Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Both scar tissue and fibrotic tissues undergo contraction. In order to evaluate the mechanistic basis for this process, contraction of three-dimensional type I collagen gels has been used as a model system (1). Fibroblasts cultured inside such a gel cause contraction. Mediators believed to promote fibrosis and wound healing can augment fibroblast-mediated contraction. Other agents can inhibit the process, suggesting potential antifibrotic activity.

Whereas fibroblasts likely play a key role in theopment of fibrosis, other cells may also participate. Several types of epithelial cells can mediate collagen gel contraction (2- 4). These include retinal epithelial cells that could contribute to retinal detachment and bronchial epithelial cells that could contribute to the airway narrowing that characterizes chronic obstructive pulmonary diseases and asthma. Fibrosis of the alveolar parenchyma is a feature of several important lung diseases, including idiopathic pulmonary fibrosis (IPF). This, therefore, raises the question of whether alveolar epithelial cells can also mediate tissue contraction and thereby contribute to the structural alteration and loss of function. The current study was designed to evaluate this question.

To accomplish this, we extracted collagen from rat tail tendons. This collagen is almost entirely pure type I collagen and was used to prepare gels made of native type I collagen fibers. A549 cells, a cell line derived from an alveolar cell carcinoma, were then plated on top of or cast into the gels. Although A549 cells have been widely used as a model of alveolar cells in vitro (5), they are derived from a cancer (6) and, therefore, for comparison, primary alveolar type II cells isolated from rat lungs were also studied. Our results demonstrate that alveolar epithelial cells can contract collagen gels and, therefore, could be active participants in processes that lead to alveolar remodeling.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials

The A549 cell line was purchased from American Type Culture Collection (ATCC; Rockville, MD) at passage 76 and grown in Ham's F-12 medium (GIBCO BRL, Grand Island, NY) with 10% fetal calf serum (FCS; GIBCO BRL). Human bronchial epithelial cells (HBEC) were prepared from a healthy volunteer by following the procedure of Kelsen and colleagues (7). They were cultured in a 1:1 mixture of LHC-9 (8) and RPMI 1640 media (LHC-9/RPMI) (9). These cells have been frozen at passage 6. For use in these experiments, they were thawed, cultured in LHC-9/RPMI medium, and used in replica experiments after five to six additional passages. Rat alveolar epithelial cells (RalvEC) were isolated by following the procedure of Dobbs and coworkers (10), and maintained in F-12 medium with 10% FCS. They were used either as primary cultures or after the first passage. HFL1, a human fetal lung fibroblast cell line, was purchased from ATCC and maintained in Dulbecco's modified Eagle's medium (DMEM; GIBCO BRL) with 10% FCS. It was used at passages 16 to 18. Recombinant human transforming growth factor- (TGF-₁; #240- B-010), recombinant human tumor necrosis factor- (TNF-; #210-TA-050), recombinant human interleukin-1 (IL-1; #201-LB-005), and recombinant human interferon- (IFN-; #285-IF-100) were purchased from R&D Systems (Minneapolis, MN). Prostaglandin E₂ (PGE₂; #p-6532) was obtained from Sigma (St. Louis, MO). Monoclonal antihuman ₁ and ₂ integrin antibodies (#12086-013 and #12087-011) were from GIBCO BRL.

Preparation of Collagen Gels

Type I collagen stock solution was prepared from rat tail tendons as previously described (11). Collagen gels were prepared by mixing collagen solution, distilled water, and 4× DMEM so that the final concentration was 1× DMEM and final collagen concentration was 1.0 mg/ml. After this, 500 µl of the mixture was cast into each well of a 24-well culture plate. The solutions gelled in approximately 20 min at room temperature. Cells were trypsinized and plated on top of the solidified gels (monolayer). Alternatively, cells were suspended in the collagen solution and the mixture was immediately cast into the wells (embedded culture). After allowing 2 to 8 h for the cells to attach and spread, gels were released from the well in which they were cast with a sterile spatula, floated in culture medium, and incubated at 37°C in a 5% CO₂ atmosphere.

