Hypercapnic Acidosis Attenuates Endotoxin-induced Acute Lung Injury

Hypercapnic Acidosis Attenuates Endotoxin-Induced Nuclear Factor-␬B Activation Kei Takeshita, Yukio Suzuki, Kazumi Nishio, Osamu Takeuchi, Kyoko Toda, Hiroyasu Kudo, Naoki Miyao, Makoto Ishii, Nagato Sato, Katsuhiko Naoki, Takuya Aoki, Koichi Suzuki, Rika Hiraoka, and Kazuhiro Yamaguchi Department of Medicine, School of Medicine, Keio University, Tokyo; and Departments of Medicine and Biomedical Research, Kitasato Institute Hospital, Tokyo, Japan

Although permissive hypercapnia improves the prognosis of patients with acute respiratory distress syndrome, it has not been conclusively determined whether hypercapnic acidosis (HA) is harmful or beneficial to sustained inflammation of the lung. The present study was designed to explore the molecular mechanism of HA in modifying lipopolysaccharide (LPS)-associated signals in pulmonary endothelial cells. LPS elicited degradation of inhibitory protein ␬B (I␬B)-␣, but not I␬B-␤, resulting in activation of nuclear factor (NF)-␬B in human pulmonary artery endothelial cells. Exposure to HA significantly attenuated LPS-induced NF-␬B activation through suppressing I␬B-␣ degradation. Isocapnic acidosis and buffered hypercapnia showed qualitatively similar but quantitatively smaller effects. HA did not attenuate the LPS-enhanced activation of activator protein-1. Following the reduced NF-␬B activation, HA suppressed the mRNA and protein levels of intercellular adhesion molecule-1 and interleukin-8, resulting in a decrease in both lactate dehydrogenase release into the medium and neutrophil adherence to LPS-activated human pulmonary artery endothelial cells. In contrast, HA did not inhibit LPS-enhanced neutrophil expression of integrin, Mac-1. Based on these findings, we concluded that hypercapnic acidosis would have anti-inflammatory effects essentially through a mechanism inhibiting NF-␬B activation, leading to downregulation of intercellular adhesion molecule-1 and interleukin-8, which in turn inhibits neutrophil adherence to pulmonary endothelial cells.

The mortality of acute respiratory distress syndrome (ARDS) remains unacceptably high at 40–60% (1). Most patients with ARDS undergo mechanical ventilation, and one of the most important concepts in the recent treatment of patients with ARDS is the recognition that a classical setting of mechanical ventilation with a relatively large tidal volume can worsen or even cause new lung injury (2). To

(Received in original form July 21, 2002 and in revised form January 28, 2003) Address correspondence to: Kazuhiro Yamaguchi, M.D., Department of Medicine, School of Medicine, Keio University, 35 Shinanomachi, Shinjukuku, Tokyo 160-8582, Japan. E-mail: [email protected] Abbreviations: activator protein-1, AP-1; acute respiratory distress syndrome, ARDS; buffered hypercapnia, BH; Dulbecco’s phosphate-buffered saline, DPBS; endothelial cell growth medium, EGM; glyceraldehyde3-phosphate dehydrogenase, G3PDH; hypercapnic acidosis, HA; human pulmonary artery endothelial cells, HPAEC; isocapnic acidosis, IA; intercellular adhesion molecule-1, ICAM-1; inhibitory protein ␬B, I␬B; I␬B kinase, IKK; interleukin-8, IL-8; lactate dehydrogenase, LDH; normocapnia, NC; nuclear factor-␬B, NF-␬B; carbon dioxide tension, PCO2; phycoerythrin, PE; tumor necrosis factor-␣, TNF-␣; TNF receptor–associated factor 6, TRAF6. Am. J. Respir. Cell Mol. Biol. Vol. 29, pp. 124–132, 2003 Originally Published in Press as DOI: 10.1165/rcmb.2002-0126OC on February 6, 2003 Internet address: www.atsjournals.org

prevent such ventilator-associated lung injury, mechanical ventilation with a low tidal volume has been introduced as an effective strategy for the management of ARDS (1). This hypoventilation technique not only reduces airway pressure, but also gives rise to a significant increase in carbon dioxide tension (PCO2) and a decrease in pH of the arterial blood, which is referred to as permissive hypercapnia (3). Some randomized clinical trials demonstrated that the hypoventilation technique may improve the survival rate in patients with ARDS, and its positive effects have been attributed to prevention of excessive stretch of the lung tissue (4, 5). In this context, hypercapnic acidosis (HA) has been regarded as a by-product to be permitted or even buffered. Respiratory frequency was increased, and sodium bicarbonate was infused to maintain normal arterial PCO2 and pH in a recently published large clinical trial conducted in North America, i.e., the ARDS Network (5). This indicates that HA was implicitly taken as a harmful factor in the ARDS Network study, though there have been few authentic studies confirming the harmful effects of HA on seriously injured lungs. On the contrary, administration of sodium bicarbonate is usually avoided in other situations with acidosis such as cardiopulmonary arrest (6). In sporadic experiments, HA has been shown to attenuate the injury caused by ischemia/reperfusion in an isolated rabbit lung model (7), and buffering pH with sodium bicarbonate significantly worsens the injury (8). HA has also been reported to protect the rat brain against ischemic stroke (9), indicating that HA may have protective effects on the inflammatory process in critically damaged organs (10). Unfortunately, however, the above studies did not address the molecular mechanisms regarding possible effects of HA on decreasing sustained inflammation in various organs. We previously reported the preliminary data showing the possibility that HA would attenuate nuclear factor-␬B (NF-␬B) activation in human pulmonary artery endothelial cells (HPAEC) (11). Extending our previous work, we attempted, in the present study, to thoroughly analyze the molecular basis of HA for modifying DNA-binding activity of transcription factors, including NF-␬B and activator protein-1 (AP-1) in HPAEC upon activation by endotoxin. In addition, we examined the effects of HA on the endotoxin-induced modification of endothelial expression of inhibitory protein ␬B-␣ (I␬B-␣) and I␬B-␤, pivotal proteins inhibiting the translocation of NF-␬B into the nucleus. Subsequently, we analyzed the effects of HA on the endothelial expression of proinflammatory proteins that are thought to be mediated by the NF-␬B pathway, i.e., intercellular

Takeshita, Suzuki, Nishio, et al.: Anti-Inflammatory Effect of Hypercapnic Acidosis

adhesion molecule-1 (ICAM-1) and interleukin-8 (IL-8). Finally, we addressed the issue of whether HA would alter the expression of neutrophil Mac-1, a ligand for ICAM-1, and the interaction between neutrophils and endotoxinactivated endothelial cells.

