of May 28, 2022.This information is current as

LevelsTargets Inflammatory Responses at Multiple ELR-CXC Chemokine Receptor Antagonism

Donald W. Cockcroft and John R. GordonXixing Zhao, Jennifer R. Town, Fang Li, Xiaobei Zhang,

http://www.jimmunol.org/content/182/5/3213doi: 10.4049/jimmunol.0800551

2009; 182:3213-3222; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/182/5/3213.full#ref-list-1

, 19 of which you can access for free at: cites 60 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2009 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on May 28, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on May 28, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

ELR-CXC Chemokine Receptor Antagonism TargetsInflammatory Responses at Multiple Levels1

Xixing Zhao,* Jennifer R. Town,* Fang Li,† Xiaobei Zhang,* Donald W. Cockcroft,‡

and John R. Gordon2*‡

The ELR-CXC chemokines play important roles in neutrophilic inflammation. We report in this study that a fully humanELR-CXC chemokine antagonist that we have generated, CXCL8(3–72)K11R/G31P (G31P), has potent anti-inflammatory effectsthat arise through its actions at multiple levels. G31P inhibited CXCL8-induced chemotactic responses and intracellular Ca2� fluxin CXCR1-transfected HEK cells and neutrophils, and responses of neutrophils to CXCR2-exclusive ligands. G31P desensitizedheterologous G protein-coupled receptors on neutrophils, 52–86% reducing their Ca2� flux and chemotactic responses to leuko-triene B4, C5a, and the bacterial tripeptide fMLP. G31P also 60–90% blocked neutrophil chemotactic responses to mediatorspresent in 10 of 12 sputum samples from cystic fibrosis or bronchiectasis subjects with bacterial pneumonia. Moreover, whereasA549 bronchial epithelial cells (which expressed CXCR1) secreted �29,000 pg/ml CXCL8 in response to in vitro endotoxinchallenge, G31P reduced this response by up to 98%, presumably by interrupting an autocrine inflammatory loop. The anti-inflammatory effects of G31P extended also to reversing the antiapoptotic influence of ELR-CXC chemokines on neutrophils. Thatthese effects were relevant in vivo was confirmed in a guinea pig model of airway endotoxemia, wherein the human form of G31P>95% blocked neutrophil infiltration into and activation within the airways, as determined by airway levels of the neutrophilprimary, secondary, and tertiary granule markers myeloperoxidase, lactoferrin, and matrix metalloproteinase-9, respectively, andthe epithelial cell marker matrix metalloproteinase-2. These data suggest that the beneficial effects of ELR-CXC chemokineantagonism arise through effects that occur at multiple levels, including epithelial cells, neutrophils, and alternate G protein-coupled receptors. The Journal of Immunology, 2009, 182: 3213–3222.

N eutrophil responses are a critical element in host defenseduring bacterial infections, for example, but neutrophilresponses can also be overtly pathogenic. Thus, during

many inflammatory diseases, neutrophils contribute more to thepathology than do the microbes themselves (1). Neutrophil recruit-ment is a complex process involving activation of local structuralcells (e.g., epithelial cells) via pathogen-associated molecular pat-tern receptors, complement cascade products, or arachidonic acidmetabolites, such that the structural cells express inflammatorymediators (e.g., IL-1, CXCL8) (2). These can in turn activate re-gional endothelial cells, which actively foster intravascular neu-trophil retention via the now classical paradigm of chemokine(e.g., CXCL8)- and adhesion molecule-mediated rolling-arrest di-apedesis (3, 4).

The ELR-CXC chemokines are a subgroup of the CXC chemo-kine family in which the amino subterminal two cysteine residues,which are separated by an alternate amino acid, are immediatelypreceded by a Glu-Leu-Arg motif. They include CXCL1–3 and

5–8 (growth-related oncogene �, �, and �; epithelial cell neutro-phil-activating peptide-78; granulocyte chemotactic protein-2;neutrophil-activating peptide-2; and IL-8, respectively) (3), whichchemoattract and activate neutrophils via two closely related Gprotein-coupled receptors (GPCR),3 the CXCR1 and CXCR2.CXCL8 binds both receptors with high affinity, whereas CXCL6binds both receptors with lower affinity (5, 6). The other ELR-CXC chemokines bind to the CXCR2, also with relatively loweraffinities (6, 7). Both the CXCR1 and CXCR2 can trigger chemo-tactic responses and intracellular Ca2� flux in neutrophils and con-tribute to elastase release (6, 8), but activation of the respiratoryburst and phospholipase D release responses is reportedly CXCR1dependent (9), whereas CXCR2 signaling is critical to matrix met-alloproteinase (MMP)-9 release (10). Moreover, these two recep-tors may be differentially involved in neutrophilic pathology invivo. For example, the CXCR1 is reportedly of more importance ininflammatory bowel diseases (11), sepsis, and acute respiratorydistress syndrome (12, 13), whereas both it and the CXCR2 playroles in synovial infiltration by neutrophils in arthritic joints (14).However, numerous alternate GPCR are involved in neutrophilicinflammation, including those for leukotriene B4 (LTB4), C5a, andfMLP, such that antagonism of either LTB4 or C5a has been shownto be of significant benefit in various inflammatory settings (15–17). The precise interrelationships of these mediators in inflam-matory responses have not been formally determined, but it has

*Department of Veterinary Microbiology, University of Saskatchewan, Saskatoon,Canada; †Department of Immunology, Dalian Medical University, Dalian, China; and‡Division of Respirology, Critical Care and Sleep Medicine, Royal University Hos-pital, University of Saskatchewan, Saskatoon, Canada

Received for publication February 19, 2008. Accepted for publication December20, 2008.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by grants from the Natural Sciences and EngineeringResearch Council of Canada and the Canadian Institutes of Health Research (toJ.R.G.), and by a research contract from Pacgen Biopharmaceuticals.2 Address correspondence and reprint requests to Dr. John R. Gordon, Division ofRespirology, Critical Care and Sleep Medicine, Room 3610, Royal University Hos-pital, 103 Hospital Drive, Box 120 R.U.H., Saskatoon, SK, Canada S7N 0W8. E-mailaddress: john.gordon@usask.ca

3 Abbreviations used in this paper: GPCR, G protein-coupled receptor; BAL, bron-choalveolar lavage fluid; bG31P, bovine G31P; CF, cystic fibrosis; COPD, chronicobstructive pulmonary disease; LF, lactoferrin; G31P, human CXCL8(3–72)K11R/G31P; LTB4, leukotriene B4; MMP, matrix metalloproteinase; MPO, myeloperoxi-dase; ROI, reactive oxygen intermediate; WBC, white blood cell.

Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00

The Journal of Immunology

www.jimmunol.org/cgi/doi/10.4049/jimmunol.0800551

by guest on May 28, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

been reported that signaling through the C5a or fMLP receptorscan effectively desensitize the ELR-CXC chemokine receptors(18). In contrast, CXCL8 reportedly poorly desensitizes someevents (e.g., intracellular Ca2� flux) associated with C5a andfMLP receptor signaling (19, 20), although it can desensitize che-motactic responses driven by these ligands (18).

Some time ago, we developed a very high-affinity broad-spec-trum ELR-CXC chemokine antagonist, bovine CXCL8(3–74)K11R/G31P (bG31P) (21, 22), based on a high-affinity bovine CXCL8analog, CXCL8(3–74)K11R (23). This antagonist could blockCXCL8 binding to neutrophils, as well as neutrophil activation byagonists present in wash fluids from the lungs of cattle with bac-terial pneumonia. bG31P fully antagonizes neutrophil responses toendotoxin challenge of cattle for more than 2 days after a singlelow-dose treatment (22), and blocks pyrexia, hemorrhagic consol-idation, and pulmonary neutrophilia in airway endotoxemia inguinea pigs even if delivered after the onset of symptoms (24). Wealso subsequently developed numerous human-bovine chimericforms of bG31P that were as effective as bG31P in blocking ELR-CXC chemokine-mediated pathology (25), and then developed afully human form of G31P. However, some of the protective ef-fects of bG31P in airway endotoxemia (24), for example, seemedto be somewhat tangentially related to its putative neutrophil-cen-tered effects. For example, bG31P treatment reduces endogenouspyrogen expression and pyrexia before the time when neutrophilswere appreciably present in the airways, suggesting that it mayhave effects on structural (e.g., epithelial) cells. Moreover, datafrom a study of the effects of G31P in aspiration pneumonia,wherein high-level bacterial colonization of the lungs occurs,confirmed its effectiveness in this environment as well (X.Zhao, J. Town, and J. Gordon, submitted for publication). Insuch cases, the bacteria would release formyl peptides (e.g.,fMLP) among other products and activate the complement cas-cade, leading to the generation of C5a.Thus, we wished to ex-plore further the mechanisms by which G31P might interactwith airway epithelial cells, but also its effect on neutrophilresponses to ligands for heterologous GPCR.

