help button home button
AJRCMB
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Published ahead of print on April 10, 2008, doi:10.1165/rcmb.2007-0186OC
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Online Supplement
Right arrow All Versions of this Article:
2007-0186OCv1
39/3/324    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Pulichino, A.-M.
Right arrow Articles by Sayegh, C. E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Pulichino, A.-M.
Right arrow Articles by Sayegh, C. E.
American Journal of Respiratory Cell and Molecular Biology. Vol. 39, pp. 324-336, 2008
© 2008 American Thoracic Society
DOI: 10.1165/rcmb.2007-0186OC

Identification of Transforming Growth Factor β1–Driven Genetic Programs of Acute Lung Fibrosis

Anne-Marie Pulichino1, I-Ming Wang3, Alexandre Caron1, James Mortimer1, Anick Auger1, Yves Boie1, Jack A. Elias4, Aileen Kartono1, Lijing Xu1, Joseph Menetski2 and Camil E. Sayegh1

1 Merck Frosst Centre for Therapeutic Research, Kirkland, Québec, Canada; 2 Merck Research Laboratories, Rahway, New Jersey; 3 Rosetta Inpharmatics, Seattle, Washington; 4 Section of Pulmonary and Critical Care Medicine, Yale University School of Medicine, New Haven, Connecticut

Correspondence and requests for reprints should be addressed to Camil E. Sayegh, PhD, 16711 TransCanada Highway, Kirkland, PQ, H9H 3L1 Canada. E-mail: camil_sayegh{at}merck.com


   Abstract
Top
 Abstract
CLINICAL RELEVANCE
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Lung fibrosis is characterized by excessive accumulation of extracellular matrix components leading to progressive airflow limitation. Distinct profibrotic pathways converge on the activation of transforming growth factor–β (TGF-β), a central growth factor implicated in most fibroproliferative diseases. Recently, enforced expression of bioactive human TGF-β1 (hTGF-β1) in lungs of transgenic mice was shown to recapitulate several key pathophysiologies observed in fibrotic disorders of the lung, including cellular inflammation, tissue fibrosis, and myofibroblast hyperplasia. Inducible expression of hTGF-β1 in this system provided a unique opportunity to characterize TGF-β–driven mechanisms that precede and/or follow the onset of inflammation and fibrosis. Using gene expression profiling in lungs, we demonstrate temporal activation of key genetic programs regulating cell movement and invasiveness, inflammation, organ remodeling, and fibrosis. Consistent with our gene expression data, multiple soluble mediators associated with inflammation and tissue remodeling were markedly elevated in the bronchoalveolar lavage fluid of mice expressing hTGF-β1. We observe significant TGF-β1–driven infiltration of F4/80+ mononuclear cells producing bioactive arginase, a marker of alternatively activated macrophages. Finally, we identified a common "fibrosis" gene signature when comparing our findings with published data derived from preclinical and clinical studies.

Key Words: TGF-β • lung fibrosis • gene expression profiling • alternatively activated macrophages • arginase



   CLINICAL RELEVANCE
Top
 Abstract
 CLINICAL RELEVANCE
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
To our knowledge, this is the most extensive analysis of TGF-β–mediated events in the lung. We identified several key genetic programs that precede or occur during the fibrotic response. We extended our findings to published mouse and human studies.

 
Fibrosis is a key determinant of disease progression and prognosis in pathologies such as idiopathic pulmonary fibrosis (IPF), with patient survival time being 2 to 5 years after diagnosis. Fibrosis is also observed in the airways of patients with asthma and chronic obstructive pulmonary disease and is suggested to contribute to progressive airflow limitation (1).

Fibrosis is believed to arise as a consequence of repeated tissue injury and deregulated wound repair responses. At least two non–mutually exclusive pathogenic routes, exaggerated Th2-dominated inflammatory responses and epithelial cell injury, are proposed to directly activate profibrotic pathways. In the latter, a profibrotic role for alternatively activated macrophages (aaMac) in patients with IPF, which are thought to arise in response to Th2 cytokines, has recently been suggested (1, 2).

Transforming growth factor (TGF)-β promotes multiple features associated with fibrosis, such as fibroblast migration, proliferation, and differentiation into myofibroblasts. In addition, TGF-β directly regulates the excessive deposition of extracellular matrix components (ECM), a hallmark of fibrosis, through activation of ECM synthesis and inhibition of ECM degradation (1). Increased expression of TGF-β is a common feature of multiple fibroproliferative diseases, suggesting a central role for TGF-β in the initiation and progression of fibrosis. Consistently, ectopic expression of TGF-β in the lungs of rodents is sufficient to induce progressive tissue fibrosis, recapitulating several pathophysiologies observed in human lung diseases (1, 3–5). Furthermore, the profibrotic role for the Th2 cytokine IL-13 has been shown to be mediated through activation of TGF-β expression (6). Tremendous insight into the mechanisms contributing to fibrosis has been gleaned from the use of the neoplastic drug bleomycin in rodents, which results in acute lung injury, inflammation, and progressive pulmonary fibrosis in susceptible strains. An essential role for TGF-β in mediating the acute lung injury and fibrosis that follows administration of bleomicin has been demonstrated using genetic and pharmacological interventions (7, 8). Taken together, various profibrotic pathogenic routes converge on the expression and/or activation of the pleiotropic growth factor TGF-β (1).

In this study, we examined profibrotic pathways induced by expression of a bioactive human form of TGF-β1 (hTGF-β1) in the lungs of mice (3). Using gene expression profiling, we demonstrated, for the first time, the temporal activation of key TGF-β–driven genetic programs regulating cell movement and invasiveness, inflammation, and organ remodeling and fibrosis in vivo. Consistent with our gene expression data, multiple soluble mediators associated with inflammation and tissue remodeling were identified in the bronchoalveolar lavage fluid (BALF) of mice expressing hTGF-β1. We noted the hTGF-β1–driven infiltration of F4/80+ mononuclear cells. These cells expressed markers consistent with an aaMac macrophage phenotype in the BALF of mice. Finally, a common gene signature was identified by comparing genetic programs activated by ectopic expression of hTGF-β1 to published gene expression data derived from bleomycin-treated animals.


   MATERIALS AND METHODS
Top
 Abstract
 CLINICAL RELEVANCE
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
TGF-β Transgenic Mice and Doxycycline Treatment
Generation of the CC10-rtTA-tTS-TGF-β1 was described previously (3). These mice were rederived and backcrossed on the B6JBom background at Taconic and renamed B6JBom-TgN(CC10-rtTA)-TgN(tTS)-TgN(TGF-β1) (Germantown, NY). Because of the use of the B6JBom background, it is plausible that mice used in our study were not genetically identical to those described previously (3). We used male and female 8- to 10-week-old transgenic (Tg) mice and wild-type (WT) littermate control mice in all experiments. No statistically significant differences in the fibrotic response were noted between male and female mice. To induce expression of active hTGF-β1 in the lungs of transgenic mice, doxycycline (DOX) was added to the drinking water at 1 mg/ml. Lungs and BALF were harvested at various time points ranging from 2 to 14 days after the addition of DOX to the drinking water (i.e., hTGF-β1 induction). All animal procedures were approved by the Animal Care Committee at the Merck Frosst Centre for Therapeutic Research (Kirkland, PQ, Canada) and were performed according to guidelines established by the Canadian Council on Animal Care.

Lung Compliance Measurement
Mice were premedicated with a combination of ketamine (Ayerst Veterinary Laboratories, Guelph, ON, Canada) and diazepam (Sandoz, Boucherville, PQ, Canada) (130/7 mg/kg, respectively) administered in a single intraperitoneal injection. Five minutes later, mice were anesthetized using intraperitoneal injection of sodium pentobarbital (40 mg/kg) (Somnotol; MTC Pharmaceuticals, Cambridge, ON, Canada). When motor responses to nociceptive stimuli were abolished, a cut and beveled 18-g needle (~ 1.2 cm long) was inserted into the trachea via tracheotomy. Surgical thread was used to secure the canulated trachea, which was connected to a small animal ventilator (FlexiVent; Scireq, Montreal, PQ, Canada). Mice were ventilated in the supine position by mechanical ventilation (150 breaths/min, 10 ml/kg, positive end-expiratory pressure of 3 cm H2O). To collect measurements under passive mechanical conditions, respiratory muscle paralysis was induced by intraperitoneal injection of pancuronium bromide (1.2 mg/kg) (Sandoz, Boucherville, PQ, Canada). Approximately 1 minute before the initiation of measurements for each mouse, the lungs were inflated at a pressure of 30 cm H2O by closing the expiratory line to standardize the volume of the respiratory system. Measurements of lung mechanics were performed by briefly interrupting (2 seconds) ventilation and applying oscillations in volume (2.5 Hz) to the airway opening. A quasi-static pressure–volume curve was performed by step inflation (0.1-ml increments) of the lungs until airway pressure reached 25 cm H2O, followed by a similar deflation step. Pressures from the normalized compliance curves were extrapolated by fitting the Salazar Knowles equation at 0.1-ml increments, and lung compliance was calculated for each animal.

