Pulmonary fibroblasts mobilize the membrane-tethered matrix metalloprotease, MT1-MMP, to destructively remodel and invade interstitial type I collagen barriers

R. Grant Rowe, Daniel Keena, Farideh Sabeh, Amanda L. Willis, Stephen J. Weiss


In acute and chronic lung disease, widespread disruption of tissue architecture underlies compromised pulmonary function. Pulmonary fibroblasts have been implicated as critical effectors of tissue-destructive extracellular matrix (ECM) remodeling by mobilizing a spectrum of proteolytic enzymes. Although efforts to date have focused on the catabolism of type I collagen, the predominant component of the lung interstitial matrix, the key collagenolytic enzymes employed by pulmonary fibroblasts remain unidentified. Herein, membrane type-1 matrix metalloprotease (MT1-MMP) is identified as the dominant and direct-acting protease responsible for the type I collagenolytic activity mediated by both mouse and human pulmonary fibroblasts. Furthermore, MT1-MMP is shown to be essential for pulmonary fibroblast migration within three-dimensional (3-D) hydrogels of cross-linked type I collagen that recapitulate ECM barriers encountered in the in vivo environment. Together, these findings demonstrate that MT1-MMP serves as a key effector of type I collagenolytic activity in pulmonary fibroblasts and earmark this pericellular collagenase as a potential target for therapeutic intervention.

  • collagenase
  • proteolysis
  • extracellular matrix

in lung disease, ranging from acute respiratory distress syndrome and idiopathic pulmonary fibrosis to asthma, the normal architecture of the pulmonary interstitium is perturbed, leading to defects in lung compliance as well as gas exchange (7, 16, 20, 39, 60). In overview, the pulmonary interstitium is a complex, mesh-like three-dimensional (3-D) extracellular matrix (ECM) composed of an interwoven network of structural proteins (including collagens and elastin) as well as glycoproteins and proteoglycans that collectively act as a scaffold upon which pneumocytes, capillaries, and interstitial cells are organized to form lung tissue (18). During the pathogenesis of lung disease, inflammatory responses initiated by infectious, toxic, or idiopathic challenges trigger a cycle of deregulated tissue injury and repair (16). Resembling, in many respects, a wound-healing response, this cycle results in the aberrant remodeling of ECM macromolecules, particularly triple-helical type I collagen, which comprises 50–60% of the lung ECM (16, 20, 24, 46, 48, 60). Accordingly, ECM homeostasis and type I collagen metabolism have been the subject of intense scrutiny in the context of lung disease (48, 60).

In the mature lung, pulmonary fibroblasts are the predominant cell type embedded within the interstitial ECM (18). Likewise, in damaged tissues, wound healing programs are activated, whereby 3-D ECM barriers are invaded and subsequently populated and remodeled by recruited fibroblasts (41). As matrix remodeling programs similar to those engaged during wound healing are induced during the progression of tissue-destructive lung diseases (14), efforts have focused on pulmonary fibroblasts as key effectors of deregulated ECM turnover, especially with regard to the expression and regulation of collagenolytic matrix metalloproteinases (MMPs) (2, 56, 58, 61). In overview, MMPs are a family of extracellular zinc-dependent metalloenzymes that collectively possess the ability to degrade multiple ECM components, inducing type I collagen (4). With regard to lung fibroblasts, studies to date have focused on several MMPs whose substrate repertoires are consistent with the expression of type I collagenolytic activity (43), with MMP-1 (collagenase-1) and MMP-2 (gelatinase A) proposed as central players in the ECM remodeling events associated with pulmonary disease in humans (23, 65). However, although the expression of multiple MMPs correlate with the onset of collagenolytic activity, the identification of the key proteolytic effectors that mediate type I collagen degradation by intact pulmonary fibroblasts remains unidentified. Furthermore, attempts to extrapolate results gleaned from analyses of the collagenolytic potential of other fibroblast populations are stymied by recent studies demonstrating that fibroblast gene-expression patterns are regulated in a tissue-specific fashion (6). As such, we have utilized pulmonary fibroblasts isolated from both mice and humans in an effort to characterize their type I collagenolytic and type I collagen-specific invasive activities. Although mouse and human lung fibroblasts express distinct sets of collagenolytic MMPs, we find that both populations employ the membrane-tethered MMP, MT1-MMP, as their central, direct-acting, mediator of type I collagenolytic activity. In addition, both mouse and human lung fibroblasts utilize MT1-MMP as the sole proteolytic effector of the invasive activity required during trafficking through type I collagen-rich 3-D barriers. Taken together, these results demonstrate that MT1-MMP acts as the dominant effector of type I collagen catabolic activity mobilized by pulmonary fibroblasts during tissue remodeling and invasion, with potential ramifications for the development and use of new therapeutic interventions.


