Among the four protease-activated receptors (PARs), PAR-1 plays an important role in normal lung functioning and in the development of lung diseases, including fibrosis. We compared the expression and functional activity of PARs in normal and fibrotic human lung fibroblasts. Both normal and fibrotic cells express PAR-1, -2, and -3, with PAR-2 showing the lowest level. There was no significant difference between normal and fibrotic fibroblasts in expression levels of PAR-1 and PAR-3, whereas a fourfold higher expression level of PAR-2 was observed in fibrotic cells compared with normal cells. Ca2+ imaging studies revealed apparently only PAR-1-induced Ca2+ signaling in lung fibroblasts. PAR-1 agonists, thrombin and synthetic activating peptide, induced concentration-dependent Ca2+ mobilization with EC50 values of 5 nM and 1 μM, respectively. The neutrophil protease cathepsin G produced a transient Ca2+ response followed by disabling PAR-1, whereas elastase did not affect Ca2+ level. PAR-1 activation by thrombin or receptor-activating peptide downregulated expression of all three PARs in lung fibroblasts, with maximal effect at 3–6 h, whereas expression returned toward basal level after 24 h. Furthermore, PAR-1 agonists dose dependently increased PGE2 secretion from lung fibroblasts and induction of cyclooxygenase-2 expression. We then found that PGE2 downregulated expression of all three PARs. The effect of PGE2 was continuously growing with time. Furthermore, PGE2 exerts its effect through the EP2 receptor that was confirmed using the selective EP2 agonist butaprost. This novel autocrine feedback mechanism of PGE2 in lung fibroblasts seems to be an important regulator in lung physiology and pathology.
- cyclic adenosine 5′-monophosphate
fibroblasts are the most abundant cells of the lung connective tissue participating in tissue repair by proliferation and secretion of matrix proteins. Any shift in the subtle regulation of lung fibroblast functions leads to a predominance of profibrotic conditions, and, as a consequence, to development of lung fibrotic disorders. Although there are various initiating factors or causes, the terminal stages are all characterized by profound fibroblast proliferation and progressive accumulation of connective tissue replacing normal functional parenchyma. This leads to dramatic changes in the lung architecture and progressive respiratory insufficiency. The pathogenesis of pulmonary fibrosis includes endothelial and epithelial cell injury, accumulation of inflammatory cells, production of inflammatory mediators, and activation of fibroblasts, many of which possess the phenotypic characteristics of myofibroblasts (35, 38, 46).
Some of the mediators participating in the development of fibrosis have been identified. Profibrotic factors include transforming growth factor-β (TGF-β), connective tissue growth factor (CTGF), platelet-derived growth factor (PDGF), oncostatin M, monocyte chemotactic factor-1, and pulmonary and activation-regulated chemokine, whereas interferon-γ, prostaglandin E2 (PGE2), epidermal growth factor, and interleukin (IL)-1β are known to be antifibrotic mediators (reviewed in Refs. 4, 15, and 44).
Among the regulators of fibroblast functions, PGE2, which is a metabolite of arachidonic acid derived via the cyclooxygenase (COX) pathway, was shown to decrease lung fibroblast proliferation, to reduce collagen levels by inhibition of its mRNA synthesis, to inhibit fibroblast transition into the myofibroblast phenotype, and to inhibit fibroblast chemotaxis (30, 33, 34, 40). Being a major eicosanoid product of lung fibroblasts (19, 22), PGE2 is thus an autocrine factor that controls cellular overactivation. The PGE2 synthesis system is significantly suppressed in fibrotic lungs (6) and in fibroblasts isolated from fibrotic lung tissue (51, 55) and therefore is not able to exert its antifibrotic functions.
In addition to the profibrotic factors mentioned above, thrombin may also contribute to the development of lung fibrotic disorders (28, 29). This multifunctional serine protease plays a critical role in thrombosis and hemostasis and exerts numerous proinflammatory cellular effects, in particular, through the activation of protease-activated receptors (PARs) (1, 16, 24). Thrombin was shown to induce lung fibroblast proliferation, to stimulate synthesis of procollagen and the profibrotic agent CTGF, and to mediate differentiation of fibroblasts into myofibroblasts (5, 10, 12, 28). All these effects of thrombin were demonstrated to be resulting from activation of PAR-1 in the fibroblasts.
PAR-1 belongs to a family of G protein-coupled receptors that are activated by specific proteases with the cleavage of the receptor's NH2-terminal extracellular domain to expose a new NH2 terminus (reviewed in Refs. 42 and 54). The newly generated NH2 terminus acts as a tethered ligand and binds to the second extracellular loop of the receptor to trigger subsequent cell signaling events. Four PARs have been identified: PAR-1, -2, -3, and -4. PAR-1, PAR-3, and PAR-4 are preferentially activated by thrombin and PAR-2 by trypsin and mast cell tryptase. All four PARs have been found in airways (32, 37). Nevertheless, there has been no systematic characterization of PARs in lung fibroblasts.
