INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Fibrosing lung diseases are important clinical disorders with no effective form of therapy. Most treatment regimens use antiinflammatory and immunosuppressive drugs such as corticosteroids, azathioprine, or cyclophosphamide with response rates not better than 20-30% (1, 2). Intensive research on the cytokines involved in chronic lung diseases suggests that an imbalance of inflammatory and antiinflammatory factors may lead to chronic tissue injury and repair (3). Hence, the development of more targeted antiinflammatory or immunosuppressive approaches involving activating or inhibitory cytokines is a reasonable approach to treatment (2). Many peptide-based therapeutics have a short half-life, which makes transient gene transfer a promising tool with which to generate high and prolonged levels at the site of delivery. The technical and safety problems still associated with efficient and sustained in vivo gene delivery are anticipated to be solved in the foreseeable future (4).

The antineoplastic antibiotic, bleomycin, is widely used in cancer treatment and can cause interstitial lung disease in humans. It has often been used to induce lung fibrosis in animal models, and these models have certain pathologic and morphologic similarities to human pulmonary fibrosis (5). One of the striking parallels between the experimental model and human disease is an overwhelming production of transforming growth factor 1 (TGF-1) during the repair process. TGF-1 is one of the key cytokines in scar formation, and can act at different levels to increase lung collagen deposition. TGF-1 is chemotactic for fibroblasts and promotes their transformation into myofibroblasts, induces the synthesis of matrix proteins and glycoproteins, and inhibits collagen degradation by induction of protease inhibitors and reduction of metalloproteases (6, 7). We have previously shown in different gene transfer models, using adenoviral vectors, that TGF-1, granulocyte-macrophage colony-stimulating factor (GM-CSF), and tumor necrosis factor (TNF-) all cause fibrotic reactions, the latter two most likely by induction of endogenous TGF-1 (8).

Different anti-TGF- approaches have been used successfully to prevent or reduce fibrotic disorders of the lung, the kidneys, or the skin. Neutralizing TGF- antibodies and soluble Type II receptors for TGF-, decorin, and Smad7 are all able to interfere with and inhibit TGF-1 at different steps in fibrogenesis (12). As decorin is an endogenous proteoglycan it seems a more natural way of inactivating TGF- than antibodies or soluble receptors.

Decorin is a ubiquitous proteoglycan with a core protein of ~ 45 kD, which was shown to have two binding sites for all TGF- isoforms and is an important negative regulator of this cytokine (20, 21). The first in vivo model using recombinant decorin to ameliorate the TGF- effects in a model of experimental glomerulonephritis in rats required four to six daily intravenous injections to be successful (12). In a hamster model of bleomycin-induced pulmonary fibrosis two intratracheal injections of recombinant decorin per week were necessary to significantly reduce fibrosis (15), demonstrating the use of this approach in lung disease and suggesting gene transfer of decorin as a way to enhance local prolonged delivery to the tissue.

We constructed an adenoviral vector carrying the gene for human decorin (AdDec). In this study we asked whether AdDec was able to abrogate TGF- effects in an in vitro bioassay, and whether transfer of AdDec to the lungs of mice can generate prolonged pulmonary levels of transgene protein high enough to prevent fibrosis induced by intratracheal bleomycin. Our results show that AdDec was able to induce the synthesis of decorin proteoglycan that can interfere with TGF- effects in vitro, and demonstrate that a single intranasal administration of the gene for decorin in a replication-deficient adenovector is able to substantially reduce tissue fibrotic pathologies and collagen accumulation in a mouse model of lung fibrosis.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Recombinant Adenovirus

The construction of the adenoviral vector (AdDec) is described in detail elsewhere (22). Briefly, full-length human decorin cDNA (generous gift of L.W. Fisher, NIH, Bethesda, MD) was cloned into a shuttle vector with a human cytomegalovirus (CMV) promoter and cotransfected with a virus-rescuing vector. The resulting replication-deficient virus was amplified and purified by CsCl gradient centrifugation and PD-10 Sephadex chromatography, and finally plaque titered on 293 cells. AdTGF^223/225 (a mutant TGF-1 that is translated into bioactive TGF-1 [9]) and control vectors (AdDL70) with no insert in the E1 region were produced in the same way.