TABLE 1

Transitional changes in medium pH and PCO2 Condition

NC

Materials and Methods Endothelial Cell Culture HPAEC obtained from four donors were purchased from Kurabo (Osaka, Japan) as cryopreserved fourth-passage culture and used at passages 7–12. The cells were cultured in an endothelial cell growth medium (EGM; Kurabo) supplemented with 10% fetal calf serum, penicillin G (100 U/ml) and streptomycin (100 ␮g/ml; GIBCO BRL, Grand Island, NY) at 37⬚C in a humidified incubator saturated with a gas mixture containing 21% O2 and 5% CO2 in N2. The endothelial cells were grown on 75-cm2 tissue culture flasks (Corning, New York, NY), and then subcultured with 0.025% trypsin-0.05 mM ethylenediamine tetraacetic acid (EDTA; GIBCO BRL). Endothelial cells were identified by their pavement-like monolayer appearance and positive immunofluorescent staining with a specific anti-human von Willebrand factor antibody (DAKO Japan, Tokyo, Japan). Cell viability always exceeded 95% as determined by the trypan-blue exclusion test.

Exposure of Endothelial Cell Monolayer to Hypercapnic Gas First, the EGM in 10-cm2 culture dishes without HPAEC were equilibrated with either normocapnic gas (21% O2 and 5% CO2 in N2) or hypercapnic gas (21% O2 and 10% CO2 in N2) for 6 h at 37⬚C in a humidified multi-gas incubator (APM-36; ASTEC, Fukuoka, Japan). Subsequently, the equilibrated medium was quickly applied onto confluent HPAEC monolayers, which were continued to be exposed to either normocapnic or hypercapnic gas for up to 24 h. Simultaneously, lipopolysaccharide (LPS: E. coli O55:B5; Sigma, St. Louis, MO) dissolved in Dulbecco’s phosphatebuffered saline (DPBS) was administered to the medium at a final concentration of 1 ␮g/ml. Under normocapnic conditions, endothelial cell viability was more than 95% without LPS and 92 ⫾ 3% after 24 h of exposure to LPS. The viability of endothelial cells in HA at 24 h was 93 ⫾ 4% without LPS and 90 ⫾ 4% with LPS. Under phase-contrast microscopy, the endothelial monolayer was morphologically normal under each experimental condition. The pH and PCO2 values in EGM before application to HPAEC were, respectively, 7.34 ⫾ 0.02 and 35.5 ⫾ 0.6 mm Hg upon exposure to normocapnic gas, and the medium pH was significantly lowered only at the time points of 12 and 24 h after incubation with confluent HPAEC (Table 1). The medium PCO2 was not altered at any time point of observation. Qualitatively the same tendency was obtained in hypercapnic gas; i.e., the pH and PCO2 in the medium without HPAEC were, respectively, 7.01 ⫾ 0.01 and 74.7 ⫾ 2.8 mm Hg, and the medium pH was appreciably reduced 12 and 24 h later, with little alteration in PCO2 during incubation with HPAEC (Table 1). No significant difference in the medium pH was observed within 3 h after exposure to normocapnic or hypercapnic gas. In other series of experiments, we adjusted the medium pH and PCO2 either to isocapnic acidosis (IA) by adding 20 ␮l of 1 M HCl to the medium (5 ml) equilibrated with normocapnic gas or to normal pH with hypercapnia, i.e., buffered hypercapnia (BH), by adding 150 ␮l of 7.5% NaHCO3 to the medium (5 ml) during exposure to hypercapnic gas. There was no difference in the medium pH between zero and 1 h in both IA and BH (Table 1). The osmolality was 300 ⫾ 2 mOsm/kg H2O under normocapnia

125

HA

IA BH

Time (h)

0 1 3 6 12 24 0 1 3 6 12 24 0 1 0 1

pH

7.34 7.33 7.31 7.30 7.28 7.20 7.01 6.98 6.95 6.94 6.93 6.87 7.02 6.99 7.32 7.30

⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

PCO2 (mm Hg)

0.02 0.02 0.01 0.01 0.02* 0.01* 0.01 0.02 0.02 0.01 0.02* 0.01* 0.02 0.02 0.02 0.01

35.5 34.4 35.4 35.4 34.5 35.5 74.7 73.5 74.8 75.1 74.7 74.6 36.3 35.3 75.2 73.9

⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

0.6 1.9 0.2 0.3 1.2 0.8 2.8 1.6 0.9 0.8 0.9 0.7 1.4 0.8 1.5 1.4

n ⫽ 3 for each condition, mean ⫾ SD. * P ⬍ 0.05 versus Time zero.

(NC), HA, and IA conditions, while the addition of NaHCO3 appreciably increased the osmolality under BH condition up to 327 ⫾ 3 mOsm/kg H2O.

Electrophoretic Mobility Shift Assay for NF-␬B and AP-1 Nuclear extracts were prepared as previously described (12) and protein concentrations were determined. Then, 7.5 ␮g nuclear protein was incubated with double-stranded 32P-labeled oligonucleotides containing the human consensus binding sequence for NF-␬B or AP-1 (Promega, Madison, WI) in binding buffer, as described previously (12). DNA–protein complexes were then resolved on 4% PAGE gel in 0.5 M Tris-borate-EDTA buffer. Dried gels were visualized by autoradiography (n ⫽ 4 for NF-␬B and for AP-1). DNA-binding activity of NF-␬B and AP-1 was examined at 0.5, 1, 2, and 3 h after initiation of LPS stimulation.

Western Immunoblot Analysis of I␬B-␣ and I␬B-␤ For immunoblot analysis, 100-mm2 dishes containing confluent HPAEC with or without LPS were exposed to either normocapnic or hypercapnic gas for various time periods, and then the cells were washed with DPBS and lysed on ice in modified radioimmunoprecipitation assay buffer containing PMSF solution and aprotinin. The cellular debris was removed by centrifugation at 15,000 rpm for 20 min, and 10 ␮g of each sample was run on 10% sodium dodecyl sulfate/polyacrylamide gel electrophoresis, and transferred onto an immobilon-polyvinylidene difluoride membrane (Immobilon-P; Millipore, Bedford, MA). The membranes were blotted with an antibody against I␬B-␣ or I␬B-␤ (Santa Cruz Biotechnology, Santa Cruz, CA), and the bands were visualized with a horseradish peroxidase-conjugated secondary antibody followed by an enhanced chemiluminescence detection system (Pierce, Rockford, IL) (n ⫽ 4 for I␬B-␣ and for I␬B-␤). The expression of I␬B proteins was examined 0.5, 1, 2, and 3 h after LPS administration.

