Filamin A Regulates Neutrophil Adhesion, Production of Reactive Oxygen Species, and Neutrophil Extracellular Trap Release

The main function of neutrophils, after being recruited into the inflamed tissue, is to eliminate pathogens. Neutrophils kill pathogens by phagocytosis through the release of reactive oxygen species (ROS) and antibacterial proteins in the phagosome, as well as by releasing secretory vesicles and granules into their surroundings. Secretory vesicles deliver β₂ integrins to the cell surface to enhance cell adhesion, whereas the released granule contents enhance cell migration and killing of the invading pathogen (1). A third mechanism used to kill invading bacteria is the release of neutrophil extracellular traps (NETs), which are composed of DNA, histones, neutrophil elastase (NE), and other proteins and enzymes that help to immobilize pathogens and facilitate their phagocytosis, as well as kill them directly (2).

β₂ integrins (CD11a/CD18, CD11b/CD18, CD11c/CD18, and CD11d/CD18) are essential for neutrophil trafficking to the site of infection, because they regulate adhesion to ICAMs expressed on endothelial cells and transendothelial migration (3). β₂ integrins are also important phagocytic receptors and mediate the binding of neutrophils to iC3b-coated bacteria in inflamed tissue (4, 5). Therefore, β₂ integrins are essential for neutrophil function in immunity, as shown by the rare genetic disorders leukocyte adhesion deficiency type I and III, which are caused by reduced or lost expression or function of these molecules (6). Patients with these disorders suffer from recurrent bacterial infections, primarily as a result of neutrophil dysfunction, because neutrophils do not express other integrins that could compensate for β₂ integrin loss (3).

Integrin functions are regulated through interactions of their cytoplasmic tails (CTs) with other cytoplasmic adapter and signaling molecules (7, 8). The β₂-chain has been shown to interact with numerous intracellular molecules, with the most important being talin, kindlin-3, filamin A (FlnA), and the 14-3-3 proteins (9–12). Factors dictating which molecules are able to bind to integrin tails and, thus, which signals are transmitted to/from integrins are the overall conformation of the α-β dimer and the phosphorylation of integrin cytoplasmic domains.

FlnA is a large (280-kDa) cytoplasmic protein that consists of 24 Ig domains and a C-terminal actin-binding domain. It cross-links actin in an orthogonal network and participates in actin cytoskeleton reorganization through its interactions with actin and a number of other cytoplasmic and membrane proteins, including integrins (13, 14). FlnA is required for proper osteoclastogenesis, because it regulates monocyte migration through Rho GTPases that regulate actin dynamics (15). FlnA is also required for cell spreading, uropod retraction, and preserving the cytoskeletal integrity in various cell types (15–19). FlnA can also function as a mechanosensor, detecting and transmitting mechanical stimuli from the extracellular matrix to the cell, and it has been shown to contribute to the mechanical stability of the plasma membrane and cell cortex (20). FlnA has been shown to interact with the CTs of β₁, β₂, β_3, and β₇, and, according to a recent study, the α_IIb integrin chains, and to function as an inhibitor of integrin activation and of integrin-dependent cell adhesion (10, 19, 21–24). FlnA has been shown to bind to the β₂ integrin tail only when the T⁷⁵⁸TT motif is unphosphorylated, whereas 14-3-3ζ binds only to the phosphorylated TTT and may outcompete talin from the β₂ integrin CT (10).

Although the effects of FlnA depletion on actin dynamics and cell migration have been studied in different cell types, the functions of FlnA in other cellular processes, such as the killing mechanisms of neutrophils, has remained less well understood. To investigate the function of FlnA in the regulation of neutrophil β₂ integrin–dependent adhesion and other cellular mechanisms, we studied a mouse model in which FlnA is lacking in neutrophils. In this study, we demonstrate that FlnA is a negative regulator of β₂ integrin–dependent neutrophil adhesion and neutrophil ROS production and that FlnA depletion leads to reduced NET release. These results indicate an important role for FlnA in the regulation of integrin adhesion and other essential neutrophil functions.

