Mac-1 (CD11b/CD18) as Accessory Molecule for FcαR (CD89) Binding of IgA1

IgA, the principal ligand for FcαRI, exists in serum as monomeric IgA and at mucosal sites as secretory IgA (SIgA). SIgA consists of dimeric IgA linked by joining chain and secretory components. Human polymorphonuclear leukocytes (PMN) and mouse PMN transgenic for human FcαRI exhibited spreading and elicited respiratory burst activity upon interaction with either serum or SIgA. However, PMN devoid of the β2 integrin Mac-1 (Mac-1−/−) were unable to bind SIgA, despite expression of FcαRI. Consistent with this, serum IgA stimulated Mac-1−/− PMN oxygen radical production, in contrast to SIgA. Binding studies showed the secretory component, by itself, to interact with Mac-1-expressing PMN, but not with Mac-1−/− PMN. These data demonstrate an essential role for Mac-1 in establishing SIgA-FcαRI interactions.

I mmunoglobulin A, representing the predominant Ab isotype in humans, is present in the circulation primarily in monomeric form (serum IgA), and in mucosal secretions as dimeric complexes, called secretory IgA (SIgA) 3 (1). SIgA exists as dimeric IgA (dIgA) containing an additional joining chain and secretory component (SC), which has been proposed to increase IgA stability. Serum IgA is produced by splenic and bone marrowderived plasma cells. Mucosal plasma cells in the lamina propria produce joining chain-associated IgA dimers (dIgA), which bind with high affinity to the polymeric IgR (pIgR) present on the basolateral surface of mucosal epithelium (2). Subsequently, dIgA-pIgR complexes are endocytosed and transported to the mucosal lumen. SC, a portion of the pIgR, is cleaved off during this transepithelial transport, leading to local SIgA production (3).
Serum IgA can initiate numerous immune effector functions via Fc␣RI, which include phagocytosis, cytotoxicity, respiratory burst activity, release of inflammatory mediators, and Ag presentation (4,5,8). Moreover, serum IgA-Fc␣RI interactions on Kupffer cells may provide a "second-line" of defense in mucosal immunity (9).
SIgA is considered crucial for the "first-line" mucosal defense (10 -12), however, its precise in vivo role remains unclear. SIgA has been shown to interact with microorganisms and to block virulence factors, which may prevent microbial invasion. Furthermore, SIgA can trigger phagocyte respiratory burst activity, degranulation, and TNF-␣ production (13)(14)(15)(16)(17). Another role suggested for SIgA is neutralization of viruses within infected epithelia (18,19). Significant correlations have been reported between specific titers of SIgA and resistance to infection (20).
PMN and macrophages are the major phagocytic populations at mucosal sites. Upon microbial invasion these phagocytes are recruited, increase receptor levels, and become activated (or "primed"). Fc and complement receptors are essential for the recognition and elimination of foreign targets. Mac-1 (CR3, CD11b/ CD18, or ␣ m ␤ 2 ) is an important PMN complement receptor that binds multiple ligands, including C3bi, ICAM-1, fibrinogen, and ␤-glucan. Mac-1 is important for leukocyte adhesion, migration, phagocytosis, cytotoxicity, and chemotaxis (21). Furthermore, Mac-1 has been proposed to act as "signaling partner" for other leukocyte receptors, including IgG receptors (22). Although Mac-1 has been implicated in FcR function, this is the first study demonstrating a crucial role for Mac-1 in FcR ligand binding.

PMN isolation
Human PMN were isolated from heparinized venous blood of healthy volunteers by Ficoll-Histopaque (Sigma-Aldrich) density gradient centrifugation. PMN purity determined by cytospin preparations exceeded 95%, and cell viability was Ͼ98%.

