Polymorphonuclear Cell Transmigration Induced by Pseudomonas aeruginosa Requires the Hepoxilin A 3

This information is current as Bryan P. Hurley, Dario Siccardi, Randall J. Mrsny and Beth of September 24, 2021. A. McCormick J Immunol 2004; 173:5712-5720; ; doi: 10.4049/jimmunol.173.9.5712 http://www.jimmunol.org/content/173/9/5712 Downloaded from

References This article cites 40 articles, 19 of which you can access for free at: http://www.jimmunol.org/content/173/9/5712.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 24, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2004 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Polymorphonuclear Cell Transmigration Induced by 1 Pseudomonas aeruginosa Requires the Eicosanoid Hepoxilin A3

Bryan P. Hurley,*† Dario Siccardi,*†‡ Randall J. Mrsny,† and Beth A. McCormick2*†

Lung inflammation resulting from bacterial infection of the respiratory mucosal surface in diseases such as cystic fibrosis and pneumonia contributes significantly to the pathology. A major consequence of the inflammatory response is the recruitment and accumulation of polymorphonuclear cells (PMNs) at the infection site. It is currently unclear what bacterial factors trigger this response and exactly how PMNs are directed across the epithelial barrier to the airway lumen. An in vitro model consisting of human PMNs and alveolar epithelial cells (A549) grown on inverted Transwell filters was used to determine whether bacteria are capable of inducing PMN migration across these epithelial barriers. A variety of lung pathogenic bacteria, including Klebsiella pneumoniae, Escherichia coli, and Pseudomonas aeruginosa are indeed capable of inducing PMN migration across A549 mono-

layers. This phenomenon is not mediated by LPS, but requires live bacteria infecting the apical surface. Bacterial interaction with Downloaded from the apical surface of A549 monolayers results in activation of epithelial responses, including the phosphorylation of ERK1/2 and secretion of the PMN chemokine IL-8. However, secretion of IL-8 in response to bacterial infection is neither necessary nor sufficient to mediate PMN transepithelial migration. Instead, PMN transepithelial migration is mediated by the eicosanoid hep- oxilin A3, which is a PMN chemoattractant secreted by A549 cells in response to bacterial infection in a protein kinase C- dependent manner. These data suggest that bacterial-induced hepoxilin A3 secretion may represent a previously unrecognized inflammatory mechanism occurring within the lung epithelium during bacterial infections. The Journal of Immunology, 2004, http://www.jimmunol.org/ 173: 5712–5720.

iseases involving bacterial-induced lung inflammation of proteases and reactive oxygen species. However, these mecha- have been described, including bacterial-associated nisms are nonspecific and can lead to lung tissue damage, which if D bronchitis, pneumonia, and chronic conditions of pa- excessive, contributes to the pathology of the disease (7). tients having a mutation in the cystic fibrosis transmembrane con- To reach the airway lumen, PMNs are required to travel through ductance regulator (1–3). Several bacterial species have been im- several distinct tissue compartments within the alveolar wall (8, 9). plicated as causes of pneumonia, including Streptococcus PMNs must first escape the alveolar capillary, which involves the pneumoniae, Chlamydia pneumoniae, Haemphilus influenzae, termination of their flow through the vessel followed by adherence by guest on September 24, 2021 Klebsiella pneumoniae, Legionella pneumophila, Escherichia coli, to the lumen surface of the . The majority of PMNs and Pseudomonas aeruginosa (1, 3–5). Pseudomonas aeruginosa migrate across the junctions that connect the endothelial cells. (PA)3 is the bacterial pathogen most associated with lung damage Once across the endothelial barrier, PMNs navigate the endothelial due to inflammation in individuals with cystic fibrosis (2). basement membrane through pre-existing holes allowing PMNs to One of the pathogenic hallmarks of both cystic fibrosis and gain access to the interstitial space. PMNs then must travel through pneumonia is the accumulation of large numbers of or the extracellular matrix that encompasses the interstitial space, polymorphonuclear cells (PMNs) in the lumen of the lower airway where fibroblasts are thought to provide further directional guid- (1, 2, 6). PMNs present in the lumen can aid in eradicating of- ance. PMNs can then interact with the epithelial basement mem- fending bacteria via potent killing mechanisms, including release brane and basolateral surface of the epithelium. Finally, PMNs migrate between the epithelial cells to the apical surface of the epithelial barrier where they gain access to the airway lumen (8, 9). *Mucosal Immunology Laboratory, Massachusetts General Hospital, Charlestown, This process involves the integrated actions of cytokines, adhesion MA 02129; †Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, MA 02155; and ‡Cardiff University, School of Pharmacy, Cardiff, molecules with specificity for particular ligands as well as highly Wales timed and compartmentalized secretion of various PMN-specific Received for publication April 12, 2004. Accepted for publication August 17, 2004. chemokines such as IL-8 (8–10). Despite significant progress in The costs of publication of this article were defrayed in part by the payment of page this area of study, the specific molecular mechanism governing charges. This article must therefore be hereby marked advertisement in accordance PMN migration through each compartment to the lumen remains with 18 U.S.C. Section 1734 solely to indicate this fact. to be defined. 1 This work was supported by the National Institutes of Health Grant DK 56754 (to B.A.M.) and by a Pilot and Feasibility Grant from the Cystic Fibrosis Foundation (to Although there are numerous reports of various bacterial prod- B.A.M.). Salary support for B.P.H. was provided by the Ruth L. Kirschstein National ucts capable of inducing IL-8 secretion in several different primary Research Service Award Individual Fellowship awarded by the National Institute of and transformed airway epithelial cell lines (5, 6, 11, 12), it is Allergy and Infectious Diseases. unclear at present whether epithelial IL-8 production by lung ep- 2 Address correspondence and reprint requests to Dr. Beth A. McCormick, Mucosal Immunology Laboratories, Massachusetts General Hospital, CNY 114 (114-3503), ithelial cells is sufficient for mediating transepithelial PMN migra- Charlestown, MA 02129. E-mail address: [email protected] tion. Thus, the objective of this study was to specifically investi- 3 Abbreviations used in this paper: PA, Pseudomonas aeruginosa; PMN, polymor- gate the molecular mechanisms that are responsible for the final phonuclear cell; T-EDTA, trypsin-EDTA; LDH, lactose dehydrogenase; PKC, protein step of PMN recruitment during bacterial infection, namely, PMN ␣ kinase C; CCL, chelerytherine chloride; CDC, cinnamyl-3,4-dihydroxy- -cyanocin- transepithelial transmigration using a reductionistic in vitro model. namate; HXA3, hepoxilin A3; MAP, mitogen-activated protein; MAPK, MAP kinase; 12-LO, 12-; 5-LO, 5-lipoxygenase. Our approach to investigate the molecular mechanisms responsible