Gel Contraction Assay

To evaluate the contraction quantitatively, the area of the collagen gels was measured as a function of time using an Optomax V image analyzer (Optomax, Burlington, MA). For A549 cells and RalvEC, gels were maintained in F-12 medium with 0.5% FCS. For comparison, gels with HBEC maintained in LHC-9/RPMI medium and gels with HFL1 maintained in serum-free DMEM were used.

Manner of Culture

To determine if culture techniques affect the collagen gel contraction, A549 cells, RalvEC, HFL1, and HBEC were either plated on top of the gels at 10⁵ cells/cm² (monolayer culture) or cast into the gels at a density of 4 × 10⁵ cells/ml (embedded culture). (The total cell number was 2 × 10⁵/gel for both conditions.) F-12 medium with 0.5% FCS was used for A549 cells and RalvEC. LHC-9/RPMI medium was used for HBEC. Serum-free DMEM was used for HFL1.

Effect of Exogenous Mediators

To determine the effect of selected exogenous mediators, the following were tested: TGF- (0 to 400 pM), PGE₂ (0 to 10 µM), and cytomix, a mixture of equal weight of TNF-, IL-1, and IFN- (0 to 10 ng/ml each). TGF- was tested because it is a potent stimulus for fibroblast-mediated contraction, PGE₂ because it is a potent inhibitor, and cytomix because it has proved a potent stimulus for many cell types. These were added to the wells at the time of plating the cells. The same concentrations of these compounds were also put into the floating media at the time of release. A549 cells, HBEC, and RalvEC were plated on top of the gel as a monolayer at 1 × 10⁵/cm². HFL1 was embedded inside the gel at 4 × 10⁵/ml.

Effect of Anti-Integrin Antibodies

To determine the dependence of A549 cell-mediated contraction on ₁ and ₂ integrins, anti-₁ and anti-₂ integrin antibodies (1:1,000 final concentration) were added to cultures of attached cells approximately 2 h after plating. After 48 h, the gels were released, floated in F-12 medium with 0.5 % FCS, and the gel size was measured as a function of time.

Hydroxyproline Assay

The amount of hydroxyproline, which is directly proportional to the collagen content, was measured by a spectrophotometric assay (12). Briefly, the collagen gels were added to 2 ml of 6 N hydrochloric acid and hydrolyzed at 110°C for 12 h. After drying and reconstituting, hydroxyproline was quantified colormetrically with Ehrlich's reagent measured at 550 nm with a spectrophotometer.

Morphologic Observation and Immunocytochemistry

For morphologic observation, A549 cells were plated on top of collagen gels and cultured with or without 5 ng/ml cytomix in F-12 medium with 0.5% FCS. After 48 h of incubation without release, the collagen gels were stained by Diff-Quik (Sigma, St. Louis, MO). For immunocytochemistry, cells were cultured in chamber slides (Falcon no. 344108; Becton Dickinson, Franklin Lakes, NJ) with or without treatment with 5 ng/ml cytomix in F-12 medium with 1% FCS for 48 h. They were fixed by 4% paraformaldehyde for 5 min and stained by a modified avidin-biotin complex method. For immunoperoxidase staining, primary antibody for cytokeratin (#C2931) was obtained from Sigma, and the primary antibodies for vimentin (#M0725) and -smooth muscle actin (#M0851) were obtained from DAKO (Carpinteria, CA). A Vectastain kit (Vector Laboratories, Burlingame, CA) was used for visualization. For -tubulin, ₁ integrin, and paxillin, immunofluorescence staining was performed by replacing the peroxidase-conjugated streptavidin in a Vectastain kit by fluoroscein isothiocyanate (FITC)-conjugated streptavidin (#9538SA; GIBCO BRL). The primary antibody for -tubulin (#T5168) was from Sigma, and the primary antibody for paxillin (#P13520) was from Transduction Laboratories (Lexington, KY).

