Prostaglandin E₂ protects murine lungs from bleomycin-induced pulmonary fibrosis and lung dysfunction

Abstract

Prostaglandin E₂ (PGE₂) is a lipid mediator that is produced via the metabolism of arachidonic acid by cyclooxygenase enzymes. In the lung, PGE₂ acts as an anti-inflammatory factor and plays an important role in tissue repair processes. Although several studies have examined the role of PGE₂ in the pathogenesis of pulmonary fibrosis in rodents, results have generally been conflicting, and few studies have examined the therapeutic effects of PGE₂ on the accompanying lung dysfunction. In this study, an established model of pulmonary fibrosis was used in which 10–12-wk-old male C57BL/6 mice were administered a single dose (1.0 mg/kg) of bleomycin via oropharyngeal aspiration. To test the role of prostaglandins in this model, mice were dosed, via surgically implanted minipumps, with either vehicle, PGE₂ (1.32 μg/h), or the prostacyclin analog iloprost (0.33 μg/h) beginning 7 days before or 14 days after bleomycin administration. Endpoints assessed at 7 days after bleomycin administration included proinflammatory cytokine levels and measurement of cellular infiltration into the lung. Endpoints assessed at 21 days after bleomycin administration included lung function assessment via invasive (FlexiVent) analysis, cellular infiltration, lung collagen content, and semiquantitative histological analysis of the degree of lung fibrosis (Ashcroft method). Seven days after bleomycin administration, lymphocyte numbers and chemokine C-C motif ligand 2 expression were significantly lower in PGE₂- and iloprost-treated animals compared with vehicle-treated controls (P < 0.05). When administered 7 days before bleomycin challenge, PGE₂ also protected against the decline in lung static compliance, lung fibrosis, and collagen production that is associated with 3 wk of bleomycin exposure. However, PGE₂ had no therapeutic effect on these parameters when administered 14 days after bleomycin challenge. In summary, PGE₂ prevented the decline in lung static compliance and protected against lung fibrosis when it was administered before bleomycin challenge but had no therapeutic effect when administered after bleomycin challenge.

idiopathic pulmonary fibrosis (IPF) is a serious and usually fatal disease that is characterized by progressive scarring of lung tissue and impaired pulmonary function (7, 21, 59). The outlook for patients suffering from this debilitating condition is poor, with a median survival of ∼2–5 yr following diagnosis (1, 2, 32–34). Unfortunately, present therapeutic strategies to treat IPF are less than satisfactory, and lung transplant is often the only viable treatment option (12, 52). Although the exact cause of IPF is unknown, several risk factors, including cigarette smoking, exposure to commonly prescribed drugs, environmental factors, and genetic predispositions are likely culprits (1). IPF is primarily an epithelial-fibroblast disorder in which disease progression is initiated by injury to pulmonary tissue followed by an inflammatory response that is characterized by an influx of eosinophils, neutrophils, lymphocytes, and macrophages that produce profibrotic cytokines such as TGF-β, IL-4, and IL-13. Such cytokines are known to stimulate epithelial-to-myofibroblast differentiation and/or act directly on fibroblasts to enhance proliferation and myofibroblast differentiation. The resulting surplus of myofibroblasts ultimately leads to increased extracellular matrix deposition, collagen production, and pathological fibrosis (7, 21, 59).

Cyclooxygenase (COX)-derived prostanoids are important biological mediators that are produced in the lung and have been shown to regulate pulmonary function under normal and pathological conditions (6, 8, 9, 15, 16). Two COX isoforms, COX-1 and COX-2, have been identified (22, 23). Whereas COX-1 is constitutively expressed, COX-2 is normally expressed at low levels but can be induced by a variety of stimuli such as bacterial endotoxin, cytokines, growth factors, and pharmacological agents (19, 28, 37, 40, 41, 43, 48). Both COX isoforms catalyze the conversion of arachidonic acid to the endoperoxide intermediate prostaglandin H₂ (PGH₂), which is further metabolized by prostanoid-specific synthases to produce biologically active prostaglandins (PGI₂, PGF_2α, PGD₂, and PGE₂) and thromboxane (50).

Several studies have examined the role of COX isoforms, as well as their prostanoid products, in the development and/or progression of pulmonary fibrosis (7). Oga et al. (42) recently showed that loss of PGF_2α signaling through its cognate receptor (prostaglandin F receptor) attenuates bleomycin-induced pulmonary fibrosis in mice, thereby demonstrating the profibrotic effects of this prostaglandin (42). On the other hand, the stable prostacyclin (PGI₂) analog, iloprost, was shown to protect against TGF-β2-induced fibrosis via PKA-dependent inhibition of the Ras/MEK/ERK pathway (51). Earlier reports showed that PGE₂ decreases fibroblast proliferation and inhibits collagen synthesis while promoting its degradation (4, 5, 13, 20, 36). Whereas most in vitro studies agree that COX-2 and PGE₂ are antifibrotic, in vivo data have been inconsistent. In a model of vanadium pentoxide-induced pulmonary fibrosis, mice lacking COX-2 experienced a more extreme inflammatory response and were more susceptible to developing fibrosis than wild-type controls or mice lacking COX-1 (6). Additional work from our laboratory showed that COX-2^−/− mice had exacerbated lung dysfunction, but not fibrosis, when exposed to bleomycin by oropharyngeal aspiration (9). In contrast, Lovgren et al. (39) argue that COX-2^−/− mice do indeed develop more severe fibrosis in addition to exacerbated loss of pulmonary function in response to bleomycin. Several factors may contribute to the discrepancies between the results of these studies, including differences in sex and genetic makeup of the mice, as well as the amount of bleomycin that was used to induce fibrosis.

