Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The extracellular matrix (ECM) is a complex and dynamic meshwork of proteins and proteoglycans that not only provides structural support to tissues but also has a profound influence on many biologic activities. These include basic processes such as cell proliferation and differentiation, as well as cell adhesion, migration, and tissue morphogenesis. Cell contact with ECM molecules plays an important role in organogenesis, differentiation, and wound healing (1, 2). Although in vivo studies of cell-ECM interactions during development and wound healing have provided some clues about the effects of ECM on tissue morphogenesis and cell function, in vitro cell-culture systems offer a more controlled environment for examining the effects of specific ECM molecules on cell behavior and function. There are at least three mechanisms by which the ECM may regulate cell behavior. One involves the composition of the ECM. The second is through synergistic interactions between growth factors and matrix molecules. The third is through the cell-surface receptors that mediate adhesion to ECM components.

Numerous cell- and tissue-culture studies conducted over the past 10 yr have shown that the chemical properties of ECM can profoundly influence cell shape and functions via mechanisms involving integrins and other cell-membrane receptors (1). Cells isolated from different organs tend to rapidly lose many of their morphologic characteristics and biochemical functions when grown on plastic surfaces. It has been demonstrated that ECM plays a fundamental role in both the control of gene expression and the induction and maintenance of tissue-specific function. More specifically, it has been shown that ECM controls gene expression in the mammary gland, in hepatocytes, and in keratinocytes (2). Other studies have shown that ECM can not only activate and enhance differentiation, but can also alter it, causing epithelium to transform into mesenchyme (3).

We have recently shown that primary and confluent human bronchial epithelial cells (HBECs) cultured in plastic dishes coated with types I + III collagen expressed both major and constitutive 92-kD gelatinase (4), as well as its specific tissue inhibitor of metalloproteinase, (TIMP-1) (5). Also, these HBECs upregulated their 92-kD gelatinase expression in response to Escherichia coli lipopolysaccharide (LPS) or to the inflammatory cytokines interleukin-1 (IL-1) and tumor necrosis factor- (TNF-). These data strongly suggest that HBECs may be actively involved in the physiologic and physiopathologic remodeling of the airway basement membrane. HBECs overlie the subepithelial basal lamina, of which type IV collagen is one of the main macromolecular components. Also, the two matrix gelatinases have specific affinity for the subepithelial lamina, and more particularly for type IV collagen. We therefore designed a study to examine the effects of type IV collagen as compared with types I + III collagen on HBEC behavior and function. Type IV collagen is representative of the normal matrix substratum under physiologic conditions, whereas types I + III collagen are representative of the substratum occasionally found under physiopathologic conditions, such as basement-membrane denudation. Morphology, proliferation, and matrix gelatinase and TIMP-1 basal expression were investigated in cell cultures plated on the two matrix substrates. The effects of specific collagen types were also investigated in response to LPS and to the proinflammatory cytokines IL-1 and TNF-. When cells were cultured on type IV collagen, we found moderate but significant and consistent restrictive modulation of 92-kD gelatinase, as well as loss of activation of both 92-kD and 72-kD gelatinase, supporting a role for matrix-cell interactions in the regulation of gelatinase expression by HBECs.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Cell Cultures

Human bronchial epithelial biopsy specimens were obtained by fiberoptic bronchoscopy from 23 patients who were being investigated for bronchopulmonary carcinoma. Biopsies were taken at a distance from the tumor. Histologic features of the bronchial mucosa were normal in all specimens. All procedures were reviewed and approved by our Henri Mondor Hospital Institutional Review Board, and written informed consent was obtained from all study subjects.

HBECs were cultured as previously described (4). Briefly, two or three explants (approximately 0.5 × 0.5 mm in size) were placed on sterile plastic dishes coated with collagen G matrix (type I and III collagen) (PolyLabo, Strasbourg, France) or human placental type IV collagen (Sigma Chemical Co., St. Louis, MO). The explants were covered with 600 µl of Dulbecco's modified Eagle's medium (DMEM): F12 (1:1) medium, and were incubated for 24 h. Two milliliters of culture medium were then added to each dish. Culture medium consisted of serum-free DMEM:F12 (1:1) supplemented with 2% Ultroser G, 1% antibiotics (penicillin G sodium: 10,000 units/ml; streptomycin sulfate: 10,000 µg/ml; and amphotericin B, 25 µg/ml); and 2 mM glutamine (Life Technologies, Gaithersburg, MD). Explants were placed in a humidified incubator at 37°C under 5% CO₂ in air. The culture medium was changed every 3 to 4 d. Explants were cultured for 2 wk, until HBECs were confluent. Also, the same explants were transferred to new, sterile, coated plastic dishes at 5- to 8-d intervals to initiate new primary HBEC cultures.