Some studies provide evidence of a correlation between PAR activation and stimulation of PGE2 production in airways (3, 36). However, despite the physiological significance of these two systems, little is known about their interaction in development of fibrotic disorders. In particular, the detailed effects of PAR activation on PGE2 release in human lung fibroblasts (hLF) and the reverse effect of PGE2 on PARs in these cells have not yet been revealed.
In the present work, we have characterized the expression and signaling of PARs in hLF obtained from normal and fibrotic tissues. We clearly show that PAR-1 is critically involved in fibroblast functioning. We have investigated the possible link between PAR-1 and PGE2 and show that PAR-1 stimulates PGE2 synthesis, whereas PGE2 downregulates PAR expression in lung fibroblasts. These data indicate a novel autocrine feedback regulation of fibroblast functions and may be important for the understanding of development of fibrotic disorders in lungs.
MATERIALS AND METHODS
Thrombin and its substrate Tos-Gly-Pro-Arg-pNA were purchased from Sigma (Deisenhofen, Germany), trypsin was from Boehringer (Mannheim, Germany), and human cathepsin G and elastase were purchased from MP Biomedicals (Irvine, CA). The synthetic thrombin receptor agonist peptide (TRag; Ala-pFluoro-Phe-Arg-Cha-HomoArg-Tyr-NH2) and human PAR-2-activating peptide (PAR-2-AP; SLIGKV; Ser-Leu-Ile-Gly-Lys-Val) were purchased from Neosystem Laboratoire (Strasbourg, France). Human PAR-3-AP (TFRGAP-NH2; Thr-Phe-Arg-Gly-Ala-Pro-NH2) was from Biosyntan (Berlin, Germany). Pertussis toxin (PTX) was from Calbiochem (La Jolla, CA). Antibodies against PAR-1 (N-19) and PAR-3 (H-111) were purchased from Santa Cruz (Heidelberg, Germany). Antibodies against PAR-2 (A5, polyclonal rabbit) were a generous gift from Dr. M. Hollenberg. Alexa 488 rabbit anti-mouse IgG antibody and fura-2 AM were from Molecular Probes (Eugene, OR). Butaprost was from Cayman Chemical (Ann Arbor, MI). Plasmid pBJ containing the complete cDNA of human PAR-4 was a generous gift from Dr. S. Couglin (San Francisco, CA). The cell culture medium was from GIBCO/BRL (Eggenstein, Germany), and FCS, penicillin, and streptomycin were from Biochrom (Berlin, Germany).
Tissue samples were obtained from diagnostic open lung biopsies (fibrotic samples) or from healthy tissue areas during pneumonectomia for tumor resection (nonfibrotic samples). Ethical approval for experiments involving the use of human tissue was given by the local ethics committee. Fibroblast cell lines were obtained by mincing freshly excised lung parenchyma into ∼1 mm3 pieces, followed by digestion with collagenase IV (1 mg/ml; Sigma) for 30 min at 37°C. Fibroblasts were cultured in Iscove's modified Dulbecco's medium supplemented with 10% FCS, 10 mM glutamine, and antibiotics at 37°C in a humidified atmosphere of 5% CO2 until they reached confluence. Confluent cultures were enzymatically passed using Accutase (Sigma) with a split ratio of 1:3 to 1:4. Experiments were performed with cells between passages 4 and 9. The fibroblasts isolated from lung tissue specimens were characterized with respect to the expression of lineage-specific marker proteins. The cells were stained with antibodies directed against Thy-1, an antigen specific for fibroblasts. These cells also expressed the fibroblast-specific antigen D7-Fib and the enzyme prolyl-4-hydroxylase, which is involved in collagen synthesis. Neither CD68, a marker of monocytes/macrophages, nor CD45, a leukocyte membrane protein, was detected (data not shown). Fibrotic fibroblasts have been described and characterized before (8). Primary hLF CCD-25Lu [American Type Culture Collection (ATCC), Rockville, MD] were cultured in MEM-Earle's medium supplemented with 10% FCS and 100 μg/ml of penicillin and streptomycin (37°C, 5% CO2).
Cytosolic Ca2+ measurement.
