FcγRIIIb Allele-Sensitive Release of α-Defensins: Anti-Neutrophil Cytoplasmic Antibody-Induced Release of Chemotaxins ¹

Human neutrophils contain many distinct granule populations that can be mobilized on inflammatory stimulation (1, 2). The release of these granules is a highly regulated process with a gradient of responsiveness to stimulation. For example, extracellular release of azurophilic granules requires a higher threshold of activation than release of specific or secretory vesicles (3, 4, 5). This is in keeping with a primary role for azurophilic granule enzymes in intracellular degradation of phagocytosed particles.

The control of azurophilic granule release is important in limiting neutrophil-mediated tissue destruction at sites of inflammation. These granules contain key enzymatic mediators of inflammation such as myeloperoxidase (MPO),⁴ elastase, collagenase, and various acid hydrolases. More recently, it has become clear that among the subpopulations of azurophilic granules expressed in neutrophils (1), there is a subpopulation of granules that contain the α-defensin antibacterial/chemotactic peptides (6, 7). The receptor-mediated regulation of α-defensin release has not been explored.

α-Defensins are a family of small (3.5- to 4.5-kDa) cationic antimicrobial peptides with three to four intramolecular cysteine disulfide bonds and are widely distributed in mammals, insects, and plants (8). In addition to their antimicrobial properties, some of these peptides are potent chemotaxins for mononuclear cells (9), including dendritic cells and CD45RA⁺ and CD8 T lymphocytes (10, 11)). Based on the pattern of their cysteine residues and disulfide connections, six α and two β forms have been characterized in humans. Human α-defensins 1, 2, 3, and 4 are found in azurophilic granules in human neutrophils and thus are termed human neutrophil peptides (HNP) (12), whereas human α-defensins 5 and 6 (HNP 5 and 6) are generated by small intestine Paneth cells (13). Based on their chemotactic activity, human neutrophil α-defensins contribute to modification of acquired immune responses by mobilization of mononuclear cells, T cells, and dendritic cells (9, 10, 11).

Activation of human neutrophils by antineutrophil cytoplasmic Abs (ANCA) is at least partially dependent on engagement of FcγRs on the surface of neutrophils. We hypothesized that engagement of FcγR on neutrophils would mobilize the α-defensin-positive azurophilic granule compartment and provide a link between ANCA-induced neutrophil activation in ANCA-associated vasculitis (in particular Wegener’s granulomatosis (WG)) and granuloma formation. Human neutrophils express two structurally distinct FcγR, FcγRIIa and FcγRIIIb (14). Both of these receptors are functionally polymorphic with the H131/R131 single-nucleotide polymorphism of FcγRIIa altering the binding of hIgG₂ (15) and mIgG1 (16) and the NA1/NA2 single-nucleotide polymorphisms of FcγRIIIb altering the quantitative functional capacity of the receptor (17, 18, 19). ANCA have been shown to engage both neutrophil FcγR upon binding to cell-associated ANCA-target Ag with preferential engagement of FcγRIIIb presumably due to its numeric predominance over FcγRIIa (20, 21, 22).

In this study, we show that cross-linking of neutrophil FcγR results in significant release of α-defensins. The magnitude is influenced by the host receptor genetics, with donors homozygous for the NA1 allele of FcγRIIIb displaying significantly greater release. This release is induced by both monoclonal and human cytoplasmic staining ANCA (cANCA) (anti-proteinase 3 (PR3)) from patients with WG, and importantly, its blockade by the Fc-specific TG19320 peptide suggests an intervention strategy that can be carried to the in vivo setting.

Whole blood anticoagulated with EDTA was obtained from healthy volunteers. All donors were genotyped for both neutrophil FcγRs (FcγRIIa and FcγRIIIb) by allele-specific PCR (23, 24). In some studies, neutrophils were isolated by centrifugation through a discontinuous Ficoll-Hypaque density gradient as previously described (25). The protocol for phlebotomy was approved by the Institutional Committee on Human Rights in Research.

