local renin-angiotensin systems (RAS) have been described for a number of tissues in which ANG II production is independent of circulating precursors (1, 38). Experimental evidence suggests that ANG II thus generated is an important regulator of the fibrotic response to tissue injury. This includes the ability of ANG II receptor antagonists and angiotensin-converting enzyme (ACE) inhibitors to attenuate cardiac and renal fibrosis in a number of animal models (6, 26).

The existence of a pulmonary RAS and the capacity for ANG II generation within the lung are suggested, but not confirmed, by previous studies. High ANG II concentrations have been demonstrated in normal rat lung (4). In addition, angiotensinogen and ANG II type 1 (AT₁) receptors are also expressed in lung tissue (2, 3, 5, 10). ACE (responsible for the majority of circulating ANG II generation) is expressed predominantly in a membrane-bound form by the endothelial cells of the pulmonary circulation but is also present in plasma. An elevation in bronchoalveolar lavage fluid (BALF) and/or serum ACE concentrations has been observed in many potentially fibrotic lung diseases, including sarcoidosis (24), idiopathic pulmonary fibrosis (42), asbestosis, silicosis (13), and acute respiratory distress syndrome (ARDS) (11, 15). However, the functional significance of increased ACE in this context is uncertain and may merely represent shedding of ACE from the activated endothelial surface. ACE expression has also been observed in macrophages from patients with sarcoidosis within granulomata (41), in circulating monocytes (33), and in BALF (36), but it is not clear whether other cell types within the lung are capable of expressing ACE or whether it is functionally relevant to disease. Our recent observation that the ACE insertion/deletion polymorphism is associated with the development of, and outcome from, ARDS also suggests a pathogenic role for RAS in acute lung injury.

ACE inhibitors attenuate endothelial activation (50), TNF activation (34), and collagen deposition (45) during experimental lung injury, possibly via a reduction in epithelial cell apoptosis (49, 51); however, it is not known whether these observations result from a reduction in lung ANG II generation. Similarly, it is not clear whether ANG II is able to modulate collagen deposition in the lung via a direct effect on fibroblast procollagen synthesis.

ANG II could influence the progression of lung injury via a number of mechanisms. Wang and colleagues identified ANG II as a proapoptotic factor for alveolar epithelial cells in vitro (48) via the AT₁ receptor (47). Furthermore, human lung fibroblasts from patients with idiopathic pulmonary fibrosis, but not from normal lung, were found to generate ANG II (47). This suggests a role for ANG II in epithelial cell survival.

We previously demonstrated that ANG II is mitogenic for human lung fibroblasts via activation of the AT₁ receptor (27), implicating ANG II in the fibroproliferative response to lung injury. However, an effect on collagen synthesis in these cells has not been studied. In vascular smooth muscle cells, the cellular actions of ANG II have been linked to the autocrine release of growth factors, such as platelet-derived growth factors, basic fibroblast growth factor, and transforming growth factor-β (TGF-β) (52). ANG II has also been shown to stimulate the autocrine release of TGF-β in adult rat cardiac fibroblasts (23). TGF-β is the most potent inducer of fibroblast procollagen synthesis described via increased procollagen gene transcription (16, 37), increased mRNA stability (35), and decreased intracellular degradation (29).

In the present study, we have tested the hypothesis that ANG II generation contributes directly to the fibroproliferative response to lung injury. We demonstrate that, in vitro, ANG II is a potent stimulator of lung fibroblast collagen production via the AT₁ receptor and that this is, in part, mediated by TGF-β. After bleomycin-induced lung injury, increased ANG II concentrations preceded a doubling of lung collagen. Lung ACE activity remained unchanged. Administration of an ACE inhibitor attenuated lung ACE activity, ANG II concentrations, and collagen deposition. Treatment with an AT₁ receptor antagonist also reduced lung collagen deposition and increased ANG II levels. Together, these data support the hypothesis that ANG II, possibly generated within the lung during acute injury, may contribute directly to lung collagen deposition via fibroblast activation. However, the efficacy of ACE inhibition in this model may also involve actions unrelated to ANG II generation.

