Antineutrophil Cytoplasmic Antibodies Preferentially Engage FcγRIIIb on Human Neutrophils¹

Antineutrophil cytoplasmic Abs (ANCA)⁴ directed against granule contents are commonly found in patients with systemic vasculitis (1, 2, 3). ANCA are distinguished by two common indirect immunofluorescence staining patterns (1, 2, 3). ANCA that show a granular cytoplasmic pattern are called cANCA (mostly anti-proteinase 3 (PR3) autoantibodies) and ANCA that show a perinuclear staining pattern are called pANCA (most commonly anti-myeloperoxidase (MPO)). The correlation of cANCA titers with disease activity in some studies suggests that ANCA may be directly involved in the pathogenesis of Wegener’s granulomatosis (WG) (1, 2, 3). ANCA targets displayed on the cell surface, due to cell priming and/or apoptosis, allow ANCA to bind to neutrophils (4, 5, 6, 7, 8). Binding of ANCA to ANCA target and subsequent engagement of Fcγ receptors results in activation of an oxidative burst and cytokine secretion in neutrophils and monocytes (6, 7, 9, 10). While these observations do not preclude direct ANCA-induced cell activation (11), the importance of FcγR engagement by ANCA in WG is highlighted by our recent observation that neutrophil FcγR are a genetic risk factor for severity of disease in patients with WG (12)⁵.

Human neutrophils express two structurally and functionally distinct Fcγ receptors, the transmembrane FcγRIIa and the glycosyl-phosphatidylinositol (GPI)-anchored FcγRIIIb which is expressed at 10-fold higher density than FcγRIIa (13). While both receptors are capable of independently inducing cell activation, recent data have highlighted differences in the signaling properties and in the functional responses induced by these receptors (14, 15, 16, 17). Of particular interest in the pathogenesis of WG, signaling through FcγRIIIb induces a different neutrophil adhesive phenotype than signaling through FcγRIIa (16), and FcγRIIIb is the dominant Fcγ receptor responsible for the immune complex-induced oxidative burst in neutrophils (18, 19).

Several groups have demonstrated engagement of FcγRIIa by ANCA (6, 7, 20). However, direct experimental evidence of FcγRIIIb engagement by ANCA has been obscured by activation-induced receptor shedding and activation-induced expression of receptor on the surface of neutrophils. To address these issues, we have used a panel of anti-FcγRIII mAb to examine the ability of neutrophil-bound ANCA to engage and block the ligand binding site of FcγRIIIb under conditions that limit activation-induced shedding of FcγRIIIb. Human cANCA and pANCA, and murine anti-PR3 and anti-MPO mAb, do indeed engage FcγRIIIb. The ability to detect this interaction is inversely proportional to the relative affinity of the reporting anti-FcγRIII mAb. Our results indicate that multiple FcγR may participate in neutrophil activation by ANCA.

Phycoerythrin (PE)-labeled F(ab′)₂ goat anti-mouse IgG, anti-mouse μ-chain, and anti-human IgG were from Biosource (Burlingame, CA) and Jackson ImmunoResearch (West Grove, PA). The anti-PR3 mAb CLB-702 (mIgG1) was from Research Diagnostics (Flanders, NJ), and the anti-MPO mAb MPO-7 (mIgG1) was from Dako (Carpinteria, CA). The anti-PR3 IgM mAb WGM3 (21) was provided by Drs. W. L. Gross and E. Csernok (Medizinische Universität zu Lübeck, Lübeck, Germany). IgG from serum (normal and ANCA positive) was prepared by protein G affinity chromatography (Pharmacia, Piscataway, NJ). The columns were eluted with 0.1 M glycine-HCL (pH 2.7) according to the manufacturer’s instructions. ANCA-positive sera included two patients positive for cANCA with biopsy-proven WG and one patient positive for pANCA with polyarteritis nodosa. The concentration of IgG in all preparations was determined by a radial immunodiffussion assay (The Binding Site, Birmingham, U.K.). FITC-labeled anti-FcγRII mAb IV.3 and anti-FcγRIII mAb 3G8 were from Medarex (Annandale, NJ). Anti-CD43 (clone DF-T1) was obtained from Dako. Anti-CD62L mAb DREG-56-FITC (mIgG1) was from Immunotech (Miami, FL), and the anti-CD11b mAb F6.2-PE (mIgG2a) was from ExAlpha (Boston, MA). Isotype controls (labeled and unlabeled) were obtained from Sigma (St. Louis, MO). The anti-FcγRIIIb mAbs 214.1, 135.9, and 30.2 (22) were biotinylated using NHS-LC-Biotin from Pierce (Rockford, IL) using standard protocols (23). PE-conjugated streptavidin for the detection of bound primary Ab was from Caltag (San Francisco, CA).

