Endotoxin-free plastic and glasswares were used in all experiments. Purified LPS and myeloperoxidase (MPO) assay reagents were obtained from Sigma (St. Louis, MO). Chemicals, such as SP600125 were purchased from Calbiochem (La Jolla, CA). M39 was obtained from Merck (Rahway, NJ). Protease inhibitor cocktail was obtained from Roche Applied Science (Indianapolis, IN) and tissue fixative was purchased from Streck Laboratories (Omaha, NE). Abs to CXCL5 and their isotype controls were obtained from R&D Systems (Minneapolis, MN), and an Ab to prosurfactant protein C was obtained from Chemicon (Temecula, CA). Primers for PCR were purchased from IDT technologies (Coralville, IA). All real-time PCR reagents and the rat GAPDH primer were obtained from Applied Biosystems (Foster City, CA). An Ab against TLR4 was obtained from Santa Cruz and TIRAP was purchased from Abcam Inc. (Cambridge, MA). An Ab against MyD88 was obtained from Imgenex (San Diego, CA), and all of the secondary Abs were purchased from Jackson Immunoresearch (West Grove, PA). ELISA reagents were obtained from R&D Systems.

MyD88 gene–disrupted (MyD88^−/−) and TLR4 gene–deficient (TLR4^−/−) mice were backcrossed 12 times with C57Bl/6, yielding mice that are genetically > 99.9% C57Bl/6 and were obtained from Shizuo Akira (Osaka University, Osaka, Japan). Therefore, C57Bl/6 mice were used as controls for MyD88^−/− and TLR4^−/− mice. TIRAP gene–disrupted mice and their littermate controls were obtained by breeding between TIRAP^+/− males and females as they were in 129 × C57Bl/6 background (14). The Institutional Animal Care and Use Committee granted approval for all studies. Eight- and 10-wk-old mice ranging from 18–28 g were used in all experiments.

The rodent model to induce lung inflammation has been described in our previous publications (9, 15, 16). Briefly, mice and rats were exposed to 300 μg/ml of Escherichia coli–derived LPS (O111:B4) in 0.9% saline by aerosolization for 20 min, because this concentration of LPS-induced lung inflammation in a mouse model mimics several key features of human acute lung injury (9, 15, 16). Control groups of mice were treated according to exactly the same protocol except they received 0.9% saline instead of LPS. At 2, 8, and 24 h after LPS or saline inhalation, mice and rats were killed using a lethal dose of pentobarbital via intraperitoneal injection. These time points were selected because LPS-induced cellular responses and pathologic features were remarkably distinct at these designated time points as described earlier (9, 15, 16). In another set of experiments, mice were pretreated with 3 mg/kg of the P38 inhibitor (M39) orally in polyethylene glycol (PEG-400) 2 h before LPS challenge as described previously (16). In the other set of experiments, animals were given the JNK inhibitor (SP600125) in PPCES vehicle (30% PEG-400/20% polypropylene glycol/15% Cremophor EL/5% ethanol/30% saline) final volume 5 μl/g, 3 h before LPS as described earlier (17).

Bronchoalveolar lavage fluid (BALF) was immediately recovered from killed mice by instillation and recovery of 3.0 ml PBS via the tracheal catheter (9, 15, 16). The total cells were enumerated by counting with a hemocytometer. Differential cell count was performed by standard light microscopy based on staining with Diff-Quik (American Scientific Products, McGrew Park, IL). The remaining BALF was centrifuged and the supernatant was collected for CXCL5 measurement. CXCL5 quantitation was performed by ELISA using undiluted BALF according to the manufacturer's recommendations.

At the designated time points, the whole lungs were removed from mice and were immediately snap-frozen, followed by storage at −70°C. They were homogenized in 2 ml of PBS supplemented with 0.1% Triton X-100 and complete protease inhibitor (1 tablet/50 ml media), and the resulting homogenates were centrifuged at 12, 000 × g for 20 min (9, 16). The supernatants were harvested, passed through a 0.22-μm filter, and used immediately for measurement of CXCL5 protein by ELISA at a dilution of 1:200.