The cytosolic Ca2+ concentration ([Ca2+]i) was measured using the Ca2+-sensitive fluorescent dye fura-2 AM. For dye loading, the cells grown on a coverslip were placed in 1 ml of HEPES-buffered saline (HBS, containing 20 mM HEPES, pH 7.4, 0.145 M NaCl, 5.4 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, and 25 mM glucose) for 30 min at 37°C, supplemented with 2 μM fura-2 AM. Loaded cells were transferred into a perfusion chamber with a bath volume of ∼0.2 ml and mounted on an inverted microscope (Axiovert 135; Zeiss, Jena, Germany). During the experiments, the cells were continuously superfused with HBS heated to 37°C.
Single cell fluorescence measurements of [Ca2+]i were performed using an imaging system from T.I.L.L. Photonics (Munich, Germany). Cells were excited alternately at 340 and 380 nm for 25–75 ms at each wavelength with a rate of 0.33 Hz, and the resultant emission was collected >510 nm. Images were stored on a personal computer, and, subsequently, the changes in fluorescence ratio (F340 nm/F380 nm) were determined from selected regions of interest covering a single cell.
RNA preparation and RT-PCR.
Total RNA was isolated and DNase treated from cultured fibroblasts with the total RNA isolation kit (RNeasy) from Qiagen (Hilden, Germany). First-strand cDNA was synthesized by reverse transcription of 1 μg of each RNA sample using Omniscript reverse transcriptase kit (Qiagen) in a final volume of 20 μl according to the manufacturer's instructions. One microliter of cDNA was then amplified by PCR using 20 pmol of each primer and Hotstar Taq Mastermix kit (Qiagen) in a final volume of 50 μl. The primers used are given in Table 1. All primer pairs amplified a fragment that crossed an intron, thereby distinguishing cDNA from genomic DNA by the size of the expected fragment after amplification. PCR conditions were as follows: denaturation 15 min at 95°C, 35 cycles at 94°C for 30 s, 55°C for 90 s, and 72°C for 60 s, and elongation at 72°C for 10 min. For E-prostanoid (EP)1-EP4 receptors, annealing temperature was 60°C. PCR products were analyzed by Tris-borate-EDTA agarose (1.5%) gel electrophoresis with ethidium bromide (10 mg/ml). Documentation was done using a still video system (Eagle Eye; Stratagene, Heidelberg, Germany). Analysis of fragment size indicated that the resulting PCR fragments were not amplified from genomic DNA.
Semiquantitative PCR amplification.
For quantification of the expression of genes of PARs, semiquantitative RT-PCR was used. The number of PCR cycles (25–30) and the amount of starting cDNA were determined for each gene to make sure that the amplification product was in a linear range. For all genes, amplifications were performed for 27 cycles using conditions described above. To compensate for relative differences between samples, in the integrity of the individual RNA samples and the variations in reverse transcription, GAPDH cDNA was also amplified and quantified for every test sample. To obtain semiquantitative results, 10 μl of each reaction were electrophoresed on 1.5% agarose gels. An Eagle Eye II video system (Stratagene) was used for gel documentation. PCR signals were normalized to the GAPDH signal of the corresponding RT product to get a semiquantitative estimation of the gene expression. PCR was done in duplicate for each sample and for each gene, and the results were averaged.
Real-time RT-PCR analysis.
cDNA was generated from 1 μg of total RNA with iScript cDNA synthesis kit (Bio-Rad) in a final volume of 20 μl according to the manufacturer's protocol. Real-time PCR was performed on the iCycler (Bio-Rad) in 25 μl of reaction volume using SYBR green PCR Master Mix (Bio-Rad) as described by the manufacturer. Amplification specificity of PCR products was confirmed by melting curve analysis and agarose gel electrophoresis. All mRNA measurements were normalized to the GAPDH mRNA level, which was unchanged in control and treated cells.
Fibroblast cells grown on coverslips were washed three times in PBS and incubated with primary antibodies against PAR-1 and PAR-3 (5–20 μg/ml) and with rabbit antiserum against PAR-2 (dilution 1:1,000) at 4°C for 1 h. Cells were then washed three times in PBS and incubated with secondary antibodies conjugated to Alexa 488 (10 μg/ml) at 4°C for 1 h. Controls were made by omitting the primary antibodies. The negative controls in this study yielded no detectable labeling. To check the specificity, the antibodies against PAR-1 were preincubated with fivefold (by weight) excess of specific antigen-blocking peptide at room temperature for 2 h. The cells were examined with a LSM510 confocal laser scanning microscope (Zeiss).
For studies of PGE2 production, fibroblasts were grown to confluence in 24-well plates, equilibrated in serum-free medium containing 0.1% BSA, and treated with TRag (50–200 μM) and thrombin (0.02–10 U/ml). In the case of synthetic peptides, aminopeptidase inhibitor (10 μM) was added to the culture medium. After 16 h of incubation, the supernatant was collected and stored at −80°C until PGE2 concentration was detected using an enzyme-linked immunoassay (Cayman Chemical) according to the manufacturer's instructions. The results were expressed as nanogram per milligram of total cell protein for each sample.