Anti-FcγRIIIb mAb 3G8 F(ab′)₂ or IgG, anti-FcγRII mAb IV.3 Fab and FITC-conjugated 3G8 (mIgG1) and IV.3 (mIgG2b) were purchased from Medarex (Annandale, NJ). F(ab′)₂ fragments of F(ab′)₂-specific goat anti-mouse IgG (GαM) for anti-FcγR mAb cross-linking, Fc-specific PE-conjugated GαM F(ab′)₂ for detection of primary Abs, and murine F(ab′)₂ were obtained from Jackson ImmunoResearch Laboratories (West Grove, PA). The anti-PR3 mAb CLB 12.8 (mIgG1) was from Research Diagnostics (Flanders, NJ), and the anti-MPO mAb MPO-7 (mIgG1) was from Dako (Carpinteria, CA). Anti-CD66b mAb CLB-B13.9-FITC (mouse IgG1 (mIgG1)) and FITC- or PE-conjugated isotype-matched controls were purchased from Caltag (San Francisco, CA). Human cANCA (anti-PR3) or perinuclear staining ANCA (pANCA) (anti-MPO) plasmas were obtained from five patients with active WG. IgG from the plasmas was purified as previously described (21, 22).

Neutrophils in washed whole blood, which eliminates isolation-induced neutrophil activation (26), were stimulated as described previously (27). Briefly, EDTA-anticoagulated blood was chilled to 4°C, washed twice in modified PBS (125 mM sodium chloride, 10 mM phosphate, 5 mM potassium chloride, 5 mM glucose, pH 7.35), and then resuspended in the original volume. Washed whole blood was keep at 4°C, and all washes were performed with modified PBS at 4°C. For neutrophil stimulation, aliquots (300 μl) of washed whole blood were incubated with saturating concentrations of the anti-FcγRII- and/or anti-FcγRIII-specific mAb fragments (1 μg/ml IV.3 Fab and/or 2 μg/ml 3G8 F(ab′)₂) for 15 min at 4°C, washed twice, and resuspended in buffer with 1.09 mM CaCl₂ and 1.62 mM MgCl₂. Cell-bound mAb fragments were cross-linked with F(ab′)₂ GαM (35 μg/ml) at 37°C for 15 min. Alternately, cells were stimulated with PMA (Sigma-Aldrich, St. Louis, MO; 10 μg/ml) as a positive control or with F(ab′)₂ GαM as a negative control. Endotoxin levels in all reagents were below detection limits (Limulus ameobocyte assay; Sigma-Aldrich) as described (22, 27). In some studies, cells were pretreated with 50 ng/ml rTNF-α (R&D, Minneapolis, MN) for 20 min at 37°C before mAb cross-linking. After stimulation, cell-free supernatants were then collected and stored at −20°C. After two washes, the remaining GαM binding sites on the cells were blocked with murine F(ab′)₂ (88 μg/ml). Aliquots of stimulated cells were incubated with primary mAb for flow cytometry for 30 min at 4°C. Unconjugated mAb were detected with PE-conjugated GαM F(ab′)₂ for 30 min. FcγR expression on neutrophil in washed whole blood was quantitated with FITC-conjugated anti-FcγRIIIb mAb 3G8 or anti-FcγRII mAb IV.3. All samples were then treated with FACS Lysing Solution (BD Biosciences, San Jose, CA) for 10 min at room temperature (RT), washed once, and analyzed by flow cytometry (FACSCalibur; BD Biosciences).

The concentration of α-defensins (HNP 1–3) in diluted cell-free supernatant was measured with the HNP 1–3 ELISA Kit (Cell Sciences, Norwood, MA) according to the manufacturer’s instructions. This assay quantitates the three principal α-defensins, HNP 1–3, that are unique to neutrophils and account for >99% of the total defensin content of these cells. The concentration of elastase in diluted cell-free supernatants was also measured by ELISA. Microtiter plates were coated with 100 μl of a rabbit polyclonal anti-elastase (Biodesign, Saco, ME) (51 μg/ml) for 2 h at RT, and the wells were blocked with 0.1% BSA in PBS (PBS/0.1%BSA). The diluted supernatants (1:25 or 1:50) or elastase standards (0.5 ng/ml to 100 ng/ml) (Sigma), all diluted in PBS/0.1% BSA, were added to wells and incubated for 2 h at RT followed by four washes with PBS, 0.1% BSA. Biotinylated rabbit polyclonal anti-elastase Ab prepared with the Biotin Protein Labeling Kit (Roche Diagnostics, Mannheim, Germany) was added at 100 μg/ml in PBS, 0.1% BSA and incubated for 2 h at RT. After four washes, plates were incubated for 30 min at RT with avidin-alkaline phosphatase (Sigma-Aldrich) diluted 1:1000 in PBS, 0.1% BSA, followed by an additional four washes. The plates were developed by the addition of 50 μl of 1 mg/ml p-nitrophenyl phosphate (Sigma-Aldrich) in 10% diethanolamine (Sigma-Aldrich). After 30 min at RT, the reaction was stopped with 50 μl of 0.1% EDTA and OD₄₀₅ was determined (VMAX; Molecular Devices, Sunnyvale, CA).