Materials and MethodsReagents and supplies

The following reagents were purchased commercially: restriction endo-nucleases BamHI and EcoRI, T4 DNA ligase, DNA polymerase, fluo-4AM, MEM (with L-glutamine), HEPES, lipofectamine 2000, and MEMnonessential amino acid solution (Invitrogen); QIAprep Spin Miniprep,QIAquick Gel Extraction Kits, QIA shredder, and RNeasy Mini Kit (Qia-gen); QuickChange Site-Directed Mutagenesis Kit and Brilliant qRT-PCRMaster Mix Kit (Stratagene Cloning Systems); glutathione-Sepharose, theexpression vector pGEX-2T, and benzamidine-Sepharose (Amersham Bio-sciences); aprotinin, dextran, endotoxin (Escherichia coli LPS, serotype0127:B8), hexadecyltrimethylammonium bromide, gelatin, isopropyl-thio-D-galactopyranoside, luminol, phenol red-free HBSS, PMSF, tetramethyl-benzidine, Triton X-100, and DTT (Sigma-Aldrich); paired anti-CXCL8capture and detection Abs and rCXCL8 (R&D Systems); rabbit anti-humanlactoferrin Ab, recombinant lactoferrin, and biotinylated anti-lactoferrinAb (ICN Biomedicals); Lymphocyte Separation Medium density gradientmedium (MP Biomedicals); Diff-Quick stain kit (American Scientific Prod-ucts); annexin V/propidium iodide staining kit (BD Biosciences); and5-wk-old female Hartley guinea pigs (Charles River Laboratories). All ofthe experiments were conducted according to the guidelines established bythe Canada Council on Animal Care, and were approved by University ofSaskatchewan animal ethics review panel.

Generation of G31P

A full-length human CXCL8 cDNA was synthesized commercially(Takara Biotech). From this, we generated a GST fusion protein constructencoding aa 3–72 of human CXCL8, then introduced a Lys11-to-Arg sub-stitution (we reported previously that bovine CXCL8(3–72)K11R is a veryhigh-affinity CXCR1/CXCR2 agonist) (23). We next introduced a Gly31 to

Pro substitution, to generate human CXCL8(3–72)K11R/G31P (G31P).Each plasmid was transformed into competent HB101 cells, and their se-quences were verified commercially (Plant Biotechnology Institute). Themethods for expressing the constructs, cleaving the target proteins fromtheir GST fusion partners, and purifying these have been described in detail(21, 22). Briefly, the recombinant bacteria were sonicated in the presenceof the protease inhibitors aprotinin (2 mg/ml) and PMSF (50 mM), and thenthe rGST fusion proteins present in the sonicates were purified by affinitychromatography using glutathione-Sepharose columns. The rCXCL8 ana-logues were cleaved from their GST fusion partners by thrombin digestion andthen further purified using benzamidine-Sepharose columns. The relative mo-lecular mass and purity of each protein were verified by SDS-PAGE and West-ern blotting with biotinylated anti-human CXCL8 Ab (21, 23).

Generation of CXCR1-transfected HEK293 cells

HEK293 cells (CRL-1573; American Type Culture Collection) were trans-fected with the gene for the CXCR1 in the vector pCDNA3.1� (MissouriS&T cDNA resource center; www.cdna.org) using lipofectamine. TheCXCR1-expressing cells were selected using gentamicin (800 �g/ml), andtheir expression of the CXCR1 gene was confirmed by FACS and quan-titative real-time PCR.

Chemotaxis assays

Chemotaxis was assessed using modified Boyden chamber microchemo-taxis assays (21–23) with either purified neutrophils or CXCR1-transfectedHEK cells. For the former, human peripheral blood leukocytes were frac-tioned on standard density gradients, and the neutrophils were harvestedfrom the bottom of the gradients and cleared of contaminating RBC byhypotonic lysis. The purified neutrophils (or HEK cells) were suspended at2 � 106/ml in PBS� (pH 7.4) (1.2 mM MgCl2, 5 mM KCl, 0.5 mM CaCl2,5 mM glucose, and 0.1% BSA). The chemoattractants (e.g., CXCL8), ei-ther alone or together with G31P, were placed in the bottom compartmentof the Boyden chamber wells and cells in the upper compartment, with thetwo compartments separated by polyvinylpyrrolidone-free, 5 (neutrophils)-or 10 (CXCR1-HEK cells)-�m pore-size polycarbonate filters. After incu-bation for 20 min (neutrophils) or 5 h (CXCR1-HEK cells) at 37°C in a 5%CO2 atmosphere, the cells that had migrated into the filters were fixed andstained using a Diff-Quick kit. The numbers of cells responding in eachwell were enumerated by direct counting of at least five �40 objectivefields, and the results were expressed as the mean number of cells per �40field � SEM.

In a series of preliminary experiments, we confirmed that CXCL1,CXCL5, and CXCL8 induced maximal neutrophil chemotactic responses atconcentrations of 100, 100, and 10 ng/ml, respectively, whereas CXCL8induced maximal CXCR1-HEK cell responses at 100 ng/ml. We also op-timized the dilution of each sputum sample (i.e., 1:1 to 1:100) required tomaximize its neutrophil chemotactic activity. We tested the effects on spu-tum-induced chemotactic responses of using G31P at 0.1–100 ng/ml, andfound that the antagonistic effects observed at 10 ng/ml were not altered byincreasing the dose of G31P (data not shown).

Reactive oxygen intermediate (ROI) release assay

We used total white blood cell (WBC) preparations for the ROI releaseassays. To generate WBC, heparin-anticoagulated human blood was mixedwith an equal volume of 6% (w/v) dextran (75 kDa) in PBS, and then thetubes were allowed to stand upright and undisturbed for �1 h at roomtemperature. When the RBC had sedimented, the upper phase, comprisingplasma, platelets, and WBC, was centrifuged and the cell pellet was re-suspended at 2.5 � 106 cells/ml HBSS. For the assays, we mixed 50 �l ofWBC, 25 �l of human CXCL8 (final concentration, 100 ng/ml) or HBSS,25 �l of HBSS or the indicated concentrations of G31P, and 50 �l ofluminol (0.3 mM). We had confirmed in preliminary experiments that 100ng/ml CXCL8 was optimal for induction of WBC ROI release, as deter-mined by chemiluminescence using a microplate spectrofluorimeter(NovoStar; BMG LABTECH). The data are expressed as the maximumsample luminescence minus that of cells exposed to HBSS alone.

Assay of intracellular Ca2� flux

Neutrophils were purified, as described above, and then washed twice withCa2�-free PBS� medium and suspended at 5 � 106 cells/ml in the samemedium. The cells were stained for 30 min at 37°C with 2 �M fluo-4 AM,and then washed twice with Ca2�-free PBS�, resuspended at 4 � 106

cells/ml in PBS� containing 3.3 nM Ca2�, and kept at room temperatureuntil used for testing. Before challenge, the cells were incubated for 15 minwith medium or G31P (100 ng/ml), and then challenged with medium orhuman CXCL1, CXCL5, or CXCL8 (each, 100 ng/ml). In preliminary

3214 ELR-CXC CHEMOKINE ANTAGONISM

by guest on May 28, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

experiments, we had confirmed that 100 ng/ml induced optimal intra-cellular Ca2� flux in these cells, as determined using a microplate spec-trofluorimeter (emission wavelength, 520 nm; excitation wavelength,488 nm).

Heterologous desensitization of GPCR

Neutrophils were purified, as above, resuspended in PBS� medium at 2 �106 cells/ml, and then stimulated with 0.1 nM C5a, 1 nM fMLP, or 1 ng/mlLTB4 in the presence of 0, 10, or 50 ng/ml G31P. The impact of the G31Pon the agonist-induced intracellular Ca2� flux and chemotactic responseswas measured, as noted above. These experiments were repeated threetimes.

Sputum samples

Sputum samples were collected for diagnostic purposes from 12 cysticfibrosis (CF) or bronchiectasis patients with bacterial pneumonia. Sputafrom two subjects with chronic obstructive pulmonary disease (COPD) orasthma were also obtained as controls. For processing, the sputum mucusplugs were dispersed by treatment with 1% DTT/HBSS, and then the cellswere sedimented and the supernatants were dialyzed extensively againstmultiple changes of PBS, as noted previously (26–28). The levels ofCXCL8 in each sample were determined by ELISA, as noted (26). Thesupernatants were aliquoted and stored at �80°C. Sputum from healthydonors did not induce any significant neutrophil chemotactic responses(data not shown).

Effects of G31P on responses of A549 bronchial epithelial cells

Human A549 bronchial epithelial cells (American Type Culture CollectionCCL185) were grown to a maximum of 90% confluence in 24-well platesin MEM (with L-glutamine) supplemented with 10% FBS, 0.01 M HEPES,and nonessential amino acids. Before stimulation, the cell monolayers werewashed and provided with fresh medium, and then exposed for 15 min atroom temperature to 1–100 ng/ml G31P. The indicated doses of LPS or0–1.28 � 106 unstimulated purified neutrophils from healthy donors wereadded to the cells, which were incubated for 16 h at 37°C before thesupernatants were collected. Total RNA was collected from the cells usingcommercial extraction kits and standard approaches. ELISA were used toquantify cytokine levels in the supernatant fluids, whereas quantitative real-time PCR was used to quantify the relative mRNA levels.

