Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Several previous studies have clearly shown morphologic injury to the alveolar epithelial barrier in the cases of patients with clinical acute lung injury, such as pneumonia (1). During inflammation, neutrophils are recruited from the circulation in response to a series of coordinated signals. These events have been widely studied and have largely been explained in recent years (3). However, little is known about the mechanisms that lead to the resolution of inflammatory events and repair from injury.

Strong evidence exists that supports the view that leukocytes are able to mediate tissue injury in inflammatory disorders (4). Recent studies have shown that an important factor in the successful resolution of inflammation is the recognition, uptake, and degradation by phagocytes of neutrophils that are undergoing deletion by apoptosis (5).

On the other hand, it is necessary to have repair of bronchial and alveolar epithelium that are damaged by recruited neutrophils and invasive microorganisms in pneumonia. This process requires the replication of the bronchial epithelium and proliferation and migration of alveolar type II cells to restore the integrity of the denuded bronchial and alveolar epithelia (10). Growth factors are considered to be involved in lung regeneration and several potent growth factors for bronchial and alveolar epithelial cells have been characterized. Hepatocyte growth factor (HGF) has been purified from the serum of partially hepatectomized rats (11), and is a unique, heparin-binding, proteolytically activated heterodimer produced from macrophages, fibroblasts, and endothelial cells in the lung (12, 13). HGF is considered to be a candidate growth factor for bronchial epithelial cells and alveolar type II cells (14, 15). After lung injury, HGF promotes the proliferation of lung epithelial cells, thereby playing an important role in restoring the integrity of the alveolar and bronchial epithelium (16, 17). Moreover, several studies have provided ample evidence for the importance of HGF in the repair process of tissues (18, 19).

In this study we report on an investigation of the kinetics of HGF production and the cellular sources of HGF in a murine model for bacterial pneumonia. We have proposed a novel pattern of HGF production, which might be first from the bronchial epitherium and then the alveolar macrophages (AMs).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals

Specific pathogen-free, 7-wk-old male Slc:ICR mice were obtained from Charles River Agricultural Cooperative Association for Laboratory Animals, Kanagawa, Japan. The mice were provided with sterile food and water ad libitum in an environmentally controlled room. The experimental protocol was approved by the Ethics Review Committee for Animal Experimentation of Nagasaki University School of Medicine.

Bacteria

Fisher immunotype 1 (It-1) Pseudomonas aeruginosa was originally obtained from Dr. M. Pollack (Uniformed Services University of the Health Sciences, Bethesda, MD). After overnight growth on brain-heart infusion agar (Difco, Detroit, MI) at 37°C, the bacteria were harvested in normal saline, resuspended in brain-heart infusion broth (Difco) containing 2% skim milk, and stored at 80°C before use.

Pneumonia Model

Pneumonia was induced via an intratracheal challenge with P. aeruginosa using the method described by Oishi and colleagues (20). The 50% lethal dose for this strain was determined to be 3 × 10⁷ colony-forming units (CFU)/mouse. In this study, 1 × 10⁷ CFU/mouse was used as the inoculation dose for the introduction of a sublethal pneumonia in mice.

Bronchoalveolar Lavage

Bronchoalveolar lavage (BAL) was performed at 0, 3, 6, 12, 24, 36, and 72 h, and at 5, 7, and 14 d after bacterial challenge in each group of five mice. Under deep anesthesia, the trachea was exposed and intubated. A 2.5-ml syringe was connected to the tracheal cannula, and the lungs were washed four times with 2 ml of Ca²⁺- and Mg²⁺-free phosphate-buffered saline (PBS) at 4°C. A 1.5-ml volume of BAL fluid (BALF) was recovered in each mouse. The cell counts in the recovered BALF were determined by using a hemocytometer. BALF was centrifuged at 150 × g for 10 min at 4°C, and the cell pellet was resuspended in 1 ml PBS. Cell morphology was determined on cell monolayers prepared by Cytospin 2 (Shandon Southern Products, Astmoor, UK) and stained with Diff-Quik. Further, to observe the macrophage engulfment of apoptotic neutrophils, cytospins for the examination of myeloperoxidase (MPO) activity by histochemistry were fixed for 15 min at 4°C in 2% glutaraldehyde. After washing in PBS, the slides were immersed for 30 min in 0.5 mg/ml diaminobenzidine in 0.05 M Tris-HCl (pH 7.6). Then 0.01% hydrogen peroxide was added, and after 45 min the slides were washed in 0.1 M phosphate buffer and exposed to cupric nitrite 0.5% for 1 min (21). The AMs themselves were routinely negative for MPO staining. Using a ×100 oil objective lens, we calculated the number of macrophages phagocytosing apoptotic neutrophils (positive for MPO in their cytoplasm) in 200 macrophages and represented the percentage of MPO-positive AMs. Recovered BALF supernatant was filter-sterilized and stored at 80°C.