Isolation of pulmonary fibroblasts.

Lungs were isolated from MT1-MMP−/−, MMP-8−/−, or MMP-13−/− mice, or their littermate controls (25, 3132, 5354), and perfused with saline via the pulmonary artery. Fibroblasts were then recovered from mouse lungs by placing minced tissues within 10-cm tissue culture-treated dishes (Corning, Corning, NY) and cultivating outgrowths in DMEM with 10% FCS, 100 U/ml penicillin, and 100 μg/ml streptomycin (all from Invitrogen, Carlsbad, CA) for 10–14 days. Fibroblasts were purified from other cell populations by differential adhesion and serial passage. Human primary lung fibroblasts [used between passages 5 and 10 and obtained under Institutional Review Board approval (IRB 00001996)] were provided by C. Hogaboam (University of Michigan).

Collagen film degradation assay.

Twelve-well tissue culture dishes were coated with a neutralized solution of acid-extracted rat tail type I collagen as described previously (53). Pulmonary fibroblasts were seeded as a 50-μl droplet in culture medium in the center of each well, and, after allowing for attachment, 1 ml of culture medium was added to each well. In these assays, cells were cultured in the absence or presence of serum alone or platelet-derived growth factor-BB (PDGF-BB, Millipore, Billerica, MA), tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), or interferon-γ (IFN-γ; all from R and D Systems, Minneapolis, MN). Assays were also performed in the presence of aprotinin, E64 (both from Sigma, St. Louis, MO), plasminogen (Enzyme Research, Swansea, Wales), or BB-2516 (British Biotechnology, Oxford, England). After 7 days, culture supernatants were collected for hydroxyproline analysis, and the cells were lysed before staining the collagen films with Coomassie Brilliant Blue (54).

Fluorescent collagen degradation and confocal microscopy.

Fibroblasts (5.0 × 104) were seeded at low density atop collagen films labeled with Alexa Fluor-594 (5354). Cell-associated collagen degradation was assessed after 3 to 7 days by confocal fluorescence microscopy.

Hydroxyproline analysis.

Culture supernatants were harvested and ethanol-precipitated for quantifying soluble hydroxyproline as described (13, 54). Concentrations of soluble hydroxyproline were transformed into total amount of degraded collagen using the known proportion of this amino acid present in the type I collagen-α(1) and -α(2) chains, assuming a hydroxyproline content of ∼13% (30).


RNA was collected for mouse or human pulmonary fibroblasts for amplification of mouse G3pdh-1, mouse collagenase A (McolA), Mmp2, Mmp8, Mmp13, Mmp14, and Mmp15 mRNAs or human GAPDH, MMP1, MMP2, MMP14, and MMP15 mRNAs as described previously (27, 29).

Human lung fibroblasts and small-interfering RNA.

Human lung fibroblasts were electroporated using a nucleofector kit and electroporation reagents (Lonza, Cologne, Germany). Small interfering RNAs (siRNAs) against MMP1, MMP14, or MMP15 as well as scrambled controls have been described previously (27, 54). MT1-MMP protein expression 48 h following siRNA electroporation was determined by Western blot analysis using a rabbit monoclonal antibody (Epitomics, Burlingame, CA).

Invasion assays.

Type I collagen hydrogels were prepared in 24-mm Transwell culture system dishes (Corning) as described previously (54). Cells were seeded in the upper well with culture medium containing 10 ng/ml PDGF-BB placed in the lower compartment of the Transwell dishes to stimulate fibroblast invasion (54). After 3 days, invasion was determined by embedding the collagen gels in paraffin, sectioning, and staining with hematoxylin and eosin. For 3-D invasion assays, 50,000 lung fibroblasts were suspended in 100 μl of type I collagen, and the mixture was allowed to gel at 37°C within a single well of a 96-well tissue culture plate. Following collagen polymerization, the fibroblast-impregnated gel was removed and placed in the center of a single well in a 24-well tissue culture plate. An outer, 500-μl, cell-free gel was then cast around the fibroblast-collagen plug (53). The culture medium was supplemented with DMEM containing 10% FCS and 10 ng/ml PDGF-BB. Fibroblast migration from the inner collagen plug into the outer, cell-free gel was monitored over 7 days by phase-contrast microscopy (53). Results are expressed as the means ± SE.