There have also been discrepancies as to which prostanoid is primarily responsible for conferring the protective effects of COX-2 in pulmonary fibrosis. The above mentioned study by Lovgren and coworkers used several in vivo genetic models to elegantly show that prostacyclin, rather than PGE₂, protects against the development of fibrosis and the decline in lung function in response to bleomycin (39). However, a separate study showed that administration of the PGE₂ analog, 16,16-dimethyl PGE₂ (dmPGE₂), protected against bleomycin-induced lung inflammation and fibrosis in mice (17). Unfortunately, this study did not examine the effect of dmPGE₂ on pulmonary function in this model.

The current study was designed to clarify the effect of exogenously administered PGE₂ or iloprost on lung function, fibrosis, and inflammation in a mouse model of bleomycin-induced pulmonary fibrosis. In addition to examining the protective effects of these two compounds, we also investigated their potential therapeutic effects by administering them 2 wk after bleomycin challenge was initiated. This is an important aspect of our study, as it has not yet been examined by other groups. Our results show that mice receiving PGE₂ before the initiation of bleomycin challenge were significantly protected from the decline in pulmonary static compliance and had attenuated inflammatory and fibrotic responses to bleomycin. However, when administered 14 days after bleomycin, PGE₂ failed to provide significant therapeutic effects on lung function or fibrosis but did attenuate the inflammatory response that is normally observed during the fibrotic stage. Although iloprost prevented excessive collagen production and inflammation in the lung, it failed to protect against the decline in pulmonary static compliance induced by bleomycin. Like PGE₂, when iloprost was administered after bleomycin, it attenuated inflammation but had no effect on lung function or fibrosis.

MATERIALS AND METHODS

Animal care and treatments.

All studies were conducted in accordance with principles and procedures outlined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Animal Care and Use Committee of the National Institute of Environmental Health Sciences. Wild-type male C57BL/6 mice (10–12 wk old) were purchased from Taconic Farms (Rockville, MD).

PGE₂ and iloprost (Cayman, Ann Arbor, MI) were dissolved in vehicle containing 15% ethanol/sterile saline and were individually delivered 1 wk before or 2 wk after bleomycin treatment via subcutaneously implanted osmotic minipumps (model 1004; Alzet, Cupertino, CA). Minipumps were filled with 100 μl of 12 mg/ml of PGE₂ or 3 mg/ml of iloprost. Considering a release rate of 0.11 μl/h, this provided a final dose of 1.32 μg/h for PGE₂, a dose that resulted in significantly elevated serum PGE₂ levels and no obvious adverse side effects. For iloprost, this provided a dose of 0.33 μg/h and also showed no harmful side effects.

For bleomycin treatments, mice were anesthetized with isoflurane/oxygen and were administered bleomycin-sulfate (Sigma-Aldrich, St. Louis, MO) dissolved in sterile, endotoxin-free saline, or an equivalent volume of sterile saline (up to 75 μl) as a vehicle control, via an oropharyngeal aspiration method previously described (8). Bleomycin was administered at a dose of 1.0 mg/kg body wt. Mice were observed and weighed at regular intervals after aspiration of bleomycin or saline until assessment of inflammation or fibrosis/lung function at 7 or 21 days postdosing, respectively.

Invasive assessment of lung function.

Lung function was measured invasively as previously described (9) using the FlexiVent system (SCIREQ, Montreal, PQ, Canada). Briefly, mice were anesthetized with urethane and paralyzed with pancuronium bromide, and a portable heart rate monitor (CardioMonitor; BAS Vetronics, West Lafayette, IN) was used to ensure that appropriate anesthesia was maintained throughout the duration of ventilation. Ventilation was maintained at a rate of 150 breaths/min, a tidal volume of 7.5 ml/kg, and a positive end expiratory pressure of 3 cm of water. Mice were allowed to acclimate to the ventilator for 2–3 min before initiation of readings, and three to four total lung-capacity functions were performed during this acclimation period to prevent atelectasis and to ensure maximum airway and alveolar recruitment.

Static compliance (Cst), a measurement of the elasticity of the lungs, was calculated from pressure-volume curves by the FlexiVent software (version 5.1) using the Salazar-Knowles equation (46). Thirty seconds after each pressure-volume curve, a 2-s perturbation at a frequency of 2.5 Hz was applied to generate data using the single-compartment model of respiratory mechanics. Only measurements with a coefficient of determination of 0.95 or greater were used, and measurements were repeated until a total of three pressure-volume curves and three single-compartment perturbations, each with acceptable coefficients of determination, were obtained. The averages of these three measurements were determined for each mouse and averaged for each experimental group.

Analysis of cellular infiltration.

