Osteopontin is a multifunctional matricellular protein abundantly expressed during inflammation and repair. Osteopontin deficiency is associated with abnormal wound repair characterized by aberrant collagen fibrillogenesis in the heart and skin. Recent gene microarray studies found that osteopontin is abundantly expressed in both human and mouse lung fibrosis. Macrophages and T cells are known to be major sources of osteopontin. During lung fibrosis, however, osteopontin expression continues to increase when inflammation has receded, suggesting alternative sources of ostepontin during this response. In this study, we demonstrate immunoreactivity for osteopontin in lung epithelial and inflammatory cells in human usual interstitial pneumonitis and murine bleomycin-induced lung fibrosis. After treatment with bleomycin, osteopontin-null mice develop lung fibrosis characterized by dilated distal air spaces and reduced type I collagen expression compared with wild-type controls. There is also a significant decrease in levels of active transforming growth factor-β1 and matrix metalloproteinase-2 in osteopontin null mice. Type III collagen expression and total collagenase activity are similar in both groups. These results demonstrate that osteopontin expression is associated with important fibrogenic signals in the lung and that the epithelium may be an important source of osteopontin during lung fibrosis.
- pulmonary fibrosis
- type I collagen
- matrix metalloproteinase-2
- transforming growth factor-β
osteopontin is a matricellular protein abundantly expressed during inflammation and repair (11). Originally considered a bone matrix protein, recent studies showed that osteopontin also functions as a cytokine regulating macrophage accumulation and type I cytokine expression at sites of injury and infection (11, 32). Osteopontin is expressed in usual interstitial pneumonitis (UIP), fibrotic lung granulomas, and multiple models of murine lung fibrosis including transgenic overexpression of tumor necrosis factor-α and intratracheal challenge with bleomycin (19, 28, 32, 37, 41). Recent cDNA microarray studies identified osteopontin as one of several genes markedly upregulated during human UIP and bleomycin injury in mice (19, 41). Compared with normal lung, osteopontin expression was increased 12-fold in lung tissue from patients with UIP and 27-fold in mouse lungs following bleomycin challenge. The sites and cellular sources of osteopontin during lung fibrosis are unknown.
Osteopontin expression begins during the inflammatory stage of bleomycin injury in mice but progressively increases to peak levels during lung fibrosis (19, 37). Mice carrying a null mutation in the epithelial-restricted integrin-β6 develop inflammation in response to intratracheal bleomycin but are protected from lung fibrosis, largely due to failure of transforming growth factor (TGF)-β1 activation at epithelial surfaces (19). Osteopontin expression is markedly attenuated following bleomycin injury in β6-integrin null mice (19). Thus osteopontin expression correlates with fibrosis rather than inflammation during postinflammatory lung fibrosis in mice. Osteopontin null (−/−) mice have been shown to develop abnormal postinflammatory repair and fibrosis in the heart, kidney, and skin (24, 38). Osteopontin deficiency is associated with a predominance of small-caliber collagen fibrils and reduced type I collagen deposition (24, 38). Osteopontin can also modulate the expression of matrix metalloproteinases (MMP)-1, -2, and -9 as well as fibroblast proliferation and migration in vitro (11, 18, 34, 35, 37, 39). Based on the known functions of osteopontin, we hypothesized that it may modulate cellular accumulation and collagen depostion and remodeling during lung fibrosis.
Lung fibrosis involves excessive lung fibroblast proliferation, extracellular matrix remodeling, and architectural distortion due to a complex interaction between numerous factors such as TGF-β and MMPs (36). Intratracheal administration of bleomycin sulfate in susceptible strains of mice induces lung fibrosis. Recent studies demonstrated close similarities in cellular gene and protein expression between the two responses (19, 41). Nevertheless, differences in the mechanism and physiological consequences of fibrosis between this model and UIP impede direct comparisons of these responses (7).