Quantitative Culture of Lung Tissue

The bacterial number in the lung was measured by a quantitative culture of lung. The lungs were removed aseptically and homogenized in 9 ml of sterile saline per gram of lung tissue before culture.

Sandwich Enzyme-Linked Immunosorbent Assay for HGF

The concentrations of murine HGF in BALF supernatants were determined by a sandwich enzyme-linked immunosorbent assay (ELISA) by using a commercial rat HGF ELISA kit according to the instructions provided by the manufacturer (Institute of Immunology Co., Tokyo, Japan). Amino-acid sequences of mouse HGF and rat HGF are 98.6% identical. Mouse HGF as well as rat HGF were detected because the antirat monoclonal antibody (mAb) and polyclonal antibody in this kit closely crossreact with mouse HGF. A 96-well, flat-bottomed microtiter plate was coated with mouse antirat HGF mAb per well. Diluted cell-free BALF samples or standards (100 µl/well) were added in duplicate, and the plates were incubated for 20 h at 20°C. The plates were washed five times with the wash buffer, 100 µl of rabbit antirat HGF immunoglobulin (Ig) G per well was added, and the plates were then incubated for 2 h at 20°C. The plates were then washed five times, and 100 µl of peroxidase-labeled goat antirabbit IgG per well was added, followed by further incubation for 2 h at 20°C. The plates were washed five times, 100 µl of substrate per well was added, and the plates were incubated for 30 min in a dark room. Finally, 50 µl of stop solution was added to each well and the plates were read at 490 nm in an ELISA reader.

Lung Harvesting for Histologic Examination

At designated time points, mice were killed during deep anesthesia and both lungs were harvested for histologic examination. Lungs were removed and inflated with 0.5 ml of 4% paraformaldehyde in PBS. After 48 h, the lungs were embedded in paraffin and stored until examined.

Isolation and Semiquantitative Reverse Transcription/ Polymerase Chain Reaction Amplification of Whole Lung Messenger RNA

Whole lungs were harvested at preselected times after the infection of bacteria, immediately homogenized with Isogen (Wako Pure Chemical Co., Osaka, Japan) and total cellular RNA was extracted. Purified RNA was quantitated by measuring the absorbance at 260 nm. Complementary DNA (cDNA) was synthesized from 2 µg of total RNA by priming with 2.5 mMol of oligo (dT) primers, 1 mM of each deoxynucleoside triphosphate (dNTP), and reverse transcriptase (RT). cDNA equivalent to 80 ng of starting RNA was used for polymerase chain reaction (PCR) using primers for mouse HGF and -actin. The primers used were as follows: mouse HGF sense and antisense were 5'-GGACAAGATTGTTATCGTGG-3', and 5'-GTTGATCAATCCAGTGTAGC-3', respectively, yielding an amplified product of 716 base pair (bp). -Actin sense primer was 5'-TCCTGTGGCATCCATGAAACT-3', and antisense primer was 5'-ACTCCTGCTTGCTGATCCAC-3', yielding an amplified product of 273 bp. PCRs were performed with 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 2 mM MgCl₂, 1 mM dNTPs, and 2.5 U of Taq DNA polymerase (Perkin-Elmer, Branchburg, NJ) in a final volume of 100 µl. Primers were added to a final concentration of 0.1 mMol. The reactions were carried out in a DNA thermal cycler (Perkin-Elmer), first incubated for 5 min at 94°C, then cycled 35 times under the following conditions; denaturing at 93°C for 45 s, annealing at 52°C for 45 s, and extension at 72°C for 80 s. After amplification, the sample was separated on a 1% agarose gel containing 0.3 mg/ ml ethidium bromide, and bands were visualized and photographed with ultraviolet transillumination. The densities of the bands were analyzed using densitometry (Epi-Light UV FA1100; Aisin Cosmos Co. Ltd., Tokyo, Japan) and Luminous Imager software (Aisin Cosmos) as described previously (22). The densitometric intensity was normalized by comparing the ratio of HGF bands with that of -actin.