Bronchoalveolar lavage fluid (BALF) was collected from mice immediately following lung-function analysis. BALF collection was performed with two 1.0-ml aliquots of Hank's Balanced Salt Solution, and recovery was ≥80% for each animal. Recovered BALF was processed and analyzed for total cell counts using a Coulter Counter (Coulter Electronics, Hialeah, FL). Cell differential counts were determined by light microscopic assessment of Romanowsky-stained cytospin (Shandon, Pittsburg, PA) preparations of BALF.

Collagen assay.

Total collagen was measured in the right caudal lobe by homogenization in RIPA buffer containing protease inhibitors. Collagen content was determined on lysates using the colorimetric Sircol collagen assay kit (Biocolor, Carrickfergus, UK) according to the manufacturer's instructions.

Histopathological analysis.

The left lung was inflated and fixed with 4% paraformaldehyde, processed, and embedded in paraffin wax. Five-micron sections were cut and stained with hematoxylin and eosin or Masson's Trichrome. Histological assessment of the extent and severity of pulmonary fibrosis was determined using the Ashcroft method (3) by a pathologist who was blinded to the treatment groups.

Real-time PCR analysis.

Total RNA was purified from lung lysates using Qiagen RNeasy Mini kits according to the manufacturer's instructions (Qiagen, Valencia, CA), which included the optional on-column treatment with Qiagen RNase-free DNase (Qiagen cat. no. 79254). RNA was then reverse-transcribed into first-strand cDNA using the ABI High Capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA). Real-time PCR reactions were performed in duplicate as follows: 2 min at 50°C, 10 min at 95°C, then 40 cycles each at 95°C for 15 s followed by 60°C for 60 s in the ABI Prism 7900 HT Sequence Detector System. GAPDH was used as an endogenous control for all samples. The ΔΔCt method was used to determine the relative levels of gene expression (38). All TaqMan probes were purchased from Applied Biosystems with the following catalogue numbers: Gapdh (Mm99999915_g1), Tnf-α (Mm00443258_m1), intercellular adhesion molecule-1 (Icam) (Mm00516023_m1), and chemokine C-C motif ligand 2 (Ccl2) (Mm00441242).

BALF cytokine analysis.

Levels of CCL2, TNF-α, and macrophage inflammatory protein (MIP)-1α in BALF were determined with a Bio-Plex mouse cytokine kit (Bio-Rad, Hercules, CA), using fluorescently labeled microsphere beads and a Bio-Plex suspension array system (Bio-Rad) according to the manufacturer's instructions.

Analysis of serum and BALF prostaglandin levels.

Serum and BALF prostaglandin levels were analyzed by liquid chromatography-tandem mass spectrometry (LC/MS/MS). Briefly, 200 μl of serum or 500 μl of BALF was applied to Amprep Octadecyl C₁₈ columns (Amersham Biosciences, Piscataway, NJ), followed by washes with 0.1% acetic acid: 5% methanol. Samples were eluted from columns with methanol into a 2-ml collection tube, dried by SpeedVac (Thermo Scientific, Waltham, MA), flooded with argon and stored at −80°C until analysis. Details of the analytical methods have been previously described (8).

cAMP ELISA.

Serum cAMP levels were measured with the R&D Systems Parameter cAMP assay per manufacturer's instructions (R&D Systems, Minneapolis, MN).

Statistical analysis.

Statistical comparisons were performed by one-way analysis of variance (31) followed by Newman-Keuls post hoc tests, using GraphPad Prism software (GraphPad Software, San Diego, CA). In all instances, statistical significance was denoted when the P value was <0.05.

RESULTS

Implantation of PGE₂ osmotic minipumps leads to significantly elevated serum PGE₂ levels.

We first wanted to confirm that implantation of osmotic minipumps containing PGE₂ led to elevated circulating PGE₂ levels throughout the duration of our study. Vehicle- or PGE₂-containing minipumps were implanted subcutaneously, and serum was obtained after 7, 14, and 28 days of treatment. Serum analysis by LC/MS/MS showed that mice receiving PGE₂-containing minipumps had significantly higher (∼2.5-fold) PGE₂ levels than mice receiving vehicle containing minipumps at all time points examined (Fig. 1A). In contrast, there were no differences between the two groups in serum levels of 6-keto-PGF_1α (the stable PGI₂ metabolite), PGF_2α, PGD₂, thromboxane B₂ (the stable thromboxane A₂ metabolite), or other eicosanoids (data not shown). There were no differences in BALF PGE₂ levels following minipump implantation (data not shown).

Fig. 1.

Surgical implantation of subcutaneous prostaglandin E₂ (PGE₂) osmotic minipumps results in elevated serum PGE₂ and cAMP levels. A: mice were given vehicle osmotic minipumps for 7 days or minipumps containing PGE₂ for the designated time points. Serum PGE₂ levels were determined by liquid chromatography-tandem mass spectrometry (LC/MS/MS). B: serum was collected 1 wk after implantation of osmotic minipumps containing either PGE₂ or iloprost. ELISA showed a similar induction in cAMP levels in both PGE₂- and iloprost-treated mice. *P < 0.05 vs. saline-treated group; n = 3–5 per group for serum PGE₂ analysis and 9–11 per group for serum cAMP analysis.