MATERIALS AND METHODS

Cell culture. Human fetal lung fibroblasts (HFL-1, American Type Culture Collection, Rockville, MD) and adult lung fibroblasts (from a histologically classified patient with usual interstitial pneumonia) were subcultured as previously described (27). Cells at passages 15–21 were seeded at a density of 1 × 10⁵ cells/well in 1 ml of Dulbecco's modified Eagle's medium (DMEM)-10% neonatal calf serum into 12-well sterile culture dishes. When the cells reached visual confluence (after 5–6 days), the culture medium was removed and replaced with 1 ml of serum-free DMEM. After a further 24 h, the medium was replaced with fresh incubation medium. For experiments involving receptor antagonists, cells were preincubated with inhibitors in 0.5 ml of DMEM for 30 min before the addition of angiotensin or control medium to a final volume of 1 ml. Ascorbate was replenished after 24 h. Cells were exposed to 10^-6–10^-11 M ANG II for 48 h.

Specific nonpeptide antagonists to the AT₁ (10 μM losartan; courtesy of Merck, Sharp, and Dohme, Hoddeson, UK) and AT₂ (10 μM PD-123319; Parke-Davis) receptors were used to determine the receptor type responsible for any change in hydroxyproline synthesis. Hydroxyproline synthesis was also assessed in the presence or absence of pan-specific neutralizing antibodies to TGF-β isoforms 1–3 (30 μg/ml final concentration; R & D Systems) and isotype-matched IgG controls at the same concentration. The concentration of antibodies was chosen with reference to the manufacturer's data as those required to inhibit 80% of TGF-β activity (1 ng/ml). Subsequently, the efficacy of each antibody was confirmed in our culture conditions by their ability to inhibit TGF-β1-induced collagen synthesis at a range of concentrations. Cells were preincubated for 30 min with antagonist or antibody before the addition of 10^-8 M ANG II. At the end of the incubation period, tissue culture plates were frozen at -40°C. Each experiment was repeated on at least three separate occasions with six replications.

Bleomycin model of lung injury. Male Lewis rats (Harlan), 6–7 wk of age and 150–200 g body wt, were anesthetized with Hypnormfentanyl citrate (0.315 mg/ml) and fluanisone (10 mg/ml, 0.3 ml) injected intramuscularly. The trachea was exposed, and bleomycin sulfate (1.5 mg/kg in 0.2 ml) was instilled under direct vision via a 24-gauge cannula. Control animals received an equal volume of 0.9% saline. Each experimental group included six animals.

Losartan, a nonpeptide AT₁ receptor antagonist, is highly water soluble and was, therefore, suitable for administration via subcutaneous osmotic minipumps in a 0.9% saline solution (Alzet, Charles River, UK) implanted 24 h before bleomycin administration. The losartan dose was adjusted to the mean predicted weight (based on growth rates in previous experiments) with a target dose of 20 mg/kg. Ramipril (a nonsulfhydryl ACE inhibitor) has a relatively low solubility in water. This precluded the use of subcutaneous osmotic minipump infusion. This fact, combined with a reduction in water intake by animals after bleomycin administration, necessitated the direct instillation of ramipril (in saline solution) via daily gavage to ensure adequate dosage (1 mg·kg^-1·day^-1); ramipril was administered 24 h before bleomycin instillation, and administration was continued until the animals were killed. Control animals received an equal volume of 0.9% saline via osmotic minipump or gavage as appropriate. Animals were killed 3, 7, 14, and 21 days after bleomycin instillation. The lungs were perfused slowly with 10 ml of a 0.9% saline-1% heparin solution. Lungs recovered for determination of hydroxyproline or ANG II/ACE levels were then dissected free from the thoracic cavity, blotted dry, weighed, and snap frozen in liquid nitrogen before storage at -80°C. For the determination of hydroxyproline, frozen lungs were crushed to a fine powder, and 100 mg were used for acid hydrolysis and subsequent reverse-phase high-pressure liquid chromatography (HPLC), as described below.