Bronchoalveolar Lavage Fluid Collection
Mice were killed using intraperitoneal overdose of sodium pentobarbital (120 mg/kg) (Somnotol; MTC Pharmaceuticals). BALF was collected with 0.5 ml of room temperature PBS injected twice via a tracheal canula and collected as fraction 1. This procedure was repeated twice using 1 ml of PBS per lavage, which was pooled as fraction 2. Samples were kept on ice until processed. Fraction 1 lavage fluid (~ 0.4 ml) was centrifuged (5,000 rpm, 10 min), and the supernatant was collected for cytokine analysis. The remaining cell pellet was pooled with fraction 2 for cellular analysis. The total number of cells present in the BALF was determined using a Cell-Dyn 3700 hematocytometer (Abbott Laboratories, Mississauga, ON, Canada). Cytospins were prepared using a cytocentrifuge (Shandon, Pittsburgh, PA) and stained with modified Wright-Giemsa stain. Manual differential cell counts were made on 100 cells/slide. The total number for each cell population was determined by multiplying the percentage of each cell population by the total cell counts.

Tissue Harvest and Histology
Left and right lung lobes were harvested separately, rapidly frozen in liquid nitrogen, and stored at –80°C for subsequent analysis. BALF collection was omitted before harvesting of lung tissues for gene expression profiling. For histological examination, animals were perfused via the left cardiac ventricle with 10 ml of PBS followed by 10 ml of 4% buffered neutral formalin (BNF). Lungs were slowly inflated with 3 ml of 4% BNF, and the trachea was ligated before the lungs were removed and stored overnight in 4% BNF. After processing and embedding in paraffin, 5-µm midtrachea longitudinal sections were affixed to microscope slides and stained with Masson's Trichrome.

Total Hydroxyproline Determination
Right lungs were homogenized using a MagNA Lyser Instrument (Roche Applied Science, Laval, PQ, Canada) in 2 N NaOH. Samples were autoclaved at 120°C for 20 minutes and maintained at 40°C. Samples were pelleted at 14,000 rpm, and the supernantant was collected. Hydroxyproline content was measured by mixing 10 µl of sample supernatant with 90 µl of chloramine-T solution (0.056 M Chloromine-T, 10% n-propanol diluted in 0.24 M citric acid, 0.88 M sodium acetate trihydrate, 1.2% glacial acetic acid, and 0.85 M NaOH) for 25 minutes. Finally, 100 µl of Ehrlich's reagent (1 M 4-(dimethylamino)benzaldehyde in 2:1 vol/vol n-propanol/perchloric acid) was added, and the reaction was incubated for 20 minutes at 65°C. Colorimetric measurement was performed at a 550-nm wavelength using a plate reader (SpectraMax; Molecular Device, Sunnyvale, CA). The concentration of hydroxyproline per microliter of sample was multiplied by the total volume of the sample to obtain the total hydroxyproline content. All chemical reagents were purchased from Sigma-Aldrich Canada Ltd (Oakville, ON, Canada)

Arginase Assay
BALF cells were pelleted, and 3 x 105 cells were resuspended in 50 µl of Tris 0.1 M, 0.4% Triton supplemented with Complete Protease Inhibitor Cocktail Tablets (Roche). Forty microliters of the cellular homogenate or BALF was used to estimate arginase activity using the QuantiChrom Arginase Assay Kit following the manufacturer's instructions (BioAssay Systems, Hayward, CA).

BALF Analysis
Levels of hTGF-β1 and IGF-1 in mouse BALF were assessed using the human TGF-β1 and mouse IGF-1 Quantikine ELISA (R&D Systems, Minneapolis, MN), respectively. Multi-analyte profiling was performed by Rules Based Medicine, Inc. (Austin, TX).

RNA Profiling
Lung samples from two time points (2 and 14 d) were collected. Total RNA was extracted from left lungs using the RNeasy Midi kit as described by the manufacturer (Qiagen, Valencia, CA). Samples were treated with DNaseI on-column (Qiagen) for 30 minutes. RNA concentration was measured using a NanoDrop ND-1000 (NanoDrop Technologies, Wilmington, DE), and RNA integrity was determined with a 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA). Samples displaying a RNA integrity number greater than 7.5 were used for profiling. Samples from WT or Tg mice treated with water only (–DOX) were combined to serve as reference pools and compared with (time-matched) individual lung samples of –DOX (self versus self) and +DOX animals. Microarray profiling was conducted using a Mouse Custom 44K array comprising 40,572 probes. The two-color microarrays were scanned using the Agilent scanner and proprietary image acquisition software allowing rigorous image quality control. Experimental quality control was performed in MATLAB (Mathworks Inc., Natick, MA). Expression data were loaded into the Resolver (Rosetta proprietary software database, available from http://www.rosettabio.com/products/resolver/default.htm) for transformation normalization and error modeling. Fluor-reversed pairs for each sample were combined to give a single log-ratio and a P value for technical variability for each biological sample compared with its appropriate control. Next, 1D and 2D clustering and classifier analysis was used to obtain an overview of the experiment. To select gene sets that are modulated by hTGF-β1, one-way ANOVA analyses were performed between time-matched (1) WT+Dox versus Tg+DOX and (2) Tg–DOX versus Tg+DOX. Genes with an ANOVA P ≤0.0005 were selected and intersected with genes that meet the following criteria: A combined absolute fold change that is ≥1.25 and a Resolver error model P ≤0.0005. Subsequently, we applied additional filtering of absolute fold change ≥1.5 and Resolver error model P ≤0.0001 in at least two samples on the intersecting sequences of the two analyses (i.e., WT+Dox versus Tg+DOX and Tg–DOX versus Tg+DOX) for each of the Day 2 and Day 14 data sets. The resulting gene signature for Day 2 and Day 14 treated animals contained 1,031 sequences and 2,343 sequences, respectively, and the union of the Day 2 and Day 14 gene sets contained 2,861 sequences. We then projected these 2,861 sequences to all controls (WT+Dox and Tg–DOX), and the Tg+Dox replicate combined sample pool and selected for sequences with a combined absolute fold change ≥1.25 and a Resolver error model P ≤0.01. The final gene signature is shown in the k-means clustering in Figure 2A. Data mining of these clusters was performed using prior biological knowledge, known pathways, and gene ontology or keyword over-representation. We performed functional analysis using Ingenuity Pathways Analysis (Ingenuity Systems, www.ingenuity.com). The significance or P value, calculated using the right-tailed Fischer's exact test, is proportional to the number of modulated genes that are associated with each functional class.

We obtained several gene lists to support our comparative analysis. Of the 469 genes identified by the Kaminski and colleagues study, 453 were successfully translated into TGI gene models to enable an accurate comparison to our gene signature (9). Raw CEL data from the Haston and colleagues studies were reanalyzed to extract the global bleomycin-induced signature in the C57bl/6 strain consisting of 670 regulated genes (10). Furthermore, comparative analysis of basal levels and bleomycin-induced gene expression changes, obtained from the study by Haston and colleagues, in the fibrosis prone C57bl/6 compared with the fibrosis-resistant C3H mouse strains allowed us to identify a unique signature of 257 regulated genes. Specifically, 157 genes were up-regulated and 84 down-regulated in C57bl/6 compared with C3H after bleomycin treatment. This signature differed from the published study because we excluded genes that showed large baseline changes in gene expression (10). Finally, published lists of genes modulated in patients with IPF were converted into mouse orthologs using reciprocal blast as a requirement and intersected with the mouse data (11, 12).

Q-PCR and TaqMan Low-Density Array
Total RNA was extracted from mouse tissues using RNeasy Mini kit (Qiagen). Samples were treated with DNase on-column (Qiagen). RNA was converted into cDNA using the Taqman reverse transcription kit or High Capacity cDNA archive kit according to the manufacturer's instructions (Applied Biosystems, Foster City, CA). Five to eight microliters of cDNA was amplified using the Taqman Fast universal PCR master mix supplemented with 0.2 units uracil N-glycosylase, and reactions were run on a ABI Prism 7900 HT detection system (Applied Biosystems). TaqMan probes and primer sets were purchased from Applied Biosystems. The PCR reactions were denatured for 2 minutes at 50°C followed by 2 minutes at 95°C and then subjected to 50 cycles of amplification (94°C for 1 second followed by 20 seconds at 60°C). TaqMan low-density array analysis was performed according to the manufacturer's protocol (Applied Biosystems). Gene expression levels were normalized to endogenous controls (Gapdh) using the {Delta}{Delta}Ct method and expressed relative to the levels observed in control mice (value set at 1).

Flow Cytometry
BALF cells were washed three times in staining buffer (1x PBS, 2.5% FBS, 0.1% sodium azide). Detection of surface markers was performed using the following antibodies: anti-F4/80-APC (eBioscience, San Diego, CA), anti-I-Ab(MHCII)-FITC (BD Bioscience, Mississauga, Ontario, Canada), anti-CD86-FITC (eBioscience), anti-CD80-FITC (eBioscience). Istoype controls were rat IgG2a-APC, mouse IgG2a-FITC, rat IgG2a-FITC, and hamster IgG-FITC (eBioscience). Briefly, 1 x 105 cells were resuspended in 45 µL of staining buffer. Nonspecific antibody binding was blocked by adding 5 µL of anti-Fc{gamma} III/II receptor (Fc block, BD Bioscience), and cells were incubated for 15 minutes on ice. Monoclonal antibodies were added (0.2–0.5 µg) in a final volume of 100 µL and incubated on ice for 30 minutes in the dark. Cells were washed three times and resuspended in 400 µL staining buffer. Cells were analyzed on a FACScalibur by gating for live cells based on forward and side scatter (Becton Dickinson, Franklin Lakes, NJ).