Normal human mammary fibroblasts cultured to confluency on plastic dishes were used as reference cells for the human TIMP-1 reverse transcription-polymerase chain reaction (RT-PCR) assay. Human alveolar macrophages (AM) were recovered from bronchoalveolar lavage (BAL) specimens from four patients with adult respiratory distress syndrome (ARDS), and were used as reference cells for the RT-PCR assay for human 92-kD gelatinase.

For measurement of gelatinase and TIMP-1 activities, as well as isolation of total cellular RNA (RNA_T), HBECs were incubated at confluence with Ultroser G-free culture medium in the presence of 0.2% lactalbumin for 24 h. These cultures were or were not subsequently exposed to LPS (1 µg/ml), IL-1 (100 U/ml), or TNF- (100 U/ml) (Sigma), respectively, for 24 additional hours.

Cell Culture on ECM

In order to assess the effect of ECM on 92-kD gelatinase and TIMP-1 production, HBECs were cultured in plastic dishes coated with either native type IV collagen or collagen G (types I + III collagen). ECM proteins (600 µg, 3 mg/ml) were added to 35-mm culture dishes. After 15 min at ambient temperature, the protein solution was aspirated from the dish, the dish was rinsed with Hanks' balanced salt solution (HBSS), and the explants were placed in the dish as described in detail earlier.

Zymography and Reverse Zymography

The HBEC culture medium was harvested and stored at 80°C until use. Collected medium was resolved with 8% sodium dodecyl sulfate-polyacrylamide-gel electrophoresis (SDS-PAGE) in the presence of 1 mg/ml porcine skin gelatin, to evaluate gelatinase activities. Twenty-fold-concentrated medium was resolved with 11% SDS-PAGE in the presence of 1 mg/ml -casein to evaluate stromelysin-type activities. The Laemmli method was followed. After electrophoresis, the gel was washed for 30 min in 2.5% Triton X-100 at room temperature to remove SDS. The gel was then incubated overnight at 37°C in reaction buffer (100 mM Tris-HCl, 10 mM CaCl₂, pH 7.4). After staining with Coomassie brilliant blue R 250, gelatin-degrading enzymes were identified as clear zones of lysis against a blue background. Molecular weights of gelatinolytic bands were estimated through the use of prestained molecular-weight markers.

Activities in the gel slabs were quantified through semiautomated image analysis (NIH image 1.52, Bethesda, MD), which quantifies both the surface area and the intensity of lysis bands on scans of the gels. Results were expressed as arbitrary units (AU): AU/24 h/10⁴ cells. To check that this method of measuring enzymatic activity on zymograms was linear over the range of activities in unknown samples, we evaluated activities for increasing volumes of culture medium, and found that arbitrary unit values obtained with the image-analysis system increased linearly with the sample volume (r = 1.00) (6).

TIMP-1 secreted into the culture medium was detected through reverse zymography (7). Briefly, 25 µl of 20-fold-concentrated conditioned media were resolved with 11.5% SDS-PAGE in the presence of 1 mg/ml porcine skin gelatin. After removal of SDS from the gel, the standard zymographic method was modified by incubation of the gel for 45 min at 37°C in conditioned medium from 4-aminophenyl mercuric acetate (APMA)-activated rabbit-skin fibroblasts, which provided a source of activated metalloproteinases capable of degrading the gelatin in the gel. The gel was then incubated and stained in the same way as for the standard zymogram. Protection of the gelatin in the gel through the presence of TIMPs led to the production of relatively dark bands against a lighter background.