Determination of fibroblast procollagen production in vitro and total lung collagen in vivo. Fibroblast procollagen production was assessed via the determination of hydroxyproline by HPLC, as previously described (29, 39). Briefly, proteins in the cell layer and medium were precipitated in ethanol and separated from free amino acids by filtration through an acid-resistant 0.45-μm pore filter (type HV, Millipore). Filters were hydrolyzed in HCl at 110°C for 16 h and decolorized with activated charcoal. After centrifugation at 300 g at 20°C for 2 min, the supernatant was removed and stored before derivatization at room temperature with 7-chloro-4-nitrobenzo-2-oxa-1,3-diazole (Sigma Chemical). Data are expressed as procollagen production with units of nanomoles of hydroxyproline per well or nanomoles of hydroxyproline per hour. We previously showed that ANG II is mitogenic for lung fibroblasts; thus the in vitro hydroxyproline data were also corrected for any changes in cell number.

For the measurement of total lung collagen, powdered lung tissue was similarly hydrolyzed in HCl and prepared for HPLC after dilution of hydrolysates (1:100). Total lung collagen was calculated with the assumption that the lung contains 12.2% (wt/wt) hydroxyproline (22) and expressed in milligrams.

ACE and ANG II determinations. ACE activity was measured as previously described (43). Briefly, thawed lung tissue supernatant was diluted with 400 μl of PBS (pH 8), and 50 μl of carbenzoxy-phenylalanyl-histidyl-leucine substrate (Z-Phe-His-Leu) were added. ACE catalyzes the hydrolysis of this substrate to yield the dipeptide histidyl-leucine (His-Leu), which fluoresces at an excitation wavelength of 365 nm and an emission wavelength of 495 nm. The reaction was stopped by transferring the sample into 1 ml of 0.28 M NaOH. o-Phthalaldehyde solution (25 μl of 2% solution in DMSO) was added to the NaOH-sample mixture that binds to the His-Leu product. After 30 min, this reaction was terminated by the addition of 1 ml of 1 M HCl. The amount of tagged His-Leu was then determined fluorometrically compared with His-Leu standards. Results were expressed as nanomoles of His-Leu per milligram of protein per minute for tissue ACE activity.

ANG II was determined by RIA. Frozen lung tissue was weighed, minced, and transferred in an iced solution of 0.1 mol/l HCl-80% ethanol (1:10, wt/vol). The tissue was homogenized with a Polytron Kinematica (Littau, Switzerland) and then subjected to sonication (Bandelin Sonoplus, Berlin, Germany). Homogenates were centrifuged, and the supernatant was diluted with 1:1 (vol/vol) 0.1 mol/l Tris-acetate buffer twice, and the supernatant was adjusted to pH 7.4. The final solution was concentrated by reversible absorption to phenylsilica cartridges (Bondelut, Analytichem). Absorbed angiotensins were eluted with methanol (3 times with 0.5 ml each time) into conical polypropylene tubes. The eluate was evaporated under vacuum rotation using a Speed Vac concentrator (Savant Instruments). Samples were reconstituted with RIA buffer and incubated with 50 μl of tracer (¹²⁵I-ANG II, 4,000 cpm) and 50 μl of ANG II antiserum for 20 h at 4°C. The antibody-antigen complex was separated from free ANG II by the addition of dextran-coated charcoal. Free radioactive ANG II (pellet) was counted in a gamma counter (1470 Wizard, Wallac ADL). Standards ranged from 1 to 250 pg/tube and were compared with the World Health Organization ANG II reference standard. The detection limit was 1 pg of ANG II per tube.