Two-dimensional migration assay.

Cells were seeded within a cloning cylinder atop a type I collagen-coated plate. After 12 h, the cylinder was removed, and cell migration from the cylinder area into the surrounding field was tracked by phase-contrast microscopy over a 4-day culture period. Results are expressed as the means ± SE.

Animal use.

All protocols used in these studies were submitted and approved by the University of Michigan Committee on Use and Care of Animals.


All statistical analyses were performed using unpaired Student's t-test, assigning a significance cutoff of <0.05. All data are expressed as means ± SE.


Collagen-degradative activity of pulmonary fibroblasts.

Pulmonary fibroblasts isolated from 3 to 8-wk-old mouse lung tissue display a bipolar, fusiform morphology under conventional, two-dimensional (2-D) culture conditions atop type I collagen films (Fig. 1A). Immunofluorescence microscopy utilizing an antibody against α-smooth muscle actin (α-SMA) revealed the presence of numerous α-SMA-containing fibers (Fig. 1A) characteristic of myofibroblasts (24, 50, 58). To assess the collagen-degrading capacity of pulmonary fibroblast cultures, the cells were cultured atop a 3-D bed of type I collagen fibrils, and the solubilization of type I collagen-associated hydroxyproline into small Mr fragments was quantified (5354). As shown in Fig. 1B, pulmonary fibroblasts display similar levels of collagenolytic activity under either serum-free conditions or in the presence of 10% FCS. Establishing pulmonary fibroblast cultures at low vs. high cell density to vary myofibroblast numbers (42) did not alter collagenolytic potential (data not shown). Collagenolytic activity was not further enhanced when serum-containing cultures were supplemented with various cytokines proposed previously to regulate pulmonary fibroblast activity, including PDGF-BB, TNF-α, IL-1β, or IFN-γ (43, 50, 56, 58) (Fig. 1B). Hence, subsequent experiments were performed in the presence of serum without exogenous growth factors in an antiprotease-rich environment that approximates that found within the interstitium at wound sites (63).

Fig. 1.

Type I collagen degradation by mouse pulmonary fibroblasts. A: pulmonary fibroblasts were isolated from mouse lung, seeded atop type I collagen films, and viewed by phase contrast microscopy (left; scale = 50 μm) or immunostained with an antibody against α-smooth muscle actin (α-SMA) (right; scale = 30 μm). B: pulmonary fibroblasts were seeded atop type I collagen films in the absence or presence of 10% FCS or various stimuli. After 7 days, culture supernatant was collected and soluble hydroxyproline quantified. Cytokines were used at the following concentrations: platelet-derived growth factor-BB (PDGF, 10 ng/ml), transforming growth factor-α (TNF-α, 10 ng/ml), interleukin-1β (IL-1β, 5 ng/ml), or interferon-γ (IFN-γ, 10 ng/ml). Results are presented as the mean ± 1 SD (n = 4). C: collagen films were incubated alone or with pulmonary fibroblasts that were cultured in the center of a collagen gel for 3 days in the absence or presence of 10% FCS, lysed, with the collagen gel visualized by staining with Coomassie Brilliant Blue. D: cells were seeded atop Alexa Fluor-594-labeled type I collagen and counterstained either with calcein (left) or Alexa Fluor-488-labeled phalloidin and TOTO-3 iodide (right, scale = 20 μm).

To directly visualize the collagenolytic potential of pulmonary fibroblasts, a droplet of cells was placed atop a layer of type I collagen fibrils such that the adherent cells were confined to the matrix substratum in an isolated island in the center of the culture dish (44). After a 6-day culture period, cells were lysed, and the culture substratum was stained with Coomassie Brilliant Blue to visualize the remaining collagen matrix (i.e., zones of collagenolysis will not stain blue with Coomassie and appear as “clear” zones). In this assay system, pulmonary fibroblast-mediated collagen degradation was restricted to subjacent zones at the site of cell attachment, demonstrating that degradation was confined to the fibroblast-collagen interface (Fig. 1C). To examine collagen degradative events at the single-cell level, pulmonary fibroblasts were next cultured atop a layer of Alexa Fluor-594-labeled type I collagen fibrils, and, after a 7-day culture period in the presence of 10% serum, collagenolysis was visualized by confocal laser microscopy. Under these conditions, fibroblasts restricted proteolytic activity to subjacent collagen fibrils associated with the ventral aspect of the cell (Fig. 1D).

