Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2008 Jun 13.
Published in final edited form as: Am J Physiol Lung Cell Mol Physiol. 2008 Feb 1;294(4):L778–L786. doi: 10.1152/ajplung.00410.2007

Hepatocyte growth factor regulates cyclooxygenase-2 expression via β-catenin, Akt, and p42/p44 MAPK in human bronchial epithelial cells

Young H Lee 1, Yuichiro J Suzuki 2, Autumn J Griffin 1, Regina M Day 1
PMCID: PMC2427436  NIHMSID: NIHMS52845  PMID: 18245266

Abstract

Hepatocyte growth factor (HGF) is upregulated in response to lung injury and has been implicated in tissue repair through its antiapoptotic and proliferative activities. Cyclooxygenase-2 (COX-2) is an inducible enzyme in the biosynthetic pathway of prostaglandins, and its activation has been shown to play a role in cell growth. Here, we report that HGF induces gene transcription of COX-2 in human bronchial epithelial cells (HBEpC). Treatment of HBEpC with HGF resulted in phosphorylation of the HGF receptor (c-Met), activation of Akt, and upregulation of COX-2 mRNA. Adenovirus-mediated gene transfer of a dominant negative (DN) Akt mutant revealed that HGF increased COX-2 mRNA in an Akt-dependent manner. COX-2 promoter analysis in luciferase reporter constructs showed that HGF regulation required the β-catenin-responsive T cell factor-4 binding element (TBE). The HGF activation of the COX-2 gene transcription was blocked by DN mutant of β-catenin or by inhibitors that blocked activation of Akt. Inhibition of p42/p44 MAPK pathway blocked HGF-mediated activation of β-catenin gene transcription but not Akt activation, suggesting that p42/p44 MAPK acts in a parallel mechanism for β-catenin activation. We also found that inhibition of COX-2 with NS-398 blocked HGF-induced growth in HBEpC. Together, the results show that the HGF increases COX-2 gene expression via an Akt-, MAPK-, and β-catenin-dependent pathway in HBEpC.

Keywords: fibrosis, signal transduction, proliferation, tissue repair, T cell factor-4 binding element

Hepatocyte growth factor (HGF), a multi-functional factor, induces cellular activities including growth, motility, and morphogenesis (26); the diverse functions of this factor are specific to the target tissue and cell type (26). HGF promotes normal repair in a number of organs, including the lung (10, 14, 16, 18). The activity of HGF as a repair factor has been demonstrated in the lung using several experimental models of lung injury and tissue repair. Liu et al. (21) reported that single injections of HGF protein prevented hydrogen peroxide-induced acute lung injury in rats. Simultaneous or delayed administration of HGF protein also prevented bleomycin-induced endothelial and epithelial cell apoptosis, appearance of fibroblast foci, and accumulation of collagen in C57BL/6 mice (38). Administration of adenovirus expressing HGF or HGF gene expression using in vivo transient plasmid transfection prevented bleomycin-induced increased collagen expression in the lung and preserved normal lung architecture (35). Thus, the administration of either HGF protein or HGF gene transfer can prevent lung injury and fibrosis, suggesting the therapeutic potentials of HGF for lung diseases.

Because epithelial repair is a critical step for normal tissue regeneration in the lung (28), the functions of HGF in growth induction and prevention of apoptosis may be important for its antifibrotic and repair activity subsequent to lung injury (30). In various cell types, the mechanism(s) of HGF-induced growth and inhibition of apoptosis involves the activation of two major signal transduction pathways: the p42/p44 mitogen-activated protein kinase (MAPK) pathway and the phosphatidylinositol 3-kinase (PI3K)/Akt pathway (6, 11). A study of cardiac remodeling following ischemia/reperfusion injury suggested that HGF prevented cardiac remodeling via PI3K/Akt signaling (34). Akt activity is often associated with the prevention of apoptosis (4), and the activation of Akt is believed to be an important part of the antiapoptotic signaling mechanism of HGF.

Here we show that HGF upregulates cyclooxygenase-2 (COX-2) in human bronchial epithelial cells (HBEpC) in a PI3K/Akt-dependent manner. HGF activates transcription of the COX-2 promoter, and this regulation requires the TBE element, but not the CRE, NF-IL-6, or NF-κB promoter elements. This activation was suppressed by a dominant negative (DN) mutant of β-catenin. HGF activates GSK-3β, the nuclear accumulation of β-catenin, and gene transcription controlled by a β-catenin-responsive promoter. A recent study in MDCK cells showed the requirement of the p42/p44 MAPK pathway for increased expression of COX-2 and subsequent increase in PGE2 (41). We also show that HGF activation of the β-catenin minimal-responsive reporter regulation also requires p42/p44 MAPK. Therefore, HGF regulation of the COX-2 promoter requires the PI3K/Akt, β-catenin, and MAPK pathways.

MATERIALS AND METHODS

Reagents

Fetal bovine serum (FBS 100–106) was from Gemini Bio-Products (Woodland, CA). DMEM medium, RPMI medium, fungizone, tissue culture antibiotics, and Dulbecco’s PBS were purchased from Invitrogen (Carlsbad, CA). Antibodies for phosphorylated proteins p42/p44 MAPK, GSK-3β (Ser9), c-Met (Tyr1234/Tyr1235), c-Met (Tyr1349), and Akt (Thr34), and unphosphorylated forms of Akt and GSK-3β were purchased from Cell Signaling Technology (Danvers, MA). Antibodies against non-phosphorylated c-Met, phosphotyrosine, and non-phosphorylated p42/p44 MAPK were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The PI3K inhibitor LY-294002 and COX-2 inhibitor NS-398 were purchased from Sigma; U0126 MEK1/2 inhibitor was from Cell Signaling Technology. Luciferase reporter-COX-2 promoter constructs from −1432 bp to +59 bp wild type (wt) and the COX-2 promoters containing the TBE element point mutation or the TBE element deletion were gifts from Dr. C. Harris (National Institutes of Health, Bethesda, MD) (1). The luciferase reporter-COX-2 promoter construct from −327 bp to +59 bp was a gift from Dr. H. Inoue (Nara Women’s Univ., Nara, Japan) (13). The TOP-FLASH β-catenin minimal responsive luciferase reporter (with thymidine kinase elements) was from MTR Scientific (Ijamsville, MD). The β-catenin (CTNNB1) DN mutant Y654/670 expression plasmid and empty vector control (p3XFlagCMV) were gifts from Dr. S. S. Monga (Univ. of Pittsburgh School of Medicine, Pittsburgh, PA)(39). Purified human HGF protein was a gift from Dr. G. vande Woude (Van Andel Research, Grand Rapids, MI).