Serum cAMP levels in PGE₂- and iloprost-treated mice.

To assess the downstream signaling effect of PGE₂ and iloprost, we measured serum cAMP levels as a biomarker of prostanoid signaling. As shown in Fig. 1B, both PGE₂ and iloprost induced a similar increase in serum cAMP levels following 1 wk of dosing. On the basis of these data, we conclude that the PGE₂ and iloprost doses chosen for our study are physiologically effective and comparable, at least from a cell-signaling standpoint.

Bleomycin treatment leads to increased BALF PGE₂ levels.

To assess the effect of bleomycin exposure on endogenous prostaglandin levels, we collected BALF from mice 7 days after saline or bleomycin challenge and measured prostaglandin levels by LC/MS/MS. As shown in Fig. 2, bleomycin treatment led to an ∼1.6-fold elevation in BALF PGE₂ levels. In contrast, no changes in 6-keto-PGF_1α levels were observed following bleomycin challenge. Similar analysis of other prostaglandins also showed no differences between saline- and bleomycin-treated animals (data not shown). These results suggest that, at least at this time point, PGE₂ (rather than PGI₂ or other prostaglandins) may mediate or be involved in mitigating the inflammatory, fibrotic, and/or physiological responses to bleomycin exposure in vivo.

Fig. 2.

Bronchoalveolar lavage fluid (BALF) PGE₂, but not 6-keto-PGF_1α, is elevated in response to bleomycin. Mice were treated once with either saline or 1 mg/kg bleomycin, and BALF PGE₂ and 6-keto-PGF_1α levels were measured 7 days later by LC/MS/MS. *P < 0.05 vs. saline-treated group; n = 6–7 per group.

Both PGE₂ and iloprost attenuate bleomycin-induced weight loss.

Over the course of 7 days, we observed an increase in body weight in animals that were administered saline, an effect that was similar in vehicle-, PGE₂- and iloprost-treated animals (Fig. 3A). The increased body weight is likely due in part to the fact that mice in this study were fed ground NIH-31 food mixed with reverse osmosis deionized (RODI) water. Essentially, this is a standard rodent chow that has been “mashed” in RODI water to make it easily digestible for animals that may experience a loss of appetite or difficulty eating, which often is the case following bleomycin treatment. It is thus provided on the cage floor in an effort to prevent excessive weight loss and mortality following bleomycin administration. Nevertheless, mice that were administered bleomycin experienced a loss of body weight in the first week. This weight loss was significantly attenuated in PGE₂- and iloprost-treated mice compared with vehicle-treated controls (Fig. 3A). After 21 days of treatment, all saline- and bleomycin-treated mice had similar body weights that were ∼10% above predosing levels (data not shown). None of the animals in this study died as a result of bleomycin treatment.

Fig. 3.

PGE₂ and iloprost attenuate bleomycin-induced weight loss and decline in lung function. Mice were given vehicle, PGE₂, or iloprost via minipumps and treated with bleomycin 1 wk later. A: body weights were recorded after 1 wk of bleomycin (Bleo) administration. B: static compliance (Cst) was measured via Flexivent 21 days after bleomycin administration. Bleomycin-induced decrease in Cst observed in vehicle-treated animals was significantly attenuated in PGE₂-treated mice. *P < 0.05 vs. vehicle + bleomycin group, †P < 0.05 vs. vehicle + saline group; for body weight analysis, n = 4–8 mice for saline groups, n = 9 for bleomycin groups. For lung function analysis, n ≥ 6 for saline groups, n ≥ 11 for bleomycin groups.

PGE₂ protects against bleomycin-induced decline in lung function.

Studies from our group and others have consistently shown that bleomycin administration leads to a significant reduction in pulmonary static compliance (C_st) in mice (9, 39, 57). To test the effect of PGE₂ and iloprost on bleomycin-induced pulmonary dysfunction, we challenged mice via oropharyngeal administration with 1 mg/kg of bleomycin and then conducted invasive assessment of lung function 21 days later. No differences in static compliance were observed among vehicle-, PGE₂-, and iloprost-treated mice that received saline, indicating that neither PGE₂ nor iloprost influenced lung static compliance under basal conditions (Fig. 3B). Relative to saline-treated controls, bleomycin administration led to reduced static compliance in all three experimental groups; however, PGE₂-treated mice maintained significantly higher static compliance following bleomycin administration than vehicle-treated mice (Fig. 3B).

These functional changes in response to bleomycin may seem small. Nonetheless, small changes can have profound effects on an animal's overall health. Importantly, when given slightly higher doses of bleomycin that induce greater reductions in Cst, animal mortality became an issue. The dose we chose for our study provided a modest level of pulmonary dysfunction but was not associated with mortality.

Cellular infiltration is attenuated in airways of PGE₂- and iloprost-treated mice.

