Hepatocyte Growth Factor May Act as a Pulmotrophic Factor on Lung Regeneration after Acute Lung Injury*

Hepatocyte growth factor (HGF) has been shown to have hepatotrophic and renotropic functions for regen- eration of the liver and kidney through its mitogenic, motogenic, and morphogenic properties. To examine the involvement of HGF in lung regeneration after acute injury, we analyzed changes of HGF mRNA, HGF activ- ity, and HGF receptors in the rat lung after lung insult and measured HGF concentration in sera of patients with various lung diseases. Following the onset of acute lung injury induced by intratracheal hydrochloride injection, a compensatory DNA synthesis occurred in the bronchial epithelium with a peak at 24 h and in the alveolar epithelium with a peak at 48 h. Expression of HGF mRNA in the rat lung remarkably increased only 3 h after the treatment and HGF activity in the lung also increased to about %fold at 6 h later. HGF receptors in the lung but not in the other noniqjured organs were down-regulated 12 h later. These marked increases in HGF mRNA and HGF activity and the concomitant down-regulation of HGF receptor occurred before the marked compensatory DNA synthesis in bronchial and alveolar epithelial cells. HGF concentration in sera of patients with various lung diseases, as

Hepatocyte growth factor (HGF)l was first detected in the plasma of partially hepatectomized rats as a potent mitogen for adult rat hepatocytes in primary culture (Ref. 1, and reviewed in Refs. 2 and 3) and was purified to homogeneity from rat platelets (4, 5) and subsequently from human plasma (61, rabbit serum (7), livers of CC14-administered rats (8), and conditioned medium of human lung fibroblasts (9). Mature heterodimer HGF derives from a single chain precursor of 728 amino acids by proteolytic processing and has four kringle do-~~ *This work was supported by a Research Grant for Science and Cancer from the Ministry of Education, Science and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The abbreviations used are: HGF, hepatocyte growth factor; BrdU, 5-bromo-2'-deoxyuridine; EGF, epidermal growth factor; TGF, translike growth factor. forming growth factor; KGF, keratinocyte growth factor; IGF, insulin-mains in the a-chain and the serine protease-like domain in the P-chain (10-13). HGF targets a wide variety of epithelial cells and has multiple biological activities. HGF stimulates proliferation of various cells, including renal tubular epithelial cells (141, keratinocytes, melanocytes, and many other epithelial cells (9, 151, whereas it inhibits growth of hepatoma and melanoma cells (16-18). HGF strongly enhances cell motility of various epithelial cells as a motogen (15, 19-22) and induces branching tubule formation of Madin-Darby canine kidney cells in collagen gel as a morphogen (23). These pleiotrophic actions of HGF are mediated through high-affinity HGF receptors with a Kd value of 20-30 PM (24, 25). The high-affinity receptor is known to be a c-met proto-oncogene product which has the intracellular tyrosine kinase domain (26-28).
HGF functions as a hepatotrophic or renotropic factor for hepatic or renal regeneration, respectively. HGF mRNA and HGF activities markedly increase in the liver or plasma of rats after various liver insults (2, 3, 8, 291, and intravenously injected recombinant HGF markedly enhances liver regeneration in mice (30). Moreover, HGF mRNA and HGF activity also increases in the kidney after renal injury (31, 32).
On the other hand, since HGF receptors and c-met mRNA are present in the lung, the possibility that HGF participates in regeneration following lung injury would have to be given attention. To produce acute lung injury, we injected intratracheally hydrochloride (HC1 injection) (331, as an experimental model of aspiration pneumonia which induces adult respiratory distress syndrome.
We described here evidence that both HGF mRNA and HGF activity markedly increased followed by marked down-regulation of the HGF receptor and subsequently by DNA synthesis of lung epithelial cells in the injured lung. Our findings lend support to the thesis that HGF may be a "pulmotrophic factor" for regeneration of an injured lung.
