Adult respiratory distress syndrome may incorporate in its pathogenesis the hyperplastic proliferation of alveolar epithelial type II cells and derangement in synthesis of pulmonary surfactant. Previous studies have demonstrated that hepatocyte growth factor (HGF) in the presence of serum is a potential mitogen for adult type II cells (R. J. Panos, J. S. Rubin, S. A. Aaronson, and R. J. Mason. J. Clin. Invest. 92: 969–977, 1993) and that it is produced by fetal mesenchymal lung cells (J. S. Rubin, A. M.-L. Chan, D. P. Botarro, W. H. Burgess, W. G. Taylor, A. C. Cech, D. W. Hirschfield, J. Wong, T. Miki, P. W. Finch, and S. A. Aaronson. Proc. Natl. Acad. Sci. USA 88: 415–419, 1991). In these studies, we expand on this possible involvement of HGF in chronic lung injury by showing the following. First, normal adult lung fibroblasts transcribe only small amounts of HGF mRNA, but the steady-state levels of this message rise substantially in lung fibroblasts obtained from animals exposed to oxidative stress. Second, inflammatory cytokines produced early in the injury stimulate the transcription of HGF in isolated fibroblasts, providing a plausible mechanism for the increased amounts of HGF seen in vivo. Third, HGF is capable of significantly inhibiting the synthesis and secretion of the phosphatidylcholines of pulmonary surfactant. Fourth, HGF inhibits the rate-limiting enzyme in de novo phosphatidylcholine synthesis, CTP:choline-phosphate cytidylyltransferase (EC 220.127.116.11). Our data indicate that fibroblast-derived HGF could be partially responsible for the changes in surfactant dysfunction seen in adult respiratory distress syndrome, including the decreases seen in surfactant phosphatidylcholines.
- phosphatidylcholine metabolism
- protein kinase C
- protein kinase A
- cytidine 5′-triphosphate:choline-phosphate cytidylyltransferase
- adult respiratory distress syndrome
the ongoing synthesis and secretion of pulmonary surfactant, a complex lipoprotein produced by alveolar epithelial type II cells, is necessary for normal lung function. Previous studies indicate that the amounts and composition of pulmonary surfactant are perturbed in adult respiratory distress syndrome (ARDS) (10) and bronchopulmonary dysplasia (18, 25), chronic lung injuries in adults and premature newborns, respectively. The etiologies of both syndromes frequently involve reparative processes characterized by increased populations of interstitial fibroblasts and alveolar epithelial cells, and the latter are frequently altered in appearance. The identity of substances responsible for these actions is uncertain, although growth factors are widely proposed as likely candidates.
Paracrine-autocrine regulation involving feedback loops between mesenchymal and epithelial cells is becoming increasingly recognized for its importance in fetal development and injury repair (32). Specifically among lung cells, Brody (5) has demonstrated that mesenchymal-epithelial contacts through pseudopods increase with fetal development, and Adamson et al. (2) have found similar contacts in chronic lung injury in adults. Fibroblasts synthesize several growth factors, including hepatocyte growth factor (HGF), that might affect these changes in surfactant synthesis and cell proliferation (32). HGF is a multifunctional peptide originally recognized for its effects on cell proliferation and its ability to affect cell motility and aggregation. More recent studies (43a) indicate a far wider range of actions, and it may also participate in malignancy, wound repair, and neovascularization concomitant with inflammatory disease. HGF is produced by mesenchymal cells, but they do not contain its receptor, c-met. In contrast, epithelial cells contain the receptor but do not produce the ligand. Thus signaling is exclusively by mesenchymal-epithelial communication (32).