Immunohistochemical Localization of Antigenic HGF

Paraffin-embedded specimens of whole lungs were cut in 3-µm sections and placed on silane-coated slides, dewaxed with xylene, and dehydrated through graded concentrations of ethanol. The tissue was then treated with 0.03% trypsin for 1 h. This procedure provides more available antigenic sites for the antibody. In the next step, the tissue was placed in 3% hydrogen peroxide for 5 min to reduce endogenous peroxidase activity. Tissue-nonspecific binding sites were blocked with normal swine serum in PBS for 30 min. Excess serum was removed by blotting, and sections were covered overnight with 1:5 dilution of rabbit polyclonal anti-HGF antibody or control rabbit IgG at 4°C. After washing with PBS, sections were then covered with the biotinylated second antibody, swine antirabbit IgG, for 40 min; rinsed in PBS; covered with peroxidase (Dako Co., Carpinteria, CA) reagent for 40 min at room temperature; and rinsed in PBS. Antigenic sites on sections were visualized by reacting these sections with a mixture of 0.05% 3,3,-diaminobenzidine tetrahydrochloride in 0.05 M Tris-HCl buffer and 0.01% hydrogen peroxide for 7 min. Sections were then counterstained with methyl green for 10 min, dehydrated in ethanol, cleared in xylene, and mounted.

Isolation of Murine AMs

AMs were obtained by BAL as previously reported with minor modifications (23). After the euthanizing of the ICR mice by an intraperitoneal pentobarbital injection, AMs were obtained by repeated lung lavage with PBS. The lavaged cells were centrifuged at 150 × g for 10 min at 4°C and then resuspended in Dulbecco's modified Eagle's medium (DMEM) (Iwaki Co. Ltd., Chiba, Japan). Cell counts and viability were determined by hemocytometer counts of trypan blue-stained aliquots. Cytospin preparations were stained with a Diff-Quik stain; and examination showed that > 98% of cells were normal lung macrophages.

Isolation of Peritoneal Neutrophils

Peritoneal exudate neutrophils were obtained after an intraperitoneal injection of casein according to the modified method of Van Epps and Garcia (24). Normal mice (7 wk old) were injected with 3.5 ml of a 2% saturated solution of casein (Sigma Chemical Co., St. Louis, MO). After 15 h, mice were anesthetized with diethylether (Wako) and peritoneal cells were recovered by lavage with 10 ml of cold (4°C) PBS. To minimize cell clumping, the lavaged cells were washed and centrifuged at 4°C and visible cell clumps were removed by adherence to glass pipettes. The washed peritoneal cells were further purified by centrifugation through mono-poly resolving medium (Dainippon Pharma. Co. Ltd., Osaka, Japan). Purified peritoneal exudate cells were 95% neutrophils, as evidenced by Diff-Quik staining.

Aging of Mice Neutrophils

Aged neutrophils were obtained using a technique previously reported by Heifets and associates (25). Neutrophils were suspended in 40 ml of a 50:50 (vol/vol) mixture of Hanks'-bovine serum albumin and 0.34 M sucrose and incubated overnight at 4°C. This medium was chosen because it allows neutrophils to age without aggregation and permits the estimation of the effect of serum on neutrophils. After aging, the neutrophils were centrifuged once (150 × g at room temperature), resuspended in DMEM, and recounted, and their viability was determined by trypan blue dye exclusion. Neutrophils aged by these techniques were routinely 95 to 97% viable.

Assessment of Apoptosis of Neutrophils by Flow Cytometry

A separate and independent assessment of apoptosis was performed using a modification of Nicoletti's protocol (26) which involves the measurement of DNA fragmentation by flow cytometry. A total of 1 × 10⁶ neutrophil cells (freshly isolated neutrophils and aged neutrophils) were suspended in 1 ml of a hypotonic fluorochrome solution (50 µg/ml propidium iodide in 0.1% sodium citrate plus 0.1% Triton X-100) and incubated overnight at 4°C in dark before analysis. DNA fragmentation analysis was carried out using an EPICS Profile II instrument (Beckman Coulter Electronics, Luton, UK). Apoptotic nuclei were measured by the appearance of a sub-G₀ peak and represented the percentage of apoptotic neutrophils in freshly isolated as well as in aged neutrophils.