Copyright © 2004 by The American Association of Immunologists, Inc. 0022-1767/04/$02.00 The Journal of Immunology 5713 for a distinct phase of bacterial-mediated inflammation in the lung was added to the basolateral chamber (bottom). After 2 h, the amount of has revealed new insight into how bacteria interact with the lung HRP that translocated to the basolateral chamber was quantified by a HRP epithelium. We have uncovered a previously unrecognized inflam- activity assay. matory pathway in the lung epithelium involving the secretion of Cell viability assay the eicosanoid hepoxilin A3 (HXA3). HXA3 is an metabolite synthesized mainly through the actions of 12-lipoxy- To determine whether the bacterial infection resulted in toxicity to the A549 monolayers, release of the enzyme lactose dehydrogenase (LDH) genases, which can induce the release of calcium from intracellular into the supernatant of monolayers under infected conditions was com- stores of various cell types, including PMNs, via an intracellular G pared with uninfected monolayers. As a positive control for cell death, protein-coupled receptor (13). This action has been reported to Triton X-100 was added to monolayers. As with the PMN transmigration ϳ result in a diverse array of biological functions in various cell assay, inverted monolayers were washed and equilibrated in HBSS for 30 min. Monolayers were then flipped to expose the apical surface and treated types, including; the stimulation of insulin secretion from pancreatic for 1 h with HBSS alone or bacteria, washed, and flipped back over into islets of Langerhans, the regulation of cell volume in platelets, and the 24-well plates containing either HBSS or Triton X-100. After 2 h, the stimulation of PMN chemotaxis (13–16). Our studies herein suggest amount of LDH that has been released into the apical and basolateral cham- bers was quantified using the LDH assay (Sigma-Aldrich). a role for HXA3 in the directed migration of PMNs across lung epi- thelial monolayers infected with pathogenic bacteria. IL-8 assay Materials and Methods Inverted monolayers were washed and equilibrated in HBSS for ϳ30 min., ϫ 6 Growth and maintenance of epithelial cells flipped over, and treated with either 2 10 PA01/monolayer (low con- centration) or 2 ϫ 108 E. coli/monolayer (high concentration) or with The A549 cell line was derived through explant culture of lung carcino- HBSS alone. After 1 h, all monolayers were washed and flipped back over Downloaded from matous tissue from a 58-year-old Caucasian male. These cells display prop- into 24-well plates and 100 ␮l of HBSS was added to the top (basolateral) erties of type II alveolar epithelial cells and also form polarized barriers chamber and 600 ␮l of HBSS was added to the bottom (apical) chamber. when grown on permeable filters (10, 17–19). A549 cells were maintained After 2 h, supernatants in the apical and basolateral chambers were col- lected and filtered (0.2 ␮M) before performing the IL-8 ELISA (Pierce/ in Ham’s F-12K medium with 2 mM L-glutamine 1.5 g/L NaHCO3, 10% FBS, and 100 U of penicillin/streptomycin. Polarized monolayers of A549 Endogen, Rockford, IL) (22). cells were grown and maintained on the underside of 0.33-cm2 collagen- coated Transwell filters to study PMN migration in the physiological ba- Detection of ERK1/2 activation http://www.jimmunol.org/ solateral to apical direction. A549 monolayers seeded on 1.5-cm2 permeable filters were infected with PMN isolation PA01 [low] or E. coli F-18 [high] on the apical surface for various times. Cells were treated with lysis buffer which includes 1% Triton X-100, 100 PMNs were isolated from whole blood anticoagulated with acid citrate/ mM NaCl, 10 mM HEPES (pH 7.6), 2 mM EDTA, 200 mM PMSF, 4 mM dextrose obtained from healthy human volunteers. The buffy coat was ob- Na3VO4, 40 mM NaF, and 1 Complete Mini protease inhibitor mixture tained by a 400 ϫ g spin at room temperature. Plasma and mononuclear tablet/10 ml lysis buffer (Roche Diagnostics, Mannheim, Germany). Cells cells were removed by aspiration, and the majority of the RBCs were were then scraped from the permeable filter and passed through a 25-gauge removed using a 2% gelatin sedimentation technique. Residual RBCs were needle, followed by centrifugation (14,000 rpm at 4°C for 30 min). The removed by lysis in cold NH4Cl lysis buffer. This technique allows for supernatant represents the Triton X-100-soluble fraction and protein con- Ͼ rapid isolation of functionally active PMNs ( 98%) at 90% purity (20). centration of each supernatant was measured by the Dc Protein Assay (Bio- by guest on September 24, 2021 Rad, Hercules, CA). Lysates were normalized for protein concentration and Bacterial strains ϳ30 ␮g was run on an 8–16% gradient polyacrylamide gel and transferred The strains used in this study were all grown overnight at 37°C under to nitrocellulose. Blots were probed with either an anti-ERK1 Ab or anti- aerobic conditions to a stationary growth phase. PA (PA01, PA14, PA103), phospho-ERK1/2 Ab (, Beverly, MA) followed by incuba- the enteric commensal E. coli (F-18), and the E. coli K12 strain (MC1000) tion with the appropriate HRP-conjugated goat anti-species Ab and detec- were grown in L broth. K. pneumoniae (ATCC 9590) was grown in nu- tion with ECL reagent (Pierce/Endogen, Rockford, IL) (23). trient broth. Bacteria were resuspended to a concentration of ϳ6 ϫ 109 bacteria/ml HBSS and diluted accordingly. LPS from E. coli and PA (100 Detection of protein kinase C (PKC) activation ␮g/ml) was obtained from Sigma-Aldrich (St. Louis, MO). A cell fractionation protocol was used to obtain membrane-soluble pro- PMN transmigration assay teins, as PKC accumulates in the membrane upon activation (24, 25). A549 monolayers seeded on 1.5-cm2 permeable filters were infected with PA01 The PMN transmigration assay using inverted cell culture monolayers of [low] or E. coli F-18 [high] on the apical surface for various times. Cells polarized cells has been described previously (20). Briefly, inverted mono- were scraped off of permeable filters in buffer containing 150 mM Tris (pH layers were washed and equilibrated in HBSS for ϳ30 min. Inverted mono- 8.0), 15 mM EDTA, 6 mM EGTA, 200 mM PMSF, 4 mM Na3VO4,40mM layers were then flipped over and treated with 25 ␮l of various concentra- NaF, and 1 Complete Mini protease inhibitor mixture tablet/10 ml buffer. tions of bacteria. Alternatively, monolayers were treated with LPS or with Scraped cells were subjected to sonication followed by centrifugation at HBSS alone. After 1 h, all monolayers were washed and flipped back over 55,000 rpm for 1 h. Cell pellet was resuspended in a lysis buffer containing into 24-well plates. For a positive control of the ability of PMNs to migrate, 0.1% Triton X-100, 0.2% SDS, 50 mM Tris (pH 8.0), 5 mM EDTA, 2 mM Ϫ8 10 ␮M fMLP was added to the bottom (apical) chamber of uninfected EGTA, 200 mM PMSF, 4 mM Na3VO4, 40 mM NaF, and 1 Complete Mini monolayers for each experiment. PMNs (1 ϫ 106) were then added to the protease inhibitor mixture tablet/10 ml buffer. Lysates were again subjected top (basolateral) chamber and the plate was placed at 37°C for 2 h. PMNs to sonication followed by centrifugation at 30,000 rpm for 10 min. Super- that fully migrated into the apical chamber were quantified by the myelo- natant was collected and concentrated using Centricon filters with a Mr peroxidase assay (20). Data are displayed as mean (SD) of at least three 30,000 cutoff. Lysates were normalized for protein concentration and run independent monolayers/condition. on an 8–16% gradient polyacrylamide gel followed by transference to nitrocellulose. Blots were probed with an anti-phospho-pan PKC Ab (Cell Assay to assess barrier integrity Signaling) followed by incubation with HRP-conjugated goat anti-rabbit Ab and detection with ECL reagent (Pierce/Endogen) (24, 25). To determine whether the bacterial infection resulted in disruption of the A549 barriers, movement of the protein HRP across infected vs uninfected Drug treatments monolayers was compared (21). As a positive control for barrier disruption, uninfected monolayers were treated with trypsin-EDTA (T-EDTA) fol- For drug treatment experiments, cells were pretreated for either 1 h with lowed by washing and addition of HRP. As with the PMN transmigration chelerytherine chloride (CCL), U0126, caffeic acid, or cinnamyl-3,4-dihy- assay, inverted monolayers were washed and equilibrated in HBSS for ϳ30 droxy-␣-cyanocinnamate (CDC; BIOMOL, Plymouth Meeting, PA) or for min. Monolayers were then flipped to expose the apical surface and treated 18 h with cycloheximide (Sigma-Aldrich before infection. The drug con- for 1 h with HBSS alone, bacteria, or with T-EDTA. Monolayers were then centration was maintained during infection, but was removed by washing washed and flipped back over into 24-well plates. HRP (0.5 ␮g/ml) in three times before addition of PMNs. Each drug at the doses presented in HBSS (100 ␮l) was added to the apical chamber (top) and 600 ␮l of HBSS this study had minimal effect on cell viability both in the presence or 5714 HXA3 MEDIATES BACTERIAL-INDUCED PMN TRANSMIGRATION absence of bacterial infection, as assessed by the barrier integrity assay described above. Also, none of these drugs had any major affect on the amount of bacteria adhering to the A549 monolayers.