RT-PCR Analysis for ICAM-1 and IL-8 LPS-induced mRNA expression of ICAM-1 and IL-8 was analyzed with a semiquantitative RT-PCR. Total RNA was extracted from

126

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 29 2003

endothelial cells using Trizol (GIBCO BRL). First-strand cDNA was synthesized from 10 ␮g RNA by SuperScript RT (GIBCO BRL). PCR reactions were performed in a thermocycler (PerkinElmer Cetus, Norwalk, CT) with 50 pmol of the 5⬘ and 3⬘ primers with 2.5 U Taq polymerase (Takara Biomedicals, Kyoto, Japan) in a total volume of 50 ␮l. The reaction buffer consisted of 10 mM Tris-HCl, 50 mM KCl, 1.5 mM MgCl2, 0.01% gelatin, and 10 mM deoxynucleotide triphosphates. PCR cycles were allowed to run for 30 s at 94⬚C, followed by a 30-s run at 55⬚C and a 1-min run at 72⬚C, and final extension at 72⬚C for 10 min. The specific primer pairs used for PCR amplification were 5⬘-TGACCATCTACA GCTTTCCGCC-3⬘ and 5⬘-GTCTGAGGTTACACGGTCCGA-3⬘ for ICAM-1, 5⬘-ATGACTTCCAAGCTGGCCGTGCT-3⬘ and 5⬘TCTCAGCCCTCTTCAAAAACTTCTC-3⬘ for IL-8, and 5⬘-TGA AGGTCGGAGTCAACGGATTTGGT-3⬘ and 5⬘-CAT GTGGG CCATGAGGTCCACCAC-3⬘ for glyceraldehyde-3-phosphate dehydrogenase (G3PDH). A 10-␮l aliquot of the amplified cDNA reaction mixture was separated by 2% agarose gel electrophoresis and then visualized by ultraviolet fluorescence after being stained with ethidium bromide (n ⫽ 4 for ICAM-1 and IL-8). To quantify the levels of mRNA, a standard curve was constructed by titrating RNA harvested from LPS-stimulated endothelial cells. For semiquantitative evaluation of ICAM-1, IL-8, and G3PDH mRNA, 28, 28, and 24 cycles, respectively, were selected. The specificity of the amplified products was validated from their predicted sizes on agarose gel. Expression of mRNA of ICAM-1 and IL-8 was assessed before and 6 and 12 h after LPS stimulation.

Expression of ICAM-1 on HPAEC Flow cytometry was applied to detect ICAM-1 expression on HPAEC. Endothelial cells were detached by treatment with 0.1% EDTA for 1 min at 37⬚C and washed with DPBS. The suspended endothelial cells were incubated for 30 min at 4⬚C with phycoerythrin (PE)-conjugated anti-human ICAM-1 monoclonal antibody (Becton Dickinson, San Jose, CA). The cells were washed three times with DPBS and fixed with 1% paraformaldehyde. The intensity of fluorescence and the light-scattering properties of the cells were examined with a FACScan flow cytometry system equipped with an argon laser (488-nm emission, 15-mW output; Becton Dickinson). PE red fluorescence was detected between 564 and 606 nm with a band-pass filter. The analysis was run simultaneously with mouse isotype control antibody (IgG2; Becton Dickinson), and the values thus obtained were subtracted. In each sample, 10,000 endothelial cells were examined. The results were expressed as the percent intensity of fluorescence relative to that under NC without LPS stimulation (n ⫽ 8). Protein levels of endothelial ICAM-1 were examined at 24 h upon exposure to LPS.

IL-8 in Medium Using the supernatants obtained from HPAEC culture media exposed to NC or HA for 24 h with or without LPS, we measured the concentration of IL-8 by means of a commercially available enzyme-linked immunosorbent assay kit (R&D Systems, Abingdon, Oxon, UK) (n ⫽ 5).

Measurement of Lactate Dehydrogenase Release into Medium To assess the endothelial cell injury caused by LPS stimulation, we measured lactate dehydrogenase (LDH) activity in the culture medium. After exposure to NC or HA with or without LPS for 24 h, the medium was centrifuged at 3,000 rpm for 10 min, and

the supernatant was collected for LDH activity measurement by spectrophotometric assay of NADH oxidation (13) (n ⫽ 5).

Neutrophil Adherence to Endothelial Cell Monolayer HPAEC in 6-well tissue culture plates (Corning) were exposed to either normocapnic or hypercapnic gas with or without LPS in a humidified multi-gas incubator for 24 h, and the culture medium was replaced with new medium equilibrated with normocapnic gas. Consequently, 100 ␮l of isolated neutrophils (5 ⫻ 105 cells/ ml) was introduced into each well. The plates were then incubated for 2 h at 37⬚C under normocapnic conditions in a humidified incubator. Nonadherent neutrophils were removed by gently washing the plates three times with prewarmed DPBS. Ten randomly selected fields were read at ⫻200 magnification under a light microscope. Neutrophil adhesion was evaluated by counting the number of neutrophils adhering to the endothelial cell monolayer (n ⫽ 8). In addition, the contribution of ICAM-1 to neutrophil adhesion was assessed by incubating endothelial cells with 10 ␮g/ml antihuman ICAM-1 (Santa Cruz) for 30 min before applying neutrophils (n ⫽ 8).

Expression of Mac-1 (CD11b/CD18) on Neutrophils Neutrophils were isolated from healthy adult volunteers (n ⫽ 5) and separated on a discontinuous gradient consisting of Histopaque 1,077 and 1,119 (Sigma). The neutrophils were suspended in DPBS at a final concentration of 5 ⫻ 105 cells/ml (purity ⬎ 98% as confirmed by modified Wright’s stain). To investigate the effects of HA on LPS-induced Mac-1 expression on neutrophils, isolated neutrophils were exposed to either normocapnic or hypercapnic gas with or without LPS stimulation for 2 h. Subsequently, the neutrophils were incubated with PE-conjugated anti-human CD11b or CD18 monoclonal antibody (Ancell, Bayport, MN) for 30 min at 4⬚C. Cells were washed three times with DPBS and fixed with 1% paraformaldehyde. The fluorescence intensity and the lightscattering properties of the cells were determined utilizing a FACScan flow cytometry system, with reference to the mouse isotype IgG as the control (IgG1 for CD11b; IgG2 for CD18). In each sample, 10,000 neutrophils were examined. Values were expressed as the percent intensity of fluorescence relative to that observed under NC without LPS stimulation.