Macrophage Purification
After BALF cell collection, red blood cells were lysed using the mouse erythrocyte lysing kit (R&D Systems, Minneapolis, MN). BALF cells were rinsed three times in P buffer (1 x PBS, 0.5% BSA, 2 mM EDTA) and incubated in 90 µL of P buffer and 10 µL of anti-Fc{gamma} III/II receptor (Fc Block, BD Bioscience) for 15 minutes at 4°C. Anti-F4/80-PE antibody (eBioscience) was added (5 µg) and incubated for 15 minutes at 4°C. Cells were washed with a 20-fold volume excess of P buffer and resuspended in 100 µL P buffer containing 10 µL of anti-PE magnetic beads (Miltenyi Biotec, Inc., Auburn, CA). Cells were incubated for 15 minutes at 4°C, washed once with a 20-fold volume excess of P buffer, and resuspended in 500 µl. Positive enrichment of F4/80+ cells was performed by passing cells twice over MACS Separation Columns (MS) following the manufacturer's instructions (Miltenyi Biotec, Inc.). Overall, a purity of 90 to 95% was achieved as measured by flow cytometry (FACScalibur, Becton Dickinson).

Statistical Analysis
Each n value corresponds to a different animal unless otherwise described. Group comparisons for two or more groups were performed using one-way ANOVA with Bonferroni post test. Student's t test was performed if only two groups were compared. All data are represented as mean ± SEM.


   RESULTS
Top
 Abstract
 CLINICAL RELEVANCE
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
Phenotypic Characterization of hTGF-β1 Triple Transgenic Mice
We used the B6JBom-TgN(CC10-rtTA)-TgN(tTS)-TgN(TGF-β1) triple-transgenic mouse described previously to examine TGF-β–driven mechanisms contributing to lung fibrosis (3). In this mouse, expression of bioactive human TGF-β1 (hTGF-β1) is induced by the reverse tetracycline-controlled transcriptional activator (rtTA) in the presence of doxycycline (DOX). In the absence of DOX, the tetracycline-controlled transcriptional silencer (tTS) actively suppresses background hTGF-β1 expression resulting in its tight temporal regulation. Furthermore, because rtTA and rtTS expression is driven by the lung epithelial clara cell protein 10 promoter (CC10), induction of hTGF-β1 levels by DOX is restricted to the lung. Consistently, 10,000 pg/ml and 750 pg/ml of hTGF-β1 were detected in the BALF of transgenic mice treated with doxycycline for 2 and 14 days, respectively (Tg+DOX) (Figure 1A and Figure E1 in the online supplement). In contrast, hTGF-β1 was not detected in the BALF of transgenic mice in the absence of DOX treatment (Tg–DOX) (Figure 1A). Tg–DOX mice were phenotypically indistinguishable from WT mice that were treated with DOX (WT+DOX) or with regular water (WT–DOX) and were therefore used as controls. Acid activation of latent TGF-β in BALF obtained from day 2 and day 14 Tg+DOX mice demonstrated dramatic increases in total levels of active TGF-β (Figure E1). Because hTGF-β1 is expressed only under the active form, these data indicate that substantial levels of endogenous latent TGF-β are induced by hTGF-β1 (3). Therefore, it is plausible that both transgenic and endogenous TGF-β contribute to the pathophysiologies observed in this system.


Figure 1
View larger version (33K):
[in this window]
[in a new window]

  		Figure 1. hTGF-β1 transgenic mice as a model of lung fibrosis. (A) Levels of hTGF-β1 measured in the bronchoalveolar lavage fluid (BALF) of wild-type (WT) and transgenic (Tg) mice given regular water (–DOX) or water supplemented with DOX (+DOX) for 14 days. Values are means ± SE. n = 4 (WT) and n = 9 (Tg). (B) Cellular infiltration kinetics in the BALF of Tg–DOX (squares) compared with Tg+DOX (triangles). Values are means ± SEM. n = 4 (Day 2), n = 6 (Day 4), n = 9 (Day 7), n = 8 (Day 10), and n = 9 (Day 14). ***P < 0.001 versus Tg–DOX mice. (C) Differential BALF cell counts from Tg mice treated with DOX for 14 days (+DOX) and Tg control mice (–DOX). Neutrophils (neutro), lymphocytes (lympho), eosinophils (eosino), macrophages or monocytes (macro/mono), and total white blood cells (total WBC) counts are displayed. Values are means ± SEM. n = 4. (D) Flow cytometric analysis of BALF cells derived from Tg+DOX animals. Isotype control staining is shown in black, and anti-F4/80-APC staining is shown in red. Pooled BALF cells obtained from three Tg+DOX animals were used. This is a representative data set out of three experiments. (E) Total hydroxyproline content in right lung lobes from Tg-DOX (white bars) and Tg+DOX (black bars) at day 7, 10, and 14 of treatment. Values shown are means ± SEM. n = 5 to n = 9. *P < 0.05 and ***P < 0.001 versus Tg–DOX mice. (F) Lung compliance measurements in Tg–DOX (white bars) versus Tg+DOX (black bars) 14 days after DOX treatment. Values are means ± SEM. n = 6. (G) Representative histological sections of left lung lobes isolated from Tg–DOX and Tg+DOX at Day 14. Sections were stained with Masson's trichrome. (Original magnification: x20.).

 
Induction of hTGF-β1 in the lungs of transgenic mice resulted in a time-dependent increase in white blood cells in BALF (Figure 1B). Increased white blood cell counts were first detected approximately 4 days (0.7 x 106 cells) after DOX treatment, reaching maximal levels (2.6 x 106 cells) by Day 10. Similar kinetics of cellular infiltration in the lung parenchyma were observed by histological analysis of lung sections (Figure 1G and data not shown). The majority of infiltrating cells in the BALF displayed a monocytic/macrophage-like phenotype as determined by histological staining and F4/80 surface expression (Figures 1C, 1D, and 4A). Furthermore, we noted the presence of infiltrating lymphocytes and the absence of significant numbers of neutrophils and eosinophils in the BALF of Tg+DOX mice (Figure 1C). Hydroxyproline measurements were used to estimate lung collagen content. A significant increase in lung hydroxyproline was observed in Tg+DOX mice by Day 7, reaching a maximal 2.5-fold increase over controls (Tg–DOX) by Day 10 to Day 14 after DOX treatment (Figure 1E). Histological characterization showed marked subepithelial fibrosis with evidence of parenchymal fibrosis (Figure 1G and data not shown). These observations are overall consistent with previously published studies (3). Finally, hTGF-β1–driven pathophysiologies resulted in decreased lung compliance, suggesting a role for fibrosis in deregulated lung mechanics (Figure 1F).

Gene Expression Profiling of hTGF-β1 Transgenic Mice
We used gene expression profiling to identify mechanisms contributing to lung fibrosis. Two time points after hTGF-β1 induction, Day 2 and Day 14, were selected in our studies. Transgenic mice treated with DOX for 2 days expressed hTGF-β1 in the BALF in the absence of detectable cellular inflammation and excessive collagen deposition. This provided us with a unique opportunity to compare early TGF-β–driven mechanisms, preceding cellular infiltration of the lungs, to hTGF-β1–dependent phases of inflammation and lung fibrosis (14 d after DOX). We ruled out a significant contribution to gene expression changes derived from genomic insertion of the transgene and ectopic lung expression of tTA and tTS by comparing the gene signatures of WT–DOX to Tg–DOX. In the latter mice, only 67 (0.2%) and 160 (0.4%) genes were modulated at Day 2 and Day 14, respectively (data not shown). Similarly, DOX treatment alone resulted in minor changes in gene expression as determined by comparing the gene signatures of WT–DOX with WT+DOX mice, respectively. Indeed, only 93 (0.2%) and 256 (0.6%) genes were modulated at Day 2 and Day 14, of which only three genes were commonly regulated by DOX at both time points.

The expression of 1,886 and 2,563 genes, accounting for 4.7% and 6.4% of the transcriptome probed on our arrays, was significantly modulated in Tg+DOX at Day 2 and Day 14, respectively (Figures 2A and 2B). Overall, 968 genes (Day 2) and 1,307 (Day 14) were up-regulated, whereas 918 (Day 2) and 1,256 (Day 14) genes were down-regulated. These genes were organized into six gene clusters according to their regulation state (up-regulated or down-regulated) and temporal regulation (Figure 2A). Clusters 4 (151 genes) and 5 (146 genes) include genes modulated only at Day 2, whereas clusters 2 (485 genes) and 6 (489 genes) are comprised of genes modulated at Day 14 after TFG-β1 induction (Figure 2A). Finally, clusters 1 (822 genes) and 3 (767 genes) represent genes that are commonly regulated at the day 2 and day 14 time points. These commonly regulated genes accounted for the majority of gene expression changes, representing 84% and 62% of all modulated genes at Day 2 and Day 14, respectively (Figure 2B). Genes showing the greatest modulation in each cluster are summarized in Table 1. Broadly, genes modulated by TGF-β1 revealed a strong functional association with cellular movement, immune response, tissue morphogenesis, and organ disease including respiratory (Figures 2C–2E).