Detection of HXA3 A549 monolayers seeded on 1.5-cm2 permeable filters were infected with PA01 in the presence or absence of 50 ␮M CDC. After 1 h, A549 cells were washed three times and HBSS was added. After 2 h, conditioned HBSS was collected and placed at Ϫ80°C. A method for the detection and quantification of HXA3 in cell supernatants has been recently described (15). Briefly, samples were thawed and passed through an Amicon ultra- filtration apparatus (Millipore, Bedford, MA) fitted with a 2000-Da cutoff membrane. Filtrate components were size fractionated by hydrophobic sur- face chromatography using methanol elution. The methanol fraction was dried under vacuum and injected onto a Vydac C18 (10 ␮m; 300 Å) semi- preparative column (10 ϫ 250 cm). A methanol gradient of 1–10% over 10 min, then 10–60% over 25 min, followed by 60–100% over 45 min was used to isolate HXA3 fractions having no detectable absorbance at 280 nm and weak absorbance at 214 nm. Samples were analyzed using a Genesis C18 (4 mm, 120 Å) analytical HPLC column (4.6 ϫ 150 mm) equilibrated with 5 mM triethylamine/acetic acid (pH 7.2). Samples were chromato- graphed using a linear methanol gradient of 0–100% over 60 min and Downloaded from analyzed using a ThermoFinnigan LCQDeca HPLC/electrospray mass spectrometer set in the negative ion mode. Compounds were characterized for retention time, UV spectra (UV6000LP photodiode array detector), and m/z signals (15).

Results http://www.jimmunol.org/ Pathogenic bacteria induce PMN transepithelial migration As a model for the airway epithelium, we chose the A549 cell line. A549 cells are alveolar epithelial cells that are capable of forming polarized barriers and are widely used as a model of the airway FIGURE 1. Bacterial-induced PMN transmigration across lung epithe- epithelial surface. In addition, these cells have been used exten- lial cells. A, A549 monolayers are infected with 2 ϫ 108 bacteria/mono- sively for PMN transepithelial migration studies (10). A549 cells layer [high] and the number of PMNs that have completely migrated across are grown on the underside of Transwell permeable supports the epithelium into the apical chamber are quantified and presented as 4 where they form cellular barriers as described in Materials and number of PMNs multiplied by 10 . B, A549 monolayers are infected with Methods. To investigate whether respiratory pathogens can induce various concentrations of bacteria. C, A549 monolayers are infected with by guest on September 24, 2021 2 ϫ 106 bacteria/monolayer [low]. HBSS represents uninfected monolay- PMN transepithelial migration across model lung epithelia, we se- ers. Each condition was performed in triplicate and data presented are a lected bacterial strains that have been associated with lung disease representative figure of an experiment performed at least twice. involving inflammation, including, PA as well as two species of Gram-negative commensal enteric bacteria, E. coli (F-18) and K. pneumoniae (1–3). We also selected a laboratory K-12 E. coli strain (MC1000), which has distinct properties from natural human cell viability and even at the [high] (2 ϫ 108 bacteria/monolayer) colonizing bacteria and might potentially provide some insight into of infection, none of the strains resulted in any significant barrier the specificity of the bacterial-induced transepithelial migration disruption as measured by HRP flux across the monolayers or cell response. We infected A549 monolayers with [high] (2 ϫ 108 death as assessed by LDH release (Table I). Although, it is currently bacteria/monolayer) and observed a potent PMN transmigration unclear why [high] of PA strains fail to induce the PMN transmigra- response to both species of Gram-negative enteric bacteria tested tion response, such data suggest that several bacterial species capable (Fig. 1A). The response to MC1000 was minimal and similar to of causing lung inflammation and disease induce PMN transmigration uninfected monolayers (HBSS). Surprisingly, the PA strains across lung epithelial cells in this in vitro model. (PA01, PA14, and PA103) had a weak response. To determine the We next sought to determine the host-pathogen interactions crit- concentration where E. coli F-18 and K. pneumoniae lose the abil- ical for the PMN transmigration response. Induction of the PMN ity to induce PMN transmigration, we performed dose-response transmigration response required live bacteria interacting with the studies. As shown in Fig. 1B, E. coli F-18 promoted PMN trans- apical surface, since there was no detectable response to treatment migration only at [high], as F-18 was unable to induce PMN trans- of the apical surface of A549 monolayers with heat-killed bacteria migration at or below a concentration of 1 ϫ 107 bacteria/mono- (Fig. 2A). We also observed that purified LPS from PA or E. coli layer. In contrast, K. pneumoniae had a much larger concentration was incapable of inducing any degree of PMN transmigration. Al- range of effective response, with an equally potent PMN transmi- though bacterial fMLP is a potent PMN chemoattractant, it does gration response occurring over several orders of magnitude. The not appear to be involved in the bacterial-induced PMN transmi- K12 E. coli MC1000 had little or no response at all concentrations gration response reported herein, since a specific antagonist of tested (Fig. 1B). Despite the fact that the PA strains were incapable fMLP, tBOC, was unable to prevent bacterial-induced migration, of inducing PMN transmigration at [high], we observed that these while effectively blocking transmigration (96% inhibition) in re- strains did, in fact, stimulate a robust response at lower concen- sponse to artificially imposed fMLP gradients across A549 mono- trations (2 ϫ 105–1 ϫ 107 bacteria/monolayer; Fig. 1, B and C). layers. Furthermore, A549 cells pretreated with the eukaryotic pro- The lack of a PMN transmigration response at a [high] of PA tein synthesis inhibitor cycloheximide were markedly reduced strains did not appear to be the consequence of PA toxicity to the (50%) in facilitating PMN transmigration in response to bacterial monolayer. We assessed A549 barrier integrity as well as A549 infection, suggesting that A549 epithelial monolayers are playing The Journal of Immunology 5715

Table I. Effect of bacterial infection of A549 monolayers on barrier integrity and cell viability

Barrier Integritya Cell Viabilityb

[high]c [low]d [high] [low]