Several reports (27, 38) show that HGF in combination with serum is a mitogen for alveolar type II cells in primary culture, a property that it shares with keratinocyte growth factor (27), acidic fibroblast growth factor (38), and transforming growth factor-α (34). HGF is produced by cloned cell lines of fetal lung fibroblasts (27, 33), and Yanagita et al. (48) have demonstrated that steady-state levels of HGF mRNA are transiently increased in response to an intratracheal injection of HCl. Here we present evidence that HGF may be involved not only in the stimulated proliferation of lung cells in ARDS but also in the reduction of alveolar pools of surfactant. First, we show that oxidative stress leading to chronic lung injury results in a sustained transcription of HGF (compared with the temporary increase found in the protocol of Yanagita et al.). The likely source of this increased HGF is lung fibroblasts. Second, cytokines that are produced in response to lung injury stimulate the transcription of HGF in lung fibroblasts. Third, HGF inhibits both the synthesis and secretion of phosphatidylcholine (PC) when presented to isolated type II cells. Furthermore, we present limited evidence on the mechanisms of these actions by HGF. We find that HGF inhibits the activity of the rate-limiting enzyme in PC synthesis, CTP:choline-phosphate cytidylyltransferase (CT; EC 18.104.22.168), possibly acting through a proline-dependent kinase. We are unable, however, to demonstrate the involvement of protein kinase (PK) C in inhibiting secretion.
Animal procedures. Adult rats were exposed to 100% oxygen at 1 atm, as described (23). In brief, adult rats (175–200 g) were exposed to 100% oxygen for up to 10 days in a stainless steel-Plexiglas chamber with access to water and food ad libitum. Oxygen was delivered at a rate of 15 ml/min and recirculated after being scrubbed for carbon dioxide. Each day the chamber was opened and cleaned, and the animals were aired for 1.5 h in the morning and late afternoon. Rats exposed continuously to 100% oxygen die within 72 h; exposure to room air extends longevity to a time course more consistent with chronic lung injuries in primates. At the end of each exposure period, the animals were removed from the chamber, transported to the laboratory, and killed for cell isolation or biochemical studies.
Animal protocols were approved by the Institutional Animal Care and Use Committee, and all animals used were cared for in accordance with institutional guidelines.
Fibroblast cell culture. Lung tissue was dissociated with 2 mg/ml trypsin (type II; Sigma) and 0.16 mg/ml DNase 1 (Sigma) in MEM for 20 min at 37°C. The cell suspensions were treated with 10% FCS to stop the enzymatic reaction, and the cells were plated in MEM containing 10% FCS, 2.5 μg/ml of Fungizone, and 50 μg/ml of gentamicin. After 60 min at 37°C, the culture medium containing floating cells was removed and replaced with fresh MEM. The adherent fibroblasts used in these studies were passaged three times and displayed a typical fibroblast morphology when examined with inverted phase microscopy. Macrophages or monocytes were not evident.
Culture of NCI-H441 cells. The cells were obtained from American Type Culture Collection as batch F-12518 of the 50th passage. The cells were maintained in McCoy's 5A medium with 10% FCS and 50 μg/ml gentamicin.
Isolation of alveolar type II cells. As described by Dobbs et al. (12), type II cells from normal adult rats were isolated as follows: lungs were perfused with DMEM, lavaged with a calcium-balanced salt solution, and filled with 12 U/ml of pancreatic elastase (Worthington). Tissue was dissociated by incubation at 37°C for 20–30 min. Cells were filtered through gauze pads, washed, and incubated on microbiological plastic coated with rat IgG. Nonadherent cells were recovered and washed by centrifugation, and 1–2 million cells/well were placed in 24-well plates on uncoated inserts (Millipore) for study of surfactant metabolism. Purity, as assessed by staining with Papanicolaou reagents, was generally ∼85% type II cells, with the populations ranging from 80 to 90%.
Study of surfactant metabolism. All experiments were conducted in a defined medium consisting of DMEM containing 10−4 M hydrocortisone, 5 × 10−4 M cAMP, 1 μM/ml insulin-transferrin-selenium (Collaborative Research), and 6 μM arginine-glycine-aspartic acid-serine (Sigma). No FCS was given to the cells at any time during the experiments.