Statistical Analysis

Each experiment includes gel contraction assays performed in triplicate. The values are shown as mean ± standard error (n = 3) unless otherwise described. Statistical comparison was done by two-sample t test.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Density-Dependent Manner of Gel Contraction by A549 Cells

A549 cells significantly contracted the gels in a time- and cell density-dependent manner when cultured in F-12 medium with 0.5% FCS (Figure 1). Similar results were observed in LHC-9/RPMI medium (data not shown). Maximal contraction was observed at a density of 10⁵ cells/cm² where gel area was reduced to 61 ± 0.5% of control after 4 d (P < 0.05). These results were consistently observed; in four separate experiments, A549 cells contracted collagen gels significantly.

View larger version (21K): [in this window] [in a new window]   Figure 1.   Effect of cell density on A549-mediated collagen gel contraction. A549 cells were plated at various densities as indicated on the figure as a monolayer on top of collagen gels. The gels were released into F-12 medium with 0.5% FCS after 3 h of incubation. Gel area was then measured daily with an image analyzer. Horizontal axis: time after release. Vertical axis: gel area expressed as a percentage of initial size.

Effect of the Manner of Culture and the Cell Types

A549 cells contracted the gel more when they were plated as a monolayer on top of the gel than when they were cast inside the gel (63.4 ± 0.5% versus 81.6 ± 1.7% of initial size; P < 0.05) (Figure 2). A similar result was observed with primary cultures of RalvEC. Thus, the alveolar epithelial cells resembled HBEC, which also contract gels more when plated on top of the gel, and contrasted with HFL1, which contracted the gels more when embedded inside the gels. Whereas alveolar epithelial cells reliably contracted the gels, there was significantly less contraction than resulted from a similar density of normal HBEC (P < 0.05). The epithelial cells (A549 cells, RalvEC, and HBEC) were observed to be well spread when they were plated on top of the gel, whereas they remained rounded when they were embedded inside the gel (data not shown). In contrast, HFL1 was well spread both when they were plated on top and embedded inside the gels (data not shown).

View larger version (58K): [in this window] [in a new window]   Figure 2.   Effect of culture technique on gel contraction by various cell types. Cells were cultured as a monolayer on top of the gels (1 × 105/cm) or cultured embedded inside the gels (4 × 105/ml). The total cell number was 2 × 105/gel for both techniques. The area of the gels was measured 72 h after release. Horizontal axis: cell types and culture technique. Vertical axis: gel area expressed as percentage of initial size. *P < 0.05.

Effect of Exogenous Mediators

The contraction caused by A549 cells was significantly augmented by TGF- in a concentration-dependent manner (Figure 3). A significant, measurable effect was observed at 100 pM, and the maximal effect was obtained at the highest concentration tested, 400 pM. The contractions by HBEC and HFL1 were also augmented by TGF-.

View larger version (16K): [in this window] [in a new window]   Figure 3.   Effect of TGF- on collagen gel contraction mediated by A549 cells and other cell types. A549 cells (circles) and HBEC (diamonds) were plated on top of the collagen gels, and HFL1 (triangles) was embedded inside the gels as described in MATERIALS AND METHODS with various concentrations of TGF- as indicated on the figure. After overnight incubation, gels were released into the culture medium containing the same concentration of TGF-. After 48 h, gel size was measured. Horizontal axis: concentration of TGF-. Vertical axis: gel size expressed as a percentage of initial size.

As reported previously (13), PGE₂ inhibited the gel contraction by HFL1 concentration dependently (Figure 4). In contrast, PGE₂ did not inhibit the contraction caused by epithelial cells (A549 cells, RalvEC, and HBEC), except at the highest concentration tested, 10⁵ M. 

View larger version (20K): [in this window] [in a new window]   Figure 4.   Effect of PGE2 on collagen gel contraction mediated by A549 cells and other cell types. A549 cells (circles), RalvEC (squares), and HBEC (diamonds) were plated on top of the collagen gels, and HFL1 (triangles) was embedded inside the gels as described in MATERIALS AND METHODS with various concentrations of PGE2 as indicated on the figure. After overnight incubation, gels were released into the culture medium containing the same concentration of PGE2. After 48 h, gel size was measured. Horizontal axis: concentration of PGE2. Vertical axis: gel size expressed as a percentage of initial size.

