matrix metalloproteinases (MMPs) are a family of zinc-dependent, neutral endopeptidases that degrade the extracellular matrix and many nonmatrix substrates (35, 37). MMPs have been implicated in remodeling of tissue in development, involution, and metastasis of cancer cells (42). The expression of MMPs also occurs in injured airway epithelium. MMP-7 and MMP-9 are produced by airway epithelial cells after mechanical wounding, and both are required for cellular migration (11, 25). However, in pathological states, many other MMPs are present including membrane type 1 MMP (MT1-MMP, MMP-14) and are expressed in the airway epithelium (27, 33, 34). Whether MT1-MMP or other MMPs produced by airway epithelial cells are required for airway epithelial repair has not been established.

MT1-MMP is a membrane-tethered MMP with fibrillar collagenase activity that is required for long bone growth and molar eruption (16, 51). In addition, MT1-MMP cleaves several proteins including α_v integrin, syndecan-1, tissue transglutaminase, and laminin-332 (formerly laminin-5) that suggest it has a role in epithelial cell migration (2, 3, 5, 9). Consistent with this speculation, MT1-MMP is essential for cancer cell migration across covalent cross-linked extracellular matrix barriers (18, 19). MT1-MMP also liberates EGF ligands (23) and activates TGF-β (31), suggesting that expression of this protease by the respiratory epithelium has important consequences. MT1-MMP is not present in the normal airway epithelium, but is upregulated in the airways of subjects with chronic obstructive pulmonary disease (COPD) and asthma (27, 33, 34). Previously, we (1) and others (20, 32) have demonstrated that MT1-MMP knockout (KO) mice fail to properly develop alveolar septations but have normal airway branching and airway epithelial cell differentiation.

Since MT1-MMP is expressed by the airway epithelium in lung diseases and is required for epithelial tumor invasion, we hypothesize that MT1-MMP is for epithelial cell migration after airway epithelial injury. To induce airway epithelial-specific injury, we exposed wild-type (WT) and MT1-MMP KO mice to the cytotoxicant naphthalene (NA). Since the stages of epithelial cell repair after NA-induced injury are similar to that seen with viral or bacterial injury only without the robust infiltration of inflammatory cells (24, 44), NA-induced injury enabled examination of epithelial repair without excess macrophages and lymphocytes, which also produce MT1-MMP (4, 41).

Rather than defective epithelial cell migration, we found that MT1-MMP KO mice have decreased airway epithelial cell proliferation. Although many factors may contribute to this phenotype, depressed keratinocyte growth factor receptor (KGFR) expression on the basal surface of airway epithelial cells correlated with the failure in epithelial repair in MT1-MMP KO mice.

MATERIALS AND METHODS

Mice.

MT1-MMP KO mice were developed as previously described (16). The mice used in these experiments were littermates from a heterozygote breeding of a mixed NIH Black Swiss/129ReJ background. Uninjured MT1-MMP KO mice live for ∼2–3 mo. All experiments were initiated when mice were 6 wk of age, a time point when Clara cells are at full adult number and maturity (22). All animal procedures were approved by the Washington University Animal Studies committee and performed in accordance with National Institutes of Health guidelines.

Antibodies.

Goat anti-CC10 antibody from Dr. G. Singh (Pittsburgh Veteran Affairs Medical Center) was used at 1:25,000 dilution. Mouse anti-forkhead homeobox protein J1 (anti-Foxj1) antibody from Dr. S. Brody (Washington University, St. Louis) was used at 1:2,000 dilution. Rabbit anti-KGFR (from Dr. T. Koji, Nagasaki University School of Medicine, Japan) was used at 1:8,000 dilution. Mouse anti-Ki-67 (BD Pharmingen, San Jose, CA) was used at 1:100 dilution. Rat anti-bromodeoxyuridine (anti-BrdU; Accurate Chemical, Westbury, NY) was used at 1:1,000 dilution, and anti-phosphotyrosine-EGFR (Tyr1045) and anti-phosphotyrosine-HER2 (Tyr1221) (Cell Signaling Technology, Danvers, MA) were used at 1:200 dilution.