Cell Culture and Bioassay for TGF-

A549 cells were plated in a 100-mm² flask and left in -modified minimal essential medium (-MEM) supplemented with 1% L-glutamine, 1% penicillin-streptomycin, 0.4% amphotericin, and 10% fetal calf serum (FCS) until confluent. The cells were infected with AdDec, AdTGF-^223/225, or AdDL70 at a multiplicity of infection (MOI) of 20 PFU/cell. After 14 h the supernatant was removed and the cells were washed six times with phosphate-buffered saline (PBS). Medium containing 1% FCS was added and supernatant generated over 72 h was saved for further analysis.

Bioactive TGF- was detected by an established bioassay (25). Mink lung epithelial cells (MLEC clone 32, kindly provided by D. Rifkin, New York, NY), with a stable transfection of an 800-bp fragment of the 5' end of the human plasminogen activator inhibitor 1 (PAI-I) gene fused to the firefly luciferase reporter gene, were cultured in six-well dishes in Dulbecco's modified Eagle's medium (DMEM) containing 1% penicillin-streptomycin, 1% L-glutamine, 10% FCS, and Geneticin (200 µg/ml; Sigma, Oakville, ON, Canada) until they were confluent. After three washes with PBS the MLECs were exposed to conditioned medium of the infected A549 cells: AdTGF-^223/225 supernatants (diluted 1:4 in medium) combined with AdDec supernatants (diluted 1:1, 1:4, and 1:9), AdDL70 (diluted 1:1), or monoclonal antibody against mouse TGF-1-3 (25 µg/ml; Genzyme Diagnostics, Markham, ON, Canada). The cells were then incubated for 16 h and washed again three times with PBS. A total of 300 µl of lysis buffer (0.1 M potassium phosphate plus 1 mM dithiothreitol [DTT], pH 7.8) was added and the cells were scraped off. The lysate was pelleted and resuspended in the same buffer. After three freeze-thaw cycles the supernatant was used for the luciferase assay. D-()-Luciferin and the firefly luciferase standard (Boehringer, Mannheim, Germany) were used and assayed by luminometer (Lumat LB 9501; Berthold Systems, Pittsburgh, PA). Data are presented in relative light units (RLU).

Animal Treatment

Six-week-old female C57BL/6 mice were obtained (Charles River Laboratories, Montreal, QC, Canada) and housed under special pathogen-free conditions. Rodent laboratory food and water was provided ad libitum. The animals were treated in accordance with the guidelines of the Canadian Council of Animal Care.

All animal procedures were performed with inhalation anesthesia with isoflurane (MTC Pharmaceuticals, Cambridge, ON, Canada). AdDec or AdDL70 (4 × 10⁸ PFU) was administered intranasally in a volume of 20 µl of PBS 48 h before administration of bleomycin. One group of animals received PBS only. Bleomycin (Blenoxane; Bristol-Meyers, Syracuse, NY) was dissolved in PBS and instilled intratracheally with a 30-gauge needle (0.1 U per mouse in a volume of 50 µl, equivalent to 5 U/kg) after a small surgical preparation. Control animals received PBS only by intratracheal injection. Mice were killed by abdominal aortic bleeding on Day 1, 4, 7, or 21 after bleomycin administration. Body weight was assessed on Day 21. All experiments involving the administration of bleomycin were repeated with a second group.

Bronchoalveolar Lavage

After opening the chest cavity, the lungs were removed and rinsed with PBS. Bronchoalveolar lavage (BAL) was performed as follows. A total of 1 ml of PBS was injected intratracheally in four aliquots of 250 µl and retrieved. The pooled fluid was centrifuged at 1,500 rpm for 10 min and the supernatant of the first 500 µl was taken for determination of TGF-1 levels. Ten microliters of a protease inhibitor was added immediately to prevent further degradation of proteins (0.1 mM phenylmethylsulfonyl fluoride). The samples were stored at 70° C. BAL cells were counted in a hemocytometer, centrifuged in a cytospin, and stained for differential cytology (Hema³-solution; Biochemical Sciences, Swedesboro, NJ). A total of 300 cells per sample was counted for differentials.