Enzyme activity assay.
Proteolytic activity of thrombin was assayed spectrophotometrically at 405 nm from hydrolysis of chromogenic substrate Tos-Gly-Pro-Arg-pNA. The final reaction mixture contained 0.05 M Tris·HCl buffer, pH 8.0, 0.1 M NaCl, and 6 mM substrate. The reaction was initiated by adding the enzyme. The amount of active thrombin was calculated using catalytic constants described in Ref. 41. Thrombin (1 U/ml) was detected to correspond to 5 nM of active site of the enzyme.
Data shown represent means ± SE. Statistical evaluation was carried out using Student's t-tests. Statistical comparison between normal and fibrotic cell lines in terms of PGE2 production was performed using the ANOVA test. Differences were considered significant when P was <0.05.
Detection of PARs in hLF by RT-PCR and immunocytochemistry.
The RT-PCR methodology was used to investigate the expression of all four PARs in hLF: normal fibroblasts (CCD-25Lu from ATCC) and cells obtained from normal and fibrotic patient tissues (NF and FF, respectively). GAPDH was used as an internal control for the RT-PCR conditions. The use of intron-flanking primers in all cases guaranteed that the amplification product was mRNA transcript and not genomic DNA signal. We showed that hLF express three of the four PAR subtypes: PCR products corresponding to PAR-1 (415 bp), PAR-2 (341 bp), and PAR-3 (452 bp) were found in both normal and fibrotic cells (Fig. 1A). No PCR fragment corresponding to PAR-4 was found in all cells tested. The PAR-4 PCR signal was confirmed using a plasmid containing full-length DNA of human PAR-4 as a positive control (Fig. 1A). The PCR specificity for PAR subtypes was controlled by sequencing the amplified products. The intensity of the bands for PAR-1 and PAR-3 was significantly higher than that of the band for PAR-2.
To quantify the relative abundance of PARs in hLF, semiquantitative RT-PCR was used. Relative amounts of PAR transcripts were calculated by normalization to the housekeeping gene GAPDH. In all cells derived from normal tissue, including CCD-25Lu cells, PAR-1 expression level was highest with some variation in the absolute values. The mean values (n = 9) for PAR-2 and PAR-3 expression were 2.1 ± 1.1% and 59 ± 15% of the PAR-1 expression level, respectively (Fig. 1B). The PAR expression profile did not change significantly from cell passage to passage. The same tendency in PAR expression level with predominant expression of PAR-1 was also observed in the fibrotic cells (Fig. 1B). Thus the relative abundance of PAR transcripts in hLF was confirmed to be PAR-1 > PAR-3 > PAR-2.
We furthermore compared the level of PAR expression in normal and fibrotic cells using quantitative real-time RT-PCR. The data presented in Fig. 1C are the means obtained with eight different cell cultures of NF and seven cultures of FF. The results are expressed as relative values, i.e., change of expression (n-fold) in fibrotic cells compared with normal cells. Both the PAR-1 and PAR-3 transcript levels were not significantly different between NF and FF (P > 0.05). However, the PAR-2 mRNA levels in fibrotic cells showed a 4.8-fold increase compared with normal cells (P < 0.01), although they remained still at a relatively low absolute value (Fig. 1B).
To check PAR expression on the protein level and to determine the receptor localization, we performed immunostaining of the cells using antibodies obtained against the NH2 terminus of the receptors. Figure 1D shows the representative pictures for PAR-1, -2, and -3 where fluorescence staining is clearly visible on the cell plasma membrane. PAR-2 staining was always weak, which probably reflects its low transcript level.
PAR-mediated Ca2+ mobilization in lung fibroblasts.
To provide evidence for functional expression of PARs in hLF, we examined the effects of PAR agonists on [Ca2+]i using NF, FF, and CCD-25Lu cells. All these cells demonstrated a comparable ability to be activated by PAR agonists: the proteases thrombin and trypsin and the corresponding synthetic receptor-activating peptides.
Short-term applications of thrombin and TRag stimulated a rapid increase of [Ca2+]i with a recovery back to basal level after agonist withdrawal (Fig. 2, A and B). At concentrations >0.5 nM, thrombin produced a concentration-dependent increase in [Ca2+]i with an EC50 value of ∼5 nM, whereas TRag caused an increase with an EC50 value of ∼1 μM (data not shown). The maximum response to TRag was comparable with that to thrombin, although TRag as an agonist was approximately two orders of magnitude less potent than thrombin.
