Smad3 Null Mice Develop Airspace Enlargement and Are Resistant to TGF-β-Mediated Pulmonary Fibrosis¹

The integrity and metabolism of extracellular matrix (ECM)³ components are central processes in lung development and homeostasis of lung function and morphology. Collagen accounts for ∼15% of the dry lung mass and is actively metabolized, with one-tenth of the total lung collagen being degraded and newly synthesized each day (1). Two different lung diseases can be considered to result from an imbalance of ECM homeostasis. Emphysema associated with the adult lung is thought to result from the progressive proteolytic destruction of ECM without adequate repair, occurring through an imbalance in proteinase-antiproteinase activity (2). In contrast, pulmonary fibrosis is characterized by excessive interstitial deposition of ECM (3), possibly also resulting from an imbalance in ECM metabolism, with a more inhibitory proteolytic microenvironment with decreased matrix removal (4). For both conditions, no treatment is known to affect the natural history of the disease.

TGF-β1 is a key cytokine involved in the process of fibrogenesis (4), and has been demonstrated to play a pivotal role in fibrosis of many different organ systems (5). We have previously shown that transient adenoviral vector-mediated gene transfer of active TGF-β1 (AdTGF-β1^223/225) to rat (6) or mouse (7) lungs leads to progressive and severe fibrosis. TGF-β causes myofibroblast differentiation, induces expression of connective tissue growth factor (CTGF), and increases the synthesis of ECM components such as collagen and fibronectin (3). Moreover, the role of TGF-β1 in inhibition of ECM degradation by up-regulation of proteinase inhibitors, such as tissue inhibitor of metalloproteinases (TIMPs) or plasminogen activator inhibitor-1 (PAI-1), has been demonstrated (8). In contrast, TGF-β1 has both positive and negative effects on synthesis of specific matrix metalloproteinases (MMP). MMPs are a large group of zinc-dependent enzymes that play a central role in the metabolism of ECM and basement membrane components (9). Among the metalloproteinases, MMP-2, MMP-9 (gelatinase A and B, respectively), and MMP-12 (macrophage elastase) have been implicated as having significant activity in the pathogenesis of emphysema (2). Recently, Morris et al. (10) showed that loss of activation of latent TGF-β in α_vβ₆ null mice causes a MMP-12-dependent emphysema. In addition, MMP-2 and MMP-9 have been shown to be present in fibrotic tissues (11).

TGF-β signaling pathways are complex, and the Smad, MAPK, PI3K, and JNK pathways have been implicated by various studies (12). Smads are a family of cytoplasmic signal transducer proteins. Among these, Smad2 and Smad3 predominantly mediate signals from activated TGF-β receptors. Once phosphorylated, they bind to Smad4, translocate to the nucleus, and activate numerous TGF-β-responsive promoters (5, 13, 14). Smad3 appears to be a crucial element in the signal transduction pathways involved in wound healing and fibrosis (15, 16, 17). Smad3 null mice are protected against radiation-induced fibrosis of the skin (18), and a second Smad3 null line demonstrates attenuated lung fibrosis induced by bleomycin (19). Overexpression of Smad7, an inhibitor of the Smad pathway that prevents the phosphorylation and activation of Smad2 and Smad3, has also been shown to protect against fibrosis in a bleomycin model (20).

In this study, we exposed Smad3 null (knockout (KO)) or wild-type (WT) mice (21) to high local concentrations of TGF-β1 in the lungs by transient overexpression of the active TGF-β1 gene using adenoviral vector gene transfer, AdTGF-β1^223/225 (6). We show in this study that loss of the Smad3 signaling pathway prevents specific TGF-β1-induced ECM gene regulation and blocks the development of progressive lung fibrosis. Moreover, we report for the first time that Smad3-deficient mice spontaneously develop increasing airspace enlargement with age, and demonstrate a marked presence of MMPs in the lung, with high acquired expression of MMP-9 and MMP-12. These data suggest that impaired regulation of both MMPs and TIMPs and other inhibitors, either under pathologic or homeostatic conditions, mediated through the Smad3 signaling pathway, may be causally related to the development of emphysema as well as pulmonary fibrosis.