Functional MPO activity was performed as previously described (28). Briefly, isolated neutrophils (2 × 10⁶/ml) were added to wells coated with IgG (100 μg/ml) and incubated at 37°C for 1 h to heterotypically cross-link with FcγRIIa and FcγRIIIb. After stimulation, cell-free supernatants were collected and stored at −20°C. Released MPO was quantitated by mixing 300 μl of dilutions of cell-free supernatants with 250 μl of 80 mM NaH₂Po₄ buffer, pH 5.4, and 80 μl of 20 mM 3,3′,5,5′-tetramethylbenzidine in dimethylformamide. Then 100 μl of 1.0 mM H₂O₂ were added, and the ΔOD₆₅₅ was recorded at 15-s intervals at 25°C for 1 min (29). Released functional MPO activity is represented as ΔOD₆₅₅/min during the liner phase of the reaction.

ANCA (mAb or human) stimulation of neutrophils in washed whole blood was performed as described above except that the washed cells were preincubated for 40 min at 37°C to induce cell surface expression of proteinase 3 (PR3) ANCA target. Preliminary studies indicated that near maximal surface PR3 was detected at this time point. Aliquots were then incubated with or without anti-PR3 mAb CLB 12.8 (10 μg/ml) or diluted cANCA-positive plasma for 45 min. After stimulation, supernatants were harvested as described above, and cells in washed whole blood were analyzed for CD66b expression by flow cytometry as described above (30).

In some studies, we used the IgG-Fc region-binding peptide (TG19320) (31) or a scrambled TG19320 peptide to block ANCA-induced neutrophil activation. In these studies, aliquots of washed whole blood were incubated with anti-PR3 mAb or human ANCA pretreated with various concentration of peptides at 37°C at 45 min.

Data were collected using a BD Biosciences FACSCalibur. The instrument was routinely calibrated using fluorescent beads (Rainbow Calibration Particles, 3.5 μm; Spherotech, Libertyville, IL). Neutrophils in whole blood were identified by characteristic light scatter properties and confirmed by FcγR characteristic expression (14) using FITC-conjugated IV.3 and 3G8. As previously reported, cell surface expression of PR3 displayed a bimodal pattern (32, 33).

Analysis of flow cytometry listmode date was done using CellQuest (BD Biosciences). The results are expressed as mean fluorescence intensity (MFI) of the histogram data.

For quantitation of azurophilic granule release in donors homozygous for FcγRIIIb alleles, experiments were performed in a matched paired donor design. The effect of activating stimuli on FcγR-mediated degranulation are presented as percent increase relative to control incubations. Data are displayed as mean ± SEM. The effects of stimuli were compared using the t test. p = 0.05 was used to reject the null hypothesis that there is no difference between the groups or conditions.

To study the release of α-defensins in response to FcγR cross-linking, we used ELISA to quantitate the concentration of HNP 1–3 in cell-free supernatants after targeted FcγR cross-linking on neutrophils in washed whole blood. No significant release of HNP 1–3 above baseline was detected in response to homotypic cross-linking of either FcγRIIa or FcγRIIIb. However, HNP 1–3 were released from neutrophils after co-cross-linking of both FcγRIIa and FcγRIIIb together and after stimulation with 10 μg/ml PMA (Fig. 1,A). Because TNF-α can prime neutrophils for enhanced stimulus-mediated degranulation (28, 34, 35), we examined the release of α-defensins induced by FcγR cross-linking in TNF-α-primed neutrophils to enhance release of azurophilic granules. TNF-α (50 ng/ml) alone induced significant release of HNP 1–3 and markedly enhanced the FcγR-mediated release of HNP 1–3 (Fig. 1 B). In the presence of TNF-α, homotypic cross-linking of FcγRIIa or FcγRIIIb and heterotypic cross-linking of FcγRIIa and FcγRIIIb induced significant release of HNP 1–3 relative to control incubation. These results indicated that cross-linking of FcγRs could induce the release of the α-defensin subpopulation of azurophilic granules in both unprimed and primed neutrophils.