Statistical Analysis Statistical difference in medium pH at different time points under NC or HA conditions was judged by one-way ANOVA followed by multiple comparisons of Fisher’s protected least-significantdifference test. Difference in medium pH under IA and BH conditions before and after incubation with HPAEC was judged by unpaired t test. Statistical significance was assessed by two-way ANOVA followed by multiple comparisons by Fisher’s protected test applied for ICAM-1 expression, medium IL-8 concentration, Mac-1 expression, medium LDH concentration, and number of adherent neutrophils. Values were expressed as mean ⫾ SD, with P ⬍ 0.05 considered to be statistically significant.

Results Effects of HA on LPS-induced Activation of Transcription Factors DNA-binding activity of NF-␬B in HPAEC was detectable under control conditions (i.e., NC without LPS). NF-␬B– DNA-binding activity under NC was significantly enhanced

Takeshita, Suzuki, Nishio, et al.: Anti-Inflammatory Effect of Hypercapnic Acidosis

127

Figure 1. (A ) Transitional changes in LPSinduced DNA-binding activity of NF-␬B under NC and HA examined by gel shift assay. CTL: NF-␬B–DNA-binding activity under control conditions without LPS (NC). Observation time: 0.5–3 h after start of LPS stimulation. See text for further details. (B ) Differential effects of HA, IA, and BH on LPS-induced DNA binding activity of NF-␬B. HPAEC were exposed to each condition for 1 h with LPS stimulation. HA, IA, and BH attenuated LPSinduced NF-␬B activation, but HA showed the greatest effect. (C ) Nuclear extracts from HPAEC without LPS (lane 1) and with 1-h LPS stimulation (lanes 2–5) under NC. Lane 3: presence of 100-fold excess of unlabeled NF-␬B. Lane 4: presence of antip50 monoclonal antibody. Lane 5: presence of anti-p65 monoclonal antibody.

at 0.5 h after LPS stimulation, and increased further at 1, 2, and 3 h (Figure 1A). This LPS-induced augmentation of NF-␬B–DNA-binding activity was substantially attenuated by HA exposure at all time points studied (Figure 1A). To elucidate whether the inhibitory effect of HA on NF␬B–DNA-binding activity can be ascribed to low pH, high PCO2 or both, endothelial cells were exposed, for 1 h, to IA (low pH and normal PCO2) or BH (normal pH and high PCO2) with LPS stimulation. Although both IA and BH inhibited LPS-induced NF-␬B–DNA binding, the degree of inhibition under the conditions of IA or BH was small in comparison with the inhibition by HA (Figure 1B). The specificity of NF-␬B–DNA binding was confirmed by almost complete displacement of NF-␬B–DNA complex in the presence of a 100-fold molar excess of unlabeled NF-␬B and supershift after applying the monoclonal antibody against p50 or p65 (Figure 1C). Similar to the case of NF-␬B, administration of LPS enhanced the DNA binding activity of AP-1 at 0.5, 1, 2, and 3 h under normocapnic conditions (Figure 2A). However, differing from NF-␬B, HA did not reduce the LPS-induced DNA-binding activity of AP-1 (Figure 2A). The specificity of AP-1 binding was also confirmed by a complete displacement of AP-1–DNA complex in the presence of a large excess of unlabeled AP-1 (Figure 2B). Inhibition of LPS-Induced I␬B-␣ Degradation by HA LPS stimulation decreased the protein level of I␬B-␣ within 1 h, but it was restored to the baseline value after 2 and 3 h under normocapnic conditions (Figure 3A). In contrast, LPS stimulation caused only modest degradation of I␬B-␣ under HA at any time point studied (Figure 3A). I␬B-␤ expression was not altered by LPS stimulation under NC, and was not modified upon exposure to HA (Figure 3B). To examine whether low pH or high PCO2 is more important in the inhibitory effect of HA on the LPS-induced I␬B-␣ degradation, endothelial cells were exposed to IA or BH for 1 h in the medium containing LPS. We found

that LPS-induced degradation of I␬B-␣ was attenuated both in IA and in BH, but the degree of attenuation caused by IA or BH was smaller than that observed for HA (Figure 3C). Downregulation of LPS-Induced ICAM-1 and IL-8 mRNA Expression by HA mRNA of ICAM-1 and IL-8 was detected under control conditions without LPS stimulation. LPS augmented endothelial ICAM-1 mRNA expression at 6 and 12 h under NC, whereas this LPS-induced ICAM-1 mRNA augmentation was inhibited by exposure to HA (Figure 4).

Figure 2. (A ) Changes in LPS-induced DNA binding activity of AP-1 under NC and HA examined by gel shift assay. CTL: AP1-DNA binding activity under control conditions without LPS (NC). Observation time: 0.5–3 h after start of LPS stimulation. See text for further explanation. (B ) Nuclear extracts from HPAEC without LPS (lane 1) and with 1-h LPS stimulation (lanes 2 and 3 ) under NC. Lane 3: presence of 100-fold excess of unlabeled AP-1.

128

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 29 2003

Figure 3. (A and B ) Transitional changes in expression of I␬B-␣ (A ) and I␬B-␤ (B ) proteins upon LPS stimulation under NC and HA examined by Western blot. CTL: I␬B proteins under control conditions without LPS (NC). Observation time: 0.5–3 h after start of LPS administration. (C ) Differential effects of HA, IA, and BH on LPS-induced I␬B-␣ degradation. HPAEC were exposed to each condition for 1 h with LPS stimulation. Although IA as well as BH reduced LPS-induced I␬B-␣ degradation, the inhibitory effects were modest as compared with that of HA.

Upregulation of IL-8 mRNA was observed after 6 and 12 h of LPS stimulation under NC, but this upregulation was attenuated in cells exposed to HA. IL-8 mRNA level was significantly lower at 12 h than at 6 h under both NC and HA. Decrease in LPS-Induced Production of ICAM-1 and IL-8 by HA The absolute fluorescence intensities (AFI) observed under NC conditions without LPS administration were taken as the measure representing the baseline (control) expression of ICAM-1 and found to be 77.6 ⫾ 14.7. The AFI after LPS administration was corrected for the baseline AFI. The relative fluorescence intensities thus calculated were used for estimating the augmented expression of ICAM-1 upon LPS stimulation under each experimental condition.