TGF-β bioassay. Active TGF-β in conditioned medium from HFL-1 cells exposed to control medium or ANG II (10^-8–10^-6 M) in serum-free conditions for 24 h was assessed using a highly quantitative bioassay based on the ability of TGF-β to induce plasminogen activator inhibitor-1 (PAI-1) gene expression in mink lung epithelial (Mv1Lu) cells stably transfected with a truncated TGF-β-responsive PAI-1 promoter fused to the luciferase reporter gene. These cells were a kind gift from Dr. D. B. Rifkin (New York University Medical Center, New York, NY). Cells were grown to 75% confluence and incubated with fibroblast-conditioned medium, naive medium, or conditioned medium spiked with a pan-specific TGF-β-neutralizing antibody (30 μg/ml final concentration) for 16 h. At the end of the incubation, the medium was removed, the cell layer was washed with cold PBS, and luciferase activity in cell lysates using passive buffer was assayed with a luciferase assay kit (Promega, Southampton, UK) according to the manufacturer's instructions on a microplate luminometer (Tropix TR717, PE Applied Biosystems). Data are expressed in relative light units per well and compared with data obtained with increasing concentrations of active TGF-β1 (0.062–1 ng/ml).

TGF-β expression by immunohistochemistry. After laparotomy and exsanguination of animals, the lung vasculature was perfused initially with heparinized PBS and then freshly prepared ice-cold fixative (buffered 4% paraformaldehyde) via the right atrium. Lungs were also fixed by intratracheal instillation of fixative at a pressure of 20 cmH₂O, immersed in fixative for 4 h at 4°C, and transferred to 15% sucrose in PBS overnight before dehydration in alcohol. Sections (5 μm thick) were cut from paraffin-embedded sections, which were dewaxed and rehydrated. Tissue endogenous peroxidase activity was blocked by incubation with 3% hydrogen peroxide in methanol. Sections were washed and incubated with a 1:6 solution of normal goat serum before incubation with rabbit polyclonal anti-TGF-β1 (0.5 μg/ml; Santa Cruz Biotechnology) overnight at 4°C. The sections were washed and incubated with biotinylated goat anti-rabbit IgG (1:200; Dako) for 60 min at room temperature, washed, and incubated with a streptavidin-peroxidase complex (1:200; Dako) for 30 min. Labeled TGF-β1 was visualized by incubating sections with a solution of 3,3′-diaminobenzidine (Vector) for 10 min. Sections were washed in water and counterstained with hematoxylin. Controls included incubation of sections with an isotype-specific, nonimmune rabbit IgG or primary antibody preincubated for 2 h with blocking peptide (2.5 μg/ml; Santa Cruz Biotechnology) instead of the primary antibody.

Statistical analysis. All collagen, ACE, and ANG II data are means ± SE. Statistical analysis was performed using an unpaired Student's t-test. For multiple comparisons, results were assessed by one-way ANOVA, and significance was determined using Bonferroni's correction. P < 0.05 was taken as significant.

RESULTS

Fibroblast procollagen production. Procollagen production over 48 h was increased in response to ANG II. This effect was concentration dependent in the range 10^-10–10^-6 and maximal at 10^-8 M for both cell types (1.46 and 1.22 nmol hydroxyproline/10⁵ cells, respectively). This represents a 184 ± 8 and 132 ± 6% increase above control for adult and fetal cells, respectively, compared with cells treated with medium alone (P < 0.01) but was significantly greater in the fetal cells at this concentration (P < 0.05). We previously demonstrated that ANG II is mitogenic for lung fibroblasts (28) and, thus, conditions that minimized proliferation were used. In addition, data were corrected for any change in cell number.