ELISA of cytokines and chemokines

Our ELISA protocols have been reported in detail previously (26). Culturesupernatants were not diluted before the assays, whereas bronchoalveolarlavage (BAL) fluids for lactoferrin assays were diluted 1/10. The data arepresented as pg/ml or ng/ml, based on recombinant protein standard curves.The cytokine ELISAs were sensitive to 5–10 pg/ml, whereas the lactoferrinassay was sensitive to 5 ng/ml. Our CXCL8 ELISA does not detect G31P,because the capture Ab does not recognize this molecule (J. R. Gordon,unpublished observation).

Quantitative real-time PCR

Total cellular RNA was quantified spectrophotometrically and stored at�80°C. For quantitative real-time PCR, we used a commercial one-stepmaster mix kit, with the thermal profile programmed as follows: first seg-ment, 50°C for 30 min, and 95°C for 10 min (1 cycle); second segment,95°C for 30 s, 62°C for 30 s, and 72°C for 30 s (40 cycles); and thirdsegment, the reading was taken at 72°C during the 30-s plateau. The se-quences of the quantitative real-time PCR primers were as follows:GAPDH forward, 5�-TGC CTC CTG CAC CAC CAA C-3� and reverse,5�-GGC CAT CCA CAG TCT TCT GG-3�; CXCL1 forward, 5�-CCCAAG AAC ATC CAA AGT GTG A-3� and CXCL1 reverse, 5�-GTC ACTGTT CAG CAT CTT TTC G-3�.

Neutrophil apoptosis

Human neutrophils were isolated, as described above, resuspended at 5 �106 cells/ml in PBS�, and then aliquoted into 96-well plates at 100 �l/well.Medium (PBS�) alone or containing G31P (100 ng/ml) was added to thecells and incubated for 15 min at room temperature, before addition ofCXCL1 or CXCL8 (100 ng/ml) or carrier medium alone. After 24 h at37°C, the cells were stained as per the manufacturer’s instructions usingFITC-labeled annexin V and propidium iodide for FACS detection of ap-optotic and necrotic cells, respectively, and then analyzed within 1 h usinga FACScan flow cytometer.

In vivo confirmation of the anti-inflammatory activities of G31P

The assessment of the effect of G31P on 15-h airway endotoxemia pathol-ogy in guinea pigs has been described previously for the bovine and hu-man-bovine chimeric forms of G31P (24, 25). Briefly, for challenge theanimals (n � 5) were given a 200-�l bolus of LPS (5 �g/kg) in sterilesaline by intranasal intubation. At 30 min before challenge, 250 �g/kgG31P or an equivalent volume (1 ml) of saline was given s.c., and at 4, 8,or 16 h postchallenge the animals were euthanized with halothane. BALwas collected using 2.0-ml vol of sterile saline. BAL total WBC and RBCwere enumerated by direct counting, and the results were expressed as themean number of cells per BAL sample (�SEM). BAL cell differentialswere determined using Wright’s solution-stained BAL cell cytospin prep-arations, and the total numbers of neutrophils were calculated using thesedata. All of the data are expressed as means � SEM. All BAL fluids wereassessed independently (either immediately or after storage at �80°C) formyeloperoxidase (MPO), MMP-9, MMP-2, and lactoferrin (LF). This ex-periment was repeated twice.

MPO assay

BAL fluids (10 �l) were diluted 10-fold with 0.5% hexadecyltrimethyl-ammonium bromide and incubated at room temperature for 10 min with 10vol of tetramethylbenzidine, before stopping the reactions with 10 vol of 1M phosphoric acid. The reactions were quantified by spectroscopy (OD,450 nm), and the data were expressed as the mean OD450� SEM. Eachsample was assayed in triplicate.

Assessment of BAL MMP-9 and MMP-2 levels

BAL levels of MMP-2 and MMP-9 were measured by gel zymography, aspreviously described (10). Briefly, 20-�l samples of BAL sample were runon 7.5% SDS-PAGE gels containing 1 mg/ml gelatin. The gels were in-cubated overnight at room temperature in 2.5% Triton X-100, 50 mMTris-HCl, 10 mM calcium chloride, and 1 mM zinc chloride; rinsed indeionized water for 30 min; and then incubated overnight at 37°C in 50mM Tris-HCl, 10 mM calcium chloride, and 1 mM zinc chloride, beforestaining with Coomassie brilliant blue. The identities of the gelatinolyticMMP-2 and MMP-9 bands were determined based on their molecularweights and comparison with standards prepared from homogenates ofuntreated HT1080 fibrosarcoma cells (American Type Culture CollectionCCL21) (29). The results are expressed as relative densitometry units foreach gelatinase band.

Statistical analysis

Two-group comparisons were made using Student’s t tests (two tailed). Weused Spearman Rank correlation analyses to assess the correlations be-tween the sputum cytokine levels and their associated responses. All resultsare expressed as the mean � SEM.

ResultsHuman CXCL8(3–72)K11R/G31P effectively antagonizes bothCXCR1- and CXCR2-dependent activation of neutrophils

To confirm unequivocally whether human CXCL8(3–72)K11R/G31P (G31P) acts at least in part through the CXCR1, we usedCXCR1-transfected HEK293 cells. We confirmed by FACS (datanot shown) and quantitative real-time PCR (Fig. 1A) that thesecells express this receptor, but not the CXCR2 (data not shown),and that 100 ng/ml CXCL8 induced chemotactic and intracellularCa2� flux responses. At 10 ng/ml, G31P �80% blocked CXCL8-induced chemotactic responses, whereas at 100 ng/ml (but not 10ng/ml; data not shown) it dramatically reduced intracellular Ca2�

flux (Fig. 1A). Given the biological differences between CXCR1-transfected HEK cells and neutrophils, we also confirmed withfreshly purified human neutrophils that G31P blocked CXCL8 (10ng/ml)-induced chemotaxis and ROI release (Fig. 1B). It is note-worthy that neutrophil superoxide release is reportedly attributableexclusively to CXCR1 signaling (15). We also found that 10 ng/mlG31P markedly attenuated intracellular Ca2� flux induced by 100ng/ml CXCL8 (Fig. 1B). We then assessed the ability of G31P toblock activation of neutrophils induced by CXCR2-exclusive li-gands. As with CXCL8, G31P dose dependently blocked chemo-tactic responses of neutrophils to CXCL1 and CXCL5 (both 100

3215The Journal of Immunology

by guest on May 28, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

ng/ml), whereas 10 ng/ml G31P �95% reduced intracellular Ca2�

flux induced by an optimal dose (100 ng/ml) of either ligand (Fig.1C). It is noteworthy that, overall, CXCL8 was a substantiallystronger agonist than either CXCL1 or CXCL5 in terms of induc-ing this intracellular Ca2� response.

G31P induces heterologous desensitization of alternate GPCR

Numerous neutrophil agonists, including C5a, LTB4, ELR-CXCchemokines, and fMLP, signal into these cells via distinct GPCR.

It has been shown previously that signaling through C5a, fMLP, orCXCL8 can reciprocally desensitize neutrophil chemotactic re-sponses to subsequently introduced heterologous ligands (18). Incontrast, CXCL8 reportedly does not effectively desensitize C5a-or fMLP-induced neutrophil activation as assessed by intracellularCa2� flux (19, 20). In assessing the relative capacity of G31P toaffect heterologous GPCR signaling, it is important to recognizethat G31P has a much higher affinity for its receptors than do theELR-CXC chemokines. Thus, bG31P inhibits by 50% the chemo-tactic responses of neutrophils to a �250-fold higher concentrationof native ligand (i.e., CXCL8) (22). We wished to assess then theextent to which G31P would desensitize neutrophil responses toligands such as C5a, fMLP, and LTB4, each of which are relevantneutrophil agonists in inflammatory conditions. Thus, we exam-ined the impact of G31P on intracellular Ca2� flux and chemo-tactic responses induced by each of these agonists (Fig. 2). Wefound that 0.1 nM C5a induced marked Ca2� flux in neutrophils,as well as chemotactic responses, and that the addition of G31Preduced C5a-mediated chemotaxis by �50%, although its impacton intracellular Ca2� flux was substantially more modest. The bac-terial tripeptide fMLP induced a much more robust Ca2� flux re-sponse than did C5a, but an essentially equivalent chemotacticresponse. The G31P treatment very substantially reduced both theCa2� and chemotactic (85% reduction) responses. LTB4 was astrong neutrophil agonist in both assays and, as with fMLP, G31Psubstantially reduced LTB4-induced Ca2� flux and chemotactic(�67% reduction) responses. Taken together, these data show thatG31P did dampen neutrophil responses to each of these heterolo-gous GCPR ligands.

