Differential Enhancement of Neutrophil Phagocytosis by Anti–Bactericidal/Permeability-Increasing Protein Antibodies

Abstract

Bactericidal/permeability-increasing protein (BPI) plays a major role in innate immunity through the ability of the N-terminal domain (NTD) to bind LPS, mediate cytotoxicity, and block LPS-induced inflammation. The C-terminal domain mediates phagocytosis of bacteria bound to the NTD. These two domains are linked by a surface-exposed loop at amino acids 231–249 for human BPI, known as the “hinge region.” Autoantibodies to human BPI are prevalent in many chronic lung diseases; their presence is strongly correlated with Pseudomonas aeruginosa and with worse lung function in patients with cystic fibrosis and bronchiectasis. Although prior literature has reported BPI neutralization effect with autoantibodies targeting either NTD or C-terminal domain, the functionality of BPI Ab to the hinge region has never been investigated. Here, we report that Ab responses to the BPI hinge region mediate a remarkably selective potentiation of BPI-dependent phagocytosis of P. aeruginosa with both human and murine neutrophils in vitro and in vivo. These findings indicate that autoantibodies to the BPI hinge region might enhance bacterial clearance.

Key Points

• Anti-BPI Abs targeting the BPI hinge region enhance neutrophil phagocytosis.

• The phagocytosis enhancement effect is dependent on FcγRII and BPI presence.

• BPI autoantibodies can have a beneficial impact on bacterial clearance.

Introduction

Antibodies are a key component of the human adaptive immune response to combat pathogens. Fc receptors for IgG (FcγR) mediate several key antimicrobial activities through phagocytosis, Ab production, and clearance of opsonized pathogens and infected host cells (1–5). Moreover, Ab–Fc receptor interactions activate downstream cytokine production and immune cell Ag presentation (4, 5). Autoantibodies to bactericidal/permeability-increasing protein (BPI), a potent antimicrobial protein residing in neutrophil azurophilic granules, are frequently (range of frequency 40–60%) found in adult patients with cystic fibrosis (CF) (6–8), bronchiectasis (BE) (9, 10), chronic obstructive pulmonary disease (COPD) (11), as well as other disorders (9, 12–20). BPI has been shown to mediate extracellular cytocidal and opsonic activity against Gram-negative bacteria (GNB) (21–23) and intracellular killing of GNB after phagocytosis via fusion with phagolysosome containing secreted BPI (24). Interestingly, autoantibodies to BPI in CF, BE, and COPD exhibit a remarkable restriction to patients with Pseudomonas aeruginosa infection. This association is curious given the association of BPI autoantibodies with worse disease outcome (6, 10, 25). Thus, the pathophysiology underlying the induction of these autoantibodies is still under investigation. Two possible explanations are: 1) the presence of anti-BPI autoantibodies confers disadvantage to the host by inhibiting the innate immune function of BPI (6, 10, 11, 26–30), or 2) the host generates these Abs in response to persistent P. aeruginosa infection to potentiate BPI activity by engaging the Fc receptor on the immune cells.

The N-terminal domain (NTD) of BPI (aa 32–230) binds to negatively charged lipid A moiety of LPS expressed on the GNB outer envelope (21, 31–33) This interaction destabilizes the bacterial membrane leading to bacterial lysis and death (31, 33–38). The C-terminal domain (CTD) of BPI (aa 250–487) contributes to LPS binding and bacterial opsonization, but these activities are not well characterized (22, 39). NTD is linked to CTD by a proline-rich hinge region spanning from aa 231 to 249 (40). Previous studies using affinity-purified anti-BPI Abs from patient serum have shown that autoantibodies to BPI targeting both the NTD and CTD could inhibit the functional activity of BPI in vitro (14, 26). Patient serum-purified Abs targeting the NTD were shown to inhibit BPI antimicrobial activity against E. coli (14), which is consistent with the notion that autoantibodies to the N terminus target the LPS-binding domains that allow binding and killing of GNB. Moreover, affinity-purified IgG with reactivity against the CTD inhibited BPI-induced human neutrophil phagocytosis of E. coli in vitro (26). These data suggested that patient autoantibodies to the C terminus block, rather than enhance, E. coli phagocytosis and, therefore, enhance bacterial survival. This model would suggest that anti-BPI responses in these conditions are consistent with a causal role in bacterial persistence in these states.