A replication-deficient adenovirus carrying a mutated TGF-β1 gene was constructed, as previously described (22). The TGF-β1 gene is mutated at positions 223 and 225 (cysteine to serine), which prevents binding to its latency-associated protein and thus yields a biologically active TGF-β1. The resulting replication-deficient virus (AdTGF-β1^223/225) was amplified and purified by CsCl gradient centrifugation and PD-10 Sephadex chromatography, and finally plaque titered on 293 cells (6). The control vector (adenovirus null vector control) with no insert in the E1 region was produced in the same way (22).

Exon 8 of the Smad3 gene was disrupted in mice of background 129SV/EV × C57BL/6 by Yang et al. (21). Smad3 heterozygous mice were bred under special pathogen-free conditions. The genotypes of both WT and Smad3 KO mice were determined by PCR analysis on tail DNA obtained from 3-wk-old animals; all experiments were performed with littermates to ascertain identical genetic backgrounds. Rodent laboratory food and water were provided ad libitum. The animals were treated in accordance with the guidelines of the Canadian Council of Animal Care. All animal procedures were performed under inhalation anesthesia with isofluorane (MTC Pharmaceuticals, Cambridge, Canada). After weaning, animals are kept in a viral- and parasite-free environment, but are exposed to normal commensal bacterial agents.

For histomorphometric assessment, untreated 3-, 5-, 8-, and 9-wk-old, and 4-mo-old animals were sacrificed by abdominal aorta bleeding. Lungs were inflated and fixed by intratracheal instillation of 10% neutral buffered formalin at a constant pressure of 20 cm of water (15 cm for 3-wk-old animals) for 5 min (five to seven animals per group).

For assessment of fibrosis, a total of 5 × 10⁸ PFU of AdTGF-β1^223/225 or adenovirus null vector control was administered intranasally in a volume of 20 μl of PBS to 8-wk-old mice. Mice were sacrificed without any treatment, 4 and 28 days after adenoviral treatment. Five to seven animals were studied at each time point. After washing with PBS, the right main bronchus was tied; the right lung was removed, rinsed in PBS again, and frozen immediately in liquid nitrogen. Tissue samples were stored at −70°C until further processing. Bronchoalveolar lavage (BAL) was then performed, as described previously (23), in the left lung (300 μl). BAL fluid was processed for cell counts and stored at −70°C until further use. The left lung was inflated and fixed in 10% formalin, as described below, for histological examination. BAL from total lung (600 μl) from 8- to 10-wk-old animals were processed, and the cell pellets were resuspended in 1 ml of TRIzol for RNA extraction. For newborn, the entire lung was frozen for RNA extraction.

After fixation in 10% buffered formalin for 24 h, longitudinal sections of the lung were paraffin embedded, sectioned, and stained with either H&E for cellular content or Masson-trichrome for matrix content.

Lungs were fixed for 24 h before embedding in paraffin and sectioning at 5 μm. Sections were stained with H&E and were examined without knowledge of genotype. Twenty random fields (×100) within 1.5 mm of the lung pleura from the left lobe were digitized using a Leica DMR microscope (Leica Microsystems, Wetzlar, Germany). All sections were analyzed by the same image-processing algorithm using Leica Qwin Image Processing Software (Leica Imaging Systems, Cambridge, U.K.). Large airways and bronchi were excluded from the analysis. Results are reported as mean cord length of the airspace and are independent of the thickness or complexity of alveolar septa.

SDS polyacrylamide gels containing gelatin (1 mg/ml; manufacturer) were used to identify proteins with gelatinolytic activity from BAL supernatants. Each lane was loaded with 25 μl of sample. We measured total protein in the BAL fluid using a standard assay (Bio-Rad, Hercules, CA). After electrophoresis, the gels were incubated in a solution of 2.5% Triton X-100 (Sigma-Aldrich, Oakville, Canada) for 30 min, and incubated overnight at 37°C in 50 mM Tris-HCl, pH 8, 10 mM CaCl₂, 150 nm ZnSO₄, and 150 mM NaCl. The gels were stained with Coomassie brilliant blue R250 (Bio-Rad) for 30 min and destained in a solution of 7.5% acetic acid and 5% methanol. Zones of enzymatic activity appeared as clear bands against a blue background. MMP-2 and MMP-9 zymography standards were run, as well as a m.w. standard ladder (Chemicon International, Temecula, CA).