G31P effectively antagonizes the inflammatory mediatorspresent in sputum from bacterial pneumonia patients

As suggested above, inflammatory diseases are marked by thesimultaneous expression of numerous neutrophil agonists. Spu-tum from CF patients, for example, can contain CXCL8 andother CXC chemokines, LTB4, formyl peptides (e.g., fMLP),and perhaps C5a (30). Thus, given our data documenting thatG31P can variably desensitize these receptors, we wished toknow just how effective it would be in blocking the agonistspresent in inflammatory samples. Thus, we obtained a bank ofsputa from 12 CF and bronchiectasis patients being treated forbacterial pneumonia and directly tested the impact of G31P ontheir neutrophil chemotactic activities (Fig. 2). The disease sta-tus of the CF donors was classified as advanced, severe, mod-erate, or mild, and that of the bronchiectasis donors as severe ormoderate, or frank bacterial pneumonia, based on clinical cri-teria. We included sputum from one asthmatic and one COPDsubject as controls. We assessed the levels of CXCL8 in eachsputum sample (Fig. 3; values above the sample bars) and foundthat these correlated positively with the ability of the samples toinduce neutrophil chemotaxis (r � 0.63 (14 degrees of free-dom), p � 0.013). At 10 ng/ml G31P, �60 –95% inhibited thechemotactic responses of neutrophils from healthy blood donorsto the sputum from 10 of the 12 CF and bronchiectasis donors.Only sputum from the two donors classified as having advancedCF (donors 1 and 2) appeared more or less resistant to theantagonistic effects of G31P in this assay. Interestingly, therelative ability of G31P to block sputum-induced neutrophilchemotactic activities (i.e., percentage of inhibition) was notsignificantly correlated with the levels of CXCL8 in the samples(r � 0.369 (14 df), p � 0.15). For example, sample 1 contained3.09 ng/ml CXCL8 and was only modestly affected by G31P,whereas sample 10 contained 3.32 ng/ml CXCL8 and its activ-ity was almost ablated by the G31P treatment. The sputum from

FIGURE 1. The ELR-CXC chemokine antagonist CXCL8(3–72)K11R/G31P operates by targeting the CXCR1 and CXCR2 on neutrophils. A, Wegenerated CXCR1-transfected HEK293 cells and used these to determinewhether G31P acts via the CXCR1. We used quantitative real-time PCR toassess CXCR1 expression by control and CXCR1-transfected HEK cellsand compared this expression with that of peripheral blood neutrophils (leftpanel). For the chemotaxis assays, responses to an optimized concentrationof CXCL8 (10 ng/ml) were assessed using modified Boyden chamber mi-crochemotaxis assays, as noted in Materials and Methods. The ability ofG31P (10 ng/ml) to antagonize this response was assessed by simulta-neously exposing the cells to the agonist and antagonist. The results areexpressed as the mean (�SEM) number of cells/�40 objective microscopefield (middle panel). For the Ca2� flux assays, neutrophils were labeledwith the Ca2� indicator dye fluo-4 AM and assayed in the presence (100ng/ml) or absence of G31P for chemokine (100 ng/ml)-induced intracel-lular Ca2� flux, as described in Materials and Methods (right panel). B, Wealso assessed whether G31P could antagonize CXCL8-dependent chemo-taxis, ROI release by leukocytes, and intracellular Ca2� flux responses bypurified human neutrophils using the same approaches as in A. C, To con-firm that G31P also acted via the CXCR2, we tested its ability to antago-nize neutrophil chemotactic and intracellular Ca2� flux responses inducedby the CXCR2-exclusive ligands CXCL1 and CXCL5 (both 100 ng/ml).Our data confirm that G31P did effectively antagonize both CXCR1- andCXCR2-dependent responses. � and ��, p � 0.05 and 0.01, respectively,relative to the chemokine treatments alone. The data shown are represen-tative of at least three experiments performed with similar results. Thedashed lines across graphs in A–C represent the mean background responsein the indicated assays.

3216 ELR-CXC CHEMOKINE ANTAGONISM

by guest on May 28, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

the asthmatic and COPD subjects also induced marked neutro-phil responses, and G31P largely blocked these too; we havereported previously that sputum from asthmatic subjects con-tains high levels of CXCL8 (27).

G31P blocks ELR-CXC chemokine-mediated antiapoptoticeffects in neutrophils

As neutrophils leave the vasculature or become effete, they rapidlyundergo programmed cell death or apoptosis as a means of pre-venting adventitious host pathology. However, during inflamma-tory events, it is in the best interest of the host to maintain these

cells in a viable state such that they can effectively engage micro-bial targets. An array of inflammatory mediators, including theELR-CXC chemokines, have antiapoptotic effects on neutrophils(31). We wished to know then whether G31P’s anti-inflammatoryeffects would extend to reversing chemokine-induced antiapoptoticprocesses, thereby allowing neutrophils to quickly progress intoapoptosis. We used FITC-annexin V binding as our assay of ap-optosis. We stained freshly purified neutrophils with this marker asa negative control (shaded signal; Fig. 4) and cells that had beencultured for 18 h in medium alone as an apoptosis-positive control(Fig. 4, upper panel). The freshly purified cells appeared to havetwo levels of annexin binding, with most cells being more or lessnegative and a smaller proportion binding moderate levels of thisapoptosis marker. When we added CXCL8 to the cells (Fig. 4,middle panel, dashed line), there was a marked diminution of thenumbers of cells that strongly bound annexin V, and the additionof G31P to these cultures fully reversed this CXCL8-induced an-tiapoptotic activity. CXCL1 also had protective effects on neutro-phils, but they were subtly different from those of CXCL8 (Fig. 4,bottom panel, dashed line). The CXCL8-protected cells hadlargely bound intermediate levels of annexin V, whereas thoseincubated with CXCL1 bound either very low or intermediate lev-els of annexin. The addition of G31P to the CXCL1 cultures in-duced a substantial proportion of the cells to up-regulate their ap-optotic processes, but nevertheless there remained a significantpopulation of cells that bound intermediate levels of annexin.

G31P antagonizes the inflammatory cycle inendotoxin-challenged bronchial epithelial cells

Exposure of the airway epithelium to bacterial endotoxin triggersa pulmonary inflammatory response, wherein these cells secreteCXCL8, for example, which directly fosters neutrophil recruitmentinto this tissue compartment. Although we have shown that G31Pantagonizes the impact of the ELR-CXC chemokines on neutrophilresponses, little attention is paid to whether epithelial cells them-selves respond to the ELR-CXC chemokines they elaborate in air-way endotoxemia, or to the neutrophils they recruit. We thus de-cided to investigate this using human type II alveolar A549epithelial cells that we challenged with bacterial endotoxin (Fig. 5)

FIGURE 3. G31P antagonizes the neutrophil chemotactic activitiespresent in sputum from CF and bronchiectasis patients with bacterial pneu-monia. Sputum samples were collected from 12 CF or bronchiectasis pa-tients with bacterial pneumonia as well as two control subjects with COPDor asthma, and assayed for their CXCL8 content (values above f). Theabilities of G31P (10 ng/ml) to antagonize chemotactic responses of neu-trophils from healthy donors to optimized dilutions of the sputa were as-sessed as in Fig. 1. The data are expressed as the mean (�SEM) numberof cells/�40 objective microscope field. The assays were repeated twiceand gave essentially the same results each time. Sputum from healthy do-nors did not induce any significant neutrophil chemotactic responses (datanot shown). With the exception of samples 1 and 2, G31P significantlyreduced the cells’ chemotactic responses. ��, p � 0.01 vs no G31P treat-ment. The dashed lines across the graph represent the mean backgroundresponse in the assay.

12.5 30.5 48.5 66.5 84.5

fMLP agonist

+G31P

12.5 29.5 46.5 63.5 80.5 97.5

LTB4 agonist

+G31P 0

4000

8000

12000

12.5 30.5 48.5 66.5 84.5

C5a agonist

+G31P

0.0

60.0

120.0

C5a 0.1nM fmlp 1nM LTB4 1ng/mL SIX

ATO

MEH

C )

MES -/ + dl eif X04/ sll ec( C5a fMLP LTB4

REL

. FLU

OR

ESC

ENC

E

TIME (sec)

* * ** **

** **

FIGURE 2. G31P engagement on neutrophils in-duces heterologous desensitization of their GPCR forC5a, LTB4, and fMLP. Purified human neutrophilswere stimulated with 0.1 nM C5a, 1 nM fMLP, or 1ng/ml LTB4, in the absence or presence of 10 or 50ng/ml G31P, as described in Material and Methods. A,The ability of G31P (10 ng/ml) to interfere with theagonist-induced intracellular Ca2� flux was assessed asin Fig. 1. B, The abilities of varying doses of G31P toblock the cells’ chemotactic responses were also as-sessed as above. G31P very substantially reduced Ca2�

flux induced by fMLP and LTB4, but was less effectivein blocking C5a-dependent Ca2� flux, and this relativeefficacy was observed also in the chemotaxis assays.� and ��, p � 0.05 and 0.01, respectively, relative tothe agonist treatments alone. The data shown are rep-resentative of at least three experiments performed withsimilar results. The dashed lines across graph in B rep-resent the mean background response in the assay.

3217The Journal of Immunology

by guest on May 28, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

in the presence or absence of G31P. The quiescent A549 cellssecreted very little CXCL8, but LPS dose dependently up-regu-lated their expression of this chemokine. Reports of CXCR1 andCXCR2 expression by epithelial cells vary, so we wished first todetermine whether our A549 cells expressed either of these recep-tors. We used quantitative real-time PCR and did indeed detectsignificant levels of the CXCR1 (Fig. 5A, right panel), but little inthe way of CXCR2. When we challenged our A549 cells with 10

ng/ml LPS in the presence of increasing doses of G31P, the cellsproduced progressively less CXCL8 (Fig. 5B). This effect ex-tended also to their expression of CXCL1, wherein G31P reducedLPS-induced CXCL1 expression by �98%, as determined byquantitative real-time PCR.