Our present studies demonstrate the expression of osteopontin protein in human and murine lung fibrosis and describe the effects of osteopontin deficiency on pulmonary fibrosis induced by bleomycin challenge in vivo. Osteopontin immunoreactivity was present in inflammatory and epithelial cells during lung fibrosis. Compared with osteopontin-sufficient mice, osteopontin −/− mice developed lung fibrosis characterized by cystic dilatation of distal airways and a reduction in type I collagen expression. In the absence of osteopontin expression, there was decreased nonlatent or active TGF-β1 as well as total and active MMP-2 expression. These results show that osteopontin expression is associated with important fibrogenic signals in the lung. The association of osteopontin with lung epithelium and important mediators of the cellular response to lung epithelial injury suggests that the epithelium may be an important source and target for osteopontin during lung fibrosis.
MATERIALS AND METHODS
Tissue samples and bleomycin treatment.
Archived lung tissue from five patients with idiopathic pulmonary fibrosis was used as approved by the Institutional Review Board of the Boston Veterans Administration Medical Center. The hisotological presence of UIP was confirmed by a pathologist using standard criteria (1). Mouse studies were approved by the Institutional Animal Care and Use Committee at Boston University School of Medicine. The generation and baseline characterization of osteopontin −/− mice (deletion of exon 4–7 on 129SVJ × Black Swiss background) were previously described, and the genotype of all mice was confirmed by PCR analysis (24). Specific pathogen-free male and female homozygous osteopontin −/− mice or age- and sex-matched wild-type littermate controls (8–12 wk, weight 20–28 g) were challenged intratracheally (IT) with either bleomycin sulfate (Blenoxane, Bristol-Myers Squibb) or sterile saline. As there are strain variations in the response to IT bleomycin, we performed preliminary experiments with a range of doses of bleomycin from 0.03 to 0.1 units. These experiments showed that the 129SVJ background was moderately sensitive to bleomycin with doses over 0.03 units resulting in a high mortality rate due to acute lung injury by day 8. Doses of 0.03 units resulted in lung fibrosis with an early mortality of <10%. Thus all experiments were performed on groups of five to seven mice using 0.03 units of IT bleomycin. Histology was assessed at 8 and 16 days, and all other analysis was performed at 16 days.
After fixation in 4% paraformaldehyde and routine processing, the extent of fibrosis was assessed by percent involvement, and the severity of fibrosis was assessed using a modified Ashcroft score: 1 = minimal fibrosis; 2 = moderate fibrosis with thickening of the alveolar walls but no architectural distortion; 3 = fibrosis with definite damage to the lung structure and formation of fibrous bands; and 4 = severe distortion of structure with large fibrous areas (including honeycomb lung) (4). Cellularity was graded histologically and by bronchoalveolar lavage (BAL) cell counts. The number and size of epithelial lined cysts were measured using a tissue micrometer at ×100 magnification in eight random fibrotic sites per mouse lung. Results were reported as mean size (μm) and mean number of cystic spaces per high-power field. Histological analysis was performed by a pathologist (J. Hayes) blinded to the experimental conditions.
For immunohistochemistry (IHC), following antigen retrieval with citrate buffer, tissue sections were stained as previously described with the following primary antibodies: human lung biopsies with mouse anti-rat osteopontin MAb, MPIIIB10 (1:150 dilution, Iowa Hybridoma Bank); mouse lungs with 1) goat anti-rat osteopontin, Op199 pAb (20 μg/ml), 2) rat anti-mouse monocyte/macrophage marker F4/80 (1:5 dilution, Serotec), 3) goat anti-mouse type I collagen polyclonal antibodies raised against fetal mouse skin (1:25 dilution, Calbiochem), and 4) goat anti-human type III collagen (10 μg/ml; Chemicon International, Temecula, CA) (14, 31, 32, 38). Control slides were incubated with appropriate normal serum or isotype-matched MAb. Staining was read as either positive or negative except for type I collagen where it was graded as follows: immunoreactivty: 0 = none; 1 = minimal; 2 = moderate; and 3 = intense staining. All histology was assessed in a blinded manner.