NA injury model.

NA (Sigma-Aldrich) was dissolved in corn oil (Sigma-Aldrich) and sterile filtered. Mice were given a single dose of NA (100 mg/kg ip) or corn oil alone as vehicle-alone controls. All exposures were performed between 8–10 AM on 6-wk-old mice. Mice were killed at 1, 2, 3, 5, 14, and 28 days after exposure. Lungs were inflation-fixed with 10% phosphate-buffered formalin to 25 cmH₂O pressure, and lobes were separated and embedded in a parasagittal manner to maximize surface area per slide. For electron microscopy experiments, lungs were inflation-fixed with 300 mosmol/kgH₂O Karnovsky fixative using electron microscopy grade paraformaldehyde and glutaraldehyde as previously described (1). For proliferation studies, mice were treated with 50 mg/kg ip BrdU (Sigma-Aldrich) 1 h before death.

Immunohistochemistry and immunofluorescence.

Paraffin-embedded lungs were cut into 5-μm sections and stained with antibodies as noted above. Vector ABC goat, rat, rabbit, and mouse-on-mouse detection kits (Vector Labs, Burlingame, CA) were used as per manufacturer's directions. Staining was developed using diaminobenzidine substrate with nickel (Vector Labs) as per manufacturer's directions. Fluorescent secondary antibodies (The Jackson Laboratory, Bar Harbor, ME) were used at a 1:400 dilution. For the Ki-67 and Foxj1 antibodies, antigen retrieval was performed using a pressure cooker (Biocare Medical, Concord, CA) in a pH 8.0 Tris buffer for 5 min. For the BrdU antibody, tissue was treated with porcine pancreatic trypsin (Sigma, St. Louis, MO) at 0.05% in PBS at 37°C for 30 min followed by 4 N HCl for 30 min before staining.

MT1-MMP in situ hybridization.

A probe specific for mouse MT1-MMP was developed as previously described (17). After vector linearization, sense and antisense digoxigenin (Dig)-labeled probes were generated using the Dig RNA labeling mix (Roche) and SP6 or T7 RNA polymerase, respectively. Five-micrometer sections were deparaffinized, rehydrated, and treated with proteinase K (5 μg/ml) for 10 min at room temperature and then fixed in paraformaldehyde. Slides were incubated in acetylation solution (triethanolamine, pH 8, HCl, acetic anhydride) for 10 min to reduce nonspecific, charge-based hybridization. Hybridization with 20 ng of probe was performed overnight at 65°C in a humidity chamber. Then, slides were sequentially washed with 5× SSC at room temperature for 5 min, 1× SSC/50% formamide at 60°C for 30 min, and TNE (10 mM Tris, pH 7.5, 500 mM NaCl, 1 mM EDTA, pH 8) for 10 min at 37°C. Sections were treated with RNase A (20 μg/ml in TNE) for 10 min and then washed again in 2× SSC followed by 0.2× SSC at 60°C. Dig detection was performed with the anti-Dig-AP antibody (Roche) and developed with the BM purple substrate (Roche) for 5 days with one change of the substrate solution as per manufacturer's directions. Slides were counterstained with tartrazine yellow.

Quantification.

All quantification was performed on terminal airway epithelium, defined as airway epithelial cells within 100 μm of the bronchoalveolar junction. Since injury and repair can be airway level-dependent, and airway size cannot be accurately determined in two dimensional sections, only airways with a visible connection to the alveolar spaces were used. Ten terminal airways (2–3 different sections at least 50 μm apart) from at least five different mice were examined. All quantified data were expressed as number of positive cells per 100-μm length of basement membrane and not as a percentage of total nuclei because nuclear number underestimates cell number in cytoplasmic stains, like CC10, on 5-μm thick sections.

For quantification of KGFR expression, individual positive cells were difficult to delineate, so image analysis software (Openlab 4.0; Improvision, Lexington, MA) was used to measure the surface area of staining in a region of interest within the last 100 μm of five terminal airways per mouse on binary black and white images (nickel-containing stain is black). For statistical analysis, an ANOVA test was performed with a Scheffé's post hoc analysis for conditions found to be significant using SPSS 13 software (SPSS, Chicago, IL).