Trypsin, known to activate both PAR-1 and PAR-2 (49), also produced an elevation of [Ca2+]i in lung fibroblasts (Fig. 2C). The cells demonstrated high heterogeneity with respect to the ability of PAR-2-AP to induce a [Ca2+]i increase even within the same experiment. PAR-2-AP-elicited Ca2+ responses were infrequently occurring and therefore could not be analyzed for the concentration-effect dependence. In CCD-25Lu cells, PAR-2-AP was completely unable to influence [Ca2+]i. In additional experiments, we checked the ability of PAR-2-AP to activate its cognate receptor using an epithelial cell line that has been shown in our previous study to produce stable Ca2+ responses to PAR-2 activation (47). Moreover, PAR-2-AP was able to induce a proliferative response in CCD-25Lu cells (data not shown). Thus, trypsin seems to induce Ca2+ responses through PAR-1, because functional activity of PAR-2 in lung fibroblasts is unclear in terms of stable Ca2+ signaling.
Application of a synthetic PAR-3-AP (TFRGAP-NH2), shown to be an effective PAR-3 activator in human smooth muscle cells (7), did not induce any effect in lung fibroblasts. To find out whether PAR-3 can be activated by thrombin, we used 100 μM TRag, which was able to induce maximal responses in lung fibroblasts, to desensitize PAR-1. The subsequent response to TRag or to saturating concentration of thrombin was attenuated to a similar extent [44 ± 2.5% (n = 102 cells) and 48 ± 3% (n = 88 cells), respectively]. In this experiment, a relatively larger response to thrombin would have indicated a partial response due to PAR-3 activation. Therefore, only PAR-1 seems to be involved in thrombin-mediated [Ca2+]i increase in hLF.
When cells were preincubated with 100 ng/ml of PTX for 24 h and then exposed to either thrombin or TRag, the Ca2+ response was reduced by 40% in both cases. The response both to thrombin, which can activate PAR-1 and PAR-3, and to TRag, which is the specific PAR-1 agonist, was reduced to the same extent. This provides an additional confirmation for the fact that only the thrombin receptor PAR-1, but not PAR-3, is able to induce Ca2+ signaling in hLF and that this signaling is partly mediated by Go/Gi type of G proteins.
Effect of neutrophil proteases and thermolysin on Ca2+ response.
We tested three proteases potentially capable of affecting PARs in hLF using cells from normal (NF and CCD-25Lu cells) and from fibrotic tissue (FF): human neutrophil-derived elastase and cathepsin G, and bacterial metalloprotease thermolysin. In Fig. 3, [Ca2+]i responses to the proteases tested and the subsequent responses to thrombin (Fig. 3, A, C, and E) and TRag (Fig. 3, B, D, and F) are shown. Thermolysin was not able to elicit a [Ca2+]i response by itself, but it reduced the subsequent response to thrombin in a concentration-dependent manner and did not affect that to TRag (Fig. 3, A and B). Thus thermolysin on hLF acts as a PAR-1-disabling protease, which makes impossible the receptor activation by thrombin. However, it does not damage the ligand-binding domain of the receptor, since the synthetic peptide was still able to elicit a Ca2+ signal.
Elastase did not affect the cells at concentrations <1 μM. At concentrations of 1 μM and higher, which is severalfold above the average physiological concentration of active elastase in inflamed lungs (∼200–400 nM) (53), the protease evoked a small [Ca2+]i signal, in some cases a hardly significant rise (Fig. 3, C and D). In all these cases, the [Ca2+]i response to thrombin and TRag was not influenced by preincubation with elastase.
The second neutrophil protease tested, cathepsin G, in contrast to elastase, was able to induce pronounced [Ca2+]i mobilization in fibroblasts at physiological concentrations (Fig. 3, E and F). The amplitude of the response to thrombin after short application of cathepsin G was reduced, whereas the response to TRag was not altered. Exposure to cathepsin G for a longer time (30 min) significantly abolished the cell response to thrombin but did not affect the response to TRag. Thus cathepsin G demonstrates both activating and disabling properties in lung fibroblasts.
Effect of PAR agonists on PGE2 release and COX-2 expression in hLF.
Experiments measuring the PGE2 release were performed with serum-deprived cells, influenced by agonists for 16 h. In most cases, 0.1–25 nM thrombin (0.02–5 U/ml) and 50–200 μM TRag caused a concentration-dependent increase of PGE2 release from both fibrotic and normal lung fibroblasts (Fig. 4, A and B) with a maximal value reached at 5 nM (1 U/ml) thrombin [693 ± 117% of control release for normal cells and 357 ± 120% for fibrotic cells (n = 4 for NF, n = 4 for FF)]. In normal CCD-25Lu cells, PAR-1 agonists showed the same tendency to induce PGE2 release. Fibrotic fibroblasts released significantly less PGE2 than normal cells (∼2-fold) after stimulation with thrombin (n = 4 for NF, n = 4 for FF, P < 0.01). The PAR-2-AP in the range of 50–200 μM did not significantly alter PGE2 release from both types of fibroblasts (n = 4 for NF, n = 4 for FF, P > 0.05; data not shown).