BAL was performed on mice 16 days after bleomycin or saline challenge. Ice-cold sterile PBS with 0.1 mM EDTA was infused IT in aliquot volumes of 700 μl and aspirated. This was repeated until a 4-ml vol was recovered from each mouse. BAL was then centrifuged at 300-g force (1,200 rpm) for 10 min, the resultant cell pellet was resuspended in 1 ml of PBS. Total cells were counted using a microcytometer. After cytospin (40,000 cells/sample) and staining with Hema 3 stain set (123–869, Fisher Diagnostics, Middletown, VA), differential cell counts were performed by counting 400 cells in each BAL sample.
Lungs were first perfused with 10 ml of ice-cold PBS, excised, weighed, and frozen in liquid nitrogen. Lungs were then homogenized in Triton solution (0.25% Triton X-100 in 10 mM CaCl2) at a concentration of 20 ml Triton solution/g lung tissue. Final lung homogenate volumes were normalized to total protein concentration as measured by Bradford assay and stored at −20°C.
RNA isolation and Northern blot analysis.
Total lung cellular RNA was isolated by guanidine thiocyanate-phenol-chloroform extraction (10). RNA loading was assessed by ethidium bromide staining of ribosomal bands and by cohybridization with GADPH. The α1(I) collagen cDNA probe was derived from a rat α1(I) collagen cDNA that specifically binds murine α1(I) collagen mRNA (13).
Total collagen, TGF-β1 ELISA, and collagenase assay.
Total lung-soluble collagen was measured using the Sircol assay (Accurate Chemical and Scientific) as previously described (16). One-hundred microliters of total lung homogenate were assayed for collagen content and results were expressed per milligram of lung. Equal amounts of total protein from lung homogenates were assessed for latent and nonlatent TGF-β1 by ELISA (TGF-β1 Emax ImmunoAssay System, Promega). This ELISA has been used previously to measure lung total and bioactive (nonlatent) TGF-β1 in mice with a reported sensitivity of 32 pg/ml (21). Similarly, equal amounts of total protein from lung homogenates were assessed for collagenase activity using the Chemicon MMP collagenase activity kit (ECM710, Chemicon International), which measures cleavage of bovine biotinylated collagen using a streptavidin-enzyme complex. This collagen substrate is readily cleaved by MMP-1, -8, and -13, as well as enzymes with less specific collagenase activity such as elastases. Lung homogenate total collagenase activity was expressed relative to a standard curve constructed with serial concentrations of activated human MMP-1 (range 1–6 ng/ml) supplied by the manufacturer.
Gelatinolytic activity was assessed in lung homogenates using gelatin zymography as previously described (23). Equal amounts of total protein (20 μg) from lung homogenates were mixed with sample buffer in the absence of reducing agent and then loaded and run on a 10% SDS-polyacrylamide gel containing gelatin (concentration 0.5 mg/ml) for 90 min. The gel was then incubated in renaturing buffer (cat. no. LC2670, Invitrogen, Carlsbad, CA) for 60 min and then in developing buffer (cat. no. LC2671, Invitrogen) for 48 h at 37°C. Nondigested gelatin was then stained with 0.1% Coomassie brilliant blue (Sigma). Negative staining showed areas of gelatinolytic activity typical of MMP-2 and -9 between 60 and 92 kDa. The relative amount and ratio of activated to total MMP-2 were measured by densitometry (ChemiImager System and AlphaEase software, Alpha Ionotech, San Leandro, CA).
Results are expressed as means ± SD. Means ± SE are used where results from several experiments are reported together. Student's t-test was used to compare two groups of animals, whereas ANOVA with post hoc analysis using Newman-Keuls test was used where appropriate in multiple group comparisons. P < 0.05 was considered significant.
Expression of osteopontin in human UIP and murine bleomycin-induced lung fibrosis.