1,10-Phenanthroline,N-α-p-tosyl-l-lysine chloromethyl ketone (TLCK), diisopropyl fluorophosphate (DFP), leupeptin, pepstatin A, and aprotinin were from Sigma. Pefabloc was from Boehringer Mannheim (Indianapolis, IN). All other standard chemicals were from Fisher Scientific (Pittsburgh, PA) except for BSA, which was from Boehringer Mannheim.

Blood was collected into heparinized tubes (Becton Dickinson, Franklin Lakes, NJ) using a protocol approved by the Institutional Committee on Human Rights in Research. All donors were healthy volunteers genotyped for the FcγRIIa allelic polymorphism (FcγRIIa-H131/R131) using an allele-specific PCR-based assay as we have described (24). Neutrophils were isolated by Ficoll gradient centrifugation at room temperature (15) and immediately used in the experiments. All washes were performed in 125 mM sodium chloride, 10 mM sodium phosphate, 5 mM potassium chloride, and 5 mM glucose, pH 7.35 (buffer 1). Cells were resuspended at a concentration of 5 × 10⁶ cells/ml in buffer 1 containing 0.1% BSA and 1 mM 1,10-phenanthroline. Before use, 1.09 mM CaCl₂ and 1.62 mM MgCl₂ was added.

The ability of unlabeled ANCA to engage the ligand binding site of FcγR was measured by the inhibition of binding of receptor-specific anti-ligand binding site mAb by ANCA. Inhibition was defined as the percent decrement in anti-FcγR mAb binding induced by ANCA compared with isotype control and is expressed as a positive number. Total ANCA binding was independently determined with a PE-labeled F(ab′)₂ goat anti-mouse IgG or anti-human IgG incubated for 15 min at 4°C.

A total of 50 μl of neutrophils in buffer 1 with 0.1% BSA, 1 mM 1,10-phenanthroline, 1.09 mM CaCl₂, and 1.62 mM MgCl₂ chloride were placed in polypropylene tubes (Falcon Labware, Franklin Lakes, NJ) and incubated with 10 μg/ml of anti-PR3 mAb, anti-MPO mAb, or isotype control (mIgG1) for 10 min at 37°C, followed by 20 min at 4°C. For the experiments with human ANCA the final concentration of total IgG was 1 mg/ml. The cells were then washed and centrifuged at 350 × g for 5 min with 2 ml of buffer 1 at 4°C containing 0.1% BSA, 1 mM 1,10-phenanthroline, 1 mM EDTA, and 1 mM sodium azide. In the inhibition assay, a saturating concentration (2 μg/ml) of biotinylated anti-FcγRIIIb mAb (135.9, 214.1, or 30.2), mAb 3G8-FITC (2 μg/ml), or mAb IV.3-FITC (2 μg/ml) was incubated for 8 min at 4°C, followed by two washes. Separate control experiments indicated that this was the minimum incubation time required for optimal mAb binding. Binding of the biotinylated mAbs was detected with PE-labeled streptavidin incubated for 15 min at 4°C.

A total of 3–4 ml of blood was collected into heparinized Vacutainer tubes (Becton Dickinson) and immediately chilled to 4°C, washed twice in buffer 1 (see above), and resuspended in the original volume. Aliquots (50 μl) of this washed whole blood were used per tube (Becton Dickinson Labware, Lincoln Park, NJ). Stimulation with mIgG1 anti-PR3 at various concentrations was performed at 37°C for the indicated times as we have previously described (6, 16). Negative control samples were incubated with an isotype control mAb, and positive controls included cells stimulated with 1 μM FMLP (Sigma). All samples were then treated with 1 ml FACS Lysing Solution (Becton Dickinson Immunocytometry, San Jose, CA) for 10 min at room temperature, washed once with 2 ml PBS, and analyzed by flow cytometry.