Immunoblotting

The supernatant of confluent HBECs was discarded and the cells were washed twice in phosphate-buffered saline (PBS). Cells were then lysed with 1% SDS for 10 min. Twenty microliters of cell lysates (4 × 10⁴ cells) were separated with 8% SDS-PAGE in the presence of 10% -mercaptoethanol. The terminally truncated form of human MT1-matrix metalloproteinase (MT1-MMP), lacking the transmembrane domain, was expressed in Escherichia coli as a fusion protein to glutathione-S-transferase (M. P. d'Ortho and G. Murphy, unpublished results). The 70-kD recombinant protein was used to generate polyclonal sheep antibody, and was used as the positive control. The blots were transferred to a Biohylon-Z⁺ membrane (Bioprobe, Montreuil, France). Nonspecific staining was blocked by incubating the transfers for 1 h at 37°C in Tris-buffered saline (TBS) containing 5% nonfat dried milk and 5% fetal calf serum (FCS). The transfers were then incubated with polyclonal sheep antibody against human MT1-MMP (M. P. d'Ortho and G. Murphy, unpublished results) diluted 5 µg/ml in TBS. The blots were washed three times in TBS-Tween (0.05% Tween 20) and incubated for 2 h with peroxidase-conjugated rabbit antisheep IgG (DAKO, Trappes, France) diluted 1:3,000 as the second antibody. The blots were visualized using an enhanced chemiluminescence (ECL) kit (Amersham, Bucks, UK).

RNA Extraction

RNA_T was extracted from HBECs, human macrophages, and fibroblasts, using Trizol reagent (Life Technologies) according to an improvement to the single-step RNA isolation method developed by Chomczynski and Sacchi (8). RNA_T was quantified at 260 nm/280 nm, and the integrity of the samples was checked with 1.5% agarose-gel electrophoresis. Reproducible amounts of 8 to 15 µg RNA_T were obtained from 10⁶ cells, and aliquots were stored in sterile microfuge tubes at 80°C until use.

Quantitative RT-PCR Assay of 92-kD Gelatinase and TIMP-1 mRNA

Primer design and synthesis. Sense and antisense primers for human 92-kD gelatinase and TIMP-1 were designed as described in our recent works (4, 5), as follows:

RT step. RT assays of 92-kD gelatinase and TIMP-1 were done with RNA_T as previously described (4, 5). Briefly, to minimize sample handling and contamination, RT and PCR steps were performed sequentially in the same reaction tube. To a final volume of 25 µl, the following compounds were added: 3 µl of 10× PCR buffer (200 mM Tris-HCl, pH 8.3; 500 mM KCl; 15 mM MgCl₂; 1 mg/ ml gelatin), 10 µl dilution buffer for RNA (1 M Tris, pH 8.3, 10 µl; 0.1 M dithiothreitol [DTT] 20 µl; RNase 1 µl; BSA 100 µl; H₂O 870 µl), 10 µl RNA_T obtained from HBEC (10 ng) cultures on both matrices, with or without exposure to LPS, IL-1, or TNF-, and 2 µl corresponding downstream primer (10 pmol). After heating for 2 min at 80°C in the thermocycler to break up secondary structures, the tubes were equilibrated at 42°C. Each sample was supplemented with 25 µl of RT mix containing 2.5 µl of 10× PCR buffer; each deoxynucleotide triphosphate (dNTP) at 1.25 mM, 16 µl; 100 mM MgCl₂, 1.5 µl; and 100 mM DT T, 4 µl, with or without 200 U of Moloney murine leukemia virus (MMLV) RT (Life Technologies). The final volume was 50 µl. The RT reaction lasted 45 min and was done at 42°C to prevent excessive mispriming and possible RNA refolding. After completion of RT, the temperature was raised to 96°C for 30 s to inactivate the enzyme and denature the RNA-DNA hybrid. The temperature was then equilibrated at 80°C.

Quantitative PCR. Quantitative RT-PCR assays were performed as previously described (4, 5), and required the availability of two specific internal DNA standards corresponding to the 92-kD gelatinase and TIMP-1 RNA_T targets. These internal DNA standards were obtained by amplification of foreign DNA fragments from the ampicillin-resistance gene in the Bluescript IISK plasmid, using two composite primers. Each composite primer was composed of the corresponding target-gene primer sequence, attached to a short segment of nucleotides that hybridized to the opposite strands of the foreign DNA fragment. Since each internal standard contained primer target RNA_T (in the primer mixture), the inner end was the internal standard primer, whereas the outer end was the target RNA_T primer. Two fragments, of 586 bp and 585 bp, were obtained as the internal DNA standards for 92-kD gelatinase and TIMP-1, respectively.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Cellular Morphology and Growth Pattern

Phase-contrast microscopic observation showed that HBECs migrated outward from the explants in an approximately monolayer fashion, regardless of the type of collagen coating on the plastic dishes. With both matrices, the outgrowth appeared within 2 d, continued for up to 2 wk, and could be maintained with functional ciliated cells for more than 2 wk. Thus, the time to confluence was identical for cells growing on types I + III collagen (15.3 ± 0.7 d) or on type IV collagen (16.0 ± 0.5 d) (Table 1). Also, the HBEC number evaluated at confluency with a hemocytometer was similar (10⁶ ± 10³) with the two culture conditions.