Scanning electron microscopy.

Tissue was prepared as previously described (1). Briefly, the right middle lobe was divided where the distal tip angles away from the rest of the lobe. The pleural surface of both the distal tip and the rest of the medial lobe were affixed to a glass coverslip using cyanoacrylate. Then, using a dissecting microscope, the airways were exposed to the level of alveolar ducts. Mucus was removed by a brief toluene wash, after which specimens were dehydrated and sputter-coated with gold and then examined with a Hitachi S-450 scanning electron microscope.

RESULTS

MT1-MMP is upregulated in airways after NA exposure.

Because MT1-MMP is expressed by the airway epithelium in various diseases and may have a role in the injury repair response, we evaluated distal airways of NA-injured mice for MT1-MMP mRNA expression by in situ hybridization. NA injury and repair is characterized in stages of injury and cell sloughing (several hours to 2 days), proliferation and migration (approximately days 2–5), and differentiation (approximately days 5–14) (43). As expected, MT1-MMP-positive cells were rare in the airways of mice receiving vehicle alone (Fig. 1A) . The few airway cells that expressed MT1-MMP appeared to be subepithelial (Fig. 1A, inset). No MT1-MMP expression was detected in airway epithelial cells. In contrast, MT1-MMP expression was rapidly upregulated in airway epithelial cells after NA exposure.

We concentrated on the terminal airway (defined as the airway epithelium within 100 μm of the bronchoalveolar junction) because this is the area of greatest injury post-NA and is a consistently identifiable structure on histological sections (i.e., estimation of larger airway size can be mistaken because of oblique sectioning resulting in comparison of areas of unequal injury). Localization of MT1-MMP mRNA was apparent at 24 h (2.1 cells per 100 μm of terminal airway) with a peak at 2–5 days (6.6 and 5.6, both statistically significant compared with vehicle; P < 0.001; Fig. 1, B–D). By 14 days after NA, the number of MT1-MMP-positive cells was the same as vehicle-only controls (0 and 0.5; Fig. 1, A and E).

Although the majority of cells that expressed MT1-MMP were airway epithelial cells, some expression was detected in subepithelial, alveolar, and perivascular cells, particularly on days 2 and 5 (Fig. 1, C and D, arrowheads). It is likely that these cells were infiltrating mononuclear leukocytes because similar cells were positive for the macrophage-specific CD147b/Mac-3 antigen on serial sections (data not shown). The expression of MT1-MMP in the airway epithelium 1–5 days after NA-induced injury is consistent with a role in airway epithelial repair and peak synthesis that occurs 2–5 days after exposure suggests that MT1-MMP is not involved in the injury phase.

MT1-MMP is required for airway epithelial repair.

Since MT1-MMP expression was increased during airway epithelial repair, we evaluated MT1-MMP KO mice for alterations in repair after NA-induced injury. Six-week-old WT and MT1-MMP KO mice were administered a weight-based dose of NA (100 mg/kg), and airway histology was evaluated at time points during injury and repair process. A lower dose of NA and younger mice than typical for this model (14, 43) were used because of the limited lifespan of MT1-MMP KO mice (2–3 mo) and expectation that modest injury would be sufficient to demonstrate a phenotype without causing mortality in the KO mice. However, at this dose, some deaths were seen in the MT1-MMP KO mice (∼15% of the treated KO mice died within 24 h of NA treatment). Death was not seen in WT or vehicle-alone KO mice during early stages. Most deaths occurred in MT1-MMP KO mice 5 g or less in body wt within 1 day of NA exposure. At later time points (28 days after NA), several MT1-MMP KO mice, in both the vehicle-alone and NA groups, died (∼20% of KO mice), suggesting that the NA model cannot be evaluated in MT1-MMP KO mice past this time point.