Cell culture

HBEpC were purchased from Cell Applications (San Diego, CA). Passage 2–8 cells were used for all experiments. HBEpC were cultured in defined medium provided by Cell Applications in 5% CO2 at 37° in a humidified atmosphere in a culture incubator.

Cell lysate

To prepare whole cell lysates, the cells were washed in cold PBS and solubilized with 50 mM HEPES solution (pH 7.4) containing 1% (vol/vol) Triton X-100, 4 mM EDTA, 1 mM sodium fluoride, 0.1 mM sodium orthovanadate, 1 mM tetrasodium pyrophosphate, 2 mM PMSF, 10 μg/ml leupeptin, and 10 μg/ml aprotinin. To prepare nuclear extracts, cells were washed in PBS and incubated for 15 min at 4°C in 10 mM HEPES (pH 7.8), 10 mM KCl, 2 mM MgCl2, 4 mM EDTA, 0.1 mM PMSF, 5 μg/ml leupeptin, 5 μg/ml aprotinin, 95 μM sodium fluoride, 0.1 mM sodium orthovanadate, and 1 mM tetrasodium pyrophosphate. Nonidet P-40 was then added to a final concentration of 0.6%, mixed vigorously, and centrifuged. Pelleted nuclei were resuspended in 50 mM HEPES (pH 7.8), 50 mM KCl, 300 mM NaCl, 0.1 mM EDTA, 0.1 mM PMSF, and 10% (vol/vol) glycerol, then mixed for 30 min at 4°C, centrifuged, and the supernatant harvested.

Western blot

Cell lysates or nuclear extracts (10 μg protein/sample) were electrophoresed through a reducing SDS polyacrylamide gel and electroblotted onto nitrocellulose membrane. Membrane was blocked and incubated with the polyclonal IgG for phosphospecific proteins. To normalize the results, the same blots were stripped and reblotted for total proteins. The levels of phosphoproteins and proteins were detected with horseradish peroxidase-linked secondary antibodies and the ECL System (Amersham Life Science, Arlington Heights, IL).

GE microarray and analysis

The GEArray Q Series Human Cancer Pathway Finder Gene Arrays (SuperArray Bioscience, Frederick, MD) was used to analyze the expression of 96 genes with chemiluminescent detection; the Array was performed according to the manufacturer’s instructions. Briefly, following treatment of cells, with or without adenoviral infection, total RNA was isolated with TRIzol (Invitrogen). RNAs were labeled and converted to cDNA; hybridization was performed as directed in the kit, followed by chemiluminescence detection. For normalization, each background-corrected data value on each array was divided by the median background-corrected data value on that array.

Adenoviral infection

The adenovirus-directed gene transfer was performed by adding the gene-carrying replication-deficient adenovirus (50 pfu/cell) to cells at 70–80% confluence in 35-mm plates. Cells were incubated with adenovirus in 0.7 ml of serum-free DMEM with gentle shaking every 15 min. After 1 h, 0.8 ml of 0.1% FBS-containing medium was added, and cells were cultured for 48 h. An adenovirus for the expression of the DN mutant Akt (T308A, S473A) was a kind gift from Dr. Kenneth Walsh (Boston Univ., Boston, MA).

RT-PCR

Total cell RNA was obtained from cultured cells using TRIzol (Invitrogen). RNA concentrations were determined spectrophotometrically at 260 nm. One microgram of RNA from each sample was reverse transcribed for 10 min at ambient temperature, followed by 30 min at 42°C, in a 20-μl reaction containing 50 units of murine leukemia virus reverse transcriptase (MuLV-RT) in 10 mM Tris·HCl (pH 8.3), 50 mM KCl, 5 mM MgCl2, 2.5 μM oligo(dT)16, 1 mM each dNTP, and 1 unit of RNase inhibitor. Samples were heated to 95°C for 10 min to inactivate the MuLV-RT and stored at −20°C. All RT reagents were purchased from Applied Biosystems (Foster City, CA). As a control for DNA contamination, RNA was treated with DNase in control experiments. Two micrograms of RNA from each sample was incubated with 10 units of DNase in supplied buffer (Boehringer Mannheim, Indianapolis, IN) for 15 min at ambient temperature before the addition of EDTA at 2.5 mM (final concentration).

One microgram of cDNA from the RT reaction was used in a 50-μl PCR reaction with 0.4 μM each forward and reverse primer, 200 μM each dNTP, and 1 unit of iTaq DNA polymerase and supplied PCR buffer (BioRad Laboratories, Hercules, CA). Control reactions were performed with RNA with no RT to ensure that the PCR product obtained was from cDNA and not genomic DNA. PCR reactions were optimized for annealing temperatures using a temperature gradient in a BioRad iCycler. The Relative RT-PCR kit for Human COX-2 was used for quantitative PCR with an internal standard according to the manufacturer’s instructions (Ambion, Austin, TX). Primers for GAPDH, used for normalization of the PCR results, were purchased from BD Biosciences Clontech (Palo Alto, CA); these sequences were: forward primer 5′-ACCACAGTCCATGCCATCAC-3′ and reverse primer 5′-TCCACCACCCTGTTGCTGTA-3′. Reactions were carried out for 30 cycles using the following conditions: 30 s, 94°C denaturation; 1 min, 56–60°C annealing; 2 min, 72°C extension. The last cycle extension was for 10 min. PCR reactions were analyzed on a 1.5% agarose gel in Tris EDTA buffer, and bands were visualized using ethidium bromide.