Phagocytosis of Apoptotic Neutrophils and HGF Production by AMs

AMs obtained from normal mice were cultured at 1.5 × 10⁵ cells/ well in a Lab-Tek chamber slide system (chamber mounted on glass slide with cover; Nalge Nunc International, Naperville, IL) and in 24-well culture plates (Nalge Nunc) for 2 d. Subsequently, 3 × 10⁵ cells of freshly isolated or aged neutrophils, which were suspended in 300 µl of DMEM without fetal calf serum (FCS), were placed on the cultured AMs in triplicate and incubated for 50 min at 37°C in 5% CO₂/95% air. In the Lab-Tek chamber slide, noningested neutrophils were removed by washing twice in PBS and then, to visualize internalized neutrophils, an MPO stain was used. The percentage of phagocytosing macrophages was calculated as described previously (5, 27, 28). In the case of the 24-well culture plates, noningested neutrophils were removed by washing twice in PBS, adding 300 µl of fresh DMEM without FCS to each well, and incubating for 24 h. After this, the culture supernatants were collected and filtered, and then HGF concentrations in each sample were measured by ELISA.

Statistical Analysis

Data are expressed as means ± standard error of the mean (SEM). Differences between groups were examined for statistical significance using the unpaired (two-tail) t test. A P value < 0.05 denotes the presence of a significant statistical difference.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Elimination of Invading Bacteria by Neutrophils and the Resolution of Inflammation by Phagocytosis of Neutrophils by AMs

When mice were challenged with P. aeruginosa It-1 strain at a dose of 1 × 10⁷ CFU/mouse, the bacterial densities reached levels of approximately 1.13 ± 0.63 × 10⁸ CFU/g of lung tissue 3 h after challenge. However, the bacterial number in the lung tissue progressively decreased during the 14-d post-challenge period (Figure 1A).

View larger version (14K): [in this window] [in a new window]   Figure 1.   Kinetics of bacterial number and airway neutrophil influx in response to intratracheal P. aeruginosa challenge. Quantitative culture of homogenized lung and BAL was performed at 0, 3, 6, 12, 24, 36, 48, and 72 h, and at 5, 7, and 14 d after intratracheal challenge of P. aeruginosa. (A) The bacterial number in the lung was determined quantitatively by culture of lung. (B) Absolute neutrophil count was calculated from a differential count determined in cytospin preparations stained with Diff-Quik stain. Each data point represents the mean ± SEM value of five mice.

The phase of neutrophil accumulation and neutrophil clearance was determined by a kinetic study of airway leukocyte influx. The baseline BALF cell count was 3.8 ± 0.8 × 10⁴ cells/ml with a differential count of 99.5% AMs. The neutrophil influx into the airway was apparent at 3 h (6.7 ± 4.6 × 10⁴/ml) after bacterial challenge and then increased, reaching a peak level at 36 h (6.8 ± 2.0 × 10⁵/ml). It then gradually decreased over a 14-d period (Figure 1B). These results indicate that the initiation and amplification of inflammation occurred within a few hours and reached a peak level at around 36 h after bacterial challenge. Further, on the basis of the kinetic study of the neutrophils, the data suggest that the resolution phase of pulmonary inflammation and the repair of lung injury commences after 36 h and continues until Day 14. Cox and coworkers (7) recently investigated the phase of neutrophil clearance in acute pulmonary inflammation as induced by an intratracheal inoculation of lipopolysaccharide, and demonstrated that AMs phagocytosed apoptotic neutrophils and became positive for peroxidase at specific time points. Accordingly, to determine the phase of macrophage phagocytosis of apoptotic neutrophils in the experiments, in the present study we stained a cytospin preparation of cells obtained by BAL with MPO stain at preselected time intervals. Figure 2B shows the kinetics of MPO-positive AMs in BALF that were recognized to contain phagocytosed apoptotic neutrophils. MPO-positive macrophages were detected from 24 h (17.2 ± 10.3% of total AMs) but their percentage reached maximum value at 36 h (43.4 ± 10.5% of total AMs) and remained at high levels up to 72 h after bacteria inoculation. The baseline AM number remained unchanged until 24 h after challenge but then increased at 36 h, reaching a peak level at Day 5 (Figure 2A). After 72 h, the percentage of neutrophils in BALF diminished gradually until the end of the observation period (14 d). These findings indicate that AM actively phagocytoses neutrophils between 24 and 72 h of infection and contributes to the resolution of neutrophilic inflammation.

