Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Lung repair after injury is characterized by an orderly progression of events to re-establish an intact alveolar-capillary interface. The initial lung insult triggers a series of responses including inflammation, epithelial and mesenchymal cell proliferation and migration, angiogenesis, extracellular matrix (ECM) deposition, apoptosis, and eventually restoration of the normal alveolar architecture. Throughout this process, ECM components are deposited, removed, or remodeled in the alveolar and interstitial compartments in an orderly manner. Degradation of the ECM through the action of various proteases removes damaged cells and matrix components, and enables cell migration and neovascularization to restore the tissue architecture (1). The matrix metalloproteinase (MMP) gene family, also known as matrixins, is thought to make significant contributions to this process. As a group, the matrixins are capable of degrading all components of the ECM (reviewed in Ref. 2). Interstitial collagenase cleaves fibrillar types I, II, III and X collagens. Gelatinases of 72 and 92 kD digest native types IV, V, VII, and X collagens; denatured types I, II, and III collagen; fibronectin; and elastin. Stromelysins-1 and -2 and matrilysin cleave fibronectin, laminin, entactin, proteoglycans, and pepsin-sensitive collagen regions (1).

Increasing evidence indicates that MMP activity has diverse biologic consequences. Although a primary function of MMPs is the cleavage and removal of matrix molecules, MMPs can also attack nonmatrix proteins to produce biologic effects such as tumor growth (3). MMP activity is essential for the initiation of epithelial cell migration on the ECM (4) and modulates epithelial cell apoptosis (5). In addition, MMPs participate in the process of neutrophil accumulation at sites of acute inflammation (6, 7). Accordingly, the expression or activation of MMPs may be crucial in modulating the inflammation and the local accumulation and degradation of the ECM in response to tissue injury.

Strict regulation of MMP activity is necessary to maintain tissue homeostasis as well as to effect tissue remodeling. Transcription of MMP genes is induced by growth factors, inflammatory cytokines, cell-matrix interactions, and cell-cell interactions (1). MMP gene expression is inhibited by transforming growth factor (TGF)-, retinoic acids, and glucocorticoids. MMPs are released from the cell as inactive zymogens that are believed to be activated in vivo by tissue or plasma proteases. An additional level of MMP regulation is provided by the metalloproteinase inhibitors ₂-macroglobulin and tissue inhibitors of metalloproteinases (TIMPs) which block the proteolytic activity of MMPs.

TIMPs are the major endogenous regulators of MMP activities in the tissue microenvironment. Four homologous TIMPs have been well characterized: TIMP-1, -2, -3, and -4 (reviewed in Ref. 2) (8). Each inhibitor noncovalently binds to the MMP with 1:1 stoichiometry. Members of the TIMP family are distinguished by differences in their efficiency of MMP inhibition, in their transcriptional regulation, and in their patterns of expression. In addition, TIMPs have been shown to possess biologic functions that are independent of MMP-inhibitory activity, including stimulation of cell proliferation (1, 9), induction or inhibition of apoptosis (10, 11), and induction of MMP expression in vitro (12).

The mechanisms that regulate MMP activity in the lung in response to injury are largely undefined. To restrict matrix remodeling to areas of lung damage and to avoid destruction of healthy tissue, the proteolytic activities of MMPs must be tightly controlled. Members of the TIMP gene family could make crucial contributions to regulating the actions of MMPs in the injured lung. Further, the altered expression of TIMPs in the pulmonary microenvironment could have important biologic effects independent of their ability to inactivate MMPs. To investigate the contribution of TIMP-1 and TIMP-2 in the pathogenesis of lung fibrosis, we examined the expression of these metalloproteinase inhibitors after bleomycin-induced lung injury in mice. We found that bleomycin injury results in a significant and durable increase in both TIMP-1 and TIMP-2 within the alveolar compartment. We also found that the expression of TIMP-1 and TIMP-2 is differentially regulated in the injured lung. TIMP-1 messenger RNA (mRNA) and protein are markedly increased in response to lung injury whereas the increase in TIMP-2 protein is unaccompanied by a change in TIMP-2 mRNA levels. TIMP-1 gene expression is spatially restricted to areas of lung damage, and both lung parenchymal cells and inflammatory cells located within the areas of tissue damage serve as in vivo sources of TIMP-1 in the injured lung. Our results implicate TIMP-1 and TIMP-2 as potentially important regulatory proteins early in the inflammatory and fibrotic responses to bleomycin-induced lung injury.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Bleomycin-Induced Lung Injury

Specific pathogen-free 8-wk-old male C57BL/6 mice weighing 24.6 ± 0.3 g (mean ± standard error [SE]) received a single dose of 0.0035 U/g of bleomycin sulfate (Pharmacia, Inc., Kalamazoo, MI) in 50 µl of sterile saline via a tracheostomy under intraperitoneal avertin anesthesia (13). Control mice received saline alone. The bleomycin dose used was consistently shown to produce pulmonary fibrosis with a mortality rate of < 10% in preliminary experiments with mice of similar genetic background.