EXPERIMENTAL PROCEDURES Animal neatments-Adult male pathogen-free Wistar rats (200-250 g of body weight) were used in this study. To induce acute lung injury, we anesthetized the animals with ether and exposed trachea. An 18gauge elaster needle was placed in the trachea and 0.1 M HCl 2 mVkg body weight was injected into the lung (33).
Assay for Histological Change and Labeling Index-Fifty m&g body weight of 5-bromo-2'-deoxyuridine (BrdU, Sigma) was intraperitoneally injected into rats, and 1 h later these rats were anesthetized with ether, the lungs were removed, fixed with 70% ethanol for 12 h, and embedded in paraffin. For histological observation, 4-pm-thick sections were deparaffinized and stained with hematoxylin and eosin. To measure DNA synthesis, the cells undergoing DNA synthesis were immunohistochemically identified using anti-BrdU antibody, as described elsewhere (30). Enzyme reaction was initiated in buffer composed of 0.2 M Tris-HC1 (pH 8.2), 250 pg/ml L-levamisole (Sigma), 0.01% new fuchsin (Merck), 0.01% sodium nitrite, 250 pg/ml naphthol AS-BI triphosphate (Sigma), and 0.25% N,N-dimethylformamide (34). The sections were washed with phosphate-buffered saline, fixed with 10% formalin, and counter-stained with hematoxylin. The labeling index was counted in cells among more than 1000 nuclei of bronchial epithelia and alveolar septa, but granulocytes and lymphocytes were excluded from counting.
Preparation ofAnti-rat HGFAntibodies-Recombinant rat HGF was purified from conditioned medium of Chinese hamster ovary (CHO) cells transfected with expression vector containing rat HGF cDNA (13).
To prepare polyclonal antibodies against rat HGF, purified rat HGF was injected into rabbit with Freund's adjuvant. IgGs were, respectively, purified from pre-immune serum and antiserum using protein-A Sepharose.
Northern Blot Analysis-Total RNA was isolated from the lung, using the acid-guanidium thiocyanate-phenol-chloroform method (35). Twenty pg of total RNA were applied to electrophoresis on 1.0% formaldehyde-agarose gels and transferred t o a Hybond-N membrane filter (Amersham). The EcoRI fragment (1.4 kilobase pairs) of rat HGF cDNA (RBC-1 clone, Ref. 13) was labeled, using the Megaprime DNAlabeling system (Amersham), according to the manufacturer's instruction. RBC-1 encodes the 5"portion including the fourth kringle domain of a-chain, the entire p-chain of HGF, and a part of 3'-noncoding region. Hybridization was performed at 42 "C for 20 h in solution composed of 50% (vh) formamide, 5 x SSPE (1 x SSPE consists of 0.15 M NaC1, 10 m M sodium phosphate buffer (pH 7.4) and 1 m M EDTA), 2 x Denhardt's solution (1 x Denhardt's solution consists of 0.02% Ficoll (Type 400, Pharmacia), 0.02% polyvinylpyrrolidone, and 0.02% bovine serum albumin (Fraction V, Sigma)), 0.5% SDS, and 50 pg/ml salmon sperm DNA. The filter was washed with 0.2 x SSPE, 0.1% SDS for 15 min at 65 "C, then was dried and autoradiographed on Fuji x-ray film. The gel was stained with ethidium bromide to estimate the amount of loaded RNAs.
Partial Purification of HGF fiom the Lung-HGF was partially purified from the lung, as described elsewhere (1,4,8). Briefly, the lung was homogenized with Polytron in 4 volumes of buffer composed of 10 m M Tris-HC1 (pH 7.5), 0.3 M NaC1, and 0.01% Tween 80 containing protease inhibitors (1 m M phenylmethylsulfonyl fluoride, 1 m M monoiodoacetate, and 1 n m EDTA). After centrifugation at 105,000 x g for 1 h, the supernatant was dialyzed against 100 volumes of the same buffer, using cellulose tube with molecular cut off >M, 30,000, and applied to a heparin-Sepharose column equilibrated with the same buffer. The column was washed with 2 volumes of the same buffer and then eluted with buffer containing 2 M NaCl. The eluate was dialyzed against buffer composed of 2 m M Hepes-NaOH (pH 7.2) and 20 m M NaCl, lyophilized, and then dissolved with H20. Samples used for HGF activity assay were passed through a 0.22-pm pore size filter (Millipore).