Neutrophils release a variety of mediators that are potential ep-ithelial cell agonists (e.g., elastase) (32), whereas epithelial cellssimilarly secrete neutrophil agonists. This suggests the possibilitythat these cells could through a putative mutual stimulatory mech-anism drive an escalating inflammatory process. We tested this bycoculturing purified otherwise quiescent neutrophils with mono-layers of resting A549 cells. Although the neutrophils were largelyunstimulated, there was a significant background release ofCXCL8 by these cells over 16 h in culture (Fig. 6). Nevertheless,simple coculture of these neutrophils with A549 cells induced asynergistic expression of CXCL8 in the cultures, and this effectwas dependent on the numbers of neutrophils added to the cultures(Fig. 6). The A549 cells did not secrete significant amounts ofCXCL8 when cultured on their own. The addition of G31P to thesecocultures dramatically (up to 98%) reduced the expression ofCXCL8 ( p � 0.01; Fig. 6).

100 101 102 103 1040

20

40

60

80

100

100 101 102 103 1040

20

40

60

80

100

100 101 102 103 1040

20

40

60

80

100

IL-8

GRO

FIGURE 4. G31P reverses the antiapoptotic influence of CXCL1 andCXCL8 on neutrophils. Purified human neutrophils were cultured for 18 hin medium alone (apoptosis positive control; top panel) or in the presenceof 100 ng/ml CXCL1 or CXCL8, either with (solid lines) or without(dashed lines) G31P (100 ng/ml), and stained with fluorochrome-labeledannexin V. The proportions of apoptotic (annexin V-binding) cells weredetermined by FACS. Negative control results obtained using freshly pu-rified annexin V-stained neutrophils (shaded tracing) are included in eachgraph. Necrotic cell death was also assessed by propidium iodide uptake;we found �3% propidium iodide-positive cells in all samples. The dataindicate that cells incubated in medium alone were largely apoptotic.Exposure to CXCL8 or CXCL1 reduced apoptosis very substantially,but addition of G31P to these cultures largely reversed the ELR-CXCchemokine-induced antiapoptotic effects. These assays were repeatedthree times.

0 1 10

100

G31P (ng/ml)

0

1000

3000

CXCL8 (pg/ml)

** *

0

5

10

CXCL1 mRNA

- 10

0 10

100

1000

no LPS

**

**

0 15

00

3000

0 .0

1 0.

1 1 10 CXC

L8 S

ECR

ETIO

N (p

g/m

l)

LPS (ng/ml)

**

** ** **

REL

ATIV

E EX

PRES

SIO

N

A549 PMN

1

10

100

1000

CXCR1CXCR2CXCR1 CXCR2

FIGURE 5. G31P antagonizes the endotoxin-induced inflammatory cy-cle in human bronchial epithelial cells. A, Monolayers of A549 bronchialepithelial cells were exposed to varying doses of endotoxin, and the culturesupernatant was collected 16 h later and assayed by ELISA for CXCL8(left panel). Expression of the CXCR1 or CXCR2 by the A549 cells andhuman neutrophils was assessed by quantitative real-time PCR using spe-cific primers (right panel). Our data indicate that peak CXCL8 expressionwas found in the A549 cultures that were challenged with 10 ng/ml LPS,and that our A549 cells expressed significant levels of CXCR1, althoughcertainly much less than that found in neutrophils. B, A549 cell monolayerswere challenged with 10 ng/ml LPS in the presence of the indicated con-centrations of G31P, and 16 h later the culture supernatants and cellularRNA were isolated and assayed for CXCL8 protein (ELISA) and CXCL1mRNA (quantitative real-time PCR). Addition of G31P to the A549 cellsdose dependently suppressed expression of both chemokines by the LPS-challenged cells. � and ��, p � 0.05 and 0.01, respectively, relative to thebaseline treatments alone. The data shown are representative of at leastthree experiments performed with similar results.

3218 ELR-CXC CHEMOKINE ANTAGONISM

by guest on May 28, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

G31P blocks airway endotoxemia-induced acute lunginflammation

The observations that a fully human form of G31P blocked in-flammatory responses at multiple levels (e.g., heterologous desen-sitization of GCPR, proapoptotic effects, blockade of the autocrineepithelial inflammatory cycle) suggested that it could be effectivealso in blocking inflammatory responses in vivo. We have shownthat bovine and bovine-human chimeric orthologues of G31P canblock an array of inflammatory pathologies, including airway en-dotoxemia (24, 25), aspiration pneumonia (X. Zhao, J. Town, F.Li, W. Li, X. Zhang, and J. Gordon, submitted for publication),ischemia-reperfusion injury (X. Zhao, J. Town, A. Yang, X.Zhang, G. Sawicki, and J. Gordon, submitted for publication), andenvironmental pollutant-induced lung injury (33). Although weknow in airway endotoxemia that bG31P treatments reduce neu-trophil infiltration of the airways, it remains possible that neutro-phils marginate in the pulmonary vasculature under the influenceof the inflammatory mediators expressed in this compartment andare activated in situ. For this reason, we wished to determinewhether we could detect markers of the individual neutrophil gran-ules or of activated structural cells in the lungs of endotoxemicanimals. We induced airway endotoxemia in guinea pigs, as re-ported previously (24, 25), and assessed the levels of MPO (1°granules), LF (2° granules), and MMP-9 (3° granules) at 4, 8, and16 h after endotoxin challenge. We confirmed that G31P did in-deed reduce airway neutrophilia (90–99% reduced at 4, 8, and16 h; p � 0.01 vs saline-treated animals). In concert with this,there were significant reductions in the BAL levels of MPO ( p �0.05, 8-h time point only), LF ( p � 0.05, all times), and MMP-9( p � 0.05, 4- and 8-h time points), which suggested that fewneutrophils that did achieve the lungs were not sufficiently acti-vated to degranulate. We coincidentally observed a strong signalfrom a 68-kDa gelatinase (Fig. 7) that comigrated with MMP-2expressed in positive control HT1080 fibrosarcoma cells (data notshown), suggesting that we had indeed strongly activated the lungepithelium and other structural cells that express this metallopro-teinase (34, 35). G31P treatment also essentially abrogated thisresponse at each time, suggesting that, as we had observed withour A549 cells, structural cells were also influenced by ELR-CXCchemokine antagonism in vivo.

DiscussionWe have generated a human orthologue of bG31P (36, 37). Weshow in this study that this G31P specifically blocks activation ofneutrophils through the CXCR1 and CXCR2, and thereby theirchemoattraction into inflammatory foci and expression of ROI. Inengineering bG31P from its parent molecule, bCXCL8, our firststep was to introduce high affinity into the molecule, which we didby N-terminal truncation and substitution of a lysine in place ofArg11 to generate bCXCL8(3–73)K11R (23). These arguments werein part based on prior work with human CXCL8 by Clark-Lewiset al. (38, 39), although the bovine form of CXCL8(3–73)K11Rappeared to have unexpectedly high agonist activity relative to itsclosest human counterpart (23). We next reduced the receptor-signaling activity of this agonist by introducing a Gly residue inplace of Pro31 to generate an antagonist that fortuitously had ex-ceptionally high affinity relative to CXCL8 (22). We hypothesizethat it is this augmented receptor affinity that allows G31P to soeffectively compete with native ELR-CXC chemokines for theirreceptors on neutrophils. It makes sense that G31P would operatevia both the CXCR1 and CXCR2 because its parent molecule wasCXCL8, a high-affinity ligand of both receptors (9, 40). Our datashowing that G31P blocks CXCL8-dependent intracellular Ca2�

flux and chemotaxis in CXCR1-transfected HEK cells confirm un-equivocally that G31P targets the CXCR1, and the fact that itblocks the activities of CXCL1 and CXCL5, which are exclusive

FIGURE 6. Coculture of quiescent A549 cells with unstimulated pe-ripheral blood neutrophils leads to synergistic CXCL8 expression. Theindicated numbers of quiescent, purified human peripheral blood neutro-phils from healthy donors were either cultured alone (PMN alone) or addedto monolayers of unstimulated A549 cells (A549 � PMN). G31P (100ng/ml) was added to half of these cultures, and 16 h later the culturesupernatants were harvested assayed for CXCL8 by ELISA. The additionof neutrophils to the A549 cells synergistically enhanced expression ofCXCL8 in a cell concentration-dependent fashion, and G31P antagonizedthis cytokine response. � and ��, p � 0.05 and 0.01, respectively, relativeto the no G31P treatments alone. The data shown are from one blood donorthat is representative of the responses of three such donors.

FIGURE 7. G31P treatments reduce neutrophil activation in airway en-dotoxemia. Airway endotoxemia was induced in guinea pigs by intranasalinstillation of bacterial endotoxin (5 �g/kg). The animals (n � 5) weretreated 30 min before challenge with either saline (f) or G31P (u; 250�g/kg) and euthanized 4, 8, or 16 h later. Differential counts were done onthe cells recovered by BAL, and the BAL fluids were assessed for thelevels of MPO, LF, and MMP-9, as specific markers of neutrophil primary,secondary, and tertiary granules, respectively, as noted in Materials andMethods. The MMP-9 graphical data were obtained from the zymogram(lower panel) by gel-scanning densitometry. The zymogram shows theBAL levels of MMP-9 (92 kDa) and MMP-2 (68 kDa), an activationmarker of lung structural cells. The G31P treatments dramatically reducedneutrophil activation in the airways as well as structural cell activation.�, p � 0.05, and ��, p � 0.01, as compared with LPS plus saline treatmentonly. Data are expressed as the mean � SEM, and are representative of tworepeat experiments.