Bleomycin initiates an inflammatory response that includes infiltration of cells into the airways, an event that has been implicated as a causative factor in the development of pulmonary fibrosis (59). We therefore examined the effects of treatment with PGE₂ or iloprost on this process during both the early inflammatory stage (day 7) and the late fibrotic stage (day 21) of bleomycin-induced pulmonary fibrosis. Relative to saline-administered controls, vehicle-, PGE₂-, and iloprost-treated mice that were administered bleomycin had elevated total, lymphocyte, and neutrophil cell numbers in BALF 7 days after dosing. Whereas PGE₂ and iloprost had modest effects on total cell numbers, both significantly attenuated lymphocyte infiltration at 7 days after bleomycin dosing. Macrophage cell numbers did not differ between saline- and bleomycin-treated animals in any of the treatment groups (Fig. 4A). We also observed a marked increase in total, lymphocyte, and macrophage cell numbers 21 days following bleomycin administration relative to control mice that received saline (Fig. 4B). Similar to our results at the 7-day time point, PGE₂ significantly attenuated lymphocyte infiltration at 21 days. In contrast, iloprost had no significant effect on lymphocyte infiltration at 21 days. Although macrophage numbers were higher in bleomycin-treated animals, no differences were observed between the three different treatment groups (Fig. 4B). Furthermore, almost no neutrophils were observed in BALF from saline- or bleomycin-treated animals in any of the treatment groups and were thus not included in Fig. 4B. Importantly, neither PGE₂ nor iloprost had significant effects on total, lymphocyte, neutrophil, or macrophage cell numbers at either of the examined time points in saline-treated animals (Fig. 4, A and B).

Fig. 4.

Cellular infiltration in PGE₂- and iloprost-treated mice. A: although no differences were observed in total cell numbers among any of the bleomycin-treated groups, PGE₂ and iloprost both inhibited lymphocyte influx into BALF 7 days after bleomycin challenge. B: no differences in total cell numbers were observed among any of the bleomycin-treated groups 21 days after challenge. In contrast, PGE₂, but not iloprost, inhibited lymphocyte infiltration into BALF at this later time point. *P < 0.05 vs. vehicle + bleomycin group, †P < 0.05 vs. vehicle + saline group; n ≥ 4 for saline groups, n ≥ 9 for bleomycin groups.

Our results showing similar levels of cellular infiltration at 7 and 21 days after bleomycin challenge may seem a bit unusual. However, these results are consistent with previous studies from our laboratory and were reproducible in multiple independent experiments. Discrepancies between the current findings and previous reports from other groups may be attributable to a variety of factors including bleomycin dosing levels, age, sex, and strain of mice, or differences in other experimental conditions.

Cell adhesion molecule and inflammatory cytokine levels following bleomycin administration.

It is well established that the early inflammatory response to bleomycin includes increased expression of several proinflammatory cytokine, chemokine, and cell adhesion molecule genes in the lung (29, 30). Therefore, we used real-time RT-PCR to quantitatively assess Tnf-α, Icam, and Ccl2 (also known as monocyte chemoattractant protein 1) gene expression changes in the lung 7 days following bleomycin administration. As shown in Fig. 5A, bleomycin administration led to significantly elevated levels of TNF-α, ICAM, and CCL2 mRNAs in vehicle-, PGE₂-, and iloprost-treated mice compared with saline-treated controls. Whereas PGE₂ and iloprost had no significant effect on bleomycin-induced TNF-α or ICAM mRNA levels, PGE₂ treatment led to significantly reduced CCL2 mRNA levels following bleomycin administration. We also observed a considerable, albeit statistically insignificant, reduction of bleomycin-induced CCL2 mRNA levels in iloprost-treated mice. It should be noted that neither PGE₂ nor iloprost had a significant effect on TNF-α, ICAM or CCL2 mRNA levels in saline-treated groups (Fig. 5A).

Fig. 5.

Cell adhesion molecule and inflammatory cytokine levels in bleomycin-treated mice. A: bleomycin challenge resulted in elevated lung TNF-α, intercellular adhesion molecule (ICAM), and chemokine C-C motif ligand 2 (CCL2) mRNA levels after 7 days (determined by real-time RT-PCR). TNF-α and ICAM mRNA levels were not affected by PGE₂ or iloprost treatment. In contrast, PGE₂ significantly reduced CCL2 mRNA levels. B: BALF CCL2 protein levels were significantly reduced in PGE₂- and iloprost-treated mice that received bleomycin (determined by Bio-Plex). *P < 0.05 vs. vehicle + bleomycin group, †P < 0.05 vs. vehicle + saline group; n ≥ 4 for saline groups, n ≥ 8 for bleomycin groups.

In addition to analysis of the above-mentioned markers at the mRNA level, we also examined CCL2, TNF-α, and MIP-1α protein levels by Bio-Plex analysis in BALF 7 days following treatment with saline or bleomycin. TNF-α and MIP-1α were undetectable in BALF from saline- and bleomycin-treated animals (data not shown). Whereas CCL2 protein levels were also undetectable in BALF from saline-treated mice, bleomycin exposure led to significantly increased CCL2 levels (Fig. 5B). Consistent with our observations at the mRNA level, treatment with either PGE₂ or iloprost significantly attenuated the increase in BALF CCL2 protein levels following bleomycin challenge (Fig. 5B). Together, these findings demonstrate an anti-inflammatory role for PGE₂ 7 days following bleomycin administration. The anti-inflammatory properties of iloprost appear to be less consistent.