Mice with targeted deletion of the X-linked Flna gene were created by crossing mice carrying the myeloid cell–specific Cre recombinase (B6.129P2-Lyz2^tm1(cre)Ifo/J) (LysMCre mice) (25) with mice having a conditional-knockout (KO) allele of the Flna gene (Flna^tm1.1Caw/J) (FlnA^LoxP/LoxP mice) (both from The Jackson Laboratory). LysMCre and/or FlnA^LoxP/y mice were used as controls. Mice were bred and maintained at the University of Helsinki, and all animal procedures were done in compliance with the Finnish Act on Animal Experimentation.

Neutrophils were purified from control and LysMCre/FlnA^LoxP/y mouse tibia and femur bone marrow by negative selection using a MACS mouse neutrophil purification kit (Miltenyi Biotec), according to the manufacturer’s instructions. Cell purity was verified by flow cytometry and was typically >85%. Neutrophils were kept cold/on ice during the purification, and they were used for experiments on the day of isolation. The presence of FlnA in the neutrophil lysates was detected from cells lysed directly in SDS-PAGE sample buffer by Western blot with Abs against Filamin A (Bethyl Laboratories). β-Actin (Cell Signaling Technology) was used as a loading control after stripping the primary Ab with 0.2 M glycine and 0.1% SDS (pH 2.2).

Escherichia coli containing GFP (strain BL-21) used in phagocytosis assays was a kind gift from S. Taira (University of Helsinki). In other assays, E. coli DH5α was used. E. coli was cultured in Luria–Bertani medium (LB) and LB plates containing ampicillin, and GFP production was induced with IPTG where necessary.

Staphylococcus aureus S11 strain (ATCC 12598) was from HAMBI collection (University of Helsinki) and was cultured in tryptic soy broth and on tryptic soy broth–agar plates. When appropriate, S. aureus was labeled with FITC by treating the overnight-cultured bacteria with 100 μg/ml FITC (isomer I; Sigma-Aldrich, St. Louis, MO) in PBS for 30 min under constant stirring at room temperature. Bacteria were then washed three times with PBS and adjusted to an OD₆₀₀ of 1.7, corresponding to a bacterial density of 1 × 10⁹ per milliliter, as assessed by using a Bürker counting chamber. FITC-labeled bacteria were freshly prepared before each experiment.

For opsonization, bacteria (1 × 10⁹ per milliliter) were washed once in opsonization buffer (PBS + 5% FBS + 0.1% glucose), incubated in the presence 10% C5-deficient serum (Sigma-Aldrich) or 20% mouse serum (for ROS assays) for 15 min at 37°C under constant end-to-end shaking, washed once with opsonization buffer, and suspended in 10–20% serum in the same buffer. A bacteria/neutrophil ratio of 45 (multiplicity of infection) was used in all experiments.

Cells were stained with the following fluorescently labeled Abs (Ab clones are in parentheses): B220-PE (RA3-6B2), CD3–Alexa Fluor 488 (145-2C11), CD11a-PE (2D7; BioLegend), CD11b-allophycocyanin (M1/70), CD11c–PE–Cy7 (N418), CD18-FITC (C71/16; BD Biosciences), F4/80–allophycocyanin–Cy7 (BM8), and Gr-1–PerCP–Cy5.5 (RB6-8C5). Fc block (anti-mouse CD16/CD32) was included in all stains. Abs were from eBioscience unless indicated otherwise. For investigation of CD11b/CD18 delivery to the cell surface, the cells were activated with LPS (Sigma-Aldrich) or PMA (Calbiochem/Merck Millipore) for 30 min prior to staining. The flow cytometry analysis was performed at the Flow Cytometry Core Facility in the Department of Biosciences, University of Helsinki, and analyzed using FlowJo software (TreeStar).

Purified neutrophils were left untreated or were cultured in RPMI 1640 medium supplemented with 10% FBS, 2 mM l-glutamine, and 100 U/ml penicillin-streptomycin in the presence of GM-CSF (10 ng/ml; PeproTech), PMA (100 nM), and 5 mM Mg + 1 mM EGTA. After 46 h of culture, the cells were stained with propidium iodide (Sigma-Aldrich), and the percentage of viable cells was detected by flow cytometry.