To determine whether stimulatory effect of sample prepared from rat lung on DNA synthesis of hepatocytes was attributable to HGF, test sample was incubated with pre-immune control I& or anti-rat HGF IgG at 4 "C for 1 h and assayed for HGF activity.
Assay for HGFActiuity-HGF activity was determined by measuring the stimulatory effect on DNA synthesis of rat hepatocytes in primary culture (4). Adult rat hepatocytes were isolated by the in situ collagenase perfusion method. The isolated cells were plated at a density of 6.25 x lo4 cells, 0.5 ml, 2 cm2 on 24-well plastic dish (Corning) coated with type I collagen and cultured in Williams' E medium containing 5% calf serum, M insulin, and M dexamethasone for 3 h. The medium was changed to serum-free Williams' E medium containing M insulin, lo-" M dexamethasone and 0.1 pgml aprotinin. After 20 h, test samples were added and the cells were cultured for 20 h. The cells were pulse-labeled with 1 pCi of 1251-deoxyuridine for 6 h, and 1251-deoxyuridine in the nuclei was measured in a y-counter. One unit of HGF activity corresponds to the stimulatory activity which gives a halfmaximal value of stimulatory effect on DNA synthesis of hepatocytes when 10 ng/ml EGF is added. Assay for '251-HGF Binding to Plasma Membranes-Plasma membranes were prepared from lung, liver, kidney, and spleen of rats by density-gradient centrifugation on Percoll, as described elsewhere (36). Human recombinant HGF was radioiodinated by a chloramine-T method, as described elsewhere (37). The lZ5I-HGF had a specific radioactivity of 70-160 mCi/mg of protein. Fifty pg of plasma membranes were incubated with 48 PM lZ5I-HGF for 1 h at 10 "C, with or without 6.4 n M unlabeled HGF in 100 pl of binding buffer (Hanks' solution containing 10 m M Hepes, 2 mg/ml bovine serum albumin). Membranes were centrifuged for 10 min at 12,000 x g at 4 "C, resuspended with 10 ml of binding buffer, and transferred to fresh tubes. lZ5I-HGF specifically bound to membranes was counted in a y-counter. All binding experiments were done in triplicate.
Assay for HGF Concentration in Human Blood-HGF concentration in human sera was measured by sandwich radioimmunoassay using monoclonal anti-human HGF antibodies. Briefly, beads coated with anti-human HGF monoclonal antibodies were incubated with one-halfdiluted human serum for 15 h at 25 "C. After washing with phosphatebuffered saline, the beads were incubated with 1251-labeled anti-human HGF monoclonal antibodies for 4 h, and lZ6I specifically bound to the beads was counted using a y-counter. HGF concentrations in human sera can be precisely determined as low as 0.01 ng/ml.2

Histological Change after Induced Acute Lung Injury"T0
induce acute lung injury, 0.1 M HCl 2 ml/kg body weight was intratracheally administrated into the lung of each rat. In the intact lung, the intima of the bronchus consisted of simple columnar epithelia and that of the thin alveolar septa consisted of simple squamous epithelia (Fig. IA). Three hours after HC1 injection, hemorrhage in the submucosal region of the bronchus and in the alveolar space was obvious, and infiltration of polymorphonuclear leukocytes was present in several regions (Fig.  1B). Twelve hours later, there was histological evidence of the onset of acute lung injury: some pulmonary edema, de-epithelialization of the bronchus and alveoli, a neutrophilic cellular response in the submucosal layer of the bronchus and perivascular area, and hemorrhage accompanying the formation of fibrin-net-forming (Fig. 1C). Thereafter, hyaline degeneration, infiltration of macrophages, and epithelial proliferation was observed at 24-48 h, and mesenchymal cell proliferation and fibrous thickening of alveoli were present 3-7 days later (not shown). Thus, intratracheal administration of HC1 produced a severe acute lung injury in rats.