Effect of PGE₂ and iloprost on bleomycin-induced collagen deposition and development of lung fibrosis.

Twenty-one days following bleomycin administration, mice typically experience a robust fibrotic response in the lung characterized by a surplus of myofibroblasts that produce and release excessive amounts of collagen, resulting in histological changes similar to those observed in patients with pulmonary fibrosis (12, 59). Biochemical assessment of lung tissue showed that vehicle-, PGE₂-, and iloprost-treated mice that received bleomycin had elevated lung collagen content relative to control mice that received saline (Fig. 6A). However, among mice dosed with bleomycin, collagen content in PGE₂- and iloprost-treated mice was significantly lower than in vehicle-treated mice. Furthermore, bleomycin exposure led to a robust inflammatory and fibrotic response characterized by collagen deposition and cellular infiltration within the alveolar walls and alveolar spaces as observed on Masson's trichrome-stained histological sections (Fig. 6B). Histopathological quantification for fibrosis (Ashcroft scores) revealed a significant protective effect of PGE₂ in bleomycin-treated animals (Fig. 6C). Although we also observed a modest reduction in Ashcroft scores in iloprost-treated animals, this group was not significantly different from vehicle-treated animals that received bleomycin (Fig. 6C). Importantly, PGE₂ and iloprost had no significant effect on lung collagen content or histopathological evidence of lung fibrosis in animals that received saline instead of bleomycin (Fig. 6, A–C).

Fig. 6.

Effect of PGE₂ and iloprost on bleomycin-induced collagen production and development of lung fibrosis. A: lung collagen content 21 days after bleomycin challenge was lower in mice treated with PGE₂ or iloprost. B: representative histology of Masson's trichrome-stained slides. C: calculated fibrosis scores based on histopathological assessment of Masson's trichrome-stained sections revealed a protective effect of PGE₂ in bleomycin-treated lungs. *P < 0.05 vs. vehicle + bleomycin group, †P < 0.05 vs. vehicle + saline group; n ≥ 6 for saline groups, n ≥ 15 for bleomycin groups.

Therapeutic effect of PGE₂ and iloprost on lung function.

Although several studies have examined the protective effects of PGE₂ and iloprost in pulmonary fibrosis, none have investigated the therapeutic effects of these compounds in a fibrosis model. We therefore administered PGE₂ or iloprost 14 days after the initiation of fibrosis with bleomycin to test for potential therapeutic effects.

As shown previously, no differences in static compliance were observed among vehicle-, PGE₂-, and iloprost-treated mice that received saline. Relative to saline-treated controls, bleomycin administration led to reduced static compliance in all three experimental groups. Interestingly, there were no significant differences between any of the bleomycin-treated groups (Fig. 7A). Thus PGE₂ and iloprost lacked a significant therapeutic effect on lung function in our model.

Fig. 7.

Therapeutic effect of PGE₂ and iloprost on lung dysfunction and cellular infiltration following bleomycin. A: mice were challenged with bleomycin and then given vehicle, PGE₂, or iloprost via minipumps 14 days later. Static compliance was measured via FlexiVent 21 days after bleomycin administration. Bleomycin-induced decrease in Cst observed in vehicle-treated animals was not affected by PGE₂ or iloprost. B: PGE₂ and iloprost both inhibited total cell and lymphocyte influx into BALF when administered 14 days after bleomycin challenge. *P < 0.05 vs. vehicle + bleomycin group, †P < 0.05 vs. vehicle + saline group; n ≥ 8 for saline groups, n ≥ 21 for bleomycin groups.

Therapeutic effect of PGE₂ and iloprost on cellular infiltration.

As shown previously, relative to saline-administered controls, vehicle-, PGE₂-, and iloprost-treated mice that were administered bleomycin had elevated total, lymphocyte, and macrophage cell numbers in BALF 21 days after dosing. Mice treated with PGE₂ or iloprost 14 days after the initiation of bleomycin challenge had significantly fewer total cells and lymphocytes, but macrophage numbers were similar in all three treatment groups (Fig. 7B). On the basis of these data, we conclude that PGE₂ and iloprost act therapeutically to attenuate inflammation associated with the later stages of bleomycin-induced pulmonary fibrosis.

We would also like to note here that, although macrophages and lymphocytes make up a majority of the total cell population in BALF following bleomycin challenge, there are several other cell types, including neutrophils, eosinophils, and basophils, that contribute to the total population. Thus the large differences observed in total cell counts following bleomycin challenge are also impacted by these cell types. Although we did assess these cell types, no differences were observed between any of the treatment groups, and thus, in an effort to make the relevant data more presentable, these data were not included in our results.

Therapeutic effect of PGE₂ and iloprost on bleomycin-induced collagen deposition and development of lung fibrosis.

Consistent with previous results, biochemical assessment of lung tissue showed that vehicle-, PGE₂-, and iloprost-treated mice that received bleomycin had elevated lung collagen content relative to control mice that received saline. There were no significant differences in collagen content between any of the bleomycin-treated groups (Fig. 8A). Furthermore, histopathological quantification for fibrosis showed no significant differences between any of the bleomycin-treated groups (Fig. 8B). Thus PGE₂ and iloprost had no therapeutic effect on collagen content or fibrotic lung injury in our model.