With both collagen matrices, HBECs were flat and polygonal in shape, and were closely apposed, as is typical of cultured epithelial cells (Figures 1A and 1B). Also, the epithelial nature of all cultured bronchial cells was confirmed by staining with antibody to cytokeratin, the characteristic component of epithelial-cell intermediate filaments.

View larger version (70K): [in this window] [in a new window]   Figure 1.   Morphologic assay of primary cultures of HBECs (magnification: ×100) in two different matrices. (A) Types I and III collagen. (B) Type IV collagen. Column 1: 4-d cell cultures; column 2: confluent HBEC culture on Day 16.

Zymography

Zymography on SDS-gelatin was done to determine whether use of two different matrix substrates translated into differences in the regulation of gelatinase expression by HBECs at the protein level (Figure 2). As in an earlier study (4), we found that: (1) the activities of gelatinase released from HBECs plated on types I + III collagen and investigated under basal conditions were detected in four main forms: a major band at 92 kD, produced by the pro- form of gelatinase B (MMP-9); a barely visible 88-kD band corresponding to the active form of gelatinase B; and two minor bands at 72 kD and 68 kD, corresponding to the pro- and active forms of gelatinase A (MMP-2), respectively; (2) exposure to LPS, IL-1, or TNF- induced marked increases in production of the 92-kD gelatinase and of its active form (88 kD). In contrast, LPS, IL-1, and TNF- all failed to modify the minor latent or active forms of gelatinase A. 

View larger version (31K): [in this window] [in a new window]   View larger version (32K): [in this window] [in a new window]   View larger version (28K): [in this window] [in a new window]   Figure 2.   Gelatinase production by HBECs cultured on collagen G (types I + III collagen) or type IV collagen under basal conditions and after stimulation by LPS, IL-1, and TNF-. Zymography was used to evaluate gelatinase production. At confluence, cultured HBECs were incubated for 24 h with Ultroser G-free medium in the presence of 0.2% lactalbumin, and were or were not exposed to 1 µg/ml LPS, 100 U/ml IL-1, or 100 U/ml TNF- for 24 additional hours. Then, each culture medium (20-µl aliquot corresponding to 104 cells) of confluent HBECs was subjected to SDS- PAGE/gelatin electrophoresis. HBECs from two subjects, cultured in control medium on collagen G (types I + III collagen) or on type IV collagen, were exposed to LPS (A), IL-1 (B), or TNF- (C). The arrows indicate the positions of 92-kD, 88-kD, 72-kD, and 68-kD gelatinase activities.

Several differences were evidenced in HBECs cultured on type IV collagen versus types I + III collagen (Figure 2): (1) Basal expression of 92-kD gelatinase was slightly but consistently reduced in the cultures from three donors. Semiautomated image analysis adjusted for cell numbers showed that the reduction was 32 ± 3%. Moreover, the 88-kD active form of gelatinase B became undetectable. (2) The increase in the production of the pro- form of 92-kD gelatinase induced by exposure to LPS or TNF- was less marked. The decreases as assessed with semiautomated image analysis were 23 ± 2% and 37 ± 6% for LPS and TNF-, respectively. No reduction in the effect of IL-1 was seen. The 88-kD active form of gelatinase B was not detectable with any of the stimulants used in the study when type IV collagen was the substratum, indicating loss of the gelatinase B activation process. (3) The 68-kD active form of gelatinase A also consistently became undetectable, indicating loss of gelatinase A activation. However, it is of interest that the total amount of 72-kD gelatinase (pro- form + active form) produced remained fairly constant (2.6 × 10⁴ ± 3 UA/24 h/10⁴ cells).

Zymography on SDS--casein was done to determine whether the use of two different matrix substrates translated into differences in the regulation of stromelysin expression by HBECs at the protein level. Two caseinolytic bands of 56 kD and 52 kD from 20-fold-concentrated culture medium, probably related to the pro- form and active form of stromelysin 1 (MMP-3), respectively, were barely detectable, and appeared unchanged whatever the substrate used (data not shown). EDTA completely inhibited the activity of all caseinase bands, in accord with previous findings that stromelysins belong to the MMP family.