Plasmid transfection and luciferase assays

The day before transfection, cells were plated at 1.4 × 105 cells per well in a 12-well plate. One microgram of DNA per well was transfected using the Fugene 6 transfection reagent (Roche Diagnostics, Indianapolis, IN) according to the manufacturer’s instructions in serum-free, antibiotic-free medium. Cotransfection of the Renilla luciferase reporter under the control of constitutive thymidine kinase promoter (pTK-Renilla luciferase) was performed to normalize for transfection differences between wells, with a ratio of luciferase reporter:Renilla of 6:1 with a total concentration of DNA equal to 1 μg; when the DN β-catenin mutant Y654/670 plasmid (or empty vector control) was cotransfected, the ratio of luciferase reporter:DN expression vector:Renilla was 6:3:1, with a total concentration of DNA equal to 1 μg. Cells were transfected for 6 h, and then medium was replaced with 0.01% FBS medium containing antibiotics. Cells were treated 1 h later with HGF and/or inhibitors for times indicated for each experiment.

Luciferase assays were performed using the Dual Luciferase Assay Kit (Promega, Madison, WI) to monitor firefly luciferase and Renilla luciferase. Transfected cells were washed twice in PBS and lysed for 50 min at ambient temperature in Passive Lysis Buffer (Promega). Cellular debris was removed by centrifugation at 14,000 g for 30 s. Cell lysates were added at a 1:5 ratio to Luciferase Assay Reagent II. Firefly luciferase activity was read in a model TD-20/20 luminometer (Turner Designs, Sunnyvale, CA). Subsequently, Stop and Glow was added and mixed by pipette, and the Renilla luciferase activity reading was taken after inactivating the firefly luciferase activity. The ratio of firefly luciferase to Renilla luciferase was observed for each well of transfection.

Cell growth assay

HBEpC were plated in triplicate at a density of 5 × 104 cells/35-mm dish in defined medium. To attached cells, HGF was added (25 ng/ml) to the medium on days 1, 2, and 4. The NS-398 COX-2 inhibitor was used at 10 μM and added to cells before the addition of HGF. Cells were trypsinized and counted using a hemacytometer on days 3 or 6.

Statistical analysis

Means ± SD were calculated, and statistical analysis was performed using one-way ANOVA, followed by the nonparametric Kruskal-Wallis test; P < 0.05 was considered statistically significant.

RESULTS

HGF upregulation of COX-2 mRNA in HBEpC is Akt-dependent

HGF is a growth factor for cells of epithelial origin; when combined with low levels of serum, HGF induces growth in cultured normal and immortalized human bronchial epithelial cells (31). We selected untransformed, non-immortalized HBEpC as a model for signal transduction in lung epithelial cells. Treatment of HBEpC with HGF resulted in activation of the c-Met receptor within 5 min, as determined by Western blot analysis of phosphorylation of the kinase domain (P-Tyr 1234/1235) (Fig. 1A).

Fig. 1.

Fig. 1

Open in a new tab

Role of Akt in hepatocyte growth factor (HGF)-induced gene expression in human bronchial epithelial cells (HBEpC). HBEpC were grown to 80% confluence and placed in serum-free medium for 16 h. Cells were treated with 25 ng/ml HGF for the indicated times for each experiment. A: cell lysates were prepared, and equal amounts of protein were used for Western blots for phosphorylated c-Met protein; blots were stripped and probed for total c-Met protein. Experiments were repeated at least 3 times; representative results are shown. Band densitometry at left shows the means ± SD, *P ≤ 0.05. B: equal amounts of mRNA from each time point were used for semiquantitative RT-PCR of COX-2 (bottom band), with a competing internal standard (top band). Densitometry shows average of several experiments of COX-2 mRNA induced by HGF treatment, normalized to the internal standard (IS), *P ≤ 0.05. Experiments were repeated at least 3 times; representative results are shown. C: cell lysates were prepared, and equal amounts of protein from each time point were used in Western blot analysis for phosphorylated Akt; the blot was stripped and probed for total Akt protein as a control. Experiments were repeated at least 3 times; representative results are shown. Densitometry at left shows means ± SD, *P ≤ 0.05. D and E: HBEpC were grown to 80% confluence and infected with either control adenovirus (AdCont) or adenovirus expressing DN Akt (AdDnAkt). Following infection, cells were placed in serum-free medium for 16 h before treatment with 25 ng/ml HGF for the indicated times. D: cell lysates were made, and Western blots were performed on equal amounts of protein to detect Akt levels. E: equal amounts of mRNA from each time point were used for semiquantitative RT-PCR of COX-2, with a competing IS. Densitometry was performed as for B. Experiments were repeated at least 3 times; representative results are shown.

The GEArray Q Series microarray was used to broadly screen the regulation of signal transduction pathways involved in proliferation, antiapoptotic activity, or other pathways potentially involved in HGF tissue repair. Array analysis showed that HGF increased the transcription of Bcl-2, Egr-1, c-fos, ornithine decarboxylase, and COX-2 (data not shown). Several other laboratories have shown that HGF-induced proliferation and inhibition of apoptosis requires COX-2 activity (5, 29, 40). We therefore focused our further studies on HGF-induced COX-2 regulation. The COX-2 message was increased approximately four- to sixfold by HGF at 4 h, as determined by quantitative RT-PCR, relative to an internal standard (Fig. 1B).

A recent study suggested a strong association between HGF amelioration of postinfarction remodeling of the heart and activation of the Akt pathway (34). We found that HGF induced phosphorylation of Akt within 3 min, sustained up to 60 min in HBEpC, as determined by Western blotting (Fig. 1C). We investigated gene regulation by HGF when the Akt pathway is blocked. HBEpC were infected with either control adenovirus or adenovirus encoding DN Akt. Expression of DN Akt adenovirus resulted in a >10-fold increase in the level of total Akt protein in the cells (Fig. 1D). HGF treatment of the cells did not affect the level of Akt expression in either infection (Fig. 1D). Overexpression of DN Akt blocked or reduced HGF-induced increase in Bcl-2, Egr-1, c-fos, ornithine decarboxylase, and COX-2 expression detected by the microarray (data not shown). Inhibition of the HGF-induced COX-2 expression was further confirmed using quantitative RT-PCR for the same time points (Fig. 1E). These results indicate that HGF regulation of COX-2 mRNA expression requires the Akt pathway.

