Neutrophil-Derived Cytosolic PLA2α Contributes to Bacterial-Induced Neutrophil Transepithelial Migration

This article is featured in In This Issue, p.2609

Prostacyclins, PGs, leukotrienes, lipoxins, and hepoxilins are inflammatory-modulating lipid mediators that are collectively termed eicosanoids. Eicosanoids are the product of oxygenation of arachidonic acid and play a significant role in multiple pathological processes, including atherosclerosis, asthma, and rheumatoid arthritis (1–3). Pharmacological targeting of eicosanoid production is a well-established therapeutic tool. Aspirin, which inhibits the enzyme responsible for generating PGs, cyclooxygenase (COX)-2, is a treatment in atherosclerosis (4). Zileuton and montelukast are therapeutic agents that are commonly used to treat allergies and asthma by targeting leukotriene synthesis (5, 6). Eicosanoids also play a key role in neutrophilic diseases, such as sepsis (7) and pneumonia (8, 9), but precise pharmacologic targeting of these pathways has not been adopted as clinical therapies. Gaining a better understanding of the mechanisms that underlie these processes may provide more focused therapeutic targets.

The generation of eicosanoids is regulated by phospholipase A2 (PLA2), a family of enzymes that is responsible for releasing arachidonic acid from cellular membranes to serve as substrate for eicosanoid synthesis. There are several isoform classes of the PLA2 family of enzymes: secretory PLA2, cytosolic PLA2 (cPLA2), calcium-independent PLA2 (iPLA2), lysosomal PLA2, platelet-activating factor acetylhydrolase, and bacterial PLA2; each has a unique physiologic role and eicosanoid-generating capacity (10). There is emerging evidence that certain PLA2 isoforms associate with specific eicosanoid-generating enzymes, establishing selectivity toward eicosanoid species generation under certain physiological contexts (10). For example, cPLA2α governs proximity to 5-lipoxygenase (5-LOX) and 5-LOX–activating protein (FLAP), resulting in leukotriene B4 (LTB4) production (11), whereas, in other contexts, cPLA2α will colocalize with COX-1 and COX-2, leading to efficient generation of PGs (12–14). To stimulate arachidonic acid release from membranes, cPLA2α translocates from the cytosol to the intracellular membrane, in response to calcium, where cPLA2α is phosphorylated by MAPKs, enhancing its catalytic activity (15). Once activated, cPLA2α liberates arachidonic acid from membrane phospholipids, which triggers downstream synthetic enzymes to oxygenate arachidonic acid, generating a variety of eicosanoids.

Multiple eicosanoids have been shown to play key roles in neutrophil recruitment to the airway (16–19). Neutrophils migrate from circulation, across the endothelium, interstitium, and epithelial layer to reach infectious insults within the airway. In certain diseases, like acute respiratory distress syndrome or cystic fibrosis (CF), this neutrophilic inflammatory response is excessive, contributing directly to airway damage (20, 21). Epithelial cells and neutrophils rely on eicosanoid signaling to mediate and coordinate the transepithelial passage of neutrophils in response to pathogenic infection of the epithelium. Pulmonary epithelial cells infected with Pseudomonas aeruginosa release the 12-lipoxygenase–generated eicosanoid, hepoxilin A3 (HxA3), in an apically directed manner, generating a chemotactic gradient that drives neutrophils from the interstitium across the epithelium and into the airway (9). We previously assessed the role of cPLA2α in epithelial production of eicosanoids and demonstrated that infection of lung epithelial cells with P. aeruginosa elicits the generation of multiple eicosanoids, including PGE₂ and HxA3 (10). Although cPLA2α drives epithelial production of PGE₂ in response to infection, epithelial cell–derived cPLA2α did not appear to exert any influence on HxA3 release or neutrophil transepithelial migration (16, 22).

After migrated neutrophils reach the airspace in response to HxA3, they produce a second eicosanoid, LTB4, which augments neutrophil transepithelial migration (19). LTB4 is generated by oxidation of arachidonic acid by 5-LOX in association with FLAP (23). cPLA2α has been demonstrated to play a direct role in LTB4 production (11), because cPLA2α-derived arachidonic acid affects the conformation of FLAP to enhance its ability to bind 5-LOX, favoring the production of LTB4. cPLA2α is prominently expressed in lung epithelium, endothelium, fibroblasts, macrophages, and neutrophils (24, 25).

Because cPLA2α exhibits selective eicosanoid-generating capacity favoring PGE₂ and LTB4, but not HxA3, synthesis, we sought to investigate whether migrated neutrophil LTB4 synthesis is dependent on the cPLA2α isoform. Understanding inflammatory mechanisms of different PLA2 isoforms in the context of neutrophilic responses is critical to identify appropriate therapeutic targets and may provide important clinical insight. Using a range of models of neutrophil transepithelial migration, we sought to determine whether cPLA2α expressed in neutrophils contributes to bacterial-induced neutrophil transepithelial migration.