View larger version (17K): [in this window] [in a new window]   Figure 2.   Serial changes in AM number and percentage of AMs that showed phagocytosed apoptotic neutrophils (stained for MPO) in BALF at the indicated time intervals. (A) AM number in BALF was determined using a hemocytometer. (B) Percentage of MPO-positive macrophages in BALF was determined using MPO stain. Percentage of phagocytosing macrophages was assessed by counting 200 macrophages. Data are means ± SEM of five independent experiments.

Biphasic Production of HGF in BALF

HGF concentrations were detected in BALF at 0 h, increased slightly up to 3 h, reached a maximum level at 6 h (3.28 ± 0.44 ng/ml), then decreased between 24 and 36 h (Figure 3). Interestingly, HGF concentration increased again at 48 h, reaching a peak level at 72 h (4.42 ± 0.82 ng/ml). Thus, the levels of HGF showed biphasic kinetics in BALF in our mouse model of pneumonia.

View larger version (9K): [in this window] [in a new window]   Figure 3.   Serial changes in HGF production in response to intratracheal P. aeruginosa challenge. HGF concentration was determined in BALF at each time point by ELISA. Each data point represents the mean ± SEM value of five mice. Note the biphasic response of HGF activity in the lung of mice with pneumonia. The conditions used for administration of P. aeruginosa are similar to those described in Figure 1.

Biphasic Expression of HGF Messenger RNA in Whole Lung

To determine the production of HGF at the messenger RNA (mRNA) level, semiquantitative RT-PCR was performed using mRNA extracted from whole lung. HGF mRNA was detectable in the lungs of uninfected mice but was detected in increased levels in lung homogenates of mice with pneumonia within 3 h of bacterial challenge. Further analysis showed that HGF mRNA expression decreased later but then increased markedly again during the period of 24 to 72 h after bacterial challenge (Figure 4A). The HGF mRNA levels at 3 h after bacterial challenge were approximately 1.9-fold higher, as evidenced by optical density measurement, than uninfected controls (Figure 4B). The mRNA levels decreased later but then increased again, reaching peak level at 72 h after bacterial challenge (2.5-fold higher than the uninfected controls). These results suggest that the biphasic kinetics of HGF are reflected by corresponding mRNA levels.

View larger version (36K): [in this window] [in a new window]   Figure 4.   Time-dependent production of HGF mRNA in lung homogenates after intratracheal instillation of P. aeruginosa. (A) Determination of HGF and -actin mRNA induction in the lung of mice with pneumonia. The top band shows the PCR product of HGF, and the bottom band shows the PCR product of -actin. (B) The band intensities were determined with Luminous Imager software, and the level of each HGF mRNA expression was normalized with mRNA level of -actin. The conditions of P. aeruginosa administration were similar to those described in Figure 1. Note the high expression of HGF mRNA in lung homogenates within 3 h after bacterial challenge, which then decreased but increased later, at 24 to 72 h after challenge.

Cellular Source of HGF in Murine Lung Tissue

To determine the cellular source of HGF in the murine lung, an immunohistochemical study was carried out on lung tissue using anti-HGF polyclonal antibody. HGF-positive cells were present only within bronchial epithelial cells at 6 h after bacterial challenge (Figures 5A and 5B). In contrast, at 72 h after instillation of bacteria the stained cells were present not only within bronchial epithelial cells but within AMs as well (Figures 5C and 5D). These findings suggest that HGF is produced by bronchial epithelial cells during the neutrophil accumulation phase, followed by production by the AM during the neutrophil clearance phase. Staining for HGF was specific, inasmuch as no staining was observed in sections of lungs at 72 h that had been incubated with purified IgG from control serum (Figure 5E) or in sections of uninfected controls which were stained using the anti-HGF antibody (Figure 5F).

View larger version (104K): [in this window] [in a new window]   Figure 5.   Immunohistochemical staining of murine lungs with anti-HGF antibody at 12 h (A and B) and 72 h (C and D) after intratracheal challenge of P. aeruginosa. (A and B) At 12 h, HGF-positive cells were present only within bronchial epithelial cells. (C and D) At 72 h, HGF-positive cells were present not only within bronchial epithelial cells but also within AMs as well. (E) Sections of lungs at 72 h after bacterial challenge that had been incubated with control IgG. (F) Sections of uninfected control lungs which stained using anti-HGF antibody. Original magnifications: A, C, E, and F, ×200; B and D, ×400. The conditions of P. aeruginosa administration were similar to those described in Figure 1.