Figure 5. (A ) Flow cytometric analysis of ICAM-1 expression on HPAEC, which were exposed, for 24 h, to either baseline (control) conditions (NC without LPS) (1), HA without LPS (2), NC with LPS (3), or HA with LPS (4) (n ⫽ 8 for each condition). Data are expressed as percent intensity of fluorescence compared with the baseline value (mean ⫾ SD). †P ⬍ 0.001 compared with expression without LPS. *P ⬍ 0.01 compared with expression under NC. (B ) IL-8 concentration in medium with and without LPS stimulation after exposure to NC and HA for 24 h examined by ELISA. 1 and 2: IL-8 concentration in medium under NC and HA without LPS, respectively (n ⫽ 5 for each condition). 3 and 4: IL-8 concentration in medium under NC and HA with LPS stimulation, respectively (n ⫽ 5 for each condition). Open bars, NC; filled bars, HA. Values are mean ⫾ SD. †P ⬍ 0.001 compared with value obtained under conditions without LPS. *P ⬍ 0.001 versus value obtained under NC with LPS.

Appreciable expression of endothelial surface ICAM-1 was detected under normocapnic and HA conditions without LPS stimulation (Figure 5A). A 7-fold increase in surface expression of ICAM-1 (719.4 ⫾ 134.3%; P ⬍ 0.001) was observed after 24 h of LPS stimulation under NC as compared with that obtained under baseline conditions without LPS. HA attenuated the LPS-induced enhancement of ICAM-1 expression (506.9 ⫾ 87.1%; P ⬍ 0.01) (Figure 5A). Exposure to HA without LPS for 24 h did not significantly alter ICAM-1 expression (82.5 ⫾ 5.5%). Without LPS stimulation, the medium concentration of IL-8 was 10.1 ⫾ 2.3 ng/ml under NC and 11.6 ⫾ 0.4 ng/ml under HA. Although LPS stimulation markedly increased IL-8 concentration under both NC (87.4 ⫾ 3.2 ng/ml; P ⬍ 0.001) and HA (42.3 ⫾ 5.9 ng/ml; P ⬍ 0.001), the latter was considerably lower than the former (P ⬍ 0.001) (Figure 5B). Effect of HA on LPS-Induced Endothelial Cell Injury LDH release into the medium without LPS stimulation did not differ between NC (42.8 ⫾ 5.1 IU/liter) and HA (40.3 ⫾ 5.9 IU/liter). Although LPS administration considerably enhanced LDH release under normocapnic conditions (123.0 ⫾ 3.9 IU/liter; P ⬍ 0.001), this LDH release was obviously inhibited by exposure to HA (83.8 ⫾ 7.4 IU/liter; P ⬍ 0.001).

Figure 4. RT-PCR analysis of mRNA expression of ICAM-1 and IL-8 in HPAEC under NC (left panel) and HA (right panel). CTL: without LPS under NC. RT-PCR analysis was performed before and 6 and 12 h after LPS administration under both conditions. G3PDH was used as the reference.

Effects of HA on Neutrophil Adherence to LPS-Activated Endothelium When LPS was not introduced, the number of neutrophils adhering to the endothelial cell monolayer did not differ

Takeshita, Suzuki, Nishio, et al.: Anti-Inflammatory Effect of Hypercapnic Acidosis

Figure 6. Neutrophil adherence to LPS-activated endothelial cells after 24-h exposure to NC and HA. 1 and 2: number of adherent neutrophils under NC and HA without LPS, respectively (n ⫽ 8 for each condition). 3 and 4: number of adherent neutrophils under NC and HA with LPS, respectively (n ⫽ 8 for each condition). 5 and 6: number of adherent neutrophils under NC and HA in the presence of both LPS and anti–ICAM-1 antibody, respectively (n ⫽ 8 for each condition). Open bars, NC; filled bars, HA. Values are mean ⫾ SD. †P ⬍ 0.05 compared with value obtained under conditions without LPS. *P ⬍ 0.05 versus value under NC with LPS. #P ⬍ 0.05 versus value in absence of anti–ICAM-1 antibody.

between NC and HA at 24 h. Although LPS stimulation for 24 h substantially augmented neutrophil adherence to the endothelial monolayer under NC, this was significantly inhibited under HA (Figure 6). In the presence of anti– ICAM-1 blocking antibody, LPS-induced neutrophil adherence was reduced under both NC and HA, resulting in no significant difference between the two conditions. Although anti–ICAM-1 antibody effectively decreased the number of neutrophils adhering to the LPS-activated endothelial monolayer under both NC and HA, it was not restored to the baseline level observed under conditions without LPS. Effects of HA on LPS-induced Expression of Neutrophil Integrins The AFI of CD11b and CD18 under baseline conditions without LPS averaged 346 ⫾ 75.7 and 50.2 ⫾ 6.5, respectively. The relative fluorescence intensities standardized for the baseline AFI were used as the quantitative measure of expression of these integrins upon LPS stimulation. Without LPS stimulation, exposure to HA for 2 h did not alter the expression of CD11b or CD18 in comparison with that observed under normocapnic conditions. LPS treatment caused a 3-fold increase in neutrophil surface expression of CD11b and CD18 under NC. HA exposure had little influence on LPS-induced neutrophil Mac-1 expression (Figure 7).

Discussion Critique of Methods There are a couple of crucial points in the present study. First, qualitatively the different interpretation of data should be made if the medium pH is considerably changed in the first few hours over which DNA-binding activities of transcription factors (NF-␬B and AP-1) and protein levels

129

Figure 7. LPS-induced neutrophil integrins under conditions with 2-h exposure to NC and HA examined by flow cytometry. Values (mean ⫾ SD) are expressed as percent fluorescence intensity relative to that obtained under baseline (control) conditions (NC without LPS). 1 and 2: NC and HA without LPS, respectively (n ⫽ 5 for each condition). 3 and 4: NC and HA with LPS, respectively (n ⫽ 5 for each condition). Open bars, NC; filled bars, HA. †P ⬍ 0.05 versus value obtained under conditions without LPS.