The H292 cell line is a human pulmonary epithelial cell line derived from a patient with mucoid pulmonary carcinoma. MLE12 cells are a murine epithelial cell line. H292 and MLE12 cells were maintained in DMEM/F12 (1:1) culture medium with 10% heat-inactivated FBS and antibiotics. Both cell lines were seeded onto inverted 24- or 96-well 3-μm Transwell inserts (Corning Life Sciences) until adherent (4 h) and then were reinverted into a Transwell receiver plate. H292 monolayers were maintained in culture medium for >7 d to ensure mature monolayers for migration assays. MLE12 monolayers were maintained in culture medium for 3–5 d prior to migration assays.

Human airway basal stem cells were isolated from discarded human airways harvested from the New England Organ Bank or Department of Pathology (Massachusetts General Hospital) under an Institutional Review Board–approved protocol (number 2010P001354), as previously described (26). The trachea and mainstem bronchi were dissected and cleared of connective tissues and blood cells, and epCAM⁺ epithelial stem cells were isolated. These cells were cultured in complete SAGM (catalog number CC-3118; Lonza) on plates precoated with laminin-enriched 804G-conditioned medium. Growth on 3-μm Transwells, required for neutrophil transepithelial assays, involved a two-step coating process: the side of the membrane for cell seeding was precoated with 804G medium only, and the opposite side was precoated with 804G medium containing 5% Matrigel (product number 354230; BD Biosciences). To seed the airway basal stem cells, the Transwell inserts were removed and flipped, and airway basal stem cells, with a seeding density > 6000 cells per square millimeter, were applied on the surface of flipped membranes for 6 h. Inserts were then returned to their original orientation. The growth medium (complete SAGM) was added to the upper and lower chambers for 1–2 d to ensure that the cells were confluent before replacing with complete PneumaCult-ALI medium (catalog number 05001; STEMCELL Technologies) (27) in the upper and lower chambers and cultured for another day. The following day, ALI medium was added only in the top chamber to initiate an airway–liquid interface (ALI) on the down-facing surface of the Transwell (count as day 0). The medium was changed every 1–2 d until differentiation was well established. Ciliogenesis was monitored by inverted-phase microscopy and could be observed after 11–15 d. ALIs used in experiments were cultured for 14–26 d to allow full maturation. Transepithelial electrical resistance was assessed using a voltmeter prior to migration assays to ensure the establishment of a polarized epithelial barrier (EVOM2 Epithelial Voltohmmeter; World Precision Instruments).

Neutrophils were isolated from healthy volunteers under an Institutional Review Board–approved protocol (number 1999P007782) at the Massachusetts General Hospital using an established isolation technique (28). Briefly, blood was drawn by venipuncture into a syringe containing anticoagulant, acid citrate/dextrose, and centrifuged at room temperature at 2000 rpm for 20 min without brake to allow layering of the buffy coat. Plasma and mononuclear cells were removed by aspiration. A 2% gelatin sedimentation technique, followed by wash with RBC lysis buffer, allowed for removal of RBCs. Cells were washed and resuspended in HBSS without calcium or magnesium (HBSS−) and suspended at a concentration of 5 × 10⁷ neutrophils per milliliter. This neutrophil-isolation technique allows for isolation of functionally active neutrophils (>98%) at 90% purity (29).

Escherichia coli (MC1000) and P. aeruginosa (PAO1) were grown aerobically in Luria-Bertani broth overnight at 37°C in a shaking incubator. Prior to the experiments, each bacterial suspension was washed and resuspended in HBSS to a concentration of 6 × 10⁷ CFU/ml.