At 2, 4, 7, 14, and 28 d after instillation, the mice were killed by exsanguination under deep anesthesia and the lungs harvested as previously described (13). In brief, the lungs were exposed by a midthoracotomy incision, and the pulmonary arteries were perfused with ribonuclease (RNAse)-free phosphate-buffered saline (PBS). The right lung was isolated with a ligature at the right hilum, resected, rinsed in RNAse-free PBS, and finely minced. The minced lung was divided into two aliquots, one each for tissue RNA and protein extraction. Each aliquot of the minced right lung was weighed, frozen in liquid nitrogen, and stored at 70°C for further analysis. The left lung was inflated with RNAse-free, 4% neutral buffered paraformaldehyde instilled at 30 cm H₂O pressure through the trachea for 120 min. The trachea was then tied and the lung immersed in the RNAse-free, 4% buffered paraformaldehyde for 24 h before embedding in paraffin. In certain experiments, bronchoalveolar lavage (BAL) was performed using 1 ml of sterile PBS before lung harvest.

RNA Isolation and Quantification

Total cellular RNA was isolated from the frozen tissue by a modification of the method of Madtes and colleagues with cesium choride density gradient centrifugation (13). Total cellular RNA isolated from phorbol myristate acetate (PMA)-stimulated murine epidermal fibroblast cell line 3T3 or parathyroid hormone- treated rat uterine fibroblasts was used as a positive control. RNA isolated from mouse liver was used as a negative control.

Probes

Complementary DNA (cDNA) was radiolabeled for use in Northern analysis. A murine TIMP-1 cDNA probe was prepared from the BamHI-HindIII restriction digest of clone pBSmouseTIMP-1, which contains the entire coding region and polyadenylation signal sequence of mouse TIMP-1 (provided by Dr. Dylan Edwards, University of Calgary, Calgary, AB, Canada). Murine TIMP-2 and TIMP-3 cDNAs were also provided by Dr. Edwards (14, 15). Murine interstitial collagenase (MMP-13) cDNA was provided by Dr. Howard Welgus, Washington University, St Louis, MO (16). Murine collagen 1(I) cDNA was provided by Dr. Malcolm Collins, University of Washington, Seattle, WA. cDNA inserts were random prime-labeled to a specific activity of 1 × 10⁹ counts per min (cpm)/µg with [-³²P]deoxycytidine triphosphate (NEN, Wilmington, DE) using a commercially available kit (Boehringer Mannheim, Indianapolis, IN) (17).

For in situ hydridization, a murine TIMP-1 complementary RNA (cRNA) probe was prepared by subcloning the XbaI-BamHI restriction digest of clone pBSmouseTIMP-1 into pBluescript SK downstream to the T₇ promoter. In vitro-transcribed antisense and sense RNA probes were labeled with [-³³P] uridine triphosphate (UTP) as described (17). The antisense TIMP-1 RNA probe, 399 bases in length, was transcribed from template DNA linearized with XbaI. Sense control probe, 363 bases in length, was generated from template DNA linearized with EcoRI.

Northern Analysis

Northern analysis was performed as previously described (17). In brief, total cytoplasmic RNA (20 µg/lane) was electrophoresed through 1% agarose/formaldehyde gels and transferred to nylon membranes (Nytran; Schleicher & Schuell, Keene, NH). The membranes were hybridized with the appropriate ³²P-labeled cDNA probe, 2 × 10⁶ cpm of probe per milliliter of hybridization solution (0.125 M Na₂HPO₄, 0.25 M NaCl, 7% [wt/vol] sodium dodecyl sulfate [SDS], 1 mM ethylenediaminetetraacetic acid [EDTA], 50% [vol/vol] foramide, 10% [wt/vol] polyethylene glycol [8,000], and 0.25 mg/ml sonicated denatured salmon sperm DNA) at 42°C for 18 h. The membranes were washed in 2× saline sodium citrate (SSC)/0.1% (wt/vol) SDS at room temperature (RT) for 15 min, twice in 0.2× SSC/0.1% SDS at 60°C for 30 min, and a final room temperature wash in 0.1× SSC/0.01% SDS. Previously hybridized nylon membranes were stripped of probe and hybridized with a 28S ribosomal RNA (rRNA) cDNA probe or an 18 rRNA oligonucleotide probe as previously described (17).

Quantitation of mRNA and Autoradiography

Quantitation of mRNA was performed as previously described (17). Autoradiographs of the hybridized membranes were made by exposing PhosphorImager storage plates at room temperature for 18 to 72 h, as needed to provide adequate intensity of bands. The plates were scanned with a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Relative quantities of hybridized probe in each band were determined using Image Quant software (Molecular Dynamics). The signal intensity for the mRNA bands of interest were divided by the signal intensity of 28S or 18S rRNA bands in the same lanes to control for variations in the quantity of RNA loaded in each lane. Conventional autoradiographs of the hybridized membranes were made by exposing Kodak XAR-2 film at 70°C with Cronex (Cronex Lightning Plus; DuPont, Wilmington, DE) intensifying screens for 7 to 14 d.