Figure 2
View larger version (38K):
[in this window]
[in a new window]

  		Figure 2. Gene expression profiling of hTGF-β1 transgenic mice. (A) Heat map of modulated genes in lungs of Tg+DOX (n = 4) mice compared with control mice (average of WT–DOX, n = 4; WT+DOX, n = 4; and Tg–DOX, n = 4) after 2 or 14 days of treatment. Modulated genes are subdivided into six distinct clusters. (B) A Venn diagram reveals temporally regulated and commonly regulated genes by hTGF-β1 in lungs of Tg+DOX mice treated for 2 or 14 days. (C) Functional classification of genes uniquely modulated at Day 2 (left side of Venn diagram) (D) or Day 14 (right side of Venn diagram) (E) and commonly regulated genes (intersect in Venn diagram). (F) Heat map of modulated genes functionally associated with cell movement or (G) extracellular matrix remodeling.

 

View this table:
[in this window]
[in a new window]

  		TABLE 1. GENES WITHIN EACH CLUSTERS SHOWING THE HIGHEST MODULATION

 
Cell Movement and Invasiveness
Genes functionally associated with cellular movement were significantly regulated by hTGF-β1. Several genes regulating cell traffic were exclusively modulated at Day 2, a time point preceding infiltration of inflammatory cells (Figure 2F). Consistently, of all genes in clusters 4 and 5, Ccl7 (MCP-3) gene expression showed the strongest modulation being induced 6.6-fold. Additional chemokines, such as Ccl11 (eotaxin) and Ccl19 (MIP-3β), were up-regulated. A recently identified chemokine-like protein, Unq473 (DMC), that has been shown to increase migration of monocytes and dendritic cells, was induced by 1.9-fold at Day 2 (13). Additional chemokines and chemokine receptors were modulated at Day 2 and Day 14 and included genes such as Ccl1, Ccl2 (MCP-1), Ccl8, Ccl12, Cxcl10, Cxcl12, and Ccr5 (Figure 2F). Several chemokines and their receptors listed above, such as Ccl2, Ccl12, and Ccr5, have been associated with lung fibrosis (14–16). Finally, consistent with the pro-oncogenic role of TGF-β1, several genes implicated in epithelial to mesenchymal transition (EMT) and fibroblast invasiveness were up-regulated at the early Day 2 time point including Snai1 (2.4-fold), Pvr (1.5-fold), Has1 (2.3-fold), and Fosl1 (2.4-fold) (Figure 2F and Table 2) (17–20).


View this table:
[in this window]
[in a new window]

  		TABLE 2. EXPRESSION LEVELS OF SELECTED GENES MODULATED AT DAY 2 AND DAY 14 IN TG+DOX MICE

 
Immune Response and Inflammation
Genes functionally associated with immune response and inflammation were modulated as early as Day 2 after TGF-β1 overexpression. Of particular interest is the up-regulation of genes encoding for cytokines such as IL-6, IL-11, and IL-13. IL-11 expression in lungs was elevated at Day 2 only (Table 2). In contrast, IL-6 and IL-13 were up-regulated at Day 2 and Day 14. IL-6 showed the strongest modulation among the signature of 1,886 regulated genes at Day 2, being up-regulated 23.6-fold. Ectopic expression of each of these cytokines listed above was shown to promote lung remodeling and fibrosis, in part by regulating TGF-β1 expression or function (6, 21). Next to IL-6 in extent of modulation at Day 2 was Tnfsf18, the gene coding for the glucocorticoid-induced TNFR-related protein ligand (GITRL), which was up-regulated 17.1 and 6.9-fold at Day 2 and Day 14, respectively (Table 2). Tnfrsf18 (GITR) was also up-regulated 1.5-fold at Day 14 (Table 2). Recently, the GITR-GITRL axis was demonstrated to regulate bleomycin-induced chronic lung injury and fibrosis (22).

Organ Remodellng and Fibrosis
Overall, the common gene signature included growth factors and both constituents and modulators of the ECM. Endogenous Tgfb1, Areg (amphiregulin), and Igf1, were variably up-regulated at Day 2 and Day 14. These growth factors have been proposed to play an important role in promoting fibroblast proliferation and fibrosis (23, 24). ECM modulators and constituents included collagens, cathepsins, metalloproteases, and antiproteases (Figure 2G). We noted the modulation of genes implicated in collagen fibrillar formation and homeostasis such as Lox and Mfap5, which were up-regulated at Day 2 (4.3- and 2.6-fold) and Day 14 (2.5- and 4.4-fold), respectively (Table 2). The former gene encodes for lysyl oxidase, an enzyme implicated in the initiation of cross-links in collagens and elastin and which overexpression is associated with fibrosis (25). The latter gene, Mfap5, encodes for microfibrillar associated protein 5, which is demonstrated to stabilize type I collagen (26).

Soluble Mediators in BALF of Tg+DOX Mice
The gene expression data revealed the up-regulation of several chemokines, cytokines, and growth factors. Therefore, we were interested in determining whether we could detect their cognate proteins in the BALF of Tg+DOX mice. To do this, expression of 69 cytokines, chemokines, growth factors, and antiproteases was simultaneously assessed using a multianalyte platform (for the complete list, refer to supplemental material). Consistent with the gene expression data, we note the elevated expression of several mediators such as IL6, CCL2, CCL7, CCL9, CCL11, CXCL2, SPP1 (osteopontin), and TIMP1 (Figure 3A). In contrast, we detected elevated expression of CCL3, CCL4, and CCL22 in the BALF but not in our gene expression studies (Figure 3A). Furthermore, we verified expression of IGF-1 and IL-13 in the BALF using ELISA. IL-13 expression was undetected in the BALF of Tg+DOX animals (data not shown), whereas elevated IGF-1 was observed at day 14 after TGF-β1 expression (Figure 3C). IGF-1 levels in the BALF correlated with lung hydroxyproline content, an indirect measure of collagen deposition and fibrosis, consistent with its reported pro-fibrotic properties (Figure 3D) (27). Finally, several factors associated with acute injury, such as serum amyloid P-component and C-reactive protein, and vascular homeostasis, such as vascular endothelial growth factor, were elevated as early as Day 2 in the BALF of Tg+DOX mice (Figure 3A and supplemental data). To determine if BALF mononuclear cells could contribute to the production of soluble mediators listed above, we examined their expression using TaqMan low-density arrays. Expression of Ccl22, Spp1, Timp1, and most MIP family members (Ccl3, Ccl4, Ccl9, Cxcl2, and Ccl19) were up-regulated in F4/80+ cells isolated from Tg+DOX mice consistent with the elevated levels of these soluble mediators in the BALF (Figure 3B).


Figure 3
View larger version (33K):
[in this window]
[in a new window]

  		Figure 3. Identification of secreted mediators in the BALF of Tg+DOX mice. (A) Multianalyte profiling of soluble mediators in the BALF of Tg+DOX treated for 2 or 14 days compared with control mice. Average values from 10 animals per group (from two independent experiments) are represented in a heat map. (B) Gene expression analysis from sorted F4/80+ mononuclear cells. Three distinct pooled (n = 4 per pool) preparations of BALF cells were used to prepare RNA from Tg–DOX control mice and Tg+DOX mice treated for 2 days. Individual RNA preparations (n = 6) were prepared from BALF cells derived from Tg+DOX mice treated for 14 days. Values are means ± SEM. *P < 0.05 and **P < 0.01 versus Tg–DOX animals. (C) IGF-1 measurements in BALF of Tg–DOX and Tg+DOX mice treated for 14 days. Values are means ± SEM of n = 15 (–DOX) and n = 19 (+DOX). **P < 0.01 versus –DOX animals. (D) Correlation between IGF-1 levels and hydroxyproline (HP) content in Tg+DOX animals. **R2 = 0.4448 and P = 0.0018.

 
F4/80+ BALF Cells in Tg+DOX Mice Express Markers Consistent with an Alternatively Activated Macrophage Phenotype
Expression of several chemokines and mediators linked with the recruitment of inflammatory cells such as monocytes was observed as early as 2 days after hTGF-β1 overexpression. Infiltration of mononuclear cells in the lung parenchyma and bronchoalveolar space mirrored the increase in collagen content in the lung (Figures 1B and 1E, and data not shown). Therefore, we were interested in characterizing these cells and in examining their potential contribution to fibrosis. Morphologically, infiltrating mononuclear cells in the BALF of Day 14 Tg+DOX mice were indistinguishable from alveolar macrophages isolated from control mice as determined by Wright-Giemsa staining (Figure 4A). We used flow cytometry to extend our characterization of these cells. Infiltrating mononuclear cells expressed the macrophage marker F4/80, albeit the surface expression of F4/80 was relatively reduced when compared with control cells (data not shown). Furthermore, similar to alveolar macrophages from control mice, these cells stained CD11b–/lo CD11chi and did not express the dendritic cell marker 33D1 (data not shown). Increased surface expression of I-Ab (MHCII), CD80, and CD86 are hallmarks of "classically" activated macrophages. Expression of these markers in the infiltrating F4/80+ cells derived from the BALF of Tg+DOX mice was not increased when compared with control cells (Figures 4B–4D).


