Possible Impedance of Luminal Reepithelialization by Tracheal Cartilage Metalloproteinases

Background  Rapid reepithelialization of respiratory epithelium after injury to the large conducting airway (eg, trachea and bronchus) is poor. Our laboratory has developed an in vitro model of the trachea that allows us to examine reepithelialization in a complex culture system. We previously described how the presence of cartilage inhibited respiratory epithelial cell (REC) migration/proliferation. In the present study, we examined the effect of cartilage-conditioned medium (CCM) on REC proliferation. We hypothesized that a potential cause of delayed reepithelialization of the large conducting airway after injury could be excessive or aberrant secretion of matrix metalloproteinases (MMPs) by cartilage.

Design  We assessed cartilage-derived MMP production and effects on REC proliferation by adding CCM to primary cultures of porcine RECs on type I collagen and determining the cell number and viability. Cartilage-conditioned medium–derived MMP activity was determined by means of gelatin zymography in pooled samples from different times during in vitro cartilage culture.

Results  We detected MMP-2 and a small amount of MMP-9 in CCM. Enzyme activity was abolished by EDTA, confirming MMP identity. Cartilage-conditioned medium inhibited REC attachment and proliferation. Addition of the MMP inhibitor GM6001 to cartilage cultures yielded CCM that did not inhibit REC growth, indicating a role for cartilage-derived MMPs in modulating REC proliferation.

Conclusion  Cartilage production and activity of MMP after injury to the large conducting airway may be a factor in the failure of luminal reepithelialization, resulting in aberrant repair.

THE LARGE conducting airway offers unique biological and clinical challenges in reestablishing normal anatomy and physiology after injury.¹ This may be due to a distinctive interplay among cartilage, submucosa (connective tissue), and respiratory epithelium. Determining the interactions among these components is central to understanding injury and repair mechanisms in the proximal airway. We have developed an in vitro model of the trachea incorporating cartilage, extracellular matrix (ECM), and respiratory epithelium, which facilitates study of the upper airway with its various tissue components in an easily manipulated and accessible system.² The purpose of this particular study was to test the hypothesis that exposed or injured cartilage secretes matrix metalloproteinases (MMPs) that may interfere with respiratory epithelial cell (REC) attachment and proliferation.

Cellular migration over an appropriate ECM after injury is part of a generalized repair process and requires regulated proteolytic degradation of the ECM. This leads to activation and/or release of matrix-bound growth factors that augment tissue repair. The MMPs constitute a family of zinc-containing neutral proteases that together are capable of degrading all components of the ECM. They are secreted in a latent form as propeptides requiring activation and are inhibited by endogenous tissue inhibitors of metalloproteinases (TIMPs).^3-6 Synthesis of MMP is transcriptionally enhanced by several growth factors and inflammatory cytokines.⁷ Although MMPs are necessary for cell migration during morphogenesis and tissue repair, inappropriate or excessive MMP production has been implicated in tissue destruction in several diseases.⁶ The proenzymes MMP-2 and MMP-9 are produced by cartilage and play a major role in ECM remodeling and loss in osteoarthritis.⁸ Previous work with our in vitro tracheal model (composite coculture, US patent 6 312 952) showed that RECs failed to spread across type I collagen overlying a cartilage layer.² This in vitro situation is analogous to clinical observations of tracheal injury where lack of reepithelialization can lead to aberrant repair (eg, stricture, dehiscence).⁹ Complex signaling interactions between the cartilage and epithelium may be a factor in the failure of RECs to migrate and proliferate appropriately. In this report, we have explored the role of MMPs in REC proliferation, as their effects on reepithelialization in complex tissue systems such as the trachea are unclear.

Chondrocytes were isolated from bovine articular cartilage and cultured on collagen inserts as described elsewhere.² Bovine and porcine cartilage-conditioned media (CCM) gave identical results, but since bovine material was more abundant, it was used for the experiments. Media harvested for gelatin zymography were cleared by means of centrifugation to remove cell debris, pooled at each time, and frozen at −20°C until assay. In some experiments, we diluted 50µM of the broad-spectrum hydroxamic acid MMP inhibitor GM6001 (N-[(2R)-2-(hydroxamidocarbonyl methyl)-4-methylpentanoyl]-L-tryptophan methylamide)^10,11 or its negative control drug (N-t-butoxycarbonyl-L-leucyl-L-tryptophan methylamide) (Calbiochem, San Diego, Calif) in dimethyl sulfoxide and added the drugs daily from days 0 through 14 of culture after switch to a serum-free medium (SFM).