Assays were adapted from the neutrophil transepithelial migration model (28). In brief, epithelial monolayers were grown on the underside of Transwell filters with 3 μm pores to accommodate neutrophil passage. This model allows for neutrophil migration from the basolateral to the apical compartment, toward chemoattractant gradients that are exogenously provided or endogenously generated by epithelial infection with the P. aeruginosa strain PAO1, triggering epithelial secretion of HxA3. To interfere with epithelial signaling pathways, epithelial cells were treated for 2 h with specific chemical inhibitors, washed, and infected with 25 μl of 6 × 10⁷ CFU/ml bacteria (MC1000 or PAO1) for 1 h prior to undergoing neutrophil transepithelial migration. Additional uninfected controls were placed in wells containing chemoattractants to create an exogenous apical chemotactic gradient with fMLF (100 nM; Sigma), IL-8 (100 ng/ml; eBioscience, San Diego, CA), or LTB4 (5 ng/ml; Enzo Life Sciences, Farmingdale, NY), capable of driving neutrophil transepithelial migration. To interfere with neutrophil signaling pathways, neutrophils were incubated in chemical inhibitors for 30–60 min at 37°C prior to placement in the basolateral compartment at a concentration of 2 × 10⁵ neutrophils per 96-well Transwell or 1 × 10⁶ neutrophils per 24-well Transwell. Neutrophils were not washed following treatment with chemical inhibitors because of concerns that additional handling of cells would impact their activation state or alter relative counts prior to migration. Migration was allowed to progress for 2 h at 37°C, 5% CO₂. After 2 h, Transwells were discarded, and migrated neutrophils were quantified using a neutrophil myeloperoxidase (MPO) activity assay (28), whereby the magnitude of MPO activity (OD at 405 nm) measured using a colorimetric enzyme assay directly correlates with the number of neutrophils migrated in a linear fashion (R² > +0.99) (9, 29, 30). Standard curves were used for each condition of neutrophil pretreatment to control for possible variability in MPO production between neutrophil groups. Fig. 1 provides a schematic diagram of the neutrophil transepithelial migration assays designed to target the epithelial response or the neutrophil response. Data are displayed to reflect the magnitude of migration as the “percentage of PAO1-infected control” for comparative purposes.

Chemical inhibitors ONO-RS-082 (ONO; Enzo Life Sciences), cPLA2α inhibitor (cPLAi; Calbiochem), U0126 (Cell Signaling Technology), bromoenol lactone (BEL; Cayman Chemical), or diphenyleneiodonium (DPI; Sigma) were suspended as per the manufacturers’ instructions and diluted to the desired concentrations. Vehicle controls consisted of DMSO diluted with HBSS at a matched dilution factor.

Following treatment with HBSS, vehicle control, or chemical inhibitor, neutrophils were counted for viability. Neutrophils were stained with trypan blue, placed on a hemocytometer, and visualized under a microscope. Live and dead cells were counted manually to determine total viability.

Neutrophils were incubated in lucigenin (Invitrogen) at a concentration of 5 × 10⁷ cells per milliliter for 15 min on ice and transferred to a black, clear-bottom 96-well plate with 4 × 10⁶ neutrophils per well. Inhibitor (BEL, DPI) or vehicle control was added for an additional 10-min treatment. Cells were stimulated with 10 μM fMLF or HBSS, and reactive oxidase species (ROS) activity was quantified via luminescence and scintillation report using a TopCount machine (PerkinElmer).

Following transepithelial migration assays, Transwells were discarded, and plates were spun at 1800 rpm for 5 min in a centrifuge. Supernatant was collected and stored at −80°C. LTB4 was measured by ELISA per the manufacturer’s instructions (Cayman Chemical).

High-resolution micro-optical coherence tomography (μOCT) is a custom-built imaging technology providing optical resolution of 2 μm in the lateral/horizontal direction and 1 μm in the axial/vertical direction through analysis of reflectivity of a sample as a function of depth. μOCT imaging methods have been reported previously (31). The μOCT instrument was inverted, with the imaging laser beam directed to the sample from below; it required a custom holder with a transparent optical coherence tomography–compatible bottom to hold the Transwell containing the epithelium and neutrophils at a distance of ∼100 μm from the glass bottom. μOCT captured time-lapse three-dimensional (3D) imaging every 10 min over 2 h, immediately after the placement of neutrophils into the basolateral compartment. The sample was maintained near 37°C during migration by placing an incandescent heat source. The 3D Viewer plug-in in ImageJ was used to render 3D μOCT volume sequences. Neutrophil migration volume was determined by measuring the number of voxels in a region of interest exceeding a brightness threshold and dividing by the bright voxels measured from an isolated neutrophil.

Experiments were performed according to the guidelines of the Institutional Animal Care and Use Committee of Massachusetts General Hospital. Femurs and tibias of cpla2α^−/− mice on a BALB/c background or a 129Sv/C57BL/6 background, as well as inbred strain-matched littermate control mice (32), were removed for the harvest of bone marrow. Bone marrow was flushed and collected in HBSS−. RBCs were lysed with RBC lysis buffer, washed, and resuspended in HBSS− to a concentration of 20 × 10⁶ cells per milliliter. Transepithelial migration was assessed using single and mixed bone marrow migration across mouse lung epithelial (MLE) cell monolayers (19). Mixed bone marrow from cpla2α^−/− mice and wild-type controls was also stained differentially with CFSE (BioLegend, San Diego, CA) and combined at a 1:1 ratio to assess the potential for compensatory signaling. Migrated cells were collected and analyzed by flow cytometry. Data were collected on an Accuri C6 flow cytometer (BD Biosciences, San Jose, CA) and analyzed with FlowJo analysis software (TreeStar, Ashland, OR).