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 previously (24). The results were calculated as micrograms of hydroxyproline per milligrams of wet lung weight using hydroxyproline standards (Sigma-Aldrich).

Lungs of untreated 6-wk-old mice were removed under sterile conditions after exsanguination of the animals. The pulmonary vasculature was perfused blood free with 10 ml of PBS through the right ventricular cavity. Large bronchi and vessels were removed; the lung parenchyma was cut into 2- to 3-mm³ pieces, plated onto dry culture dishes, and incubated at 37°C for 30 min to ensure attachment. Eagle’s MEM containing 10% FCS, 1% l-glutamine, and 1% penicillin/streptomycin was then added, and the tissue was left for 7 days without further handling. After 7 days, fibroblasts were seen to be growing out of the tissue. Cells were used between passages 2 and 4. Primary pulmonary fibroblasts from Smad3 KO and WT mice were grown to confluence in 25-cm² flasks (BD Biosciences, Sandy, UT) using DMEM with 10% FCS. Cells were then rested in DMEM with 1% FCS for 12 h and exposed to medium containing 1 ng/ml human rTGF-β1 (R&D Systems, Minneapolis, MN). At four time points between 0 and 24 h, supernatants were removed and cells were harvested with TRIzol (Invitrogen Life Technologies, Burlington, Canada)

Cells lysed in TRIzol (Invitrogen Life Technologies) were processed for RNA and protein, according to the manufacturer’s instructions.

Frozen lung samples were homogenized in 7 ml of TRIzol (Invitrogen Life Technologies), and RNA was extracted. RNA integrity and concentration were determined with a microgel bioanalyzer (Agilent 2100; Agilent, Waldbronn, Germany). RNA (1 μg) was DNase treated, then reverse transcribed using a standard protocol (Invitrogen Life Technologies). Quantitative real-time PCR was conducted using an ABI Prism 7700 Sequence Detector. Negative control samples (no template or no reverse transcriptase) were run concurrently. Primers (Mobix, Hamilton, Canada) and probes (Applied Biosystems, Foster City, CA) are shown in Table I. Results were normalized to GAPDH, which was measured using previously optimized probe and primers (Applied Biosystems).

Data are shown as mean ± SEM. For evaluation of group differences, we used Student’s t test. A p value <0.05 was considered significant.

As we found that lung tissue from untreated adult (4-mo-old) Smad3 KO mice showed histological evidence of marked airspace enlargement, we analyzed them at different ages. At 3 wk of age, there were no major histologic differences between lungs of KO and WT mice. In both genotypes, there was minor evidence of uncompleted alveologenesis with some enlarged terminal bronchioles. However, even at this early time point, histologic examination suggested that there already was a minor increase in alveolar airspace in Smad3 KO mice (Fig. 1,A). We confirmed this small, but nonsignificant difference in alveolar size using quantitative histomorphology (Fig. 1,B). This increase became obvious in all KO animals from 8 wk of age onward, and at 4 mo of age, the airspace enlargement in the peripheral airspace in lungs of Smad3 KO mice was very evident. This enlargement was widespread throughout the parenchyma, but was not homogeneous, because enlarged alveoli were interspersed with normal-size alveoli (Fig. 1,A). The differences in airspace in 8-wk-old and 4-mo-old mice were confirmed with quantitative histomorphology, which demonstrated, respectively, a 22 and 36% increase in mean cord length of the airspace in Smad3 KO lungs compared with WT (p < 0.001) (Fig. 1 B).

Untreated adult 8-wk-old KO mice showed a significant 35% increase in total cell count in BAL fluid compared with WT littermates (Table II). However, there was no significant difference in the differential cell count, except for a slight increase in neutrophils in KO animals compared with WT (respectively 1.6 and 0.3%). There was no evidence of abnormal cell morphology or tissue destruction.