The AT₁ receptor antagonist losartan completely inhibited ANG II-induced procollagen production in HFL-1 cells (Fig. 1; P < 0.01). In contrast, the angiotensin type 2 receptor antagonist PD-123319 had no effect. Similarly, losartan and PD-123319 alone had no effect on procollagen production. Incubation with anti-TGF-β antibodies led to a significant inhibition of ANG II-induced procollagen synthesis (67 ± 3%, P = 0.01; Fig. 2). The antibody also successfully inhibited TGF-β-induced procollagen synthesis (87 ± 5%, P = 0.001), demonstrating the efficacy of the antibody in our culture conditions. No effect was observed with the IgG isotype control antibodies at the same concentration (data not shown).

Subsequently, ANG II-induced TGF-β expression in vitro was confirmed by the ability of conditioned medium from ANG II-exposed lung fibroblasts to induce PAI-1 gene expression in Mv1Lu cells at a range of concentrations, maximal at 10^-7 M ANG II (P < 0.0001; Fig. 3). This effect was completely blocked by incubation with a TGF-β-neutralizing antibody before the addition of the conditioned medium. TGF-β, at a range of concentrations, also induced PAI-1 expression in a dose-responsive fashion, maximal at 1 ng/ml (9,740 relative light units, P < 0.000001), which was also completely inhibited by preincubation with a TGF-β-neutralizing antibody.

Changes in ACE and ANG II during bleomycin injury. ACE activity was determined 3, 7, and 14 days after bleomycin administration and expressed per milligram of wet lung weight. No significant changes in ACE activity were observed throughout the experimental period: 5.4–5.5 and 4.9–5.4 nmol His-Leu·min^-1·mg lung wt^-1 for saline- and bleomycin-treated animals, respectively.

ANG II concentrations were determined by RIA 3, 7, 14, and 21 days after saline or bleomycin instillation (Fig. 4). ANG II was increased at all time points after bleomycin administration and was maximal by 14 days of administration, when an 80% increase in ANG II was observed (P < 0.01).

Lung collagen after bleomycin-induced injury. No signifi-cant changes in lung collagen were seen 3 and 7 days after bleomycin administration. By 14 days, lung collagen content was increased by 87 ± 4% (P < 0.01). This is consistent with previous observations from our laboratory (14). No additional increase was observed by 21 days; thus, for subsequent blocking experiments, the lung collagen at 14 and 21 days only was examined. No change in lung collagen was observed in control animals receiving intratracheal saline at any time point.

Effect of ramipril and losartan on lung RAS activation. Ramipril produced a dramatic reduction in ACE activity at each time point studied (Fig. 5). The reduction was maximal by 3 days and was maintained throughout the experimental period (P < 0.01 for all time points). Residual activity was observed but is common with the assay system used, inasmuch as the tissue sample is diluted with buffer to attain optimal assay conditions (1:30). This leads to a dissociation of ramipril from the enzyme and accounts for a loss of up to 40% of observed inhibition, which more than accounts for the residual activity observed in these samples (43). It is likely, therefore, that almost all ACE activity was inhibited by ramipril. No change in ACE activity was observed in response to losartan.

Figure 6 shows changes in ANG II concentration in response to ramipril and losartan. Ramipril significantly reduced ANG II concentration at all time points studied in saline- and bleomycin-treated animals (38–62% reduction, P < 0.01 for all time points). By contrast, all animals receiving losartan demonstrated an increase in lung ANG II at all time points studied (P < 0.05 for all time points).

Effect of ramipril and losartan on lung collagen. At 14 days after bleomycin administration, lung collagen deposition was significantly inhibited by losartan (Fig. 7A), representing a 39 ± 7% reduction in the excess lung collagen deposited (P < 0.05). Interestingly, there was also a significant reduction in collagen content in control animals receiving saline intratracheally (25 ± 7%, P < 0.05). At 21 days, a similar trend toward a reduction in lung collagen occurred (28 ± 7% of deposited collagen); however, this failed to reach statistical significance (P = 0.14). No difference was seen between control groups receiving intratracheal saline at this time point.

Ramipril also produced a significant reduction in lung collagen in bleomycin-treated animals at 14 days (39 ± 14% reduction in bleomycin-induced lung collagen, P < 0.05; Fig. 7B). Similar to losartan-treated animals, a significant reduction in lung collagen was also observed in control animals receiving ramipril and saline intratracheally at this time point (26 ± 4%, P < 0.05).