In an attempt to better understand the ability of humoral autoimmunity to modulate innate immunity by BPI, we examined the effect of defined Abs to the NTD, hinge, and CTD of BPI. To our great surprise, Abs to human BPI (hBPI) targeting aa 227–254 that comprise the hinge region (aa 231–249) markedly enhanced neutrophil-mediated phagocytosis of P. aeruginosa in vitro and in vivo. This activity required the presence of the Fc domain because it was not observed with F(ab′)₂ antisera and was blocked by Abs to FcγRII. Lastly, we showed that enhanced phagocytosis by anti–227–254 IgG was dependent on, as well as synergistic with, the presence of BPI to enhance bacterial clearance in vivo. Together, these findings provide novel evidence that BPI autoantibodies can have a beneficial impact on bacterial clearance.

Materials and Methods

Mice

Age-matched male and female (6–8 wk old) mice on the C57BL/6 background were used for all experiments. Heterozygous Bpi^+/− (C57BL/6NJ-Bpi^em1(IMPC)J/Mmjax) mice were obtained from The Jackson Laboratory (MMRRC Stock: 42125-JAX; Bar Harbor, ME) and bred in our facility. Tail DNA genotyping was performed on pups between 8 and 14 d of age following the specific strain protocol provided by The Jackson Laboratory (MMRRC Stock: 42125-JAX). Bpi^−/− mice were used. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Research Council (41). All experiments were conducted according to protocols approved by Dartmouth College’s Institutional Animal Care and Use Committee (IACUC Protocol 00002007).

Human neutrophil isolation and lysate generation

Healthy human neutrophils were obtained with consent approved by the Committee for the Protection of Human Subjects of the Geisel School of Medicine at Dartmouth. Whole blood was diluted 1:1 in RPMI 1640 and centrifuged over Ficoll gradient (GE Healthcare). The erythrocyte pellet was diluted 1:1 in RPMI 1640, and neutrophils were separated by 5% dextran sedimentation at 4°C for 1 h. After RBC lysis using 1× BD Pharm Lyse (BD Bioscience), neutrophils were resuspended in RPMI 1640, counted, and used in phagocytosis assays as described later. To generate cell lysate, we resuspended neutrophil pellet in citrate-phosphate buffer (0.2M Na₂HPO₄ + 0.1M citric acid, pH 3.0) and frozen/thawed three times (7). Lysate protein concentration was determined by Pierce BCA Protein Assay Kit (Thermo Fisher).

Bacterial culture and phagocytosis assays

P. aeruginosa strains PAO1 and PA14 were obtained from Dr. Brent Berwin. Bacteria were cultured overnight in lysogeny broth (LB) at 37°C and subcultured for 3 h in LB (1:30) at 37°C. Phagocytosis assay of live P. aeruginosa bacteria was performed as previously described (30, 42, 43). In the presence of 1% heat-inactivated FBS (HyClone), neutrophils (250,000 cells in RPMI 1640, with or without 30-min preincubation with BPI Abs and with or without peptides) were incubated with P. aeruginosa (PAO1 or PA14) for 45 min at 37°C (with or without 1% heat-inactivated human plasma) at 10 or 25 multiplicities of infection (MOIs). The cells were then treated with gentamicin (67 μg/ml) for 20 min at 37°C and washed twice in RPMI 1640. Neutrophils were resuspended in lysis buffer (0.1% Triton X-100 in PBS), and lysed contents (10 μl) were plated on 1.5% LB agar plate. Bacterial colonies (CFUs) were counted after overnight incubation at 37°C, representing the number of live bacteria phagocytosed by the neutrophils. BPI Abs used were anti–61–75 (affinity-purified rabbit polyclonal; Sigma), anti–227–254 (mouse monoclonal; Santa Cruz), anti–227–254 (affinity-purified rabbit polyclonal; Pacific Immunology, Ramona, CA), and anti–248–487 (rabbit polyclonal; ABclonal). BPI peptides used were aa 113–137, 227–254, and 234–248 (Shanghai Royobiotech).