Coverslips were coated with 6 μg/ml ICAM-1 (Bio-Techne) or iC3b (Calbiochem). Phorbol 12,13-dibutyrate (PDBu)-activated neutrophils were seeded onto coated cover slips and left to adhere in whole media for 30 min before fixation with 4% paraformaldehyde. F-Actin was stained with 1:20 TRITC-phalloidin in 0.1% saponin/1% FCS/PBS (both from Sigma-Aldrich) for 1 h. Slides were imaged using a Leica SP5 II confocal microscope (Leica Microsystems) with LAS AF Lite Software and 561 laser (10% laser power). Z stacks were taken with the following parameters: spectral range: 570–779 nm, max QD405/488/561/635 mirror, Smart Gain 800 V, Smart Offset 0.0%, pinhole 111.49 μm, zoom 1.00, objective 63×, z distance 8.003 μm, 55 steps, 0.15 μm step size, line average 4, and format 512 × 512. Area, fluorescence intensity, and background were measured with ImageJ and used to calculate corrected total cell fluorescence (CTCF) values as previously described (26). A total of 25–100 cells per condition was measured for each animal.

Static adhesion assays were performed as previously described (27). Briefly, the integrin ligand ICAM-1 or iC3b (each 6 μg/ml) was coated onto 96-well MaxiSorp plates (Nunc/Thermo Fisher Scientific) by overnight incubation at 4°C. Purified neutrophils (1.5 × 10⁶ cells per milliliter) were resuspended in adhesion medium (RPMI supplemented with 0.1% BSA, 40 mM HEPES, and 2 mM MgCl₂) and added to the plate. Cells were allowed to adhere for 15 min at 37°C before gentle washing to remove unbound cells. Bound cells were lysed and detected with phosphatase substrate p-nitrophenyl phosphate (Sigma-Aldrich).

Neutrophil-adhesion assays under shear flow were performed essentially as described (28). Briefly, ibidi VI^0.4 μ-Slides (ibidi) were coated overnight at 4°C with 6 μg/ml ICAM-1, 5 μg/ml CXCL-1 (Bio-Techne), and 30 μg/ml E-selectin (Bio-Techne) or with 6 μg/ml ICAM-1 alone. Neutrophils (0.7 × 10⁶ cells per milliliter) were treated with 200 nM PDBu (Sigma-Aldrich) or were left untreated and were allowed to flow over the coated slides with continuous shear flow of 0.3 dyne/cm² (for ICAM-1–coated slides) or 3 dyne/cm² (for ICAM-1+E-selectin+CXCL-1–coated slides) over a 5-min period. Cells were monitored by microscopy (Hamamatsu ORCA-Flash2.8), and the number of adhered cells in the field of view was determined by manual counting.

Purified neutrophils were incubated with opsonized or nonopsonized GFP-expressing E. coli or FITC-labeled S. aureus for 1 h at 37°C, washed and the percentage of neutrophils that had phagocytosed bacteria was measured with flow cytometry. Absence of extracellular bacteria was verified by killing with 100 μg/ml gentamicin and/or by microscopy.

Purified neutrophils were incubated with opsonized E. coli or S. aureus for 10 min at 37°C to allow the initiation of phagocytosis, the nonphagocytosed bacteria were killed with 100 μg/ml gentamicin (time point 0 min), and the neutrophils were allowed to kill bacteria for 15 or 30 min. Neutrophils were transferred on ice, washed with ice-cold opsonization buffer, and lysed with 0.1% Triton X-100 in LB. The amount of remaining living bacteria was determined by plating serial dilutions and counting colonies after overnight incubation at 37°C. Survival percentage (number of colonies at 15 or 30 min divided by the number of colonies at 0 min) was calculated, and the results are expressed as survival percentage of FlnA-KO neutrophils relative to that of control cells.

Purified neutrophils were treated with 5-(and-6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate (CM-H2DCFDA; Thermo Fisher Scientific), according to the manufacturer’s instructions. Cells were activated with 100 ng/ml PMA or opsonized E. coli or S. aureus immediately prior to analysis. ROS production was analyzed by flow cytometry on the FITC channel. Samples were measured every 10 min for 70 min.

To study NE production, purified neutrophils were suspended in 1% BSA/RPMI 1640, seeded onto a 96-well plate (1 × 10⁶ cells per well), and activated with LPS (10 μg/ml) or PMA (100 ng/ml) in the presence of 10 ng/ml GM-CSF for 4 h. Elastase activity was detected in cell supernatants with a Neutrophil Elastase Activity Assay Kit (Cayman Chemical), according to the manufacturer’s instructions.