HGF regulation of the COX-2 promoter requires TBE cis element and β-catenin activation

Regulatory cis elements in the human COX-2 promoter that control its gene expression have been analyzed by several laboratories, and a number of factors have been shown to regulate COX-2 expression (1, 2) (Fig. 2A). To identify transcriptional factors required for HGF regulation, we utilized two luciferase-COX-2 promoter constructs, one encoding only the core promoter, from −327 bp to +59 bp (with the transcriptional start site at +1), and the other encoding −1432 bp to +59 bp. Results showed that HGF increased luciferase expression approximately fourfold from the −1432 bp promoter at 17 h; this increase was detectable within 6 h (data not shown). However, the truncated promoter, with only −327 bp upstream of the transcriptional start site, was not induced by treatment with HGF at 17 h (Fig. 2B), or at shorter time points (data not shown). This result indicates that the necessary positive cis regulatory factors for HGF activation of the COX-2 promoter are in distal locations of the promoter.

Fig. 2.

Fig. 2

Open in a new tab

HGF regulation of the COX-2 promoter and the role of the TBE element. A: schematics of the COX-2 promoter: top, with predicted regulatory elements indicated (1, 2, 22); bottom, with the segments of the promoter used for luciferase reporter constructs indicated. B: HBEpC were grown to 80% confluence and transiently transfected with a COX-2-luciferase reporter construct: WT promoter −1432 bp to +59 bp, truncated WT promoter −327 bp to +59 bp, −1432 bp to +59 bp with either a deletion mutation in the TBE element, or −1432 bp to +59 bp with a point mutation in the TBE element. Six hours following transfection, cells were placed in serum-free medium for 1 h before treatment with 25 ng/ml HGF for 17 h. Luciferase assays were performed on equal amounts of protein. Firefly luciferase activity was normalized to Renilla activity to correct for transfection differences between samples. Results show means of 3 replicates ± SD; *P < 0.05. Experiments were performed at least 3 times, and representative results are shown.

The upstream human COX-2 promoter includes a TBE element, an AP-1, and several AP-1-like elements, an NF-κB element, and several STAT1 and HIF binding sites (Fig. 2A) (1, 2). Some of these elements have been hypothesized to regulate COX-2 expression in response to hypoxia and proinflammatory cytokines (2), but only the TBE element has been directly proven to be required for gene regulation, in this case by Wnt (1). An extensive study of the COX-2 TBE element using deletion and point mutations in luciferase reporter assays and band shift assays demonstrated the binding of β-catenin and the requirement of β-catenin activation for its function (1). We found that HGF failed to increase expression from a full-length promoter when the TBE element was deleted (Fig. 2B) or when the TBE element contained a single point mutation (Fig. 2B).

Wnt activation of the COX-2 promoter via the TBE element requires the nuclear accumulation of β-catenin protein (1). Previous work in trophoblasts showed that the pathway for HGF activation of LEF1/Tcf-4 involves the phosphorylation and inactivation of GSK-3β downstream of the PI3K/Akt pathway. GSK-3β inhibition leads to accumulation of β-catenin in the cytoplasm and subsequent transport to the nucleus where it interacts with LEF1/Tcf-4, forming an active transcription factor complex. In HBEpC, HGF induced sustained phosphorylation of GSK-3β, detectable within 3 min (Fig. 3A). Increased concentrations of β-catenin protein could be detected in the nucleus within 30 min and was sustained for at least 1 h (Fig. 3B). This nuclear translocation occurred with a similar time course of activation of SP3 in the nucleus, which our laboratory has previously demonstrated occurs downstream of p42/p44 MAPK activation by HGF (7). β-Actin was monitored as a control for protein loading.

Fig. 3.

Fig. 3

Open in a new tab

Role of β-catenin for COX-2 promoter activation by HGF. A and B: HBEpC were grown to 80% confluence and placed in serum-free medium overnight. Cells were treated with 25 ng/ml HGF for the indicated time points. A: cell lysates were prepared, and equal amounts of protein were subjected to SDS-PAGE. Western blots were performed for phosphorylated GSK-3β protein; the blot was stripped and probed for total GSK-3β protein. Densitometry, below, shows means ± SD, *P ≤ 0.05. B: nuclear extracts were prepared, and equal amounts of protein were subjected to SDS-PAGE followed by Western blotting for β-catenin; the blot was stripped and probed for SP3 and stripped again and probed for β-actin. Densitometry, below, shows means ± SD, *P ≤ 0.05. C: HBEpC were grown to 80% confluence and transfected with the COX-2-luciferase reporter construct from −1432 bp to +59 bp cotransfected with either a plasmid expressing the mutant Y654/670 β-catenin (DN β-cat) or the p3XFlagCMV control (empty vector). Lysates were made 17 h following treatment with HGF (25 ng/ml). Luciferase assays were performed on equal amounts of protein, and firefly luciferase activity was normalized to Renilla activity to correct for transfection variations between samples. Results show means of 3 replicates ± SD; *P < 0.05.

A recent report by Zeng et al. (39) identified specific β-catenin mutations that blocked the activation of β-catenin downstream of the c-Met receptor. The authors showed that the double mutant Y654F/Y670F completely blocked HGF-induced nuclear translocation of a coexpressed FLAG-β-catenin chimera and prevented subsequent complex formation with TCF-4 (39). We utilized the DN Y654F/Y670F β-catenin to determine whether this mutant could block HGF-induced expression of the wt COX-2 promoter. Coexpression of DN Y654F/Y670F β-catenin reduced basal levels of expression of the wt −1432 bp to +59 bp COX-2 promoter and completely blocked increased expression by HGF (Fig. 3C). As a control, we show that cotransfection of an empty expression vector had no effect on HGF-induced activation of the wt COX-2 promoter. Together, our data suggest that HGF regulates the COX-2 promoter by activation of the TBE element through the activation of β-catenin.

Role of PI3K in HGF-induced β-catenin activation

HGF activation of the PI3K pathway has been demonstrated to be required for Akt activation. Since we found that the TBE element was required for HGF activation of the COX-2 promoter, we wished to confirm that this activation occurred downstream of the PI3K/Akt pathway. HGF upregulation of the wt COX-2 promoter was abrogated by the pretreatment of cells with the PI3K inhibitor LY-294002 (Fig. 4A, left). Control experiments show that LY-294002 alone does not significantly alter basal activity of the COX-2 wt promoter (Fig. 4A, right). The minimal β-catenin-responsive promoter TOPFLASH was activated approximately twofold by HGF within 6 h; this activation was sustained for at least 17 h, in agreement with our findings that the COX-2 promoter also had sustained activation by HGF (compare Fig. 4B and 3B). Inhibition of the PI3K/Akt pathway by LY-294002 also blocked HGF-induced activation of the minimal β-catenin responsive promoter (Fig. 4B, left). Control experiments show that LY-294002 alone does not significantly alter basal activity of the TOPFLASH promoter (Fig. 4B, right).