For rabbit polyclonal anti-hBPI 227–254 production, BPI peptide (aa 227–254; Shanghai Royobiotech) was supplied to Pacific Immunology Company for custom Ab production in rabbits. Rabbit sera exhibiting high Ab titers against BPI, tested by direct ELISA, were affinity purified using Sulfolink Coupling Resin (Thermo Scientific) and peptide following the manufacturer’s protocol. Concentration of the purified Ab was determined using NanoDrop spectrophotometer (Thermo Scientific).

FcγRII and FcγRIII blockade

Human neutrophils (250,000 cells/treatment) were preincubated with Ab targeting FcγRII (CD32, 1 μg/ml; eBioscience) and/or FcγRIII (CD16, 1 μg/ml; eBioscience) for 15 min at 37°C. After washing with media, neutrophils were treated with anti-BPI Abs targeting aa 227–254 or 321–450 (1 μg/ml; Santa Cruz) for 30 min at 37°C before adding P. aeruginosa PAO1 MOI 10. Phagocytosis assay was performed as described earlier.

Ab Fc digestion and F(ab′)₂ purification

F(ab′)₂ Abs were generated as described in the manufacturer’s protocol from anti-hBPI 227–254 mouse Ab (2 mg/ml; Santa Cruz) using the Pierce Mouse IgG₁ Fab and F(ab′)₂ Micro Preparation Kit (Thermo Fisher). Sample desalting preparation, digestion, and purification were done according to the manufacturer’s instructions. To detect the purified F(ab′)₂ fractions, 0.1 μg of Ab was run on SDS-PAGE (12% acrylamide gel) in nonreducing conditions before transferring to a nitrocellulose membrane. After blocking in TBST + 3% BSA, the membranes were probed with goat anti-mouse IgG (H+L) peroxidase-labeled secondary Ab (1:50,000 in TBST + 1% BSA; Bio-Rad) to detect digested and purified Ab. SuperSignal West Pico (Thermo Fisher) was used for protein detection via the Bio-Rad Chemidoc MP system and software (Bio-Rad).

BPI detection by immunoblot and immunofluorescence

Immunoblotting against human neutrophil lysate (10 μg, generated by repeated freeze–thaw cycles in citrate-phosphate buffer, pH 3.0) (44) or hBPI (0.2 μg purified hBPI; Athens Research and Technology), with or without preincubation with BPI peptides (aa 227–254 or 234–248, 0.2 or 1 μg/ml; Shanghai Royobiotech), was performed after resolution by SDS-PAGE (12% acrylamide gel) and transfer to a nitrocellulose membrane. After blocking in TBST + 3% BSA, the membranes were probed with anti-hBPI Ab (aa 227–254, 1:1000; Santa Cruz) in TBST + 1% BSA overnight at 4°C, washed, and incubated with goat anti-mouse peroxidase-labeled secondary Ab (1:50,000 in TBST + 1% BSA; Bio-Rad). Secondary Ab-only blots were performed as controls. The membrane was imaged as described earlier.

Immunofluorescence (IF) using human neutrophils was performed following phagocytosis assay using tdTomato-expressing P. aeruginosa PA14 3 × 10⁶ CFUs. After 45 min, human neutrophils were transferred to 13-mm round coverslips to adhere, fixed in 4% paraformaldehyde (Fisher Scientific) for 20 min, washed, permeabilized using 0.5% Triton X-100, and blocked in 5% donkey serum (Sigma-Aldrich). hBPI was detected using either anti-hBPI targeting aa 248–487 (1:200; Abclonal) followed by donkey anti-rabbit Alexa Fluor 647 (1:500; Abcam), or anti-hBPI targeting aa 227–254 (1:200; Santa Cruz) followed by donkey anti-mouse Alexa Fluor 488 (1:500; Jackson Immunoresearch). Anti-hBPI Abs against aa 227–254 and 61–75 were detected using donkey anti-mouse Alexa Fluor 488 (1:500; Jackson Immunoresearch) or donkey anti-rabbit Alexa Fluor 647 (1:500; Abcam), respectively. Samples were mounted using ProLong Gold Antifade Mount with DAPI (Thermo Fisher Scientific) and visualized with the laser point scanning confocal microscope (ZEISS LSM 800; Zeiss), 63×. Intracellular and extracellular bacteria were manually counted from IF images taken. The number of intracellular PA14 tdTomato bacteria per 100 neutrophils was quantified.