DNA Synthesis in Bronchial Epithelia and Alveolar Septa following Lung Injury-To determine the time course of regenerative cell proliferation after acute lung injury, cells undergoing DNA synthesis were identified by the incorporation of BrdU followed by immunochemical staining (Fig. 2). In the untreated rat lung, there were few cells undergoing DNA synthesis (0.95% in bronchus and 1.42% in alveoli). However, there was a marked increase in the number of cells undergoing DNA synthesis following HC1 treatment. The cells undergoing DNA synthesis were predominantly simple columnar cells in the bronchial epithelia (Fig. 2 A ) and alveolar type I1 epithelial cells in alveolar septa, characterized by a round nucleus and a comparatively small cytosol protruding into the alveolar space (Fig.  2 B ) . Several endothelial cells with a small flat nucleus and macrophages in the alveolar space also underwent DNA synthesis, although the number of these cells undergoing DNA synthesis was much fewer than in epithelial cells.
The number of cells undergoing DNA synthesis in the bronchial epithelia and the alveolar septa remarkably increased from 24 h following HCl treatment (Fig. 3). Maximum values of the labeling index were 13.0% at 24 h in bronchial epithelia and 11.7% at 48 h in alveolar septa. Then, labeling indexes markedly decreased 72 h after HC1 treatment and returned to almost normal levels within 1 week.

Change in HGF mRNA in Lung after HCl-induced Lung
Injury-% determine whether or not HGF has a role in regeneration of the acutely injured lung, we examined HGF mRNA in the lung after HC1 treatment (Fig. 4). HGF mRNA was detectable in the intact lung, and it markedly increased with a peak at only 3 h after HC1 injection; the maximum level was &fold higher than normal. The high level of HGF mRNA expression in the lung continued for up to 12 h, and HGF mRNA levels reverted to normal at 24 h in two of three rats after HC1 injection.
HGF Activity in Lung-HGF activity in the injured lung increased as early as 3 h after HCl injection and reached the maximum at 6 h (Fig. 5). The maximum value was 34-fold T. Kawamoto, H. Iwaki, K. Sekiguchi, and T. Nakamura, submitted for publication.

FIG. 1. Histology of the lung after intratracheal HCI injection.
Acute lung injury was induced by intratracheal injection of 0.1 M HCI (2 m l k g of body weight). The lungs were fixed with 70% ethanol and embedded in paraffin. Four-pm sections were deparaffinized and stained with hematoxylin and eosin. A, normal lung; B, 3 h after HCI treatment; C, 12 h after HCI treatment. Burs represent 200 pm.
higher than that of the normal. HGF activity gradually decreased from 12 h after the treatment and reverted to normal within 48 h after the injection. Because HGF activity in the lung was almost completely abrogated by the addition of antiserum against recombinant rat HGF (Fig. 5) but not by preimmune serum (data not shown), HGF activity in the lung was attributable to HGF. These results suggested that HGF was newly synthesized in the injured lung following increase in HGF mRNA expression after the HC1-induced injury. Change in Binding of I2'jl-HGF to the Plasma Membranes -We reported that rapid down-regulation of cell surface HGF receptor was found only in the injured organs after liver or renal insult and not in intact organs (24, 25). Since down-  Hours after treatment

FIG. 3. Time courses of DNA synthesis in bronchial and alveolar epithelial cells in rat lung after HC1-induced injury.