At 21 days, a significant reduction in excess lung collagen was observed in ramipril-treated animals (64 ± 19%, P < 0.05). No difference was seen in control animals receiving saline intratracheally.

TGF-β expression. Immunohistochemical examination of lung TGF-β expression was performed at key time points. In untreated animals receiving saline intratracheally, staining was observed within the alveolar interstitium, in alveolar macrophages, and in submesothelial connective tissue at 14 and 21 days (Fig. 8A). Bronchial epithelial staining was generally weak or absent, and alveolar type II cells were negative for staining. After bleomycin administration, strong staining was seen in alveolar macrophages as well as in the cytoplasm of cells and in extracellular matrix within fibrotic lesions (Fig. 8B). Weak staining continued to be seen in the bronchial epithelium and subepithelial matrix.

After 14 days of treatment with losartan, staining was weaker than for saline-treated controls, but small regions of intense focal staining were still present (Fig. 8C). Bronchial epithelial staining remained weak, while macrophages were still strongly stained. Similar changes were also seen at 21 days.

After treatment with ramipril, staining within fibroblastic lesions was absent and staining within macrophages was markedly reduced (Fig. 8D). Again, similar changes were seen at 21 days. In control animals treated with ramipril but receiving saline intratracheally, the normal staining seen in the alveolar interstitium, alveolar macrophages, and submesothelial connective tissue was also markedly reduced.

No staining was seen with control IgG antibodies at any time point.

DISCUSSION

In this study, we have demonstrated the ability of ANG II to stimulate lung fibroblast procollagen production via the AT₁ receptor, the autocrine production of TGF-β, and the inhibition of lung collagen deposition by an ACE inhibitor and an AT₁ receptor antagonist. Our preliminary observations also suggest a modulation of TGF-β after lung injury by these agents. The increase in ANG II observed in injured lung strongly suggests the local generation of this peptide, but further study is required. Taken together, these data support the hypothesis that ANG II is an important regulator of fibroproliferation in the lung.

ANG II induced a net increase in fibroblast procollagen production, with a maximal effect on collagen observed at a concentration much lower than that previously described for proliferation (27). Similar differences in concentration response have been observed in cardiac fibroblasts (53), but the importance is uncertain. It may be that the collagen response is more biologically important in vivo. Alternatively, effects on proliferation and matrix synthesis may have complementary roles at different phases of the repair response.

Local tissue concentrations of ANG II are difficult to estimate because of the short half-life of ANG II; however, recent studies reporting the synthesis of ANG II by cardiac and pulmonary fibroblasts in vitro (9, 46) support the hypothesis that such concentrations might be reached at the cell surface. In the present study, a significant increase in procollagen production was obtained at concentrations as low as 10^-10 M, and tissue ANG II concentrations of this magnitude have been reported in rat lung (4). Stimulation at such concentrations over longer periods of time could result in significant increases in collagen synthesis in vivo.

Although collagen is a major constituent of the interstitial matrix in the lung (30), the ability of ANG II to enhance the synthesis of other matrix proteins such as fibronectin or proteoglycans (crucial to collagen structure and mesenchymal cells) deserves attention in future studies, particularly because the synthesis of these proteins may be coregulated.

The increase in pulmonary ANG II concentration observed in injured lungs and, more importantly, the increase in noninjured lungs after losartan administration suggest that ANG II may be locally generated, particularly because it was fivefold higher than the ANG II concentration normally seen in the rat circulation (4). Increased ANG II production is a characteristic effect of AT₁ receptor antagonist administration (32). It is possible that ANG II could be sequestered into the lung, but it seems unlikely in the uninjured lung; furthermore, the contribution of circulating ANG II is believed to be negligible in other tissues (4, 40).