Uptake of hBPI and anti-hBPI 227–254 Ab into Bpi^−/− mouse neutrophils in vivo was evaluated by IF after PA14 (3 × 10⁶ CFUs, 3 h postinfection [hpi]) i.p. infection in the presence of anti-hBPI 227–254 IgG (56.4 μg or 1212 nM; Santa Cruz) and exogenous hBPI (10 μg or 606 nM; Athens Research). Mice were lavaged for peritoneal fluid and cells 3 hpi, and cells were left to adhere on 13-mm coverslips and processed as mentioned earlier. hBPI was detected using anti-hBPI 248–487 Ab (1:200; Abclonal) followed by donkey anti-rabbit Alexa Fluor 647 (1:500; Abcam). Ab to hBPI anti–227–254 was detected using donkey anti-mouse Alexa Fluor 488 (1:500; Jackson ImmunoResearch). Samples were mounted and imaged as described earlier.

Bacterial killing assay by hBPI

Purified hBPI (5 μg/ml or 90 nM; Athens Research) was preincubated with or without anti-BPI Abs (90 nM): anti–61–75 (Sigma), 64–76 (Antibodies-online), 227–254 (Santa Cruz), or 248–487 (Abclonal) for 20 min at 37°C, before adding to 10⁸ CFUs of PAO1 (in 100 μl PBS). After 30 min at 37°C, the samples were diluted 1:10,000 in ice-cold LB before plating (10 μl) onto LB agar plates. Plates were incubated overnight at 37°C before CFUs were counted. CFU data were normalized to no-BPI, no-Ab treatment (100% survival). PAO1 viability was also determined by IF staining using Live/Dead BacLight Bacterial Viability Kit (Thermo Fisher), according to the manufacturer’s instructions.

Detection of reactivity to BPI by ELISA

ELISA plates coated with 10 μg/ml P. aeruginosa PA14 lysate in PBS (2 h, 37°C) were blocked in 5% BSA and normal goat serum (Sigma) overnight at 4°C. Purified hBPI (5 μg/ml; Athens Research) was added (1 h, room temperature [RT]) followed by anti-hBPI Abs (0.4 μg/ml anti–61–75 or –227–254; 1 h at RT). Goat anti-rabbit or -mouse (H+L) HRP conjugated (1:50,000; Bio-Rad) followed by the substrate (R&D Systems) was used to detect Ab signal and develop the ELISA. Absorbance (450 nm) was determined using ELISA reader (Epoch; BioTek). Ab binding to BPI was determined by ELISA using hBPI-coated (5 μg/ml; Athens Research) wells (PBS, 2 h 37°C) and the same Abs to BPI as above. Reactivity of patient sera to BPI was detected using the same BPI-coated plates and goat anti-human (H+L) HRP conjugate (1:50,000; Bio-Rad). For competitive ELISA between anti–227–254 and patient serum, plates were coated with P. aeruginosa LPS (10 μg/ml; Sigma) for 2 h at 37°C, blocked, and incubated with hBPI (5 μg/ml, 1 h at RT). Patient or healthy sera were added (1:100, 1 h at RT), followed by anti–227–254 Ab (0.4 μg/ml; Santa Cruz). Anti–227–254 reactivity was detected using Goat Anti-Mouse (H+L) HRP conjugate (1:50,000; Bio-Rad) as mentioned earlier. Anti–227–254 from a different species (rabbit; Pacific Immunology) was used in place of serum as a negative control for mouse anti–227–254 Ab reactivity. Mouse anti–227–254 was used in place of serum for positive control.

Mouse in vivo i.p. infection with P. aeruginosa

For in vivo i.p. sublethal infection, mice (WT or Bpi^−/−) were injected with 3 × 10⁶ CFUs of subcultured PA14 with anti-BPI Abs 61–75 (Sigma), 227–254 (Santa Cruz), or isotype controls (i.p., 300 μl, 28 μg, or 1212 nM). hBPI (10 μg or 606 nM; Athens Research) in PBS was injected i.p. 15 min postinfection. Mice were euthanized 3 hpi for anti–61–75 treatment or 1 hpi for anti–227–254 treatment. Peritoneal fluid was collected, and 10 μl of cell-free supernatant was plated (1:10) on 1.5% LB agar plate. Plates were incubated overnight at 37°C, and recovered CFUs were determined by colony counts. Data were normalized to no-BPI, no-Ab treatment (100%).