Figure 4
View larger version (33K):
[in this window]
[in a new window]

  		Figure 4. Phenotypic characterization of infiltrating mononuclear cells in the BALF of Tg+DOX mice. (A) Representative Wright-Giemsa staining performed on BALF cells from Tg–DOX and Tg+DOX mice treated for 14 days. (B–D) Flow cytometric analysis of gated F4/80+ BALF cells. (B) I-Ab (MHCII) surface expression. Isotype control stain is shown in gray, and specific I-Ab staining is shown in black for Tg–DOX and red for Tg+DOX mice. (C) Isotype control (gray line) and CD80 (black line) expression in Tg–DOX (left panel) and Tg+DOX mice (right panel). (D) Isotype control (gray line) and CD86 (black line) expression in Tg–DOX (left panel) and Tg+DOX mice (right panel). Data shown is representative of three independent experiments. For each experiment, a single preparation of pooled BALF cells from Tg–DOX animals and three separate pools (n = 3 per pool) of BALF cells from Tg+DOX mice were used.

 
To define the potential activation state of infiltrating F4/80+ mononuclear cells, we examined the expression of genes associated, at least in vitro, with classically, type II, and alternatively activated macrophages (2, 28). In contrast to classically activated macrophages, aaMac only minimally up-regulate MHCII and co-stimulatory molecules necessary for antigen presentation to T lymphocytes and express markers such as arginase, Retlna/b (Fizz1/2), and Chi3l3/4 (Ym1/2) (28, 29). Genes linked to alternately activated macrophages such as Arg1, Arg2, Retnla (Fizz1), Retnlb (Fizz2), Chi3l3 (Ym1), and Chi3l4 (Ym2) were up-regulated in lungs of Tg+DOX mice (Figure 5A). Furthermore, we examined the expression of selected genes in purified F4/80+ cells derived from the BALF of Tg+DOX mice. Similar to the gene signature observed in lung homogenates, we observed elevated expression of Arg1, Arg2, Fn1, and Igf1 in F4/80+ cells from Tg+DOX compared with control mice (Figures 5B, 5C, and E1). Elevated Arg1 and Arg2 gene expression correlated with increased arginase activity that was detected in BALF supernatant and F4/80+ cell extracts (Figures 5D and 5E). Taken together, these data suggest that infiltrating F4/80+ mononuclear cells may include aaMAcs.


Figure 5
View larger version (20K):
[in this window]
[in a new window]

  		Figure 5. Arginase activity is elevated in the lungs Tg+DOX mice. (A) Relative expression levels of genes associated with aaMac in lungs of Tg+DOX mice. (B) Relative Arginase 1 (Arg1) and (C) Arginase 2 (Arg2) mRNA expression levels normalized to expression levels of Gapdh in purified F4/80+ mononuclear cells isolated from the BALF of Tg–DOX and Tg+DOX after 2 or 14 days of treatment. Three distinct pooled preparations (n = 4 per pool) of BALF cells were used to prepare RNA from Tg–DOX control mice and Tg+DOX mice treated for 2 days. Individual RNA preparations (n = 6) were prepared from BALF cells derived from Tg+DOX mice treated for 14 days. Values are means ± SEM. *P < 0.05 compared with Tg–DOX animals. (D) Arginase activity in BALF from Tg–DOX and Tg+DOX treated for 2 and 14 days. Values are means ± SEM. n = 4 (–DOX), and n = 6 (+DOX). **P < 0.01 versus Tg–DOX animals. (E) Arginase activity measure in BALF cell lysates. Values are means ± SEM. n = 3. **P < 0.01 versus Tg–DOX animals.

 
Comparison of Lung Gene Expression Signatures Derived from hTGF-β1 Tg and Bleomycin-Induced Fibrosis Models Reveals a Common Signature
We compared the gene expression signature derived from Tg+DOX mice to selected published data obtained from bleomycin-treated animals (9, 10). To simplify our analyses, we consolidated the data obtained from kinetic experiments by obtaining the union of gene lists that are either up-regulated or down-regulated at any given time point. First, we compared the published bleomycin-induced gene signatures and identified 120 shared genes, comprising 18% and 35% of the gene signature in the studies by Haston and colleagues and Kaminski and colleagues, respectively (Table 3) (9, 10). Of these, 115 showed conserved directionality in gene expression changes, and five genes were anticorrelated (Table 3). In addition, Kaminski and colleagues identified approximately 52 genes that were up-regulated in C57bl/6 wild-type but not in fibrosis-resistant β6–/– mice after bleomycin treatment (9). These genes were intersected with the list of 157 "fibrosis-specific" up-regulated genes identified by Haston and colleagues (10). This filter did not result in additional enrichment because only 20 commonly regulated genes were identified (Table 3). We identified 367 and 129 genes that intersected with the studies of Haston and colleagues and Kaminski and colleagues comprising 55% and 38% of the gene signatures identified in these respective studies when compared with gene lists from Tg+DOX mice. Remarkably, the directionality of gene expression was conserved with 354 out of 367 genes and 116 out 129 genes being modulated similarly in these studies respectively. Equivalent data were observed at the individual Day 2 and Day 14 time points (refer to the online supplement). Finally, we performed three-way analysis to isolate the common gene signature from bleomycin and Tg+DOX studies and identified 47 up-regulated and 21 down-regulated genes (Figure E3 and supplemental data). Overall, up-regulated genes clustered in two canonical pathways, complement and coagulation factors and ECM–receptor interactions. In particular, up-regulated genes included growth factors such as Areg, Igf-1, Spp1, and Tgfbi; ECM components such as Tnc and Eln; proteases such as Catsk and Mmp14; and anti-proteases such as Serpine2. Down-regulated genes included antioxidants such as Gsta3 and Pon1 and multiple genes implicated in electron transport and metabolism processes.


View this table:
[in this window]
[in a new window]

  		TABLE 3. COMPARATIVE ANALYSIS OF GENE EXPRESSION SIGNATURES DERIVED FROM BLEOMYCIN-TREATED AND TG+DOX MICE

 
Comparative Analysis of Gene Expression Signatures Derived from Preclinical Fibrosis Models and Patients with IPF
To determine whether regulated genes in preclinical models of lung fibrosis are conserved in humans, we intersected gene signatures derived from bleomycin-treated and Tg+DOX mice with the data obtained from two independent studies using patients with IPF (11, 12). Only eight genes were shared in the Zuo and colleagues and Selman and colleagues IPF studies (Table 4). Furthermore, analysis of the human data revealed a small number of genes that intersected with data obtained form preclinical models with an overall trend for conserved directionality in gene expression changes observed with the up-regulated IPF gene set from the Selman and colleagues study (Table 4 and supplemental data) (11).


View this table:
[in this window]
[in a new window]

  		TABLE 4. COMPARATIVE ANALYSIS OF GENE EXPRESSION SIGNATURES DERIVED FROM PRECLINICAL MOUSE MODELS OF LUNG FIBROSIS AND PATIENTS WITH IDIOPATHIC PULMONARY FIBROSIS

 

   DISCUSSION
Top
 Abstract
 CLINICAL RELEVANCE
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
Two non-mutually exclusive pathogenic routes, epithelial cell injury and exaggerated Th2-dominated inflammatory responses, are proposed to contribute directly to excessive extracellular matrix deposition leading to tissue fibrosis (30). Nonetheless, the precise sequence of events and the nature of mechanisms contributing to fibrosis are poorly defined. In the current study, we exploited hTGF-β1 transgenic mice to specifically examine TGF-β–driven mechanisms contributing to lung fibrosis. Consistent with previously published observations, ectopic expression of hTGF-β1 in the lungs of transgenic mice resulted in a time-dependent inflammatory response characterized by massive infiltration of F4/80+ monocytic/macrophage-like cells, an increase in lung collagen content, and decreased lung compliance (3). Using gene expression profiling, we identified a common signature of genes that were similarly modulated by hTGF-β1 at Day 2 and Day 14 time points after DOX treatment. This common signature comprised the majority of genes modulated by hTGF-β1 and included genes functionally associated with tissue remodeling and ECM regulation, suggesting that some pro-fibrotic mechanisms are rapidly activated by hTGF-β1 before the recruitment of inflammatory cells. Nonetheless, we identified genes for which expression was temporally regulated by hTGF-β1. These were associated with key genetic programs regulating cell movement and invasiveness, inflammation, and organ remodeling and fibrosis.

We used real-time PCR to validate the data obtained by microarray hybridization and showed comparable gene expression levels for at least 40 modulated genes. Furthermore, we compared RNA levels of a subset of genes encoding for soluble mediators to their cognate protein levels in BALF and observed, overall, consistent modulation of mRNA and protein levels. Nonetheless, certain chemokines, such as CCL3, -4, and -22, were elevated in BALF in the absence of detectable RNA expression changes in lung homogenates. For such genes, it is plausible that a small cellular subset contributed to localized elevation of those factors. Alternately, these factors may be regulated by post-transcriptional mechanisms, which have been suggested to regulate expression of cytokines such as TNF-{alpha} (31). Taken together, these results highlight the importance of combining multiple analytical platforms in examining global biological processes.