Modulation of 92-kD Gelatinase mRNA Levels in HBECs by Matrix Substratum

As previously described (4), 92-kD gelatinase mRNA levels were evaluated with quantitative RT-PCR done on HBEC RNA_T through coamplification with 10⁴ molecules of the corresponding specific internal DNA standard (Figure 3 and Table 2).

View larger version (23K): [in this window] [in a new window]   Figure 3.   Expression of 92-kD gelatinase by HBECs exposed to LPS, IL-1, and TNF- on two matrices. Ten nanograms of RNAT from HBECs cultured with or without 1 µg/ml LPS, 100 U/ml IL-1, or 100 U/ml TNF-, and 10 ng RNAT from AM (MAC), were reverse-transcribed and coamplified with a constant amount (104 molecules) of the specific internal DNA standard. Two serial bands of the expected sizes (491 bp for internal DNA standard and 303 bp for HBECs and MAC) were observed. The bands of the internal standard were the same across the range of amplified samples, indicating a constant amplification rate. HBEC bands were diminished in the presence of type IV collagen as compared with type G (I + III) collagen, indicating restrictive modulation of the transcription level of 92-kD gelatinase.

Under basal conditions, coamplification and scanning analysis of autoradiograms showed a 30% decrease in the 92-kD gelatinase mRNA level of HBECs cultured on type IV collagen as compared with types I + III collagen (1 × 10⁴ versus 1.5 × 10⁴ mRNA molecules/10 ng RNA_T), reflecting the decrease observed at the protein level.

As expected, HBECs cultured on types I + III collagen in the presence of the proinflammatory cytokines IL-1 and TNF- showed a net increase in the level of 92-kD gelatinase mRNA (2.4 and 2.5 mRNA molecules/10 ng RNA_T, respectively, versus 1.5 mRNA molecules/10 ng RNA_T for HBECs in the absence of cytokines). In contrast, when HBECs were cultured on type IV collagen, the increase was reduced by about 20 to 50% as compared with HBECs cultured on types I + III collagen (1.9 × 10⁴ and 0.9 × 10⁴ mRNA molecules/10 ng RNA_T for HBECs cultured in the presence of IL-1 and TNF-, respectively). This result was in accordance with the decrease in the total (pro- + active) form of gelatinase observed at the protein level.

Also as expected, HBECs cultured on types I + III collagen in the presence of LPS showed only a small increase in the level of 92-kD gelatinase mRNA (2 × 10⁴ versus 1.5 × 10⁴ mRNA molecules/10 ng RNA_T), owing to mRNA stabilization by LPS (4). However, when cells were cultured on type IV collagen, the increase in the 92-kD gelatinase mRNA level was unchanged as compared with that observed on types I + III collagen, suggesting that mRNA stabilization by LPS may be modulated to some extent by cell-matrix interactions.

Immunoblotting

Because MT1-MMP is known to be involved in the activation process of 72-kD gelatinase, immunoblotting of MT1-MMP was done to determine whether use of two different matrix substrates translated into differences in its regulation by HBECs. MT1-MMP was clearly detected in HBEC lysates, but its expression appeared similar whatever the type of collagen substrate used (Figure 4).

View larger version (46K): [in this window] [in a new window]   Figure 4.   Immunoblotting of MT1-MMP. Aliquots of HBEC lysates (4 × 104 cells) were analyzed with immunoblotting to define more precisely the MMP property of HBEC transmembrane MT1-MMP. Lane 1: 200 ng of the 70-kD recombinant-protein MT1-MMP used as the positive control. Lanes 2 and 3: 20-µl HBEC lysates stimulated with IL-1 (100 U/ml) and LPS (1 µg/ ml), respectively, in the presence of collagen G (types I + III). Lanes 4 and 5: 20-µl HBEC lysates from the same subjects stimulated with IL-1 (100 U/ml) and LPS (1 µg/ml), respectively, in the presence of type IV collagen. The blots were incubated with polyclonal sheep antibody (generated from the 70-kD recombinant protein MT1-MMP). The specificity of the 70-kD band recognized by the immune serum was confirmed by the negative result obtained without the first antibody (Lane 6).