Statistical analyses

Data were analyzed using GraphPad Prism 6 software. Student paired and unpaired t tests with Welch correction were applied to determine the significant difference between two datasets. One-way ANOVA with multiple comparisons was applied for more than two datasets. A p value <0.05 was considered significant.

Results

Ab to BPI hinge region enhances P. aeruginosa phagocytosis by human neutrophils

We investigated the effect of Abs to BPI targeting the N terminus (anti–61–75), C terminus (anti–248–487), and the hinge region (anti–227–254) on phagocytosis of P. aeruginosa by human neutrophils. Only the presence of the mouse mAb to the BPI hinge region (227–254) resulted in up to 10-fold higher CFU recovery from human neutrophil lysates, indicating increased phagocytosis of P. aeruginosa in vitro (Fig. 1A). Abs to the N and C termini did not enhance phagocytosis (Fig. 1A). Combination of anti–227–254 and anti–248–487 Abs did not inhibit the phagocytosis enhancement effect of the hinge Ab (Fig. 1B). To confirm the species independence of our observation, the mouse mAb effect was compared with rabbit affinity-purified antisera with identical specificity at different MOIs (Fig. 1C). Comparable increases in PAO1 CFUs recovered from human neutrophils were seen with both antisera and were significantly higher than no treatment (MOI 10: p = 0.0005 for mouse and p = 0.0022 for rabbit; MOI 25: p < 0.0001 for mouse and p = 0.0004 for rabbit species) or isotype controls (MOI 10: p = 0.0004 for mouse and p = 0.0043 for rabbit; MOI 25: p = 0.0002 for mouse and p = 0.0006 for rabbit species) (Fig. 1C). Intracellular and extracellular P. aeruginosa were quantified by IF imaging of tdTomato P. aeruginosa, BPI, and different Abs to BPI. The presence of mouse anti–227–254 increased P. aeruginosa uptake by human neutrophils relative to rabbit anti–61–75 NTD Ab (Fig. 1D). This was reflected in a significant increase in intracellular P. aeruginosa with anti–227–254 IgG treatment (p < 0.0001 compared with isotype and p = 0.0001 compared with anti–61–75 IgG treatment), compared with no change with anti–61–75 IgG or isotype control (Fig. 1E). As predicted, the number of extracellular bacteria decreased with anti–227–254 IgG, compared with isotype controls and anti–61–75 IgG (Fig. 1E). In contrast with their specific modulation of P. aeruginosa uptake by human neutrophils, none of these Abs affected direct bactericidal activity of BPI toward P. aeruginosa (Supplemental Fig. 1).

FIGURE 1.

Ab to hBPI targeting aa 227–254 enhances P. aeruginosa phagocytosis in human neutrophils. (A) Colony-forming units (CFUs) recovered from human neutrophils (n = 3) infected with P. aeruginosa PAO1 (MOI 25) in the presence of hBPI IgG Abs targeting aa 61–75, 227–254, and 248–487 (1 μg/ml). (B) CFUs recovered from human neutrophils (n = 3) infected with P. aeruginosa PAO1 (MOI 10) in the presence of hBPI IgG Abs targeting aa 61–75, 227–254, 248–487, or a combination of 227–254 and 248–487 (1 μg/ml). (C) Colony-forming units (CFUs) recovered from human neutrophils (n = 12, 3, and 7) infected with P. aeruginosa PAO1 (MOI 10, 25) in the presence of hBPI IgG Abs targeting aa 227–254 from mouse or rabbit (1 μg/ml), as well as their isotype controls. (D) Representative IF images of human neutrophils with (left and right columns) or without (middle column) infection with tdTomato-expressing PA14 (MOI 10, 45 min) in the presence of mouse anti-hBPI 227–254 (top row) or rabbit anti-hBPI 61–75 Ab (bottom row), stained with DAPI (blue), anti-hBPI Abs (red or green), and tdTomato P. aeruginosa (magenta). Anti–227–254 treatment shows increased P. aeruginosa uptake relative to anti–61–75 treatment. Fluorescent images were obtained using a 63× oil immersion objective. Scale bars: 20 μm. (E) Quantification of intracellular (number of images analyzed for each column from left to right; n = 22, 4, 12, and 5) P. aeruginosa PA14 count per 100 human neutrophils (top) or extracellular (number of images analyzed for each column from left to right; n = 22, 4, 12, and 5) P. aeruginosa count (bottom) in the presence of mouse anti–227–254 or rabbit anti–61–75 (1 μg/ml), as well as their respective isotype controls. (F, left) Reactivity of BPI Abs (0.4 μg/mL rabbit anti–61–75 or mouse anti–227–254) to BPI (5 μg/ml) that was bound to P. aeruginosa lysate (10 μg/ml). n = 3. (right) Reactivity of BPI Abs (0.4 μg/ml rabbit anti–61–75 or mouse anti–227–254) to BPI-coated (5 μg/ml) ELISA. n = 3. (G) Recovered CFUs from human neutrophils infected with PAO1 (MOI 10) in the absence or presence of mouse anti-hBPI 227–254 Ab (0.5 μg/ml), preincubated with or without BPI peptides aa 113–137, 227–254, or 234–248 (0.05 μg/ml). Data were analyzed by unpaired t test with Welch correction. Error bars represent mean ± SEM. ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05.