MMP-dependent type I collagen degradation by pulmonary fibroblasts.

As various members of the serine, cysteine, and metalloprotease families have been proposed to participate directly or indirectly in type I collagenolytic processes (52), class-specific inhibitors of these proteases were utilized to identify the effectors mobilized by pulmonary fibroblasts (54). Pulmonary fibroblasts cultured in the presence of either aprotinin [an inhibitor of serine proteases (26)] or E64 [a cysteine protease inhibitor (54)] retained full collagenolytic activity (Fig. 2, A and B). In contrast, BB-2516, a broad-spectrum peptidomimetic MMP inhibitor (5), completely abrogated fibroblast subjacent type I collagen degradation and hydroxyproline release (Fig. 2, A and B), identifying MMPs as the dominant purveyor of type I collagen degradative activity mobilized by mouse pulmonary fibroblasts.

Fig. 2.

Matrix metalloproteinase (MMP)-mediated type I collagen degradation by pulmonary fibroblasts. A: fibroblasts were cultured in the center of type I collagen films in 10% serum under control (no inhibitor) conditions or in the presence of 100 μg aprotinin, 100 μM E64, or 5 μM BB-2516 for 7 days when cells were lysed and type I collagen substratum stained with Coomassie Brilliant Blue. B: hydroxyproline solubilization was quantified in the presence of the various protease inhibitors and presented as percent control measurements. Results presented as mean ± 1 SD (n = 3; *P = 0.001).

MT1-MMP mediates local type I collagenolysis in mouse pulmonary fibroblasts.

Among the 23 members of the murine MMP family, only a subset of the proteases have been reported to participate in type I collagenolysis: mColA, MMP-2, MMP-8, MMP-13, and the membrane-anchored MMPs, MT1-MMP (MMP-14) and MT2-MMP (MMP-15) (4). As assessed by RT-PCR, MMP expression in mouse pulmonary fibroblasts was restricted to MMP-2, MMP-13, and MT1-MMP under standard culture conditions (Fig. 3A). To determine the relative roles of each of these MMPs in collagenolysis, pulmonary fibroblasts were isolated from transgenic mouse strains, wherein the genes encoding each of these type I collagenases were disrupted (25, 3132, 54) and the cells cultured atop type I collagen substrata. Although wild-type (WT), MMP-2-deficient, and MMP-13-deficient fibroblasts displayed indistinguishable patterns of collagenolytic activity, MT1-MMP-deficient mouse pulmonary fibroblasts were unable to degrade subjacent collagen (Fig. 3B). Similar results were obtained when subjacent collagenolysis was assessed by confocal laser microscopy or hydroxyproline solubilization (Fig. 3C). Although MT1-MMP can process MMP-2 or MMP-13 zymogens to their active forms (46, 53), type I collagenolysis was unaffected in MMP-2- or MMP-13-deficient fibroblasts, precluding a requirement for either an MT1-MMP/MMP-2 or MT1-MMP/MMP-13 proteolytic cascade under these conditions.

Fig. 3.

Membrane type-1 (MT1)-MMP is required for type I collagenolysis by pulmonary fibroblasts. A: RNA was isolated from wild-type (WT) mouse lung fibroblasts and mRNA levels of G3pdh-1, mouse collagenase A (McolA), Mmp2, Mmp8, Mmp13, Mmp14, and Mmp15 were measured by RT-PCR. B: pulmonary fibroblasts isolated from WT mice or mice with targeted genetic deficiencies in Mmp2, Mmp13, or Mmp14 were cultured in the center of type I collagen substrata for 7 days, when residual type I collagen was visualized by staining with Coomassie Brilliant Blue. C: mouse fibroblasts were cultured atop substrata of Alexa Fluor-594-labeled type I collagen for 7 days, and subjacent collagenolysis was examined by confocal laser microscopy (scale = 20 μm). Type I collagen degradation by each fibroblast strain was quantified by hydroxyproline measurement in the culture supernatant (mean ± 1 SD; n = 3; *P ≤ 0.0002). D: fibroblasts were cultured either atop a type I collagen substratum (left, scale = 50 μm) or embedded within a type I collagen hydrogel (right, scale = 200 μm) and cell morphology assessed after 7 days by phase contrast microscopy. E: hydroxyproline solubilization was measured WT fibroblasts under the indicated culture conditions as well at MT1-MMP-deficient fibroblasts cultured within 3-D type I collagen (mean ± 1 SD; n = 3; *P = 0.0015 compared with WT 3-D cells).