Cells were incubated in phosphate-free defined medium for 1 h. Medium was then removed, and the cells were given 50 μCi/well of [32P]orthophosphate in the same defined medium, together with either rat HGF (5 ng/ml) or solvent. In separate experiments testing the effects of inhibiting tyrosine kinases, cells received HGF plus sodium orthovanadate (SOV; 20 μM) or SOV alone. Cells were incubated for up to 48 h at 37°C in a CO2incubator, and at varying times, medium was removed from the wells and cells were recovered from the inserts by gentle trituration. After a wash in DMEM, the cells were suspended in 2 ml of H2O and lysed by freeze-thaw. PCs were isolated with the following protocol. Lipids in cells and medium were extracted and dissolved in a small volume of 1:1 (vol/vol) chloroform-methanol, and the phospholipids were separated by silica gel thin-layer chromatography as previously described (23). Phospholipids were detected by a 2-(p-toluidino)naphthalene-6-sulfonic acid (TNS) spray, and the PC band was recovered by scraping and transesterified with 1% H2SO4 in methanol for 2 h at 70°C together with an internal standard of a known amount of diheptadecanoyl PC. Methyl esters were extracted with hexane, and PC content was quantified by gas-liquid chromatography. The lower phase of the extract containing the radioactive glycerolphosphocholine was used to quantify the incorporation of orthophosphate by scintillation counting.
Extraction and analysis of RNA. Steady-state levels of HGF mRNA in lung fibroblasts were determined by Northern blots or RNase protection assay (RPA). Confluent monolayers were lysed with guanidinium isothiocycanate, and total RNA was isolated as described by Chomczynski and Sacchi (9). For Northern blots, ∼50 μg of total RNA were applied for electrophoresis in 1% agarose-formaldehyde gels and transferred to 0.45 μM Nytran in 10× sodium chloride-sodium citrate (0.15 M NaCl and 15 mM Na3C6H5O7, pH 7.0). A 1.4-kb cDNA fragment of rat HGF (41) coding for portions of the α- and β-chains was labeled with [32P]dCTP with a random-labeling kit (Pharmacia). Hybridization was performed overnight at 42°C using ∼1 × 106counts ⋅ min−1(cpm) ⋅ ml−1 of labeled cDNA. The blot was washed for 15 min once with 0.1% sodium dodecyl sulfate (SDS) in 0.1× sodium chloride-sodium phosphate-EDTA (SSPE; 0.15 M NaCl, 10 mM Na3PO4, and 1 mM EDTA, pH 7.4), followed by two washes with 0.1% SDS in 6× SSPE. The radioautographs were evaluated by comparison with the total amount of loaded RNA and quantified by staining with ethidium bromide and gel scanning.
RPA was done with an RPA II kit supplied by Ambion following the manufacturer's recommended procedures. A 200-bp antisense cRNA probe to nucleotides 1626–1827 of rat HGF was synthesized by RT-PCR and a Maxiscript kit (Ambion) with RNA isolated from rat liver. The probe was sequenced, and the sequence was confirmed to be identical to authentic rat HGF (41).
Assay of PKA activity. Analyses were conducted with commercially available kits following instructions supplied by the manufacturer (GIBCO BRL, Life Technologies). Briefly, 2–6 million cells were incubated in culture with 5 ng/ml HGF or with the buffer control for 1–2 h with the culture conditions described inStudy of surfactant metabolism. The cells were recovered, washed by centrifugation, and extracted with 20 mM Tris, 0.5 mM EDTA, 0.5 mM EGTA, 0.5% Triton X-100, and protease inhibitors. PKA activity was assayed directly from the lysates using a PKA-specific activator as described in the kit.