The right main bronchus was tied and the lung was removed, rinsed in PBS again, and frozen immediately in liquid nitrogen. Tissue samples were stored at 70° C until further processing for total RNA extraction or hydroxyproline determination. The left lung was fixed for histologic examination.

RNA Extraction and mRNA Analysis

Frozen lung samples were homogenized in 4 ml of Trizol with a tissue homogenizer. Chloroform (0.8 ml) was added and samples were centrifuged at 3,000 rpm for 30 min. The aqueous layer was aspirated and RNA was precipitated with 2 ml of isopropanol. After centrifuging at 9,000 rpm for 10 min and washing the pellet with 75% ethanol, total RNA was dissolved in RNase-free water and the concentration was determined in a spectrophotometer at 260 nm.

Decorin mRNA was determined by the Northern blot technique. Total RNA extracts of A549 cells (15 µg) or total lung homogenate (30 µg) were separated on a 1% formaldehyde gel and transferred to a nylon membrane (ICN Pharmaceuticals, Montreal, PQ, Canada). Blots were hybridized overnight at 52° C with a 1.6-kb cDNA probe for human decorin, stringently washed, and exposed for 2h (A459 cells) or 48 h (lung homogenate) to film (XAR; Eastman Kodak, Rochester, NY).

Determination of TGF- Levels in BAL

Total TGF-1 was determined after acid activation, using an enzyme-linked immunosorbent assay (ELISA) (R&D Systems, Minneapolis, MN), performed according to the recommendations of the manufacturer. This assay detects TGF-1 across a number of species. The sensitivity of the assay is 7 pg/ml.

Histologic Examination

After fixation in 10% buffered formalin for 24 h a longitudinal section of the lung was paraffin embedded, sectioned, and stained with hematoxylin and eosin (H&E) and Masson trichrome. The slides were examined by two reviewers in a blinded fashion, using a fibrosis score similar to the score published by Madtes and coworkers (26): Grade 0 = normal lung; Grade 1 = sparse fibrosis (fine connective tissue fibrils in less than 50% of the area); Grade 2 = mild fibrosis (fine fibrils in 50-100% of the area or patchy peribronchial scars); Grade 3 = moderate fibrosis (fine fibrils in 100% of the area and/or patchy peribronchial and parenchymal scars involving less than 50% of the area); Grade 4 = severe fibrosis (disseminated scars involving 50-100% of the section).

Hydroxyproline Assay

Frozen lung samples were homogenized in 5 ml of deionized water. One milliliter of the homogenate was hydrolyzed in 2 ml of 6 N HCl for 16 h at 110° C. Hydroxyproline content was determined by a colorimetric assay described earlier (27). Briefly, the reaction was started by adding 1 ml of chloramine-T solution to 400 µl of sample (diluted with 1.6 ml of water after adjusting the pH to 7.0). The reaction was stopped with 1 ml of 70% perchloric acid, and 1 ml of dimethylbenzaldehyde solution was added. After an incubation period of 20 min at 60° C the optical density (OD) was determined within 30 min at a wavelength of 557 nm. The results were calculated as micrograms of hydroxyproline per milligram wet lung weight, using hydroxyproline standards (Sigma).

Statistical Analysis

Data are shown as means ± SEM unless otherwise mentioned. For evaluation of group differences we used the Student t test assuming unequal variances. A p value less than 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Gene Expression of Human Decorin in Cell Culture and Lungs

Using the Northern blot technique we demonstrated that A549 cells were successfully infected with AdDec and expressed mRNA for human decorin (Figure 1, left). In mice treated with AdDec, mRNA signal for human decorin was found in total lung homogenate 3 d after infection (Figure 1, right). Decorin mRNA was present but reduced 7 d after infection. No mRNA for human decorin was detectable in A549 cells and lungs treated with control virus.

View larger version (65K): [in this window] [in a new window]   Figure 1.   Human decorin mRNA in cells and tissue. Total mRNA extracts of A549 cells infected with AdDec show a strong positive signal for human decorin (2-h exposure) while no signal was detectable with control virus AdDL (left). In lungs treated with AdDec, human decorin mRNA was present 3 and 7 d after infection (right, 48-h exposure). There was no signal after infection with control virus. Equal loading was confirmed by ethidium bromide staining of rRNA.