of I␬B were measured. Fortunately, however, the medium pH did not differ significantly between 0 and 3 h under both normocapnic and hypercapnic conditions with LPS stimulation (Table 1), suggesting that the medium pH would be fairly constant over the observation periods in which activities of transcription factors and protein levels of I␬B were examined. Based on these facts, we considered that the difference in LPS-evoked activation of NF-␬B–DNA binding and that in protein levels of I␬B between normocapnia and hypercapnia can be interpreted to be essentially induced by the difference in pH and/or PCO2 between the two conditions but not by the transitional change in pH and/or PCO2 under each experimental condition (Figures 1 and 3). Second, the increased osmolality in buffered hypercapnia may exert an influence on activities of transcription factors and/or protein levels of I␬B, because Gharzouli and coworkers have shown that hyperosmolality has a protective effect on the cell injury (14). However, Gharzouli and colleagues used the solution with an extremely high osmolality (600 mOsm/kg H2O), about the double of the osmolality established in buffered hypercapnia of the present study (327 mOsm/kg H2O). The difference in osmolality between buffered hypercapnia and other experimental conditions was only 27 mOsm/kg H2O in the present study. In addition, we have preliminarily demonstrated that the solution with a high concentration of glucose (330 mOsm/kg H2O) inhibits LPS-induced NF-␬B activation on HPAEC, but the mannitol solution with the same osmolality does not (Takeshita and coworkers, unpublished data). Thus, we consider that such a small difference in osmolality observed in buffered hypercapnia and other experimental conditions plays no important role in modifying the DNA-binding activity of NF-␬B and the expression of I␬B proteins. Inhibitory Effect of Hypercapnic Acidosis on LPS-Induced Activation of NF-␬B–Associated Signaling Pathway Protective effects of hypercapnic acidosis against oxidantmediated injury have recently been recognized in vivo. Shi-

130

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 29 2003

bata and coworkers demonstrated that hypercapnic acidosis reduced ischemia-reperfusion injury in isolated rabbit lungs (7). Laffey and colleagues showed that ventilating rabbits with hypercapnic gas reduced tumor necrosis factor (TNF)-␣ concentration in bronchoalveolar lavage fluid and nitrotyrosine production, as well as apoptosis in lung tissue (15). In accordance with the in vivo studies qualitatively (7, 15), we found that LPS-induced endothelial cell injury, judged from LDH release into the medium, was distinctly suppressed by exposure to hypercapnic acidosis. These protective effects of hypercapnic acidosis may be reasonably explained on the basis of the fact that DNA binding activity of NF-␬B is significantly inhibited by a mechanism mediated through a decrease in I␬B-␣ degradation under hypercapnic acidosis (Figures 1 and 3). LPS, a constituent of the cell wall of gram-negative bacteria, induces intracellular signaling through toll-like receptor 4 (16). Upon activation by LPS, toll-like receptor 4 subsequently recruits MyD88/IL-1 receptor-associated kinase, leading to recruitment of TNF receptor–associated factor 6 (TRAF6), which in turn activates downstream kinases. The involved kinases include NF-␬B–inducing kinase and mitogen-activated protein kinase/ERK kinase 1, each of which is capable of activating the I␬B kinase (IKK) complex. In addition to the NF-␬B pathway, LPS may also activate AP-1-related signals, because mitogen-activated protein kinase/ERK kinase 1 in association with TRAF6 can activate c-jun–NH2-terminal kinase (16). The findings obtained in the present study are highly consistent with those reported in the literature (16) concerning NF-␬B and AP-1 activation by LPS under normocapnic conditions (Figures 1 and 2). Interestingly, however, hypercapnic acidosis exerted little influence on LPS-induced AP-1 activation (Figure 2), though it obviously attenuated LPS-induced DNA-binding activity of NF-␬B (Figure 1). Taken together, these findings suggest that hypercapnic acidosis specifically inhibits the signal transduction associated with NF-␬B, probably at sites downstream from TRAF6. NF-␬B is found in the cytoplasm in an inactive form joined with regulatory proteins called I␬B, in which the important isoforms are I␬B-␣ and I␬B-␤ (17). I␬B proteins are phosphorylated by the IKK complex and subsequently degraded, thus allowing NF-␬B to translocate into the nucleus, bind to specific promoter sites, and activate target genes (17). The inhibitory effect of I␬B-␣ and I␬B-␤ on NF-␬B activation differs with cell type and insult on the cell. Zen and coworkers demonstrated that LPS induced I␬B-␣ degradation within 1 h, but did not cause I␬B-␤ degradation up to 2 h in human umbilical vein endothelial cells (18), consistent with our results obtained for HPAEC (Figure 3). The findings observed in the present study clearly demonstrated that hypercapnic acidosis reduces LPS-induced degradation of I␬B-␣, but it exerts little influence on I␬B-␤ degradation in HPAEC (Figure 3). These results suggest that the inhibitory effect of hypercapnic acidosis on LPS-induced NF-␬B activation is mainly caused by the suppression of degradation of I␬B-␣, but not I␬B-␤. The present study showed that isocapnic acidosis as well as buffered hypercapnia also attenuated LPS-induced NF-␬B activation and I␬B-␣ degradation, though the degree of

inhibition caused by either isocapnic acidosis or buffered hypercapnia was smaller than that observed for hypercapnic acidosis (Figures 1 and 3). The differential effects of hypercapnic acidosis and isocapnic acidosis on the NF-␬B activation pathway may be attributable to the difference in intracellular pH kinetics. In fact, we have recently demonstrated that the intracellular pH reduction of HPAEC is more rapid and stable under extracellular acidification by hypercapnia than that by HCl loading (19). In contrast to our findings, Bellocq and colleagues showed that isocapnic acidosis (pH ⫽ 7.0) established by addition of HCl into the medium would activate NF-␬B, leading to an increase in expression of the inducible isoform of nitric oxide synthase and TNF-␣ production in macrophages harvested from rat peritoneum (20). Xu and Fidler reported that isocapnic acidosis (pH ⫽ 6.6) achieved by 2-(N-morpholino) ethane-sulfonic acid and Tris treatment would enhance IL-8 transcription through activation of NF-␬B and AP-1 in ovarian cancer cells (21), though the behavior of I␬B proteins was not analyzed in these studies (20, 21). Although it is not easy to elucidate the precise reason for the difference in results between the present study and previous studies (20, 21), one possibility is that the cell species used in previous studies differed from those in the present study. To see whether the attenuation of NF-␬B activation by hypercapnic acidosis actually plays an important role in reducing LPS-induced inflammation, we investigated the mRNA and protein levels of ICAM-1 and IL-8, which are thought to be mainly regulated by the NF-␬B–related pathway (22, 23). The LPS-induced upregulation of mRNA expression of ICAM-1 and IL-8 in HPAEC was clearly inhibited by exposure to HA for 6 or 12 h (Figure 4), followed by a significant decrease in protein levels of these substances at 24 h (Figure 5). The findings of the present study is highly consistent with those reported by Coakley and associates, who demonstrated that intracellular acidification induced by hypercapnia inhibited LPS-evoked IL-8 release and oxidant production in human neutrophils (24). Neutrophil adherence to LPS-activated HPAEC was partially inhibited by hypercapnic acidosis, and the difference in neutrophil adherence to activated HPAEC between normocapnia and hypercapnic acidosis was abolished by pretreatment with anti–ICAM-1 monoclonal antibody (Figure 6), confirming that the inhibitory effect of hypercapnic acidosis on LPS-induced neutrophil adherence to an activated endothelial cell layer may be due at least partly to downregulation of the LPS-induced increase in endothelial ICAM-1 expression. In endotoxin-induced lung injury, endothelial ICAM-1 is upregulated and plays a critical role in the accumulation of neutrophils in the inflamed lung (25–27). Watanabe and coworkers reported that anti– ICAM-1 antibody prevented shock induced by LPS in rabbits (28), consistent with the findings in LPS-challenged ICAM-1–deficient mice (29). Endothelial ICAM-1 is particularly important for leukocyte kinetics in the intact lung and various types of pulmonary inflammation such as oxygen toxicity and bleomycin-induced injury (30–32). In addition, IL-8 from activated endothelial cells is crucial for neutrophil accumulation in inflamed tissue (33). The findings of the present study suggest that the LPS-