Fig. 4.

Fig. 4

Open in a new tab

Role of phosphatidylinositol 3-kinase (PI3K) for HGF-induced β-catenin-responsive promoter activity. HBEpC were grown to 80% confluence and transfected with the either the COX-2-luciferase reporter construct from −1432 bp to +59 bp (A) or with the TOPFLASH β-catenin-responsive reporter (B). Left: 6 h following transfection, cells were placed in serum-free medium for 1 h before treatment with 25 ng/ml HGF for the indicated times ± 20 min pretreatment with 10 μM LY-294002 (LY). Right: 6 h after transfections, cells were placed in serum-free medium for 1 h before treatment with 10 μM LY-294002 alone. Luciferase assays were performed on equal amounts of protein. Firefly luciferase activity was normalized to Renilla activity to correct for transfection differences between samples. Results show means of 3 replicates ± SD; *P < 0.05.

Role of p42/p44 MAPK in HGF activation of β-catenin

We and others have shown that some cross talk exists between p42/p44 MAPK and PI3K activation by HGF (6). A recent report in MDCK cells showed that HGF-induced expression of COX-2 mRNA and subsequent production of PGE2 required the p42/p44 MAPK pathway (41). Thus we investigated whether HGF activation of p42/p44 MAPK was also required in HBEpC for activation of the β-catenin-responsive promoter. Downstream signaling by HGF in HBEpC caused the sustained phosphorylation of p42/p44 MAPK (Fig. 5A), in agreement with findings by Takami et al. (31).

Fig. 5.

Fig. 5

Open in a new tab

Role of MAPK for HGF-induced β-catenin-responsive promoter activity. A and B: HBEpC were grown to 80% confluence and placed in serum-free medium overnight. Cells were treated with 25 ng/ml HGF for the indicated time points, and cell lysates were prepared; where indicated, cells were treated with 20 μM U0126 or 20 μM LY-294002 for 20 min before addition of HGF. Equal amounts of protein were subjected to SDS-PAGE, and Western blots were performed for phosphorylated p42/p44 MAPK (B) or phosphorylated Akt (C); the blots were stripped and reprobed for the respective total protein. Densitometry for MAPK shows means ± SD, *P ≤ 0.05. C: HBEpC were grown to 80% confluence and transfected with the TOPFLASH β-catenin-responsive reporter. Six hours following transfection, cells were placed in serum-free medium for 1 h before treatment with 25 ng/ml HGF for 6 h ± pretreatment with U0126 (20 μM, left); control experiments used transfected cells treated with U0126 alone (right). Luciferase assays were performed on equal amounts of protein, and firefly luciferase activity was normalized to Renilla activity to correct for transfection differences between samples. Results show means of 3 replicates ± SD; *P < 0.05.

Activation of Akt was investigated in the presence of either the PI3K inhibitor LY-294002 or the MEK1/2 inhibitor U0126. We found that the MEK1/2 inhibitor does not inhibit HGF activation of Akt, whereas the PI3K inhibitor completely blocks HGF-induced phosphorylation of Akt (Fig. 5B). Inhibition of the p42/p44 MAPK pathway by U0126 completely abolished HGF-induced activation of the β-catenin responsive TOPFLASH reporter (Fig. 5C, left). Control experiments show that U0126 alone reduced basal activity of the TOPFLASH promoter by ~20% (Fig. 5C, right). We also found that although LY-294002 eliminated HGF-induced phosphorylation of GSK-3β, U0126 had no effect on its activation by HGF (data not shown). Together, our results show that, in addition to the PI3K/Akt pathway, the MEK/ERK pathway is required for HGF-induced β-catenin activation. The mechanism by which p42/p44 is involved in activation of β-catenin does not involve cross talk between the MEK/MAPK pathway and the PI3K/Akt pathway.

COX-2 activity is required for HGF-induced growth in HBEpC

Previous work has shown that COX-2 activity is required for HGF-induced proliferation and inhibition of apoptosis in a variety of cancer cell types (5, 40). We examined the requirement of COX-2 activity for HGF-induced growth in the primary HBEpC. Treatment of cells with the COX-2 inhibitor NS-398 blocked HGF-induced growth at both 3 days and 6 days as determined by cell counting (Fig. 6). However, the growth of HBEpC induced by the defined medium was not inhibited by the presence of the COX-2 inhibitor, suggesting that the cell growth inhibition by NS-398 was specific for HGF-induced growth.

Fig. 6.

Fig. 6

Open in a new tab

COX-2 activity is required for HGF-induced proliferation in HBEpC. HBEpC were split on day 1 to 5 × 104 cells/35-mm dish in defined medium. HGF was added (25 ng/ml) to the medium on days 1, 2, and 4. The NS-398 COX-2 inhibitor (10 μM) was added to cells before the HGF. Cells were counted on days 3 or 6. Experiments were done in triplicate; data show means ± SD; *P < 0.05. Representative data are shown.

DISCUSSION

The present study demonstrated that HGF stimulates COX-2 gene expression in a mechanism that requires both the PI3K/Akt and p42/p44 MAPK pathways in HBEpC. Downstream of these two pathways, HGF regulation of the COX-2 promoter requires β-catenin activation and binding to the TBE element. The activation of β-catenin transcriptional activity has been demonstrated to occur through the concerted activation of several signaling pathways (3). The activation of PI3K/Akt pathway and subsequent inactivation of GSK-3β lead to increases in cytosolic β-catenin protein levels; this is normally followed by β-catenin NH2-terminal serine and threonine phosphorylation and transport into the nucleus where it interacts with other transcription factors to regulate gene transcription (3). The mechanism by which p42/p44 MAPK regulates β-catenin activation downstream of HGF is not known; Ding et al. (9) showed that β-catenin activation by the hepatitis B virus protein X requires p42/p44 MAPK to dock with and phosphorylate GSK-3β at Thr43 (9). A report by Zhang et al. (41) showed the requirement of the MAPK pathway for HGF regulation of COX-2. Here we show that HGF-induced β-catenin activation requires both PI3K/Akt and MAPK and that these pathways do not converge at the level of Akt activation. The mechanism by which MAPK participates in β-catenin activation downstream of HGF is currently under investigation.