Abrogation of Active TGF- by Adenovirus-derived Decorin In Vitro

We wanted to show that the transgene-derived decorin is able to interfere with active TGF- in an appropriate bioassay. We generated A549 supernatants containing high amounts of bioactive TGF-1 by infecting the cells with AdTGF^223/225 (3.6 ng/ml, determined by ELISA). This supernatant diluted 1:4 induced a PAI-I/luciferase activity of ~ 3 × 10⁴ RLU in the bioassay (~ 5 × 10^-11 mg/ml luciferase standard). The supernatants generated by infecting A549 cells with AdDec and AdDL70 resulted in luciferase activities similar to those of A549 supernatants without virus treatment (negative control). When we combined supernatant of AdTGF^223/225-infected cells (approximately 700 pg of TGF- per milliliter) with supernatant of AdDec-infected cells we observed a dose-dependent reduction of the TGF--induced luciferase activity (Figure 2). Addition of TGF- antibody reduced the luciferase activity to the same extent as the highly concentrated AdDec supernatant. Combination of supernatants of AdTGF^223/225- and AdDL70- infected A549 cells did not result in altered TGF- activity in the bioassay. We were able to reproduce these results with recombinant human TGF-1 (R&D Systems) at 500 pg/ml and supernatants of AdDec-infected A549 cells.

View larger version (19K): [in this window] [in a new window]   Figure 2.   Dose-dependent abrogation of TGF- bioactivity by transgene-derived decorin. A549 supernatants infected with AdTGF223/225 induced the highest luciferase activity (PAI-I gene induction; +control ); addition of supernatants generated by infection of A549 cells with AdDec almost reduced the luciferase activity to that of TGF--free supernatants (-control  ), p < 0.001 for AdDec 1:1 and 1:4, p = 0.12 for AdDec 1:9. Addition of A549 supernatants infected with control virus (AdDL70) had no effect on active TGF- (p = 0.4). Columns represent the average of three assays for each condition.

Effect of AdDec on Cell Counts and TGF- Levels in BAL during Bleomycin-induced Pulmonary Fibrosis

As previously shown by others (review in [5]), intratracheal injection in bleomycin results in an increase in total BAL cells from Day 4 onward, which is still detectable 21 d after injection. There was a high number of neutrophils and macrophages with enlarged and vacuolized cytoplasm at Days 4 and 7; on Day 7 total eosinophils were elevated as well (Table 1). The cell differential returned to normal proportions after 3 wk, with macrophages still showing more cytoplasmic vacuoles. The intranasal pretreatment with AdDec 2 d prior to bleomycin administration did not alter the initial increase or differential of inflammatory cells recovered in the BAL. At 3 wk, however, AdDec-treated animals had significantly lower total cell counts and lower neutrophils than controls. The administration of AdDec with intratracheal PBS instead of bleomycin 2 d later had no significant effect on BAL cells compared with PBS-only controls (data not shown).

Total TGF- as measured by ELISA was increased almost 3-fold in all bleomycin-treated mice on Day 4 compared with PBS-treated controls, and was still elevated on Day 7. The levels returned almost to baseline by Day 21. No significant difference was seen between pretreatment with AdDec or PBS (Figure 3). We were unable to detect any spontaneous active TGF- by ELISA or by bioassay.

View larger version (13K): [in this window] [in a new window]   Figure 3.   Total TGF- in BALF. Significant increase in total TGF- levels on Days 4 and 7 after intratracheal injection of bleomycin (Day 4 p < 0.03, Day 21 p < 0.008 bleo versus control); there was no difference between mice treated with bleomycin/PBS (open columns) and bleomycin/AdDec (solid columns); n = 5-10 per group.