In vehicle-alone-treated-mice, the terminal airway epithelium appeared similar between WT and MT1-MMP KO mice, with dome-shaped Clara cells lining the airways to the level of the bronchoalveolar junction (Fig. 2, A and F, bronchoalveolar junction on the right). Because there was no difference in the vehicle-only mice at any time point, all vehicle-only time points were analyzed as a single condition.

NA-exposed MT1-MMP KO mice demonstrated a similar degree of injury to the WT mice with both showing loss of airway Clara cells (Fig. 2, B and G) and squamation of the residual uninjured airway epithelial cells (Fig. 2, C and H) without evidence of basement membrane exposure. However, as early as day 5, where the airway epithelial cells started to demonstrate a cuboidal appearance in WT mice (Fig. 2D, arrowheads), MT1-MMP KO airway epithelial cells were still extensively squamated (Fig. 2I). The degree of squamation varied from airway to airway, but the presence of significant areas of squamated epithelium in MT1-MMP KO mice was still evident at day 28 (Fig. 2J). By day 28, WT mice have columnar, dome-shaped Clara cells lining the luminal surface of all distal airways (Fig. 2E, arrows).

To confirm these findings, we examined the ultrastructural appearance of the airway surface using scanning electron microscopy of microdissected airways at day 14, a time point when most of the airways of WT mice had recovered a normal-appearing epithelium. Although MT1-MMP KO airways contained some dome-shaped cytoplasmic protrusions typical of Clara cells (Fig. 2K, arrows), no cilia were present in the terminal airways of MT1-MMP KO mice. Cell-cell junctions were intact (Fig. 2K, inset, arrowheads), but the airway was covered with epithelial cells that had no discernable ultrastructural appearance (cilia or dome-shaped cytoplasmic protrusions). In contrast, in WT mice, Clara cells were abundant, and cilia were short but plentiful (Fig. 2, L and inset). This suggests MT1-MMP is required for a stage of repair before terminal differentiation and that MT1-MMP KO airways are arrested in that stage.

Airway epithelial injury is not altered in MT1-MMP KO mice.

Since MT1-MMP KO mice are smaller than age- and strain-matched WT controls, we examined the extent of injury to rule out the possibility that the lack of airway repair was the result of more severe injury in the MT1-MMP KO mice. We used a weight-based dose and have previously demonstrated that the ratio of lung-to-body weight is maintained in MT1-MMP KO mice (1). We determined the extent of injury by quantifying the number of CC10-positive Clara cells in the terminal airway, since NA induces Clara cell-specific destruction. Other measures of severity of acute lung injury cannot be used in this model as NA does not result in leakage of serum proteins as seen in acute alveolar injury.

There was no difference in the density of distal airway Clara cells between WT and MT1-MMP KO mice in any of the vehicle-only mice (Fig. 3A, vehicle). The loss of terminal airway Clara cells was similar in MT1-MMP KO and WT mice at 2 days with a small, insignificant increased loss of Clara cells in WT mice (Fig. 3A; P = 0.24). Thus the severity of injury was unrelated to MT1-MMP genotype. However, by 5 days after NA, the number of Clara cells in the WT mice began to exceed that of MT1-MMP KO mice, and the differences were statistically significant at 14 and 28 days even though WT mice had returned to baseline density by 14 days (P = 0.67).

We also determined the ciliated epithelial cell density in the same terminal airway sections to ensure MT1-MMP KO mice did not have altered damage to this cell population resulting in failure to produce cilia in their terminal airways after injury (Fig. 3B). Since cilia are lost after injury, the nuclear marker of ciliated cells, Foxj1, was used to identify cells of a ciliated lineage. Ciliated cells are much less abundant than Clara cells in the terminal airway and decrease at slightly later times after NA (day 5; P < 0.001), but the decrease in Foxj1-positive cells was not significantly different between WT mice and MT1-MMP KO mice (Fig. 3B; P = 0.53).