HBSS 30 (1) NDe 0.4 (0.1) ND T-EDTA 210 (75) ND ND ND Triton X-100 ND ND 85.5 (4.9) ND PA01 24 (12) 34 (8) 0.6 (0.4) 0.1 (0.1) MC1000 24 (8) 30 (9) 0.7 (0.3) 0.3 (0.2) Klebsiella 30 (6) 25 (6) ND ND F-18 30 (26) 26 (7) ND ND

a Barrier integrity is assessed by the amount (nanograms per milliliter) of HRP (MW 45 kDa) that crosses the A549 monolayers in 2 h following treatment with the conditions listed above. Monolayers were treated with T-EDTA as a positive control to intentionally disrupt the barrier integrity of the monolayer. There is no significant difference among uninfected, untreated monolayers (HBSS), and all infected mono- layers. All conditions were performed in triplicate with SDs reported in parentheses. Each experiment was performed at least twice. b Cell viability was assessed using the LDH assay. Values are reported as percent release of LDH into the supernatant as a percentage of total LDH (cell associated plus Downloaded from released in supernatant). Monolayers were treated with Triton X-100 as a positive control to intentionally disrupt A549 cell viablity. There is no significant difference among uninfected, untreated monolayers (HBSS), and all infected monolayers. All conditions were performed in triplicate with SDs reported in parentheses. Each ex- periment was performed at least twice. c [high] refers to 2 ϫ 108 bacteria/monolayer. d [low] refers to 2 ϫ 106 bacteria/monolayer. e ND, Not done.

FIGURE 2. PMN transmigration vs IL-8 secretion. A549 monolayers http://www.jimmunol.org/ are treated with buffer alone (HBSS), 2 ϫ 106 bacteria/monolayer [low] of PA01, or an equivalent concentration of PA01 that has been heat killed (20 an active role in the process of bacterial-induced PMN transmi- min. at 70°C), and monolayers are assessed for either PMN transmigration gration. Based on the above observations, it is our contention that (A) or IL-8 secretion (B). A549 monolayers are treated with buffer alone (HBSS), 2 ϫ 108 bacteria/monolayer [high], or 2 ϫ 106 bacteria/monolayer the ability of different species of Gram-negative lung pathogenic [low] of PA01 and monolayers are assessed for either PMN transmigration bacteria to induce PMN transmigration across airway epithelial (C) or IL-8 secretion (D). Each condition was performed in triplicate and monolayers may represent a conserved innate mechanism exhib- data presented are a representative figure of an experiment performed at ited by the lung epithelium in response to various bacterial species least twice. that are capable of colonizing the lung. It remains to be determined by guest on September 24, 2021 whether PMN transmigration occurs in response to Gram-positive lung pathogenic species such as S. pneumoniae. Epithelial cells at several mucosal sites have been shown to produce PMN chemoat- stimuli responsible for inducing the PMN transmigration response. tractants in response to various stimuli including bacterial infec- In addition, the IL-8 secretion response occurred to a similar de- tion (11, 12, 20, 21). Since the lung epithelial cells appear to play gree with either [high] or [low] PA01 infection (Fig. 2D), whereas an active role in this process, we next sought to identify a PMN the PMN transmigration response only occurred with [low] PA01 chemoattractant responsible for mediating the PMN transmigration infection (Fig. 2C). Taken together, these data indicate that bac- process. terial-induced secretion of IL-8 from A549 cells is not sufficient to Pathogenic bacteria induce IL-8 secretion mediate PMN transmigration. IL-8 is a well-known PMN chemoattractant that is secreted by the epithelial cells of several mucosal surfaces upon interaction with Pathogenic bacteria induce ERK1/2 phosphorylation bacteria (5, 11, 20, 21). The signaling cascade leading to IL-8 To further understand signaling events involved in bacterial inter- secretion has been well studied and thought to represent an innate action with the apical surface of A549 monolayers that may lead to response by mucosal surfaces upon the sensing of infection (2, 26, IL-8 production, we investigated whether PA01 and/or E. coli F-18 27). The signaling cascade is initiated by epithelial cell recognition activates the MAPK/ERK pathway. The MAPK/ERK pathway is of pattern motifs presented by ubiquitous bacterial surface prod- comprised of several serine/threonine kinases involved in the reg- ucts such as LPS or the bacterial protein flagellin through TLRs ulation of a wide range of cellular responses including the produc- (12, 26–28). Such specific recognition is followed by signaling tion of proinflammatory molecules such as IL-8 (29). ERK1 and 2 through mitogen-activated protein (MAP) kinases and NF␬B lead- kinases are phosphorylated by activated MEK1/2 as a consequence ing to the epithelial secretion of IL-8 (2, 26–28). As shown in Fig. of the phosphorylation cascade of the active MAPK/ERK pathway 2, B and D, PA01 induces a significant increase in IL-8 secretion (29). For these studies, we selected concentrations of the respec- compared with uninfected A549 monolayers (HBSS). IL-8 de- tive bacteria that are capable of inducing PMN transmigration, i.e., picted in Fig. 2, B and D, represents the quantity of IL-8 recovered [low] PA01 and [high] E. coli F-18. PA01 (2 ϫ 106 bacteria/ in the basolateral compartment, since the quantity of IL-8 in the monolayer) and F-18 (2 ϫ 108 bacteria/monolayer) infection of apical compartment was undetectable. Unlike results observed A549 monolayers both result in an increase in the phosphorylated with the PMN transmigration response (Fig. 2A), heat-killed PA01 form of ERK1 and 2 within A549 cells compared with uninfected displayed a comparable, if not greater, IL-8 secretion response (HBSS) A549 monolayers (Fig. 3), suggesting that the MAPK/ compared with infection with live bacteria (Fig. 2B). This indicates ERK pathway is activated upon infection (29). The amount of total that the bacterial stimuli inducing the IL-8 response differ from the ERK1 protein is unchanged in response to infection. 5716 HXA3 MEDIATES BACTERIAL-INDUCED PMN TRANSMIGRATION

induced PMN transepithelial migration in vitro. This notion is fur- ther substantiated by the fact that the addition of neutralizing IL-8 Abs failed to inhibit bacterial-induced PMN transmigration (data not shown). In addition, the majority of the IL-8 secreted by A549 monolayers is recovered in the basolateral chamber, which is counter to the direction of the PMN migration. Taking together all of the above data, we hypothesize that although IL-8 is produced in response to bacterial infection, IL-8 is not responsible for the specific step of transepithelial migration of PMNs to the airway FIGURE 3. PA and E. coli induce phosphorylation of ERK1/2. A549 lumen. Rather IL-8 is more likely required for the recruitment of monolayers were treated with buffer alone (HBSS), E. coli F-18 [high] PMNs from the blood stream through the interstitial space (5, 8, ϫ 8 ϫ 6 (2 10 bacteria/monolayer), or PA PA01 [low] (2 10 bacteria/mono- 20, 22). Hence, PMN transepithelial migration must involve a dis- layer). Cell lysates were collected, normalized for protein concentration, tinct PMN chemoattractant. subjected to Western blotting, and probed with either an anti-phospho- ERK1/2 or anti-ERK1. Figure represents an experiment performed at least three times with similar results. PKC inhibitors block bacterial-induced PMN transmigration Previous studies from our laboratory have reported that PKC ac- tivity is required for bacterial-induced PMN transmigration across MAP/ERK inhibitors do not block bacterial-induced PMN intestinal epithelial cells (25). Using the PKC inhibitor CCL (30), Downloaded from transmigration we show in Fig. 5A that CCL pretreatment of A549 cells nearly Since the evidence above suggests that the MAP/ERK pathway is completely ablates PA01 and significantly reduces F-18-induced activated in response to bacterial infection, we next sought to de- PMN transmigration. Furthermore, PMN transmigration in re- termine whether this signaling pathway contributes to either the sponse to an imposed fMLP gradient is unaffected by CCL pre- IL-8 response or the PMN transmigration response described treatment of A549 cells, demonstrating that the drug pretreatment above. We treated A549 monolayers with the MEK1/2-specific of A549 cells does not nonspecifically interfere with the ability of inhibitor U0126 (29) before infection with either F-18 or PA01: PMNs to migrate across the monolayer (data not shown). Also, http://www.jimmunol.org/ one set of experiments was assayed for IL-8 secretion and another CCL treatment at the same doses had no effect on the ability of was assayed for PMN transmigration. U0126 (20–40 ␮M) effi- PA01 or F-18 to induce ERK1/2 phosphorylation or IL-8 secretion ciently reduces the PA01-and F-18-stimulated IL-8 secretion by from A549 cells (data not shown and Fig. 5B). These data suggest 50–75% (Fig. 4A). This result was not surprising because IL-8 that PKC activity within A549 cells is necessary for the bacterial- secretion from many cell types has been shown to be dependent on induced PMN transmigration response. the MAP/ERK pathway (29). However, at doses of U0126, which Since PKC activity appeared to play a significant role in bacte- efficiently blocked the IL-8 response, there was no significant ef- rial-induced PMN transmigration, we next sought to determine