Besides its role in cell growth, COX-2 has been implicated in abnormal fibrotic repair. Studies have shown that patients with pulmonary fibrosis have reduced expression of COX-2 in cells cultured from their lungs and reduced PGE2 (36). Reduced PGE2 due to COX-2 enzyme deficiency sensitizes mice to bleomycin-induced pulmonary fibrosis, and animals with pulmonary fibrosis have lower PGE2 (15, 25). In other experiments, animals were protected from bleomycin-induced pulmonary fibrosis by either direct upregulation of COX-2 or by treatment with drugs that increased COX-2 (23, 24). The antifibrotic effect of COX-2 was shown to be dependent on PGE2 production, a downstream product of COX-2 (12). In agreement with this, in cell culture, the negative regulation of fibroblast proliferation by lung epithelial cells was shown to be dependent on COX-2 expression and the production of PGE2 (20, 41). In a more direct experiment, treatment of primary lung fibroblasts and MDCK cells with HGF prevented TGF-β1-induced myofibroblast transdifferentiation and attenuated TGF-β1-induced collagen I production; this activity was found to be dependent on COX-2 activity and production of PGE2 (17, 41).

The function of HGF in normal tissue repair has received increasing attention, not just for its actions in the lung, but also in the liver, kidney, and heart (10, 14, 16, 19, 33). Our current findings from the gene array showed that HGF regulated several other genes besides COX-2 in an Akt-dependent manner, including c-fos, Egr-1, and ornithine decarboxylase. The HGF regulation of c-fos via the PI3K/Akt pathway has been studied previously (32). HGF has also been shown to regulate the activity of ornithine decarboxylase and its downstream product, intracellular polyamines, through PI3K, protein kinase C, and pp60 c-src (8). HGF regulation of Egr-1 has been shown to occur via protein kinase C and ERK MAPK, but the role of Akt is unknown (37). The activation of these genes, all with roles in cell growth as well as inhibition of apoptosis, together with COX-2, likely makes up an important combined and complementary regulation of cellular activities.

Here, we found that inhibition of COX-2 enzyme activity blocked HGF-induced cell growth but not growth induced by the defined medium for these cells. This suggests that production of prostaglandins is also important for HGF-induced growth in primary epithelial cells. The requirement of COX-2 activity for HGF-induced growth of lung epithelial cells may have implications for the use of corticosteroids following lung injury, as these drugs may have the unintended side effects of antagonizing the repair function of the endogenous repair factor HGF.

Acknowledgments

We thank Dr. D. P. Bottaro (National Institutes of Health, Bethesda, MD) for discussions related to this work. The opinions expressed in this document are those of the authors and do not reflect the views of the Uniformed Services University of the Health Sciences, the Department of Defense, or the U. S. Federal Government.

GRANTS

This work was supported in part by National Institutes of Health (NIH) Grant R01-HL-073929 and by Uniformed Services University of the Health Sciences CO75LE (to R. M. Day) and by NIH Grants R01-HL-067340 and R01-HL-072844 (to Y. J. Suzuki).