DNA synthesis was detected by BrdU uptake and following immunohistostaining using anti-BrdU monoclonal antibody. Labeling index (percent of cells undergoing DNA synthesis) was determined by counting labeled nuclei among more than 1000 nuclei of bronchial ( 0 ) and alveolar epithelial cells (0). regulation of growth factor receptors occurs immediately after ligand binding, these results suggest that HGF immediately produced at the injured site and other intact organs such as lung, spleen, and kidney (29,38,39) exerts biological activities specifically at the injured site, as a trophic factor for regeneration. Therefore, to examine whether HGF newly synthesized in the lung subjected to insult is involved in lung regeneration, we HGF was partially purified from the lung using heparin-Sepharose. HGF activity was determined by measuring stimulatory effects on DNA synthesis of adult rat hepatocyte in primary culture. Test samples were preincubated with (0) or without ( 0 ) anti-rat HGF I& and assayed for HGF activity. One unit of HGF activity corresponds to the stimulatory effect on DNA synthesis of hepatocytes by adding 10 ndml EGF. Each value represents the mean + S.D. of duplicate experiments on three rats in each group. prepared plasma membranes from injured lungs and intact organs and analyzed HGF receptor. Specific binding of HGF was saturable at about 100 PM (not shown), and the amount of specific binding of HGF to plasma membranes from the lung decreased to 68% of that of normal lung at 3 h after HCI treatment, to 26% a t 6 h, and almost completely disappeared during 12-48 h after HCI injection, as shown in Fig. 6A. The binding of l2"I-HGF recovered to 26% of normal levels at 72 h after the treatment and to 88% 1 week after the insult. Scatchard analysis of lZhI-HGF binding to plasma membranes prepared from normal and HC1-treated rat lung at 3 h after the injection revealed the Kd value to be 58 PM and the number of binding sites was 1530 sitedpg of plasma membranes protein in the normal lung, whereas 1050 sitedpg of plasma membrane protein with K d of 51 PM were present in lung 3 h after treatment ( Fig. 6.4, inset). Therefore, the decrease in specific binding of HGF after the induced injury was not due to change in affinity to the receptor, but rather to a decrease in the number of HGF receptors.
To determine whether down-regulation of HGF receptors was specific to the injured lung, we examined ""I-HGF binding to plasma membranes purified from liver, kidney, and spleen after lung insult. As shown in Fig. 6B. there was no significant change in 12"II-HGF binding in the plasma membranes from these organs. These results clearly show that rapid down-regulation of the HGF receptor occurs specifically in the injured lung.
HGF Concentmtion in Sera of Patients with Lung Dineaxes "To examine changes in HGF levels in sera of patients with various lung diseases, we measured HGF concentration in sera from healthy donors and from patients with lung diseases, using a highly sensitive radioimmunoassay for HGF (Fig. 7). Elevated levels of HGF in sera of patienL9 with lung diseases were found; HGF concentration in sera of 197 healthy donors was 0.33 * 0.10 ng/ml, whereas HGF concentration in sera of 22 patients with lung diseases was 1.22 * 1.01 ng/ml. Among patients with various lung diseases, HGF levels were 4.10 ng/ml in one with silicosis, 2.84 ndml in one with lung abscess, and in seven with pneumonia the level was 1.06-2.80 ng/ml.

DISCUSSION
As the lung is a respiratory organ, exposure to various exogeneous pathogens is inevitable. In alveolar injury, alveolar type I epithelial cells are predominantly damaged, de-epithelialization occurs, and alveolar type I1 cells actively proliferate to reconstruct the epithelium (40). In case of bronchial injury, remained bronchial epithelial cells, Clara cells, and basal cells are thought to proliferate and differentiate into the ciliated simple columnar cells and to compensate for the bronchial epithelia (41). Thus compensatory proliferation and subsequent differentiation of alveolar type I1 and bronchial epithelial cells seem to be essential for lung regeneration. Growth factors are considered to be involved in lung regeneration and several potent growth factors for bronchial and alveolar epithelial cells have been characterized. Proliferation of bronchial epithelial cells is stimulated by epidermal growth factor (EGF), transforming growth factor-a (TGF-a) (42), keratinocyk growth factor (KGF) (43) and insulin-like growth factor-I (IGF-I) (42), but it is inhibited by transforming growth factor-@ (TGF-P) (44). DNA synthesis of alveolar type I1 cells is stimulated by insulin, EGFITGFa, and acidic fibroblast growth factor (aFGF) (40,45, 46). EGF enhances the formation of branching tubule structure in organ cultures of chick embryo (47), and an intravenous injection of EGF in uiuo induces hyperplasia and metaplasia of bronchial epithelial cells (48).