Respiratory epithelial cells were isolated from porcine tracheal tissue using elastase digestion as described by Hicks et al¹² and plated on top of type I collagen–coated dishes at 0.2 × 10⁶ cells/cm² in SFM. Media were collected as described for gelatin zymography.

We assessed the effects of CCM on REC proliferation by adding pooled serum-free CCM (days 3-14 after switching from serum-containing media) diluted to a ratio of 1:1 with fresh SFM, every 2 to 3 days, at media changes. At selected intervals, we harvested cultures for cell counting using a hemocytometer. Cell viability was assessed by means of trypan blue exclusion, and cells stained blue were considered damaged or nonviable. In selected experiments, pooled CCM was serially diluted (at ratios of 1:1, 1:10, 1:100, and 1:1000) and used to treat RECs before determining the cell number after 10 days of growth. Preliminary experiments indicated that 50µM GM6001 abolished MMP activity in CCM without affecting REC growth or morphology. We also added GM6001 directly to cultured RECs daily for an additional control group. The final volume of dimethyl sulfoxide added to cultures was negligible (0.25% vol/vol) and did not affect cell growth or viability.

Latent proteases were activated by pretreatment with 4-aminophenyl-mercuric acetate (APMA)¹³ or by storage of samples at −20°C for at least 7 days (autoactivation).¹⁴ Storage of samples resulted in activation equivalent to APMA treatment. Collected medium was mixed with 2× sample buffer (20mM dithiothreitol, 4% sodium dodecyl sulfate, 50mM Tris, 20% glycine, and 0.002% bromphenol blue) and underwent electrophoresis without prior heating as described,¹⁵ under nonreducing conditions through 7.5% acrylamide gels containing 1-mg/mL gelatin as substrate. Zymogram gels were then washed twice in 2.5% Triton X-100, and incubated in developing buffer (50mM Tris hydrochloride, 0.2M sodium chloride, 10mM calcium chloride, 0.02% [wt/vol] polyoxyethylene 23 lauryl ether (Brij 35; Sigma-Aldrich Corp, St Louis, Mo [pH, 7.5]) overnight at 37°C. Duplicate gels were incubated in developing buffer with 20mM EDTA to chelate the zinc ions and inactivate the MMPs. Gels were rinsed, stained with Coomassie brilliant blue R-250, and destained until clear areas of enzyme activity were detected against a blue background.¹⁴ Purified MMP-2 and MMP-9 standards serving as positive controls (Chemicon International, Inc, Temecula, Calif) and kaleidoscope-prestained molecular weight standards (Bio-Rad Laboratories, Hercules, Calif) were included in every gel. Gels were photographed using Kodak 665 negative/positive film (Eastman Kodak Co, Rochester, NY).

Experiments were performed at least twice on independent REC preparations. Because of small variations in cell growth, the data were expressed as the percentage of the control cell number (set to 100%) for each experiment. Observed growth trends were identical in all experiments. Data were analyzed by means of analysis of variance or t test, as appropriate, with a level of statistical significance at P<.05, using Instat (Graphpad, San Diego, Calif) statistical software.

We collected CCM on days 3 to 14 after switching from serum-containing media to SFM and performed gelatin zymography. Several major bands of gelatinolytic activity were detected (Figure 1A). Bands were observed at approximately 105, 92, 88, 72, and 68 kd. Based on comparisons with purified MMP standards run in parallel, the areas of enzyme activity at 92 and 88 kd corresponded to pro–MMP-9 and active MMP-9, respectively. The band of activity at 72 kd corresponded to pro–MMP-2, and the 65-kd band corresponded to activated MMP-2. The higher molecular weight form may represent an additional gelatinase.¹⁵ When duplicate zymograms were developed in the presence of the chelator EDTA, enzyme activity was abolished and the clear areas of lysis then disappeared, confirming enzyme activity as belonging to MMPs (Figure 1B).

Serum-free CCM exhibited considerable MMP activity (Figure 1A, all lanes, and Figure 2, lane 4) that diminished over time in culture, and by day 14 was barely detectable (Figure 1A). Serum-containing CCM also contained substantial MMP activity (not shown). The SFM from day 3 cultured RECs contained a negligible amount of MMP activity (Figure 2A, lane 1), whereas that from day 7 cultures contained small amounts of MMP-9 and MMP-2 (Figure 2B, lane 1). All nonconditioned media lacked any MMP activity (not shown).

The number of attached RECs obtained after 7 days in culture was reduced by 67% (average of 6 experiments) in cultures treated with CCM compared with SFM controls (Figure 3), suggesting that cartilage-derived MMPs interfered with epithelial cell attachment/proliferation. Cell viability and general morphology were not affected by the addition of CCM (not shown).