A critical role for cell apoptosis in the lung, preceding the onset of inflammation and fibrosis, has been demonstrated in hTGF-β1 transgenic mice. In these studies, marked elevation of Egr1 and small but significant changes in Bax and Casp3 expression levels were observed shortly after ectopic hTGF-β1 induction (3, 32). Our results show that the expression of genes functionally associated with apoptosis were modestly modulated in the lungs of Tg+DOX mice at Day 2, such as Egr1, Casp3, and Casp4, or not modulated, such as pro-apoptotic members of the Bcl-2 family (Bax, Bid, Bak, and Bad). Furthermore, in preliminary studies, we did not observe increased apoptosis in situ in Tg+DOX mice treated with DOX for 2 days. Because of the transient nature of the reported apoptotic "wave," it is plausible that the discrepancy in our data reflects differences in the timing of the onset of apoptosis in these mice.

We observed a decrease in BALF levels of bioactive TGF-β levels in Day 14 (750 pg/ml) compared with Day 2 (10,000 pg/ml) Tg+DOX mice. Consistent with the decrease in hTGF-β protein levels, we noted a 7- to 10-fold decrease in hTGF-β1 mRNA levels in the lungs of Tg+DOX mice treated for 14 days compared with 2 days (Figure E4). Expression of hTGF-β1 in this construct is regulated by the Scgb1a1 gene (CC10) promoter. Expression of endogenous Scgb1a1 was reduced by at least 10-fold in the lungs of Day 14 animals (Figure E5), consistent with the concomitant decrease of hTGF-β1. Two nonmutually exclusive hypotheses, silencing of transgene expression and loss of CC10+ transgene expressing cells, may explain this decrease in hTGF-β1 expression. In the latter, in situ TUNEL assays did not reveal extensive loss of airway epithelial cells (data not shown). A decrease in hTGF-β1 expression levels in BALF was not reported in a previous study using these Tg mice (3). The discrepancy in our data may be due to the higher concentrations of DOX that we used and consequently higher levels of hTGF-β1 at Day 2 (10,000 pg/ml versus 2,000 pg/ml in the study by Lee and colleagues, Ref. 3) or to differences in the genetic background of these mice because we used mice rederived and backcrossed on a B6JBom background. More extensive analysis is needed to define the mechanisms leading to decrease hTGF-β1 expression in this model.

An important number of genes functionally associated with cellular movement were modulated as early as Day 2 after DOX treatment. Distinct mechanisms, in addition to the proliferation of local fibroblasts, have been proposed to contribute to increased fibroblast numbers at the sites of fibrotic lesions. EMT is a process characterized by the loss of epithelial local cell adhesion, increased cellular mobility, and adoption of a mesenchymal phenotype resulting in tissue invasion. Consistently, we observed elevated gene expression levels of Snai1, a transcriptional repressor of E-cadherin that promotes EMT (17). Recently, circulating mesenchymal fibrocytes, which, upon recruitment to the site of injury, differentiate into myofibroblasts under the influence of local growth factors, have been proposed to promote fibrosis (1, 33). The expression of chemokines and chemokine receptors implicated in the recruitment of fibrocytes such as Ccl1, Ccl12, and Ccr5 was elevated after ectopic TGF-β expression in the lungs of hTGF-β1 transgenic mice (14, 34). Finally, expression of multiple genes regulating fibroblast and epithelial cell traffic was observed. There was an increase in Spp1 (osteopontin) mRNA in lungs, which is a common feature in patients with IPF (35). In addition, we detect elevated expression of Has1 and 2 (hyaluronan synthase 1 and 2). Elevated hyaluronan, regulating cell traffic, inflammation, and ECM composition, precedes fibrosis and disease in several pathologies (36). Recently, Spp1-driven Has2 induction was shown to promote breast cancer malignancy, thereby linking common mechanisms implicated in ECM remodeling, cellular recruitment, and cancer invasiveness (37).

Massive infiltration of F4/80+ mononuclear cells was observed in the BALF of mice ectopically expressing hTGF-β1. A direct role for TGF-β1 in the recruitment of monocytes has been previously demonstrated (38). In addition, hTGF-β1 expression resulted in the induction of genes functionally associated with recruitment of leukocytes and inflammation at Day 2 after DOX treatment, a time point preceding the recruitment of inflammatory cells. These included chemokines such as Ccl2, 3, 4, and 7 and matrix metalloproteinases (MMPs), which can facilitate monocyte transmigration into the lung (38, 39). Further, ectopic hTGF-β1 expression resulted in the expression of several proinflammatory cytokines, such as IL-6, IL-11, and IL-13. Enforced expression of each of these cytokines was shown to promote lung remodeling and fibrosis (6, 21). We detected elevated IL-6 levels in the BALF of transgenic mice treated with DOX, consistent with data demonstrating induced IL-6 expression in TGF-β–treated monocytes.

Expression of "classic" activation markers, such as MHCII and B7 adaptor molecules, was not induced on the surface of infiltrating F4/80+ mononuclear cells compared with alveolar macrophages from control mice. Furthermore, expression of Arg1, Arg2, Chi3l4 (Ym2), and Retnla (Fizz1) was elevated in lungs of mice ectopically expressing hTGF-β1. Expression of these markers has been associated with development of aaMacs in vitro (2). Expression of aaMAc markers such as Arg1, Arg2, and Fn1 was elevated in purified F4/80+ mononuclear cells derived from Tg+DOX mice compared with control mice (Figures 5 and E1A). Moreover, we observed elevated arginase activity in the BALF of mice ectopically expressing hTGF-β1. Arginase is an enzyme that is implicated in the biosynthesis of proline, an abundant amino acid in collagen, and elevated arginase activity has been associated with tissue remodeling (40). In addition, we observed increased gene expression levels of Igf1 in isolated BAL macrophages derived from Day 14 Tg+DOX (Figure E1B) and observed a positive correlation between IGF-1 protein levels in BALF and collagen content in these mice (Figures 3C and 3D). A requirement for a Th2-biased cytokine environment favoring the development of aaMAc has been reported to induce IGF-1 expression by macrophages (27). Taken together, these results suggest the possibility that infiltrating F4/80+ mononuclear cells may comprise aaMac. Additional experiments are required to determine the contribution of macrophages to fibrosis in hTGF-β1 transgenic mice. Emerging literature supports a role for alternatively activated macrophages in preclinical models of fibrosis and clinical fibrotic diseases (41–43). In patients with IPF, elevated expression levels of alternatively activated macrophage markers, such as arginase I, IGF-1, and CCL18, were identified in BAL-derived macrophages, and CCL18 expression in these cells correlated negatively with airway function (42–44). Conditioned supernatants derived from alternatively activated macrophages increased collagen production by normal human lung fibroblasts in vitro, suggesting a pathological role for alternatively activated macrophages in IPF (43).

Deregulated ECM turnover is thought to be an important feature of fibrosis. We noted increased expression of multiple ECM genes as early as Day 2 after hTGF-β1. The latter is consistent with a role for TGF-β1 in directly inducing the expression of multiple ECM genes (3). Furthermore, TGF-β1 activated expression of growth factors, such as CTGF and Inhba, which have been demonstrated to synergize with TGF-β or to independently promote ECM production (45). In addition to elevated ECM synthesis, the balance between MMPs and tissue inhibitor of MMPs (TIMPs) is suggested to regulate the progression of fibrosis via the regulation of ECM turnover. Consistently, TIMPs gene expression is up-regulated in lungs from patients with IPF (46). We observed elevated expression of Timp1 and 4 in Tg+DOX mice. Timp1 expression was particularly elevated in F4/80+ mononuclear cells, suggesting a potential role for these cells in regulating ECM turnover. Finally, we note increased expression of several MMPs, including Mmp12, which has been recently genetically linked to TGF-β–driven lung fibrosis (32). Increased activity of selected MMPs is believed to promote fibrosis through recruitment of inflammatory cells by direct remodeling of the ECM or by releasing chemokines and other biologic modulators (32).