Assay of PKC activity. H441 cells were washed twice with ice-cold PBS and scraped in 200 μl of homogenization buffer [20 mM Tris ⋅ HCl (pH 7.5), 0.25 M sucrose, 0.5 mM EDTA, 0.5 mM EGTA, 25 μg/ml of leupeptin, 25 μg/ml of pepstatin, 1 mM phenylmethylsulfonyl fluoride, and 0.2 mM SOV]. To separate cytosolic and membrane fractions, the cells were homogenized and centrifuged at 40,000 rpm for 30 min at 4°C (Beckman). The membrane pellet was dissolved in 100 μl of homogenization buffer containing 0.05% Triton X-100, kept on ice for 30 min, and centrifuged as above. Cytosolic and solubilized-membrane fractions (10 μl each) were incubated with 2 μg of myelin basic protein peptide, acids 4–14 (MBP4–14), 0.8 μg/ml of diolein, 8 μg/ml of phosphatidylserine, 20 μM ATP, and 1 μCi of [32P]ATP in 50 μl of kinase buffer [20 mM Tris ⋅ HCl (pH 7.5), 10 mM magnesium acetate, 1 mM CaCl2, and 50 μg/ml of leupeptin]. The peptide substrate MBP4–14 has been shown to be resistant to phosphatase digestion (13) and is a specific substrate for PKC (50). The reaction was allowed to proceed for 15 min at 30°C. Aliquots (10 μl) of the reaction mixtures were applied to phosphocellulose paper disks, and the disks were washed three times with 1% acetic acid and twice with water before counts were measured in a scintillation counter. Nonspecific activity in the cytosolic and membrane fractions was estimated by subtracting activity after phosphatidylserine and diolein were omitted and 1 mM EDTA was substituted for CaCl2. Activities were calculated as the amount of phosphorylation of MBP4–14 per microgram of cell protein per minute.
Quantification of CT activity. Activity was measured as described by Vance et al. (43) as the amount of [14C]phosphocholine converted to [14C]CDPcholine. Approximately 2 million H441 cells were placed overnight in McCoy's 5A medium without FCS. The following morning, the cells were washed and given new medium containing 10 ng/ml of human HGF (R&D Systems) or equivalent volumes of control medium. Cells were recovered at various times after HGF was provided, washed twice, and assayed for CT activity. Results of different experiments were compared by expressing data as percent of control activity at each sampling time. Similar procedures were followed with freshly isolated type II cells placed overnight in DMEM to allow stabilization and treated the next morning with 10 ng/ml of HGF.
Quantitation of diacylglycerol content. H441 cells were plated in 12-well plates in McCoy's 5A medium with 10% FCS. The morning of the experiment, the cells were washed twice with medium without FCS and equilibrated for 1 h in fresh McCoy's without FCS containing ∼100,000 cpm of [9,10-3H] palmitic acid. HGF was added after 1 h (time 0 for the timing of the experiment), and cells were harvested at varying times after isotope was added. Cells were washed twice and scraped into 0.8 ml of H2O. Lipids were extracted and spotted together with 1,2-diacylglycerol (DAG) on silica gel thin-layer plates. The plates were developed with a solvent system of ethyl ether-benzene-ethanol-acetic acid (40:50:2:0.2), followed by ethyl ether-hexane (6:94) (14). 1,2-DAG was clearly separated from 1,3-DAG and cholesterol in this solvent system. The lipids were visualized by spraying with TNS, and the 1,2-DAG spot was scraped and counted. Time-matched control experiments were run at all times, and changes in 1,2-DAG induced by HGF were quantified by differences from these time-matched control experiments.
Activation of p42/p44 mitogen-activated PK. Cells seeded in 60-mm tissue culture dishes were treated with HGF (10 ng/ml) for various time periods. The cells were washed twice with PBS and lysed in 200 μl of Tris ⋅ HCl buffer (pH 7.4) containing 25 mM Tris, 1% igepal, 150 mM NaCl, 50 mM NaF, 200 μM SOV, and 1 mM phenylmethylsulfonyl fluoride. The lysates were probe-sonicated for 5 s and centrifuged at 14,000 rpm for 10 min to remove the cell debris. Lysate protein was separated on 10% SDS-polyacrylamide, and the gels were probed with anti-phospho-specific p42/p44 mitogen-activated PK (MAPK; New England Biolabs) used at a 1:1,000 dilution in Tris-buffered saline containing 5% BSA and 0.2% Tween 20. Equal protein loading in the gels was confirmed by blotting the membranes with anti-p42/p44 MAPK. The proteins of interest were detected by an enhanced chemiluminescence kit (Amersham). Anti-rabbit IgG-horseradish peroxidase secondary antibody (Santa Cruz Biotechnology, Santa Cruz, CA) was used at 1:5,000 dilution in the same buffer.