Using IHC, we showed abundant expression of osteopontin protein in lung biopsies demonstrating UIP. In all cases, immunoreactivity for osteopontin was seen exclusively in cellular components of the fibrotic response and not the fibrotic matrix (Fig. 1, A–D). Positive staining was seen in macrophages, lymphocytes, and both bronchiolar and alveolar epithelium (Fig. 1, C and D). No staining was seen when an isotype control antibody was used (Fig. 1E). Cubiodal epithelial cells typical of hyperplastic type II pneumocytes exhibited strong immunoreactivity for osteopontin (Fig. 1, C and D). There was no osteopontin signal in fibroblast foci (Fig. 1D). As we previously reported, osteopontin immunoreactivity was present in some alveolar macrophages but not airway epithelium or type II pneumocytes in normal human lung (not shown) (31). We next characterized the expression of osteopontin in postinflammatory lung fibrosis 16 days after bleomycin challenge in mice, a time we previously showed to be associated with established lung fibrosis in this strain of mice (Fig. 2C). Again, osteopontin signal was present in macrophages and epithelial cells but not the fibrotic matrix (Fig. 1, F and G). No staining was seen when an isotype control antibody was used (Fig. 1H). Consistent with prior reports, osteopontin immunoreactivity was not present in normal mouse lung (not shown) (37). Thus epithelia and macrophages demonstrate immunoreactvity for osteopontin in UIP and bleomycin-induced lung injury in mice.
Osteopontin −/− mice develop cystic lung fibrosis in response to bleomycin challenge.
At early stages (day 8), bleomycin induced diffuse pulmonary inflammation and alveolitis in both osteopontin +/+ and osteopontin −/− mice. This was characterized by prominent inflammatory infiltrates composed of lymphocytic aggregates and macrophages. By day 16, lung injury was characterized by diffuse fibrosis with persistent but decreased inflammatory cell infiltrates of macrophages and lymphocytes (Fig. 2, C–F). Lung fibrosis was diffuse but predominantly peribronchiolar and subpleural, although considerable heterogeneity was present with some mice exhibiting fibrosis of almost the entire lobe (fibrosis % lung involvement: osteopontin +/+: 23 ± 18%, mean ± SD, vs. osteopontin −/−: 31 ± 26%, mean ± SD; modified Ashcroft score: osteopontin +/+: 2.33 ± 0.5, mean ± SD, vs. osteopontin −/−: 2.5 ± 0.4, mean ± SD; n = 15, P = 0.3). These findings are typical of bleomycin-induced lung fibrosis in mice and were similar in osteopontin +/+ and osteopontin −/− mice. There was also no significant difference between the numbers of inflammatory cells between the groups of mice (Fig. 3A). In particular, the numbers of macrophages as measured by BAL cell count (Fig. 3A) and lung IHC for the macrophage marker F4/80 (not shown) were similar in osteopontin +/+ and osteopontin −/− mice.
There was an increase in the number of cystic epithelial lined air spaces within the fibrotic areas of lung in osteopontin −/− mice compared with osteopontin +/+ controls (Figs. 2, C–F, and 3B). There were 11.6 ± 2 (mean ± SD) spaces per ×100 field in osteopontin −/− mice compared with 5.5 ± 3 (mean ± SD) spaces per high-power field in osteopontin +/+ mice (P = 0.004; Fig. 3B). In addition to an increase in number, these air spaces were also more dilated in osteopontin −/− mice. The average diameter of the cystic spaces in osteopontin −/− was 1,090 ± 210 μm (mean ± SD) compared with 880 ± 96 μm (mean ± SD) in osteopontin +/+ mice (P = 0.04; Fig. 3B). There was no significant difference in the number or size of terminal bronchioles in osteopontin −/− compared with osteopontin +/+ mice. Results are representative of three experiments (n = 15/group).
Reduced type I collagen expression in response to bleomycin challenge in osteopontin −/− mice.
