Regulation of Extracellular Chromatin Release from Neutrophils

Antibodies and Supplies

Anti-citrullinated histone H3 rabbit antibodies (Ab 5103) were obtained from AbCam (Cambridge, Mass., USA) and anti-total-histone H3 rabbit antibodies (07-690) from Upstate Biotechnology (Lake Placid, N.Y., USA). LPS, zymosan, apocynin and horseradish peroxidase (HRP)-conjugated secondary antibodies to rabbit or mouse Ig were purchased from Sigma (St. Louis, Mo., USA). Goat anti-rabbit AF 648, annexinV AF488 and Sytox Orange were obtained from Invitrogen (Carlsbad, Calif., USA). H₂O₂ was obtained from Fisher Scientific (Fairlawn, N.J., USA). SecinH3 was received from Merck KGaA (Darmstadt, Germany). Recombinant human TNF was supplied by Biosource (Carlsbad, Calif., USA). Lipoteichoic acid (LTA) was a kind gift from Dr. David Hasty (VA, Memphis, Tenn., USA).

Neutrophil Isolation and Treatments

Neutrophils were isolated from blood of healthy donors purchased from Lifeblood Biological Services (Memphis, Tenn., USA) and used in accord with protocols approved by the University of Tennessee institutional review board. Neutrophils were purified following modified methods of Wang et al. [19] and Sergeant et al. [20]. Briefly, neutrophils were enriched at room temperature in the supernatant from a isolymph sedimentation and recovered from an Isolymph density gradient (Gallard-Schlesinger, Plainview, N.Y., USA) under endotoxin-free conditions.

Polypropylene tubes containing neutrophils were transferred to ice, the remaining erythrocytes were lysed in ice-cold hypotonic (0.2%) NaCl for 30 s, and the solution was rendered physiological saline by addition of hypertonic (1.6%) NaCl. At this point, neutrophil viability was typically >98%, as assessed by Trypan Blue dye exclusion.

Neutrophils were suspended at a density of 2 × 10⁶/ml in PBS supplemented with 0.1% glucose and 0.5% heat-inactivated human serum, but without Ca⁺⁺ or Mg⁺⁺. Neutrophils were stimulated with increasing concentrations of TNF, LPS, zymosan, LTA or H₂O₂ in the presence of 2 mM calcium at 37°C for 2 h. Following the incubations, cells were washed with cold PBS and cell lysates were prepared in SDS-lysis buffer (2% SDS in 62.5 mM Tris, pH 6.8, supplemented with 5% 2-mercaptoethanol and 10% glycerol). Lysates were sonicated and protein concentrations were measured by using a Nanodrop 1000 spectrophotometer (Thermo Scientific, Wilmington, Del., USA).

Western Blotting

Cell lysates were analyzed as previously described [18]. Briefly, samples were separated on 12% SDS-PAGE and proteins were transferred to nitrocellulose membranes. Membranes were blocked for 1 h at room temperature with 5% BSA in TBST (25 mM Tris pH 7.2, 150 mM NaCl and 0.1% Tween) and rinsed before incubation at 4°C with primary antibodies diluted in TBST. Subsequently, membranes were washed and incubated with goat-anti-rabbit secondary antibody conjugated to HRP for 1 h at room temperature, washed 3 times with TBST and twice with TBS alone. The HRP activity was detected by using chemiluminescence reagent plus (PerkinElmer Life Sciences Inc., Boston, Mass., USA). Relative intensities of immunoreactive bands were quantitated using image analysis software (Adobe, San Jose, Calif., USA). Experiments were performed at least three separate times with consistent results.