As noted in the histological sections, there was a persistent decrease in Clara cell number of the MT1-MMP KO mice (Fig. 3A). At day 14, when WT mice reconstituted to near baseline Clara cell density (Fig. 3C), MT1-MMP KO airways had only small clusters of CC10-producing Clara cells (Fig. 3D, arrows) with large intervening areas that lacked Clara cells including the bronchoalveolar junction (Fig. 3D, arrowheads) compared with the bronchoalveolar junction in vehicle-treated MT1-MMP KO mice (Fig. 3E, arrowheads). This suggests that MT1-MMP was not required for differentiation into Clara cells but rather MT1-MMP was necessary for the reconstitution of normal epithelial cell density.

MT1-MMP is required for airway epithelial cell proliferative responses.

Since the cell density was decreased in the MT1-MMP KO terminal airways from days 5–28 (Figs. 3A and 2K), we postulated that there was a defect in distal airway epithelial cell proliferation. We examined proliferation in the terminal airway epithelium by immunostaining for expression of the Ki-67 antigen, a marker of entry into the cell cycle. There was no measurable proliferation in WT or KO mice after vehicle-only on day 1 (data not shown), but significant increases in Ki-67-positive cells were present in WT mice on days 2 and 5 (Fig. 4). In MT1-MMP KO mice, there were significantly fewer Ki-67-positive cells at days 2 and 5 than WT littermates. Proliferation seen in the terminal airways of KO mice was increased only at day 5 compared with vehicle-treated KO mice. In addition, there was essentially no Ki-67 staining in WT or MT1-MMP KO epithelium at days 14 and 28 (data not shown) despite persistent decreased cell density in the MT1-MMP KO mouse terminal airways. This result suggests that MT1-MMP expression that started at day 1 and peaked at days 2–5 both correlated with and was required for distal airway epithelial proliferation.

To confirm these findings, additional mice (5 KO, 5 WT) were pulsed with BrdU 1 h before tissue collection. Colocalization of BrdU and CC10 by immunofluorescence demonstrated that most of the proliferating cells in the terminal airways of WT mice 2 days after NA exposure were CC10-positive Clara cells (3.8 BrdU-positive cells per terminal airway, 93% Clara copositive; Fig. 5F, white arrows). However, MT1-MMP KO airways had very few BrdU-positive cells especially in the terminal airways (0.7 cells, 64% Clara; Fig. 5H, white arrowheads). The slight increase in proliferation seen with BrdU when compared with Ki-67 likely represents the sensitivity of the method as the numbers in both the KO and WT mice increased. However, many of the BrdU-positive cells in MT1-MMP KO mice did not appear to be epithelial with a subepithelial location and lack of CC10 staining (Fig. 5I, gray beveled arrows). Accordingly, the failure of MT1-MMP KO mice to repair terminal airways after NA-induced injury involves a failure in the proliferative response of the Clara cells in the distal airway epithelium.

EGF-mediated signaling is not responsible for defective airway epithelial repair in MT1-MMP KO mice.

Since EGF ligands and receptors increase and redistribute after NA-induced injury (45), and MMPs can liberate EGF ligands (23, 40), we performed immunohistochemistry for the phosphorylated form of the EGF receptor (pEGFR) in WT and MT1-MMP KO mice. At all time points examined, there was no difference in the expression of the pEGFR between WT and MT1-MMP KO mice (Fig. 6). Although we found a significant amount of pEGFR in the epithelium at baseline, staining intensity increased at day 2 but decreased below baseline by day 5. This finding matches previous localization of EGFR and EGFR ligands after NA (45) but suggests MT1-MMP is not required for EGFR-mediated signaling in this model.

Since inducible EGF family signaling may be more dependent on human epidermal growth factor receptor 2 (HER2) heterodimers (47), we evaluated HER2 localization with a phospho-specific antibody. Similar to pEGFR, there was some baseline HER2 phosphorylation with increases at day 2, but there was no difference between MT1-MMP KO and WT terminal airways at any time point examined (data not shown). This suggests that the proliferation of terminal airway epithelium after NA-induced injury is not EGFR-dependent as the MT1-MMP KO mice have abundant pEGFR and pHER2 but minimal proliferation. This led us to conclude some additional or alternative process other than EGFR-mediated signaling must be MT1-MMP-dependent and necessary for epithelial proliferation after NA-induced injury.