3219The Journal of Immunology

by guest on May 28, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

CXCR2 agonists (9, 40), similarly confirms its interactions withthese receptors.

We also documented in this study that G31P desensitizes het-erologous GPCR in a quantitatively better fashion than CXCL8(18), probably in part because of its high-affinity interactions withCXCR1 and CXCR2 (22). Thus, G31P blocked both intracellularCa2� flux and chemotactic responses of neutrophils to signalingvia CD88/C5aR1 (the C5a receptor), BLT1 (the LTB4 receptor),and Fpr1 (the fMLP receptor), although it was not as effective indesensitizing C5a-dependent Ca2� flux as that associated withLTB4 or fMLP signaling. It has been reported that in responding tosignals generated in inflammatory settings, neutrophils adopt a hi-erarchical response pattern, with responsiveness to agonists imme-diately responsible for recruitment from the vasculature (e.g.,CXCL8, LTB4) being superceded as the cells traverse the bloodvessel wall by responsiveness to those that would draw the cell tothe pathogen itself (e.g., C5a, fMLP) (41). This effect is reportedlybrought about by C5a- and fMLP-driven activation of MAPK inneutrophils, which dampens Akt activation, a critical event in re-sponses to CXCL8 and LTB4 (41). In the context of a bacterialinfection, for example, this process would make a great deal ofsense, because as the neutrophil moves into the tissues and towardthe pathogen, it must necessarily concentrate its resources on thattarget. In support of this, it had also been reported that whereas thefMLP and C5a receptors can cross-desensitize one another and theCXCL8 receptor, CXCL8 signaling cannot reciprocally desensitizeCa2� flux or a number of other intracellular events downstream ofthe fMLP or C5a receptors (19, 20). At odds with this, however, isthe report that whereas the CXCR2 cannot desensitize the CD88 orFpr1, CXCR1 signaling can do so, albeit at significantly lowerlevels than that seen with C5a or fMLP-induced CXCL8 receptordesensitization (42). Also at odds is the report that CXCL8 sig-naling desensitizes neutrophil chemotactic responses to subsequentC5a or fMLP exposures (18). Our own data clearly indicated thatG31P desensitizes neutrophil responses to fMLP, C5a, and LTB4,as noted most likely related to high-affinity interactions with itsreceptors. This could in part explain its effectiveness in dampeningthe pathology associated with airway endotoxemia in our model. Itis known that endotoxin activates C5 to generate C5a (43), and itseems reasonable to suggest that commercial endotoxin prepara-tions such as we used would likely be contaminated with bacterialtripeptides (e.g., fMLP). Furthermore, given that LTB4 antagonismis of significant benefit in airway endotoxemia (17, 44), it is alsoreasonable to suggest that this agonist would also have beenpresent in our endotoxemic animals, although we did not assessthis. This suggests that the neutrophilic pathology we observedcould well be related to the combined effects of multiple agonists,including ELR-CXC chemokines, C5a, LTB4, and perhaps fMLP.We hypothesize that the ability of G31P to ameliorate this pathol-ogy is attributable to its ability to not only block the CXCR1 andCXCR2, but also desensitize these other GPCR.

Another important finding in this study was that G31P also an-tagonized the responses of airway epithelial cells to bacterial en-dotoxin. Our data show that A549 cells express the CXCR1, al-though epithelial expression of the CXCR1 or CXCR2 is notuniform. Although unstimulated BEAS2B or colonic epithelialcells do not express the CXCR-1 or CXCR-2 (45, 46), epithelialcell lines from CF patients (IB3-1 and CFBE41o-) do express thesereceptors (47). In contrast, hypoxic Caco-2 and HT-29 bronchialepithelial cells express the CXCR1 (but not the CXCR2) and un-dergo chemotactic responses to CXCL8 signaling (48). This couldsuggest that inflammatory signals up-regulate expression of thesereceptors, but in A549 cells, at least, LPS stimulation does not altertheir expression of the CXCR1 or CXCR2 (J. Town, unpublished

observation). The blockade of CXCL8 expression in LPS-stimu-lated epithelial cells by G31P leads us to speculate that CXCR1ligands are autocrine inflammatory mediators for these cells. It hadbeen reported previously that CXCL8 is an autocrine growth factorfor pancreatic tumor (49) and melanoma (50, 51) cells, but, as faras we are aware, this is the first demonstration that ELR-CXCchemokine expression can be a self-perpetuating event for epithe-lial cells. We also found that simple coculture of quiescent humanneutrophils with A549 cells strongly induced CXCL8 expressionin these cultures and that G31P interrupts this response, dramati-cally reducing chemokine production. We did not determine whichpopulation of cells was activated in these cocultures, and therefore,responsible for the cytokine expression. A number of mediatorsreleased from activated neutrophils (e.g., cathepsin G, elastase,defensins) can induce the release of CXCL8 from respiratory ep-ithelial cells and, reciprocally, a number of epithelial products areneutrophil agonists (52, 53). Although neutrophils can be readilyactivated by the procedures used in their isolation (54, 55), whencultured alone neither the epithelial cells or neutrophils secretedsubstantial amounts of CXCL8, indicating that the interactions be-tween these two cells were instrumental in their induction, and thatendotoxin contamination of our culture medium by itself wouldnot explain our observations. More importantly, G31P was able tointerfere with these intercellular interactions sufficiently to ame-liorate their influence on one another. The remarkable degree towhich G31P dampened neutrophilic inflammation in our airwayendotoxemic animals suggests that its ability to block the interac-tions between bronchial epithelial cells and those neutrophils thatdid find their way into the airways could be another integral com-ponent of its therapeutic effects.

Neutrophils are continuously generated in the bone marrow andreleased into the circulation at exceptionally high rates, whichmeans of course that they need to be equally quickly disposed ofif a homeostatic balance is to be maintained. Thus, neutrophils arepreprogrammed to undergo apoptotic death quite rapidly, and thisis only offset if inflammatory events dictate that there is an ex-tended need for increased numbers of viable cells. A number ofinflammatory mediators have antiapoptotic effects on neutrophils,including the ELR-CXC chemokines (31, 56), such that cells ex-posed to these mediators remain viable for extended periods oftime, and thereby can potentially contribute to the regional anti-microbial activities. This translates also into potential contribu-tions to adventitious pathogenic processes, such that the ability ofG31P to countermand ELR-CXC chemokine-driven antiapoptoticeffects would be of significant benefit to the host by virtue ofencouraging functional removal of these cells from the inflamma-tory environment.

We had previously shown that bovine and a human-bovine chi-meric form of G31P are effective in blocking neutrophilic pathol-ogy associated with airway endotoxemia (24, 25), and we con-firmed in this study for the first time that the human orthologueG31P was similarly effective. In this study, we assessed the kinet-ics of neutrophil infiltration of the airways as well as their activa-tion in situ in the lungs. Our data showed that G31P can dramat-ically reduce neutrophil recruitment and activation, inasmuch aswe found high levels of the primary, secondary, and tertiary gran-ule markers MPO, LF, and MMP-9, respectively, in lungs of thesaline-treated endotoxemic animals, but not in the G31P-treatedanimals. Previous studies have shown that a number of MMP arepresent in the airways during experimental or clinical acute lunginjuries or respiratory distress (34, 35), and we found both MMP-2and MMP-9 in our BAL samples. Although one might argue thatthe multiple gelatinases detected in our zymograms (i.e., 92 and 68kDa) could have been attributable to processing of high m.w.

3220 ELR-CXC CHEMOKINE ANTAGONISM

by guest on May 28, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

MMP-9 to smaller forms, such processing of MMP-9 reportedlyleads to generation of 86- and 82-kDa gelatinases (i.e., not a 68-kDa activity) (57). MMP-2 is expressed as a 72-kDa zymogen thatis processed to a 68-kDa gelatinase (58). MMP-2 is mainly pro-duced by resident cells during acute lung injury, including, forexample, type II epithelial cells or endothelial cells (35), such thatthe observation that G31P blocked its expression further supportsour hypothesis that this antagonist acts not only on the neutrophil,but also on epithelial or other resident cells. Endothelial cells, an-other source of MMP-2 (35), also express the CXCR2 (59), socould potentially also fall under the influence of G31P.