Donors homozygous for the NA1 allele of FcγRIIIb display a quantitatively greater FcγR-dependent phagocytic capacity (17) and greater FcγRIIIb enhancement of FcγRIIa-specific phagocytosis relative to NA2-homozygous donors (19). To test whether the NA1 allele might have a greater capacity to mobilize azurophilic granules relative to the NA2 allele, we assessed the influence of FcγRIIIb-NA1/NA2 polymorphism on the quantitative magnitude of release of α-defensins in response to heterotypic cross-linking of FcγRIIa and FcγRIIIb on TNF-α-primed neutrophils. Indeed, heterotypic cross-linking from donors homozygous for the NA1 allele of FcγRIIIb showed significantly greater release of HNP 1–3 (Fig. 2).

There was no difference in the concentration of released HNP 1–3 in washed whole blood supernatants after incubation with TNF-α alone in these donors (HNP1–3 (ng/ml): 198.7 ± 26.42 vs 196.1 ± 38.07 for NA1/1 vs NA2/2, p = 0.4340; seven pairs). The surface expression of FcγRIIa and FcγRIIIb on neutrophils was identical in the two groups (Fig. 3), with the mean MFI for FcγRIIa being 56.3 ± 3.2 vs 55.1 ± 2.7 (p = 0.571; seven pairs) and the mean MFI for FcγRIIIb being 366.4 ± 47.4 vs 377.6 ± 49.4 (p = 0.688; seven pairs) for NA1/1 and NA2/2 donors, respectively. FcγRIIa is also functionally polymorphic. The H131/R131 polymorphism alters the binding of human IgG2 (15) and murine IgG1 (16) which results in differences in ligand-induced FcγRIIa-mediated functions (36, 37). Because we have used Fab or F(ab′)₂ for engagement of FcγRs on neutrophils, FcγRIIa alleles should not influence the magnitude of α-defensin release in this system. Indeed, there was no difference in the release of HNP 1–3 in donors homozygous for the H131 allele and the R131 allele of FcγRIIa when donors were matched for the FcγRIIIb NA genotype (data not shown).

There was little evidence of elastase release or display of MPO and PR3 on the cell surface after FcγR cross-linking on unprimed neutrophils in washed whole blood. However, FcγR induced release of elastase was strikingly enhanced by TNF-α (Fig. 4,A). Likewise, cell surface expression of MPO and PR3, both ANCA target Ags found in azurophilic granules, was significantly up-regulated in response to FcγR cross-linking after TNF-α priming (Fig. 4, B and C). Expression of MPO after receptor-specific stimulation was also significantly greater in NA1/1 donors relative to NA2/2 donors (Fig. 5), as was FcγR-stimulated release of soluble MPO, measured functionally in the medium (Table I). In contrast to MPO, PR3 remains mainly membrane bound and is released only in minute amounts into the extracellular medium (38). Together, our results demonstrate that FcγR-mediated degranulation of azurophilic granules, including the α-defensin subpopulation, was quantitatively greater in donors homozygous for the NA1 allele relative to donors homozygous for the NA2 allele of FcγRIIIb.

ANCA bind neutrophil-associated ANCA target and engage FcγR with preferential binding to FcγRIIIb (20, 21, 22). Although resting neutrophils in washed whole blood expressed little to no PR3 on the cell surface, incubation at 37°C for 40 min induced detectable cell surface PR3 expression as measured by the anti-PR3 mAb CLB12.8 (33). Incubation of cells with CLB12.8 for an additional 45 min at 37°C induced further degranulation as assessed by an increase in CD66b expression (22), the magnitude of which was dependent on the density of PR3 expression (Fig. 6,A). This activation also showed FcγRIIIb allele sensitivity (Fig. 6 B).