In Situ Hybridization

In situ hybridization was performed as described by Strandjord and colleagues (18). Lung sections of 5 µm, fixed in 4% (wt/vol) paraformaldehyde in PBS, were deparaffinized and rehydrated sequentially in graded ethanol. The sections were postfixed with 4% (wt/vol) paraformaldehyde at 4°C for 30 min, permeabilized in 20 µg/ml proteinase K (Boehringer Mannheim) at RT for 7.5 min, postfixed with 4% paraformaldehyde at 4°C for 5 min, and then acetylated in 0.1 M triethanolamine buffer/0.25% acetic anhydride for 10 min at RT before dehydrating.

The treated sections were prehybridized in a solution of 50% (vol/vol) deionized formamide, 0.6 M NaCl, 90 mM Tris-HCl (pH 8), 10 mM EDTA, 1× Denhardt's solution (0.02% [wt/vol] Ficoll-400, 0.02% [wt/vol] polyvinylpyrolidone, and 0.1% [wt/vol] bovine serum albumin fraction V), 500 µg/ml sheared and denatured salmon-sperm DNA, and 500 µg/ml denatured yeast transfer RNA (tRNA) for 4 h at RT in a moist chamber containing 50% (vol/vol) formamide and 4× SSC (1× SSC is 0.15M NaCl/ 0.015 M Na citrate). The prehybridization solution was removed and the sections were hybridized with a denatured [³³P]UTP-labeled sense- or antisense-oriented TIMP-1 cRNA probe at a concentration of 8 × 10⁶ cpm/ml in a solution of 50% (vol/vol) deionized formamide, 0.6 M NaCl, 90 mM Tris-HCl (pH 8), 10 mM EDTA, 1× Denhardt's solution, 100 µg/ml sheared and denatured salmon-sperm DNA, 83 µg/ml denatured yeast tRNA, 10% (wt/vol) dextran sulfate, 0.1% (wt/vol) SDS, and 10 mM dithiothreitol (DTT), at 55°C overnight in a moist chamber.

The hybridized sections were washed briefly with a solution of 50% formamide, 1× SSC, and 10 mM DTT at RT. They were then washed in the same solution at 55°C for 30 min, followed by a wash in a solution of 0.5× SSC at RT for 30 min. Nonhybridized RNA probe was digested by RNAse A (20 µg/ml) for 30 min at 37°C. The RNAse-treated slides were washed twice with 3.5× SSC at RT, and finally with 0.1× SSC at 65°C for 2 h. The washed sections were dehydrated through ethanol solutions containing 0.3 M sodium acetate, and were subsequently air-dried. The slides were dipped in Kodak NTB-2 emulsion diluted 1:1 with water, and were air-dried. They were exposed for 4 wk at 4°C, developed with Kodak D-19 developer, and counterstained with hematoxylin. Bright- and darkfield color photomicrographs were produced with a Nikon E600 photomicroscope and Kodak Ektachrome 64T transparency film. The slides were scanned and digitalized on a Kodak photo CD. The digitalized images were then processed by means of Adobe Photoshop for Windows 5.0 software (Adobe Systems, Inc. Mountain View, CA). Identical parameters were used for all photomicrographs and the images were printed on a Phaser II SDX dye sublimation printer (Tektronix, Beaverton, OR)

Immunohistochemistry

Mouse-lung macrophages were identified by immunohistochemistry using an affinity-purified rat monoclonal immunoglobulin (Ig) G_2a antibody, BM8 (Bachem Bioscience, King of Prussia, PA). An avidin-biotin complex (ABC) peroxidase technique was used as previously described (17). In brief, 5-µm sections of lung fixed with 4% (wt/vol) paraformaldehyde were deparaffinized and rehydrated. Endogenous peroxidase activity was blocked by incubation of the sections in 0.6% hydrogen peroxide in 80% methanol at RT for 15 min. The sections were incubated overnight at 4°C with rat antimouse macrophage antibody at a final concentration of 2 µg/ml with 2% (vol/vol) rabbit serum (Vector Laboratories, Burlingame, CA) in PBS. Primary antibody was detected with biotinylated rabbit antirat Ig (Vector Laboratories) at a 1:200 dilution. Bound antibody was visualized with ABC peroxidase (Vector Laboratories). The sections were counterstained with methyl green. Color photomicrographs were produced with a Nikon E600 photomicroscope and processed as described earlier. As a negative control, adjacent serial sections were stained in the absence of primary antibody.