Reverse Zymography

Reverse zymography of 20-fold-concentrated conditioned media from HBECs was done to determine whether two different matrix substrates differentially regulated TIMP-1 expression at the protein level (Figure 5).

View larger version (43K): [in this window] [in a new window]   Figure 5.   Reverse zymography of concentrated conditioned medium of HBECs. At confluence, cultured HBECs were incubated for 24 h with Ultroser G-free medium in the presence of 0.2% lactalbumin, and were treated with or without 1 µg/ml LPS, 100 U/ml IL-1, and 100 U/ml TNF- for 24 additional hours. After being concentrated 20-fold by lyophilization, 25 µl of media were subjected to 11.5% SDS-PAGE/gelatin electrophoresis. Lane 1: unstimulated cells (control) on collagen G (types I + III collagen); Lanes 2, 3, and 4: cells stimulated with LPS (1 µg/ml), IL-1 (100 U/ml), and TNF- (100 U/ml), respectively, on collagen G; Lanes 5 to 8: same conditions as for Lanes 1 to 4 but on type IV collagen. Arrows indicate migration of molecular-weight standards.

When HBECs were cultured on types I + III collagen, TIMP-1 activity was constitutively expressed under basal conditions, remained unchanged in the presence of LPS or IL-1, and decreased slightly in the presence of TNF-. When HBECs were cultured on type IV collagen, no changes in TIMP-1 were observed from culture on types I + III collagen, whether under basal conditions or in the presence of LPS or of proinflammatory cytokines.

Comparative Investigation of TIMP-1 mRNA Levels According to the Matrix Substratum

TIMP-1 mRNA levels were evaluated with quantitative RT-PCR done on HBEC RNA_T through coamplification with 5 × 10⁴ molecules of the corresponding specific internal standard (Figure 6 and Table 3).

View larger version (18K): [in this window] [in a new window]   Figure 6.   Evaluation of TIMP-1 mRNAs from cultured HBECs with quantitative RT-PCR. Ten nanograms of RNAT from HBECs cultured with or without 1 µg/ml LPS, 100 U/ml IL-1, or 100 U/ml TNF-, and 2.5 ng RNAT from fibroblasts (FB), were reverse-transcribed and coamplified with a constant amount (5 × 104 molecules) of the specific internal DNA standard. Two serial bands of the expected sizes (585 bp for internal DNA standard and 400 bp for HBECs and FB) were observed. Quantitative PCR produced similar HBEC bands for each culture condition and with both collagen matrices, indicating absence of modulation of the transcription level of TIMP-1.

Under basal conditions as well as in response to LPS or to the proinflammatory cytokines IL-1 and TNF-, TIMP-1 mRNA levels showed little or no change (8.8 × 10⁴ ± 0.2 mRNA molecules/10 ng RNA_T) with either types I + II or type IV collagen as the matrix substrate.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

We found that human placental type IV collagen, as compared with types I + III collagen used as the matrix substratum for HBEC cultures, was associated with: (1) less production of major 92-kD gelatinase under basal conditions; (2) less upregulation of 92-kD gelatinase under inflammatory conditions; (3) loss of activation of the pro- forms of 92-kD and 72-kD gelatinases; and (4) maintenance of TIMP-1 production. All these data support the current concept that the ECM regulates gene expression in differentiated cells (2, 9).

Restrictive 92-kD Gelatinase Regulation on Type IV Collagen Substratum

In our study, HBECs cultured on type IV collagen were more representative of the homeostatic cell phenotype, whereas HBECs cultured on types I + III collagen mimicked the resorptive phenotype. The resorptive phenotype, characterized by greater basal 92 kD gelatinase expression at both the protein and the mRNA level, may indicate an ability of HBECs to actively participate in matrix remodeling and wound repair. Lung-injury-induced basement-membrane denudation may put HBECs in contact with components of the ECM such as collagen types I and III, which can be degraded by the triple helicase activity of gelatinase as distinct from that of interstitial collagenase (10). Recent work by Buisson and colleagues (11) showed that the 92-kD gelatinase produced by human surface respiratory epithelial cells from nasal polyps and cultured on type I collagen actively contributed to wound repair in the respiratory epithelium.

Conceivably, the homeostatic phenotype of HBECs characterized by a smaller degree of basal 92-kD gelatinase expression at both the protein and the mRNA levels might be maintained by interaction with type IV collagen, which is a major component of the underlying basement membrane. Our data are in accordance with those in previous work demonstrating that constitutive 92-kD gelatinase expression by human keratinocytes was enhanced by type I collagen and decreased by type IV collagen matrices (9).