fect on PMN transmigration (Fig. 4B). The MEK1-specific inhib- whether PKC is activated in A549 cells in response to bacterial by guest on September 24, 2021 itor PD98059 was also capable of blocking IL-8 production while infection. We treated A549 monolayers with either buffer alone having no significant effect on the PMN transmigration response (HBSS), 1 ␮M PMA as a positive control, or with PA01 for 15–45 (data not shown). These results suggest that bacterial-induced IL-8 secretion appears to be an unnecessary event for the bacterial-

FIGURE 4. MEK1 inhibitor blocks IL-8 secretion, but not PMN trans- FIGURE 5. PKC inhibitor blocks PMN transmigration, but not IL-8 se- migration. A549 monolayers were pretreated with various doses (micro- cretion. A549 monolayers were pretreated with various doses (micromolar) molar) of the MEK1 inhibitor U0126. A549 monolayers were then treated of the PKC inhibitor CCL. A549 monolayers were then treated with buffer with buffer alone (HBSS), E. coli F-18 [high] (2 ϫ 108 bacteria/mono- alone (HBSS), E. coli F-18 [high] (2 ϫ 108 bacteria/monolayer), or PA layer), or PA PA01 [low] (2 ϫ 106 bacteria/monolayer). Monolayers are PA01 [low] (2 ϫ 106 bacteria/monolayer). Monolayers are assessed for assessed for either IL-8 secretion (A) or PMN transmigration (B). Each either PMN transmigration (A) or IL-8 secretion (B). Each condition was condition was performed in triplicate and the data presented are represen- performed in triplicate and the data presented are representative of an ex- tative of an experiment performed at least three times. periment performed at least three times. The Journal of Immunology 5717 min. After treatment, cells are lysed and separated into cytoplas- mic and membrane fractions (20). Previous studies have shown that upon activation, PKC becomes phosphorylated and relocalizes to the (20, 21). Fig. 6 depicts the amount of phos- phorylated PKC in the cellular membrane fraction of A549 cells treated under various conditions and clearly demonstrates that in- fection with PA01 results in an increase in the amount of phos- phorylated PKC in the membrane fraction compared with unin- fected monolayers.

Bacterial-induced PMN transmigration involves HXA3 Since PKC is activated within A549 cells upon infection with PA01 and since PKC activity appears to be involved in the process of PA01-induced PMN transmigration, the mechanism mediating bacterial-induced PMN transmigration across lung epithelial cells displays similarities to bacterial-induced PMN transmigration across intestinal epithelial cells (15, 25). A recent report from our group has identified a PKC-dependent PMN chemoattractant pro- duced by intestinal epithelial cells in response to Salmonella ty- Downloaded from phimurium infection that is responsible for mediating PMN trans- FIGURE 7. PMN transmigration is blocked by the 12-LO inhibitor, but epithelial migration (15). This molecule, previously termed not by the 5-LO inhibitor. A549 monolayers were pretreated with various pathogen-elicited epithelial chemoattractant, was identified as doses (micromolar) of the 12-LO inhibitor CDC (A) or the 5-LO inhibitor HXA (15). HXA is an eicosanoid metabolized from arachidonic caffeic acid (B). A549 monolayers were then treated with buffer alone 3 3 ϫ 6 acid (13). The enzyme that is largely responsible for the synthesis (HBSS) or PA PA01 [low] (2 10 bacteria/monolayer). Monolayers are assessed for PMN transmigration. Each condition was performed in trip- of HXA from arachidonic acid is 12-lipoxygenase (12-LO). http://www.jimmunol.org/ 3 licate and the data presented are representative of an experiment performed 12-LO converts arachadonic acid to 12-hydroxyperoxyeicosatet- at least twice. ranoic acid (12-HpETE), which can then be converted to HXA3 (13). We therefore investigated whether bacterial-induced PMN transmigration across the lung epithelial cells requires 12-LO ac- tivity. Pretreatment of A549 monolayers with CDC, a 12-LO-spe- lateral chamber. Of the total number of PMNs added to the apical cific inhibitor (31), significantly blocked PMN transmigration in- surface, 18.4 Ϯ 5.1% migrated to the apical chamber in response duced by bacteria, but did not significantly interfere with PMN to a gradient of HXA3. A different arachadonic acid metabolite migration in response to a fMLP gradient (Fig. 7A and data not A4 (50 ng/ml) placed apically is incapable of directing shown). An inhibitor of the 5-lipoxygenase (5-LO) pathway, caf- PMNs across the epithelium as the amount of PMNs migrating to by guest on September 24, 2021 feic acid (32), which converts arachidonic acid into the apical chamber (1.6 Ϯ 0.3%) is no different from the number rather than hepoxilins is unable to block bacterial-induced PMN of spontaneously migrating PMNs, where no eicosanoid gradient is transmigration (Fig. 7B). These data suggest that the 12-LO enzy- added (1.2 Ϯ 0.1). Based on this data, we hypothesize that A549 cells release HXA in response to bacterial infection and this mol- matic pathway, which is the major synthetic pathway of HXA3,is 3 involved in the process of bacterial-induced PMN transmigration ecule is capable of directing PMN transepithelial migration. across lung epithelial cells.