HGF is synthesized and secreted by mesenchymal cells, such as macrophages, endothelial cells, and fibroblasts, and controls proliferation and morphogenesis of a broad spectrum of epithelial cells. It has recently been revealed that HGF is a potent mitogen for rat alveolar type I1 cells in primary culture (461. Therefore, we hypothesized that HGF may have important roles in regulating growth of lung epithelium and in regeneration of the lung, as a paracrine factor. In the present study, we obtained evidence which suggests that HGF has a role in regeneration of the lung following acute lung injury. Following intratracheal HCl injection, HGF mRNA and HGF activity in lung remarkably increased as early as 3-6 h later, and the HGF receptor on the plasma membranes of lung was markedly down-regulated to an undetectable level 12 h later. Since the down-regulation of HGF receptor is probably caused by internalization of the HGF receptor following HGF binding, our results suggest that HGF exerts biological activities in the lung. Following marked changes in HGF mRNA, HGF activity, and the receptor in lung, there was marked increase in DNA synthesis of bronchial epithelial cells and alveolar epithelial cells during 24-48 h after the onset of lung injury. Thus the rapid and sequential induction of HGF mRNA and HGF activity that precede pulmonary epithelial cell proliferation strongly suggests that HGF may have a "trophic" role for the regeneration of lung following acute lung injury. Based on the pleiotrophic actions of HGF as mitogen, motogen (stimulation of cell motility), and morphogen, we propose that HGF may enhance lung regeneration through its multiple biological activities, e.g. not only by stimulating proliferation of bronchial and alveolar epithelial cells, but also by constructing normal tissue architecture of the bronchus and alveolus.
Although the HGF receptor on plasma membranes of the lung was markedly down-regulated, the HGF receptor in liver, kidney, and spleen did not change after lung injury. Thus HGF synthesized after lung injury specifically binds to the receptor in lung, but not to that in other intact organs. We reported that HGF mRNA is markedly induced in the intact lung following partial hepatectomy and unilateral nephrectomy in rats (39) and that the HGF receptor was specifically down-regulated in the injured organ but not in the intact lung (25). Our previous results suggested that HGF synthesized in the intact lung after hepatic or renal injury acts as a hepatotrophic or renotropic factor for regeneration, in an endocrine fashion. Therefore, taken together with present results, HGF synthesized in lung appears to contribute to the regeneration of the lung itself and also to regeneration of distal organs.
HGF concentrations in sera of patients with various hepatic and renal diseases were significantly higher than those in healthy donors. We reported direct evidence that HGF functions as an hepatotrophic and renotropic factor in uiuo; administration of recombinant HGF into mice with hepatic or renal injury remarkably stimulated regeneration of the liver or kidney and prevented the onset of severe hepatic or renal dysfunction (30, 49). All these findings mean that HGF can be considered to treat subjects with hepatic and renal diseases, as well as being a diagnostic for these diseases. Since HGF concentration in sera of patients with various lung diseases were much higher than those in healthy donors, HGF seems to be involved in regeneration of lung also in human. Following acute or chronic lung injury, overgrowth of fibroblasts and overproduction of extracellular matrix often causes lung fibrosis which results in a decrease in respiratory functions. Since HGF is produced by mesenchymal cells and acts predominantly on epithelial cells, HGF is a putative "pulmotrophic factor" for regeneration of the lung, preventing onset of lung fibrosis. Our preliminary experiment using experimental animals has indeed proved that administration of HGF accelerate lung regeneration in uiuo and plans are being designed to clinical administration of HGF to treat subjects with lung disorders. Dr. K. Oshima. and M. Sueihara (Fukuoka Universitv. Fukuoka. JaDan) 24. Higuchi, O., and Nakamura, T. (1991)