Confocal Microscopy

Neutrophils were allowed to settle for 30 min onto glass coverslips that were precoated with 0.001% poly-L-lysine and placed in wells of 6-well tissue culture plates at 37°C in 5% CO₂. The cells were treated with LPS or alternative treatments and incubated for 1 h at 37°C. The coverslips were washed inside the wells with ice-cold HBSS, the cells were fixed with 6% PFA in HBSS for 15 min at room temperature and blocked overnight with blocking solution (HBSS with 10% FBS, 1% BSA, 0.05% Tween 20 and 2 mM EDTA) at 4°C. The coverslips were washed with wash buffer (HBSS with 3% FBS), incubated with rabbit anti-citrullinated histone H3 antibodies (diluted 1:100 in wash buffer) for 30 min at 4°C, washed again, and incubated with goat-anti-rabbit IgG coupled with AF 647, Sytox Orange, and annexin V coupled to AF 488 for 30 min at 4°C. Annexin V is a useful marker for the plasma membrane of cells treated as above with paraformaldehyde and low concentrations of Tween 20 [21, 22]. Following addition of fluorochromes, coverslips were washed, mounted on slides in wash buffer containing 50% glycerol, and analyzed by confocal microscopy, as previously described [23]. Percentages of cells releasing NETs were determined by counting between 300 and 400 cells for each of the treatments and averaging the results. Experiments were performed at least 3 separate times with consistent results.

Inflammatory Stimuli Induce Histone Deimination and NET Formation

Neutrophils are exquisitely sensitive to signals of an ongoing infection, among them cytokines and microbial breakdown products that activate neutrophils [1]. As examples of these substances, we used TNF, the quintessential proinflammatory cytokine, LPS and LTA, components of the Gram-negative and Gram-positive cell walls, respectively, and zymosan, the carbohydrate cell wall of fungi, to stimulate neutrophils in vitro. We also used hydrogen peroxide, a substance produced by the respiratory burst in neutrophils, and known to be essential for production of NETs [5]. Each of these stimuli induced robust histone deimination (fig. 1a) that was well above the background level of deimination observed in buffer alone. Higher concentrations of stimuli induced more deimination. The amount of total protein in the lysates was carefully adjusted prior to electrophoresis, and equal loading was confirmed by measuring total H3 histone (fig. 1a).

In parallel experiments, we asked whether neutrophils respond to inflammatory stimuli by production and release of NETs. Previous studies had demonstrated that treatments employed in figure 1a induce the formation of NETs [6, 10, 18]. Our previous data indicated that fewer than half of NETs react with antibodies to deiminated histone H3 [18]. We used this observation here to explore the regulation and timing of histone deimination.

In contrast to unstimulated neutrophils that have rounded cellular outlines and lobed nuclei (fig. 1b), neutrophils stimulated with hydrogen peroxide released NETs as early as 1 h after addition of the stimulus (fig. 1c). As highlighted in figure 1c, one of the neutrophils (indicated by the violet arrow) cast a NET that reacted intensely with the antideiminated histone H3 antibody (red) and the Sytox orange stain for DNA (blue). The overlap of the 2 fluorescence signals results in an intense violet color. In contrast, a second neutrophil, immediately adjacent to the first, cast a NET that failed to react with the antibody (fig. 1c, blue arrow). This result suggested that histone deimination is accomplished prior to the rupture of the plasma membrane and ceases upon NET release.

Disruption of the Cytoskeleton Inhibits Histone Deimination and NET Release

Large-scale transitions in cell morphology almost always involve contributions of the cytoskeleton. In neutrophils, cell attachment, migration and phagocytosis are among the key processes requiring the functions of the actin cytoskeleton [24]. Microtubules carry out important functions in the transport of cell organelles, including lysosomes and granules, and organelle fusion with membranes [25]. Clearly, one might predict that the release of NETs requires the participation of both actin and tubulin networks, although it is not obvious at which step the cytoskeleton provides a critical contribution. Similarly, histone deimination may require signals involving the cytoskeleton, although this has not previously been examined.

To test the involvement of microtubules and actin filaments in histone deimination and NET formation, we incubated neutrophils in nocodazole, a drug that interferes with tubulin polymerization into microtubules, or cytochalasin D, a drug that disrupts the polymerization of actin filaments. Pretreatment with nocodazole prior to LPS addition to neutrophils inhibited 53% of histone deimination relative to LPS alone (fig. 2a), and cytochalasin D inhibited 47% of the deimination observed with LPS alone (fig. 2a). Apocynin, a drug that inhibits NADPH oxidase and thus blocks production of reactive oxygen [26], was slightly more effective in blocking deimination (64% reduction relative to LPS alone). Because cytochalasin D and nocodazole substantially reduced histone deimination, we asked whether NET release would likewise be affected.