Protein Extraction and Western Analysis

Lung protein was isolated by the method of Webb and associates (19). Frozen lung samples were pulverized under liquid nitrogen and then lysed in 50 mM Tris (pH 7.6), 1% SDS, 1 µM phenylmethylsulfonyl fluoride, and 10 µg/ml leupeptin for 24 h at 4°C. The lysate was sheared by passage through a 21-gauge needle and then centrifuged at 10,300 × g, 4°C, for 10 min. The supernatant was removed and stored at 70°C. Protein concentration was determined by the bicinchoninic acid assay (Pierce, Rockford, IL) as described by the manufacturer. Lung protein (10 µg per lane) or BAL fluid (BALF) (5 µl per lane) were electrophoresed under reduced conditions on 12% SDS-polyacrylamide gel and transferred to a nitrocellulose filter (Hybond-C Extra, Amersham Life Sciences, Arlington Heights, IL). The nitrocellulose filter was incubated with either murine monoclonal antihuman TIMP-1 antibody (1 µg/ml; Oncogene Research Products, Cambridge, MA) or murine monoclonal antihuman TIMP-2 antibody (1 µg/ml; Oncogene Research Products) overnight at 4°C. The filter was incubated with sheep antimouse IgG (Fab)₂-horseradish peroxidase (1:5,000; Amersham Life Sciences) in blotto for 1 h at 22°C. The secondary antibody was reacted with 5-amino-2, 3-dihydro-1, 4-phthalazinedione (SuperSignal West Dura; Pierce) substrate and the reaction products were visualized on XAR-5 autoradiographic film. The exposed film was scanned using an Epson 636 scanner and relative quantities of TIMP-1 or TIMP-2 immunoreactive protein in each lane were determined using Image Quant software (Molecular Dynamics).

Reverse Zymography

TIMP reverse zymography of lung protein (30 µg) or BALF (40 µl) was performed by electrophoresis on 10% SDS-polyacrylamide gel containing 1 mg/ml gelatin and 160 ng/ml recombinant human gelatinase A (Oncogene Research Products) as previously described (20). The gel was washed in 2.5% Triton X-100 for 3 h and then incubated for 18 h in 50 mM Tris (pH 7.5), 200 mM NaCl, 5 mM CaCl₂, and 0.02% Brij-35 at 37°C before staining with Coomassie blue.

Data Presentation and Statistical Analysis

At each fixed day, the treatment and control groups were compared using the two-sample t test assuming unequal variances (SPSS 6.0; SPSS, Inc., Chicago, IL). The mean value of the ratio of TIMP or metalloproteinase to 18S rRNA signal intensity for the test and control animals at each time point was calculated. The mean value of this ratio for the bleomycin-treated animals was expressed as a percentage of the mean value of the control animals at that time point. The standard error of the ratio of treatment and control means was calculated using the delta method where the SE of the ratio is [(1/B²) × (S_A)²/N_A + A²/B⁴ × (S_B)²/ N_B]^1/2; A and B are the group means; S is the group standard deviation; and N is the number of observations (17) .

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

TIMP and MMP mRNA in Normal and Bleomycin-Injured Mouse Lung Tissue

Total cellular RNA extracted from the lungs of control and bleomycin-injured mice were examined for the presence of TIMP-1, -2, and -3 as well as interstitial collagenase (MMP-13) and procollagen 1(I). TIMP-1 mRNA was detected at low levels as a 0.8 Kb message in the lungs of control animals (Figure 1). TIMP-1 gene expression was strikingly increased at Day 2 and remained elevated for at least 4 wk after bleomycin instillation. Lung steady-state TIMP-1 mRNA levels were 1,470% (n = 5, P < 0.05), 1,031% (n = 5, P < 0.05), 807% (n = 5, P < 0.05), 695% (n = 5, P < 0.05), and 979% (n = 5, P < 0.05) of control values at days 2, 4, 7, 14, and 28, respectively, after bleomycin exposure (Figures 1 and 2). TIMP-2 transcripts were detected as two distinct bands at 1.0 and 3.5 Kb in control lungs. Lung steady-state mRNA levels for TIMP-2 were unchanged during the first 4 wk after bleomycin injury. TIMP-3 mRNA was identified at 4.5 Kb in the lungs of control mice and its steady-state mRNA level was not altered after bleomycin administration.

View larger version (86K): [in this window] [in a new window]   Figure 1.   Steady-state mRNA levels of TIMP-1, TIMP-2, and TIMP-3 in lung after bleomycin injury. Northern analysis of mouse lung RNA isolated at 4 d after the intratracheal administration of bleomycin or saline (controls). Cellular RNA (20 µg) isolated from individual animals was electrophoresed, blotted onto a nylon membrane, and then hybridized with 32P-labeled TIMP cDNAs or 32P-end labeled 18S oligonucleotide. After high-stringency washes, the blots were exposed to XAR-5 film. Cellular RNA (20 µg) from mouse liver and PMA-treated NIH3T3 cells served as negative and positive controls, respectively.

View larger version (18K): [in this window] [in a new window]   Figure 2.   Temporal changes in TIMP-1, TIMP-2, and TIMP-3 steady-state mRNA levels after bleomycin administration. Mean values (± SE) for the ratio of signal intensities of tissue inhibitor of metalloproteinase/18S for bleomycin-injured mice are expressed as a percentage of the mean values for control (saline-treated) animals. TIMP-1 ( filled diamonds), TIMP-2 (open triangles), TIMP-3 (open circles). Five bleomycin-injured and five control animals were analyzed at each time point; *P < 0.05.