KGFR is highly upregulated early after NA-induced injury in WT mice but not in MT1-MMP KO mice.

Since intrapulmonary administration of KGF causes robust Clara cell proliferation, and KGFR is present on airway epithelial cells (6, 13, 28), we examined the expression of KGFR (also called FGFR2 splice variant IIIb) after NA-induced injury. We used an antibody raised to a KGFR peptide sequence conserved across species (49, 50). KGFR was minimally present in vehicle-only mice (Fig. 7A), but as early as 1 day after NA exposure, the expression of KGFR was extensively upregulated throughout the airways in the expected basal epithelial location in the WT mice (Fig. 7B). KGFR expression decreased by day 5 (Fig. 7D), suggesting that KGFR expression preceded the proliferative phase of repair, which peaked at day 5 in our model (Fig. 4). This time course is consistent with data on exogenous KGF treatment where proliferation is maximal 2 days after treatment (13).

In sharp contrast, the MT1-MMP KO mice demonstrated only minimal increases in KGFR expression in the peripheral airways at days 1 and 2 (Fig. 7, G and H). The KGFR expression in the MT1-MMP KO mice at day 5 (Fig. 7I) was similar to the WT mice, but no further KGFR expression was identified at later time points in either strain. The expression of the KGFR coincides both in time and place with the MT1-MMP mRNA expression pattern in WT mice (Fig. 1), consistent with KGFR expression and MT1-MMP production being related in airway epithelial cells.

We attempted to quantify these differences using the area of staining in the terminal airways as defined previously (Fig. 8). Although this approach may not reflect quantity of KGFR protein, as sections were developed to equal times and may have resulted in over-saturation of some samples, it provides some objective measure. The KGFR expression measured by this method was elevated in the KO mice at day 2 but was significantly less than the area seen in the WT airways at the same time point or day 1.

DISCUSSION

MMPs are produced by airway epithelium in lung diseases, and several play a role in epithelial regeneration (11, 25). We have investigated MT1-MMP using the NA model of acute small airway epithelial injury and repair. Our data demonstrate that MT1-MMP is produced by the distal airway epithelium after injury. Furthermore, the data suggest that MT1-MMP is responsible for the small airway epithelial cell proliferation and not airway epithelial cell migration as initially expected. Importantly, the severity of injury was not modulated by MT1-MMP production. The lack of proliferation after NA exposure in MT1-MMP KO mice despite normal EGFR-mediated signaling suggests that KGFR- and not EGFR-mediated signals are necessary for the proliferative response seen in this model. The phospho-specific antibodies we used could cross-react with other EGFR family members, however, the consistency of temporal regulation of pEGFR with previous publications of EGFR expression suggests our results are valid (45). Consistent with this, asthmatic airways where EGFR phosphorylation is elevated do not demonstrate increased proliferation (10, 12). Unlike pEGFR, the expression of KGFR was similar in localization and temporally associated with MT1-MMP expression in this model but greatly decreased in MT1-MMP KO mice. Since KGF can specifically induce proliferation in the small airway epithelium (13), we suspect a KGFR-mediated proliferative response is dependent on local MT1-MMP production in the airway epithelium after acute injury but cannot rule out an indirect interaction. Specificity of this KGFR antibody for the antigen has been demonstrated in human specimens, and the temporal-regulated expression on the basal surface of airway epithelial cells is consistent with what would be expected for KGFR, but cross-reactivity with other FGF receptors in the mouse cannot be completely ruled out.

NA exposure as a model of acute airway injury.

NA is an aromatic hydrocarbon that is converted into a toxic metabolite only in airway Clara cells. Although NA-induced injury is dose dependent, injury also varies with airway level with the most severe injury in the terminal airways where Clara cells are most abundant (46). We concentrated on the terminal airway (defined as the airway epithelium within 100 μm of the bronchoalveolar junction) because of the differential injury and because it is a consistently identifiable structure on histological sections.