Data were reported as mean ± SD, and individual conditions were compared with the appropriate control using an unpaired two-tailed Student t test. To analyze multiple comparisons, we used one-way ANOVA. In case of significant differences, the Bonferroni posttest was used to assess differences within the group, or the Dunnett test was used to compare individual conditions with control. A p value ≤ 0.05 was considered statistically significant.

Eicosanoids are critical mediators of neutrophil transepithelial migration and have been shown to be produced by infected epithelial cells and migrating neutrophils. Because PLA2 activity plays a key role in eicosanoid generation (16, 22), its contribution was assessed in both cell types following epithelial infection with PAO1. Epithelial cells were pretreated with escalating doses of ONO, a general PLA2 inhibitor that nonspecifically targets all isoforms, for 2 h prior to infection. Following a 1-h infection, untreated neutrophils were placed in the basolateral compartment and allowed to migrate from the basolateral to the apical aspect of the epithelium for 2 h (Fig. 1). Epithelial PLA2 inhibition reduced PAO1-induced neutrophil transepithelial migration in a dose-dependent manner (Fig. 2A), consistent with previous studies (22). Inhibition of epithelial cells by PLA2 inhibitor (ONO) did not impact responses to HBSS buffer alone or to infection with nonpathogenic E. coli MC1000 (Fig. 2B). The contribution of PLA2 activity in neutrophils was then assessed by treatment of neutrophils with PLA2 inhibitor (ONO) in escalating doses for 30 min prior to migration. ONO- or vehicle-treated neutrophils were placed in the basolateral compartment of the Transwell and allowed to migrate across PAO1-infected epithelium (Fig. 1). Inhibition of neutrophil PLA2 activity was also associated with decreased migration in response to PAO1-infected epithelium (Fig. 2C), again in a dose-dependent manner. As was observed with inhibition of epithelial PLA2 activity, interference of neutrophil PLA2 activity resulted in reduced migration to PAO1 infection without impacting responses to buffer or to infection with the nonpathogenic E. coli strain MC1000 (Fig. 2D). Neutrophil treatment with ONO did not affect cell viability or disrupt the ability of neutrophils to produce MPO (Supplemental Fig. 1). Additionally, the effects of PLA2 activity inhibition were additive when the epithelium and neutrophils were simultaneously treated with ONO (Fig. 2E), suggesting that PLA2 activity is functionally important in neutrophil transepithelial migration in the epithelial and neutrophil signaling processes.

Cytosolic PLA2, specifically the cPLA2α isoform, is expressed in lung epithelium and neutrophils (24, 25) and has been associated with the synthesis of eicosanoids in a number of models (33–35). We previously investigated the PLA2 isoforms responsible for HxA3 generation and found that cPLA2α did not appear to impact HxA3 generation, despite the fact that it was expressed and operative in the epithelium, leading to the generation of PGE₂ (16). This selective eicosanoid-generating capacity emphasizes the importance of critically evaluating isoform-specific contributions of PLA2 activity within each cellular context where eicosanoids are generated. Although cPLA2α was dispensable for epithelial generation of HxA3, we considered the possibility that the cPLA2α isoform might be required in the neutrophil to promote neutrophil transepithelial migration, given the requirement for general PLA2 activity (Fig. 2C, 2D). Neutrophils were treated with escalating doses of cPLAi prior to assessment of transepithelial migration. cPLA2α inhibition in the neutrophil resulted in decreased transepithelial migration in response to apical epithelial infection with PAO1 in a dose-dependent manner (Fig. 3A). Migration to the exogenous chemoattractant IL-8, which does not trigger neutrophil LTB4-mediated augmentation, was not affected by cPLA2α inhibition (Fig. 3B). Neutrophil treatment with cPLAi did not affect cell viability or prevent MPO production (Supplemental Fig. 1). Prior studies have shown that cPLA2α inhibition in epithelial cells reduces the liberation of arachidonic acid and the subsequent generation of certain eicosanoids, including PGE₂. Inhibition of cPLA2α within the epithelium did not impact 12-lipoxygenase metabolites or neutrophil transepithelial migration (16). Consistent with these studies, we found that treatment of epithelial cells with cPLAi did not reduce PAO1-induced neutrophil transepithelial migration (Fig. 3C). Inhibition of cPLA2α did not impact responses to buffer alone or infection with MC1000, and it did not serve to inhibit neutrophil transepithelial migration toward an imposed IL-8 gradient. Thus, neutrophil cPLA2α activity is critical in mediating PAO1-induced neutrophil transepithelial migration, whereas epithelial cPLA2α activity is dispensable for this process.