To determine whether A549 cells secrete HXA3, we analyzed Discussion supernatants from the apical surface of polarized monolayers of Bacterial species such as E. coli, K. pneumoniae, and PA present

A549 cells for the presence of HXA3 using HPLC separation fol- in the airway are capable of causing severe inflammation in dis- lowed by mass spectrometry as described in Materials and Meth- eases such as pneumonia and cystic fibrosis (1–3). One aspect of ods (15). A549 monolayers incubated in buffer alone did not pro- the inflammatory process involves the infiltration of lung tissue by Ͻ duce any detectable HXA3 ( 1.0 pmol). In response to infection neutrophils (2, 5, 6, 8, 33). Neutrophils are capable of crossing ϳ with PA01, however, A549 cells produced 210 pmol HXA3. both the endothelial and epithelial barriers in response to an in- Secretion of HXA3 was prevented by pretreating A549 cells with fection in the airspace (8). In this study, we have modeled the the 12-LO inhibitor CDC (Ͻ 1.0 pmol). directed migration of neutrophils across lung epithelial monolay-

We next verified that HXA3 acts as a PMN chemoattractant ers; a distinct phase of the complex recruitment process across the A549 barrier; we added commercially available HXA3 that occurs during lung inflammation. We have shown that several (50 ng/ml; Calbiochem, San Diego, CA) to the apical chamber of bacterial species stimulate PMN transmigration across lung epi- A549 monolayers with PMNs simultaneously placed in the baso- thelial cell monolayers. Each bacterial strain we investigated in this study displays a distinct profile in terms of concentration re- quired for induction of transmigration. For example, the intestinal commensal E. coli (F-18) induces a PMN transmigration response only at [high] infection (2 ϫ 108 bacteria/monolayer), whereas the intestinal commensal K. pneumoniae results in a robust PMN FIGURE 6. PA induced activation of PKC. A549 monolayers were transmigration not only at 2 ϫ 108 bacteria/monolayer, but also at treated with buffer alone (HBSS), 1 ␮M PMA, PA PA01 [low] (2 ϫ 106 bacteria/monolayer), or E. coli K12 MC1000 [low] (2 ϫ 106 bacteria/ lower concentrations over several orders of magnitude. A large monolayer). Membrane fractions were collected, normalized for protein response to K. pneumoniae occurs with as little as 100,000 bacte- concentration, subjected to Western blotting, and probed with anti- ria/monolayer. The K12 E. coli strain (MC1000), which is less phospho-PKC. Figure represents an experiment performed at least three likely to be pathogenic in the lung, is not capable of inducing times with similar results. substantial PMN transmigration at any concentration tested. 5718 HXA3 MEDIATES BACTERIAL-INDUCED PMN TRANSMIGRATION

In contrast to E. coli and K. pneumoniae, PA strains display a interfere with the PMN transmigration response. Finally, we have unique feature in which a productive PMN transmigration re- shown that inhibitors of PKC ablate bacterial-induced PMN trans- sponse occurs with a bell curve distribution over a range of infec- migration, but have no effect on IL-8 production. Based on this tion concentrations. At higher concentrations (2 ϫ 108 bacteria/ evidence, IL-8 does not appear to contribute to PMN migration monolayer), the PMN transmigration response is minimal. In across infected lung epithelial monolayers. Since IL-8 does con- contrast, lower concentrations of ϳ1 ϫ*105–1 ϫ 107 bacteria/ tribute to overall PMN recruitment during lung infection and in- monolayer results in substantial PMN transmigration. Below a flammation (8, 9, 10, 33), we hypothesize that the primary role concentration of ϳ1 ϫ 105 bacteria/monolayer, the PMN transmi- played by IL-8 during PMN infiltration of lung tissue may be to gration response is diminished. These data imply that there is guide PMNs through the interstitium up to the epithelium, rather something unique about the interaction of higher concentrations of than guiding PMNs across the epithelium into the airspace (15, PA with the epithelial monolayer that renders them incapable of 20). This assertion is based on the fact that epithelial cells produce facilitating PMN transepithelial migration. This result cannot be IL-8 in response to bacterial infection predominately from the ba- explained simply by destruction of A549 cell barriers due to in- solateral surface of the A549 monolayers. fection with these [high] PA, since barrier properties and cell vi- Previous studies from our group have demonstrated that PMN ability remained intact under these infection conditions. Several migration across intestinal epithelial cell monolayers infected with possibilities might explain this result. For example, PA can par- S. typhimurium requires PKC (25). Furthermore, we have recently ticipate in quorum sensing, a process in which the density of or- demonstrated that the chemoattractant responsible for S. typhi- ganisms in a given population can alter the behavior of individual murium-induced PMN migration across intestinal epithelial mono- Downloaded from organisms within the population (34). Also, Pseudomonas species layers is the eicosanoid HXA3 and this molecule is produced are capable of forming biofilms when enough bacteria are present, and/or secreted in a PKC-dependent manner (15). This arachidonic and these structures can interact with various surfaces in a unique acid metabolite is one of the products of the 12-LO pathway and manner when compared with interactions by unorganized popula- has been previously shown to stimulate Ca2ϩ release within PMNs tions of bacteria (35). Another possibility is that PA, in contrast to as well as to possess PMN chemotactic activity (13, 16). Likewise, E. coli or K. pneumoniae, has the capacity to constitutively pro- we demonstrate in this report that PA infection of A549 cells also

duce a substance (i.e., inhibitor) that interferes with induction of results in activation of this novel signaling pathway. PKC is acti- http://www.jimmunol.org/ the PMN migration process and is only effective if enough PA is vated within A549 monolayers in response to PA. An inhibitor of present to allow the accumulation of the appropriate amount of the PKC pathway completely blocked bacterial-induced PMN that substance. Future experiments will determine the mecha- transmigration, suggesting that PKC is involved in mediating this nism underlying the inability of [high] PA to promote PMN process. Moreover, analogous to infected intestinal epithelial cells, transmigration. A549 lung epithelial cells are capable of producing HXA3 in re- Irrespective of the unique profiles of individual bacterial strains sponse to bacterial infection. Pretreatment of A549 cells with a and species with respect to the concentration required to induce specific inhibitor of the 12-LO pathway before infection with bac- PMN transmigration, it is clear that several different species that teria resulted in both a reduction in PMN transmigration as well as by guest on September 24, 2021 are relevant to lung inflammation demonstrate the ability to induce a reduction in the amount of HXA3 secreted by the A549 mono- PMN transmigration across lung epithelial cells. This observation layers. In addition, HXA3 acts as a chemoattractant for PMNs led to the question of how this process is orchestrated mechanis- when a gradient is established across A549 monolayers. Thus, lung tically. Several reports have been published demonstrating that PA epithelial cells produce HXA3 in response to bacterial infection and/or products of PA are capable of initiating the secretion of the and this PMN chemoattractant is responsible for guiding PMNs PMN chemokine IL-8 (2, 6, 11, 12). This phenomenon is believed across the lung epithelium. At present it is unclear whether sig- to be important in the PMN recruitment process. Mechanistic stud- naling through PKC is important to initiate production of HXA3 ies of bacterial-induced PMN transepithelial migration response and/or in orchestrating secretion of HXA3 from the apical surface described herein have revealed that although the A549 monolayers of A549 monolayers. Thus, future experiments will further delin- produce IL-8 in response to bacterial infection, it does not appear eate the role of PKC as well as other details of HXA3 production to be responsible for mediating the transepithelial migration of the and secretion. In this regard, it will be important to determine PMNs. This conclusion is based on several observations. First, the whether any of the 12-LO enzymes are specifically activated by majority of IL-8 produced in response to PA01 is secreted into the infection. There are at least four distinct genes that encode 12-LO, basolateral chamber of the A549 monolayers, counter to the di- some of which are expressed in epithelial cells (36, 37). These rection of PMN migration. Second, neutralizing Abs to IL-8 have include platelet-type 12-LO, leukocyte-type 12-LO, epidermal- no effect on bacterial-induced PMN transmigration. This result is type 12-LO, and 12-(R)-LO (36). In addition, 15-lipoxygenase 1 is consistent with a report of the inability of IL-8-neutralizing Abs to capable of synthesizing 12-HpETE as a minor product and has prevent C. pneumoniae-induced PMN transmigration across pri- been shown to be expressed in A549 cells in response to IL-4 and mary human airway epithelial cells (5). Also calling into question IL-13 (38). Alternatively, bacterial infection could simply stimu- the role for IL-8 in mediating PMN transepithelial migration is the late the production of increased amounts of HXA3 precursors such fact that heat killing PA before the addition to A549 monolayers as arachidonic acid, thus providing more substrate for a constitu- results in an IL-8 response that is equal to, if not greater than, live tively active 12-LO enzyme. It is also plausible that both an in- PA infection, whereas heat-killed PA are completely unable to crease in 12-LO activity and an increase in the amount of available induce PMN transepithelial migration. In addition, PA stimulates 12-LO substrate may occur in conjunction. significant IL-8 production at [high] despite the fact that [high] PA It is noteworthy to mention that this novel inflammatory path- are not capable of facilitating PMN transepithelial migration. The way involving PKC, 12-LO, and the production and secretion of