Data was collected using a FACScan (Becton Dickinson Immunocytometry) that was routinely calibrated using fluorescent beads (Sphero Rainbow; Spherotech, Libertyville, IL). Neutrophil doublets and higher aggregates (typically <5%) were detected by analysis with anti-CD43 (which allowed elimination of events with fluorescence greater than the dominant singlet peak) and light scatter properties. The results are expressed as mean fluorescence intensity (MFI) of the histogram data. Data are presented as ΔMFI above the MFI of the appropriate isotype control.

Analysis of flow cytometry listmode data was done using CellQuest (Becton Dickinson Immunocytometry). The mean channel fluorescence of histogram data was compared using Student’s t test and analysis of variance. Two-tailed paired-sample t tests or Wilcoxon’s paired-sample test were performed for the analysis of the inhibition of binding data, and a probability of 0.05 was used to reject the null hypothesis that there is no difference between the groups or pairs.

Several reports have shown that ANCA bound to neutrophil-associated ANCA target can engage FcγRIIa via their Fc region (6, 7, 20). Data documenting ANCA engagement of the other constitutive neutrophil FcγR (FcγRIIIb) have been limited by activation-induced, metalloprotease-dependent shedding of FcγRIIIb and the mobilization of preformed intracellular stores of the receptor to the cell surface (25, 26). To minimize these effects, we tested a panel of protease inhibitors and found that shedding of FcγRIIIb induced by FMLP was inhibited completely by the metalloprotease inhibitor 1,10-phenanthroline (27). Other protease inhibitors such as Pefabloc, DFP, and TLCK were less effective, and no inhibition was found with aprotinin, pepstatin A, and leupeptin (data not shown). The inhibition of FMLP-induced FcγRIIIb shedding by 1,10-phenanthroline was dose dependent with an IC₅₀ of ∼0.4 mM and the maximum inhibition reached at 1 mM (results not shown). The quantitative expression of FcγR surface expression and ANCA engagement of FcγR (see below) was not different in the presence of 1 mM and 5 mM 1,10-phenanthroline. Expression of FcγRIIa was unaltered by activation in the presence or absence of 1,10-phenanthroline. Therefore, 1 mM 1,10-phenanthroline was used in all subsequent experiments with no effect on cell viability, as detected by trypan blue exclusion, over the time course of these studies.

We have previously shown that anti-PR3 and anti-MPO engage and block the ligand binding site of FcγRIIa. To determine whether the presence of 1,10-phenanthroline alters the interaction between cell bound ANCA and neutrophil FcγR, we determined the inhibition of the binding of the anti-FcγRIIa mAb IV.3 by anti-PR3 mAb in the presence of 1,10-phenanthroline. PR3 and MPO expression was induced by repletion of divalent cations (Ca²⁺ and Mg²⁺) at 37°C (Fig. 1,A) as we have previously described (6). Freshly isolated neutrophils, or cells that were incubated in parallel at 4°C, did not bind anti-PR3 or anti-MPO. Using an mIgG1 monoclonal anti-PR3, a correlation between the amount of PR3 expression and inhibition of mAb IV.3 binding was found for neutrophils from donors homozygous for R131 allele (the mIgG1 binding allele) of FcγRIIa (Fig. 2). In all of the experiments over a range of PR3 target densities, the binding of the anti-PR3 mAb resulted in significant inhibition of the binding of mAb IV.3 (p < 0.05, n = 6). As a control, IgM anti-PR3 mAb did not inhibit the binding of anti-FcγRIIa mAb IV.3 to neutrophils (1.2 ± 2.4% inhibition (n = 4), p > 0.05). These results are in agreement with our earlier study (6) and confirm that ANCA bound to neutrophils can engage FcγRIIa in the presence of 1,10-phenanthroline.