HGF Production by Mouse AMs Was Induced by Phagocytosis of Apoptotic Neutrophils

The phase of the second HGF peak was closely associated with an increased percentage of MPO-positive AMs in BALF (Figures 2B and 3). In addition, AMs represent one of the cellular sources of HGF during this period (Figure 5C). On the basis of this observation, we hypothesize that the second HGF peak was the result of production by AMs that had phagocytosed apoptotic neutrophils. To test this hypothesis, we examined the issue of whether mouse AMs that phagocytize aged neutrophils, but not fresh neutrophils, produce HGF in vitro. For this purpose, we first assessed the apoptosis of neutrophils by flow cytometry. The percentage of apoptotic cells among the aged neutrophils (70.2 ± 2.7%) was significantly higher than that among freshly isolated neutrophils (45.9 ± 10.85%; P < 0.05, Figure 6A). In the next step, we determined the percentage of AMs that phagocytosed apoptotic neutrophils via a determination of examining histochemical peroxidase activity. The percentage of MPO-positive AMs incubated with aged neutrophils (47.5 ± 6.9%) was significantly higher than those incubated with freshly isolated neutrophils (15.8 ± 4.5%; P < 0.05, Figure 6B). Lastly, AMs that were incubated with aged neutrophils produced significantly higher concentrations of HGF (0.95 ± 0.04 ng/ml) than did those that had been incubated with freshly isolated neutrophils (0.37 ± 0.01 ng/ml; P < 0.05, Figure 6C). Thus, the population of aged neutrophils contained more apoptotic cells, which upregulated HGF production by AMs.

View larger version (22K): [in this window] [in a new window]   Figure 6.   HGF production from murine macrophages that were incubated with freshly isolated neutrophils or aged neutrophils in vitro. (A) Comparison of percentage of apoptotic neutrophils between aged neutrophils and freshly isolated neutrophils by using flow cytometry. Data are means ± SEM of five independent experiments. (B) Comparison of percentage of MPO-positive macrophages (stained for MPO) between macrophages incubated with aged neutrophils and freshly isolated neutrophils. MPO-positive AMs that were calculated as the percentage of phagocytosing macrophages was determined by counting 200 macrophages. Data are means ± SEM of three independent experiments. (C) Comparison of HGF production by macrophages between macrophages incubated with aged or freshly isolated neutrophils. As a control, macrophages were incubated for 24 h without any stimulus. Data are expressed as the mean production of HGF ± SEM for three experiments. * and **Significantly different (P < 0.05).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Because the lung is exposed to various exogenous pathogens, bacterial pneumonia is a common occurrence in clinical practice. In the case of bacterial pneumonia, alveolar type I cells and bronchial epithelial cells are injured. Compensatory proliferation and differentiation of alveolar type II cells and the remaining bronchial epithelial cells is an important factor in lung regeneration (14, 15). A variety of growth factors appear to be involved in post-inflammatory lung regeneration, and several potent growth factors for bronchial and alveolar epithelial cells have been isolated and characterized. For example, potent growth factors for alveolar type II epithelial cells have been characterized, including epidermal growth factor (EGF), transforming growth factor (TGF)-, keratinocyte growth factor (KGF), acidic (aFGF) and basic fibroblast growth factor, and HGF (29, 30). Panos and colleagues (31) have shown that HGF and KGF induce the synthesis of DNA in alveolar type II cells, and that HGF stimulates DNA synthesis more strongly than does KGF. Shiratori and associates (32) demonstrated that HGF is a more potent stimulator of DNA synthesis than is aFGF, TGF-, or EGF. HGF strongly induces the regeneration of lung tissue compared with other growth factors. Further, Yaekashiwa and coworkers (19) reported that exogenously administered HGF was able to inhibit lung fibrosis induced by bleomycin in mice. This conclusion was based on their findings which showed that the administration of HGF delayed bleomycin-induced lung injury, stimulated the synthesis of DNA in type II epithelial cells, and promoted the turnover of damaged epithelial cells. This turnover was shown to diminish inflammatory changes and to suppress fibrotic changes in the lung. Therefore, the mechanism by which HGF is regulated during lung injury is important in terms of our understanding of lung regeneration.