In studies using the cytotoxic drug bleomycin, distinct genetic programs regulating lung inflammation and fibrosis were identified by comparing wild-type mice with genetically engineered β6–/– mice that are protected from pulmonary fibrosis or by comparing mouse strains C57bl/6 and C3H, which vary in their susceptibility to bleomycin-induced pulmonary fibrosis, respectively (9, 10). We compared gene expression signatures reported in these studies with the genetic programs that are induced after the selective induction of bioactive hTGF-β1 in lungs of Tg mice. Intersecting genes showed almost exclusive conservation in the directionality of gene expression changes in these models. These data suggest common pathways leading to fibrosis and are consistent with genetic and pharmacological studies demonstrating an essential role for TGF-β in mediating fibrosis after the administration of bleomcin (7, 8). Nonetheless, we identified a large subset of modulated genes that were unique to each study. Multiple nonmutually exclusive factors may account for these "nonconcordant genes," such as differences in the administration route of bleomycin, differences in microarray chips and data analysis, and the induction of biological pathways, which are distinct in bleomycin-treated mice compared with Tg mice ectopically expressing high levels of hTGF-β1. Identification of robust gene signatures in human diseases such as IPF has been limited by small sample sizes, heterogeneity in biopsies, and the presence of potentially confounding factors such as smoking and therapy. Only a small number of modulated genes were shared between two independent IPF studies (Table 4). When compared with the mouse orthologs obtained from Tg+DOX mice, we observed a small number of genes, such as spp1, that showed conserved regulation in both species. Multiple genes induced by ectopic expression of hTGF-β1 in the lungs of transgenic mice were also induced by exogenous TGF-β in primary human fibroblasts, indicating conserved genetic pathways in human and mouse (data not shown). Therefore, it is important to recognize that the limited overlap in gene signatures between patients with IPF and preclinical models of disease may be consequent to the fact that these animal models may reflect acute lung injury rather than chronic and well established fibrosis (47). Indeed, genes associated with immune responses and inflammation are significantly associated with Day 2 and Day 14 Tg+DOX gene signatures. Furthermore, we note the down-regulation of Gsta3 gene expression in bleomycin-treated and Tg+DOX animals, consistent with a critical role for TGF-β in mediating acute lung injury via depletion of intracellular glutathione (8). Finally, previous studies demonstrated that removal of DOX from the drinking water of Tg+DOX animals for 14 days resulted in rapid disappearance of hTGF-β1 and reversal of fibrosis after 30 days (3). Herein, we extended on these observations by demonstrating the return to baseline expression levels of genes modulated by hTGF-β1 30 days after DOX removal, indicating that, in contrast to patients with IPF in whom fibrosis is driven by endogenous mechanisms, continuous expression of exogenous hTGF-β1 in lungs of Tg+DOX animals is required to sustain a profibrotic program (refer to supplemental data).

In conclusion, we identified multiple TGF-β–driven genetic programs that are consistent with mechanisms that directly regulate fibrosis (e.g., through modulation of ECM components and its turnover) and indirectly by recruiting mesenchymal cells, which potentially regulate tissue repair. More work is required to examine whether mesenchymal cells are recruited to the lungs of these mice and to define a functional role for alternately activated macrophages in regulating fibrosis. Finally, to improve the predictability of our animal models relative to human diseases, it is important to examine how gene expression changes are altered after prolonged activation of ectopic hTGF-β1 expression in lungs of transgenic mice and whether prolonged transgenic exposure results in the activation of endogenous pathways that are able to independently sustain histopathological alterations in lungs.


   Acknowledgments
 
The authors thank Dr. Haston for kindly providing raw CEL microarray data profiling; Denis Normandin, Simon Wong, Mariane Meilleur, and Étienne Daigneault for tissue collection; Carole Desroches for histological sections; Cindy Jones and Myriam Desbiens for cell counts; Karen Ng for genotyping; and Alex Therrien, Michael Crackower, and Francois Gervais for critical review of the manuscript.


   Footnotes
 
This work was supported by a fellowship from the Natural Sciences and Engineering Research Council of Canada (A.-M.P).

This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org

Originally Published in Press as DOI: 10.1165/rcmb.2007-0186OC on April 10, 2008

Conflict of Interest Statement: C.E.S. is a full-time employee of Merck & Co. and holds stock options in Merck & Co., Inc. that are worth approximately $9,000. A.M.P. post-doctoral fellow has been financed by Merck and co. I.W. is a full-time employee of Merc & Co. and holds stock options in Merck & Co., Inc. that are worth approximately $12,000. A.C. is a full-time employee of Merck & Co. and holds stock options in Merck & Co., Inc. that are worth approximately $1,000. A.A. is a full-time employee of Merck & Co. and holds stock options in Merck & Co., Inc. that are worth approximately $20,000. Y.B. is a full-time employee of Merck & Co. and holds stock options in Merck & Co., Inc. that are worth approximately $5,000. J.E. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. A.K. was a full-time employee of Merck and Co. from June 2005 to August 2006. A.K. does not hold stock options of Merck & Co., Inc. L.X. is a full-time employee of Merck & Co. and holds stock options in Merck & Co., Inc. that are worth approximately $5,000. J.P.M. is a full-time employee of Merck & Co. and holds stock options in Merck & Co., Inc. that are worth approximately $4,000.