Materials. Rat HGF, as previously described (41), was obtained by recombinant technology by Dr. T. Nakamura, Division of Biochemistry, Osaka University Medical School, Osaka, Japan. Polyclonal antiserum to this recombinant rat HGF was produced in rabbits by conventional methods. Other materials were obtained from commercial suppliers as noted in the text.
Statistical analysis. To compare experiments, specific activities (in cpm/μg PC for metabolic experiments) or enzyme activity units (activity/μg protein) are expressed as a ratio of HGF to control values. These data were tested for significance with a one-factor ANOVA where time and treatment were a combined grouping. Individual means for each group and time were compared with Fisher's protected least squares difference test, and significance was accepted for P ≤ 0.05. We also tested all data at each time point using an unpaired t-test, selecting for groups that differed from a population mean of 1 (no difference between specific activities of HGF and control groups). The results were essentially identical to those obtained with the ANOVA and are not presented.
HGF mRNA and protein are increased in lung tissue and fibroblasts exposed to oxidant stress. HGF may be transcribed in alternative forms (6). The principal transcript is ∼6 kb; an alternative transcript of ∼2.4 kb, which codes for a truncated form of HGF, competes for membrane receptors and is an inhibitor. The results of a Northern blot from four animals exposed to varying periods of 100% oxygen are shown in Fig. 1. HGF mRNA was very low in the normal animal but markedly increased with exposure to high oxygen. Only the larger transcript was found in either the normal animal or those exposed to 100% oxygen. The smaller (6) transcript was not detected (data not shown).
After determining the size of the HGF transcript, we used RPA to quantify changes in steady-state levels of HGF mRNA in RNA extracted from lung tissue. The results from one of two experiments are shown in Fig. 2 A. HGF mRNA was not detected by RPA in 10 μg of total RNA of nonexposed animals but was evident after 48 h of exposure to 100% oxygen and markedly increased after 72 h. Using quantitative densitometry, we estimate that HGF mRNA increased by >10-fold. The time course of HGF protein was identical to that of the mRNA and is shown in Fig. 2 B.
We hypothesized that fibroblasts may be a particularly important source of HGF acting on type II cells. Therefore, we assayed steady-state HGF mRNA in fibroblasts isolated from animals exposed to 100% oxygen for 2–6 days and then maintained in culture through two passages. Shown in Fig. 3 are the results obtained from five experiments. HGF mRNA was not detected in normal animals but became notable after 3 days of exposure (based on four animals that were examined) and was still prominent at 5–6 days (based on two animals). The results suggest that the source of the increased HGF after injury may be, in part, fibroblasts.
Tumor necrosis factor-α and interleukin-1β induce the transcription of HGF in fibroblasts. We sought possible mechanisms for the increase in HGF found in animals with lung injury induced with 100% oxygen. Cytokines are produced early in the inflammatory phase of lung injury, and cytokines have been reported to induce HGF in stabilized fetal cell lines (40). We studied, therefore, whether cytokines would be able to induce HGF mRNA in lung fibroblasts of normal, uninjured adult animals.