After killing, mouse and rat lungs were perfused and fixed using streck tissue fixative (Streck Laboratories) overnight at room temperature (9). Fixed lungs were mounted in paraffin, and 5 μM sections were made and affixed to microscopic slides. For immunohistochemistry, the sections were deparaffinized with xylene and rehydrated via a series of decreasing concentrations of ethanol. These sections were treated with 5% hydrogen peroxide in tris-buffered saline (pH 7.4) followed by treating with serum to block nonspecific staining. The lung sections were then incubated with anti-rat CXCL5 antibody (Ab). The binding to CXCL5 was amplified using a Vectastatin kit (Vector Laboratories, Burlingame, CA) and detected with diaminobenzidine subsrate (Promega Laboratories, Madison, WI). Sections were scanned using an Olympus BX51 upright microscope. Another section from the same lung lobe was stained with an Ab against prosurfactant protein C (proSP-C), a molecule specifically expressed in AEII cells (18, 19).

At 2, 8, and 24 h after challenge, total RNA was extracted from the lungs, and RT-PCR was performed for CXCL5. The sequences of the primers were: CXCL5, 5′-CTCAGTCATAGCCGCAACCGAGC-3′ and 5′-CGCTTCTTTCCACTGCGAGTGC-3′ (260 bp); and GAPDH, 5′-ATG-GAT- GAC-GAT-ATC-GCT-C-3′ and 5′-GAT-TCC-ATA-CCC-AGG-AAG-G-3′ (550 bp). A total of 8 μl were separated on a 1% agarose gel containing ethidium bromide, and bands were visualized and photographed using an ultraviolet transilluminator.

AEII cells were isolated and cultured from specific pathogen–free male Sprague-Dawley rats (Harlan, Indianapolis, IN). The cells were cultured in Dulbecco's modified Eagle's medium (DMEM) with 5% rat serum as previously described (18, 19). The purity of AEII cells was confirmed as > 96% using immunofluorescence microscopy with antibodies against cytokeratin and pro-surfactant C as respective markers for epithelial cells and AEII cells (18, 19). The cells were then stimulated with varying concentration of ultra pure LPS (E. coli LPS: O111:B5) for 18 h and proceeded for immunocytochemical analysis, or the supernatants were harvested for CXCL5 protein determination. In another set of experiments, an Ab against the extracellular domain of TLR4 or goat IgG (isotype control) was preincubated with cells at a final concentration of 40 μg to both the apical and basolateral sides 1 h before LPS stimulation. This is the minimal concentration that abolishes LPS-induced cytokine expression, including CXCL5 (S. Jeyaseelan and G. S. Worthen, unpublished data). To inhibit p38 and JNK kinases, cells were incubated with 10 μM and 5 μM respective concentrations of M39 (p38 kinase inhibitor) and SP600125 (JNK kinase inhibitor), because these inhibitors are highly selective at these concentrations as we described in the past (16). AEII cells were dedifferentiated to become type I–like cells in DMEM supplemented with 5% fetal bovine serum in the absence of keratinocyte-derived growth factor (KGF) as previously described (18, 19).

Mature murine bone marrow–derived neutrophils were isolated from mouse femurs as described in our previous publication (20). In our previous studies, we have also shown that the marrow neutrophils (> 95% purity) have equivalent functional responses and recirculation patterns as compared with peripheral murine neutrophils (20). A total of 5 × 10⁶ mature neutrophils were stimulated with 30 ng of LPS and incubated for 4 h at 37°C. After the incubation, the supernatant was centrifuged and collected for CXCL5 protein determination by ELISA.

Culture of alveolar macrophages (AMs) were isolated and were grown at a density of 2 × 10⁶ cells/ml/well on a 6-well plate for 3 d using RPMI 1640 media containing 10% fetal bovine serum and were stimulated with 30 ng of LPS. After 18 h incubation, the supernatants were harvested and used to detect the CXCL5 protein by ELISA.

Immunocytochemistry was performed as described earlier (21) with minor modifications. Briefly, AEII cells were grown in DMEM with 5% rat serum. The cells were fixed with alcohol and were immunostained for murine CXCL5 using goat anti-rat CXCL5 IgG (R&D Systems) or isotype-matched control Ab followed by FITC-labeled rabbit anti-goat IgG (Jackson Immunoresearch). CXCL5-producing cell (CPC) index was calculated as immunofluorescence-positive cells/total cells × 100. A total of 3,000 cells in 15 microscopic fields at a magnification of ×100 were enumerated for the calculation.