References

  • 1.Araki Y, Okamura S, Hussain SP, Nagashima M, He P, Shiseki M, Miura K, Harris CC. Regulation of cyclooxygenase-2 expression by the Wnt and ras pathways. Cancer Res. 2003;63:728–734. [PubMed] [Google Scholar]
  • 2.Bazan NG, Lukiw WJ. Cyclooxygenase-2 and presenilin-1 gene expression induced by interleukin-1beta and amyloid beta 42 peptide is potentiated by hypoxia in primary human neural cells. J Biol Chem. 2002;277:30359–30367. doi: 10.1074/jbc.M203201200. [DOI] [PubMed] [Google Scholar]
  • 3.Brembeck FH, Rosario M, Birchmeier W. Balancing cell adhesion and Wnt signaling, the key role of beta-catenin. Curr Opin Gen Dev. 2006;16:51–59. doi: 10.1016/j.gde.2005.12.007. [DOI] [PubMed] [Google Scholar]
  • 4.Brunet A, Bonni A, Zigmond MJ, Lin MZ, Juo P, Hu LS, Anderson MJ, Arden KC, Blenis J, Greenberg ME. Akt promotes cell survival by phosphorylating and inhibiting a Forkhead transcription factor. Cell. 1999;96:857–868. doi: 10.1016/s0092-8674(00)80595-4. [DOI] [PubMed] [Google Scholar]
  • 5.Chen JH, Wu CW, Kao HL, Chang HM, Li AF, Liu TY, Chi CW. Effects of COX-2 inhibitor on growth of human gastric cancer cells and its relation to hepatocyte growth factor. Cancer Lett. 2006;239:263–270. doi: 10.1016/j.canlet.2005.08.026. [DOI] [PubMed] [Google Scholar]
  • 6.Day RM, Cioce V, Breckenridge D, Castagnino P, Bottaro DP. Differential signaling by alternative HGF isoforms through c-Met: activation of both MAP kinase and PI 3-kinase pathways is insufficient for mitogenesis. Oncogene. 1999;18:3399–3406. doi: 10.1038/sj.onc.1202683. [DOI] [PubMed] [Google Scholar]
  • 7.Day RM, Thiel G, Lum J, Chevere RD, Yang Y, Stevens J, Sibert L, Fanburg BL. Hepatocyte growth factor regulates angiotensin converting enzyme expression. J Biol Chem. 2004;279:8792–8801. doi: 10.1074/jbc.M311140200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Desiderio MA, Pogliaghi G, Dansi P. Hepatocyte growth factor-induced expression of ornithine decarboxylase, c-met, and c-myc is differently affected by protein kinase inhibitors in human hepatoma cells HepG2. Exp Cell Res. 1998;242:401–409. doi: 10.1006/excr.1998.4073. [DOI] [PubMed] [Google Scholar]
  • 9.Ding Q, Xia W, Liu JC, Yang JY, Lee DF, Xia J, Bartholomeusz G, Li Y, Pan Y, Li Z, Bargou RC, Qin J, Lai CC, Tsai FJ, Tsai CH, Hung MC. Erk associates with and primes GSK-3beta for its inactivation resulting in upregulation of beta-catenin. Mol Cell. 2005;19:159–170. doi: 10.1016/j.molcel.2005.06.009. [DOI] [PubMed] [Google Scholar]
  • 10.Dohi M, Hasegawa T, Yamamoto K, Marshall BC. Hepatocyte growth factor attenuates collagen accumulation in a murine model of pulmonary fibrosis. Am J Respir Crit Care Med. 2000;162:2302–2307. doi: 10.1164/ajrccm.162.6.9908097. [DOI] [PubMed] [Google Scholar]
  • 11.Fan S, Ma Y, Wang JA, Yuan RQ, Meng Q, Cao Y, Laterra J, Goldberg I, Rosen E. The cytokine hepatocyte growth factor/scatter factor inhibits apoptosis and enhances DNA repair by a common mechanism involving signaling through phosphatidyl inositol 3′ kinase. Oncogene. 2000;19:2212–2223. doi: 10.1038/sj.onc.1203566. [DOI] [PubMed] [Google Scholar]
  • 12.Hodges RJ, Jenkins RG, Wheeler-Jones CP, Copeman DM, Bottoms SE, Bellingan GJ, Nanthakumar CB, Laurent GJ, Hart SL, Foster ML, McAnulty RJ. Severity of lung injury in cyclooxygenase-2-deficient mice is dependent on reduced prostaglandin E(2) production. Am J Pathol. 2004;165:1663–1676. doi: 10.1016/S0002-9440(10)63423-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Inoue H, Yokoyama C, Hara S, Tone Y, Tanabe T. Transcriptional regulation of human prostaglandin-endoperoxide synthase-2 gene by lipo-polysaccharide and phorbol ester in vascular endothelial cells. Involvement of both nuclear factor for interleukin-6 expression site and cAMP response element. J Biol Chem. 1995;270:24965–24971. doi: 10.1074/jbc.270.42.24965. [DOI] [PubMed] [Google Scholar]
  • 14.Kawaida K, Matsumoto K, Shimazu H, Nakamura T. Hepatocyte growth factor prevents acute renal failure and accelerates renal regeneration in mice. Proc Natl Acad Sci USA. 1994;91:4357–4361. doi: 10.1073/pnas.91.10.4357. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Keerthisingam CB, Jenkins RG, Harrison NK, Hernandez-Rodriguez NA, Booth H, Laurent GJ, Hart SL, Foster ML, McAnulty RJ. Cyclooxygenase-2 deficiency results in a loss of the anti-proliferative response to transforming growth factor-beta in human fibrotic lung fibro-blasts and promotes bleomycin-induced pulmonary fibrosis in mice. Am J Pathol. 2001;158:1411–1422. doi: 10.1016/s0002-9440(10)64092-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Kitta K, Day RM, Ikeda T, Suzuki YJ. Hepatocyte growth factor protects cardiac myocytes against oxidative stress-induced apoptosis. Free Radic Biol Med. 2001;31:902–910. doi: 10.1016/s0891-5849(01)00663-3. [DOI] [PubMed] [Google Scholar]
  • 17.Kolodsick JE, Peters-Golden M, Larios J, Toews GB, Thannickal VJ, Moore BB. Prostaglandin E2 inhibits fibroblast to myofibroblast transition via E. prostanoid receptor 2 signaling and cyclic adenosine monophosphate elevation. Am J Respir Cell Mol Biol. 2003;29:537–544. doi: 10.1165/rcmb.2002-0243OC. [DOI] [PubMed] [Google Scholar]
  • 18.Kosai K, Matsumoto K, Funakoshi H, Nakamura T. Hepatocyte growth factor prevents endotoxin-induced lethal hepatic failure in mice. Hepatology. 1999;30:151–159. doi: 10.1002/hep.510300102. [DOI] [PubMed] [Google Scholar]
  • 19.Kosai K, Matsumoto K, Nagata S, Tsujimoto Y, Nakamura T. Abrogation of Fas-induced fulminant hepatic failure in mice by hepatocyte growth factor. Biochem Biophys Res Commun. 1998;244:683–690. doi: 10.1006/bbrc.1998.8293. [DOI] [PubMed] [Google Scholar]
  • 20.Lama V, Moore BB, Christensen P, Toews GB, Peters-Golden M. Prostaglandin E2 synthesis and suppression of fibroblast proliferation by alveolar epithelial cells is cyclooxygenase-2-dependent. Am J Respir Cell Mol Biol. 2002;27:752–758. doi: 10.1165/rcmb.4857. [DOI] [PubMed] [Google Scholar]