IL-8 response due to infection with PA is mediated in part by the HXA3 appear to occur at multiple mucosal sites (i.e., intestine and MAPK pathway ERK1/2. Inhibitors that sufficiently block the lung) (15). Inhibition of the 12-LO pathway during S. typhimurium ERK1/2 pathway result in a 50–75% reduction in the amount of infection of human intestinal xenografts results in a dramatic re- IL-8 produced in response to PA when A549 monolayers are pre- duction in inflammation and subsequent tissue destruction demon- treated with such inhibitors. ERK1/2 pathway inhibitors do not strating the importance of this pathway in a more complex disease The Journal of Immunology 5719 model (15). This pathway may represent a conserved innate im- References mune mechanism for the detection and eradication of pathogens 1. McIntosh, K. 2002. Community-acquired pneumonia in children. N. Engl. interfacing with host mucosal surfaces. Intestinal epithelial cells J. Med. 346:429. are fairly restrictive in terms of the bacterial stimulus needed to 2. Lyczak, J. B., C. L. Cannon, and G. B. Pier. 2002. Lung infections associated with cystic fibrosis. Clin. Microbiol. Rev. 15:194. trigger HXA3 production. S. typhimurium, through its effector mol- 3. Marrie, T. J., M. J. Fine, D. S. Obrosky, C. Coley, D. E. Singer, and ecule, SipA, activates HXA3 secretion whereas commensal E. coli W.N. Kapoor. 1998. Community-acquired pneumonia due to Escherichia coli. strains such as F-18 (which does not posses SipA) are incapable of Clin. Microbiol. Infect. 4:717. 4. Tateda, K., T. A. Moore, M. W. Newstead, M. C. Tsai, X. Zeng, J. C. Deng, stimulating PMN transmigration across intestinal monolayers (39). G. Chen, R. Reddy, K. Yamaguchi, and T. J. Standiford. 2001. PMN recruitment Initiation of PMN transmigration across lung epithelial cells ap- in a murine model of Legionella infect. Immunity 69:2017. pears to be more promiscuous as several bacterial species (i.e., E. 5. Jahn, H. U., M. Krull, F. N. Wuppermann, A. C. Klucken, S. Rosseau, J. Seybold, J. H. Hegemann, C. A. Jantos, and N. Suttorp. 2000. Infection and activation of coli, K. pneumoniae, and PA) all are capable of inducing this phe- airway epithelial cells by Chlamydia. J. Infect. Dis. 182:1678. nomenon. PMN transmigration across lung epithelial cells has also 6. Pizurki, L., M. A. Morris, M. Chanson, M. Solomon, A. Pavirani, I. Bouchardy, been shown to occur in response to C. pneumoniae (5). The intes- and S. Suter. 2000. CFTR does not affect PMN migration across cystic fibrosis tinal mucosa is colonized by a plethora of commensal bacterial airway epithelial monolayers. Am. J. Pathol. 156:1407. 7. Weiss, S. 1989. Tissue destruction by neutrophils. N. Engl. J. Med. 320:365. organisms, yet does not exist in a constant state of inflammation, 8. Burns, A. R., C. W. Smith, and D. C. Walker. 2003. Unique structural features whereas the lower airway epithelium is not generally considered to that influence neutrophil emigration into the lung. Physiol. Rev. 83:309. be chronically colonized with bacteria except under pathological 9. Wagner, J. G., and R. A. Roth. 2000. Neutrophil migration mechanisms, with an conditions (1, 40). Thus, lung epithelial cells may react to the emphasis on the pulmonary vasculature. Pharmacol. Rev. 52:349. 10. Smart, S. J., and T. B. Casale. 1994. Pulmonary epithelial cells facilitate TNF- Downloaded from presence of bacteria in a more indiscriminant manner or there ␣-induced neutrophil chemotaxis: a role for cytokine networking. J. Immunol. could be dampening responses that occur in the gut (26). Future 152:4087. studies will determine whether E. coli, K. pneumoniae, and PA 11. DiMango, E., H. J. Zar, R. Bryan, and A. Prince. 1995. Diverse Pseudomonas aeruginosa gene products stimulate respiratory epithelial cells to produce inter- share a bacterial trigger for HXA3 production or whether they leukin-8. J. Clin. Invest. 96:2204. posses discreet factor(s) which harness this or perhaps other path- 12. Hybiske, K., J. K. Ichikawa, V. Huang, S. J. Lory, and T. E. Machen. 2004. ways leading to PMN transepithelial migration. Our data thus far Cystic fibrosis airway epithelial cell polarity and bacterial flagellin determine host response to Pseudomonas aeruginosa. Cell Microbiol. 6:49. suggest that at least the PA strain PA01 and the commensal E. coli http://www.jimmunol.org/ 13. Pace-Asciak, C. R., D. Reynaud, P. Demin, and S. Nigam. 1999. The hepoxilins: strain F-18, albeit at different concentrations of infection, initiate a a review. Adv. Exp. Med. Biol. 447:123. common host cell pathway to promote PMN transepithelial 14. Margalit, A., Y. Sofer, S. Grossman, D. Reynaud, C. R. Pace-Asciak, and migration. A.A. Livne. 1993. Hepoxilin A3 is the endogenous lipid mediator opposing hy- Proc. Natl. Acad. Sci. USA 90:2589. A recent review by Burns et al. (8) provides a very thorough and potonic swelling of intact human platelets. 15. Mrsny, R. J., A. T. Gewirtz, D. Siccardi, T. Savage, B. P. Hurley, J. L. Madara, informative breakdown of issues governing the overall PMN re- and B. A. McCormick. 2004. Identification of hepoxilin A3 in inflammatory cruitment process from the capillaries to the airway lumen, which events: a required role in neutrophil migration across intestinal epithelial. Proc. involves transendothelial transmigration, transmatrix migration, Natl. Acad. Sci. USA 101:7421. 16. Sutherland, M., T. Schewe, and S. Nigam. 2000. Biological actions of the free migration through the interstitium adjacent to fibroblasts, and acid of hepoxilin A3 on human neutrophils. Biochem. Pharmacol. 59:435. transepithelial transmigration. Our approach to construct a reduc- 17. Rosseau, S., J. Selhorst, K. Wiechmann, K. Leissner, U. Maus, K. Mayer, by guest on September 24, 2021 tionist model system for a distinct phase in the PMN recruitment F. Grimminger, W. Seeger, and J. Lohmeyer. 2000. Monocyte migration through the alveolar epithelial barrier. J. Immunol. 164:427. process (transepithelial migration) has resulted in the identification 18. Bose, S., A. Malur, and A. K. Banerjee. 2001. Polarity of human parainfluenza of a novel inflammatory pathway, which now provides a testable virus type 3 infection in polarized human lung epithelial A549 cells: role of hypothesis in unraveling the complex multifactored mechanism microfilament and microtubule. J. Virol. 75:1984. that occurs during bacterial-induced PMN recruitment to the air- 19. Foster, K. A., C. G. Oster, M. M. Mayer, M. L. Avery, and K. L. Audus. 1998. Characterization of the A549 cell line as a type II pulmonary epithelial cell model space. Additional experiments using alternative in vitro cell culture for drug metabolism. Exp. Cell Res. 243:359. models and more complex in vivo models will be performed to 20. McCormick, B. A., P. M. Hoffman, J. Kim, D. K. Carnes, S. I. Miller, and address both the prevalence of the HXA secretion process as well J. L. Madara. 1995. Attachment of S. typhimurium to IECs imprints the matrix 3 with gradients chemotactic for PMNs. J. Cell Biol. 131:1599. as to ascertain the specific anatomic location of action for both 21. Hurley, B. P., M. Jacewicz, C. M. Thorpe, L. L. Lincicome, A. J. King, IL-8 and HXA3 in mediating PMN recruitment during the disease G. T. Keusch, and D. W. Acheson. 1999. Shiga toxins 1 and 2 translocate dif- process. ferently across polarized intestinal epithelial cells. Infect. Immun. 67:6670. 22. Hurley, B. P., C. M. Thorpe, and D. W. Acheson. 2001. Shiga toxin translocation In summary, we have presented evidence for a novel inflamma- across intestinal epithelial cells is enhanced by neutrophil transmigration. Infect. tory pathway that occurs within lung epithelial cells. Bacterial in- Immun. 69:6148. teraction with airway epithelial cells activates PKC and results in 23. Kohler, H., S. P. Rodrigues, and B. A. McCormick. 2002. Shigella flexneri in- the production and secretion of HXA , likely through the action of teractions with the basolateral membrane domain of polarized model intestinal 3 epithelium: role of lipopolysaccharide in cell invasion and in activation of the the 12-LO enzymatic pathway. Future studies will be geared to- mitogen-activated protein kinase ERK. Infect. Immun. 70:1150. ward evaluating the relevance of this pathway using in vivo mod- 24. Song, J. C., C. M. Hanson, V. Tsai, O. C. Farokhzad, M. Lotz, and els of lung inflammation. Discovery of this pathway uncovers the J. B. Matthews. 2001. Regulation of epithelial transport and barrier function by distinct protein kinase C isoforms. Am. J. Physiol. 281:C649. potential for a novel therapeutic strategy to modulate the inflam- 25. Criss, A. K., M. Silva, J. E. Casanova, and B. A. McCormick. 2001. Regulation matory process in the lung through specific inhibitors of the 12-LO of Salmonella-induced neutrophil transmigration by epithelial ADP-ribosylation pathway in an effort to minimize the pathological inflammatory factor 6. J. Biol. Chem. 276:48431. 26. Neish, A. S., A. T. Gewirtz, H. Zeng, A. N. Young, M. E. Hobert, V. Karmali, processes that occur during bacterial pneumonia and/or bacterial A. S. Rao, and J. L. Madara. 2000. Prokaryotic regulation of epithelial responses infections associated with cystic fibrosis. by inhibition of I␬B-␣ ubiquitination. Science 289:1560. 27. Didierlaurent, A., J. C. Sirard, J. P. Kraehenbuhl, and M. R. Neutra. 2002. How the gut senses its content. Cell Microbiol. 4:61. Acknowledgments 28. Schroeder, T. H., M. M. Lee, P. W. Yacono, C. L. Cannon, A. A. Gerceker, We thank the following individuals for their expert technical assistance: D. E. Golan, and G. B. Pier. 2002. CFTR is a pattern recognition molecule that Mike Pazos, Natecia Williams, William J. Nadeau, Maytinee Lilaonitkul, extracts Pseudomonas aeruginosa LPS from the outer membrane into epithelial cells and activates NF-␬B translocation. Proc. Natl. Acad. Sci. USA 99:6907. and Milton Silva. We also thank Dr. Stephen Lory and Lisa Tatterson for 29. Ballif, B. A., and J. Blenis. 2001. Molecular mechanisms mediating mammalian helpful discussions. We thank Karen L. Mumy for critical review of this mitogen-activated protein kinase (MAPK) kinase (MEK)-MAPK cell survival manuscript. signals. Cell Growth Differ. 12:397. 5720 HXA3 MEDIATES BACTERIAL-INDUCED PMN TRANSMIGRATION