PMN binding and spreading on IgA
To exclude interference with complement, all Ab preparations were heatinactivated before use (30 min, 56°C). Glass slides (Menzel, Braunschweig, Germany) were coated with 0.5% (w/v) BSA (Boehringer Mannheim, Mannheim, Germany), 100 g/ml human serum IgA (ICN Pharmaceuticals, Costa Mesa, CA), dIgA (kindly provided by Dr. C. van Kooten, Leiden University Medical Center, Leiden, The Netherlands), SIgA (Sigma-Aldrich), or IgG (CLB, Amsterdam, The Netherlands) for 3 h at 37°C, and rinsed with PBS. Isolated human or mouse PMN (2 ϫ 10 5 cells) were incubated in RPMI 1640 medium (with 10% FCS) on coated slides for 10 or 30 min at 37°C. Cells were fixed in 3.7% paraformaldehyde, and stained for actin with phalloidin-FITC (1:200; Sigma-Aldrich), or stained for Fc␣RI and Mac-1 (see Colocalization studies) for 30 min at 20°C. Samples were mounted in Mowiol (with 2.5% 1,4-diazobicyclo-[2.2.2]-octane), and PMN binding and spreading was analyzed by confocal laser scanning microscopy using a Leitz DMIRB fluorescence microscope (Leica, Voorburg, The Netherlands) interfaced with a Leica TCS4D confocal laser microscope (Heidelberg, Germany). Cell morphology was imaged just above (0.2 m) the coated surfaces. PMN diameters were measured orthogonally, and then averaged. At least 30 PMN from three different experiments were analyzed. In Mac-1 blocking experiments, PMN were preincubated with 10 g/ml anti-CD11b mAb for 30 min at 4°C before SIgA binding. In additional experiments, PMN were preincubated with 0.1 M N-acetyl-D-glucosamine (NADG; Sigma-Aldrich) for 10 min at 20°C, and plated on SIgA-coated slides in the presence of NADG. Cytochalasin D (Sigma-Aldrich) was used at 10 g/ml to study the role of actin microfilament polymerization in Fc␣RI-IgA binding.

Respiratory burst measurements
The luminol-ECL method was used for analysis of real-time respiratory burst activity. Polystyrene tubes were coated with PBS, 100 g/ml serum IgA, or SIgA for 3 h at 37°C, and blocked with HEPES ϩ buffer (containing 20 mM HEPES, pH 7.4, 132 mM NaCl, 6 mM KCl, 1 mM MgSO 4 , 1.2 mM NaH 2 PO 4, 1 mM CaCl 2 , 5 mM glucose, and 0.5% (w/v) BSA) for 1 h at 37°C. Isolated mouse PMN (4 ϫ 10 5 cells) in HEPES ϩ buffer were incubated in these tubes, and placed in a 953 LB Biolumat (Berthold, Wildbad, Germany). Luminol (150 M) was injected in all tubes and light emission was recorded continuously for 30 min at 37°C. As a positive control, PMN were stimulated with 100 ng/ml PMA (Sigma-Aldrich), and PMN incubated with luminol only served as negative control. In blocking experiments, PMN were incubated with Fc␣RI mAb My43 supernatant during the assay (25).

Statistical analyses
Data are expressed as means Ϯ SD. Statistical significance was determined by two-tailed unpaired Student's t tests. Values of p Ͻ 0.05 were considered significant.

Role of Mac-1 in Fc␣RI-IgA binding
To assess Mac-1 involvement in Fc␣RI-ligand binding, we studied PMN from human Fc␣RI-Tg mice crossed with CD11b knock-out mice (Mac-1 Ϫ/Ϫ ). Because Fc␣RI represents a low affinity IgAR, we analyzed PMN binding to IgA-coated surfaces. Tg Mac-1-expressing PMN efficiently bound and spread onto serum IgA (IgA), dIgA, and SIgA-coated slides within 10 min (Fig. 1, upper panels). PMN spreading capacity on IgA-coated surfaces varied among different IgA forms, and was quantified by measuring PMN diameters (18 Ϯ 4.2 m on serum IgA, 15 Ϯ 3.3 m on dIgA, and 14.6 Ϯ 3.7 m on SIgA). Tg Mac-1 Ϫ/Ϫ PMN bound properly to serum IgA and dIgA, but did not interact with SIgA (Fig. 1, lower panels). Furthermore, Mac-1 Ϫ/Ϫ PMN, exhibited impaired spreading on slides coated with serum IgA (12.1 Ϯ 2.5 m) or dIgA (11.9 Ϯ 4.2 m). PMN of Ntg Mac-1 ϩ/Ϫ littermates, expressing Mac-1, but not Fc␣RI, served as controls and did not bind human IgA. These data were supported by studies with human PMN. Binding and spreading capacity of human PMN on serum IgA, dIgA, and SIgA was similar to that of Tg Mac-1 ϩ/Ϫ PMN. However, F(abЈ) 2 of ␣-Mac-1 mAb inhibited spreading of human PMN on SIgA, but not on serum IgA slides (data not shown). In addition, all phagocytes bound IgG-or BSA-coated surfaces (Fig. 1), irrespective of Fc␣RI/Mac-1 expression. Binding of both human and Tg mouse PMN to SIgA was abolished in the presence of cytochalasin D. These results indicate Mac-1 to be required for phagocyte-Fc␣RI binding of SIgA.