To determine whether the increase in PGE2 production is mediated by de novo synthesis of the inducible form of cyclooxygenase, COX-2, we assessed the steady-state transcript level of COX-2 using real-time RT-PCR analysis. For these experiments, we used CCD-25Lu cells since we observed that these cells possess the same properties as fibroblasts obtained from donor tissues. Thrombin (0.5 nM) and TRag (100 μM) dramatically increased COX-2 mRNA levels in CCD-25Lu cells (Fig. 5). After 3 h of incubation with thrombin and TRag, the COX-2 transcript level was increased 11- and 18-fold, respectively (data are the means of 3 independent experiments). After 6 h of stimulation with PAR-1 agonists, COX-2 mRNA level began to decline with an average five- to ninefold increase above the control level. These data demonstrate that the increase in PGE2 production through PAR-1 activation is due to newly synthesized COX-2.
PAR-1 agonists and PGE2 modulate PAR expression.
Three- to six-hour treatment of CCD-25Lu lung fibroblasts with thrombin (0.1 U/ml) and TRag (100 μM) resulted in statistically significant reduction of expression of PAR-1 by nearly 50%. Figure 6 shows the data of the expression levels of PAR-1 in hLF under the influence of PAR-1 agonists. PAR-2 and PAR-3 expression levels were changed similarly with the same time dependence (data not shown). After 24 h of incubation with PAR-1 agonists, mRNA levels exhibited a tendency to return to basal level. Proteolytic activity of thrombin was retained up to 24 h, as determined using the chromogenic substrate Tos-Gly-Pro-Arg-pNA (data not shown).
Interestingly, application of 50 nM PGE2 for 6 h caused two- to threefold reduction of mRNA steady-state levels of PAR-1, but unlike in the case of PAR agonists, this inhibitory effect became stronger (up to 3- to 4-fold reduction) after 24 h of PGE2 influence (Fig. 6). A comparable PGE2 effect on expression of PAR-2 and PAR-3 was also detected (data not shown).
PGE2 exerts its biological effects via four types of prostanoid receptors: EP1, EP2, EP3, and EP4. So far, PGE2 receptor expression has not yet been completely characterized in hLF isolated from adult subjects. RT-PCR analysis revealed stable expression of EP2 in all types of lung fibroblasts (NF, FF, and CCD-25Lu), whereas EP1 and EP4 showed weak bands visible only after 35 cycles of amplification. No sign of expression of EP3 could be detected (Fig. 7). Thus most likely PGE2 exerts its effects in lung fibroblasts mainly through EP2.
Because EP2 as well as EP4 is known to be coupled to the activation of adenylyl cyclase and increase of the cAMP level, we first tested the ability of the cAMP analog dibutyryl cAMP (dBcAMP) and whether it could mimic the effect of PGE2 in these cells. dBcAMP (1 mM) produced a decrease in PAR mRNA expression comparable with that of 50 nM PGE2. To further prove that this effect was due to EP2 signaling, we used the selective EP2 agonist butaprost. Application of butaprost (5 μM) resulted in a decrease in PAR mRNA expression, where the degree and the time course of reduction were similar to the effects observed with PGE2 (Fig. 6). These results implicate the involvement of EP2 in the downregulation of PAR expression in hLF by PGE2.
PARs in airways are known to stimulate tissue repair and regeneration and are also involved in the development of inflammatory processes and the induction of fibrosis (reviewed in Refs. 9, 11, and 37). The importance of PARs in the pathogenesis of both fibrosis and inflammation in lung is supported by the observation of increased PAR expression at both the mRNA and protein level in fibrotic and inflamed lung tissues (29, 32). However, little is known about changes in PAR expression and their functioning in lung fibroblasts during development of fibrosis. In our work, we addressed this question using hLF derived from normal and fibrotic lung tissues. We reveal a reciprocal influence between PARs on one side and the cytoprotectant PGE2 on the other side.