Total collagen levels in lungs from mice 16 days after bleomycin challenge showed increased collagen deposition in both osteopontin +/+ and osteopontin −/− mice compared with saline-treated controls (percent increase over saline 153 ± 22%, mean ± SD, in osteopontin +/+ mice vs. 117 ± 21%, mean ± SD, in osteopontin −/− mice; Fig. 3C). Collagen accumulation was less in osteopontin −/− mice compared with osteopontin +/+ mice, but this was not statistically significant (osteopontin +/+ 6.2 ± 0.9 μg·mg lung−1, mean ± SD, vs. osteopontin −/− 5.5 ± 1.3 μg·mg lung−1, mean ± SD; n = 6, P = 0.1). There was, however, a significant decrease in type I collagen between the two sets of mice (Fig. 4, A and B). By IHC, there was decreased expression of type I collagen in the fibrotic response seen in osteopontin −/− mice compared with osteopontin +/+ controls (IHC grade: osteopontin +/+ 2.5 ± 0.1, mean ± SD, vs. osteopontin −/− 1.4 ± 0.15, mean ± SD; n = 8, P = 0.004; Fig. 4, A and B). In contrast, type III collagen expression was similar in both sets of mice (Fig. 4, C and D).
These findings could reflect deceased expression type I collagen or increased degradation by type I collagenases that are widely expressed during lung fibrosis (36). By Northern blot analysis of total lung mRNA, we showed that α1 (I) collagen mRNA expression was similar at baseline in both sets of mice but was significantly higher in osteopontin +/+ (osteopontin +/+ 0.75 ± 0.1 relative densitometry to GAPDH, mean ± SD) compared with osteopontin −/− mice (0.35 ± 0.25, mean ± SD, relative densitometry to GAPDH; n = 3, P < 0.05) following bleomycin treatment (Fig. 5). In contrast, there was no difference in collagenase activity between osteopontin −/− and osteopontin +/+ bleomycin-treated mice (osteopontin −/− 116 ± 20 ng/ml, mean ± SD, vs. osteopontin +/+ 122 ± 18 ng/ml, mean ± SD, of MMP-1 activity per mg of lung, n = 6, P = 0.25).
Osteopontin −/− mice have reduced active TGF-β1 expression during bleomycin-induced lung fibrosis.
TGF-β1 is a potent stimulator of type I collagen production and so we next determined whether the expression of TGF-β1 was altered in these mice. By ELISA we found a significant decrease in nonlatent or active TGF-β1 levels in whole lung homogenates obtained during the fibrotic phase of bleomycin-induced lung injury. In three experiments involving a total of 16 mice per group, nonlatent TGF-β1 levels were 1,043 ± 144 pg/ml (mean ± SE) in osteopontin +/+ mice compared with 606 ± 97 pg/ml (mean ± SE) in osteopontin −/− mice (P = 0.015; Fig. 5). There was a small but nonsignificant decrease in total TGF-β in osteopontin −/− compared with osteopontin +/+ mice (osteopontin +/+ 2,171 ± 545 pg/ml, mean ± SE, vs. osteopontin −/− 1,753 ± 689 pg/ml, mean ± SE; P = 0.35). No significant differences were found in either total or active TGF-β levels in saline-treated control animals (active TGF-β: osteopontin +/+ 132 ± 120 pg/ml, mean ± SE, vs. osteopontin −/− 92 ± 85 pg/ml, mean ± SE; P = 0.35).
Reduced expression and activation of MMP-2 in osteopontin −/− mice.
Similar to collagenases, gelatinases are expressed during murine bleomycin-induced lung fibrosis and may regulate several aspects of lung fibrosis including TGF-β activation (23, 40). In view of these data, we next assessed gelatinase activity in bleomycin-treated osteopontin −/− and osteopontin +/+ mice. Compared with saline-treated controls, bleomycin treatment resulted in increased expression and activation of both MMP-2 and -9 in both groups of mice. However, there was a significant decrease in MMP-2 expression in osteopontin −/− mice compared with osteopontin +/+ mice (gelatinolytic activity: osteopontin +/+ 6,291 ± 1,200, mean ± SD, vs. osteopontin −/− 2,167 ± 510, mean ± SD; n = 5, P < 0.05; Fig. 6). There was also a significant decrease in active MMP-2 in osteopontin −/− mice (gelatinolytic activity: osteopontin +/+ 2,304 ± 800, mean ± SD, vs. osteopontin −/− 689 ± 350, mean ± SD; n = 5, P < 0.05; Fig. 6). This represented a significant reduction in the percent active form of MMP-2 (osteopontin +/+ 36 ± 4%, mean ± SE, vs. osteopontin −/− 32 ± 2%, mean ± SE, n = 10, P < 0.05). Expression of MMP-9 was similar in osteopontin +/+ and osteopontin −/− mice (data not shown). Results are representative of two experiments (n = 10 per group).