We used the model of NA-induced Clara cell injury to study the role of MT1-MMP in small airway epithelial repair because the epithelial repair occurs in a stepwise fashion that allows dissection of the repair process. The dose of NA and strain of mouse used determine the severity of injury, but most sensitive strains of mice do not demonstrate basement membrane exposure and leakage of serum proteins into the lung (24).

NA is found in diesel exhaust and cigarette smoke, but we acknowledge it is not a significant independent cause of human disease. Rather, it is a useful model of noninflammatory airway epithelial-specific injury and repair. However, it is worthy to note that in human diseases where the airway epithelium is injured (e.g., viral bronchiolitis or toxic inhalations), there is typically an intense inflammatory reaction, and additional factors produced by inflammatory cells modulate both the injury and repair processes. Inflammatory cells are a significant source of MT1-MMP, and epithelial cell production of MT1-MMP may be less important when they are present. However, the active MT1-MMP is membrane associated, and local cell surface activity may be critical if nondiffusible substrates are important (e.g., same cell modification of cell surface proteins).

MT1-MMP in airway epithelial repair.

It is notable that several MMPs have been demonstrated to be required for airway epithelial cell migration (11, 25), but MT1-MMP, a known promigratory protease (21), functions in a role other than cell migration in this model. Much of the data on MMPs in airway epithelial cell migration used mechanical wounding of large airway epithelial cells and not toxic injury to small airway epithelial cells like our model. Recent work using MMP-7- or MMP-9-specific inhibitors in a tracheal xenograft model of injury repair (which is more similar to the NA model) demonstrated no defect in migration or proliferation when either MMP-7 or -9 were inhibited but rather a metaplastic undifferentiated epithelium (8). This is important because many in vitro models of airway epithelial repair operate on the premise that the airway is injured similar to the skin with mechanical wounding when, in reality, diffuse toxic or infectious insults are more common. In a toxic injury model, the migration seen is predominately cell spreading, which may not require proteases. Our data do not rule out the existence of a defect in airway epithelial cell migration in MT1-MMP KO mice since the model depends on passive spreading of uninjured cells, and the proliferation defect is before any epithelial sheet migration during the reconstitution phase.

MT1-MMP-deficient mice in a pure C57BL/6 background have no defect in skin wound healing (30) in contrast to our finding of decreased airway epithelial cell proliferation after NA. The differences in results could be related to the strain of mice used and/or targeting construct; however, it has been suggested that in fibrin-independent skin wound healing, MT1-MMP is not produced by basal keratinocytes (29). The toxic injury to NA should be similar to a model of fibrin-independent wound healing, and in our model, we see clear production of MT1-MMP in airway epithelial cells. Basal cells of the large airway may be more similar to basal keratinocytes, and repair of large airway toxic injury could be MT1-MMP-independent but would not be tested in the Clara cell-dependent NA injury model.

Unlike other MMP-deficient mice, MT1-MMP KO mice appear abnormal within the first 10 days after birth. We (1) have previously demonstrated that these mice have normal airway development but fail to develop the adult component of alveoli. However, the mice do not appear to die of a respiratory cause and in a mixed strain can live to 3 mo of age. Although we cannot completely rule out the debilitated state of the MT1-MMP KO mice resulting in inadequate injury repair, the injury was similar in WT and KO mice, and enhanced injury resulting in a failure to proliferate without an exaggerated inflammatory response seems unlikely. Ultimately, an airway-specific conditional MT1-MMP KO mouse will be needed to definitively prove airway epithelial cell and not inflammatory cell or fibroblast production of MT1-MMP is an absolute requirement for epithelial repair after NA-induced injury.

KGFR in airway epithelial repair.

Our discovery of an acute upregulation of KGFR expression after NA-induced injury that precedes the proliferative response suggests the presence of KGFR ligand-mediated signaling. The abundant proliferation seen in WT mice but not in the proliferation-deficient MT1-MMP KO mice that fail to upregulate KGFR confirms a relationship between MT1-MMP, KGFR, and subsequent proliferation. Increased KGFR ligands (FGF-2, -7, -10, or -22) have not been described in this or other airway injury models, but KGF can increase airway epithelial wound healing and does induce proliferation in airway epithelial cells (13, 48). The peak in proliferation days after maximal KGFR staining with subsequent decrease in KGFR expression at day 5 is consistent with KGFR-to-KGF ligand interaction on days 1 and 2 post-NA followed by downregulation of KGFR as has been described in KGF instillation models (13, 38).