Interstitial collagenase (MMP-13) mRNA was detected at low levels as a 2.9 Kb transcript in control lungs, and steady-state mRNA levels of this metalloproteinase were unchanged by bleomycin administration at all days tested (Figure 3). Procollagen 1(I) transcripts were detected as two distinct bands at 4.8 and 7.0 Kb in control lungs as previously reported (21). Lung steady-state mRNA levels for procollagen 1(I) increased to 313% (n = 5, P < 0.05) and 245% (n = 5, P < 0.05) of control values at Days 7 and 14, respectively, after bleomycin injury.

View larger version (43K): [in this window] [in a new window]   Figure 3.   Interstitial collagenase and procollagen 1(I) steady-state mRNA levels in lung after bleomycin injury. Northern analysis of mouse lung RNA isolated 7 d after the intratracheal administration of bleomycin or saline. Cellular RNA (20 µg) isolated from individual animals was electrophoresed and hybridized with 32P-labeled MMP-13 or procollagen 1(I) cDNAs or with 32P-labeled 18S oligonucleotide as described in Figure 1. Cellular RNA (20 µg) from parathyroid hormone-stimulated rat uterine fibroblasts (UTR) and PMA-treated NIH 3T3 cells serve as positive controls.

In Situ Hybridization

To determine the spatial distribution and cellular sources of TIMP-1, in situ hybridization was performed on bleomycin-injured and saline-treated lungs. Using a specific [³³P]-labeled antisense riboprobe, we found infrequent alveolar mononuclear cells expressing low levels of TIMP-1 mRNA in control lungs (Figure 4A). Specificity of TIMP-1 hybridization was confirmed using TIMP-1 riboprobe transcribed in the sense direction (Figure 4B). In contrast to normal lung, TIMP-1 mRNA was markedly increased in a distinct distribution at Day 4 after bleomycin administration. TIMP-1 mRNA was detected primarily in perivascular and periairway regions of the injured lung (Figures 4C and 4D). In areas of lung remote from bleomycin damage, no appreciable increase in TIMP-1 mRNA was detected (Figures 4E and 4F). TIMP-1 transcripts were most prominent in mononuclear inflammatory cells within the areas of tissue damage (Figure 4G). TIMP-1 mRNA was also identified in alveolar interstitial cells adjacent to areas of inflammatory cell accumulation (Figure 4H). In comparison, TIMP-1 mRNA was virtually absent from epithelial cells lining the airway lumen (Figure 4D). Immunohistochemistry performed on adjacent serial sections of bleomycin-injured lung with a monoclonal antibody against murine macrophages suggested that pulmonary macrophages may serve as a cellular source of TIMP-1 mRNA in vivo (Figures 4I and 4J), however the cell type has not been unambiguously determined by this method.

View larger version (78K): [in this window] [in a new window]   Figure 4.   Localization of TIMP-1-expressing cells by in situ hybridization in bleomycin-injured and control mouse lung. (A) Mouse lung at Day 4 after saline instillation hybridized with 33P-labeled antisense riboprobe to murine TIMP-1 shows only background activity, demonstrating the low level of TIMP-1 gene expression in normal lung. a = airway (darkfield illumination; original magnification: ×25). (B) A section of injured mouse lung at Day 4 after bleomycin instillation hybridized with 33P-labeled sense riboprobe shows only background activity, compared with specific hybridization by the TIMP-1 antisense riboprobe of a serial tissue section (D). a = airway, v = blood vessel (darkfield illumination; original magnification: ×25. (C) Mouse lung at Day 4 after bleomycin instillation stained with hematoxylin and eosin shows a focus of injury in the periairway region (arrow). (Brightfield illumination, original magnification: ×25.) (D) TIMP-1 mRNA-positive cells are found in the area of injury on a serial tissue section adjacent to C using 33P-labeled antisense riboprobe to TIMP-1. Note the virtual absence of TIMP-1 message in the epithelial cells lining the airway lumen. (Darkfield illumination; original magnification: ×25.) (E) Mouse lung at Day 4 after bleomycin instillation stained with hematoxylin and eosin shows an area remote from bleomycin damage. (Brightfield illumination; original magnification: ×25.) (F ) A serial tissue section adjacent to E hybridized with 33P-labeled antisense riboprobe to TIMP-1 shows only background activity in the area of lung remote from bleomycin damage. (Darkfield illumination; original magnification: ×25.) (G) TIMP-1 localizes to mononuclear cells within the alveolar lumen (arrows) at Day 4 after bleomycin instillation using 33P-labeled antisense riboprobe to TIMP-1. (Brightfield illumination, original magnification: ×250.) (H ) TIMP-1 mRNA is found in spindle-shaped cells (arrowheads) within the alveolar septa at Day 4 after bleomycin instillation using 33P-labeled antisense riboprobe to TIMP-1. (Brightfield illumination; original magnification: ×250.) (I ) TIMP-1 mRNA is localized to the cytoplasm of alveolar cells (arrows) at Day 4 after bleomycin treatment using 33P-labeled antisense riboprobe to TIMP-1. (Brightfield illumination; original magnification: ×500.) (J) Immunostaining with rat antimouse macrophage monoclonal antibody on a serial section adjacent to I suggests that pulmonary macrophages (arrow) may be a cellular source of TIMP-1 mRNA in the injured lung. (Brightfield illumination; original magnification: ×500.)