Interestingly, ACE activity did not change significantly. Indeed, some previous studies even demonstrate a reduction in whole lung ACE and the transpulmonary generation of ANG II from angiotensin I after experimental lung injury, including animals exposed to chronic hypoxia (17, 19) and monocrotaline (18), despite the ability of ACE inhibitors to attenuate pulmonary hypertension and vascular remodeling in this context (20, 31). It may be that the differential expression of ACE and other ANG II-generating enzymes within specific cells at sites of lung injury/repair may overshadow gross changes in tissue concentrations (31). Alternatively, ACE activity may not be the rate-limiting step in ANG II generation within the lung. Changes in ANG II may, for example, result from increased angiotensinogen synthesis.

Ramipril proved highly effective at inhibiting lung ACE throughout the experimental period; however, the reduction in ANG II was not complete, despite near-complete ACE inhibition. Although some residual ACE activity cannot be excluded, these data suggest that ACE represents approximately half of the total ANG II-generating capacity in the lung and implies the presence of other pathways for ANG II production. Chymase is one such enzyme, reported to be responsible for a significant proportion of ANG II-generating activity in cardiac membranes (44). However, the enzymes responsible for ANG II generation may differ between species and in pathological states.

The magnitude of the reduction in lung collagen achieved after ramipril and losartan treatment at 14 days was similar. Why only ramipril produced a significant change in collagen at 21 days is uncertain, particularly inasmuch as the increased lung ANG II concentration implies the efficacy of losartan in the lung at this time point. Ramipril's ability to inhibit bradykinin degradation may also have contributed to its efficacy. Alternatively, differences in pharmacokinetics may explain these observations. Ramipril binds to ACE and dissociates slowly (7, 8). It is a potent and prolonged inhibitor of tissue ACE, particularly pulmonary ACE (12), and its effectiveness within the developing fibrotic lesion may thus be greater than that of losartan, despite changes in ANG II.

An unexpected finding was the reduction in control lung collagen seen with ramipril and losartan. Collagen turnover in rat lung has been estimated to be as high as 10% per day in young animals (21), but it is likely to be lower in the animals used for these experiments. Thus one explanation for these observations is that ANG II modulates normal lung collagen turnover, stimulating basal collagen synthesis by lung fibroblasts. Thus, in the presence of inhibitors, continuing matrix degradation would result in a net reduction in lung collagen. Alternatively, these agents may be altering collagen metabolism indirectly, affecting nutritional status or gut absorption of essential nutrients, such as vitamin C, required for the normal synthesis of collagen. However, there is little evidence for this.

The ability of losartan to inhibit lung collagen deposition firmly supports a role for ANG II in the fibroproliferative response to lung injury. However, the direct inhibition of ANG II-induced fibroblast procollagen synthesis is not the only explanation for this effect. ANG II may lead to a reduction in collagen deposition by influencing other mechanisms at the initial stages of lung injury, in particular epithelial repair (45), as well as the inflammatory response (25, 34) and vascular permeability.

Conclusions. ACE inhibitors and AT₁ receptor antagonists have proven important therapies in the treatment of cardiovascular and renal diseases in which tissue remodeling and fibrosis occurs. It is now accepted that their clinical efficacy results from effects on circulating and tissue-based RAS. Our present observations also demonstrate the potential of these agents to attenuate collagen deposition after lung injury, possibly via a direct effect on lung fibroblast procollagen production. There are, however, a number of important cellular targets for ANG II in the lung, and the exact mechanism by which a pulmonary RAS contributes to the injury response remains to be clarified. It is also likely that other enzymes and effector molecules of the RAS system, such as chymase and bradykinin, play additional roles.

The local generation of ANG II within the lung is suggested but not confirmed by the present findings and that ACE is one, but not the only, enzyme responsible for its production. Future studies aimed at defining additional cell types responsible for RAS component expression would be highly informative and could enable the development of effective therapeutic regimens in humans.

FOOTNOTES

• The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.