Takeshita, Suzuki, Nishio, et al.: Anti-Inflammatory Effect of Hypercapnic Acidosis

induced abnormal accumulation of inflammatory leukocytes in the pulmonary circulation mediated by endothelial ICAM-1 may be attenuated under sustained hypercapnic conditions. Effects of HA on Neutrophil Mac-1 Expression Mac-1 is an important adhesion molecule expressed on the neutrophil surface and acts as the counterpart of endothelial adhesion molecule, ICAM-1, leading to biological adherence of neutrophils along the endothelial cell layer in association with ICAM-1 (34). In contrast to endothelial ICAM-1, the expression of neutrophil integrins was not influenced by hypercapnic acidosis (Figure 7). Mac-1 is stored in intracellular granules of neutrophils, and the cell surface expression is rapidly upregulated upon exposure of neutrophils to various stimuli (34). The findings obtained in the present study suggest that hypercapnic acidosis exerts little influence on the signal pathways regulating Mac-1 expression on neutrophils. Clinical Significance Summarizing the experimental findings observed in the present study, hypercapnic acidosis inhibits endotoxin-induced neutrophil adherence to pulmonary endothelial cells through a specific molecular mechanism of attenuating I␬B-␣ degradation, which in turn neutralizes the DNA-binding activity of NF-␬B. However, hypercapnic acidosis does not appear to modify the DNA binding activity of AP-1. The expression of ICAM-1 and production of the chemoattractant IL-8 are distinctly decreased under hypercapnic acidosis. The hypercapnia-associated decrease in these signals finally suppresses endothelial cell injury as well as neutrophil adherence to an activated endothelial cell layer, though hypercapnic acidosis does not reduce the expression of Mac-1 on neutrophils. Although the protective effects of the hypoventilation technique for treating ARDS patients have been considered to be the consequence of a low tidal volume decreasing excessive mechanical stretch of lung tissue (4, 5), the findings of the present study indicate that the benefits are provided not only by decreased stretch, but also by coexisting hypercapnic acidosis having anti-inflammatory effects. These facts suggest that the protective effects of the hypoventilation technique during treatment of ARDS patients may be enhanced when coexisting hypercapnic acidosis is not corrected either by increasing respiratory frequency or by adding sodium bicarbonate. Acknowledgments: This work was partly supported by Keio University Grantin-Aid for Encouragement of Young Medical Scientists (to Dr. Kei Takeshita).

6. 7.

8. 9. 10. 11.

12.

13. 14.

15.

16. 17. 18. 19.

20.

21. 22. 23. 24.

25.

26.

References 1. Ware, L. B., and M. A. Matthay. 2000. The acute respiratory distress syndrome. N. Engl. J. Med. 342:1334–1349. 2. Dreyfuss, D., and G. Saumon. 1998. Ventilator-induced lung injury: lessons from experimental studies. Am. J. Respir. Crit. Care Med. 157:294–323. 3. Feihl, F., and C. Perret. 1994. Permissive hypercapnia: how permissive should we be? Am. J. Respir. Crit. Care Med. 150:1722–1737. 4. Amato, M. B., C. S. Barbas, D. M. Medeiros, R. B. Magaldi, G. P. Schettino, G. Lorenzi-Filho, R. A. Kairalla, D. Deheinzelin, C. Munoz, R. Oliveira, T. Y. Takagaki, and C. R. Carvalho. 1998. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N. Engl. J. Med. 338:347–354. 5. The Acute Respiratory Distress Syndrome Network. 2000. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute

27. 28.

29.

30.