Lung fibroblasts were isolated and placed in culture for two passages. About 10 million cells were given varying concentrations of either tumor necrosis factor (TNF)-α or interleukin (IL)-1β for 24 h, the cells were then harvested, and total RNA was isolated. HGF mRNA was quantified by RPA. The results of one experiment are shown in Fig.4 and are representative of those from two independent experiments. HGF mRNA was undetectable in control fibroblasts. However, in cells exposed to either 1 ng/ml of TNF-α or 1 ng/ml of IL-1β, HGF mRNA was abundant. Using densitometry of the gels to estimate the changes induced by the cytokines, we found that 1 ng/ml of TNF-α increased HGF mRNA by >300-fold (320 ± 83, mean ± range of 2 experiments) and 1 ng/ml of IL-1β increased HGF mRNA by >3-fold (3.25 ± 0.95).
Changes in surfactant metabolism induced by HGF. The time course of the changes seen in cell and medium pools from one experiment is shown in Fig. 5. Because HGF activates signaling pathways by tyrosine phosphorylation of c-met (24), we included 20 μM SOV (39) in all wells, both control and HGF treated. Cells accumulated radioactivity in PC pools throughout the entire course of the 45- to 48-h experiments, although the rates of accumulation plateaued after 24 h. In contrast, specific activities in medium increased with time. The data suggest sustained metabolic viability throughout the course of the experiment, with a probable tendency toward equilibration between new synthesis and reuptake in the cell pool. The mean changes induced by HGF from all experiments pooled and tested for significance are seen in Fig.6. HGF combined with SOV decreased specific activities in cells to 60–75% of control values at 5–48 h (P ≤ 0.05). In medium, the specific activity at 5 h was reduced but was not significant. At all other times, however, the inhibition of medium specific activity was greater than the inhibition in cells. Specific activities of medium from HGF plus SOV-treated cells were generally reduced to 40–80% of control values in the time period between 17 and 48 h and in occasional individual experiments, even less. When we compared control cells with and without SOV, we found that SOV by itself did not markedly affect the synthesis or secretion of phospholipids (data not shown).
HGF used alone also reduces cell and medium specific activities, but the inhibition is less than when SOV is included (Fig.7). Cell specific activity at 5 h was ∼65% of control specific activity (P ≤ 0.05) but was reversed at 17–48 h to ∼85–90% of control specific activity (P ≥ 0.1). In medium, there were no significant changes at 5 h. From 17 to 48 h, however, HGF induced decreases in medium specific activities, with significant reductions from 17 to 48 h, which varied from 65 to 85% of control value.
Amounts and composition of surfactant PCs are unchanged by HGF.Neither HGF nor HGF-SOV increased the amount of cellular PC, concomitant with a possible stimulation of cell division, but this would not be expected because we used nonsynchronized cell cultures, and none contained serum, which is necessary for cell division (27). Medium pools in some experiments were reduced (in one experiment, to ∼35% of the corresponding control pools), but the results were varied and were not significant in the averaged experiments.
The compositions of the cell and medium PC pools were routinely measured. Neither was affected by HGF. Within the limits of our detection, the medium was composed of 80–90% disaturated PC and the cells >60%.
HGF inhibits the activity of CT. CT is the rate-limiting enzyme in the de novo synthesis of PC (29), and we evaluated the effects of HGF on its activity. Because type II cells are difficult to obtain in large quantities, we used H441 cells to develop a time course of the effects of HGF on the activity of CT (Fig.8). H441 cells are frequently used for the study of surfactant because they synthesize several of the proteins of surfactant (47) and demonstrate other characteristics of a surfactant-producing cell (26). The results show that HGF inhibits membrane-bound CT activity between 0 and 1 h. The effects are relatively rapid because the 0-h data, which more accurately reflect the ∼5 min required for chilling the cells to ∼5°C and processing, show a reduced (but not significant) activity in all samples. The inhibition was reversed by 3 h. This time course for the change in CT activity, therefore, is compatible with the observed inhibition in cellular synthesis of PC found after 5 h of HGF but reversed after 17 h.