In these studies, we show immunoreactivity for osteopontin in epithelial cells and inflammatory cells during lung fibrosis. In addition, mice deficient in osteopontin expression develop altered bleomycin-induced lung fibrosis characterized by cystic dilated air spaces, decreased type I collagen expression, and a reduction of active TGF-β1 and MMP-2.
We found osteopontin immunoreactivity in inflammatory and epithelial cells but not the fibrotic matrix in UIP. Similar results were found in postinflammatory lung fibrosis induced by bleomycin in mice. Our results suggest that the lung epithelium may be an important source of osteopontin in response to epithelial injury and lung fibrosis. This is consistent with known expression of osteopontin by epithelial cells lining the bronchial, gastrointestinal, and genitourinary tracts (8). As well as bronchiolar epithelium we found osteopontin immunoreactivity in hyperplastic type II pneumocytes. In UIP, distal airway and alveolar epithelial cells undergo cuboidalization and hyperplasia and are prominent sources of fibrogenic cytokines such as TGF-β1 and platelet-derived growth factor BB (3, 20, 36). They may also provide critical signals to induce fibroblasts to migrate, proliferate, and ultimately develop a profibrotic phenotype (36). These results show that osteopontin is part of the epithelial response during lung fibrosis and suggest that it may play a role in epithelial repair and regeneration during lung fibrosis.
Some studies in osteopontin-deficient mice have shown altered inflammatory cell and, in particular, macrophage, accumulation as well as Th1 cytokine expression during inflammatory responses such as granulomatous inflammation (5, 9, 30, 32). Others involving wound repair and myocardial infarction have shown normal inflammation but aberrant fibrosis (24, 38). We show that despite normal macrophage accumulation, there is aberrant lung fibrosis in the absence of osteopontin. The nature of the lung injury may account for these different results. Models of granulomatous inflammation are usually intravenously delivered, causing a gradually evolving hypersensitivity response (22). In contrast, IT bleomycin results in acute severe epithelial damage followed by acute lung injury and fibrosis (36). It is possible that different signals and mediators regulate inflammatory cell recruitment in these types of lung inflammatory responses.
Several studies have shown prominent osteopontin expression during postinflammatory lung fibrosis in mice, and osteopontin can augment platelet-derived growth factor BB-induced lung fibroblast proliferation and migration in vitro (19, 28, 37). We now show that osteopontin deficiency is associated with altered fibrosis characterized by increased numbers of cystic air spaces. These findings may reflect airway dilatation due to an altered fibrotic response similar to the postinfarction-dilated cardiomyopathy that was recently reported in osteopontin −/− mice (38). Fibrosis characterized by cystic dilatation of distal air spaces has been described in mice following cadmium injury in a process that can be exacerbated by inhibiting the lysyl oxidase enzyme (29). Therefore, abnormal collagen matrix assembly can result in air space dilatation. Interestingly, electron microscopy has shown abnormal collagen fibril composition in both the heart and skin of osteopontin −/− mice (24, 38). We did not perform ultrastructural analysis, but our findings of reduced type I collagen expression in fibrotic osteopontin-deficient lungs are similar to that reported in cardiac postinflammatory fibrosis in osteopontin −/− mice (38). Collectively, these studies suggest that osteopontin may be required to develop a typical fibrotic scar with adequate tensile strength to prevent airway dilatation. A recent study demonstrated that deficient activation of TGF-β resulted in increased MMP-12 expression with associated airway dilatation and emphysema (27). Further studies are necessary to clarify the relationship of osteopontin deficiency to these findings.