It is interesting to note that at Day 2 after bleomycin administration, TIMP-1 mRNA was detected in alveolar septal cells and mononuclear inflammatory cells in the periairway regions, despite minimal histologic evidence of lung injury. At Days 14 and 28 after bleomycin administration, TIMP-1 mRNA was found in cells associated with focal areas of ECM accumulation within the alveolar space as well as the alveolar interstitium. Mononuclear inflammatory cells and alveolar interstitial cells continued to be the predominant cellular sources of TIMP-1 transcripts weeks after the initial lung insult.

TIMP Protein Content in Normal and Bleomycin-Injured Mouse Lung

Reverse zymography demonstrated the presence of metalloproteinase inhibitory activity of approximate relative molecular mass (M_r) of 21 to 28 kD in the lung extracts of control mice (Figure 5). The level of metalloproteinase inhibitory activity was increased in lung extracts recovered 4 d after bleomycin administration, suggesting induction of TIMP expression. To confirm the identity and relative amounts of TIMPs in these lung specimens, we performed Western blot analysis on BALF and lung protein extracts of control and bleomycin-treated mice. TIMP-1 immunoreactive protein of approximate M_r 28 kD was detected at low levels in the BALF recovered from control mice (data not shown). Bleomycin administration resulted in 220-fold (n = 4, P = 0.016) and 151-fold (n = 4, P = 0.012) increases in TIMP-1 protein levels in the BALF recovered at Days 4 and 28, respectively, relative to control animals (Figure 6A). TIMP-1 immunoreactive protein was also detected in the lung extracts of all control animals (Figure 6B). At 4 and 28 d after bleomycin instillation, there was a trend toward increased TIMP-1 protein content in the lung extracts; however, this level did not achieve statistical significance. Thus, Western blot analysis demonstrated a dramatic and durable increase in TIMP-1 protein in the fibrotic lung that paralleled the observed increase in TIMP-1 mRNA. TIMP-1 expression was markedly increased in the alveolar compartment of the lung early in the course of bleomycin-induced lung injury. Moreover, the increase in TIMP-1 mRNA and protein was sustained for at least 4 wk after the initial lung insult.

View larger version (38K): [in this window] [in a new window]   Figure 5.   Reverse zymography of BALF or lung tissue collected at 4 d after the intratracheal administration of bleomycin or saline (control). BALF (40 µl) or lung extract (30 µg), isolated from individual mice, was electrophoresed on 10% polyacrylamide containing 1 mg/ml gelatin and 160 ng/ml gelatinase A. Areas of metalloproteinase inhibition visualize as dark bands stained with Coomassie blue. Recombinant human TIMP (hTIMP)-1 (200 pg) serves as a positive control. Molecular weight markers are indicated at right. The results shown are representative of four identical experiments.

View larger version (29K): [in this window] [in a new window]   Figure 6.   Western analysis for TIMP-1 immunoreactive protein in BALF (A) and lung tissue (B) collected at Days 4 and 28 after the intratracheal administration of saline (control) or bleomycin. BALF (5 µl) or lung protein extract (10 µg), isolated from individual mice, was electrophoresed under reduced conditions on a 10% SDS-polyacrylamide gel. Recombinant hTIMP-1 (20 pg) serves as a positive control. Molecular weight marker is indicated at right. The results shown are representative of three identical experiments.

Although lung steady-state mRNA levels for TIMP-2 were unchanged after bleomycin treatment, we found a significant increase in TIMP-2 immunoreactive protein in the BALF that persisted for at least 4 wk. TIMP-2 protein of approximate M_r 21 kD was present in low levels in the BALF isolated from control mice (data not shown). Bleomycin instillation produced 20-fold (n = 4, P = 0.026) and 103-fold (n = 4, P = 0.013) increases in TIMP-2 protein concentrations in the alveolar lining fluid at Days 4 and 28, respectively, versus control mice (Figure 7A). In contrast to the BALF, TIMP-2 protein levels were not significantly altered in the lung extracts at any point after bleomycin administration (Figure 7B). The TIMP-2 immunoreactive protein identified in the lung extracts of bleomycin-injured and control mice appeared as two distinct sizes of ~ M_r 21 kD and 18 kD. This size difference could be the result of differences in the reduction state of TIMP-2 in the samples analyzed, as previously reported (22).