MMP-9 activity was barely detectable in the BAL fluid from 8-wk-old WT animals. However, there was significant MMP-9 activity, as shown by zymography in Smad3 KO mice, even though the total protein concentration in BAL fluid was not significantly different between KO and WT. MMP-2 activity was present at similar levels in the BAL of both mouse strains (Fig. 2,A). Cells harvested from BAL fluid from adult Smad3 KO mice (95% macrophages) showed an 8.5-fold increase (p < 0.05) in MMP-12 mRNA expression compared with WT animals (Fig. 2,B). The MMP activities measured by zymography were consistent with mRNA expression in total lung RNA from 8-wk-old untreated mice (Fig. 2,C). Quantitative PCR analysis showed no difference in MMP-2 mRNA expression in total lung RNA between Smad3 KO and WT mice, whereas MMP-9 mRNA expression was significantly higher (∼2-fold, p < 0.05) as was MMP-12 expression (2.5-fold, p < 0.05) in Smad3 KO mice compared with WT. There was no significant difference in MMP-9 and MMP-12 expressions between KO and WT in total lung mRNA from newborn animals (day 1; Fig. 2,D). In contrast to the different patterns of MMP expression, expression of several genes implicated in ECM metabolism (elastin, procollagen, CTGF, PAI-1, TIMP-1) was similar in untreated 8-wk-old KO and WT mice (Figs. 2,C and 5 A).

TGF-β1 gene transfer and expression were confirmed by measurement of active cytokine in BAL fluid by ELISA, 4 days after infection, and were not significantly different between KO and WT mice (1397 ± 352 and 1685 ± 458 pg/ml, respectively). Four days after AdTGF-β^223/225 administration, there was no significant difference between WT and KO in either the total or the differential cell count (Table II) in BAL fluid.

At day 28, Smad3 KO or WT animals treated with control adenovector were not distinguishable from age-matched naive animals (data not shown), as we have previously described in other animal models (6, 25). However, following treatment with the adenoviral vector expressing active TGF-β1 (AdTGF-β1^223/225), histology of lungs from WT mice showed a fibrotic pattern characterized by marked deposition of abnormal amounts of ECM at day 28, predominantly in peribronchial tissue with patchy areas within the parenchyma (Fig. 3,A). In contrast, lungs from similarly treated Smad3 KO mice did not show collagen accumulation nor ECM deposition (Fig. 3 A) and, in fact, were not different from lungs from Smad3 KO mice after treatment with the control virus.

To confirm and validate the histological findings, tissue fibrosis was quantified by analysis of hydroxyproline content (Fig. 3 B). WT mice treated with AdTGF-β1^223/225 had a significantly higher hydroxyproline content in the lung (65% increase, p < 0.05) at day 28 compared with mice treated with control vector. In contrast, hydroxyproline concentration in the lung of Smad3 KO mice was not significantly different at 28 days after AdTGF-β1^223/225 administration compared with lungs of animals treated with null adenovirus.

To determine whether the lack of fibrotic response was evident at the level of the fibroblast, we used primary cultures of lung fibroblasts from Smad3 WT and KO mice to investigate differences in gene expression of fibrosis-related genes in response to exposure to rTGF-β1. CTGF mRNA expression at basal level was 2-fold lower in Smad3 KO lung fibroblasts than in WT. After treatment with rTGF-β1, there was a significantly increased expression of CTGF in WT fibroblasts at 3, 6, and 24 h (3.6-, 4-, and 3.3-fold increase, respectively), which was significantly higher at each time point than in Smad3 KO fibroblasts. PAI-1 mRNA expression was also lower in untreated Smad3 KO lung fibroblasts (p < 0.05) compared with WT. At 6 h after rTGF-β1, PAI-1 expression was still 2.2-fold lower in Smad3 KO fibroblasts (Fig. 4,B). TIMP-1 gene expression (Fig. 4,C) was not significantly different at basal level in either cell type. However, at 6 h after treatment with rTGF-β1, there was a significant (1.7-fold) increased expression (p < 0.03) in WT fibroblasts that was not seen in KO fibroblasts. TGF-β1 mRNA was never expressed at enhanced levels after rTGF-β1 exposure in Smad3 null fibroblasts (Fig. 4 D), whereas there was a significant increase in TGF-β1 gene expression in WT fibroblasts at 6 and 24 h (p < 0.03).