Figure 4 (representative experiment) illustrates that CCM from untreated and negative control drug–treated cartilage reduced the number of RECs after 7 days in culture, but that media derived from cartilage treated with 50µM of the MMP inhibitor GM6001 did not diminish REC growth. The yield of REC cultures that were treated with 50µM of GM6001 was similar to that of the control cultures (data not shown). The CCM serially diluted up to 1:1000 in SFM was fed to RECs, but only the 1:1 ratio of CCM-SFM significantly inhibited proliferation (mean ± SEM, 39.7% ± 8.8% of controls; P<.001). The CCM diluted to ratios of 1:10 to 1:1000 did not significantly inhibit proliferation (P>.05; data not shown). Zymography showed that GM6001 markedly reduced pro–MMP-2 activity and virtually eliminated MMP-2 and enzyme activities at higher molecular weights by day 7 of REC culture (Figure 2, lane 3), in keeping with its broad spectrum of inhibitory action on MMPs. The negative control drug diminished pro–MMP-2 activity by a small amount and markedly reduced other bands of activity by day 7 (Figure 2B, lane 2), a probable nonspecific effect of the drug.

The results of our study demonstrate that MMPs are present in CCM and inhibit the growth of primary cultured RECs on type I collagen. Addition of the MMP inhibitor GM6001 during cartilage culture yielded conditioned medium that was deficient in MMPs, as detected by means of gelatin zymography, and that no longer impaired REC attachment and proliferation. Furthermore, dilution of CCM beyond a 1:1 ratio resulted in the restoration of cell numbers to control levels by day 10. These results indicate a direct role for MMPs in modulation of REC attachment and proliferation. The negative control drug did not ameliorate the inhibitory effects of MMPs on REC attachment and growth, as expected for the negative control. We still detected MMP-2 by means of zymography in the negative control drug–treated group, and thus it may be a major inhibitory component of cartilage-derived MMPs. These data underscore the negative influence of MMPs, especially MMP-2, on REC growth.

Failure of confluent REC growth has been observed in our in vitro model of the trachea that incorporates cartilage, type I collagen, and tracheal epithelial cells.² Matrix metalloproteinases-2 and -9 are produced by cartilage⁸ and various epithelial cell types, including those of the airway,^16-18 suggesting that MMPs normally present in the large conducting airway may modulate the ability of RECs to sense and respond to an appropriate and functional basement membrane or ECM. Our data suggest that exogenous MMPs from cartilage inhibit REC attachment and ability to grow. One possible mechanism is that cartilage-derived MMPs degrade ECM receptors and signals, thereby disrupting cell-matrix interactions and impairing the ability of the cells to migrate and proliferate when stimulated by inflammatory cytokines released during the injury repair process.^18,19 Under normal conditions, TIMPs present in the immediate cellular environment are sufficient to bind to and inactivate MMPs. However, when activated MMPs are present in excess of TIMPs, the balance is shifted toward degradation. The notion of a second injury resulting from cytokine-stimulated MMP production and activity has been described for corneal injuries²⁰ and may also be relevant for the trachea. Thus, in the context of tracheal injury and repair, the contributions of cartilage-derived MMPs should also be considered in the overall process. Injury to the upper airway that violates the basement membrane and damages the underlying ECM and cartilage, such as that seen in trauma or planned surgical procedures, results in a cascade of events aimed at the timely repair of the wound. However, transcription and activation of pro-MMPs from the damaged cartilage and epithelium may be stimulated, resulting in excessive degradation of the ECM and disruption of normal repair. Although likely not the sole causative agent of impaired reepithelialization in tracheal injury, MMPs in all probability play a role in delayed or inhibited REC proliferation. The potential contribution and interactions of the tissues underlying the tracheal epithelium must be considered in strategies for facilitating upper airway wound healing.

Mild injury leads to local repair with a balance of MMPs/TIMPs and normal reepithelialization. Severe injury results in secretional cartilage and epithelial MMPs in excess of TIMPs and leads to secondary injury and impaired reepithelialization. A possible role for MMPs in influencing upper airway reepithelialization after injury is illustrated in Figure 5.

Corresponding author and reprints: Wesley L. Hicks, Jr, MD, Department of Head and Neck Surgery, Roswell Park Cancer Institute, Elm and Carlton streets, Buffalo, NY 14263 (e-mail: Diane.Biggie@roswellpark.org).

Accepted for publication July 18, 2002.