Effects of AdDec on Tissue Fibrotic Responses and Hydroxyproline Content in the Lung during Bleomycin-induced Lung Fibrosis

The application of AdDec without consecutive bleomycin did not result in any sign of altered extracellular matrix deposition. Some of the lungs showed scattered foci of lymphocytic inflammation around airways and vessels (Figure 4B), which is most likely virus associated and is observed in AdDL70 controls in the same way. The injection of bleomycin, however, induced severe morphologic alterations in the lungs. The acute and subacute stage of the bleomycin injury on Days 4 and 7 was characterized by an increase in inflammatory cells within the interstitium and alveolar spaces. Predominant cells were neutrophils and activated macrophages with cytoplasmic vacuoles. On Day 7 there was patchy infiltration of eosinophils, which was not consistent in all animals. Pretreatment with AdDec had no impact on the acute inflammation on Day 4. Twenty-one days after bleomycin injection, the groups treated with PBS and the control vector AdDL70 showed severe fibrosis with dense collagen accumulation, which was patchy and somewhat predominant around the bronchi because of the mode of delivery of bleomycin (Figures 4C and 4D). The AdDec-treated animals, however, had a markedly lower degree of fibrosis with a similar distribution pattern. The scars appeared to be less dense (Figures 4E and 4F). Using a semiquantitative, blinded histology score, we showed that the group pretreated with AdDec had significantly lower grades of fibrosis than did PBS or AdDL70-pretreated controls (Figure 5A). However, the AdDec-pretreated animals were not completely free of scars and interstitial collagen accumulation. The grade of fibrosis we obtained by this procedure correlated well with the hydroxyproline concentration of the lungs (r² = 0.26, p < 0.02; Figure 5C).

View larger version (150K): [in this window] [in a new window]   View larger version (152K): [in this window] [in a new window]   View larger version (148K): [in this window] [in a new window]   View larger version (143K): [in this window] [in a new window]   View larger version (145K): [in this window] [in a new window]   View larger version (133K): [in this window] [in a new window]   Figure 4.   Histology of lungs on Day 21 (representative sections, Masson trichrome). (A) Lung injected with PBS and no adenoviral vector (normal control). (B) Mice injected with PBS and AdDec 2 d previously showed normal lungs with some perivascular and peribronchial lymphocytic infiltrates. (C and D) Bleomycin induced a severe distortion of the lung architecture with dense, collagen-rich scars. Administration of control virus AdDL70 before bleomycin resulted in a similar histological appearance (not shown). (E and F  ) Pretreatment with AdDec markedly reduced the fibrotic response to bleomycin but was not able to completely abolish scar formation. Original magnification: (A-C and E ) ×20; (D and F  ) ×50.

View larger version (11K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   View larger version (7K): [in this window] [in a new window]   Figure 5.   Effect of transgene-derived decorin on the fibrotic response of bleomycin-treated lungs. (A) Fibrosis microscopically quantified from Grade 0 (no fibrosis) to Grade 4 (severe fibrosis), p = 0.02 bleo/PBS versus bleo/ AdDec, p = 0.002 bleo/AdDL versus bleo/AdDec. (B) Hydroxyproline concentration in the homogenized lung tissue, p < 0.003 bleo/PBS and bleo/AdDL versus no bleo, p = 0.01 bleo/PBS versus bleo/AdDec, p = 0.0005 bleo/AdDL versus bleo/AdDec. (C ) Correlation between fibrosis score and lung hydroxyproline concentration. (Numbers in parentheses reveal number of animals per group.)

Lungs exposed to bleomycin had approximately 2-fold higher hydroxyproline concentrations on Day 21 compared with normal lungs with PBS treatment alone (Figure 5B). Treatment with AdDL70 or AdDec only (no bleomycin) had no effect on the collagen content of the lungs. Pretreatment with AdDec 2 d before bleomycin significantly reduced hydroxyproline from 2.56 µg/mg in bleomycin/PBS and 3.05 µg/mg in bleomycin/ AdDL70 controls to 1.96 µg/mg (p = 0.01 and p = 0.0005, respectively). The small increase in hydroxyproline concentration in AdDL70-pre-treated animals over PBS control was not significant (p = 0.1).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study we used a replication-deficient adenoviral vector expressing the gene for human decorin (AdDec) in a bleomycin model of lung injury to alter the fibrogenic response. The adenovector was previously shown to be bioactive in an ex vivo organ culture model of lung morphogenesis (22). Here we demonstrate that a single administration of AdDec is able to reduce the fibrotic response of the lung after intratracheal injection of bleomycin as shown by histology and a marked reduction of the collagen content of the lungs. Furthermore, we provide data indicating that this effect is probably mediated by interference between the gene product and active TGF-.