In conclusion, our data indicate that G31P is a potent antagonistof neutrophil activation induced by ELR-CXC chemokines. It tar-gets not only the CXCR1 and CXCR2, but also heterologousGPCR (e.g., CD88/C5aR1, BLT1, and Fpr1) expressed onCXCR1- or CXCR2-positive cells, and it interrupts the autocrineepithelial cell inflammatory cascade triggered by bacterial endo-toxins. Thus, the beneficial effects of G31P flow from its ability tosimultaneously target multiple inflammatory processes. We spec-ulate that the ability of G31P to antagonize heterologous GPCRmay be a very important part of its anti-inflammatory activities,such that ELR-CXC chemokine antagonism in the absence of suchactivities (e.g., by anti-chemokine Abs) may provide somewhatless effective protection. An obvious question to ask when onewishes to block neutrophil participation in biological responses iswhether there are deleterious effects. For example, in bacterialpneumonia, one might expect that these cells may be required forsuccessful elimination of the pathogen(s). However, we havefound that G31P very significantly attenuates pathology in aspira-tion pneumonia, but at the same time does not reduce control of themicrobial populations that enter the airways (X. Zhao, J. Town, F.Li, W. Li, X. Zhang, and J. Gordon, submitted for publication).Blocking neutrophil recruitment in bovine pneumonic pasteurel-losis similarly improves outcomes very significantly (60). Wewould predict based on our cumulative results that a broad array ofinflammatory diseases would thus benefit from such ELR-CXCchemokine antagonism, as might other processes in which inflam-mation per se is not necessarily prevalent, but in which the ELR-CXC chemokines have nevertheless been implicated.

DisclosuresJohn Gordon is the inventor of G31P, which has been licensed to PacgenBiopharmaceuticals for commercial development. He will receive royaltiesif G31P is ever successfully commercialized.

References1. Nathan, C. 2002. Points of control in inflammation. Nature 420: 846–852.2. Smith, R. S., E. R. Fedyk, T. A. Springer, N. Mukaida, B. H. Iglewski, and

R. P. Phipps. 2001. IL-8 production in human lung fibroblasts and epithelial cellsactivated by the Pseudomonas autoinducer N-3-oxododecanoyl homoserine lac-tone is transcriptionally regulated by NF-�B and activator protein-2. J. Immunol.167: 366–374.

3. Baggiolini, M. 1998. Chemokines and leukocyte traffic. Nature 392: 565–568.4. Springer, T. A. 1994. Traffic signals for lymphocyte recirculation and leukocyte

emigration: the multistep paradigm. Cell 76: 301–314.5. Wolf, M., M. B. Delgado, S. A. Jones, B. Dewald, I. Clark-Lewis, and

M. Baggiolini. 1998. Granulocyte chemotactic protein 2 acts via both IL-8 re-ceptors, CXCR1 and CXCR2. Eur. J. Immunol. 28: 164–170.

6. Wuyts, A., P. Proost, J. P. Lenaerts, A. Ben Baruch, J. Van Damme, andJ. M. Wang. 1998. Differential usage of the CXC chemokine receptors 1 and 2 byinterleukin-8, granulocyte chemotactic protein-2 and epithelial-cell-derived neu-trophil attractant-78. Eur. J. Biochem. 255: 67–73.

7. Ahuja, S. K., and P. M. Murphy. 1996. The CXC chemokines growth-regulatedoncogene (GRO) �, GRO�, GRO�, neutrophil-activating peptide-2, and epithe-lial cell-derived neutrophil-activating peptide-78 are potent agonists for the typeB, but not the type A, human interleukin-8 receptor. J. Biol. Chem. 271:20545–20550.

8. Chuntharapai, A., and K. J. Kim. 1995. Regulation of the expression of IL-8receptor A/B by IL-8: possible functions of each receptor. J. Immunol. 155:2587–2594.

9. Jones, S. A., M. Wolf, S. Qin, C. R. Mackay, and M. Baggiolini. 1996. Differentfunctions for the interleukin 8 receptors (IL-8R) of human neutrophil leukocytes:NADPH oxidase and phospholipase D are activated through IL-8R1 but not IL-8R2. Proc. Natl. Acad. Sci. USA 93: 6682–6686.

10. Chakrabarti, S., and K. D. Patel. 2005. Regulation of matrix metalloproteinase-9release from IL-8-stimulated human neutrophils. J. Leukocyte Biol. 78: 279–288.

11. Gijsbers, K., G. Van Assche, S. Joossens, S. Struyf, P. Proost, P. Rutgeerts,K. Geboes, and J. Van Damme. 2004. CXCR1-binding chemokines in inflam-matory bowel diseases: down-regulated IL-8/CXCL8 production by leukocytes inCrohn’s disease and selective GCP-2/CXCL6 expression in inflamed intestinaltissue. Eur. J. Immunol. 34: 1992–2000.

12. Cummings, C. J., T. R. Martin, C. W. Frevert, J. M. Quan, V. A. Wong,S. M. Mongovin, T. R. Hagen, K. P. Steinberg, and R. B. Goodman. 1999.Expression and function of the chemokine receptors CXCR1 and CXCR2 insepsis. J. Immunol. 162: 2341–2346.

13. Goodman, R., C. Cummings, C. Frevert, J. Quan, and T. Martin. 1999. Functionalsignificance of CXCR2 down-regulation on neutrophils from patients with severesepsis. Chest 116: 111s–112s.

14. Podolin, P. L., B. J. Bolognese, J. J. Foley, D. B. Schmidt, P. T. Buckley,K. L. Widdowson, Q. Jin, J. R. White, J. M. Lee, R. B. Goodman, et al. 2002. Apotent and selective nonpeptide antagonist of CXCR2 inhibits acute and chronicmodels of arthritis in the rabbit. J. Immunol. 169: 6435–6444.

15. Park, K. W., M. Tofukuji, C. Metais, M. E. Comunale, H. B. Dai, M. Simons,G. L. Stahl, A. Agah, and F. W. Sellke. 2000. Attenuation of endothelium-de-pendent dilation of pig pulmonary arterioles after cardiopulmonary bypass isprevented by monoclonal antibody to complement C5a. Anesth. Analg. 89:42–48.

16. Crooks, S. W., D. L. Bayley, S. L. Hill, and R. A. Stockley. 2000. Bronchialinflammation in acute bacterial exacerbations of chronic bronchitis: the role ofleukotriene B4. Eur. Respir. J. 15: 274–280.

17. Wollert, P. S., M. J. Menconi, B. P. O’Sullivan, H. Wang, V. Larkin, andM. P. Fink. 1993. LY255283, a novel leukotriene B4 receptor antagonist, limitsactivation of neutrophils and prevents acute lung injury induced by endotoxin inpigs. Surgery 114: 191–198.

18. Blackwood, R. A., K. T. Hartiala, E. E. Kwoh, A. T. Transue, and R. C. Brower.1996. Unidirectional heterologous receptor desensitization between both thefMLP and C5a receptor and the IL-8 receptor. J. Leukocyte Biol. 60: 88–93.

19. Richardson, R. M., H. Ali, E. D. Tomhave, B. Haribabu, and R. Snyderman.1995. Cross-desensitization of chemoattractant receptors occurs at multiple lev-els: evidence for a role for inhibition of phospholipase C activity. J. Biol. Chem.270: 27829–27833.

20. Tomhave, E. D., R. M. Richardson, J. R. Didsbury, L. Menard, R. Snyderman,and H. Ali. 1994. Cross-desensitization of receptors for peptide chemoattractants:characterization of a new form of leukocyte regulation. J. Immunol. 153:3267–3275.

21. Li, F., X. Zhang, and J. R. Gordon. 2002. CXCL8(3–73)K11R/G31P antagonizesligand binding to the neutrophil CXCR1 and CXCR2 receptors and cellular re-sponses to CXCL8/IL-8. Biochem. Biophys. Res. Commun. 293: 939–944.

22. Li, F., X. Zhang, C. Mizzi, and J. R. Gordon. 2002. CXCL8(3–73)K11R/G31Pantagonizes the neutrophil chemoattractants present in pasteurellosis and mastitislesions and abrogates neutrophil influx into intradermal endotoxin challenge sitesin vivo. Vet. Immunol. Immunopathol. 90: 65–77.

23. Li, F., and J. R. Gordon. 2001. IL-8(3–73)K11R is a high affinity agonist of theneutrophil CXCR1 and CXCR2. Biochem. Biophys. Res. Commun. 286:595–600.

24. Gordon, J. R., F. Li, X. Zhang, W. Wang, X. Zhao, and A. Nayyar. 2005. Thecombined CXCR1/CXCR2 antagonist CXCL8(3–73)K11R/G31P blocks neutro-phil infiltration, pyrexia, and pulmonary vascular pathology in endotoxemic an-imals. J. Leukocyte Biol. 78: 1265–1272.

25. Zhao, X., F. Li, J. R. Town, X. Zhang, W. Wang, and J. R. Gordon. 2007.Humanized forms of the CXCR1/CXCR2 antagonist, bovine CXCL8(3–73)K11R/G31P, effectively block ELR-CXC chemokine activity and airway endotoxemiapathology. Int. Immunopharmacol. 7: 1723–1731.

26. Gordon, J. R., V. A. Swystun, F. Li, X. Zhang, B. E. Davis, P. Hull, andD. W. Cockcroft. 2003. Regular salbutamol use increases CXCL8 responses inasthma: relationship to the eosinophil response. Eur. Respir. J. 22: 118–126.

27. Gordon, J. R., and F. Li. 2002. bIL-8(3–73)K11R/G31P antagonizes human neu-trophil responses to ELR-CXC chemokines, as well as in vivo bacterial endo-toxin-induced inflammatory responses in cattle. FASEB J. 16: A1079.

28. Swystun, V. A., J. R. Gordon, E. B. Davis, X. Zhang, and D. W. Cockcroft. 2000.Mast cell tryptase release and asthmatic responses to allergen increase with reg-ular use of salbutamol. J. Allergy Clin. Immunol. 106: 57–64.