View larger version (27K): [in this window] [in a new window]   Figure 7.   Western analysis for TIMP-2 immunoreactive protein in BALF (A) and lung tissue (B) collected at Days 4 and 28 after the intratracheal administration of saline (control) or bleomycin. BALF (5 µl) or lung protein extract (10 µg), isolated from individual mice, was electrophoresed as described in Figure 6. Recombinant hTIMP-2 (20 pg) serves as a positive control. Molecular weight marker is indicated at right. The results shown are representative of three identical experiments.

To determine whether bleomycin lung injury resulted in systemic increases of TIMP-1 and TIMP-2 we compared plasma levels of these metalloproteinase inhibitors in control and treated mice. TIMP-1 and TIMP-2 immunoreactive proteins were detected in control plasma but their levels were not altered at Day 4 or 28 after bleomycin treatment (Figure 8). We conclude that the observed increases in TIMP-1 and TIMP-2 proteins in the alveolar compartments of bleomycin-injured animals were not merely the consequence of systemic increases in these inhibitors.

View larger version (25K): [in this window] [in a new window]   Figure 8.   Western analysis for TIMP-1 and TIMP-2 immunoreactive proteins in plasma collected at Day 4 after the intratracheal administration of saline (control) or bleomycin. Plasma (0.5 µl), isolated from individual mice, was electrophoresed as described in Figure 6. Recombinant hTIMP-1 (20 pg) and hTIMP-2 (20 pg) serve as positive controls. Molecular weight marker is indicated at right. The results shown are representative of three identical experiments.

The relative increases in BALF TIMP-1 and TIMP-2 immunoreactive proteins were compared with that for total protein. At Days 4 and 28 after bleomycin instillation, BALF total protein was increased 9.7-fold (n = 4, P < 0.002) and 13.8-fold (n = 4, P < 0.001), respectively, compared with saline controls (Figure 9). In contrast, BALF TIMP-1 protein was increased 220- and 151-fold compared with saline controls at Days 4 and 28, respectively, whereas TIMP-2 protein levels were increased 20- and 103-fold. These data indicate that the increases in TIMP-1 and TIMP-2 proteins in the alveolar lining fluid after bleomycin injury were substantially higher than the increase in total protein in the alveolar compartment. Further, because TIMP-1 has a molecular weight that is 6 kD larger than that of TIMP-2, the disproportionate increase in TIMP-1 compared with TIMP-2 in the alveolar compartment could not be explained solely by extravasation of these metalloproteinase inhibitors from the vascular space of the injured lung. Increased local production of TIMP-1 protein appeared to contribute to the overall increase observed in the lung after bleomycin administration.

View larger version (28K): [in this window] [in a new window]   Figure 9.   Temporal changes in BALF total protein, TIMP-1, and TIMP-2 levels after bleomycin administration. For each day, data are expressed as the fold increase in concentration of total protein or in the signal intensity of TIMP-1 and TIMP-2 in the BALF. The mean values (± SE) for the total protein concentration (filled column) or signal intensities for TIMP-1 (hatched column) and TIMP-2 (open column) for bleomycin-injured mice (n = 4) are expressed as a fold increase of the mean values for control mice (n = 4) at each time point. Means were compared through a two-sample t test, assuming unequal variances. (*P < 0.01 versus saline control; **P < 0.02 versus saline control; ***P < 0.03 versus saline control.)

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this study we demonstrate that TIMP-1 mRNA and protein expression are selectively and markedly increased in bleomycin-induced pulmonary fibrosis. Further, high levels of TIMP-1 are present early after bleomycin exposure and persist for weeks after the initial insult. We provide in situ evidence that TIMP-1 mRNA is expressed in inflammatory foci of the injured lung and localized in cells within the alveolar and interstitial compartments. In conjunction with the increase in TIMP-1 steady-state mRNA levels, we observed a greater than 150-fold increase in TIMP-1 immunoreactive protein in the alveolar lining fluid recovered at Days 4 and 28 after bleomycin treatment. Although we observed no increase in lung steady-state TIMP-2 mRNA levels, there was a greater than 20-fold increase in TIMP-2 protein in the alveolar lining fluid recovered at Days 4 and 28 from bleomycin-injured lungs. The local increase in TIMP-1 and TIMP-2 proteins was accompanied by an increase in MMP inhibitory activity of a molecular size consistent with TIMP-1 and TIMP-2 in the injured lung. The large and sustained increase in TIMP-1 and TIMP-2 activity strongly implicates these molecules as potential regulators of the events leading to ECM accumulation in the fibrotic lung.

The involvement of TIMP-1 in the fibrotic response is further reinforced by the observation that TIMP-1 gene expression was spatially restricted in the injured lung. As early as Day 2 after bleomycin instillation, TIMP-1 mRNA localized to alveolar and interstitial cells of the periairway and perivascular regions of the lung. As the lesions evolved, TIMP-1 mRNA was identified in many of the cells within these areas, but not in areas of lung that are remote from the site of tissue injury.