For the analyses of released DNA, purified neutrophils were suspended in RPMI 1640 + 2% FBS including 100 ng/ml GM-CSF (1 × 10⁶ cells per milliliter), aliquoted on a black 96-well plate (OptiPlate; PerkinElmer), and activated with LPS (10 μg/ml) or PMA (100 nM) or incubated with opsonized or nonopsonized E. coli or S. aureus. SYTOX Green Nucleic Acid Stain (Thermo Fisher Scientific) was added to a final concentration of 2.5 μM, and fluorescence in the supernatant was measured after 2–4 h using an EnSpire fluorometer (PerkinElmer).

For NET imaging, 2 x 10⁶ cells in 1 ml of RPMI + 2% FBS containing GM-CSF neutrophils were seeded onto cover slips coated with poly-d-lysine hydrobromide (10 μg/ml; Sigma-Aldrich) and left to adhere for 4 h. Neutrophils were left untreated or were activated with PMA or opsonized bacteria. After 4 h, cells were incubated with SYTOX Green (1 μM) for 10 min. One milliliter of 4% paraformaldehyde was added directly to the cell media. Histone H3 (1:2000; product code ab176842; Abcam) and elastase (1:200; catalog number bs-6982R; Bioss) were stained on nonpermeabilized cells.

Slides were imaged using a Leica SP5 II confocal microscope (Leica Microsystems) with LAS AF Lite Software, with a 488 laser (20% laser power) and a 561 laser (10% laser power). Z stacks were taken with the following parameters: QD405/488/561/635 mirror, pinhole 60.66 μm, zoom 1.00; objective 20×, z distance 10.078 μm, 0.3 μm step size, line average 2, and format 512 × 512. For SYTOX Green, the following parameters were used: spectral range 498–666 nm, Smart Gain 600 V, Smart Offset −1.0%; 561 (H3 and elastase) 599–763 nm, Smart Gain 900 V, and Smart Offset 0.0%.

Data are presented as mean ± SEM. The Student t test (GraphPad) and two-way ANOVA were used for statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001.

To investigate the role of FlnA in integrin regulation and neutrophil functions, we generated a myeloid-specific FlnA-null LysMCre/FlnA^LoxP/y mouse to induce specific deletion of FlnA in myeloid cells (such as neutrophils). FlnA-KO neutrophils show significantly reduced levels of FlnA compared with control cells (Fig. 1A). However, there is a residual band of FlnA in the neutrophils purified from the bone marrow of LysMCre/FlnA^LoxP/y mice. Incomplete gene deletion in neutrophils is likely to contribute to the residual band in FlnA-KO neutrophils, because the efficacy of deletion in LysMCre mice has been reported to be 75% in bone marrow–derived neutrophils (25). In addition, imperfect purification of neutrophils and, thus, other cell types present in the neutrophil preparation may contribute to the presence of the residual FlnA band (the purity of neutrophils isolated from bone marrow by magnetic labeling was 85–95%).

LysMCre/FlnA^LoxP/y mice appear to have relatively normal immune cell development, as seen in flow cytometry analysis of total cell numbers (Fig. 1B) and proportion of Gr-1⁺ cells in lymphoid tissues (Fig. 1C). The proportions of other myeloid and lymphoid cells were also normal in blood, bone marrow, and secondary lymphoid organs (Fig. 1D). The survival of FlnA-depleted neutrophils ex vivo is normal (Fig. 1E). These results show that FlnA is not needed for LysM⁺ immune cell development or immune cell distribution in vivo or neutrophil survival ex vivo.

Because FlnA is an important actin cross-linking protein and also regulates cell size (13, 16), the effects of FlnA depletion on neutrophil cell size and F-actin content were analyzed. The cells were allowed to spread on integrin ligand ICAM-1 or iC3b and were stained with phalloidin, and their size and actin content were analyzed by microscopy. There was a small, but significant, increase in cell size in FlnA-null neutrophils on integrin ligands compared with control neutrophils (Fig. 2A). However, the FlnA-depleted neutrophils had normal levels of F-actin in proportion to their size, expressed as CTCF of phalloidin-bound TRITC (Fig. 2B, 2C). Therefore, although actin content appears normal in FlnA-depleted neutrophils, cells display increased spreading on integrin ligands.