The use of bleomycin to cause pulmonary fibrosis is an established model in experimental lung research (5). A single intratracheal injection of this drug leads to an acute interstitial and intraalveolar inflammation with predominantly neutrophil granulocytes and activated macrophages and upregulation of cytokine mRNA for TNF-, GM-CSF, interleukin 6 (IL-6), and IL-8 (5, 7, 28). After the acute event other cytokines such as TGF- and connective tissue growth factor (CTGF) are upregulated in order to repair the damage (5, 7, 29). Elevated total TGF- levels were easily detected in our bleomycin model, but we were unable to detect the active form. Similar findings have been previously described (7, 18) and this may be due to a lack of sensitivity of the assays used, or BALF cytokine levels may not represent tissue concentrations. Because it is thought that the reparative process is overwhelming and thus induces fibrosis, different ways to interfere with TGF- activity have been investigated. Systemic administration of neutralizing antibodies to TGF- were used in a model of glomerulonephritis as well as in a bleomycin model of lung fibrosis and reduced the degree of collagen accumulation in both (14, 30). In another model of kidney fibrosis, intramuscular liposome-mediated gene transfer of soluble TGF- receptor fused to IgG-Fc prevented renal fibrogenesis (17). These experiments support the importance of TGF- in the development of scar tissue. However, the problems of a systemic blockade of TGF- are to be considered because of the pleiotropic actions of this cytokine (6, 31).

Other successful approaches have tried to bypass these concerns by locally inhibiting TGF- in the lung. Two reports describe the use of recombinant decorin, a proteoglycan with TGF--binding properties, and a recombinant soluble receptor for TGF- in a hamster model of fibrosis (15, 19). Both were able to significantly reduce the fibrogenic effects of bleomycin when given twice a week by repeated intratracheal injections. This procedure is undoubtedly an improvement over systemic application of decorin, which needed four to six daily intravenous injections to be effective (30); however, its clinical application is questionable. The development of an aerosol device containing decorin might be a solution, but that could face troublesome technical problems because of the size and biochemical properties of the molecule (32), which make manufacture and delivery a concern.

It has been shown in numerous experiments that gene transfer can initiate efficient production of transgene protein locally in the lung (2, 3, 7). Although expression is transient, gene transfer still is more effective in the delivery of proteins with a short half-life compared with recombinant molecules. The duration of expression can last 2-3 wk, depending on the vector system used (4, 7). We are aware that chronic inflammatory disorders such as pulmonary fibrosis likely will need repeated gene transfer, which still faces problems because of the immunogenicity of vectors. These hurdles, however, are anticipated to be overcome in the foreseeable future (4).

We suggest that by means of gene transfer a sustained local production of decorin around the airways could be induced that could bind and inactivate TGF- in an appropriate model of experimental fibrosis. We constructed a replication-deficient adenovirus carrying the gene for human decorin and have shown in an ex vivo model of lung morphogenesis that the adenovector can successfully transfer the gene for human decorin to the embryonic lung (22). In this model endogenous TGF- modulates lung development and addition of TGF-1 to the tissue inhibits bronchial branching. This TGF--mediated response was completely abolished by treatment with AdDec or TGF- antibodies, suggesting that AdDec interferes with TGF- activity directly. We successfully delivered AdDec to the lungs of adult rats and found the transgene product in BAL fluids, using Western blot analysis (33). No effect on lung morphology was observed in these experiments. Successful gene transfer was shown in the present study by the Northern blot technique, showing mRNA for human decorin in A549 cells and in lungs infected with AdDec. Gene expression was demonstrated for at least 7 d, but with our experience in earlier experiments using similar adenovectors and better quantifiable gene products, we suppose that transgene decorin was produced for a duration of 10-14 d (8). The presence of bioactive gene product was confirmed in vitro, showing that supernatant of AdDec-infected cells abrogates TGF- activity in a bioassay in which TGF- induces a luciferase reporter gene under the control of the PAI-I promoter (25).