Similarly, we demonstrated that apposition of HBECs to type IV collagen in the presence of LPS or of the proinflammatory cytokines IL-1 and TNF- was associated with less upregulation of 92-kD gelatinase, at both the protein and the mRNA levels, than was their apposition to types I + III collagen. Also, apposition of HBECs to type IV collagen induced loss of gelatinase activation. These data confirm that, in our cell-system model, the ECM can regulate 92-kD gelatinase activation and gene expression, under both basal and stimulated conditions.

The specific nature of the 92-kD gelatinase upregulation observed with both collagen matrices was demonstrated by the absence of any changes in the other investigated proteins, such as 72-kD gelatinase and TIMP-1. TIMP-1 production remained fairly constant regardless of the inflammatory mediator or collagen substratum used. Also, although synthesized as a minor enzyme, the 72-kD gelatinase was produced at a constant rate, and the sum of 72-kD pro- form and 68-kD active form produced with types I + III collagen as the matrix substrate was similar to the amount of the 72-kD pro- form produced with type IV collagen.

Loss of 72-kD and 92-kD Gelatinase Activation on Type IV Collagen Substratum

In addition to synthesis and inhibition, activation is a critical event in the regulation of gelatinase activity, and may be essential for ECM degradation. The progelatinases, like other members of the MMP family, are secreted as latent compounds that require activation. We found that activation of both the 92-kD and the 72-kD gelatinases by HBECs was clearly increased when types I + III collagen were used as the substratum, both under basal conditions and in response to LPS, IL-1, or TNF-, whereas the activation disappeared when type IV collagen was the substrate. The physiologic mechanisms responsible for progelatinase activation are not completely understood, but may involve other proteases, including other MMPs. Recent studies have identified matrilysin (12) and a plasma transmembrane-dependent activation specific for 72-kD gelatinase and belonging to the subfamily of MT-MMP (13). Several enzymes have been shown to activate the pro- form of 92-kD gelatinase, including stromelysin-1 (14, 15), plasmin (16), and kallikrein (17). In our cell-system model, we consistently found both the 88-kD and 68-kD active forms of the 92-kD and 72-kD gelatinases, respectively, suggesting that the 92-kD pro- form may be activated at least in part by the active form of 72-kD gelatinase. In accord with this hypothesis is the recent work suggesting (18) that activation of the pro- form of 92-kD gelatinase may be enhanced by the active species of 72-kD gelatinase.

We have also sought to determine the ability of HBECs to actively produce either MMP-3 or MT-MMP in the presence of types I + III collagen. Investigation of MMP-3 with casein zymography has shown that expression of this MMP was positive but weakly detectable in HBEC culture medium even when the latter was 20-fold concentrated, and was not enhanced in the presence of types I + III collagen as compared with type IV collagen. Concomitant investigation with immunoblotting of MT1-MMP in cell lysates showed that this MMP was clearly expressed by HBECs, but that its expression was not upregulated in the presence of types I + III collagen as compared with type IV collagen. These results allowed us to suggest that neither MMP-3 nor MT1-MMP may contribute to the respective differences in 92-kD or 72-kD gelatinase activation induced by cell-matrix interactions under our experimental conditions.

The persistent activation of 92-kD gelatinase, and the production of this enzyme in greater amounts by HBECs plated on types I + III collagen, support the concept of an HBEC resorptive phenotype with this substrate. Similarly, the loss of gelatinase activation and the reduced production of 92-kD gelatinase by HBECs plated on type IV collagen argue in favor of a homeostatic phenotype of HBECs grown on type IV collagen.

TIMP-1 Expression Did Not Change with Collagen Type

The contribution of individual proteinases to the activation of progelatinases is probably determined not only by the amount and accessibility of the enzyme, its state of activation, and the rate of formation of active species, but also by the presence of specific inhibitors. The apparent lack of detectable TIMP-2 secreted by HBECs (5) in response to LPS, IL-1, or TNF-, when types I + III collagen were the substrate may promote the induction/endogenous activation of 72-kD gelatinase. The concomitant stable or slightly decreased production of TIMP-1 may facilitate activation of the 92-kD gelatinase pro- form in response to LPS or proinflammatory cytokines. Under these stimulated conditions, the imbalance between 92-kD gelatinase and its TIMP-1 inhibitor was more in favor of the proteinase when the cells were cultured on types I + III collagen, further amplifying the resorptive phenotype.