Respiratory burst activity
To study the significance of PMN interaction with immobilized IgA, we investigated respiratory burst activity of mouse (Ntg Mac-1 ϩ/Ϫ , Tg Mac-1 ϩ/Ϫ , Tg Mac-1 Ϫ/Ϫ ) and human PMN. Both serum IgA and SIgA induced comparable oxygen radical production in Tg Mac-1 ϩ/Ϫ PMN (Fig. 2, left graph). However, Tg Mac-1 Ϫ/Ϫ PMN were stimulated by serum IgA, but not by SIgA (Fig. 2,  middle graph). Importantly, Mac-1-expressing and Mac-1-deficient PMN exhibited similar respiratory burst activity in response to PMA. Ntg Mac-1-expressing PMN did not initiate respiratory activity in response to serum IgA or SIgA (Fig. 2, right graph). Human PMN exhibited comparable oxygen radical production upon interaction with serum IgA or SIgA, albeit that serum IgAinduced respiratory burst activity was significantly faster (data not shown). Blocking studies with My43, an IgM recognizing the Fc␣RI ligand-binding domain (25), decreased oxygen radical production close to background levels (data not shown), showing IgAinduced respiratory burst to depend on interaction with Fc␣RI. These data demonstrate requirement of both Fc␣RI and Mac-1 for SIgA-induced respiratory burst activity.

Colocalization of Mac-1 and Fc␣RI
Because our data pointed to an essential role for Mac-1 in Fc␣RI-IgA interactions, we next studied the distribution of Mac-1 and Fc␣RI in PMN membranes. Isolated Tg Mac-1 ϩ/Ϫ PMN were  incubated on surfaces coated with BSA, serum IgA, or SIgA, whereupon Fc␣RI and Mac-1 were stained with FITC (green), and TRITC (red), respectively. Tg Mac-1 ϩ/Ϫ PMN mediated spreading on both serum IgA (Fig. 3A, upper panel) and SIgA (Fig. 3A,  middle panel). Fc␣RI was expressed on the plasma membrane of PMN, but also in filipodia-like outgrowth of cells. A similar staining was found for Mac-1, and colocalization was indicated in yellow ( Fig. 3; merged pictures). Tg Mac-1 ϩ/Ϫ PMN did not spread on BSA-coated slides, and exhibited overall membrane staining of Fc␣RI and Mac-1 (Fig. 3A, lower panel). As a control, Fc␣RI and ␤ 2 integrin LFA-1 membrane staining was examined, which revealed no colocalization (Fig. 3B). Importantly, experiments performed with human PMN revealed Fc␣RI and Mac-1 expression to colocalize as well (see below).

Mac-1 binds SC
Our results demonstrated only Mac-1-expressing PMN capable of binding SIgA-coated surfaces. Next, we evaluated the capacity of Mac-1 ϩ/Ϫ and Mac-1 Ϫ/Ϫ PMN to bind aggregated IgA by flow cytometry. Ntg Mac-1 ϩ/Ϫ PMN were ineffective in binding serum IgA-complexes, but interacted well with SIgA complexes. On the contrary, Ntg Mac-1-deficient PMN were unable to bind SIgA complexes (data not shown). These observations pointed to a role for the SC in PMN-SIgA binding. Therefore, we examined the capacity of recombinant SC to interact with isolated human and mouse PMN. As shown in Fig. 4A, Mac-1-expressing PMN readily bound SC, irrespective of the presence of Fc␣RI. However, binding of SC to Mac-1-deficient Tg and Ntg PMN, was abrogated. SC interacted with human PMN as well (Fig. 4B). These data show SC, either recombinant or present in its normal configuration (i.e., complexed with dIgA), to be capable of interacting with Mac-1.