Cytomix (a mixture of TNF-, IL-1, and IFN-) inhibited the gel contraction by HFL1 but augmented the contraction by A549 cells in a concentration-dependent manner, except at the highest concentration tested, 10 ng/ml (Figure 5). HBEC-mediated contraction was also augmented by cytomix, except at the highest concentration.

View larger version (17K): [in this window] [in a new window]   Figure 5.   Effect of cytomix on collagen gel contraction mediated by A549 cells and other cell types. A549 cells (circles) and HBEC (diamonds) were plated on top of the collagen gels, and HFL1 (triangles) was embedded inside the gels as described in MATERIALS AND METHODS with various concentrations of cytomix as indicated on the figure. After overnight incubation, gels were released into the culture medium containing the same concentration of cytomix. After 48 h, gel size was measured. Horizontal axis: concentration of cytomix. Vertical axis: gel size expressed as a percentage of initial size.

While there was some variability in magnitude of contraction from experiment to experiment, the effect of these mediators on A549 cell-mediated contraction was consistent (Table 1).

Effect of Anti-Integrin Antibodies

Anti-₁ integrin antibody partially but significantly inhibited the gel contraction caused by A549 cells (P < 0.05) when added 48 h before release (Figure 6). Anti-₁ integrin antibody also significantly inhibited the contraction caused by HBEC (P < 0.05). In contrast, anti-₂ integrin antibody did not affect the contraction caused by either A549 cells or HBEC.

View larger version (39K): [in this window] [in a new window]   Figure 6.   Effect of anti- integrin antibodies on epithelial cell- mediated contraction of collagen gels. A549 cells and HBEC were plated as a monolayer on top of the gels. After 2 h to permit the cells to attach to the gels, anti-1 integrin (gray bars) or anti-2 integrin antibody (filled bars) was added to the well (1:1,000 dilution); control: open bars. After 48 h, gels were released. After an additional 72 h, gel size was measured by image analysis. Horizontal axis: cell types and treatment. Vertical axis: gel size expressed as percentage of initial size. Presence of antibody is indicated with shading as described on the figure. *P < 0.05.

Hydroxyproline Content of the Gels throughout the Contraction Process

To determine if alveolar epithelial cells degraded the collagen matrix on which they were cultured, the amount of hydroxyproline in each gel was measured. No differences were observed after contraction with any of a variety of conditions (Figure 7), even though the degree of contraction was quite different.

View larger version (30K): [in this window] [in a new window]   Figure 7.   Hydroxyproline content and the size of the gels contracted by A549 cells. The amount of hydroxyproline (OH-P, [mg/gel], open bars) of each gel was measured after the contraction by A549 cells. Control (0 h) shows the size (area, filled bars) and OH-P amount at the time of release. Other groups show the area and OH-P of the control, cytomix-treated (5 ng/ml) and TGF--treated (100 pM) gels after 48 h from release. Left vertical axis corresponds to filled bars and indicates the area of the gels as percentage of initial control. Right vertical axis corresponds to the open bars and indicates OH-P amount (mg/gel).

Appearance and Phenotypic Change of the Cells

The A549 cell has a polygonal shape and sheetlike pattern in normal monolayer culture, which is compatible with its epithelial origin. With cytomix treatment, the shape was changed dramatically (Figure 8). The cells acquired a spindle shape and resembled mesenchymal cells. This was observed when they were cultured on top of collagen gels and also in routine plastic dish culture. The expression of cytokeratin and vimentin was equally positive both with and without cytomix treatment as assessed by immunocytochemistry (data not shown). Smooth muscle actin expression was also positive, although there was a faint decrease in cells treated with cytomix (data not shown). However, by immunofluorescent staining, there was a dramatic change in -tubulin expression after cytomix treatment. -Tubulin was almost negative in the control condition but was strongly positive in cells treated with cytomix (Figures 9a and 9b). Paxillin was observed in plaques likely reflecting adhesion complexes and was equally positive both with or without cytomix (Figures 9c and 9d). ₁ Integrin expression was diffusely positive both with or without cytomix (Figures 9e and 9f).