131

lung injury and the acute respiratory distress syndrome. N. Engl. J. Med. 342:1301–1308. Arieff, A. I. 1998. Efficacy of buffers in the management of cardiac arrest. Crit. Care Med. 26:1311–1313. Shibata, K., N. Cregg, D. Engelberts, A. Takeuchi, L. Fedorko, and B. P. Kavanagh. 1998. Hypercapnic acidosis may attenuate acute lung injury by inhibition of endogenous xanthine oxidase. Am. J. Respir. Crit. Care Med. 158:1578–1584. Laffey, J. G., D. Engelberts, and B. P. Kavanagh. 2000. Buffering hypercapnic acidosis worsens acute lung injury. Am. J. Respir. Crit. Care Med. 161:141– 146. Vannucci, R. C., J. Towfighi, D. F. Heitjan, and R. M. Brucklacher. 1995. Carbon dioxide protects the perinatal brain from hypoxic-ischemic damage: an experimental study in the immature rat. Pediatrics 95:868–874. Laffey, J. G., and B. P. Kavanagh. 1999. Carbon dioxide and the critically ill: too little of a good thing? Lancet 354:1283–1286. Takeshita, K., Y. Suzuki, K. Nishio, T. Aoki, O. Takeuchi, K. Toda, N. Sato, K. Naoki, H. Kudo, and K. Yamaguchi. 1999. Hyperoxia and hypercapnic acidosis differentially alter nuclear factor-kappa B activation in human pulmonary artery endothelial cells. Adv. Exp. Med. Biol. 471:265–270. Suzuki, Y., K. Nishio, K. Takeshita, O. Takeuchi, K. Watanabe, N. Sato, K. Naoki, H. Kudo, T. Aoki, and K. Yamaguchi. 2000. Effect of steroid on hyperoxia-induced ICAM-1 expression in pulmonary endothelial cells. Am. J. Physiol. Lung Cell. Mol. Physiol. 278:L245–L252. Wroblewski, F., and J. S. LaDue. 1955. LDH activity in blood. Proc. Soc. Exp. Biol. Med. 90:210–213. Gharzouli, K., A. Gharzouli, S. Amira, and S. Khennouf. 2001. Protective effect of mannitol, glucose-fructose-sucrose-maltose mixture, and natural honey hyperosmolar solutions against ethanol-induced gastric mucosal damage in rats. Exp. Toxicol. Pathol. 53:175–180. Laffey, J. G., M. Tanaka, D. Engelberts, X. Luo, S. Yuan, A. K. Tanswell, M. Post, T. Lindsay, and B. P. Kavanagh. 2000. Therapeutic hypercapnia reduces pulmonary and systemic injury following in vivo lung reperfusion. Am. J. Respir. Crit. Care Med. 162:2287–2294. Zhang, G., and S. Ghosh. 2001. Toll-like receptor-mediated NF-␬B activation: a phylogenetically conserved paradigm in innate immunity. J. Clin. Invest. 107:13–19. Tak, P. P., and G. S. Firestein. 2001. NF-␬B: a key role in inflammatory diseases. J. Clin. Invest. 107:7–11. Zen, K., A. Karsan, T. Eunson, E. Yee, and J. M. Harlan. 1998. Lipopolysaccharide-induced NF-␬B activation in human endothelial cells involves degradation of I␬B-␣, but not I␬B-␤. Exp. Cell Res. 243:425–433. Yamaguchi, K., K. Suzuki, K. Naoki, K. Nishio, N. Sato, K. Takeshita, H. Kudo, T. Aoki, Y. Suzuki, A. Miyata, and H. Tsumura. 1998. Response of intra-acinar pulmonary microvessels to hypoxia, hypercapnic acidosis, and isocapnic acidosis. Circ. Res. 82:722–728. Bellocq, A., S. Suberville, C. Philippe, F. Bertrand, J. Perez, B. Fouqueray, G. Cherqui, and L. Baud. 1998. Low environmental pH is responsible for the induction of nitric-oxide synthase in macrophages: evidence for involvement of nuclear factor-␬B activation. J. Biol. Chem. 273:5086– 5092. Xu, L., and I. J. Fidler. 2000. Acidic pH-induced elevation in interleukin 8 expression by human ovarian carcinoma cells. Cancer Res. 60:4610–4616. Roebuck, K. A., and A. Finnegan. 1999. Regulation of intercellular adhesion molecule-1 (CD54) gene expression. J. Leukoc. Biol. 66:876–888. Roebuck, K. A. 1999. Regulation of interleukin-8 gene expression. J. Interferon Cytokine Res. 19:429–438. Coakley, R. J., C. Taggart, C. Greene, N. G. McElvaney, and S. J. O’Neill. 2002. Ambient PCO2 modulates intracellular pH, intracellular oxidant generation, and interleukin-8 secretion in human neutrophils. J. Leukoc. Biol. 71:603–610. Beck-Schimmer, B., R. C. Schimmer, R. L. Warner, H. Schmal, G. Nordblom, C. M. Flory, M. E. Lesch, H. P. Friedl, D. J. Schrier, and P. A. Ward. 1997. Expression of lung vascular and airway ICAM-1 after exposure to bacterial lipopolysaccharide. Am. J. Respir. Cell Mol. Biol. 17:344–352. Kumasaka, T., W. M. Quinlan, N. A. Doyle, T. P. Condon, J. Sligh, F. Takei, A. Beaudet, C. F. Bennett, and C. M. Doerschuk. 1996. Role of the intercellular adhesion molecule-1 (ICAM-1) in endotoxin-induced pneumonia evaluated using ICAM-1 antisense oligonucleotides, anti-ICAM-1 monoclonal antibodies, and ICAM-1 mutant mice. J. Clin. Invest. 97:2362– 2369. Czermak, B. J., M. Breckwoldt, Z. B. Ravage, M. Huber-Lang, H. Schmal, N. M. Bless, H. P. Friedl, and P. A. Ward. 1999. Mechanisms of enhanced lung injury during sepsis. Am. J. Pathol. 154:1057–1065. Watanabe, S., N. Mukaida, N. Ikeda, M. Akiyama, A. Harada, I. Nakanishi, H. Nariuchi, Y. Watanabe, and K. Matsushima. 1995. Prevention of endotoxin shock by an antibody against leukocyte integrin beta 2 through inhibiting production and action of TNF. Int. Immunol. 7:1037–1046. Xu, H., J. A. Gonzalo, Y. St Pierre, I. R. Williams, T. S. Kupper, R. S. Cotran, T. A. Springer, and J. C. Gutierrez-Ramos. 1994. Leukocytosis and resistance to septic shock in intercellular adhesion molecule 1-deficient mice. J. Exp. Med. 180:95–109. Sato, N., Y. Suzuki, K. Nishio, K. Suzuki, K. Naoki, K. Takeshita, H. Kudo,

132

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 29 2003

N. Miyao, H. Tsumura, H. Serizawa, M. Suematsu, and K. Yamaguchi. 2000. Roles of ICAM-1 for abnormal leukocyte recruitment in the microcirculation of bleomycin-induced fibrotic lung injury. Am. J. Respir. Crit. Care Med. 161:1681–1688. 31. Nishio, K., Y. Suzuki, T. Aoki, K. Suzuki, A. Miyata, N. Sato, K. Naoki, H. Kudo, H. Tsumura, H. Serizawa, S. Morooka, Y. Ishimura, M. Suematsu, and K. Yamaguchi. 1998. Differential contribution of various adhesion molecules to leukocyte kinetics in pulmonary microvessels of hyperoxiaexposed rat lungs. Am. J. Respir. Crit. Care Med. 157:599–609.

32. Yamaguchi, K., K. Nishio, T. Aoki, Y. Suzuki, N. Sato, K. Naoki, K. Takeshita, and H. Kudo. 1998. Do adhesion molecules importantly regulate leukocyte kinetics within intraacinar microvessels of the lung? Histol. Histopathol. 13:1089–1102. 33. Huber, A. R., S. L. Kunkel, R. F. Todd iii, and S. J. Weiss. 1991. Regulation of transendothelial neutrophil migration by endogenous interleukin-8. Science 254:99–102. 34. Kishimoto, T. K., and R. Rothlein. 1994. Integrins, ICAMs, and selectins: role and regulation of adhesion molecules in neutrophil recruitment to inflammatory sites. Adv. Pharmacol. 25:117–169.