cPLA2α activity is regulated posttranslationally through phosphorylation by the kinase ERK1/2 within a variety of cells, including neutrophils (24, 36–38). We tested whether inhibition of ERK in neutrophils would alter their ability to migrate across the PAO1-infected epithelium. Neutrophils were pretreated with the ERK inhibitor U0126 in escalating doses prior to migration. Inhibition of ERK activity decreased migration across the PAO1-infected epithelium in a dose-dependent manner (Fig. 4A). Inhibition of neutrophil-derived ERK activity revealed reduced migration to epithelial PAO1 infection but not an imposed chemotactic gradient of IL-8 (Fig. 4B). Neutrophil treatment with the ERK inhibitor U0126 did not reduce cell viability or MPO production (Supplemental Fig. 1). In agreement with previous studies (39), no effect on PAO1-induced neutrophil transepithelial migration was observed following epithelial treatment with the ERK inhibitor U0126 (Fig. 4C). These findings support the hypothesis that neutrophil-derived ERK1/2 is important for mediating PAO1-induced neutrophil transepithelial migration, most likely through kinase-associated activation of cPLA2α, because treatment of neutrophils with ERK1/2 inhibitor did not impair transepithelial migration toward a gradient of IL-8.

To assess the specificity of neutrophil cPLA2α activity on the process of neutrophil transepithelial migration, we examined the effect of neutrophil treatment with an inhibitor of distinct PLA2 isoforms. iPLA2β and iPLA2γ are present in neutrophils (40, 41) and epithelial cells (42, 43). To assess the impact of iPLA2 isoforms on neutrophil transepithelial chemotaxis, neutrophils were pretreated with the iPLA2-specific inhibitor BEL or vehicle control. There was no impact on migration in response to PAO1 infection over a range of BEL doses (Fig. 5A). Neutrophil treatment with BEL did not reduce migration to IL-8 (Fig. 5B), affect cell viability, or disrupt MPO production (Supplemental Fig. 1). Because BEL did not have any impact on any conditions in the migration assay, we sought to confirm that it was indeed functional at the dose examined. In neutrophils, iPLA2 isoforms have been previously demonstrated to mediate ROS production, because pretreatment with the iPLA2-specific inhibitor BEL abrogated this response (40, 41). Using a similar BEL pretreatment dose, we confirmed a significant inhibition of fMLF-induced ROS activity, as measured by a ROS-sensitive probe emitting chemiluminescence (Fig. 5C). The fMLF-induced ROS activity was completely abrogated by pretreatment with the NADPH oxidase inhibitor DPI, as expected (44) (Fig. 5C). Taken together, these results indicate that the neutrophil-derived signaling responsible for orchestrating chemotaxis is PLA2 activity selective, involving cPLA2α, but not iPLA2, isoform activity.

Having demonstrated that neutrophil-derived cPLA2α function is necessary for a robust migration response, we sought to assess whether this migration phenotype was mediated through cPLA2α-dependent generation of LTB4. To test this hypothesis, neutrophils were treated with cPLAi, ERK inhibitor (U0126), or vehicle control and then applied to the basolateral surface of epithelial monolayers, which were left untreated and exposed to an imposed apical-oriented gradient of IL-8 or were infected with PAO1. Following a 2-h incubation, supernatant from the apical compartment was collected, and LTB4 was quantified by ELISA, whereas neutrophil migration was assessed by MPO activity. Substantial LTB4 release was observed in the context of PAO1-induced migration. Both LTB4 release and total neutrophil migration were significantly reduced by pretreating neutrophils with cPLA2α or ERK inhibitors (Fig. 6A, 6B). Previous studies suggest that neutrophils represent the primary source of LTB4 in the epithelial–neutrophil coculture (19). IL-8–induced neutrophil migration did not evoke significant LTB4 generation (Fig. 6A), despite the occurrence of robust migration (Fig. 6B), consistent with previous studies (19). Because of the differences in total numbers of neutrophils migrating with and without inhibitor pretreatment, we normalized LTB4 release relative to the number of transmigrating neutrophils. Inhibition of neutrophil-derived cPLA2α or ERK by pharmacological compounds decreased the relative LTB4 production per migrated neutrophil in the setting of PAO1-induced migration (Fig. 6C). These findings support the role of cPLA2α activity in neutrophil-mediated LTB4 synthesis during migration. LTB4 release during PAO1-induced neutrophil transepithelial migration is necessary for augmenting the magnitude of migrating neutrophils, which are initiated to breach the epithelial barrier by epithelial-derived HxA3.