The sensitivity to detect ANCA binding to FcγRIIIb will be dependent on the relative affinity of the reporting anti-FcγRIII mAb. The commonly used anti-FcγRIII mAb 3G8 has high affinity for the ligand binding site of FcγRIIIb and blocks immune complex binding to neutrophils completely (28). We considered the possibility that an anti-FcγRIII mAb with lower relative affinity might be more suitable in this assay and selected three mAb with relatively low affinities for the ligand binding site of FcγRIIIb (mAbs 214.9, 135.9, and 30.2) (22). Within this group of mAb, the relative affinities are mAb 214.9 < 135.9 < 30.2 based on their differential capacity to inhibit the binding of immune complexes to neutrophils and on the various degrees by which their binding could be competitively inhibited by an excess of other high affinity anti-FcγRIIIb mAb (such as mAb 3G8) (22). Using these mAbs as reporters, binding of anti-PR3 mAb to the surface of neutrophils resulted in inhibition in the binding of these anti-FcγRIII mAb relative to cells incubated with an isotype control (Fig. 3,A, Table I). In a panel of normal donors, the binding of the anti-FcγRIIIb reporter mAbs was significantly decreased by anti-PR3 (Fig. 4,A, Table I) and by anti-MPO (Fig. 4,B, Table I). In the same panel of normal donors, there was a range of binding of the high affinity mAb 3G8 after binding of anti-PR3 (Figs. 3,B and 4A) or anti-MPO (Fig. 4,B), but the average binding was not altered by binding of either anti-PR3 (Table I) or anti-MPO to the neutrophil surface (Table I) presumably due to the ability of mAb 3G8 to displace ANCA Fc region engagement of FcγRIIIb. It is also important to note the ability of anti-PR3 and anti-MPO, at low ANCA target density (MFI < 100), to significantly inhibit the binding of the reporter anti-FcγRIII mAb 214.1, 135.9, and 30.2 but not the anti-FcγRIIa reporter mAb IV.3 (Fig. 2). This finding indicates that ANCA may preferentially engage FcγRIIIb on the surface of neutrophils.

The magnitude of reduction of ANCA-induced anti-FcγRIIIb binding (mAb 214.1 ≥ 135.9 > 30.2 > 3G8) is inversely proportional to the relative affinity of the reporting anti-FcγRIIIb mAb. However, there was no correlation between the magnitude of PR3 or MPO expression and the magnitude of the reduction in net anti-FcγRIII mAb binding (Fig. 4). Even with blockade of receptor shedding, translocation of intracellular FcγRIIIb to the cell surface precludes a precise quantitative relationship between ANCA binding and net reduction of anti-FcγRIIIb binding. Such translocation of new receptor to the membrane should result in an underestimation of the actual amount of FcγRIIIb engagement by ANCA. Indeed, the binding of mAb 3G8 was found on average to be slightly increased after binding of anti-PR3 (Fig. 3,B, Table I) or anti-MPO (Table I) indicating that there was a net increase in receptor number after incubation with these murine ANCA.

Using a monoclonal anti-PR3 IgM ANCA as a negative control, no significant inhibition of the binding of any anti-FcγRIII mAb was observed (Table I). Finally, there was no significant difference in binding of any of the anti-FcγRIII mAbs to neutrophils incubated with an isotype control (murine IgG1) compared with the samples incubated with buffer alone.

We next determined if engagement of FcγRIIIb was generalizable to human ANCA. Accordingly, IgG from two plasmas containing cANCA activity, one plasma containing pANCA activity and one plasma from a disease-free healthy control donor was isolated as described in Materials and Methods. Both cANCA and pANCA containing IgG fractions bound to ANCA target positive neutrophils relative to normal serum human IgG (Fig. 1,B). We have previously shown that human cANCA binding to neutrophils significantly blocks the detection of the FcγRIIa ligand binding site with the anti-FcγRIIa mAb IV.3 (6). Likewise, cANCA binding to neutrophils significantly inhibited the binding of the two lowest affinity anti-FcγRIIIb mAbs 214.1 and 135.9 (Table II) when compared directly with normal human serum IgG, which did not inhibit the binding of these mAb. cANCA also inhibited the binding of the higher affinity mAb 30.2, but the magnitude of inhibition did not reach statistical reliability with our sample size (Table II). As with cANCA, we observed significant engagement of FcγRIIIb by pANCA. Indeed, binding of all three lower affinity anti-FcγRIIIb mAb was significantly inhibited by pANCA binding to neutrophils (Table II).