AEII cell lysates were separated by SDS-PAGE using precast Tris-glycine gels with an 18% acrylamide gradient (Bio-Rad, Hercules, CA). Proteins were then transferred to nitrocellulose membranes for 4 h at 85 V at 4°C, and the membrane was blocked with TBS containing 4% nonfat dry milk (NFM) for 1 h. The blot was incubated with goat anti-CXCL5 Ab (diluted 1/200, in TBS containing 4% NFM/0.05% Tween). The blot was then washed three times with PBS and developed using donkey anti-goat-HRP–conjugated secondary Ab (Jackson Immunoresearch) at a 1/4,000 dilution in TBS/Tween and developed with ECL plus Western blot detection system (Amersham Pharmacia Biotech, Piscataway, NJ). The same procedure was employed for the detection of anti-TLR4 (1/200 dilution of 200 μg/ml), -MyD88 (1/200 dilution of 500 μg/ml), and -TIRAP (1/200 dilution of 200 μg/ml) Ab. Appropriate secondary antibodies at a dilution of 1:5,000 concentrations were used.

For comparison between LPS- and saline-treated groups, one-way ANOVA was used (a P value < 0.05 was considered significant). We used Kaleidograph (version 6.0) (Synergy Software, Reading, PA) to make these calculations. Data expressed as mean ± SD.

In our previous studies, we have shown that aerosolization of C57Bl/6 mice with 300 μg/ml LPS for 20 min induced inflammatory features at 8 and 24 h, whereas cytokine and chemokine expression peaks at 2 h (9). Therefore, we used 8 and 24 h after LPS challenge for the demonstration of neutrophil accumulation, while 2 h after LPS was the time point selected to measure CXCL5 expression in the current investigation. Although our earlier study indicates that an anti-CXCL5 Ab attenuates neutrophil accumulation in the lung during LPS-induced inflammation at 30 μg (9), in the present study we have used CXCL5 blocking Ab over a range of concentrations. Administration of a CXCL5 blocking Ab 2 h before LPS challenge attenuated LPS-induced WBC increase and neutrophil sequestration in lungs at 8 and 24 h in a concentration-dependent manner (Figure 1)

, whereas no attenuation was observed with the pretreatment of isotype-matched control Ab (Figure 1). Furthermore, Anti-CXCL5 Ab or isotype-matched control Ab without LPS challenge at the maximal concentration (90 μg) did not influence total WBC or neutrophil counts either at 8 or 24 h (data not shown). These observations suggest that CXCL5 exerts substantial non-redundant effects on neutrophil trafficking in the lung during LPS-induced inflammation.

To determine the cells that expressed CXCL5 protein in the lung in vivo, we performed immunohistochemical analysis in normal mouse and rat lung and in lungs of LPS-aerosolized mouse. Twenty-four hours after challenge, CXCL5 immunoreactivity was not detected in saline-aerosolized (normal) lung (Figure 2A)

, whereas it was detected in lungs of LPS-aerosolized animals (Figure 2E). The specific signal was noted in the alveolus of the lung and was not observed in bronchial epithelial cells, alveolar epithelial type I-like (AEI) cells, neutrophils or macrophages (Figure 2E). Positively labeled cells appear to be AEII cells based on their location in the alveolus and proSP-C staining (Figure 2F). We have included several controls to ascertain that this immunoreactivity was specific for CXCL5: (1) the absence of immunoreactivity seen in sections where isotype-matched control serum was used (Figures 2C and 2G); and (2) no reactivity noted when primary anti-CXCL5 Ab was not included in the immunohistochemistry procedure (data not shown). In a similar manner, immunoreactivity was noted only in LPS-aerosolized rat lungs at 24 h, and the specific reactivity was observed in AEII cells based on their location (Figures 3A–3B) . No immunoreactivity was observed in LPS-challenged or saline-challenged lungs when isotype-matched control Ab was used (Figures 3C–3D). These data demonstrate that AEII cells are the primary source of CXCL5 in the lungs of mice and rats.