In vivo, fibroblasts localized within the pulmonary interstitium are embedded in a 3-D fibrillar network of type I collagen fibrils (18). As accumulating evidence indicates that cell behavior, including fundamental processes such as proliferation, motility, and differentiation, is regulated via distinct mechanisms in the 3-D as opposed to the 2-D ECM environment (9, 28, 52, 64), fibroblasts were alternatively cultured within, rather than atop, 3-D gels of type I collagen. Compared with the morphology of spindle-shaped cells observed under 2-D culture conditions, pulmonary fibroblasts dispersed in 3-D collagen hydrogels assumed a multipolar, stellate morphology (Fig. 3D). Nevertheless, 3-D-embedded fibroblasts likewise mobilize an MMP-dependent, BB-2516-sensitive process to degrade type I collagen to a degree similar to that observed under 2-D culture conditions (Fig. 3E). More importantly, as observed in 2-D culture, 3-D embedded fibroblasts continued to employ MT1-MMP as the dominant collagenolytic effector (Fig. 3E). Hence, pulmonary fibroblasts use MT1-MMP as a direct-acting collagenase to effect type I collagenolytic behavior under conditions similar to those likely to be encountered in the in vivo setting.

MT1-MMP is required for pulmonary fibroblast infiltration of type I collagen barriers.

Mounting evidence suggests that pericellular collagenolytic activity is an essential prerequisite for cell movement within type I collagen-rich interstitial matrices (38, 5354). As such, WT or MT1-MMP-deficient fibroblasts were cultured atop 3-D collagen barriers and stimulated to express invasive behavior in the presence of the fibroblast chemoattractant, PDGF-BB (58). Whereas WT fibroblasts were able to effectively invade the type I collagen gel by a BB-2516-dependent process, MT1-MMP-deficient cells fail to display invasive activity (Fig. 4A). Similarly, when embedded in a 3-D collagen plug surrounded by an outer, cell-free type I collagen matrix, pulmonary fibroblasts continue to exhibit an absolute requirement for MT1-MMP in supporting 3-D-specific invasive behavior (Fig. 4B). Furthermore, the collagen-invasive activity of WT pulmonary fibroblasts was unaffected when serum-derived MMP-2 was depleted by gelatin-Sepharose chromatography (28, 33) or when invasion was monitored in the presence of the inhibitor, tissue inhibitor of metalloproteinase-1, a potent inhibitor of the secreted collagenases, but not MT1-MMP (Ref. 26 and data not shown). Importantly, a required role for MT1-MMP in 3-D invasive activity was not due to changes in fibroblast viability (i.e., 94% viability in WT fibroblast cultures vs. 94% viability in MT1-MMP-deficient fibroblasts after a 5-day culture period under 3-D conditions as assessed by ApopTag in situ apoptosis detection). In contrast to the requirement for MT1-MMP in supporting invasion through 3-D type I collagen barriers, neither WT fibroblasts cultured in the presence of BB-2516 nor MT1-MMP-deficient fibroblasts display defects in motile responses as assessed under 2-D conditions atop collagen-coated surfaces (Fig. 4C).

Fig. 4.

MT1-MMP is required for pulmonary fibroblast invasion of type I collagen extracellular matrix barriers. A: WT fibroblasts with or without 5 μM BB-2516 or MT1-MMP−/− fibroblasts were cultured atop type I collagen hydrogels in the upper chamber of a Transwell tissue culture plate insert with a chemotactic stimulus of 10 ng/ml PDGF-BB in the lower well for 3 days. Collagen gels were then fixed, sectioned, and stained with hematoxylin and eosin and examined by light microscopy (scale = 100 μm). Number of invading cells were quantified. A representative example of 3 individual experiments is presented. B: WT fibroblasts with or without 5 μM BB-2516 or MT1-MMP−/− fibroblasts were cultured within small (100 μl) collagen gels embedded in an outer, cell-free collagen gel, and invasion from the inner into the outer gel in the presence of 10% serum and 10 ng/ml PDGF-BB was monitored for 10 days. Invasion was assessed by phase-contrast microscopy (top, scale = 200 μm) or following staining with phalloidin (green, marking F-actin) or TOTO-3 iodide (red, marking nuclei, bottom, scale = 100 μm). Results are presented as the mean ± 1 SD (n = 3). C: WT fibroblasts with and without 5 μM BB-2516 and MT1-MMP−/− fibroblasts were cultured in the center of type I collagen substrata. Cell migration atop the 2-D substratum was monitored over 4 days by phase contrast microscopy (scale = 1 mm). Representative results of 3 independent experiments are presented. The migratory front is indicated by black lines.