To determine whether the binding of ANCA to the surface of neutrophils triggers cell activation in a manner that is consistent with engagement of FcγRIIIb, we quantitated the time-dependent changes in surface expression of the β₂ integrin CD11b and CD62L (L-selectin) induced by ANCA. Cross-linking of FcγRIIa results in the simultaneous up-regulation of CD11b and shedding of CD62L whereas FcγRIIIb results in the up-regulation of CD11b without altering CD62L expression (16). Accordingly, ANCA engagement of FcγRIIIb before FcγRIIa should elicit an initial phenotype characterized by increased CD11b and preserved CD62L expression followed by CD62L shedding. Binding of anti-PR3 mAb to neutrophils in washed whole blood, performed as we have previously described (29), resulted in the time-dependent change in expression of both CD11b and CD62L (Fig. 5). After 30 min incubation, we observed significant up-regulation of CD11b expression relative to unstimulated cells (Fig. 5, A and B) and only at the later time points did we observe a significant loss of CD62L expression (Fig. 5 C). Addition of an isotype control mAb to whole blood did not induce any significant changes in the expression of either CD11b or CD62L. These results are consistent with a model of neutrophil activation in which ANCA initially engage FcγRIIIb followed by engagement of FcγRIIa.

FcγR have been shown to play a role in ANCA-induced neutrophil and monocyte activation (6, 7, 9, 20), perhaps in conjunction with direct ANCA-induced stimulation (4, 11, 30). However, it is not clear which FcγR are involved in ANCA-induced activation. Engagement of only FcγRIIa on neutrophils and monocytes has been clearly shown (6, 7, 9, 20). This study shows that ANCA on neutrophils engage FcγRIIIb as well as FcγRIIa and that engagement of FcγRIIIb is observed for both cANCA (anti-PR3) and pANCA (anti-MPO). The importance of engagement of FcγRIIIb as well as FcγRIIa is underscored by the observation that FcγRIIIb has distinct functional properties relative to FcγRIIa (15, 16), that FcγRIIIb can modulate the quantitative functional capacities of other cell surface receptors such as FcγRIIa and CD11b/CD18 (15, 31, 32), that FcγRIIIb and FcγRIIa engage distinct signal transduction pathways (14, 17), and that alleles of FcγRIIIb are a genetic risk factor for the severity of disease in patients with WG (12)⁵.

The ability to detect FcγRIIIb engagement by both human and murine ANCA is dependent 1) on the relative affinity of the reporter anti-FcγRIIIb mAb, 2) on the degree of receptor shedding, and 3) on surface expression of pre-formed intracellular receptor stores. We have used a panel of lower relative affinity anti-FcγRIII mAb (22) in this study to address the first issue and the inclusion of the metalloprotease inhibitor 1,10-phenanthroline to prevent activation-induced FcγRIIIb shedding. The affinity of 3G8 is particularly high and greater than any of the other mAb used in this study (28). mAb 3G8 is capable of competing effectively with multivalent immune complexes and most probably it is capable of displacing the Fc portion of ANCA from the low affinity binding site of FcγRIIIb. This competitive advantage results in no inhibition of binding of mAb 3G8 by ANCA, and due to up-regulation of FcγRIIIb surface expression induced by priming and/or by ANCA during the incubation, a net increase in mAb 3G8 binding to the receptor was found in some experiments (Fig. 3,B and Table I). This increase illustrates that the quantitative inhibition of binding found with other antireceptor mAb clearly underestimates the actual engagement of FcγRIIIb by ANCA on the neutrophil surface.