Previous studies in animal models and patients with fibrotic lung disease have suggested that TIMP-1 may be involved in the pathogenesis of pulmonary fibrosis. Swiderski and coworkers found that lung steady-state TIMP-1 mRNA levels increased during the first 3 wk after bleomycin injury in mice (23). Similarly, increased TIMP-1 concentrations were identified in BALF recovered from patients with interstitial lung disease as well as from patients with the acute respiratory distress syndrome (ARDS) (24, 25). Our study is the first to evaluate TIMP-1 and TIMP-2 expression in the lung at both the mRNA and protein levels in a well established model of bleomycin-induced injury, to compare TIMP-1 and TIMP-2 protein levels in the interstitial and alveolar compartments of the injured lung with those in the vascular compartment, to demonstrate the presence of metalloproteinase inhibitory activity in the injured lung, and to localize TIMP-1 gene expression to areas of lung injury.

The fibrotic response to bleomycin-induced lung injury is characterized by extensive ECM remodeling. After bleomycin injury, steady-state mRNA levels for ₁(I) procollagen and ₁ (III) procollagen are increased, and total lung collagen content is increased approximately 2-fold (21, 26). The increase in total lung collagen content is the result of both increased collagen production (26) and decreased collagen degradation (27). We have demonstrated that the increase in TIMP-1 and TIMP-2 in the injured lung precedes the induction of procollagen 1(I) gene expression and is not accompanied by a detectable increase in the expression of the interstitial collagenase (MMP-13) gene. Our results strongly support the concept that collagen accumulation in this model of lung fibrosis is the consequence of increased collagen gene expression and decreased collagen degradation due to the imbalance between collagenase production and TIMP accumulation in the injured lung.

Our in situ hybridization and immunohistochemistry results suggest that pulmonary macrophages may be a source of TIMP-1 after bleomycin injury; however, the cell type has not been unambiguously determined using these methods. Similarly, immunohistochemistry of lung biopsies obtained from patients with idiopathic pulmonary fibrosis (IPF) demonstrated TIMP-1 immunoreactive protein in alveolar macrophages (28). Cultured alveolar macrophages and peripheral blood monocytes are known to express the TIMP-1 gene in an inducible manner (1, 29). Our results also indicate that lung fibroblasts may be a cellular source of TIMP-1 in the fibrotic lung. These findings are consistent with the observation that TIMP-1 immunoreactive protein localized to fibroblasts in the lung biopsies of patients with IPF and with ARDS (28). In addition, cultured lung fibroblasts have been shown to express TIMP-1 in an inducible manner (30, 31). The identification of TIMP-1 mRNA within cells in areas of inflammation and fibrosis suggests that TIMPs may function to modulate the inflammatory response and to stabilize matrix components deposited in the injured lung.

Our study demonstrates that TIMP-2 protein is markedly increased in the injured lung in the absence of a significant increase in TIMP-2 steady-state mRNA levels. This observation suggests that TIMP-2 expression is regulated at a post-transcriptional level in this model, and demonstrates that TIMP-1 and TIMP-2 are differentially regulated after bleomycin lung injury. In support of this concept, TIMP-2 expression by cultured mesothelial cells in response to TGF- stimulation has been shown to be post-transcriptionally regulated (32). Although lung steady-state TIMP-2 mRNA levels were unchanged after bleomycin administration, this result does not preclude the possibility of increased TIMP-2 gene expression in the microenvironment of the injured lung. In previous studies, in situ hybridization has demonstrated TIMP-2 mRNA in mesenchymal cells of the visceral pleura and interalveolar septa of normal mouse lung (33). TIMP-2 immunoreactive protein has been identified in fibroblasts of lung biopsies from patients with IPF and with ARDS (28). Cultured alveolar macrophages, primary lung fibroblasts, and mesothelial cells have been shown to secrete TIMP-2 and may serve as cellular sources of this antiprotease in the bleomycin-injured lung (31, 32, 34).

TIMP-1 and TIMP-2 are multifunctional molecules that have the potential to modify a number of cellular activities, in addition to their ability to modulate matrix turnover. TIMP-1 and TIMP-2 have been demonstrated to stimulate the proliferation of cultured fibroblasts (1, 9). TIMP-1 induces the secretion of collagenase by cultured fibroblasts (12). TIMP-1 and TIMP-2 have been shown to modulate acute inflammation in vivo. TIMP-1-deficient mice display an augmented resistance to Pseudomonas aeruginosa infection that is neutrophil- and complement-dependent (35). In rodent models of lipopolysaccharide-induced and immune complex-mediated alveolitis, exogenously administered TIMP-2 inhibits neutrophil accumulation in the lungs (6, 7). In addition, both TIMPs are capable of regulating lymphocyte apoptosis in vitro (10, 11).

Our studies implicate TIMP-1 and TIMP-2 as potentially important modulators of the events that lead to ECM remodeling of the bleomycin-injured lung. The dramatic and durable accumulation of TIMP-1 and TIMP-2 may regulate the inflammatory response as well as the remodeling of ECM components after lung injury. Additional studies, including experiments with TIMP-1- and TIMP-2-deficient mice, will be useful in defining how TIMP-1 and TIMP-2 function in the pathogenesis of pulmonary fibrosis.