FlnA has been shown to bind to β₂ integrin cytoplasmic domains and regulate their functions (10, 21), but the role of FlnA in β₂ integrin regulation in primary cells is unknown. Therefore, β₂ integrin expression and function were studied in FlnA-null neutrophils purified from the bone marrow of LysMCre/FlnA^LoxP/y mice. The expression of β₂ integrins (CD11a, CD11b, CD11c, and CD18) was normal in FlnA-null neutrophils (Fig. 3A). One feature of neutrophil activation is the transport of β₂ integrin–containing secretory vesicles to the cell surface to rapidly enhance adhesion (29). However, there was no difference in the PMA-induced surface expression of CD11b/CD18 between FlnA-KO and control neutrophils (Fig. 3B), indicating that FlnA does not regulate integrin surface expression in neutrophils.

To study whether FlnA regulates β₂ integrin–dependent functions in mouse neutrophils, cell adhesion to coated integrin ligands under static conditions was studied. Interestingly, our results show that FlnA-null nontreated neutrophils show higher adhesion to β₂ integrin ligands ICAM-1 and iC3b than control cells (Fig. 3C). Together with the cell spreading data (Fig. 2A), these results indicate that FlnA acts as a negative regulator of β₂ integrin–mediated adhesion and spreading in murine neutrophils.

β₂ integrins and their interaction with the integrin-regulator kindlin-3 are essential for leukocyte adhesion under shear-flow conditions, such as those found in blood flow (28, 30). To examine whether deletion of the integrin interactor FlnA had an effect on neutrophil adhesion under flow conditions, shear-flow adhesion assays were conducted using FlnA-null neutrophils. However, phorbol ester–stimulated adhesion to ICAM-1 alone under low shear-flow conditions and chemokine-stimulated adhesion to cocoated ICAM-1 and E-selectin under high shear-flow conditions were not affected by FlnA depletion in neutrophils (Fig. 3D, 3E). Together, these results indicate that, although FlnA negatively regulates β₂ integrins under static conditions, it is dispensable for β₂ integrin–mediated neutrophil adhesion under shear-flow conditions.

One of the main functions of neutrophils is the phagocytosis of pathogens, a process that is dependent on integrins and the actin cytoskeleton. We studied the phagocytosis of opsonized and nonopsonized E. coli (Gram-negative) and S. aureus (Gram-positive) bacteria by flow cytometry. Surprisingly, we did not observe differences in phagocytosis between control and FlnA-depleted neutrophils (Fig. 4A), showing that, although adhesion to iC3b is enhanced in FlnA-KO neutrophils, phagocytosis is not affected. In addition, killing of the engulfed pathogens is not affected in FlnA-KO neutrophils (Fig. 4B). These results show that, although major roles for filamin as an actin cross-linker in cells have been described, it is dispensable for neutrophil-mediated phagocytosis and killing of bacteria.

After entry of neutrophils into tissues, the intracellular release of ROS is an important mechanism that these cells use to destroy engulfed material like bacteria (1). To study the role of FlnA in the production of ROS in neutrophils, we used a cell-permeable dye, CM-H2DCFDA, which becomes fluorescent upon ROS production in cells and can be quantified using flow cytometry. Interestingly, upon PMA activation, FlnA-null neutrophils exhibited increased ROS production (Fig. 5A), showing that FlnA is a negative regulator of ROS production in neutrophils.

Although PMA-induced ROS production is integrin independent, ROS production upon bacterial activation in β₂ integrin (CD18)^−/− and kindlin-3^−/− neutrophils has been previously shown to be defective (11, 31, 32). Therefore, to investigate ROS production in a more physiological integrin-dependent setting, we next studied whether ROS production upon bacterial activation was affected in FlnA-KO neutrophils. In line with the lack of effect of FlnA depletion on phagocytosis (above), ROS production in response to bacterial phagocytosis was not affected in FlnA-KO neutrophils (Fig. 5B, 5C). In conclusion, ROS production in response to PMA, but not in response to bacterial stimulation, is increased in FlnA-KO neutrophils.