Anti-PR3 mAb also induced release of α-defensins (HNP 1–3) (Fig. 7,A). α-Defensin release by cells incubated alone (44.1 ± 8.8 ng/ml; Fig. 7,A) was not different from cell incubated with an isotype control mAb (45.7 ± 7.3 ng/ml). Similarly, purified IgG from cANCA- and pANCA-containing plasmas, but not purified IgG from control donors that lack ANCA activity, resulted in significant release of HNP 1–3 (Fig. 8), and the magnitude of release was comparable with that of the mAb ANCA-induced release (Fig. 7,A). This release was further enhanced by pretreatment of the washed whole blood with TNF-α (Fig. 8).

To demonstrate conclusively that the α-defensin-releasing activity in ANCA-containing plasma is indeed due to IgG-Fc region engagement of FcγR, we used the Fc-region binding peptide TG19320, a peptide previously shown to block IgG-Fc region binding to FcγR both in vitro and in vivo (31). α-Defensin release induced by the anti-PR3 mAb 12.8 was completely blocked by this Fc region-blocking peptide (Fig. 7,A). Likewise, ANCA-induced release by four cANCA-containing WG plasmas was completely inhibited by the Fc region-binding peptide (Fig. 7,B). Pretreatment of the cANCA plasmas, or anti-PR3 mAb 12.8, with a nonfunctional scrambled TG19320 had no effect on the anti-PR3 induced neutrophil activation (results not shown). The TG19320 studies, and the results in Fig. 8 using purified IgG from cANCA and pANCA positive plasma, clearly demonstrate that IgG in WG derived plasma are responsible for inducing α-defensin release. Thus, ANCA-mediated neutrophil activation, initiated by monoclonal or polyclonal cANCA, induces significant release of the α-defensins HNP1–3 and is completely dependent on FcγR interactions.

ANCA-mediated cross-linking of neutrophil FcγR induces extracellular release of α-defensins capable of recruiting mononuclear, dendritic, and T cells to ANCA-induced granuloma. The magnitude of release is affected by the FcγRIIIb genotype, which may explain the association of FcγRIIIb genotype with disease phenotype (39). The ability of the TG19320 Fc-blocking peptide, which has been used effectively in some models of autoimmunity (31) to block this release, reveals a potential novel therapeutic approach to ANCA-related diseases.

Because isolation procedures can induce neutrophil activation (26, 33), we used a washed whole blood assay system, which avoids density gradient centrifugation and hypotonic lysis. We also avoided cytochalasin B, which enhances release of neutrophil granule constituents but which is clearly nonphysiological. Previously, we have reported that FcγR-specific stimulation induces exocytosis of secretory vesicles in isolated neutrophils (27). Our current data demonstrate that FcγR engagement initiated release of the α-defensin subpopulation of azurophilic granules, an effect that was enhanced by TNF-α. This effect was also sensitive to the functional NA1/NA2 polymorphism of FcγRIIIb (Table I), as has been seen with specific granules (30). α-Defensins HNP1–3 are specific for neutrophils, and elastase, although found in monocytes and basophils, predominates in neutrophils (40). Thus, measurement of release of both α-defensins and elastase into the supernatant was feasible in the washed whole blood paradigm. Of course, FcγRIIIb is expressed by neutrophils and not by monocytes, providing another parameter of cell type specificity.

Because subpopulations of azurophilic granules are found in neutrophils (1) and the dynamics of azurophilic constituent release vary (38), we measured several other azurophilic constituents either on the cell surface or in cell-free supernatant as markers for release of azurophilic granules. Display of cell-associated PR3 occurred more readily than other azurophilic markers, in part because PR3 is contained in secretory vesicles as well as in azurophilic granules and in part because very little is released into supernatant (38). Because myeloperoxidase is also released from monocytes, we used isolated neutrophils for measurement of both cell associated and soluble MPO in supernatants. Our results clearly show that FcγR cross-linking is able to induce the release of azurophilic granules. Interestingly, up-regulation of CD63, a commonly used membrane marker for release of azurophilic granules, was less sensitive than other markers (data not shown). TNF-α was able to enhance release of all azurophilic constituents induced with FcγR cross-linking.