To determine whether the gene regulation seen after rTGF-β1 stimulation of lung fibroblasts in vitro is representative of that in vivo, we examined fibrosis-related gene expression by quantitative RT-PCR in the lungs of Smad3 WT and KO mice at 4 days after exposure to AdTGF-β1^223/225. Nontreated Smad3 KO mice or WT expressed similar levels of mRNA for collagen, CTGF, TIMP-1, and PAI-1 (Fig. 5,A). At day 4 after AdTGF-β1^223/225 administration, there was a significant increase in expression of procollagen 3A1 (1.8-fold, p < 0.05), CTGF (2.6-fold, p < 0.05), TIMP-1 (16-fold, p < 0.01), and PAI-1 (5-fold, p < 0.03) in WT animals compared with control, whereas there was no similar increased expression in KO animals (Fig. 5 B).

We have observed two distinct, but likely related processes occurring in the pulmonary matrix of Smad3 KO mice: 1) untreated Smad3 KO mice, kept in conventional animal housing, develop an age-related increase in alveolar spaces, and by 4 mo of age, show clear evidence of progressive lung airspace enlargement; 2) in contrast to WT littermates, exposure of Smad3 KO mice to enhanced levels of active TGF-β1, by adenovirus gene transfer to the lung, fails to induce matrix deposition and progressive pulmonary fibrosis. We suggest that both these processes are caused by modulation of TGF-β signaling through the Smad3 pathway, and the outcome is determined by imbalances in ECM metabolism induced by changes in spontaneous or pathologic gene expression.

In this study, we observed that naive Smad3 KO mice appear to develop age-related severe lung airspace enlargement. Smad3-deficient mice show increased mortality between 1 and 8 mo due to primary defects in immune function with inflammatory lesions (21). However, tissues such as lung, brain, heart, and kidneys are not commonly involved, and these mice do not normally die from respiratory failure. To our knowledge, no descriptions have been reported of airspace enlargement in aging Smad3 KO mice. At 4 mo, emphysematous lesions were extensive, morphologically typical, and constant between animals (Fig. 1 A), with the mean cord length of the airspace in KO mice being 36% greater than that of WT.

TGF-β is an important growth factor, among others, in lung development, having a pronounced inhibitory impact on branching morphogenesis (26, 27). Smad3 null mice, however, do not display these developmental defects at birth. Explanations include a postulated compensatory role of Smad2- or other Smad3-independent signaling pathways downstream of TGF-β in development of the lung (27). We observed only mild, nonsignificant airspace enlargement at 3 wk in Smad3 KO mice (Fig. 1 B). However, emphysematous changes increase dramatically with time, reaching significant alveolar enlargement by 8–9 wk and easily evident enlargement by 4 mo of age.

During development of emphysema in the human, there is a decrease in total ECM components of the lung parenchyma (2), which could result from either a decrease in ECM production, an increase in ECM degradation, or both. The steady state level of procollagen 3A1 expression, which is at least in part dependent on the Smad3 pathway and TGF-β1 stimulation (16), was not significantly lower in the Smad3 KO mice than in the control group at birth or at adult age. Moreover, we did not find any significant difference in elastin expression at birth or at 8 wk between KO and WT mice, despite the fact that the Smad3 pathway has been implicated in the expression of elastin dependent on TGF-β stimulation (28). Fibronectin expression has been described both as being dependent (29) or not dependent on Smad3 pathway signaling (17).