Fig. 8.

Therapeutic effect of PGE₂ and iloprost on bleomycin-induced collagen production and development of lung fibrosis. Lung collagen content (A) and calculated fibrosis scores (B) were not altered in mice receiving PGE₂ or iloprost 14 days after bleomycin challenge. †P < 0.05 vs. vehicle + saline group; n ≥ 8 for saline groups, n ≥ 21 for bleomycin groups.

DISCUSSION

This was the first study designed to compare the protective vs. therapeutic effects of exogenously administered PGE₂ or iloprost in a mouse model of pulmonary fibrosis. We used a well established model of bleomycin-induced pulmonary fibrosis and examined inflammatory, fibrotic, and lung functional endpoints in mice receiving osmotic minipumps containing vehicle, PGE₂, or iloprost either before or after bleomycin administration. Our results demonstrate that PGE₂ protects against the airway inflammation, lung fibrosis, and decline in lung function observed in this model. In contrast, iloprost did not protect against pulmonary dysfunction and had less pronounced effects on the inflammatory and fibrotic responses to bleomycin. The therapeutic effects of both PGE₂ and iloprost were minimal, with significant beneficial effects only observed on cellular infiltration associated with later stages of fibrosis in this model. Thus the major finding from this study is that PGE₂ modulates the early injury phase but not the later fibrotic phase in the bleomycin-induced model of pulmonary fibrosis.

A number of other reports have observed anti-inflammatory and antifibrotic effects of PGE₂ in the lung; however, our report is the first to demonstrate PGE₂-mediated protection from the decline in lung function associated with bleomycin administration. Although it seems logical that reduced inflammation and fibrosis would translate to improved pulmonary function, the association between histological/biochemical and functional defects is not always apparent. For example, a recent report from our laboratory showed that mice lacking COX-2 had exacerbated lung dysfunction, but not increased fibrosis, in response to bleomycin (9). Several possibilities for this phenomenon exist, including regional differences in collagen deposition, differences in collagen maturity, and/or variable effects of prostanoids on airway smooth muscle cells. In comparison, several studies have shown increased lung fibrosis in COX-2^−/− mice in response to various stimuli (6, 30, 39). Possible explanations for these discrepancies include experimental design differences among studies such as differences in the doses of profibrotic agents used and in genetic background, age, and sex of the animals used. Our present study differs from the previously published reports in that we specifically tested the effects of PGE₂, whereas other studies that used COX-2 knockout mice likely tested effects of several prostaglandins simultaneously, possibly confounding and/or masking the effects of PGE₂.

Three comparable studies are worth considering in more detail. Lovgren et al. (39) used several genetic models to show that prostacyclin, rather than PGE₂, is responsible for both the histological and functional protective effects of COX-2 in bleomycin-induced pulmonary fibrosis. Differences between genetic models, which chronically alter gene expression levels, and administration of a drug for a relatively short duration cannot be underestimated. Furthermore, the amount of bleomycin used in our study (1 mg/kg) differs from the 0.025 and 0.05 units of bleomycin used by Lovgren et al. (39), regardless of body weight. In the present study, BALF prostacyclin levels were not altered by bleomycin treatment, but PGE₂ in BALF was highly elevated in bleomycin-challenged mice, suggesting a key role for PGE₂, but not prostacyclin, in this model. This observation could indeed be due to the fact that prostacyclin is primarily produced by endothelial cells, whereas PGE₂ is produced by many cell types, including epithelial cells, macrophages, and many other cell types that are either in or contiguous to the airway. It is thus possible that PGE₂ is more likely to be detected in BALF, and prostacyclin would more likely be detected in other compartments, such as urine. Another relevant study that should be noted is the one by Failla and coworkers (17), which showed that treatment with the PGE₂ analog, dmPGE₂, protects mice from bleomycin-induced pulmonary inflammation and fibrosis (17). Although this study utilized a similar experimental protocol, there are key differences between their report and ours. First, we used PGE₂ rather than an analog and showed that its stability allows it to be effectively administered via osmotic minipumps for up to 4 wk. Second, our study utilized mice on a C57BL/6 genetic background, whereas the Failla study used CD mice. This is important because many other comparable studies, including the one by Lovgren et al. (39), also used C57BL/6 mice. Considering the variability in bleomycin response in different mouse strains, this is an important experimental variable that should be taken into consideration. Finally, the most important difference between our study and the one by Failla et al. (17) is that, in addition to examining the role of PGE₂ in the inflammatory, histological, and biochemical response to bleomycin, we also studied the effect of PGE₂ on lung function. Importantly, our data demonstrate significant protective effects of PGE₂ on this endpoint. We believe that it is important to measure lung function in any model of pulmonary fibrosis because patients suffering from this condition often have profound changes in pulmonary function, and this significantly influences their quality of life. Indeed, by the time pulmonary function is compromised, the lungs have already undergone significant histological and biochemical deterioration. The prognosis for patients with reduced lung function attributable to fibrosis is generally poor, suggesting that functional assessment may be a more accurate predictor of disease status and survival than traditional pathological analysis of fibrotic lung disease. Finally, a very recent study by Zhu and coworkers (61) showed that one intraperitoneal injection of iloprost protected mice from bleomycin-induced pulmonary fibrosis and lung dysfunction. The most obvious difference between this study and ours is that we systemically treated mice with iloprost for the duration of the study, whereas Zhu et al. only gave a single intraperitoneal injection of iloprost at the time that bleomycin challenge was initiated. Another consideration is that our study dosed mice with 1 mg/kg of bleomycin, whereas Zhu et al. used 2 to 3 mg/kg of bleomycin, a dose that we have generally found to be lethal.