29. Sariahmetoglu, M., B. D. Crawford, H. Leon, J. Sawicka, L. Li, B. J. Ballermann,C. Holmes, L. G. Berthiaume, A. Holt, G. Sawicki, and R. Schulz. 2007. Reg-ulation of matrix metalloproteinase-2 (MMP-2) activity by phosphorylation.FASEB J. 21: 2486–2495.

30. Mackerness, K. J., G. R. Jenkins, A. Bush, and P. J. Jose. 2008. Characterizationof the range of neutrophil stimulating mediators in cystic fibrosis sputum. Thorax63: 614–620.

31. Glynn, P. C., E. Henney, and I. P. Hall. 2002. The selective CXCR2 antagonistSB272844 blocks interleukin-8 and growth-related oncogene-�-mediated inhibi-tion of spontaneous neutrophil apoptosis. Pulm. Pharmacol. Ther. 15: 103–110.

32. Chen, H. C., H. C. Lin, C. Y. Liu, C. H. Wang, T. Hwang, T. T. Huang, C. H. Lin,and H. P. Kuo. 2004. Neutrophil elastase induces IL-8 synthesis by lung epithelialcells via the mitogen-activated protein kinase pathway. J. Biomed. Sci. 11:49–58.

3221The Journal of Immunology

by guest on May 28, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

33. Podechard, N., V. Lecureur, E. Le Ferrec, I. Guenon, L. Sparfel, D. Gilot,J. R. Gordon, V. Lagente, and O. Fardel. 2008. Interleukin-8 induction by theenvironmental contaminant benzo (a) pyrene is aryl hydrocarbon receptor-depen-dent and leads to lung inflammation. Toxicol. Lett. 177: 130–137.

34. D’Ortho, M., P. Jarreau, C. Delacourt, I. Macquin-Mavier, M. Levame, S. Pezet,A. Harf, and C. Lafuma. 1994. Matrix metalloproteinase and elastase activities inLPS-induced acute lung injury in guinea pigs. Am. J. Physiol. 266: L209–L216.

35. Fligiel, S. E., T. Standiford, H. Fligiel, D. Tashkin, R. Strieter, R. Warner,K. Johnson, and J. Varani. 2006. Matrix metalloproteinases and matrix metallo-proteinase inhibitors in acute lung injury. Hum. Pathol. 37: 422–430.

36. Li, A., S. Dubey, M. L. Varney, and R. K. Singh. 2002. Interleukin-8-inducedproliferation, survival, and MMP production in CXCR1 and CXCR2 expressinghuman umbilical vein endothelial cells. Microvasc. Res. 64: 476–481.

37. Li, C., and R. M. Jackson. 2002. Reactive species mechanisms of cellular hyp-oxia-reoxygenation injury. Am. J. Physiol. 282: C227–C241.

38. Clark-Lewis, I., B. Dewald, M. Loetscher, B. Moser, and M. Baggiolini. 1994.Structural requirements for interleukin-8 function identified by design of analogsand CXC chemokine hybrids. J. Biol. Chem. 269: 16075–16081.

39. Clark-Lewis, I., C. Schumacher, M. Baggiolini, and B. Moser. 1991. Structure-activity relationships of interleukin-8 determined using chemically synthesizedanalogs: critical role of NH2-terminal residues and evidence for uncoupling ofneutrophil chemotaxis, exocytosis, and receptor binding activities. J. Biol. Chem.266: 23128–23134.

40. Jones, S. A., B. Dewald, I. Clark-Lewis, and M. Baggiolini. 1997. Chemokineantagonists that discriminate between interleukin-8 receptors: selective blockersof CXCR2. J. Biol. Chem. 272: 16166–16169.

41. Heit, B., S. Tavener, E. Raharjo, and P. Kubes. 2002. An intracellular signalinghierarchy determines direction of migration in opposing chemotactic gradients.J. Cell Biol. 159: 91–102.

42. Richardson, R. M., B. C. Pridgen, B. Haribabu, H. Ali, and R. Snyderman. 1998.Differential cross-regulation of the human chemokine receptors CXCR1 andCXCR2: evidence for time-dependent signal generation. J. Biol. Chem. 273:23830–23836.

43. Smedegard, G., L. X. Cui, and T. E. Hugli. 1989. Endotoxin-induced shock in therat: a role for C5a. Am. J. Pathol. 135: 489–497.

44. Fink, M. P., B. P. O’Sullivan, M. J. Menconi, P. S. Wollert, H. Wang,M. E. Youssef, and J. H. Fleisch. 1993. A novel leukotriene B4-receptor antag-onist in endotoxin shock: a prospective, controlled trial in a porcine model. Crit.Care Med. 21: 1825–1837.

45. Dwinell, M. B., L. Eckmann, J. D. Leopard, N. M. Varki, and M. F. Kagnoff.1999. Chemokine receptor expression by human intestinal epithelial cells. Gas-troenterology 117: 359–367.

46. Farkas, L., M. C. Hahn, M. Schmoczer, N. Jentsch, K. Kratzel, M. Pfeifer, andC. Schulz. 2005. Expression of CXC chemokine receptors 1 and 2 in humanbronchial epithelial cells. Chest 128: 3724–3734.

47. Boncoeur, E., V. S. Criq, E. Bonvin, T. Roque, A. Henrion-Caude,D. C. Gruenert, A. Clement, J. Jacquot, and O. Tabary. 2008. Oxidative stressinduces extracellular signal-regulated kinase 1/2 mitogen-activated protein kinasein cystic fibrosis lung epithelial cells: potential mechanism for excessive IL-8expression. Int. J. Biochem. Cell Biol. 40: 432–446.

48. Sturm, A., D. C. Baumgart, J. H. d’Heureuse, A. Hotz, B. Wiedenmann, andA. U. Dignass. 2005. CXCL8 modulates human intestinal epithelial cells througha CXCR1 dependent pathway. Cytokine 29: 42–48.

49. Takamori, H., Z. G. Oades, O. C. Hoch, M. Burger, and I. U. Schraufstatter. 2000.Autocrine growth effect of IL-8 and GRO-� on a human pancreatic cancer cellline, Capan-1. Pancreas 21: 52–56.

50. Norgauer, J., B. Metzner, and I. Schraufstatter. 1996. Expression and growth-promoting function of the IL-8 receptor � in human melanoma cells. J. Immunol.156: 1132–1137.

51. Schadendorf, D., A. Moller, B. Algermissen, M. Worm, M. Sticherling, andB.B. Czarnetzki. 1994. IL-8 produced by human malignant melanoma cells invitro is an essential autocrine growth factor. J. Immunol. 151: 2667–2675.

52. Khair, O. A., R. J. Davies, and J. L. Devalia. 1996. Bacterial-induced release ofinflammatory mediators by bronchial epithelial cells. Eur. Respir. J. 9:1913–1922.

53. Skerrett, S. J., H. D. Liggitt, A. M. Hajjar, R. K. Ernst, S. I. Miller, andC. B. Wilson. 2004. Respiratory epithelial cells regulate lung inflammation inresponse to inhaled endotoxin. Am. J. Physiol. 287: L143–L152.

54. Link, A., B. Hummel, H. Schwerdt, J. Schwamborn, F. Jung, and H. Schieffer.1997. Influence of neutrophil separation on the expression of adhesion molecules.Clin. Hemorheol. Microcirc. 17: 175–180.

55. Venaille, T., N. Misso, M. Phillips, B. Robinson, and P. Thompson. 1994. Effectsof different density gradient separation techniques on neutrophil function. Scand.J. Clin. Lab. Invest. 54: 385–391.

56. Kobayashi, S. D., J. M. Voyich, A. R. Whitney, and F. R. DeLeo. 2005. Spon-taneous neutrophil apoptosis and regulation of cell survival by granulocyte mac-rophage-colony stimulating factor. J. Leukocyte Biol. 78: 1408–1418.

57. Ramos-DeSimone, N., E. Hahn-Dantona, J. Sipley, H. Nagase, D. L. French, andJ. P. Quigley. 1999. Activation of matrix metalloproteinase-9 (MMP-9) via aconverging plasmin/stromelysin-1 cascade enhances tumor cell invasion. J. Biol.Chem. 274: 13066–13076.

58. Growcott, E. J., K. G. Spink, X. Ren, S. Afzal, K. H. Banner, and J. Wharton.2006. Phosphodiesterase type 4 expression and anti-proliferative effects in humanpulmonary artery smooth muscle cells. Respir. Res. 7: 9.

59. Reutershan, J., M. A. Morris, T. L. Burcin, D. F. Smith, D. Chang, M. S. Saprito,and K. Ley. 2006. Critical role of endothelial CXCR2 in LPS-induced neutrophilmigration into the lung. J. Clin. Invest. 116: 695–702.

60. Slocombe, R., J. Malark, R. Ingersoll, F. J. Derksen, and N. E. Robinson. 1985.Importance of neutrophils in the pathogenesis of acute pneumonic pasteurellosisin calves. Am. J. Vet. Res. 46: 2253–2258.

3222 ELR-CXC CHEMOKINE ANTAGONISM

by guest on May 28, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from