We have characterized the PAR expression pattern in human normal and fibrotic fibroblasts at the mRNA and protein level. Three PAR subtypes were identified in both normal and fibrotic cells: PAR-1, PAR-2, and PAR-3, with predominant expression of PAR-1 and less PAR-3, and a relatively minor level of PAR-2. Evaluation of PAR transcript levels revealed a significant (4-fold) increase of PAR-2 in fibrotic cells compared with normal ones. However, even this increased value of PAR-2 expression in fibrotic fibroblasts was still a minor population. PAR-1 and PAR-3 showed comparable expression levels in both NF and FF. A significant increase of PAR-1 expression in lung tissue of rats with bleomycin-induced pulmonary fibrosis was detected by Howell et al. (29). Nevertheless, it was recognized that this increase was associated predominantly with macrophages in interstitial inflammatory and fibroproliferative foci. Our data clearly indicate that PAR-1 and PAR-3 expression levels in lung fibroblasts are not changed during fibrosis development. Apparently, only PAR-2 expression is upregulated in fibrotic fibroblasts. The induction of PAR-2 expression under profibrotic conditions could be a feature of fibroproliferative disorders. This suggestion is supported by another recent report. Gruber et al. (25) showed that in different human fibroblast cell lines, the PAR-2 expression, which is relatively low under normal conditions, can be stimulated by the profibrotic factors PDGF and TGF-β1. Furthermore, we have seen that activation of PAR-2 in lung fibroblasts induced proliferation. Similarly, it was shown that tryptase induces lung, but not dermal, fibroblast proliferation via activation of PAR-2 (2). Thus from these data we suggest that PAR-2 may contribute to the development of pulmonary fibrosis and inflammation in addition to PAR-1.
PAR-1 activation by thrombin and TRag in both normal and fibrotic fibroblasts resulted in dose-dependent elevation of [Ca2+]i, which was partially PTX sensitive. We have demonstrated that thrombin-induced Ca2+ responses in hLF were mediated by PAR-1, not by PAR-3, which is also a thrombin receptor, despite the relatively high expression level of the latter. These data support the conclusion that PAR-1 is a major receptor mediating mitogenic, profibrotic, and proinflammatory effects of thrombin.
PAR-2-AP failed to produce stable Ca2+ responses in all the fibroblasts investigated in our work, although PAR-2 was obviously shown to be present on the cell plasma membrane. Because semiquantitative RT-PCR data demonstrated a very low level of PAR-2 expression compared with both PAR-1 and PAR-3 in hLF, one can expect that possible activation of the receptor PAR-2 results in rather weak [Ca2+]i signals. Another possibility could be the lack of a connection of PAR-2-induced intracellular signaling to the cytosolic Ca2+ elevation cascade in hLF.
Several human and animal studies have documented the important role of thrombin in acute and chronic lung diseases. Increased levels of active thrombin have been demonstrated in bronchoalveolar lavage fluid from patients with pulmonary fibrosis associated with systemic sclerosis (27) and in animals with bleomycin-induced pulmonary fibrosis (28). Such long-term exposure of lung fibroblasts to thrombin can influence PAR expression in the cells. To reveal this, we measured mRNA expression levels of PAR-1, PAR-2, and PAR-3 in hLF exposed to thrombin for 24 h. In our experiments, thrombin caused a significant decrease in mRNA level of all three PARs, with the same time dependency. TRag completely mimicked the effect of thrombin, demonstrating the contribution of PAR-1 activation in the regulation of PAR expression. The effect reached its maximum at 3–6 h of incubation with PAR-1 agonists and disappeared at 24 h. This suggests the existence of some compensatory mechanism, which counterbalances the downregulation of PAR expression by thrombin.
In addition to thrombin, some other proteases are also involved in the pathogenesis of lung fibrotic disorders and inflammation. Increased levels of mast cell tryptase, which is known to activate PAR-2, were found in bronchoalveolar lavage fluid from patients with fibrosing alveolitis (13). The pathogenesis of many lung diseases is characterized by a neutrophil-predominant inflammation, associated with excessive release of neutrophil proteases, such as elastase and cathepsin G (45, 52). In addition, there is evidence that elastase-deficient mice are resistant to bleomycin-induced fibrosis (21). Depending on the cell type, these neutrophil proteases were shown to activate or disable PARs (18, 20, 50). We wanted to elucidate whether elastase and cathepsin G influenced PAR-1 in hLF.
In our experiments, elastase at physiological concentrations was not able to induce Ca2+ responses in lung fibroblasts from both normal and fibrotic tissue and did not influence the subsequent responses to thrombin and trypsin. Cathepsin G induced a rise in [Ca2+]i and simultaneously depressed the subsequent responses to thrombin, but not to TRag, demonstrating the property of a PAR-1-disabling protease. This is important for regulation of PAR activity in hLF by the protease on the level of receptor integrity. There is also evidence that cathepsin G can mediate Ca2+ signaling through receptors other than PARs, in particular via activation of the bradykinin receptor B2 (26), which is known to be expressed in hLF (17). The question of whether mediation of Ca2+ signaling induced in hLF by cathepsin G is realized through PAR-1 or another receptor needs to be elucidated further.