MT1-MMP is the central type I collagenase in human pulmonary fibroblasts.

Although mouse fibroblasts exclusively rely on MT1-MMP to degrade or invade type I collagen barriers, mice do not possess a gene-encoding MMP-1, the dominant interstitial collagenase present in human fibroblasts (54). Indeed, as opposed to mouse fibroblasts, human pulmonary fibroblasts express MMP-1 rather than MMP-8 or MMP-13 (i.e., the primary collagenases detected in mouse lung fibroblasts; Fig. 3A) and express both MT1-MMP as well as the membrane-anchored collagenase MT2-MMP (Fig. 5A). Nevertheless, human lung fibroblasts, like their mouse counterparts, degraded subjacent type I collagen by a process insensitive to aprotinin or E64 but completely inhibited by BB-2516 (Fig. 5, B and C). While subsets of human fibroblast populations possess the ability to convert plasminogen to plasmin, which can, in turn, activate the secreted collagenases (53), increased collagenolytic activity was not detected when human lung fibroblasts were cultured in the presence of plasminogen (Fig. 5B).

Fig. 5.

Human lung fibroblasts utilize MMPs to degrade type I collagen. A: expression levels of the type I collagenolytic MMPs were assessed by semiquantitative RT-PCR. B: human lung fibroblasts were cultured atop a type I collagen substratum in the presence of 10% serum without protease inhibitors or in the presence of 100 μg/ml aprotinin, 100 μM E64, 5 μM BB-2516, or 20 μg/ml plasminogen. After 7 days, cells were lysed, and the substrata were stained with Coomassie Brilliant Blue. C: human lung fibroblasts were cultured atop Alexa Fluor-594-labeled type I collagen in the presence of the indicated inhibitors for 7 days, and subjacent collagenolysis was examined by confocal laser microscopy (scale = 20 μm).

To determine whether MT1-MMP-dependent pericellular collagenolysis is a mechanism conserved in human pulmonary fibroblasts, siRNAs were directed against MMP1, MMP14, or MMP15. Specific targeting of each transcript as well as MT1-MMP protein expression was confirmed without detectable off-target effects on other MMP mRNAs (Fig. 6, A and B), and the silenced fibroblasts were cultured atop fluorescently labeled type I collagen gels. In accordance with results obtained with mouse lung fibroblasts, silencing of MT1-MMP completely abrogated collagenolysis, whereas targeting of other MMPs had no discernable effect on subjacent degradation or hydroxyproline release (Fig. 6, C and D). Notably, a compensatory role for MT2-MMP in type I collagenolysis was not observed when MT1-MMP expression was silenced (Fig. 6, C and D). Moreover, abrogation of MT1-MMP activity in human fibroblasts by transfection with an MMP14 siRNA nearly completely inhibited fibroblast infiltration of 3-D type I collagen barriers (i.e., SCR 182.7 ± 12.8 μm invasion; siMMP14 35.3 ± 0.9 μm invasion; P = 0.0003; Fig. 6E). Taken together, these results identify MT1-MMP as the key type I collagenase mobilized in both mouse and human pulmonary fibroblasts.

Fig. 6.

Human lung fibroblasts require MT1-MMP for type I collagenolysis. A: human lung fibroblasts were transfected with siRNAs directed against the transcripts encoding MMP1 and MMP14 or a scrambled MMP14 sequence (SCR), and levels of the various MMP transcripts were measured by semiquantitative RT-PCR. B: Western blot analysis demonstrating silencing of MT1-MMP protein expression in human pulmonary fibroblasts. C: human lung fibroblasts transfected with the indicated siRNAs were cultured atop Alexa Fluor 594-labeled type I collagen substrata for 7 days, when subjacent collagen degradation was examined by confocal laser microscopy (scale = 30 μm). D: human lung fibroblasts transfected with the indicated siRNAs were cultured atop type I collagen substrata for 7 days, and hydroxyproline levels in the culture supernatant were quantified (mean ± 1 SD; n = 3; *P = 0.0001 vs. SCR condition). E: human lung fibroblasts were transfected with the indicated siRNAs and cultured within a 3-D collagen matrix embedded within an outer, cell-free collagen gel in the presence of 10 ng/ml PDGF-BB for 10 days. White broken line indicates the interface of the cell-containing and cell-free gels, and the black broken line indicates the invasion front (scale = 200 μm).