The effects of cytochalasin D and of nocodazole on NET production and release were striking. Just over 30% of cells treated with LPS in the absence of cytoskeletal inhibitors released NETs (fig. 2b), while others lost the constrained lobed structure of the nucleus and were thus poised for NET release. Nearly one third of the cells releasing NETs (32% in the experiment shown) reacted positive for deiminated histone H3 (fig. 2b).

In contrast, cells that were pretreated with nocodazole prior to addition of LPS had a largely normal polymorphonuclear granulocyte appearance, in that nearly two thirds of the cell population did not appreciably differ from controls (compare fig. 2c to fig. 1b). Remaining cells had progressed to the disintegration of the nucleus, yet overall only 5% of cells had released NETs (fig. 2e). Importantly, less than 3% of all cells showed pyknotic nuclei, indicating that few cells entered apoptosis under the conditions used in our experiments. Nocodazole treatment also greatly reduced reactivity with the antideiminated histone H3 antibody. Similarly, only 9% of cells treated with cytochalasin D progressed to NET release, and most retained a naïve neutrophil morphology (fig. 2d).

Remarkably, cells that were treated with cytochalasin D prior to incubation with LPS displayed a unique characteristic. In these cells, the nuclear envelope had disintegrated and allowed chromatin to mix with the cytoplasmic contents, yet the plasma membrane remained an effective barrier to the release of NETs (fig. 2d). Neutrophils whose NET release was restrained by cytochalasin D treatment were able to reach twice the diameter of unstimulated cells (fig. 2d, insert). It is possible that actin filaments play a role in allowing the rupture of the plasma membrane and the deployment of NETs. In addition, the actin network may function to push the chromatin through a breach in the plasma membrane. Further studies will be required to relate the disruption of the actin cytoskeleton to the increase in cell diameter.

Cell Attachment and Integrin Signaling Assist in Histone Deimination and NET Release

Because the majority of neutrophils treated with cytochalasin D did not show evidence of responding to LPS treatment, we wondered whether the cytoskeleton is required to sense and transmit external signals that culminate with NET release. Leukocyte integrins are cell surface receptors whose cytoplasmic domains engage the actin cytoskeleton [27].

To test whether integrin binding contributes toward histone deimination and NET release, we inhibited Mac-1 integrin, the complement receptor composed of the α_M/β₂ integrin subunits [28], with M1/70, an IgG antibody that inhibits Mac-1 binding [29]. Mac-1 binds a variety of ligands, including poly-L-lysine that was used to coat the coverslips for this experiment [30]. The antibody was administered 30 min before addition of LPS. We observed a 51% decrease in histone deimination compared to deimination observed in the absence of M1/70 or the presence of a control IgG (fig. 3a and c).

To test one likely pathway linking Mac-1 to the actin cytoskeleton, we inhibited cytohesin-1, the guanine nucleotide exchange factor that associates with the β₂ integrin subunit and, when phosphorylated, tightly binds to the actin cytoskeleton [31]. Recently, secinH3, a specific antagonist of cytohesin-1 function was developed [32], and we tested its effect on the induction of histone deimination by LPS (fig. 3b). SecinH3 inhibited 73% of the LPS-induced histone deimination (fig. 3b and c), whereas Wortmannin, the phosphoinositide-3-kinase inhibitor [33], had a negligible effect (fig. 3b). We interpret these results as evidence that integrin engagement and signaling mediated by cytohesin-1 facilitate LPS-induced histone deimination.

The effect of Mac-1 inhibition by M1/70 on NET release was dramatic. While cells responded to LPS treatment by rounding their multilobed nuclei, becoming reactive with the anti-deiminated histone antibody, and ultimately releasing NETs (fig. 3d), cells preincubated with M1/70 showed few of the dramatic responses to LPS (fig. 3e). Therefore, our data suggest that integrin Mac-1 engagement and signaling through the cytohesin-1 pathway are central to the inflammatory response in neutrophils in that they contribute to the regulation of histone deimination and NET release.