The beneficial role of PGE₂ in the lung is also supported by data obtained using other experimental models of pulmonary fibrosis and inflammation. For example, PGE₂ has been shown to inhibit fibroblast proliferation and collagen synthesis, regulate wound closure in airway epithelium, and inhibit TGF-β-induced fibroblast to myofibroblast transition (35, 47). PGE₂ also prevents LPS-induced NF-κB activation in monocytes, thereby inhibiting TNF-α production while stimulating increased levels of the anti-inflammatory cytokine, IL-10 (11, 55, 56). Whereas we observed a significant increase in TNF-α and ICAM gene expression in bleomycin-treated animals, neither PGE₂ nor iloprost had a significant effect on this response. Interestingly, we were unable to detect TNF-α or MIP-1α in BALF from saline- or bleomycin-treated animals. However, in support of our findings showing reduced lymphocyte infiltration in PGE₂-treated animals, multiple studies have shown that PGE₂ is involved in the regulation of lymphocyte trafficking and differentiation in a variety of models (44, 53). This is also consistent with our observation of reduced lung CCL2 mRNA levels and BALF CCL2 protein levels in PGE₂-treated animals. Of note, CCL2 is known to recruit T lymphocytes to sites of injury and infection (10).

Although the role of inflammation and lymphocyte infiltration in the development of pulmonary fibrosis is debatable, several studies have demonstrated potent profibrotic effects of cytokines such as IL-4 and IL-13 that are produced by lymphocytes (14, 18, 49, 60). Examination of various inflammatory cells has also shown that increased eosinophils, neutrophils, and T lymphocytes are associated with poorer prognosis in patients suffering from IPF. In fact, a 2007 study by Parra et al. (45) showed that the most important predictor for survival in patients with IPF/usual interstitial fibrosis was the number of CD3⁺ T lymphocytes present in lung tissue. Thus, whether lymphocytes significantly contribute to the development of fibrosis or simply serve as a marker of overall pulmonary health, our results demonstrating reduced lymphocytes in PGE₂-treated mice support the concept that targeting this eicosanoid pathway may benefit patients afflicted with IPF or other forms of pulmonary fibrosis.

The fact that PGE₂ is such an effective protective agent against pulmonary fibrosis but has little therapeutic effect when administered during the fibrotic stage is interesting and deserves discussion. Our results show that the only endpoint in which PGE₂ or iloprost acted therapeutically was on infiltration of total cells and lymphocytes. This observation, along with the results from our 7- and 21-day protective studies, suggests that both PGE₂ and iloprost act potently to reduce early inflammation and injury associated with fibrosis but have little effect when administered at the later fibrotic phase. If we are to believe that inflammation contributes to the development of fibrosis and subsequent decline in lung dysfunction, then it is possible that PGE₂ or iloprost treatment must be initiated earlier to be effective therapeutic agents. Dosing in our therapeutic experiments did not occur until animals were well into the fibrotic stages of disease, and, although inflammation was reduced, it may have been too late to have had an effect on fibrosis and lung function. It is also possible that, if we extended our study for another 1–2 wk, PGE₂ and iloprost would have had more time to impact fibrosis and thus would have been more effective from a therapeutic standpoint.

Finally, it is clear that COX-derived eicosanoids, especially PGE₂, have a profound effect in attenuating the inflammatory, fibrotic, and functional changes observed in pulmonary fibrosis. Previous studies have reported an induction of COX-2 expression in response to bleomycin exposure, which could result in significant alterations in the overall prostanoid profile (24). In our hands, the only COX-2 metabolite that is significantly altered in response to bleomycin challenge is PGE₂, suggesting that it is likely the main COX metabolite involved in mediating the response to bleomycin. Although COX-2 induction leads to elevated PGE₂ levels in a variety of disease states, the lung is unique in that PGE₂ serves beneficial actions as an anti-inflammatory mediator, whereas it is mostly a proinflammatory factor in other tissues (54).

In summary, our data show that exogenously administered PGE₂ via osmotic minipumps rescues mice from the lung functional decline, pulmonary inflammation, and lung fibrosis that are normally observed after bleomycin challenge. Furthermore, we show that iloprost selectively protects against some inflammatory and fibrotic endpoints but not against the associated decline in lung function. Interestingly, both PGE₂ and iloprost had little therapeutic effect when administered 2 wk after initiation of bleomycin challenge. Further investigation comparing the protective and therapeutic properties of these two compounds is warranted.

GRANTS

This work was supported with funds from the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences (Z01 ES025043 and ES050167). J. Card was the recipient of a Senior Research Training Fellowship from the American Lung Association of North Carolina.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.