that macrophages elicited by thioglycollate injection upregulated α_Dβ₂ by ~4-fold after 72 h indicating that physiologic activators in the inflamed peritoneum also induce α_Dβ₂ synthesis in vivo. The anti-α_D antibodies increased the number of macrophages recruited in the peritoneal cavity after 72 h of inflammatory stimulation. Thus, the data with blocking anti-α_D antibodies in vitro and in vivo are entirely consistent and, therefore, the blockade of excessive adhesive bonds formed between α_Dβ₂ and its ligands may explain the increased recovery of macrophages from the peritoneal cavity. However, different mechanisms may account for this effect. First, the anti-α_D antibodies potentially may inhibit increased α_Dβ₂-mediated monocyte adhesion to the ECM proteins in the mesenteric wall as monocytes move from circulation into the peritoneal cavity, thus resulting in their augmented migration. Nevertheless, a pattern of α_Dβ₂ upregulation seems to argue against the involvement of α_Dβ₂ during post-endothelial monocyte migration. Since α_Dβ₂ synthesis occurs slowly over a 72 h period it is reasonable to assume that monocytes upregulate α_Dβ₂ only after they have traversed the peritoneal wall and develop into inflammatory macrophages. Therefore, the most probable interpretation is that α_Dβ₂ is responsible for enhanced macrophage adhesion to the peritoneal wall and their retention in the peritoneal cavity. Hence, the increased number of macrophages recovered from the peritoneal fluid likely reflects the inhibitory effect of anti-α_D antibodies on adhesion of inflammatory macrophages to the peritoneal wall. Further studies using different models of inflammation may help to define how α_Dβ₂ contributes to monocyte/macrophage recruitment and their retention at sites of inflammation.

The increased expression of α_Dβ₂ on macrophage foam cells [14] suggests that α_Dβ₂ may play roles in the development of atherosclerosis. Based upon our present results, it is tempting to speculate that upregulation of α_Dβ₂ on differentiated macrophages can increase cell adhesiveness, resulting in the retention of macrophages in the atherosclerotic lesions and thereby evoking the development of atherosclerosis. It appears that α_Dβ₂ upregulation on inflammatory macrophages contrasts that of α_Mβ₂. Analogous to our data that α_Mβ₂ is not upregulated on IC-21 cells, surface expression of this integrin on elicited peritoneal macrophages was augmented to a lesser extent than that of α_Dβ₂ (1.5- versus 4-fold). Furthermore, expression of α_Mβ₂ has been downregulated in atherosclerotic lesion-derived macrophages [32] and α_Mβ₂-deficiency has not influenced the development of atherosclerosis [33]. However, since β₂-deficiency does protect against atherosclerosis [34], other β₂ integrins appear to play roles in lesion progression. Given the remarkable pattern of α_Dβ₂ upregulation by atherogenic lipids, as well as overexpression of α_Dβ₂ ligands such as CCN1 in the atherosclerotic lesions [35] the potential role of this integrin in disease progression merits further investigation.