IF staining confirmed that the presence of P. aeruginosa was required for the anti–227–254 Ab uptake into human neutrophils (Fig. 1D). Because no anti–61–75 BPI IgG was detected inside the neutrophils, we investigated whether BPI binding to P. aeruginosa inhibited engagement of this Ab, but not anti–227–254 Ab. Indeed, anti–61–75 IgG was unable to bind BPI already bound to P. aeruginosa (Fig. 1F), consistent with the known ability of this antisera to block LPS binding (14). In contrast, anti–227–254 Ab was still able to engage BPI bound to P. aeruginosa (Fig. 1F).

To verify the specificity of the Ab and its effect on phagocytosis enhancement, we neutralized the Ab with peptides targeting the same region. When preincubated with peptides corresponding to aa 227–254 or 234–248, anti–227–254 Ab binding to BPI was markedly inhibited (Supplemental Fig. 2). Preincubation with the relevant peptides specifically reduced the ability of the anti–227–254 Ab to enhance phagocytosis (Fig. 1G). These data suggest that the specificity of the V region of the Ab targeting BPI aa 227–254 is important in its phagocytosis induction effect.

Anti-BPI 227–254 IgG enhances BPI-mediated phagocytosis via FcγRII

To investigate whether anti-BPI 227–254 IgG enhances BPI-mediated phagocytosis of P. aeruginosa in an Fc receptor–dependent fashion, we first performed the phagocytosis assay in the presence of 1% heat-inactivated human plasma and FBS. The addition of 1% plasma reduced the phagocytosis enhancement effect of anti–227–254 by ∼64% (Fig. 2A). These data suggested that this effect of plasma was mediated by FcγR blockade. We thus generated F(ab′)₂ of anti–227–254 IgG (Supplemental Fig. 3). Anti–227–254 F(ab′)₂ failed to enhance phagocytosis of P. aeruginosa by neutrophils (Fig. 2B). To establish which Fc receptor is required, we pretreated human neutrophils with Abs to block FcγRII (CD32) and/or FcγRIII (CD16) before anti–227–254 IgG and P. aeruginosa treatments. Blocking of FcγRII, and not FcγRIII, prevented the increase in P. aeruginosa CFUs recovered from neutrophils seen with anti–227–254 IgG, suggesting that anti–227–254 IgG enhances BPI-mediated phagocytosis via engagement of FcγRII (Fig. 2C).

FIGURE 2.

The phagocytosis enhancement effect by anti–hBPI 227–254 is FcγRII mediated. (A) Recovered CFUs (number of experiments analyzed for each column from left to right; n = 3, 6, and 6) from human neutrophils infected with P. aeruginosa PAO1 (MOI 10) in the presence or absence of mouse anti-hBPI 227–254 Ab (1 μg/ml) and human heat-inactivated plasma (1%). (B) Recovered CFUs (n = 3) from human neutrophils infected with P. aeruginosa PAO1 (MOI 10) in the absence or presence of anti–227–254 mouse or rabbit IgG Ab (0.5 μg/ml or 3.33 nM) and anti–227–254 mouse F(ab')₂ Ab (0.37 μg/ml or 3.33 nM). (C) Recovered P. aeruginosa CFUs from human neutrophils pretreated with FcγRII (CD32) and/or FcγRIII (CD16) Abs, in the presence of anti–227–254 or mouse isotype control (n = 3). Data were analyzed by unpaired t test with Welch correction. Error bars represent mean ± SEM. **p < 0.01, *p < 0.05.