Pulmonary fibroblasts have been implicated as key effectors of lung ECM turnover given the central role that these cells play in mediating the synthesis and turnover of type I collagen, the dominant component of the pulmonary interstitium (41, 43, 50, 54, 61). However, despite several studies demonstrating that pulmonary fibroblasts remodel type I collagen, as well as considerable investigation into the regulation of collagen synthesis and MMP expression in these cells (2, 1112, 2021, 37, 43, 46, 56, 58, 65), the functional proteolytic enzymes underlying type I collagen catabolism have remained undefined. Degradation of type I collagen by “true” collagenases that cleave within triple-helical domains occurs when the molecule is specifically cleaved between glycine-775 and isoleucine-776 in the α1(I) chain and glycine-775 and leucine-776 within the α2(I) chain, resulting in the loss of tertiary structure at 37°C (4, 34). Multiple MMPs, including MMP-1, MMP-2, MMP-8, MMP-9, and MMP-13, as well as MT1-MMP and MT2-MMP, have been implicated in mediating type I collagenolytic activity (1, 34, 53). Furthermore, cathepsin K, a lysosomal cysteine protease possessing type I collagenolytic activity, is also expressed by pulmonary fibroblasts (8, 62). Given the multiple components that potentially participate in type I collagenolysis, as well as the oft assumed functional redundancy of proteolytic networks, a priori one might expect that identification of a single key protease required is unlikely. However, although the mouse lung fibroblasts used in our study expressed MMP-13, MMP-2, and MT1-MMP, only pulmonary fibroblasts derived from genetically modified mice deficient in MT1-MMP activity displayed significant defects in type I collagen degradation. Although MT1-MMP has been proposed to exert its ECM degradative effects via direct activation of pro-MMP-2 or pro-MMP-13 (46, 57), fibroblasts deficient in MMP-2 or MMP-13 exhibited an unaffected type I collagenolytic activity relative to WT cells.

Compared with the soluble collagenases expressed by pulmonary fibroblasts, why does the singular loss of MT1-MMP result in a complete defect in type I collagen catabolism? Within interstitial fluid, modeled by the presence of serum in our culture system, soluble proteases are secreted into an environment rich in antiproteases at concentrations far in excess of their targets, resulting in efficient inhibition of protease activity (53). Indeed, even in the presence of exogenous plasminogen, which, following its proteolytic processing to plasmin, activates multiple type I collagenolytic MMPs (53), an increase in collagen degradative activity was not observed in pulmonary fibroblasts. In contrast, the directed secretion of an intracellularly activated, membrane-anchored MMP activity to sites of cell-ECM contact results in the intimate positioning of the active protease to its subjacent targets relative to secreted MMPs, thereby minimizing exposure to soluble antiproteases and resulting in more efficient proteolysis (5254). Indeed, recent studies have demonstrated that MT1-MMP function can be recapitulated by tethering a mutant MMP-1 construct to the cell surface (53), highlighting the fact that cell surface-localized MMP activity is the key requirement for effective pericellular ECM degradation within an antiprotease-rich environment. However, in pathological conditions wherein the protease shield is compromised, such as in α1-antitrypsin deficiency or during prolonged oxidative stress, bulk ECM degradation by soluble proteases could contribute to widespread ECM remodeling, perturbing tissue architecture (16, 46, 63). Thus soluble and membrane-tethered MMPs likely participate in two functionally distinct proteolytic axes that are separately responsible for bulk and localized ECM turnover, respectively (53). Although we have implicated MT1-MMP as the key regulator of pulmonary fibroblast pericellular collagenolysis, the role of nonspecific, bulk ECM degradation by secreted MMPs in the context of chronic lung diseases requires further investigation. Furthermore, as MMP substrates include a variety of targets that lie outside the ECM per se, including growth factors, cytokines, and adhesion molecules (19, 47, 49), it seems likely that MT1-MMP-independent proteolytic effectors play important roles in controlling inflammatory responses within the pulmonary bed (51).