View larger version (146K): [in this window] [in a new window]   Figure 8.   Effect of cytomix on the shape of A549 cells. A549 cells were pretreated with or without 5 ng/ml cytomix for 24 h and plated on top of the collagen gels. They were incubated for 24 h with the same concentration of cytomix without being released. (a) Control condition without cytomix. (b) Treated with 5 ng/ml cytomix. Diff-Quik staining. Bars indicate 50 µm.

View larger version (114K): [in this window] [in a new window]   Figure 9.   Immunofluorescence staining. A549 cells were treated with or without 5 ng/ml cytomix and plated in chamber slides. (a, c, and e) Control cells without cytomix treatment. (b, d, and f ) Cells treated with 5 ng/ml cytomix. Panels a and b were stained for -tubulin. Panels c and d were stained for paxillin. Panels e and f were stained for 1 integrin. Immunofluorescence staining using avidin-biotin complex-FITC method. Bars indicate 50 µm.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The current study demonstrates that A549 cells can cause contraction of three-dimensional gels made of native type I collagen. In addition to A549 cells, which are derived from a bronchoalveolar cell carcinoma and often used as a model of alveolar type II epithelial cells, primary cultures of RalvEC were also able to contract native type I collagen gels. The contraction of collagen gels by A549 cells could be modulated by external stimuli. Finally, whereas the contraction of collagen gels by A549 cells resembled the contraction mediated by fibroblasts in some respects, there were also some distinct differences. Together, these observations suggest that alveolar epithelial cells may contribute directly to the remodeling of the extracellular matrices of the pulmonary parenchyma.

Type II alveolar epithelial cells can proliferate and differentiate into type I cells (14, 15). Type I cells cover most of the alveolar surface but are not believed to proliferate. As a result, type II cells are believed to play a major role in the maintenance of alveolar integrity. A549 cells resemble type II cells in a number of important features, and because they are readily cultured and are derived from a human source, they are widely used as a model of type II cells. Primary cultures of type II cells have also been prepared from both animal and human sources. The current study found a similar ability of A549 cells and primary rat type II cells to contract collagen gels. Whether type I cells would also have this capability is unknown. Such an activity of type I on type II cells, for example, may play a role in repair of alveolar pores (16).

Like fibroblasts, epithelial cells have the capacity of producing and degrading collagen (17). The current study demonstrates that, like fibroblasts, epithelial cells can also reorganize an extracellular matrix by contraction. Interestingly, epithelial cells produce a different spectrum and may produce less extracellular matrix than do fibroblasts and, under the conditions used in the current study, are less potent contractors of matrix. It is possible, therefore, that tissue remodeling consequent to the actions of epithelial cells will differ from that mediated by fibroblasts.

Epithelial cells and fibroblasts also differ in their relationship to extracellular matrices. Fibroblasts in vivo are surrounded on all sides by an extracellular matrix comprising connective tissue macromolecules. In contrast, epithelial cells demonstrate polarity. One surface is in contact with a subjacent connective tissue matrix, whereas the opposite side is freely exposed to lumenal contents. This polarity is necessary for epithelial cell function and, interestingly, to epithelial cell-mediated collagen gel contraction. Fibroblasts contract gels best when cast inside the gels. In contrast, A549 cells and RalvEC resemble bronchial epithelial cells and contract the gels best when plated on the surface of the gel. When cast into the gels, the epithelial cells remain rounded in marked contrast both to their ready spreading on the surface of the gels and to the markedly elongated shape assumed by fibroblasts cast into gels.