If PAO1-induced neutrophil transepithelial migration is augmented by neutrophil-derived LTB4 release through ERK1/2 activation of cPLA2α, then inhibition of iPLA2 isoforms (which do not impact PAO1-induced neutrophil transepithelial migration) would not be expected to impact LTB4 release. Neutrophils pretreated with BEL release equivalent amounts of LTB4 compared with vehicle control during PAO1-induced neutrophil transepithelial migration (Fig. 6D). As described above, BEL does not impact the magnitude of PAO1-induced neutrophil transepithelial migration (Fig. 5B). The LTB4 released per neutrophil was also quite similar between vehicle and BEL-treated neutrophils (Fig. 6E). In contrast, the general PLA2 inhibitor (ONO) significantly impaired neutrophil transepithelial migration (Fig. 6F), along with LTB4 release (Fig. 6G), when neutrophils were incubated with this compound prior to PAO1-induced neutrophil transepithelial migration. When evaluating LTB4 released per neutrophil (Fig. 6H), a clear and significant decrease is associated with neutrophils pretreated with the general PLA2 inhibitor (ONO). Thus, LTB4 release represents a critical functional attribute of PLA2 activity that is required for augmenting PAO1-induced neutrophil transepithelial migration, and the specific isoform cPLA2α, rather than iPLA2 isoforms, exhibits this behavior in this context.

To further evaluate the role of neutrophil-derived cpla2α in bacterial-induced neutrophil transepithelial migration, while avoiding concerns associated with pharmacological inhibition, a murine transepithelial migration assay system was used to leverage the availability of mice with targeted gene deletions (19). Bone marrow was isolated from cpla2α^−/− mice and littermate controls on a BALB/c background and was used as a source of neutrophils in our assay. MLE monolayers (MLE12) were grown on inverted Transwells and infected with PAO1, as described previously. We have previously demonstrated that bone marrow neutrophils are selectively recruited across MLE12 epithelial monolayers in response to PAO1 infection (19). Although bone marrow neutrophils from cpla2α^−/− mice and wild-type controls migrated across MLE12 monolayers, the magnitude of cpla2α^−/− neutrophils migrating across was significantly decreased in response to PAO1-infected MLE12 monolayers (Fig. 7A). Furthermore, migrated neutrophils of cpla2α^−/− mice had minimal LTB4 production compared with wild-type neutrophils (Fig. 7B). Defective migration of cpla2α^−/− bone marrow cells was not observed when migration was driven by an applied gradient of LTB4 (5 ng/ml placed in the apical well) across uninfected MLE12 monolayers (Fig. 7C). This result suggests that cpla2α^−/− bone marrow neutrophils are capable of effectively responding to LTB4 signals, if provided; however, because cpla2α^−/− bone marrow cells are not capable of generating LTB4 during migration, the magnitude of migration is suppressed, because they must rely exclusively on epithelial-elicited HxA3 to move across the MLE12 barrier. If this hypothesis is correct, then mixing wild-type and cpla2α^−/− bone marrow cells prior to initiating migration would serve to rescue the defect in cpla2α^−/− bone marrow cells, because wild-type cells are capable of providing the LTB4 gradient. To address this, wild-type and cpla2α^−/− bone marrow was differentially labeled and mixed in equal proportions. This mixed population was used as a source of neutrophils. In response to epithelial infection with PAO1, within the mixed population, cpla2α^−/− cells migrated as effectively as wild-type neutrophils, because migrated ratios were similar to input ratios (Fig. 7D). This suggests that the presence of wild-type neutrophils rescues any defect in migration exhibited by cpla2α^−/− neutrophils alone and is consistent with previous studies in which neutrophils lacking the ability to synthesize LTB4 fail to augment transepithelial migration; however, when mixed with neutrophils capable of synthesizing LTB4, they are fully capable of responding to the LTB4 signal and migrating across the epithelium (19). Neutrophils isolated from cpla2α^−/− mice on a 129Sv/C57BL/6 background were also investigated to ensure that these findings were not strain specific. Similar to the experiment using bone marrow from BALB/c mice, cpla2α^−/− neutrophils exhibited reduced transepithelial migration in response to PAO1-infected MLE12 monolayers and failed to synthesize LTB4 (Supplemental Fig. 2). Taken together, these data further support the importance of neutrophil-expressed cPLA2α in neutrophil-mediated production of LTB4 and its role in coordinating maximal transepithelial migration.