Here we investigated whether AdDec is able to interfere with the fibrogenic response to intratracheally injected bleomycin, an experimental model known to be associated with TGF- overproduction (5, 28). Because we wanted to guarantee that a "decorin layer" was established around the bronchi when the fibrogenic drug starts its action, AdDec was given 48 h before bleomycin. A similar procedure was used by others to deliver an adenovector with the negative intracellular TGF--regulating protein Smad7 to the lung by intratracheal injection at the same time that subcutaneous bleomycin infusion was started (18).

In our experiments AdDec had no influence on the acute toxicity of bleomycin, as suggested by the profile of inflammatory cells in the BAL during the first week. This is not surprising, because TGF- is thought to be a "second-line" cytokine in the pathogenesis of the bleomycin injury and responsible for repair and not acute damage (5, 7). Similar observations have been reported by other groups using TGF- inhibitors in the bleomycin model (15, 18). Further confirmation of this hypothesis is provided by an upregulation of TNF- and IL-6 mRNA in all of our bleomycin-treated animals independent of delivery of AdDec or AdDL70 (data not shown). However, lung fibrogenesis after 3 wk was markedly reduced in the mice pretreated with AdDec compared with the control groups. The histopathology revealed much smaller areas of lung with pronounced collagen deposition in the treated animals. The peribronchial scars were less dense. We were able to quantify these observations by determining a fibrosis score, which showed a significantly better outcome for the AdDec group. In addition, we determined the total collagen content of the lungs by measuring the hydroxyproline concentration. This showed a significant reduction of fibrosis in treated versus control animals. We want to point out that there was still some matrix accumulation in the AdDec-pretreated bleomycin animals as reported by other investigators using similar techniques and approaches (15, 18). Besides technical considerations (some lung areas might come in contact with the fibrogenic stimulus only and not the inhibitor), it is likely that, in addition to TGF-, other cytokines must be blocked to abolish fibrosis completely.

We believe that decorin could be a promising therapeutic agent because it is an endogenous matrix component. Decorin is a 100-kD proteoglycan with a core protein of ~ 45 kD, by which it binds and inhibits (active) TGF- (21). The results shown here confirm and extend the data shown previously using decorin protein with repeat administration (15). It is likely that the adenovector used in our experiments to express the decorin gene induced the local production of decorin and exerted its effects by binding and sequestering TGF- in this bleomycin model. This is indirectly suggested by the ability of the adenovector and its gene product to interfere with TGF- activity in the bioassay. The physiologic role of proteoglycans in the lung (other than being just an element of the matrix) must still be determined. Several animal models suggest important immunoregulatory properties by binding and inhibiting overwhelming cytokine amount (15, 30) what is supported by the presented work and by a study using decorin to reduce neointimal formation after balloon injury (34).

There are controversial reports concerning proteoglycans in human lungs, because their appearance in normal lungs of different species is not uniform. One report notes considerable quantities of decorin and biglycan in adult rat lung (35), whereas another study examining normal human tissue was not able to show significant amounts (36). However, increased extracellular deposition of decorin in scar tissue and intracellular expression in myofibroblasts were seen in fibrotic human lungs, suggesting a potential role in pulmonary fibrosis (37). In human adenocarcinoma, decorin was found in central fibrosis and seemed to negatively regulate the scar formation induced by TGF- (38).

In summary, we showed that a single delivery of an adenovirus vector expressing decorin was able to substantially reduce the profibrotic effects of intratracheally administered bleomycin, an experimental model known to be driven by TGF-. We provide data that decorin is inhibiting TGF- effects in vitro in an established bioassay and suggest that the observed antifibrogenic properties of decorin in vivo are due to binding and sequestering of TGF- to the extracellular matrix of the lung. Because of evidence of increased TGF- and decreased decorin in human pulmonary fibrosis, the delivery of decorin might be useful as a future therapy. We suggest that interrupting the chronic inflammatory processes of fibrosis by transient, but prolonged, transfer of the decorin gene to lung tissue may allow the normal process of repair and homeostasis to return the lung to normal structure and function. While adenovirus vectors currently induce immune reactivity, limiting repeat use, other delivery systems are being developed that could overcome these limitations. We propose that safe and repeatable gene transfer could become the preferred method to locally produce and deposit the required amount of decorin at the site of damage to limit and reverse the process of fibrosis in the lung.