Anti–227–254 IgG synergistically enhances BPI-mediated bacterial clearance in vivo

The differential effects of anti-BPI IgG on P. aeruginosa phagocytosis were investigated in i.p. infection of BPI-deficient (Bpi^−/−) mice in vivo. Abs to hBPI were coadministered to BPI-deficient mice infected with P. aeruginosa with or without exogenous hBPI (Fig. 3A). In the presence of exogenous BPI, anti–61–75 Ab treatment resulted in lower bacterial clearance, compared with isotype treatment (Fig. 3B), suggesting BPI neutralization through its ability to block NTD binding to LPS (14). In contrast, treatment with anti–227–254 Ab in the presence of BPI augmented bacterial clearance in the peritoneum compared with isotype treatment (Fig. 3C). IF staining revealed that Bpi^−/− mouse neutrophils relied on the presence of exogenous BPI and anti–227–254 Ab to enhance P. aeruginosa uptake into Bpi^−/− neutrophils (Fig. 3D). Together, these data provide a model by which anti-BPI hinge IgG (227–254) engages BPI bound to P. aeruginosa and mediates uptake of this complex in an FcγRII-dependent manner (Fig. 3E).

FIGURE 3.

Phagocytosis enhancement effect seen with anti–227–254 is dependent on BPI, and anti–227–254 synergistically enhances BPI bacterial clearance effect in vivo. (A) Mice were administered 10 μg hBPI 15 min after PA14 infection (3 × 10⁶ CFUs), with anti-BPI or isotype antibodies, via i.p. route. Peritoneal lavage was performed for PA14 colony count and IHC. (B and C) Peritoneum CFUs of infected Bpi^−/− mice treated with (B) anti–61–75 (n = 3) or (C) anti–227–254 (n = 3) antibody via i.p. route, with or without exogenous hBPI. Peritoneal lavage was performed 3 (B) or 1 hpi (C) for PA14 colony count and IHC. (D) IF images of Bpi^−/− mouse peritoneal immune cells infected in vivo with tdTomato PA14 (3 × 10⁶ CFUs, 3 hpi) with (top) and without (bottom) BPI treatment, in the presence of mouse anti-hBPI 227–254 antibody, stained with DAPI (blue) and anti-hBPI antibody (red or green). Fluorescent images were obtained using a 63× oil immersion objective. Scale bars: 20 μm. (E) Model diagram depicting Fc-mediated phagocytosis enhancement effect in the presence of anti–227–254 antibody that can bind to multiple BPI molecules at a time, resulting in more P. aeruginosa being trapped and phagocytosed by human neutrophils. Data were analyzed by unpaired t test with Welch correction. Error bars represent mean ± SEM. **p < 0.01, *p < 0.05.

Discussion

The prevalence of autoantibodies to BPI in chronic diseases, such as CF, BE, COPD, and inflammatory bowel disease, and their restriction to P. aeruginosa infection are not well elucidated. In an attempt to confirm prior studies reporting autoantibody-mediated inhibition of BPI function, we created a model system of the interactions between Ab, BPI, and P. aeruginosa. To our surprise, we observed different functional effects using specific Abs to the NTD, hinge, and CTD of BPI. Most notably, anti-BPI IgGs specific to the hinge region were shown to enhance neutrophil phagocytosis of P. aeruginosa, and the enhancement effect was Fc dependent and mediated by FcγRII. Using mice deficient in BPI, we showed that anti-BPI 227–254 IgG enhances phagocytosis of P. aeruginosa in vivo by acting synergistically with exogenously provided hBPI. These findings suggest beneficial impact of autoantibodies to BPI that specifically target the BPI hinge region. In contrast, Abs to the NTD and CTD had little or no effect on P. aeruginosa phagocytosis.