Collagen deposition reflects a balance among expression, remodeling, and degradation of collagen matrix (36). Net collagen accumulation is regulated by TGF-β expression and activation as well as the presence and activity of various collagenases. We found that there was a decrease in α1 (I) collagen mRNA and no increase in collagenase activity in osteopontin −/− mice. The expression of type III collagen, another major fibrillar collagen expressed during lung fibrosis, was not different between the two sets of mice. These data suggest that there may be a specific deficiency in type I collagen expression in osteopontin −/− mice.
TGF-β is the major stimulus for type I collagen expression during lung fibrosis. In bleomycin lung injury, the activity of TGF-β is regulated at both transcriptional and posttranslational levels during bleomycin-induced lung injury. We demonstrated a reduction in active TGF-β1 but not total TGF-β in osteopontin −/− mice. Although osteopontin has previously been shown to be a TGF-β1 response gene, our data suggest that osteopontin may in fact function upstream of TGF-β1 by regulating its activation (11, 12, 17, 19, 26). A prior study showed that total TGF-β1 was reduced in osteopontin −/− mice following renal tubulointerstitial injury, but this has not been reported in other organs, and no prior study has shown a reduction in TGF-β1 activation (30). The activation of TGF-β from its latency-associated peptide (LAP) can involve conformational shifts due to the interaction of LAP with thrombospondin and integrins or proteolytic cleavage by MMPs and plasmin (2, 40). The β6-integrin is a critical activator of TGF-β1 in experimental lung fibrosis, and osteopontin expression is dramatically attenuated in β6-null mice during bleomycin-induced lung fibrosis (19). Thus osteopontin appears to function downstream of the β6-integrin in this response perhaps contributing to TGF-β1 activation. The mechanism by which osteopontin participates in the activation of TGF-β1 is unknown.
The gelatinases, MMP-2 and MMP-9, are expressed during lung fibrosis and can degrade gelatin and collagen as well as activate TGF-β and regulate cellular recruitment (6, 15, 23, 40). Our findings of reduced MMP-2 expression and activation in osteopontin −/− mice when challenged with bleomycin are supported by in vitro work that shows that osteopontin can induce and activate MMP-2 in carcinoma cell lines (34, 35). A recent study showed that transgenic mice overexpressing osteopontin develop arterial intimal thickening with increased expression and activation of MMP-2 (18). Similar to osteopontin, MMP-2 expression peaks during later stages of lung fibrosis and lung epithelia, including hyperplastic type II cells, which are a prominent source of MMP-2 in both UIP and murine models of fibrosis initiated by intratracheal bleomycin (15, 23). These data suggest that osteopontin and MMP-2 function at similar stages of lung fibrosis and that MMP-2 expression may be regulated by osteopontin during lung injury in vivo. Osteopontin can induce the expression of both membrane-type MMP-1 (MMP-14) and tissue inhibitor of metalloproteinases-2 and thus regulate MMP-2 activation in vitro (34). Interestingly, active MMP-2 is among the proteases with the capacity to activate TGF-β1 and it is therefore possible that impaired TGF-β1 activation could reflect deficient MMP-2 activity in osteopontin −/− mice (40). Further studies are necessary to clarify the role of osteopontin in regulating TGF-β1 and MMP-2 expression and activation in the lung. Nevertheless, the closely related expression pattern of osteopontin with TGF-β1 and MMP-2 supports a potentially important role for osteopontin in lung repair and fibrosis.
This work was supported by Grants HL-04343 and HL-63339 from the National Institutes of Health.
We acknowledge E. Lucey for technical assistance in administering bleomycin to the mice and K. Singh for technical assistance with immunohistochemical staining for type I collagen.
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.
- Copyright © 2004 the American Physiological Society