The quantitative difference in release of α-defensins induced by the NA1 and NA2 alleles of FcγRIIIb is not due to differential binding of 3G8 F(ab′)₂ to FcγRIIIb alleles (41) or to quantitative differences in FcγRIIa or FcγRIIIb expression. Furthermore, the FcγRIIa-H131/R131 polymorphism does not influence quantitative magnitude of release of α-defensins in response to receptor-specific cross-linking. Thus, the differences in granule release seen between NA1- and NA2-homozygous donors relates directly to differences in the function of these two different alleles. The same relationship is seen with ANCA-induced release.

In the washed whole blood paradigm, quantitative differences in cell surface PR3 expression correlate with quantitative degranulation induced by anti-PR3 (Fig. 6). We found heterogeneous levels of cell surface PR3 expression on neutrophils among our donors that appeared to be a stable individual characteristic, as reported by others (32, 33). Therefore, to establish the role of FcγRIIIb genetics in anti-PR3-induced activation, we had to control for the level of PR3 expression among FcγRIIIb-homozygous donors to demonstrate the predicted difference in degranulation (Fig. 6).

Although the contribution of FcγRIIIb alleles to ANCA-induced activation support an important role for FcγR, we also explored the ability of an IgG-Fc region blocking peptide, TG19320, which blocks IgG binding to FcγR (31), to alter ANCA-induced activation. Strikingly, TG19320 peptide, but not its corresponding scrambled peptide, completely blocked cANCA-induced release of α-defensins. The same results were shown using multiple cANCA-containing plasmas and from isolated IgG preparations from patients with WG (Fig. 7 B).

ANCA are typically observed in the circulation of patients with WG, microscopic polyangiitis, and Churg-Strauss syndrome. The two major ANCA targets are PR3 and MPO, and cANCA (anti-PR3) occurs in 80–90% of patients with WG (42) whereas other less common ANCA specificities have been reported (43, 44). ANCA can directly activate neutrophils (20, 21, 22, 45), and some data indicate that ANCA titers may parallel disease activity (46). The Fab components of the ANCA IgG molecule ligate ANCA targets. Bound ANCA IgG triggers neutrophils through FcγRIIIb and FcγRIIa; FcγR-independent pathway(s) of activation may also play a role (47). It is likely that the polyclonal nature of the ANCA autoantibody repertoire results in the direct clustering of multiple FcγR and ANCA target, eliciting functional responses. The ability of monoclonal anti-PR3 (so called monoclonal cANCA) to also directly stimulate neutrophils suggests that there are preformed clusters of PR3 (ANCA target) that facilitate FcγR cross-linking or possibly that there are small preformed mAb aggregates. It is also possible that the direct cross-linking of FcγR-PR3 results in functional responses. It is also interesting to speculate on the role that other FcγR might play in ANCA-induced cell activation. ANCA have been demonstrated to activate monocytes in an FcγR-dependent manner (48), implicating FcγRIIa but also FcγRIa. Likewise, activated neutrophils exposed to IFN-γ or IL-10 will also express FcγRIa, making this an additional potential neutrophil FcγR targeted by ANCA. The recent description of the expression of the inhibitory FcγR, FcγRIIb, on monocytes (49) raises the intriguing suggestion that the magnitude of ANCA-induced cell activation could be modulated by changes in the balance of expression of the activating and inhibitory FcγR. The expression of FcγRIIb on human neutrophils is under active investigation.

Our results documenting ANCA-induced α-defensin release from neutrophils provides a mechanism to link ANCA-FcγR interactions with the recruitment of key elements of the acquired immune response to ANCA-induced granuloma. The FcγRIIIb allele sensitivity of this is consistent with previous observations that FcγR alleles are significantly associated with disease activity in patients with WG (39, 50, 51). Furthermore, the ability of a small stable and soluble peptide inhibitor (TG19320) of IgG-FcγR binding presents the possibility of a new strategy in the treatment of WG and also suggests possible approaches to the treatment and study of Ig-mediated diseases.

We thank Dr. Richard Jones (University of Alabama Capstone Medical Center, Tuscaloosa, AL) for kindly providing WG plasmas and clinical data and Debbie McDuffie for expert technical assistance with the MPO functional assay and the ELISAs.