Received in original form May 24, 2007

Accepted in final form March 4, 2008


   References
Top
 Abstract
 CLINICAL RELEVANCE
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Keane MP, Strieter RM, Belperio JA. Mechanisms and mediators of pulmonary fibrosis. Crit Rev Immunol 2005;25:429–463.[Medline]
  2. Gordon S. Alternative activation of macrophages. Nat Rev Immunol 2003;3:23–35.[CrossRef][Medline]
  3. Lee CG, Cho SJ, Kang MJ, Chapoval SP, Lee PJ, Noble PW, Yehualaeshet T, Lu B, Flavell RA, Milbrandt J, et al. Early growth response gene 1-mediated apoptosis is essential for transforming growth factor beta1-induced pulmonary fibrosis. J Exp Med 2004;200:377–389.[Abstract/Free Full Text]
  4. Sime PJ, Xing Z, Graham FL, Csaky KG, Gauldie J. Adenovector-mediated gene transfer of active transforming growth factor-beta1 induces prolonged severe fibrosis in rat lung. J Clin Invest 1997;100:768–776.[Medline]
  5. Wynn TA. Common and unique mechanisms regulate fibrosis in various fibroproliferative diseases. J Clin Invest 2007;117:524–529.[CrossRef][Medline]
  6. Lee CG, Homer RJ, Zhu Z, Lanone S, Wang X, Koteliansky V, Shipley JM, Gotwals P, Noble P, Chen Q, et al. Interleukin-13 induces tissue fibrosis by selectively stimulating and activating transforming growth factor beta(1). J Exp Med 2001;194:809–821.[Abstract/Free Full Text]
  7. Kapoun AM, Gaspar NJ, Wang Y, Damm D, Liu YW, O'young G, Quon D, Lam A, Munson K, Tran TT, et al. Transforming growth factor-beta receptor type 1 (TGFbetaRI) kinase activity but not P38 activation is required for TGFbetaRI-induced myofibroblast differentiation and profibrotic gene expression. Mol Pharmacol 2006;70:518–531.[Abstract/Free Full Text]
  8. Pittet JF, Griffiths MJ, Geiser T, Kaminski N, Dalton SL, Huang X, Brown LA, Gotwals PJ, Koteliansky VE, Matthay MA, et al. TGF-beta is a critical mediator of acute lung injury. J Clin Invest 2001;107:1537–1544.[Medline]
  9. Kaminski N, Allard JD, Pittet JF, Zuo F, Griffiths MJ, Morris D, Huang X, Sheppard D, Heller RA. Global analysis of gene expression in pulmonary fibrosis reveals distinct programs regulating lung inflammation and fibrosis. Proc Natl Acad Sci USA 2000;97:1778–1783.[Abstract/Free Full Text]
  10. Haston CK, Tomko TG, Godin N, Kerckhoff L, Hallett MT. Murine candidate bleomycin induced pulmonary fibrosis susceptibility genes identified by gene expression and sequence analysis of linkage regions. J Med Genet 2005;42:464–473.[Abstract/Free Full Text]
  11. Selman M, Pardo A, Barrera L, Estrada A, Watson SR, Wilson K, Aziz N, Kaminski N, Zlotnik A. Gene expression profiles distinguish idiopathic pulmonary fibrosis from hypersensitivity pneumonitis. Am J Respir Crit Care Med 2006;173:188–198.[Abstract/Free Full Text]
  12. Zuo F, Kaminski N, Eugui E, Allard J, Yakhini Z, Ben Dor A, Lollini L, Morris D, Kim Y, DeLustro B, et al. Gene expression analysis reveals matrilysin as a key regulator of pulmonary fibrosis in mice and humans. Proc Natl Acad Sci USA 2002;99:6292–6297.[Abstract/Free Full Text]
  13. Pisabarro MT, Leung B, Kwong M, Corpuz R, Frantz GD, Chiang N, Vandlen R, Diehl LJ, Skelton N, Kim HS, et al. Cutting edge: novel human dendritic cell- and monocyte-attracting chemokine-like protein identified by fold recognition methods. J Immunol 2006;176:2069–2073.[Abstract/Free Full Text]
  14. Moore BB, Murray L, Das A, Wilke CA, Herrygers AB, Toews GB. The role of CCL12 in the recruitment of fibrocytes and lung fibrosis. Am J Respir Cell Mol Biol 2006;35:175–181.[Abstract/Free Full Text]
  15. Capelli A, Di Stefano A, Gnemmi I, Donner CF. CCR5 expression and CC chemokine levels in idiopathic pulmonary fibrosis. Eur Respir J 2005;25:701–707.[Abstract/Free Full Text]
  16. Inoshima I, Kuwano K, Hamada N, Hagimoto N, Yoshimi M, Maeyama T, Takeshita A, Kitamoto S, Egashira K, Hara N. Anti-monocyte chemoattractant protein-1 gene therapy attenuates pulmonary fibrosis in mice. Am J Physiol Lung Cell Mol Physiol 2004;286:L1038–L1044.[Abstract/Free Full Text]
  17. Cano A, Perez-Moreno MA, Rodrigo I, Locascio A, Blanco MJ, del Barrio MG, Portillo F, Nieto MA. The transcription factor snail controls epithelial-mesenchymal transitions by repressing E-cadherin expression. Nat Cell Biol 2000;2:76–83.[CrossRef][Medline]
  18. Itano N, Atsumi F, Sawai T, Yamada Y, Miyaishi O, Senga T, Hamaguchi M, Kimata K. Abnormal accumulation of hyaluronan matrix diminishes contact inhibition of cell growth and promotes cell migration. Proc Natl Acad Sci USA 2002;99:3609–3614.[Abstract/Free Full Text]
  19. Tkach V, Tulchinsky E, Lukanidin E, Vinson C, Bock E, Berezin V. Role of the Fos family members, C-Fos, Fra-1 and Fra-2, in the regulation of cell motility. Oncogene 2003;22:5045–5054.[CrossRef][Medline]
  20. Reymond N, Imbert AM, Devilard E, Fabre S, Chabannon C, Xerri L, Farnarier C, Cantoni C, Bottino C, Moretta A, et al. DNAM-1 and PVR regulate monocyte migration through endothelial junctions. J Exp Med 2004;199:1331–1341.[Abstract/Free Full Text]
  21. Kuhn C III, Homer RJ, Zhu Z, Ward N, Flavell RA, Geba GP, Elias JA. Airway hyperresponsiveness and airway obstruction in transgenic mice: morphologic correlates in mice overexpressing interleukin (IL)-11 and IL-6 in the lung. Am J Respir Cell Mol Biol 2000;22:289–295.[Abstract/Free Full Text]
  22. Cuzzocrea S, Ronchetti S, Genovese T, Mazzon E, Agostini M, Di Paola R, Esposito E, Muia C, Nocentini G, Riccardi C. Genetic and pharmacological inhibition of GITR-GITRL interaction reduces chronic lung injury induced by bleomycin instillation. FASEB J 2007;21:117–129.[Abstract/Free Full Text]
  23. Wang SW, Oh CK, Cho SH, Hu G, Martin R, Demissie-Sanders S, Li K, Moyle M, Yao Z. Amphiregulin expression in human mast cells and its effect on the primary human lung fibroblasts. J Allergy Clin Immunol 2005;115:287–294.[CrossRef][Medline]
  24. Krein PM, Winston BW. Roles for insulin-like growth factor I and Transforming growth factor-beta in fibrotic lung disease. Chest 2002;122:289S–293S.[CrossRef][Medline]
  25. Counts DF, Evans JN, Dipetrillo TA, Sterling KM Jr, Kelley J. Collagen lysyl oxidase activity in the lung increases during bleomycin-induced lung fibrosis. J Pharmacol Exp Ther 1981;219:675–678.[Abstract/Free Full Text]
  26. Lemaire R, Korn JH, Shipley JM, Lafyatis R. Increased expression of type I collagen induced by microfibril-associated glycoprotein 2: novel mechanistic insights into the molecular basis of dermal fibrosis in scleroderma. Arthritis Rheum 2005;52:1812–1823.[CrossRef][Medline]
  27. Wynes MW, Riches DW. Induction of macrophage insulin-like growth factor-I expression by the Th2 cytokines IL-4 and IL-13. J Immunol 2003;171:3550–3559.[Abstract/Free Full Text]
  28. Edwards JP, Zhang X, Frauwirth KA, Mosser DM. Biochemical and functional characterization of three activated macrophage populations. J Leukoc Biol 2006;80:1298–1307.[Abstract/Free Full Text]
  29. Raes G, Noel W, Beschin A, Brys L, de Baetselier P, Hassanzadeh GH. FIZZ1 and Ym as tools to discriminate between differentially activated macrophages. Dev Immunol 2002;9:151–159.[CrossRef][Medline]
  30. Lee CG, Kang HR, Homer RJ, Chupp G, Elias JA. Transgenic modeling of transforming growth factor-beta(1): role of apoptosis in fibrosis and alveolar remodeling. Proc Am Thorac Soc 2006;3:418–423.[Abstract/Free Full Text]
  31. Anderson P, Phillips K, Stoecklin G, Kedersha N. Post-transcriptional regulation of proinflammatory proteins. J Leukoc Biol 2004;76:42–47.[Abstract/Free Full Text]
  32. Kang HR, Cho SJ, Lee CG, Homer RJ, Elias JA. Transforming growth factor (TGF)-Beta1 stimulates pulmonary fibrosis and inflammation via a Bax-dependent, Bid-activated pathway that involves matrix metalloproteinase-12. J Biol Chem 2007;282:7723–7732.[Abstract/Free Full Text]
  33. Quan TE, Cowper SE, Bucala R. The role of circulating fibrocytes in fibrosis. Curr Rheumatol Rep 2006;8:145–150.[Medline]
  34. Ishida Y, Kimura A, Kondo T, Hayashi T, Ueno M, Takakura N, Matsushima K, Mukaida N. Essential roles of the CC chemokine ligand 3-CC chemokine receptor 5 axis in bleomycin-induced pulmonary fibrosis through regulation of macrophage and fibrocyte infiltration. Am J Pathol 2007;170:843–854.[Abstract/Free Full Text]
  35. Pardo A, Gibson K, Cisneros J, Richards TJ, Yang Y, Becerril C, Yousem S, Herrera I, Ruiz V, Selman M, et al. Up-regulation and profibrotic role of osteopontin in human idiopathic pulmonary fibrosis. PLoS Med 2005;2:e251.[CrossRef][Medline]
  36. Noble PW, Jiang D. Matrix regulation of lung injury, inflammation, and repair: the role of innate immunity. Proc Am Thorac Soc 2006;3:401–404.[Abstract/Free Full Text]
  37. Cook AC, Chambers AF, Turley EA, Tuck AB. Osteopontin induction of hyaluronan synthase 2 expression promotes breast cancer malignancy. J Biol Chem 2006;281:24381–24389.[Abstract/Free Full Text]
  38. Li MO, Wan YY, Sanjabi S, Robertson AK, Flavell RA. Transforming growth factor-beta regulation of immune responses. Annu Rev Immunol 2006;24:99–146.[CrossRef][Medline]
  39. Ping G, Wang JM, Zack Howard OM, Oppenheim JJ. Biology of chemokines: lymphocyte trafficking in health and disease. Basel: Birkhäuser; 2006.
  40. Ricciardolo FL, Zaagsma J, Meurs H. The therapeutic potential of drugs targeting the arginase pathway in asthma. Expert Opin Investig Drugs 2005;14:1221–1231.[CrossRef][Medline]
  41. Duffield JS, Forbes SJ, Constandinou CM, Clay S, Partolina M, Vuthoori S, Wu S, Lang R, Iredale JP. Selective depletion of macrophages reveals distinct, opposing roles during liver injury and repair. J Clin Invest 2005;115:56–65.[CrossRef][Medline]
  42. Mora AL, Torres-Gonzalez E, Rojas M, Corredor C, Ritzenthaler J, Xu J, Roman J, Brigham K, Stecenko A. Activation of alveolar macrophages via the alternative pathway in herpesvirus-induced lung fibrosis. Am J Respir Cell Mol Biol 2006;35:466–473.[Abstract/Free Full Text]
  43. Prasse A, Pechkovsky DV, Toews GB, Jungraithmayr W, Kollert F, Goldmann T, Vollmer E, Muller-Quernheim J, Zissel G. A vicious circle of alveolar macrophages and fibroblasts perpetuates pulmonary fibrosis via CCL18. Am J Respir Crit Care Med 2006;173:781–792.[Abstract/Free Full Text]
  44. Uh ST, Inoue Y, King TE Jr, Chan ED, Newman LS, Riches DW. Morphometric analysis of insulin-like growth factor-I localization in lung tissues of patients with idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 1998;158:1626–1635.[Abstract/Free Full Text]
  45. Rodgarkia-Dara C, Vejda S, Erlach N, Losert A, Bursch W, Berger W, Schulte-Hermann R, Grusch M. The activin axis in liver biology and disease. Mutat Res 2006;613:123–137.[CrossRef][Medline]
  46. Pardo A, Selman M. Matrix metalloproteases in aberrant fibrotic tissue remodeling. Proc Am Thorac Soc 2006;3:383–388.[Abstract/Free Full Text]
  47. Borzone G, Moreno R, Urrea R, Meneses M, Oyarzun M, Lisboa C. Bleomycin-induced chronic lung damage does not resemble human idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2001;163:1648–1653.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 J. Immunol.Home page
S. L. Lin, A. P. Castano, B. T. Nowlin, M. L. Lupher Jr., and J. S. Duffield
Bone Marrow Ly6Chigh Monocytes Are Selectively Recruited to Injured Kidney and Differentiate into Functionally Distinct Populations
J. Immunol., November 15, 2009; 183(10): 6733 - 6743.
[Abstract] [Full Text] [PDF]

Home page
 J. Leukoc. Biol.Home page
R. D. Stout, S. K. Watkins, and J. Suttles
Functional plasticity of macrophages: in situ reprogramming of tumor-associated macrophages
J. Leukoc. Biol., November 1, 2009; 86(5): 1105 - 1109.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Online Supplement
Right arrow All Versions of this Article:
2007-0186OCv1
39/3/324    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Pulichino, A.-M.
Right arrow Articles by Sayegh, C. E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Pulichino, A.-M.
Right arrow Articles by Sayegh, C. E.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Proc. Am. Thorac. Soc. Am. J. Respir. Crit. Care Med.
Copyright © 2008 American Thoracic Society.