The balance between proteinases, especially MMPs, and their inhibitors, TIMPs, is also thought to be critical in the process of emphysema. MMP-2, MMP-9, and MMP-12 have been implicated in this ECM-destructive disease (30, 31, 32). Both MMP-2 and MMP-9 are capable of elastolysis as well as collagenolysis (33), while MMP-12 is elevated in mice exposed to cigarette smoke and MMP-12 KO mice are protected from cigarette smoke-induced emphysema (30), and alveolar macrophages release more MMP-9 in chronic obstructive pulmonary disease patients and cigarette smokers (34, 35). Osteopetrotic mice, deficient in macrophage CSF, develop emphysema and have higher levels of MMP-2, MMP-9, and MMP-12 in BAL fluid and alveolar macrophages than in controls (36), while lung surfactant protein D gene-inactivated mice have airspace enlargement associated with an increase in the same three MMPs (37). Finally, the balance of MMPs to TIMPs within the local tissue microenvironment is important in the process of emphysema, as indicated in TIMP-3 KO mice, which develop progressive airspace enlargement (38).

Smad3 is required for the TGF-β-mediated inhibition of MMP-12 expression in alveolar macrophages (39, 40), and monocytes exposed to TGF-β show inhibited release of MMP-9 (41). Moreover, both MMP-2 and MMP-9 can activate TGF-β from its inactive latent proform (42), identifying a positive feedback (homeostatic) mechanism designed to closely regulate the levels of these two highly active enzymes. Our data now demonstrate a regulatory role for the Smad3 pathway in the spontaneous development of emphysema. We have shown an age-dependent increase in MMP-9 and MMP-12 activity in BAL from untreated Smad3 KO mice, compared with WT littermates. These changes in enzymatic activity correlate with up-regulation of MMP-12 mRNA in BAL cells (95% macrophages), and with elevated levels of MMP-9 and MMP-12 mRNA in total lung RNA. Whereas TGF-β1-induced MMP-9 production has been suggested to be more dependent on Ras/MAPK pathways (43, 44), our findings suggest that Smad3 is also important in TGF-β1 inhibition of MMP-9 expression. Of note, MMP-2 has been shown to be dependent on Smad2, and not Smad3, for stimulated expression (45), consistent with our observation that levels of MMP2 are not changed in Smad3 KO mice. Our data suggest that this intrinsic acquired difference in expression of MMP-9 and MMP-12, which is maintained even after TGF-β stimulation (data not shown), may be, at least in part, responsible for the progressive ECM destruction.

Together, these data suggest that the homeostatic balance of matrix synthesis and degradation in lungs of WT mice is perturbed in Smad3 KO mice to favor degradation of matrix (Fig. 1 B). As these mice are housed in conventional conditions, repeated exposure to commensal agents or cage dust may result in tissue damage and inflammation at the microlevel in the lung. In the absence of Smad3, physiologic repair by TGF-β is blunted, resulting in progressive alveolar damage and airspace enlargement.

TGF-β1 is a key cytokine in the fibrosis of various tissues, including lung. We have already shown the dramatic profibrotic activity of TGF-β1 in rat (6) and mouse lung (25) by adenoviral gene transfer of active TGF-β1. Involvement of the Smad 3 signaling pathway, known to be activated by TGF-β and also by activin, is suspected to be a major contributing factor in fibrogenesis (15) and other aspects of tissue remodeling.

In this study, we have demonstrated that transient overexpression of active TGF-β1 does not induce fibrosis at any time point examined out to day 28 in Smad3 KO mice in contrast to the accumulation of ECM and progressive fibrosis seen in WT mice. However, TGF-β1 autoinduction was blocked in Smad3 KO primary lung fibroblasts in vitro (data not shown), supporting previously reported findings in Smad3 KO embryo fibroblasts, keratinocytes, and macrophages (17, 45). TGF-β1 autoinduction is thought to be an important aspect in the maintenance of the fibrotic process over time (46).