To evaluate these mechanisms in a model that is more closely associated with the physiology present within the human airways, a primary human airway coculture model was used. This model was recently developed ex vivo through the use of human airway basal stem cells cultured at an ALI (26). ALI culturing of primary human airway cells allows for the generation of a functional mucociliary epithelial barrier (45, 46), which exhibits multicellular architecture, appropriate cell-specific morphology, and physiologic functioning that resembles the airway surface in vivo (26). To take advantage of this complex airway morphology, we also incorporated imaging with μOCT. This new imaging technique provides 1-μm resolution and two-dimensional and 3D perspectives that can be analyzed over time and does not require fixation or staining. These attributes enable real-time imaging of microanatomical processes at the mucosal surface (31). Human airway basal stem cells derived from a healthy donor and cultured under ALI conditions were infected on the apical surface with PAO1, and primary human neutrophils were applied to the basolateral compartment, as described with the previous coculture models. During the 2-h migration, neutrophils can be seen migrating across the monolayer in clumps before dropping to the slide, which defines the bottom of the apical compartment (Fig. 8A). To assess the effect of neutrophil-derived cPLA2α on PAO1-induced neutrophil migration across human airway basal stem cell–derived epithelium, neutrophils were treated with cPLAi or vehicle control before application to the basolateral compartment. Migration in response to infection was significantly reduced in the setting of cPLAi-treated neutrophils, as observed by μOCT imaging (Fig. 8B) (Supplemental Video 1). This decrease in neutrophil penetration of the epithelium with cPLA2α-inhibited neutrophils can be quantified using a computer algorithm that assigns values representing the degree of neutrophil emergence into apical space over time (Fig. 8C). Standard MPO assays also confirm total decreased migration across PAO1-infected human airway basal stem cell–derived epithelium when neutrophil cPLA2α is inhibited (Fig. 8D). Lastly, neutrophils treated with cPLAi have decreased LTB4 production following migration across PAO1-infected human airway basal stem cell-derived epithelium (Fig. 8E), even when corrected for the numbers of neutrophils migrated (Fig. 8F). Together, these findings support the mechanism that neutrophil cPLA2α augments migration across infected mature primary basal stem cell–derived human airway mucosa through its involvement in synthesizing LTB4.

Transepithelial migration, the final step in a neutrophil’s journey from circulation to the airway, is a key microanatomical event that occurs during the inflammatory response at mucosal surfaces. Eicosanoids have been shown to be central to the process of neutrophilic breach of the epithelial mucosal barrier, delivering neutrophils to the site of infection (17, 19). However, in certain diseases this neutrophilic response may be maladaptive, worsening rather than resolving disease. Therefore, understanding the mechanisms driving eicosanoid-mediated neutrophil transepithelial migration may offer novel therapeutic targets to modulate the inflammatory process.

CF is a classic example of a disease plagued by chronic airway inflammation, with bouts of neutrophil infiltration that perpetuate pulmonary disease. Corticosteroids, which target PLA2 activity on the lipid membrane by inhibiting ERK-mediated phosphorylation (47), have been shown to reduce neutrophil inflammation in CF, but the numerous side effects outweigh the potential benefit (48). Nonsteroidal anti-inflammatories, such as ibuprofen, have been shown to reduce eicosanoid production and reduce airway inflammation in CF (49); however much like corticosteroids, the side effects often make long-term treatment intolerable (50). LTB4 activity has been targeted by blocking the LTB4 receptor (BTL1) with BLT1 antagonists given orally to CF patients; however, negative outcomes were documented, halting further examination of this approach (51, 52). Precise targeting of specific signaling pathways in the relevant tissue location may be the key to providing effective therapeutic benefit without the accompanying adverse effects.

cPLA2α has been implicated in pulmonary disease, including asthma, acute lung injury, and fibrosis (15). cPLA2α has been suggested to mediate allergic responses, including airway hyperresponsiveness and anaphylaxis, in mice following allergic stimulation (53). cPLA2α also appears to play a role in mediating acute lung injury in mouse models. Following insult by endotoxin or acid aspiration, cPLA2α-deficient mice are protected from protein leakage and neutrophil infiltration into the lung (33). Additionally, cPLA2α has been suggested to play a role in pulmonary fibrosis, because cPLA2α-deficient mice are protected from the inflammatory response and fibrosis brought on by bleomycin exposure (33). Preclinical studies have assessed the effect of a cPLAi, PF-5212372, on animal models of allergic asthma but have not considered neutrophilic responses (54, 55). Therefore, cPLA2α has been shown to play a role in the pulmonary disease, but a precise mechanism of action has not been described. Understanding its role in neutrophil transepithelial migration through LTB4 amplification may help us to better delineate the underlying mechanisms of disease.