Previous studies have suggested that ₁ integrin is required for optimal contraction of collagen gels by both fibroblasts (20) and bronchial epithelial cells (4). ₁ Integrin also appears necessary for optimal contraction of collagen gels by alveolar epithelial cells as antibodies to ₁ integrin inhibited A549 cell-mediated gel contraction. Consistent with this, A549 cells expressed surface ₁ integrin as previously reported (21), as well as cytoplasmic staining similar to that observed in other cell types (22). Thus, although epithelial cell and fibroblast spreading and contraction of collagen gels differ in important ways, some mechanisms may be shared.

A549 cells also differed from fibroblasts in their response to external modulators of collagen gel contraction. TGF- is a cytokine thought to play an important role in tissue repair and previously demonstrated to stimulate fibroblast-mediated contraction of collagen gels (23). It also stimulated contraction of collagen gels mediated by A549 cells and HBEC. In contrast, PGE₂, a potent concentration-dependent inhibitor of fibroblast-mediated gel contraction (13), had no effect on A549 cells or HBEC contraction of collagen gels. Finally, cytomix inhibited the contraction of collagen gels mediated by fibroblasts but stimulated that mediated by epithelial cells.

Cytomix is a mixture of TNF-, IL-1, and IFN-. It has been previously reported to be a potent stimulator of inducible nitric oxide synthase in A549 cells (24, 25). Cytomix also resulted in a marked change in the morphologic appearance of A549 cells, leading us to explore the changes induced in the cytoskeleton and adhesion proteins of A549 cells by cytomix. Minimal changes were noted in staining for cytokeratin, vimentin, -smooth muscle actin, plaquelike structures of paxillin, or ₁ integrin. In contrast, cytomix induced a marked increase in expression of -tubulin. Although this observation does not define the mechanism by which cytomix leads to augmented contraction of A549 cells, it does suggest that interactions with specific components of the cytoskeleton may mediate such effects.

Remodeling of the extracellular matrix requires several interconnected steps, including matrix molecule production, fiber formation, fiber organization, and matrix degradation. The present study demonstrates that alveolar epithelial cells can cause fiber reorganization, at least in terms of contraction, and hence density. Other studies have demonstrated the ability of alveolar epithelial cells to both produce and degrade extracellular matrices (26). In the current study, total collagen content of the gels was largely unchanged. The gels used were made from purified type I collagen. Matrix components synthesized by the cells, however, could have been present in the gels and degradation of these compounds would not be detected by the assay methods used. Small amounts of collagen synthesis or degradation would not have been observed with the assay method used. Larger degrees of degradation, for example in excess of 20% of the starting material, are readily observed by the methods used (27). Therefore, degradation of this degree does not appear to be involved in the contraction of collagen gels under the assay conditions used. It remains possible, of course, that degradation of collagen could be a mechanism for alveolar epithelial cell matrix remodeling under other conditions.

Fibrosis in the alveolar structures is an important feature of many diseases, including IPF, bronchiolitis obliterans organizing pneumonia (BOOP), and adult respiratory distress syndrome (ARDS). The outcome of these diseases is often closely related to the remodeling associated with foci of fibrosis. In IPF, the foci of intralumenal fibrosis are incorporated into the alveolar walls and covered with regenerated epithelial cells. This leads to a permanent alteration of alveolar architecture and associated loss of function (28). In BOOP, the polypoid-shaped, intralumenal organizing process is also covered by regenerated epithelial cells but is generally completely absorbed without causing permanent alterations in alveolar structure (29). If this process takes place effectively, the lesion may resolve without irreversible functional loss (30). In ARDS, the long-term outcome of individuals who survive the first several weeks is variable. Some individuals recover completely. Others are left with permanently scarred, functionally compromised lungs, while yet others, who appear to have initially developed fibrosis, may have substantial restoration of normal structure and function (31). The ability of alveolar epithelial cells to reorganize the extracellular matrix may play a crucial role in determining such outcomes.

In summary, A549 cells and rat alveolar type II epithelial cells are capable of mediating the contraction of native type I collagen gels. Through such mechanisms, alveolar epithelial cells could participate in the tissue remodeling that characterizes both the maintenance of normal tissue structures and the development of tissue alterations that contribute to disease.