Two mechanisms for the regulation of CT have been proposed: 1) increases in phosphorylation, which reduce the binding of CT to lipid membranes, thereby decreasing activity (3); and 2) changes in the content of activating lipids, namely DAG and fatty acids (20, 42). Changes in DAG content may, in turn, depend on the activation of PKA (20), which phosphorylates acetyl-CoA carboxylase and inhibits its activity (19), thereby reducing fatty acid synthesis and DAG content.
PKA activity is stimulated by HGF. We measured the effects of HGF on PKA activity in primary cultures of type II cells. Cells were treated for 1 h with 5 ng/ml HGF. Control cells were treated similarly, except that HGF was omitted. The results from two experiments showed that HGF increased PKA activity by 40 and 70%. These effects on PKA are consistent with those reported by Grumbles et al. (17), who found that HGF induced a modest but statistically significant increase in PKA activity in bone cells from rat tibia.
DAG content is reduced in cells treated with HGF but not at times consistent with CT inhibition. H441 cells were treated with HGF for times varying from 1 to 17 h, and DAG content was measured (Fig.9). DAG content was reduced by ∼50% at 3 and 4 h, consistent with changes in PKA activity at 1 h. However, there were no detectable changes in DAG content at 1 h. This mechanism, therefore, cannot explain the rapid decrease in CT activity at the 1-h and 0-h time points.
HGF rapidly activates p42/p44 MAPK and effects of HGF on CT are reversed by an inhibitor of proline-dependent kinases. CT contains numerous serines clustered near the carboxy terminus (21), and seven of these have consensus sequences for putative substrates of proline-directed kinases. HGF has been reported to activate p42/p44 MAPK in hepatocytes (1, 15) and H441 cells (8), providing a possible mechanism for the rapid inactivation of CT by HGF. We confirmed that HGF activated p42/p44 MAPK under the conditions in which we conducted our experiments, as shown in Fig. 10. HGF phosphorylates p42/p44 MAPK in a time-dependent manner, suggesting a corresponding time course of kinase activity. Abundance of phosphorylated MAPK peaked by 10 min to 180 ± 19% (SE) of control level (n = 3 experiments) and returned to about control levels at 1 h. Total p42/p44 MAPK was not changed.
If p42/p44 MAPK phosphorylates CT and subsequently reduces its activity, we would expect that an inhibitor of this MAPK would mitigate this effect. We used 40 μM olomoucine, a purine derivative, because, at low concentration, it is a relatively specific inhibitor of proline-dependent kinases, including p42/p44 MAPK (IC50 for p44 MAPK is 25 μM), but it does not affect other types of kinases, such as PKC (44). Olomoucine has been used previously in human keratinocytes to demonstrate that the phosphorylation of CT, and reduction in its activity, may involve a proline-dependent kinase (46). Isoolomoucine, which is inactive at this concentration, was used as a negative control. The results are shown in Fig.11. Olomoucine blocks the inhibition of CT activity by HGF at 1 h (P ≤ 0.01), whereas isoolomoucine was less effective, and its effects were not significant.
HGF does not inhibit PKC activity. In a now classic and oft quoted paper, Sano et al. (36) demonstrated that the secretion of surfactant PC in isolated type II cells is stimulated by agonists that stimulate PKC. In the past 15 years, several other investigators have identified other purported secretagogues of PC, and several of these, when investigated, appear to converge on PKC early in their respective signaling pathways. [See, for instance, work on low- and high-density lipoproteins (30), ATP (7, 16, 31), and endothelin-1 (37).] Thus it would appear that PKC has an important role in the signaling pathways used by several secretagogues of surfactant PC and that it occupies an early upstream position in those pathways. To look for a direct connection between changes in PKC activity and the inhibition of PC secretion by HGF, we assayed PKC activity in the membrane fraction of lysates of H441 cells at times varying from 17 to 41 h. At these times, we found a significant reduction in the specific activity of PC in the medium. Membrane-associated PKC activity in these cells (expressed as a percentage of control cells) was as follows: 17 h, 93.7 ± 3.8%; 24 h, 105.3 ± 9.5%; 41 h, 87.7 ± 3.3%. HGF does not significantly reduce membrane-associated PKC activity in the time period when secretion is inhibited.