Our data suggest that neutrophil cPLA2α plays a key role in neutrophil–epithelial cross-talk when tested using human and mouse epithelial coculture models. Human neutrophils treated with chemical inhibition targeting PLA2 reveal that this enzyme activity is required for neutrophil signaling in PAO1-induced neutrophil transepithelial migration. Further, chemical inhibition of the PLA2 subtype, cPLA2α, or the kinase required for activation of cPLA2α, ERK, reduces neutrophil transepithelial migration, suggesting that neutrophil-derived cPLA2α plays a key role in bacterial-triggered neutrophil transepithelial migration. Chemical inhibition of related PLA2 isoforms of the iPLA2 class within the neutrophil did not impact neutrophil migration function. To examine our findings without the requirement of a pharmacological agent, we used mice genetically deficient in cpla2α to assess transepithelial migration in the setting of infection. Mouse epithelium relies on release of HxA3 to initiate bacterial-induced transepithelial migration (19, 56). We found that neutrophils from cpla2α^−/− mice had reduced migration across the infected mouse epithelium, with corresponding reduced levels of LTB4 production. When responding to an exogenously applied gradient of LTB4, no difference was observed between cpla2α^−/− and wild-type–derived neutrophils, indicating that neutrophils deficient in cpla2α were not inherently impaired in their ability to migrate. Further, migration of a mixed population of cpla2α^−/− and wild-type–derived neutrophils across infected mouse epithelial monolayers revealed that the presence of wild-type neutrophils rescued the defective migration of cpla2α^−/− neutrophils, because both populations were equally represented following migration. These results support the hypothesis that cPLA2α drives neutrophil production of LTB4, which augments the magnitude of neutrophil transepithelial migration. As long as the LTB4 signal is provided, cpla2α^−/− neutrophils can migrate across the epithelium with equal efficiency as wild-type cells. Because cPLA2α is highly conserved between mice and humans, sharing 95% amino acid identity (57), function is highly likely to be preserved across species, further supporting the role of neutrophil cPLA2α in transepithelial signaling.

Investigating the precise mechanisms of transepithelial migration in vivo is difficult because it is technically challenging to tease apart the signaling patterns specific to each step of the migratory process, including transendothelial migration and migration through the extracellular matrix. To simulate the inflamed human airway, we used primary human airway basal stem cells grown in ALI-culturing conditions, which recapitulates the human transepithelial migration step in a more physiologic manner when paired with human neutrophils. Epithelium derived from human airway basal stem cells does not rely on cell immortalization, and it incorporates functional pulmonary features of mucus production and beating cilia, thus providing appropriate physiological context for these basic cellular mechanisms. Using this primary cell system, we found that blocking cPLA2α in neutrophils resulted in decreased transepithelial migration across P. aeruginosa–infected epithelium compared with controls. Our data further support our hypothesis that neutrophil cPLA2α plays a key role in P. aeruginosa–induced neutrophil transepithelial migration in humans.

Using a novel cellular imaging technique, μOCT, we were able to capture detailed cellular mechanisms governing neutrophil transmigratory behavior. μOCT uses optical interference with a reference beam to determine the depths of back-scattered light generated by a sample, creating a cross-sectional map of optical reflectance. μOCT has been successfully used in real time to characterize airway microanatomy, including cilia beat function and mucous layer depth, in a detailed 3D manner (31, 58). Imaging with this modality is unique because it does not require sample labeling or manipulation, and it provides cellular dynamics in real-time imaging. With μOCT, we readily observed the impact of cPLA2α inhibition on neutrophil migration. Neutrophils migrated across the epithelial barrier in organized clusters at various sites throughout the mucosal surface. When pretreated with cPLAi, the number of clusters per square millimeter appeared to decrease; however, it is unclear whether the neutrophil cluster size was impacted to any significant degree. The dynamic migratory process unveiled by μOCT imaging reveals a wealth of novel visual information that warrants further investigation. In addition, μOCT image-based data can be converted to quantitative metrics using computer algorithms that assign values to defined observations. Basic quantitative metrics, such as total migration, was demonstrated in this study; however, the opportunity exists to expand and develop novel measurements based on neutrophil speed, cluster numbers or size, cluster formation, and detachment dynamics, among any number of other possibilities (30).

In summary, neutrophil-derived cPLA2α plays a key role in amplifying the total neutrophilic response to bacterial-induced neutrophil migration by participating in the synthesis of LTB4 during the migratory process. The specific PLA2 isoform required for the generation of HxA3 within the epithelium following bacterial infection remains an open question that will be important to discern in future studies. cPLA2α has been implicated in mediating disease processes, but it has not been assessed specifically in the context of neutrophil recruitment. Furthermore, understanding the role of cPLA2α may be of benefit to specific diseases associated with neutrophil inflammation, such as CF or acute respiratory distress syndrome. Targeting pathways involved in neutrophil–epithelial cross-talk may offer therapeutic benefit, but a more detailed understanding of the mechanisms involved is required. Identification of the contribution of cPLA2α in neutrophil signaling offers an additional key piece of information in the study of neutrophilic airway epithelial breach.

We thank Eileen O’Leary, Brigham and Women’s Hospital Renal Division Laboratory Manager, for technical assistance, as well as Rhianna M. Hibbler and Kevin S. Gipson, members of the Mucosal Immunology and Biology Research Center, for critical and editorial review of the manuscript.

The authors have no financial conflicts of interest.