Involvement of Mac-1 lectin-binding site in SIgA binding
We next evaluated the region of Mac-1 involved in SIgA-Fc␣RI binding. Human PMN were plated on immobilized serum IgA or SIgA in the absence (control) or presence of NADG, which interacts with the Mac-1 lectin-binding domain (31), and were stained for both Fc␣RI and Mac-1 (Fig. 5). Human PMN exhibited spreading on both serum IgA and SIgA, like Tg Mac-1 ϩ/Ϫ PMN. Incubation of PMN with NADG resulted in a blockade of PMN binding to SIgA-coated surfaces, whereas binding to serum IgA was barely affected by NADG (Fig. 5, lower panel). The partial colocalization of Fc␣RI and Mac-1 observed in control PMN bound to serum IgA-coated slides was decreased in the presence of NADG. PMN incubated on BSA-coated surfaces in the presence of NADG exhibited no change in binding. Furthermore, NADG impaired binding of Tg Mac-1 ϩ/Ϫ to SIgA, but not to serum IgA (data not shown). In contrast, Abs directed against the I-domain of Mac-1, although partly inhibiting PMN spreading, did not inhibit binding of human PMN to serum and SIgA (data not shown). These results implicate the Mac-1 lectin-binding domain to be involved in Fc␣RI-SIgA interaction.

Discussion
The significance of IgA in immune defense is well-recognized, but incompletely understood (32). Although a number of studies focused on the molecular characterization of Fc␣RI and IgA, relatively little is known about the biology of Fc␣RI-IgA interactions. The present study documents a new role for Mac-1, a ␤ 2 integrin, in ligand binding of Fc␣RI. Mac-1 was found to be crucial for interaction of PMN Fc␣RI with SIgA, and subsequent PMN activation. Our data support Mac-1 binding to SC to underlie this phenomenon, and implicate the Mac-1 lectin-binding domain to be involved.
Modulation of receptor-ligand interactions by accessory molecules is well-documented for a number of immune receptors, including cytokine receptors, integrins, and FcR (28,(33)(34)(35)(36). IgG binding to Fc␥RI (CD64) and Fc␥RIII (CD16a) is dependent on association with the FcR ␥-chain (36 -38). In this study, Mac-1 is identified as a novel accessory protein, crucial for Fc␣RI-SIgA interaction. In SIgA, SC is covalently ligated to the C␣2 and C␣3 domains within the dIgA Fc region (39), which may interfere with the affinity of Fc␣RI for SIgA (40 -42). The present study demonstrates SC binding to PMN in a Mac-1-dependent manner. When one appreciates the structure of SC, Mac-1 interaction with SC seems rational. SC represents a heavily glycosylated protein belonging to the Ig superfamily (3). Moreover, we observed Mac-1 lectin-binding domain, bearing carbohydrate-binding specificity (31), involved in PMN binding of SIgA.
Importantly, Mac-1 modulates PMN function upon IgA binding, as indicated by stimulation of a potent respiratory burst. Our data show requirement of both Fc␣RI and Mac-1 for SIgA-induced respiratory burst activity, because Mac-1-deficient Tg PMN and Ntg PMN were unable to elicit oxygen radical production in response to SIgA. A role for Mac-1 in Fc␥R functions, including respiratory burst has been shown before (28,(43)(44)(45). CD18 interactions with the actin cytoskeleton and associated proteins may enable Mac-1 signaling (46,47). Furthermore, the present work is consistent with earlier studies documenting a role for ␤ 2 integrins in superoxide production of eosinophils mediated by SIgA, and not serum IgA (48). SC by itself was described to activate eosinophil functions, but not PMN functions, which was proposed to relate to an additional (15 kDa) "SC receptor" on eosinophils (49). Indeed, we observed Mac-1-deficient eosinophils capable of interacting with both SC and SIgA (our unpublished data).
In conclusion, this study provides evidence for a crucial role of Mac-1 in phagocyte SIgA binding. Enhanced IgA responses upon incubation of PMN with GM-CSF, TNF-␣, or IL-8 have been reported (50 -52), although the underlying mechanisms remain to be addressed. We propose Mac-1 involvement in these phenomena to be likely, because of the well-recognized ability of these cytokines and chemokines to activate Mac-1 (53-55). Requirement of two molecules interacting with one ligand may provide immune effector cells with an extra way to regulate their activity. When "first-line" defense  fails, mucosal inflammation leads to PMN recruitment and priming, accompanied by up-regulation of Mac-1, which may trigger more potent Fc␣RI-mediated responses. We hypothesize Mac-1 to increase affinity of Fc␣RI-SIgA interaction via binding of SC.