In addition to ROS production, an important microbicidal neutrophil function is the release of NETs. NETs consist of DNA, histones, and antimicrobial proteins like NE and cathepsin G, as well as several other proteins that are released by neutrophils to help trap and kill bacteria (2). NET release was imaged by microscopy of neutrophils that were stained for DNA (Fig. 6F), NE, and histone H3 (Supplemental Fig. 1). Quantification of NETs was performed through two independent experimental set-ups: by studying NE activity in cell supernatants and by measuring DNA release from the activated neutrophils using SYTOX Green, a fluorescent DNA-binding dye that cannot enter living cells. We have previously shown that the production of NE is increased in neutrophils expressing dysfunctional β₂ integrins (33), and we show in this article that neutrophil Mg²⁺/EGTA treatment, which activates integrins, reduces NE activity (Fig. 6A). These results indicate that functional activated β₂ integrins negatively regulate NET production in neutrophils. Because FlnA negatively regulates β₂ integrin function, we studied the effect of depleting FlnA on NET production. Interestingly, LPS-stimulated FlnA-KO neutrophils display reduced NE activity compared with control cells (Fig. 6B). In addition, FlnA-KO cells showed lower amounts of DNA in the cell supernatant after activation with LPS (Fig. 6C) and PMA (Fig. 6D), confirming that FlnA is needed for NET release. In accordance with these results, FlnA-KO neutrophils produced significantly lower amounts of NETs than did wild-type cells after incubation with E. coli (Fig. 6E). However, FlnA depletion did not significantly alter NET levels upon activation with Gram-positive S. aureus. Altogether, the results show that FlnA is required for efficient neutrophil NET production in murine neutrophils and that this regulation may be controlled through regulation of integrin activity.

FlnA is an essential cytoskeletal protein that cross-links actin in a rigorous network and connects the cytoskeleton to the cell surroundings through its interactions with transmembrane adhesion molecules and receptors (1, 2). Because of the important role of FlnA in cellular functions, FlnA-KO mice display embryonic lethality due to inefficient cell migration during cardiac, skeletal, and vascular development (34, 35). However, the in vivo role of FlnA in the development and function of myeloid immune cells remains incompletely understood. FlnA-deficient neutrophils have been shown to have a defect in uropod retraction during chemotaxis, resulting in a defect in neutrophil recruitment to peritoneum after sodium periodate injection in vivo (19), but their bactericidal neutrophil functions have not been assessed. In this article, we show that immune cell numbers and localization in lymphoid tissues appear normal in the LysMCre/FlnA^LoxP/y mouse. In addition, Gr-1⁺ neutrophils are found in normal numbers in bone marrow, lymph nodes, spleen, and peripheral blood of these animals, and they display normal survival in ex vivo survival assays, indicating that FlnA is not required for normal neutrophil development in vivo.

When analyzing the FlnA-depleted neutrophils further, we noticed the enlarged size of the cell spread on β₂ integrin ligands, in contrast to the reports from FlnA-depleted platelets on collagen and fibroblasts on fibronectin, which show severe defects in spreading (16, 17, 36). Furthermore, FlnA-deficient neutrophils seem to have normal levels of F-actin and relatively normal actin distribution compared with FlnA-null platelets and fibroblasts, for which the cytoskeletal integrity appears to be seriously compromised (16, 17, 36). These results show that FlnA is not needed for cell spreading on β₂ integrin ligands in neutrophils. This is in line with previous results showing that neutrophils lacking FlnA are able to polarize and extend pseudopods normally (19), whereas fibroblasts lacking FlnA show severe defects in these early stages of cell migration (37).

FlnA has been shown to be a negative regulator of β₂ integrin adhesion in the Jurkat T cell line using a small interfering RNA approach (11), but its role in regulating β₂ integrins in primary leukocytes has remained poorly understood. In this article, we show that FlnA indeed functions as a negative regulator of β₂ integrin–dependent cell adhesion under shear-free conditions in primary murine neutrophils. The depletion of FlnA in neutrophils increases β₂ integrin–mediated adhesion to the integrin ligands iC3b and ICAM-1. The increased size of FlnA-null cell spread on β₂ integrin ligands further supports the conclusion that FlnA negatively regulates β₂ integrin–mediated neutrophil adhesion in the absence of shear force. Alternatively, the enlarged cell size may reflect disruption of actin cytoskeleton integrity in these cells, although F-actin levels appeared to be normal. Phagocytosis of bacteria, a process that is dependent on the actin cytoskeleton, was not affected in FlnA-null neutrophils, further suggesting that adhesion-induced actin cytoskeleton reorganization in neutrophils functions relatively normally in the absence of filamin and indicating that other actin cross-linking proteins may instead be important in regulating these processes in primary neutrophils.