We have shown that both cANCA and pANCA, as well as murine mAb with corresponding reactivities, can engage FcγRIIIb. However, it is important to note the ability of the anti-MPO and anti-PR3, at low ANCA target density (MFI < 100), to significantly inhibit the binding of the reporter anti-FcγRIIIb mAbs 214.1, 135.9 and 30.2 (Fig. 4). Since engagement of FcγRIIa apparently requires higher ANCA densities (Fig. 2), this observation suggests that FcγRIIIb might be preferentially engaged by ANCA when ANCA target is limiting. Such preferential engagement is consistent with the nearly 10-fold higher density of surface expression of FcγRIIIb relative to FcγRIIa (33). This numerical advantage in surface receptor expression could be further amplified by activation-induced translocation of a pre-formed intracellular pool of FcγRIIIb to the cell surface with a net gain in expression even in the context of receptor shedding. Together, these points also suggest that initial neutrophil triggering of cell responses may occur via FcγRIIIb. Indeed, anti-PR3 binding to neutrophils induced a transient phenotype characterized by increased CD11b and preserved CD62L expression (Fig. 5), a phenotype that is induced by cross-linking of FcγRIIIb but not FcγRIIa (16). Furthermore, we have recently shown that the more functionally active allele of FcγRIIIb, the NA1 allele, is a significant risk factor for the development of renal disease in patients with WG (12)⁵. It is important to note that the significance of engagement of FcγRIIIb by ANCA bound to ANCA target on the cell surface would be the transient induction of a qualitatively different adhesive phenotype not necessary a difference in the kinetics of neutrophil activation. Our previous data have shown that the kinetics of FcγRIIa- and FcγRIIIb-induced up-regulation of CD11b expression is identical (16). It is also worth noting that the level of membrane associated ANCA target on neutrophils from patients with WG (5, 34) is considerably lower than the levels achieved after maximal induction in this study, a finding that favors FcγRIIIb in ANCA-induced neutrophil activation.

The ability of ANCA to engage FcγR is influenced by the subclass of the ANCA and by the allelic variants of FcγR expressed by the host. For example, only the FcγRIIa-H131 allele binds human IgG2 well (24). FcγRIIIa-V176 binds IgG1 and IgG3 with 10-fold greater efficiency than the FcγRIIIa-F176 allele (35), and some evidence suggests that FcγRIIIb is responsible for neutrophil responses triggered by IgG3 (36). Our finding of a significant association between FcγRIIIb alleles and renal disease in patients with WG does not preclude a role for polymorphisms in FcγRIIa expressed on neutrophils and monocytes and FcγRIIIa expressed on monocytes that can alter the binding of IgG. These polymorphisms may impact on the development of other ANCA-associated vasculitides and are clearly important for further study in the context of ANCA-associated vasculitis. The nature of the ANCA target may also play a role in the neutrophil activating potential of ANCA. For example, studies by Csernok et al. (5) have shown that PR3 preferentially binds to neutrophil membranes relative to MPO, a finding that is consistent with the higher level of PR3 target expression relative to MPO on neutrophils from patients with WG (5, 34).

In conclusion, we find that ANCA engage FcγRIIIb on the surface of neutrophils as detected by the inhibition of binding of anti-FcγRIIIb mAb when the shedding of this receptor is blocked, even at relatively low detectable levels of ANCA-target expression. The initial and perhaps preferential engagement of FcγRIIIb will clearly influence the signal transduction pathways and the specific cellular responses induced by ANCA. For example, cross-linking of FcγRIIIb has been shown to induce a pro-adhesive neutrophil phenotype (16) and FcγRIIIb is the dominant Fcγ receptor responsible for the immune complex-induced oxidative burst in neutrophils (18, 19). It is also likely that ANCA binding to monocytes will lead to engagement of the available repertoire of available FcγR (FcγRI, FcγRII, and FcγRIIIa). A better understanding of the dynamics of ANCA-FcγR interactions on both neutrophils and monocytes may lead to new therapeutic approaches in the systemic vasculitides such as the use of soluble FcγR and/or peptide mimics to block ANCA engagement of FcγR on the cell surface.

We thank Patsy Redecha for the purification of human IgG from serum, Andy Beavis for his expert assistance with flow cytometry, and Drs. E. Csernok and W. Gross for the IgM anti-PR3 mAb. We are grateful for the advice and discussions with our colleagues at the Hospital for Special Surgery where much of this work was performed.