30. Herbert, J. M., J. M. Augereau, J. Gleye, and J. P. Maffrand. 1990. Chelerythrine 2002. Pseudomonas aeruginosa anaerobic respiration in biofilms: relationships to is a potent and specific inhibitor of protein kinase C. Biochem. Biophys. Res. cystic fibrosis pathogenesis. Dev. Cell 3:593. Commun. 172:993. 36. Yamamoto, S., H. Suzuki, M. Nakamura, and K. Ishimura. 1999. Arachidonate 31. Arora, J. K., T. W. Lysz, and P. S. Zelenka. 1996. A role for 12(S)-HETE in the 12-lipoxygenase isozymes. Adv. Exp. Med. Biol. 447:37. response of human lens epithelial cells to epidermal growth factor and insulin. 37. Shannon, V. R., W. F. Stenson, and M. J. Holtzman. 1993. Induction of epithelial Invest. Ophthalmol. Vis. Sci. 37:1411. arachidonate 12-lipoxygenase at active sites of inflammatory bowel disease. 32. Alanko, J., A. Riutta, and H. Vapaatalo. 1992. Effects of catecholamines on Am. J. Physiol. 264:G104. eicosanoid synthesis with special reference to prostanoid/ ratio. Free 38. Brinckmann, R., M. S. Topp, I. Zalan, D. Heydeck, P. Ludwig, H. Kuhn, Radical Biol. Med. 13:677. W. E. Berdel, and J. R. Habenicht. 1996. Regulation of 15-lipoxygenase expres- 33. Mizgerd, J. P. 2002. Molecular mechanisms of neutrophil recruitment elicited by sion in lung epithelial cells by interleukin-4. Biochem J. 318:305. bacteria in the lungs. Semin. Immunol. 14:123. 39. Lee, C. A., M. Silva, A. M. Siber, A. J. Kelly, E. Galyov, and B. A. McCormick. 34. Chun, C. K., E. A. Ozer, M. J. Welsh, J. Zabner, and E. P. Greenberg. 2004. 2000. A secreted Salmonella protein induces a proinflammatory response in ep- Inactivation of a Pseudomonas aeruginosa quorum-sensing signal by human air- ithelial cells, which promotes neutrophil migration. Proc. Natl. Acad. Sci. USA way epithelia. Proc. Natl. Acad. Sci. USA 101:3587. 97:12283. 35. Yoon, S. S., R. F. Hennigan, G. M. Hilliard, U. A. Ochsner, K. Parvatiyar, 40. Hooper, L. V. 2004. Bacterial contributions to mammalian gut development. M. C. Kamani, H. L. Allen, T. R. DeKievit, P. R. Gardner, U. Schwab, et al. Trends Microbiol. 12:129. Downloaded from http://www.jimmunol.org/ by guest on September 24, 2021