Leukocytes in the blood stream are subjected to shear forces, and tensile forces acting on β₂ integrin–ligand bonds have been reported to regulate β₂ integrin–mediated adhesion and downstream functions through mechanotransduction (38, 39). Filamin is an important mechanosensor in cells. For example, in platelets, depleting FlnA disrupts platelet adhesion under arterial flow conditions (16). However, when shear flow adhesion assays were conducted with the FlnA-null neutrophils, no difference between control and FlnA-KO neutrophils could be observed. These results indicate that FlnA does not negatively regulate β₂ integrins under these conditions but also is not required for neutrophil integrin-mediated adhesion under shear flow conditions, in contrast to the situation in platelets. It is possible that FlnA is also required for adhesion strengthening under shear flow in leukocytes, as it is in platelets, but because integrin-mediated adhesion is increased in these cells, this may compensate for a possible defect in the ability of FlnA-null neutrophils to withstand externally applied forces. These findings and our results suggest that the role of FlnA in the regulation of cell adhesion under shear stress is different in platelets and in leukocytes. Indeed, platelet adhesion is not mediated through β₂ integrins, but through GPIbα, possibly explaining the differences in the observed adhesion profiles in FlnA-depleted platelets and leukocytes.

Integrin-mediated phagocytosis has been shown to be required for the oxidative burst in response to E. coli and S. aureus in neutrophils (32). Phagocytosis was normal in FlnA-KO neutrophils (Fig. 4), leading to normal ROS production downstream of phagocytosis of E. coli and S. aureus (Fig. 5B, 5C). In contrast, PMA-induced ROS production in nonadherent neutrophils, which is integrin independent, was increased in FlnA-KO neutrophils. These results indicate that the PMA-induced increase in ROS production in FlnA-KO neutrophils does not result from the activation of β₂ integrins; rather, it may be associated with alterations in the actin cytoskeleton in these cells. Indeed, parts of the ROS-production machinery (the NOX complex) are tightly associated with the cytoskeleton in neutrophils, and they are redistributed upon cell activation to the cortical cytoskeleton (40). Alterations in the actin framework due to deletion of FlnA may cause changes in NOX complex assembly and, further, changes in ROS production in PMA-activated neutrophils.

In contrast to ROS production, the release of NETs was decreased in FlnA-null neutrophils upon activation with PMA, LPS, or opsonized or nonopsonized E. coli. We have previously shown that NE production is increased by expression of nonfunctional β₂ integrins in neutrophils (33), suggesting that activated β₂ integrins may serve as negative regulators of NE production. Indeed, we show in this article that, like FlnA depletion, activation of integrins externally with Mg²⁺/EGTA also significantly decreased NE release. These results suggest that the regulation of NET release by FlnA may take place through enhancing the immunosuppression exerted by β₂ integrins. We (28, 41) and other investigators (42) previously reported that β₂ integrins in myeloid cells, such as dendritic cells and macrophages, as well as neutrophils, may serve as immunosuppressors; however, the role of FlnA in the immunoregulation through β₂ integrins has not been previously addressed. However, it is also possible that FlnA plays a more direct role in NET formation, because actin and FlnA can be found in the nucleus (43, 44), and NETs have also been shown to contain actin (45). Indeed, the actin cytoskeleton is important for NET dynamics (46) and, in particular, actin degradation by NE has been shown to promote NE delivery to the nucleus prior to NET formation (47). Because filamin is an actin cross-linking protein, several processes related to NET release may be affected by deletion of filamin in neutrophils. Clearly, these hypotheses require further mechanistic studies.

Altogether, our results show that FlnA does not affect myeloid immune cell development or distribution or integrin expression in mice in which FlnA has been conditionally depleted in LysM⁺ myeloid cells. Surprisingly, FlnA is not required for β₂ integrin–mediated adhesion under shear flow conditions, for phagocytosis of bacteria, or for bacterially induced ROS production. However, FlnA negatively regulates β₂ integrin–mediated cell adhesion in shear-free conditions and PMA-induced oxidative burst, and it is required for proper formation of NETs in primary murine neutrophils. Therefore, our current findings provide novel evidence for an important role for FlnA in the regulation of β₂ integrin adhesion, ROS production, and NET formation in primary murine neutrophils.

The authors have no financial conflicts of interest.