Accumulating evidence indicates that distinct patterns of regulation of central fibroblast functions active during tissue remodeling, such as proliferation or migration, are recruited in cells embedded within the confines of a constraining 3-D ECM relative to classic 2-D culture conditions (10, 28). Although recent work has identified required roles for MT1-MMP in dermal fibroblast infiltration of type I collagen-rich ECMs, fibroblasts exhibit marked variability in transcriptional profile depending on anatomical site (6), raising the possibility that distinct mechanisms regulate fibroblast activity in a tissue-specific manner. Somewhat surprisingly, given the genetic diversity of fibroblasts as a function of tissue location (6), we find that MT1-MMP maintained its dominance as the key effector of type I collagen degradation in pulmonary fibroblasts, indicative of a global role for this enzyme as the central type I collagenase in fibroblasts operative in diverse anatomic locations.

Fibroblast traffic through ECM plays an integral role in tissue injury and repair (41). MT1-MMP is ideally suited to support such invasive processes, given its membrane localization, thus allowing the protease to focally direct degradation to sites of cell ECM contact (38). Indeed, MT1-MMP-deficient mouse pulmonary fibroblasts exhibited a complete defect in type I collagen-invasive activity. Recent reports conclude that MT1-MMP can modulate cell motility independently of its proteolytic activity by regulating intracellular signaling pathways (15, 36, 45); thus, the observed inhibition of type I collagen infiltration in the setting of MT1-MMP depletion might result from such a catalysis-independent, motility-regulating pathway. However, we observed identical 2-D, planar migration of WT and MT1-MMP-deficient fibroblasts when cultured atop a type I collagen substratum, indicating that such protease-independent motility-modulating mechanisms are not operative in pulmonary fibroblasts. Moreover, although it has been proposed that, in the absence of MMP activity, cells adopt a protease-independent, “amoeboid” compensatory mechanism of ECM invasion (22), no such effect was observed in MT1-MMP-depleted pulmonary fibroblasts. This result is attributable to the description of protease-independent amoeboid migration exclusively in cells cultured within ECM constructs lacking the covalent cross-links present in native ECMs, resulting in the construction of connective tissue barriers whose relevance to in vivo ECM structure remains questionable (55). Therefore, our findings suggest that localized, highly specific ECM degradation by MT1-MMP allows cells to execute specific, controlled “tunneling” through tissue barriers while preserving sufficient ECM structure to create a substratum permissive for efficient migration (52).

Although we have identified MT1-MMP as the key enzyme active in type I collagenolysis operative within pulmonary fibroblasts, further studies are required to assign a mechanistic role to this process in models of chronic lung diseases such as idiopathic pulmonary fibrosis, asthma, or emphysema. Whereas fibrotic lung diseases such as idiopathic pulmonary fibrosis are characterized by excessive accumulation of type I collagen-rich tissue (16), destructive pulmonary diseases such as emphysema occur when unbridled breakdown of normal interstitial ECM occurs (46). Given the potent ability of MT1-MMP to trigger digestion of ECM-bound type I collagen fibrils, a model might be proposed wherein MT1-MMP actively contributes to the progression of lung pathology characterized by excessive ECM catabolism, while playing a protective role in diseases, wherein excessive ECM deposition occurs by catabolically counterbalancing upregulated type I collagen anabolism and deposition. Thus either activation or inhibition of MT1-MMP activity could play a central role in the pathogenesis of these two classes of pulmonary disease, respectively. Indeed, MT1-MMP seems to be overexpressed in emphysema, bronchial asthma, and bronchiectasis compared with healthy lung tissue (40, 46). Moreover, MT1-MMP expression has been observed in mouse models of idiopathic pulmonary fibrosis and in human idiopathic fibrosis tissue (23, 35), possibly reflective of a compensatory, procollagenolytic response to the aberrant, excessive ECM deposition characteristic of this disease. Alternatively, despite the collagenolytic potential of MT1-MMP, other investigators posited a profibrotic role for the proteinase as a consequence of its ability to activate latent TGF-β1 (59). Subsequent research using genetically modified mouse models of destructive and fibrotic lung diseases should interrogate MT1-MMP as a potential therapeutic target in pulmonary pathology. With the development of highly specific monoclonal antibodies directed against human MT1-MMP (17), such investigation could define a novel therapeutic route for management of chronic lung diseases.


This work was supported by NIH Grant CA71699 and the Breast Cancer Research Foundation to S. J. Weiss. Work performed in this study was also supported by MDRTC Cell and Molecular Biology Core NIH Grant P60 DK020572.


No conflicts of interest, financial or otherwise, are declared by the authors.


Present address for R. G. Rowe: Dept. of Medicine, Children's Hospital Boston, Boston, MA 02115. Present address for D. Keena: Dept. of Internal Medicine, William Beaumont Hospital, Oakland University, Rochester, MI 48309.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
View Abstract