Our laboratory (6) has examined KGFR expression of Clara cells in the past and demonstrated that KGFR is present at the mRNA level in laser-captured Clara cells in mice but actually is downregulated 7 days after bleomycin. These findings are consistent with our current data as there is little KGFR expression after initial repair (day 5) in either WT or MT1-MMP KO mice in the NA model. We suspect that KGFR ligation stimulates receptor ligand internalization, inducing a negative feedback on KGFR synthesis. The regulation of KGFR expression in this model or other models of airway epithelial injury is unknown, but airway fibroblasts produce KGF in response to proinflammatory cytokines (7). In future studies, we plan to determine whether KGFR synthesis or protein half-life is altered after NA-induced injury and whether either or both are MT1-MMP dependent.

NA-induced injury will result in oxidative damage, and the KGF pathway does alter the severity of oxidative injury in alveolar type II cells via an antiapoptotic effect (39). Many of the signaling pathways downstream of KGFR activation are similar with other receptor tyrosine kinases activated in this model so that it is difficult to separate downstream antiapoptotic effects from proliferative effects (26, 36). However, the clear decrease in epithelial cell proliferation and the lack of signs of ongoing injury in the KO mice suggest that the defect in MT1-MMP KO mice is in repair and not injury prevention.

Although KGF has been shown to accelerate airway epithelial injury repair in vitro, to our knowledge, this is the first demonstration of the KGF pathway being involved in airway epithelial cell repair in vivo. It is possible that the ligand in our model is another member of the FGF family, as the KGFR can be activated by FGF-2, -10, and -22 as well. Early analysis indicates KGFR is upregulated in other airway injury models (data not shown), suggesting that, at least in small airway epithelium, upregulation of KGFR occurs after injury.

MT1-MMP as a modulator of KGFR-mediated proliferative responses.

The interaction between MT1-MMP and KGFR on the basal surface of injured airway epithelial cells has not been determined but likely is indirect. The fact that MT1-MMP expression continues at a time point that KGFR expression has decreased to baseline is more consistent with MT1-MMP interacting with a KGFR ligand. FGF receptors require heparan-sulfate proteoglycans (HSPGs) for proper dimerization, and FGF ligands can be sequestered by cell surface- or matrix-associated HSPGs. MT1-MMP substrates include the prominent cell surface HSPG, syndecan-1, and the matrix-associated HSPG perlecan as well as many other HSPG-containing molecules. Cleavage of HSPG-containing proteins could liberate KGFR ligands and/or favor KGFR dimerization. Although syndecan-1 distribution is altered during NA-induced injury, we did not see any difference in syndecan-1 localization between MT1-MMP KO mice and littermates (data not shown).

We believe our data reveal a novel role of MT1-MMP in airway epithelial repair via a KGFR-dependent role in Clara cell proliferation. We expect this finding will have important implications for other models where airway epithelial repair to acute toxic or infectious injury occurs and may raise concern regarding the use of broad spectrum MMP inhibitors in lung or other diseases.

NOTE ADDED IN PROOF

The recent demonstration of a role for KGFR on circulating epithelial stem cells in large airway epithelial repair could also be relevant to our model of small airway epithelial repair in this manuscript (Gomperts BN, Belperio JA, Fishbein MC, Keane MP, Burdick MD, Strieter RM. Keratinocyte growth factor improves repair in the injured tracheal epithelium. Am J Respir Cell Mol Biol 37: 48–56, 2007).

GRANTS

This study was supported by National Heart, Lung, and Blood Institute Grants RO1-47328 (J. J. Atkinson and R. M. Senior) and PO1–56419 (R. M. Senior) and the Alan A. and Edith L. Wolff Charitable Trust/Barnes-Jewish Hospital Foundation (R. M. Senior).

FOOTNOTES

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