The functionality of BPI and its inhibition by BPI autoantibodies have mainly been studied in vitro using Abs to BPI targeting either NTD or CTD (14, 26, 45–48). The effects of anti-BPI Abs have never been studied in vivo, nor have the functional consequences of Abs to the BPI hinge region. The observed domain-specific differences of autoantibodies on BPI function might be accounted for by two basic models: 1) enhanced ability of BPI to bind to P. aeruginosa after its interaction with BPI Abs, and/or 2) the ability of BPI Abs to bind to BPI after the BPI–P. aeruginosa interaction. Our data suggest that once BPI has formed a complex with P. aeruginosa, hinge Ab has better access to BPI than the Ab targeting the BPI NTD. This notion aligns with the fact that BPI interacts with P. aeruginosa via LPS, whose binding domains mostly reside in the N terminus (21, 49), thus inhibiting the N-terminal Ab from binding to BPI. The fact that hinge Ab can still bind to the BPI–P. aeruginosa complex implies that the hinge Ab allows the NTD of BPI to bind and neutralize LPS. A corollary of this interpretation is not only does the BPI–LPS interaction still permit the hinge Ab to bind to BPI, but it also enables Fc-dependent phagocytosis. This Fc–Ab –BPI –P. aeruginosa complex therefore leads to enhanced bacterial uptake by neutrophils (Fig. 3E, model). Moreover, the lack of inhibitory effect of hinge and CTD Abs combination to enhance phagocytosis relative to hinge Ab alone suggests that the hinge anti-BPI Ab appears to act independently from CTD to facilitate P. aeruginosa phagocytosis. Recombinant BPI lacking the hinge domain (aa 231–249) may serve as a better control for epitope specificity. However, altering the aa of BPI may lead to protein conformational change that could alter BPI function. In addition, the presence of anti-hinge BPI Ab may be required for this functional effect, because it may play a role in the formation of multimeric BPI leading to enhanced P. aeruginosa phagocytosis via an Fc-dependent manner.

The ability of the anti-hinge BPI Ab to promote more efficient phagocytosis of P. aeruginosa raises the question of the immunomodulatory effects of anti-BPI autoantibodies in CF, BE, and COPD (26–29). Autoantibodies to BPI in two cohorts of patients with CF were largely specific to the C terminus (26, 29). In transporter associated with Ag presentation–deficient patients (i.e., patients with BE, granulomatosis, vasculitis, and chronic lung infection), autoantibodies to BPI were reactive to both the C- and the N-terminal portions involved in the antimicrobial activity of BPI (50). No studies reported autoantibodies specific to the BPI hinge region, nor were they observed in our own studies (J. Theprungsirikul, unpublished observations). Because only the hinge-targeting Ab, but not the N- and C-terminal Abs, enhanced P. aeruginosa uptake by neutrophils, we can speculate that the absence of hinge-specific Abs may contribute to worse disease in patients with chronic P. aeruginosa infection (26, 29, 50). A single study previously reported that affinity-purified BPI Abs from patient sera reactive to an overlapping, but not identical, BPI peptide (aa 231–487) inhibited BPI-induced phagocytosis of E. coli in vitro (26). Although these results appear to contradict our findings with the lack of effect of BPI Ab targeting aa 248–487, they may be a result of the differences in BPI epitopes and the polyclonality of the Ab response to BPI. Furthermore, no in vivo evidence of their functional effects was explored. Conformational change of BPI in a way that inhibits BPI interaction with P. aeruginosa after CTD Ab binding to CTD could account for the results showing that CTD Ab had no effect relative to hinge Ab.

The enhanced phagocytosis mediated by the hinge-targeting BPI Abs justify further investigation of their potential to enhance BPI function in the conditions of autoantibody-mediated neutralization of BPI activities. FcγR enables neutrophils to interact with, respond to, and promote clearance of monomeric or aggregated Igs, Ag–Ab immune complexes, opsonized (Ab-coated) particles, and infected cells through phagocytic and cytotoxic mechanisms, as well as the induction of adaptive immune responses through the modulation of dendritic cell function (3, 51). Our data demonstrate the ability of BPI to effectively bind to GNB, and the interaction with BPI hinge Ab does not affect the BPI–P. aeruginosa interaction, resulting in Fc-mediated bacterial clearance. Therefore, this hinge Ab is not only predicted to be protective against P. aeruginosa but also a promising therapy molecule to enhance bacterial phagocytosis and promote bacterial clearance in patients. Studies that confirm a functional role without any drawbacks from using this Ab as a therapeutic molecule would have obvious clinical potential.

This article is featured in Top Reads, p.751

Disclosures

The authors have no financial conflicts of interest.