  • 21.Liu XL, Sato S, Dai W, Yamanaka N. The protective effect of hepatocyte growth-promoting factor (pHGF) against hydrogen peroxide-induced acute lung injury in rats. Med Electron Microsc. 2001;34:92–102. doi: 10.1007/s007950170003. [DOI] [PubMed] [Google Scholar]
  • 22.Liu Y, Borchert GL, Phang JM. Polyoma enhancer activator 3, an ets transcription factor, mediates the induction of cyclooxygenase-2 by nitric oxide in colorectal cancer cells. J Biol Chem. 2004;279:18694–18700. doi: 10.1074/jbc.M308136200. [DOI] [PubMed] [Google Scholar]
  • 23.Lovgren AK, Jania LA, Hartney JM, Parsons KK, Audoly LP, Fitzgerald GA, Tilley SL, Koller BH. COX-2-derived prostacyclin protects against bleomycin-induced pulmonary fibrosis. Am J Physiol Lung Cell Mol Physiol. 2006;291:L144–L156. doi: 10.1152/ajplung.00492.2005. [DOI] [PubMed] [Google Scholar]
  • 24.Molina-Molina M, Serrano-Mollar A, Bulbena O, Fernandez-Zabale-gui L, Closa D, Marin-Arguedas A, Torrego A, Mullol J, Picado C, Xaubet A. Losartan attenuates bleomycin induced lung fibrosis by increasing prostaglandin E2 synthesis. Thorax. 2006;61:604–610. doi: 10.1136/thx.2005.051946. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Ogushi F, Endo T, Tani K, Asada K, Kawano T, Tada H, Maniwa K, Sone S. Decreased prostaglandin E2 synthesis by lung fibroblasts isolated from rats with bleomycin-induced lung fibrosis. Int J Exp Pathol. 1999;80:41–49. doi: 10.1046/j.1365-2613.1999.00096.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Rubin JS, Bottaro DP, Aaronson SA. Hepatocyte growth factor/scatter factor and its receptor, the c-met proto-oncogene product. Biochim Biophys Acta. 1993;1155:357–371. doi: 10.1016/0304-419x(93)90015-5. [DOI] [PubMed] [Google Scholar]
  • 28.Selman M, Thannickal VJ, Pardo A, Zisman DA, Martinez FJ, Lynch JP., 3rd Idiopathic pulmonary fibrosis: pathogenesis and therapeutic approaches. Drugs. 2004;64:405–430. doi: 10.2165/00003495-200464040-00005. [DOI] [PubMed] [Google Scholar]
  • 29.Siegfried JM, Gubish CT, Rothstein ME, de Oliveira PE, Stabile LP. Signaling pathways involved in cyclooxygenase-2 induction by hepatocyte growth factor in non small-cell lung cancer. Mol Pharmacol. 2007;72:769–779. doi: 10.1124/mol.107.034215. [DOI] [PubMed] [Google Scholar]
  • 30.Singh-Kaw P, Zarnegar R, Siegfried JM. Stimulatory effects of hepatocyte growth factor on normal and neoplastic human bronchial epithelial cells. Am J Physiol Lung Cell Mol Physiol. 1995;268:L1012–L1020. doi: 10.1152/ajplung.1995.268.6.L1012. [DOI] [PubMed] [Google Scholar]
  • 31.Takami K, Takuwa N, Okazaki H, Kobayashi M, Ohtoshi T, Kawasaki S, Dohi M, Yamamoto K, Nakamura T, Tanaka M, Nakahara K, Takuwa Y, Takizawa H. Interferon-gamma inhibits hepatocyte growth factor-stimulated cell proliferation of human bronchial epithelial cells: upregulation of p27(kip1) cyclin-dependent kinase inhibitor. Am J Respir Cell Mol Biol. 2002;26:231–238. doi: 10.1165/ajrcmb.26.2.4643. [DOI] [PubMed] [Google Scholar]
  • 32.Takeuchi K, Shibamoto S, Nagamine K, Shigemori I, Omura S, Kitamura N, Ito F. Signaling pathways leading to transcription and translation cooperatively regulate the transient increase in expression of c-Fos protein. J Biol Chem. 2001;276:26077–26083. doi: 10.1074/jbc.M102704200. [DOI] [PubMed] [Google Scholar]
  • 33.Taniyama Y, Morishita R, Aoki M, Hiraoka K, Yamasaki K, Hashiya N, Matsumoto K, Nakamura T, Kaneda Y, Ogihara T. Angiogenesis and antifibrotic action by hepatocyte growth factor in cardiomyopathy. Hypertension. 2002;40:47–53. doi: 10.1161/01.hyp.0000020755.56955.bf. [DOI] [PubMed] [Google Scholar]
  • 34.Wang Y, Ahmad N, Wani MA, Ashraf M. Hepatocyte growth factor prevents ventricular remodeling and dysfunction in mice via Akt pathway and angiogenesis. J Mol Cell Cardiol. 2004;37:1041–1052. doi: 10.1016/j.yjmcc.2004.09.004. [DOI] [PubMed] [Google Scholar]
  • 35.Watanabe M, Ebina M, Orson FM, Nakamura A, Kubota K, Koinuma D, Akiyama K, Maemondo M, Okouchi S, Tahara M, Matsumoto K, Nakamura T, Nukiwa T. Hepatocyte growth factor gene transfer to alveolar septa for effective suppression of lung fibrosis. Mol Ther. 2005;12:58–67. doi: 10.1016/j.ymthe.2005.02.019. [DOI] [PubMed] [Google Scholar]
  • 36.Wilborn J, Crofford LJ, Burdick MD, Kunkel SL, Strieter RM, Peters-Golden M. Cultured lung fibroblasts isolated from patients with idiopathic pulmonary fibrosis have a diminished capacity to synthesize prostaglandin E2 and to express cyclooxygenase-2. J Clin Invest. 1995;95:1861–1868. doi: 10.1172/JCI117866. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Worden B, Yang XP, Lee TL, Bagain L, Yeh NT, Cohen JG, Van Waes C, Chen Z. Hepatocyte growth factor/scatter factor differentially regulates expression of proangiogenic factors through Egr-1 in head and neck squamous cell carcinoma. Cancer Res. 2005;65:7071–7080. doi: 10.1158/0008-5472.CAN-04-0989. [DOI] [PubMed] [Google Scholar]
  • 38.Yaekashiwa M, Nakayama S, Ohnuma K, Sakai T, Abe T, Satoh K, Matsumoto K, Nakamura T, Takahashi T, Nukiwa T. Simultaneous or delayed administration of hepatocyte growth factor equally represses the fibrotic changes in murine lung injury induced by bleomycin. A morphologic study. Am J Respir Crit Care Med. 1997;156:1937–1944. doi: 10.1164/ajrccm.156.6.9611057. [DOI] [PubMed] [Google Scholar]
  • 39.Zeng G, Apte U, Micsenyi A, Bell A, Monga SP. Tyrosine residues 654 and 670 in beta-catenin are crucial in regulation of Met-beta-catenin interactions. Exp Cell Res. 2006;312:3620–3630. doi: 10.1016/j.yexcr.2006.08.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Zeng Q, McCauley LK, Wang CY. Hepatocyte growth factor inhibits anoikis by induction of activator protein 1-dependent cyclooxygenase-2. Implication in head and neck squamous cell carcinoma progression. J Biol Chem. 2002;277:50137–50142. doi: 10.1074/jbc.M208952200. [DOI] [PubMed] [Google Scholar]
  • 41.Zhang A, Wang MH, Dong Z, Yang T. Prostaglandin E2 is a potent inhibitor of epithelial-to-mesenchymal transition: interaction with hepatocyte growth factor. Am J Physiol Renal Physiol. 2006;291:F1323–F1331. doi: 10.1152/ajprenal.00480.2005. [DOI] [PubMed] [Google Scholar]

RESOURCES