To confirm that AEII cells are able to express CXCL5 at the transcriptional and translational levels, we measured the levels of CXCL5 mRNA and protein in primary culture. We used cultured AEII cells of rat origin instead of mouse origin because (1) immunostaining observations reveal the same cellular origin of CXCL5 (Figure 3), and (2) they are well characterized and recovered in sufficient yield to perform all of our experiments. A very low level of CXCL5 mRNA was detected in unstimulated AEII- or AEI-like cells stimulated with LPS (Figure 4A)

. On the other hand, an 8.7-fold increase of CXCL5 was noted in AEII cells to 30 ng of LPS at 2 h (Figure 4A). We also measured immunoreactive CXCL5 in the cells and supernatant of rat AEII cells using Western blot and ELISA, respectively (Figures 4B and 4C). We did not notice significant CXCL5 protein expression in unstimulated AEII cells or LPS-stimulated AEI-like cells (Figure 4B). However, AEII cells secrete a large amount of CXCL5 in response to LPS in a concentration-dependent manner, with maximal expression at 30 ng of LPS (Figure 4C). Therefore, we used 30 ng of LPS in all of our subsequent experiments. To validate our immunohistochemical observations that myeloid cells do not express CXCL5, we used isolated murine neutrophils and rat alveolar macrophages. We did not observe CXCL5 protein secretion in isolated primary alveolar macrophages (Figure 4D) or neutrophils in response to LPS (Figure 4E). These results show that isolated AEII cells, but not myeloid cells, predominantly express CXCL5. To further demonstrate that AEII cells are the predominant source of CXCL5, we performed immunocytochemical analysis in AEII cells 18 h after LPS stimulation to validate the observation that AEII cells do produce CXCL5. We did not detect CXCL5 secretion in unstimulated AEII cells (Figure 5) . By contrast, 35% of the cultured AEII cells produce CXCL5 (Figure 5). In our control experiments where isotype control Ab was used instead of primary Ab, no immunoreactivity was observed in AEII cells (Figure 5).

To gain mechanical insights of LPS-induced CXCL5 expression in the lung at the molecular level, we examined the role of TLR4-dependent signaling in the expression of CXCL5 in the lung after LPS challenge. We particularly wished to delineate the role of the pattern recognition receptor TLR4, and its adaptor molecules MyD88 and TIRAP in the TLR4 signaling cascade using TLR4^−/−, MyD88^−/−, and TIRAP^−/− mice. In comparison with their wild-type (WT) mice, TLR4^−/−, MyD88^−/−, and TIRAP^−/− mice showed very low protein expression of CXCL5 in BALF (Figures 6A and 6B)

and lung homogenates (Figures 6C–6D) in response to LPS.

Because TLR4-dependent signaling activates JNK and p38 kinases downstream, we also examined their role in the expression of CXCL5 in the lung using inhibitors for JNK and p38 kinases 2 h before LPS challenge in mice. Inhibitors of both JNK and p38 kinases significantly attenuated CXCL5 expression in the lung (Figures 6E–6H). These observations indicate that the TLR4-dependent signaling is important in CXCL5 expression in the lung.

To demonstrate whether the TLR4 signaling cascade is required for mediating CXCL5 expression in isolated primary AEII cells in response to LPS, we performed experiments at the transcriptional and translational levels. Our RT-PCR and Western blot data demonstrate the presence of TLR4, MyD88 and TIRAP in isolated primary AEII cells both at mRNA and protein levels (Figure 7A)

. AEII cells were pretreated with 40 μg of anti-TLR4 Ab, and 10 μM of p38 (M39) and 5 μM of JNK kinase (SP600125) inhibitors 2 h before 30 ng of LPS stimulation. Anti-TLR4 Ab abolished LPS-induced CXCL5 expression from AEII cells (Figure 7B). By contrast, isotype-matched control to each Ab had no inhibitory effect on CXCL5 protein secretion induced by LPS (Figure 7B). Furthermore, both p38 and JNK kinase inhibitors attenuated LPS-induced CXCL5 protein secretion (Figure 7B). The vehicle (DMSO) had no inhibition in LPS-induced CXCL5 protein level (Figure 7B). Collectively, these data demonstrate that LPS-induced expression of CXCL5 by AEII cells requires activation of the TLR4 signaling cascade.