Both acute and chronic lung diseases are associated with pathogenic bacteria that secrete a wide variety of factors, including zinc metalloproteases, which were shown to contribute to the pathophysiology of lung diseases (39). The prototype of this metalloprotease family is the thermophilic bacterial protease thermolysin, which was shown to act as a PAR-disabling protease in different cells (47, 48). Only in human bronchial epithelial (HBE) cells, we showed thermolysin to induce Ca2+ responses in a concentration-dependent manner (47). In hLF, thermolysin functions as a disabling protease for PAR-1, abolishing the subsequent activation of the receptor by thrombin without affecting the responsiveness to TRag.
The pathogenesis of pulmonary fibrosis is incompletely understood. However, enhanced fibroblast proliferation and collagen synthesis are known to be regulated by a complex interaction between stimulatory and inhibitory agents (44). Here, we analyzed the possible regulation through PAR activation, studying the release of PGE2, a crucial inhibitory mediator of lung fibroblast pathology (23, 30, 43). We have found that activation of PAR-1 in both normal and fibrotic hLF stimulated PGE2 release. The ability of fibrotic cells to release PGE2 upon thrombin stimulation was significantly impaired compared with normal cells. A similarly reduced capacity of PGE2 synthesis in fibrotic lung fibroblasts was reported for other stimuli, in particular for tumor necrosis factor-α (51), TGF-β1 (31), phorbol myristate acetate, LPS, and IL-1β (55). This effect was shown to be due to the diminished ability of fibrotic cells to upregulate COX-2 mRNA level (31, 55). In our study, we showed that despite differences in the value of PGE2 release, both cell types (NF and FF) displayed a clear tendency to increase PGE2 synthesis upon PAR-1 activation.
Activation of PAR-2 did not induce any significant changes in PGE2 release, even in fibrotic cells, despite increased PAR-2 expression in this cell type. Thus our data showing different effects produced by PAR-1 and PAR-2 ligands in Ca2+ signaling and PGE2 release led to the conclusion that activation of PAR-1 and PAR-2 in hLF clearly results in different cellular events. Additional studies are required to further characterize the contribution of PAR-2 to the development of lung fibrotic disorders.
PGE2 induces intracellular signaling via four different types of prostanoid receptors: EP1, EP2, EP3, and EP4, and has been shown to exert its effects in lung fibroblasts in an autocrine fashion (14, 34). We detected EP1, EP2, and EP4 receptor expression in both normal and fibrotic hLF, whereas EP3 was not found to be expressed in these cells. However, in embryonic hLF, EP2, EP3, and EP4, but not EP1, receptor expression was found (14, 34). Nevertheless, in both cases, a clear predominance of EP2 was revealed.
We investigated the PAR expression level in hLF, exposed to PGE2 for a long time period, to reveal whether prostaglandin was able to influence the PAR system. In contrast to the fast development of the effect of thrombin and its gradual decrease with time, the PGE2-induced downregulation developed more slowly and was permanent, reaching a three- to fourfold reduction of the PAR expression level after 24 h of incubation. The PGE2 effect in hLF was mediated mainly through the increase of the cAMP level, as it was mimicked by a stable membrane-permeant cAMP analog. Moreover, the selective EP2 agonist butaprost completely mimicked the effect of PGE2. This suggests that the EP2 receptor activation in hLF causes PGE2-induced downregulation of PAR expression.
Thus the stimulation of PGE2 synthesis in lung fibroblasts as well as long-term exposure of the cells to PGE2 from other sources can result in a suppression of cell responsiveness to PAR activation. In addition to its ability to decrease cell proliferation and collagen synthesis, PGE2 can protect hLF from activation by thrombin and other PAR agonists on the level of PAR expression. Our data further contribute to the understanding of the mechanisms through which PGE2 regulates lung fibroblast functions in a protective manner.
In conclusion, we present a novel autocrine feedback mechanism of regulatory action of PGE2 in hLF. PAR-1 activation induces PGE2 synthesis in hLF, whereas PGE2 by itself downregulates PAR expression through the EP2 receptor and the subsequent increase of cAMP level. This new pathway enhances the antifibrotic effectiveness of PGE2.
This work was supported by Deutsche Forschungsgemeinschaft Grant Re 563/11 and Bundesministerium für Bildung und Forschung (01ZZ0107).
The PAR-2 antibody was a generous gift from Dr. M. Hollenberg. The authors thank Dr. Fariba Sedehizade and Mohan Tulapurkar for help with confocal laser scanning microscopy.
↵* E. Sokolova and Z. Grishina contributed equally to this work.
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. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society