These results may have relevance to the pathophysiological changes that occur with chronic lung injury. Well documented in this condition (as in injuries in other tissues) is an early surge of inflammatory cytokines, prominently TNF-α and IL-1β, followed by increased amounts of growth factors (11, 22) accompanying an increased proliferation of epithelial cells and fibroblasts. Our data are consistent with these in vivo observations because they show increased levels of an alveolar epithelial mitogen, HGF, and because they demonstrate that this induction may be through cytokine stimulation. Our data also help explain the reduction in pulmonary surfactant also seen in ARDS, a seemingly paradoxical finding in light of the increased numbers of type II cells (10).
We used H441 cells to explore possible biochemical pathways that might be interrupted by HGF. These cells have been widely used to study surfactant metabolism. Because these experiments required very large numbers of cells, their utilization in this context is very attractive. Based on results found in these cells, we hypothesize that the effects of HGF on surfactant synthesis may be due to a reduction in CT activity either via its phosphorylation by a MAPK or through the MAPK-dependent phosphorylation of an intermediate that ultimately impacts CT activity. Consistent with this hypothesis, the effects of HGF on CT are reversed by an inhibitor of proline-dependent kinases, which would include all three classes of MAPK. We have not attempted, however, to discern which of the proline-dependent kinases are actually involved. Such studies would probably involve multiple inhibitors and the use of dominant negatives in transfected cells and are beyond the scope of our current experiments. Clearly, the relevant kinase may be one other than p42/p44 MAPK. The effects of HGF on surfactant are also increased when protein tyrosine phosphatases are inhibited by SOV, presumably in part by potentiating tyrosine phosphorylation on the HGF receptor (24), although other intermediaries in the signal transduction pathway would also be affected. An alternative mechanism proposed for the regulation of CT activity is through changes in the content of activating lipids, namely DAG and fatty acids (20). We found a decrease in DAG content induced by HGF after 3–4 h, but this change cannot explain the decrease in CT activity at 1 h. We also observed an increase in PKA activity after 1 h, and CT is a substrate for PKA phosphorylation in vitro (29, 35). It is uncertain whether this mechanism has relevance in intact cells (20).
We have been unable to provide a mechanistic explanation for the inhibition of secretion that is observed at times >17 h. This is a time span likely to involve mRNA transcription and new protein synthesis, but such proteins have not been identified. Other papers also indicate that secretion is inhibited only after long exposure to a cytokine or growth factor; witness the effects on surfactant found for TNF-α (4, 47) and transforming growth factor-β (45). PKC is a known secretagogue of surfactant PC, and we reasoned that this long time course might involve the synthesis of a regulator protein that inactivates PKC just before the observed inhibition of secretion. This was not found to be the case. PKC activity was not reduced between 17 and 41 h, implying that PKC is not involved in this regulation. An alternative explanation (and one less likely in our opinion) is that the inhibition is through a PKC isoform in relatively low abundance compared with other PKC types, and the changes in it are not detected by a relatively nonspecific assay of PKC activity.
This is, to our knowledge, the first report that HGF inhibits lung PC synthesis and secretion, possibly, in part, through the inhibition of CT. The effects of HGF in ARDS may extend beyond localized actions in the lung. Lung cells not only synthesize increased amounts of HGF in localized lung injury, but they also appear to release HGF into the circulation as they respond to humoral substances released by extrapulmonary tissues undergoing acute inflammatory injury (48, 49). Thus HGF produced by the lung in response to localized injury may result in wider effects reflected in systemic changes in PC metabolism.
Financial support for this work came from National Heart, Lung, and Blood Institute Grants HL-43704 and HL-52648.
Address for reprint requests and other correspondence: R. J. King, Dept. of Physiology, Univ. of Texas Health Science Center at San Antonio, 7703 Floyd Curl Dr., San Antonio, TX 78284-7756 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society