CTGF is a TGF-β1-responsive gene and is able to induce synthesis of matrix components such as collagen and fibronectin (47). We found that Smad3 is important in determining the steady state expression level of CTGF mRNA, as was shown previously in mesangial cells (48) and suggested in human lung fibroblasts (49). In vivo, the response of CTGF and procollagen 3A1 to AdTGF-β1^223/225 administration was almost completely abrogated in KO compared with WT mice, consistent with the histology seen at 28 days after exposure, indicating that induced ECM synthesis is impaired in Smad3 KO mice. However, the Smad pathway alone cannot account for the overall response to overexpression of TGF-β. Although the basal level of CTGF mRNA expression was lower in Smad3 KO lung fibroblasts, they did respond to TGF-β with increased expression of CTGF mRNA, but never as pronounced as the WT fibroblasts (Fig. 2). These findings are consistent with reports implicating alternate signaling pathways involving Ras, protein kinase C, or MAPK in the induction of CTGF by TGF-β1 (48, 49).

Imbalance in expression of MMPs and their inhibitors (TIMPs) is thought to be a major contributor to the fibrotic process. High expression of TIMPs, creating a nonfibrolytic environment, has been associated with matrix accumulation, and the importance of inhibition of MMP and impaired collagenolysis in fibrosis has been confirmed in animal and human studies (50, 51, 52, 53, 54, 55). We have previously demonstrated this in mice prone to fibrosis, in which significant up-regulation of TIMP-1 could explain the propensity to develop fibrosis after TGF-β1 exposure (25). In this current study, we found that induction of expression of TIMP-1 mRNA, in response to TGF-β1 stimulation, was significantly impaired in Smad3 KO mice. The role of Smad3 in induction of TIMP-1 gene expression by TGF-β1 has also been demonstrated in human dermal fibroblasts and confirmed in Smad3 KO mouse embryo fibroblasts (16). Likewise, we found that induction of the antiproteinase PAI-1 by TGF-β1 is very dependent on Smad3, as suggested from previous work (14, 56, 57).

Interestingly, our data are in accordance with reports on the function of lefty, a member of the TGF-β family and a potent inhibitor of TGF-β signaling. In a fibroblastic tumor model in vivo (58, 59), it was shown that lefty decreased collagen type I mRNA expression and simultaneously increased collagenolytic, gelatinolytic, elastolytic, and caseinolytic activities, similar changes as seen in the current studies with Smad3 KO mice. Moreover, in a recent study, Morris demonstrated that the loss of α_vβ₆ integrin in mice induces a MMP-12-dependent emphysema, related to the lack of activation of latent TGF-β (10). The importance of Smad3 in the pathogenesis of fibrosis is evident from the lack of fibrogenesis seen when the KO mice are exposed to exogenous levels of active TGF-β (Fig. 1), while the WT littermate progresses to widespread fibrosis. Similar findings were previously reported for modulation of bleomycin-induced fibrosis in Smad3 KO mice (60). However, equally important is the fact that providing excess levels of active TGF-β did not prevent the progressive alveolar space enlargement (data not shown), indicating that Smad3 signaling pathways are involved either directly or indirectly in the process of both fibrosis and emphysema.

In summary, we have provided substantial data highlighting the importance of TGF-β1 and signaling through the Smad3 pathway in the maintenance of lung ECM. We have shown that two diverse and apparently opposite disease processes, fibrosis and emphysema, are each associated, at a fundamental level, with modulation of ECM metabolism. Fibrosis and emphysema represent two extremes of matrix metabolism, with fibrosis resulting from decreased (inhibited) ECM metabolism and deposition, and emphysema from increased matrix metabolism. More specifically, TGF-β appears to play a pivotal role in both processes, with signaling through the Smad3 pathway key to both disease processes. The loss of Smad3 protects against fibrosis resulting from pathologic expression of TGF-β, but confers susceptibility to emphysema by interfering with the normal physiologic (homeostatic) action of TGF-β in maintaining alveolar structure. This also suggests that in addition to the recognized association of susceptibility for emphysema to polymorphism in the MMP-12 gene, we should suspect a similar association might exist to polymorphism in the Smad3 gene, already known to be associated with the incidence of ovarian cancer (61, 62).

We thank Jane Ann Schroeder, Duncan Chong, and Xueya Feng for their invaluable technical help, and Mary Jo Smith for outstanding preparation of histology.