In the present study, a biphasic pattern of HGF mRNA expression and protein synthesis in the lung after intratracheal bacterial challenge has been demonstrated. These results suggest that two different mechanisms are involved in the stimulation of HGF production in this model. Yanagita and colleagues (16) reported that HGF mRNA and HGF activity in lung increased markedly as early as 3 to 6 h after the onset of acute lung injury induced by an intratracheal injection of hydrochloride in the rat. They also demonstrated a marked increase in DNA synthesis in bronchial epithelial cells and alveolar epithelial cells during a 24 to 48-h period after the onset of lung injury. They indicated that the rapid and sequential induction of HGF mRNA and HGF activity before pulmonary epithelial cell proliferation indicates that HGF acts as a pulmotorophic factor on lung regeneration after acute lung injury. The first peak of HGF production in our model may correspond to previous findings reported by Yanagita and associates (16). During the early stages of pneumonia, our immunohistochemical studies indicated the localization of HGF protein in the bronchial epithelium. Tsao and coworkers (33) have previously reported that HGF is an autocrine factor in normal human bronchial epithelial cells in vitro. Bronchial epithelial cells act as biologic barrier in the lung. It is, therefore, important for the remaining bronchial epithelium to produce HGF, which would then act in an autocrine manner to repair the damaged epithelium rapidly.

Matsumoto (34) demonstrated that tumor necrosis factor (TNF)- and interleukin (IL)-1 upregulated HGF gene expression in both skin and human fetal lung fibroblasts. In our model, the production of TNF- (an important cytokine that initiates the secretion of inflammatory mediators) increased, reaching a maximal level at 6 h after intratracheal bacterial challenge (data not shown). One explanation for the early presence of HGF observed in our study may be its upregulation by proinflammatory cytokines such as TNF- and IL-1, which are derived from bronchial epithelial cells.

More importantly, we found an association of the second peak of HGF production with the increased phagocytosis of neutrophils by AMs in the airway. In addition, one cellular source of HGF in this phase was shown to be AMs. Further, we demonstrated that AMs that phagocytosed neutrophils induced HGF production in vitro. These lines of evidence indicate that AMs, which are involved in the phagocytosis of neutrophils, generate HGF during the phase of neutrophil clearance in the airways of infected mice. HGF has been reported to stimulate DNA synthesis in alveolar epithelial cells, in a dose-dependent manner, reaching a maximum level of HGF at concentrations of 5 to 10 ng/ml (31). In this study, the peak level of HGF was 4.4 ± 1.8 ng/ml in BALF at 72 h after challenge. This level of HGF, thus, is sufficient to stimulate DNA synthesis in alveolar epithelial cells. In our study, the second peak of HGF production reached a higher concentration that persisted over a longer period (about 24 h) compared with the first peak of HGF. Therefore, HGF generated by AMs that phagocytized apoptotic neutrophils could play an important role in lung regeneration through release of HGF in a paracrine fashion. Moreover, the number of AMs gradually increased from 36 h after challenge, reaching a maximum level at 5 d and then declining up to 14 d after challenge. An increase in the number of AMs may have an advantageous effect on the production of large amounts of HGF as well as for the efficient clearance of apoptotic neutrophils in the airways of infected mice.

Fadok and colleagues (35) demonstrated that AMs phagocytosed apoptotic cells, produced TGF-, and reduced the levels of proinflammatory cytokines. In their study, the removal of apoptotic cells by VnR and CD36, rather than by opsonic receptors such as FcR or CR3, led to increased levels of TGF-. In our study, a phagocytic assay was determined using serum-free medium. This finding may indicate that the mechanism for the recognition of apoptotic cells is similar to that described in the study by Fadok and associates (35).

In summary, the in vivo and in vitro data herein strongly suggest that bronchial epithelial cells produce HGF during the neutrophil accumulation phase, and that macrophage engulfment of apoptotic neutrophils produces HGF during the neutrophil clearance phase. HGF produced by bronchial epithelial cells may repair the damaged epithelium in an autocrine fashion, and HGF produced by AMs acts in a paracrine fashion in a similar manner. These two different types of cells are essential for the regulation of the healing and repair of damaged bronchial and alveolar epithelium during bacterial pneumonia. Therefore, we conclude that AMs in particular are capable of acting as key cells in the repair of damaged pulmonary epithelium through the phagocytosis of apoptotic neutrophils and intimate involvement with inflammatory events, which include the production of HGF.