Concerning the 24 to 25-kD TIMP-3 known as "epithelium TIMP," all our assays previously done with immunoblotting (5) failed to demonstrate its presence even in 20-fold concentrated culture medium of HBECs under either basal or inflammatory conditions. This result, in conjunction with the lack of detection of TIMP-2 and TIMP-3 with reverse zymography, strongly suggests that, at present and in our cell-culture system, TIMP-1 appears as the predominant inhibitor of MMPs.

Potential Role of Integrins in Modulating Gelatinase Expression by HBECs

ECM ligands, their receptors, and ECM proteases and proteinase inhibitors all participate in the construction, maintenance, and remodeling of the ECM by cells. It is well known that the ECM provides cells with instructive clues, and that signals originating at the cell surface via ECM- ligand/integrin-receptor interactions may directly affect gene expression (2). As early as 1989, signal transduction through the fibronectin receptor was shown to induce interstitial collagenase (MMP-1) and stromelysin-1 (MMP-3) gene expression by fibroblasts (19). Subsequent studies indicated that MMP-1 expression by osteogenic cell lines was regulated by the collagen receptor ₂₁ integrin (20). Also, Arner and coworkers demonstrated that signal transduction through integrin receptors upregulated MMP expression by chondrocytes (21).

As with other cells, epithelial cells must receive clues from the ECM via integrins to develop normally, establish epithelium polarity, and repair injury to the epithelium. In vivo, normal adult human airway epithelium constitutively expresses at least five different integrins, including ₅₁, ₂₁, and ₃₁ (22). The ₅₁ integrin is involved in wound repair of airway epithelium (23), whereas the ₂₁ and ₃₁ receptors for collagen may be regulated by growth factors. To our knowledge, there is no evidence to date that integrin specificity is related to collagen type. When we compared HBEC proliferation and morphology according to the type of collagen used as the substrate, we failed to detect any important differences in these two parameters. This strongly suggests that interactions between matrix ligands (types I + III collagen or type IV collagen) and related integrin receptors may be largely similar, and that any differences may induce only limited changes in the organization of the internal cytoskeleton network, with no important modifications in matrix remodeling. This hypothesis is consistent with the moderate difference in gelatinase expression with the two substrates observed in our study. Nevertheless, although moderate, the restrictive modulation of 92-kD gelatinase production in response to type IV collagen (versus types I + III collagen) was consistent, significant, and always accompanied by loss of activation of both gelatinases, reflecting loss of some of the proteinases involved in this activation process.

Previous work (24) has shown that type IV collagen, via its ₃(IV) chain, downregulates neutrophil activation via a decrease in ·O₂ production at the same time as it decreases elastase and 92-kD gelatinase secretion, thus restricting the potential for damage as neutrophils cross the capillary wall. In addition to the major ₁(IV) and ₂(IV) chains of type IV collagen, a minor ₃(IV) chain has been evidenced in most basement membranes, except that of mouse Engelbreth-Holm-Swarm (EHS) tumor. Thus, the ₃(IV) chain present in human placental type IV collagen that we used as a matrix substrate for HBEC cultures might be of relevance in restricting the degrading properties of HBECs. Also, the expression of ₃(IV) mRNA at high levels in lung tissues (25) supports the concept that this chain may really act in vivo in a self-protection mechanism. Moreover, a recent report (26) proposed that the ₃(IV) chain of type IV collagen found around some lung-invasive tumor clusters may impede tumor invasion and may be considered part of the general remodeling of the ECM observed in cancer.

In conclusion, our enzymatic and RT-PCR data show that primary cultures of HBECs exhibit differential gelatinase expression in response to the collagen type used as a matrix substrate. The same differences were evidenced under basal conditions and during exposure to LPS or proinflammatory cytokines. The restrictive modulation of 92-kD gelatinase expression, as well as the loss of both 92-kD and 72-kD gelatinase activation in the presence of type IV collagen, strongly suggest that this type of collagen is associated with the homeostatic HBEC phenotype, and limits the ability of HBECs to degrade the matrix. Thus, the ECM underlying HBECs may modulate matrix remodeling by cells, particularly during inflammatory processes such as acute lung injury.