Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

While the mechanisms underlying the initiation of acute airway inflammation following challenge with allergens are reasonably well understood, far less is known about the process underlying longer-term structural changes occurring in the airways of patients with asthma. Key features of this airway remodeling include smooth muscle hyperplasia and hypertrophy as well as an alteration and increase in extracellular matrix (ECM) (1). In vitro studies have focused on airway myofibroblasts, which are the source of ECM in the airways and are the natural precursors for smooth muscle cells (SMCs) themselves (4).

Using primary culture of airway myofibroblasts derived from explants or dispersed cell preparations from human airway tissue, a number of groups have defined important mitogenic stimuli for these cells. Agents such as thrombin, platelet-derived growth factor (PDGF), epidermal growth factor (EGF), histamine, and mast cell tryptase have all been shown to increase either DNA synthesis or cell number (5) and are thus thought to contribute to the pathogenesis of airway remodeling.

Biopsy and postmortem studies report increases in deposition of fibronectin, collagen I, III, and V, laminin, tenascin, versican, and hyaluronan in the airways of individuals with asthma (6). Not only do airway myofibroblasts produce extracellular matrix proteins and glycoproteins, but Johnson and coworkers have demonstrated that airway SMCs in culture can be stimulated to increase production of fibronectin, perlecan, laminin ₁, and chondroitin sulfate above baseline by the addition of asthmatic serum (10). Less is known about the effects of this altered matrix on the SMCs. Hirst and colleagues have studied the effect of different matrix components on mitogen-stimulated SMC proliferation and phenotype (11). They report that fibronectin and collagen I enhance proliferation and encourage expression of the nuclear proliferation marker Ki67, whereas cells grown on laminin divide more slowly but express contractile proteins. This points to the fact that cell-matrix interactions, in addition to growth factors and cytokines, have important effects on myofibroblast phenotype and cell cycle control.

Cells interact with the ECM via a family of cell-surface glycoprotein receptors called integrins. These exist as heterodimers of  and  transmembrane subunits in a variety of combinations (12). The extracellular domains recognize short peptide sequences (e.g., Arg-Gly-Asp [RGD] found on some ECM proteins [e.g., fibronectin and vitronectin]), while the intracellular domains are involved in the formation of focal adhesion complexes and in signaling events with implications for cell adhesion, migration, differentiation, and cell survival.

To date, little is known about survival signals for airway SMCs, yet the number of airway SMCs present in the airway wall is dependent upon the balance between cell proliferation and cell death. The aim of the studies described below was to investigate the effect of known mitogens and different ECM components on cell survival and to examine the effect of matrix factors on airway SMC morphology. We characterized integrin expression in human airway SMCs smooth muscle cells and examined which integrin subtype(s) were involved in the survival signal transduction to gain further insight into the mechanisms underlying airway remodeling.

Here we report that airway SMCs have low background rates of apoptosis due to a strong survival signal that results from their interaction with the ECM. We conclude that changes in the ECM in the inflamed airway may be important in airway remodeling both by contributing to altered airway wall morphology and by contributing to the survival signal of airway SMCs.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Primary cultures of human airway SMCs were prepared from explants of trachealis muscle obtained from individuals without respiratory disease within 12 h of death as previously described (13). Briefly, a segment of trachea was removed from immediately above the carina and a strip of trachealis dissected clear of surrounding tissue. After initial washes in Dulbecco's modified Eagle's medium (DMEM) containing penicillin G (200 U/ml), streptomycin (200 µg/L), and amphotericin B (0.5 µg/L), the overlying mucosa was dissected free from the airway smooth muscle under sterile conditions. 0.2 × 0.2 cm explants of airway muscle were excised and placed in 6-well plates. After allowing explants to adhere, DMEM containing the antibiotics, amphotericin B, 10% fetal bovine serum (FBS), and glutamine (2mM) were added to just cover explants. This medium was replaced regularly until muscle cell growth occurred (usually between Days 7-10), when culture medium was changed to DMEM containing 10% FBS and 2mM glutamine. As cells were approaching confluence in some parts of the well, explants were removed and 24 h later cells were harvested by trypsinization and plated in a 25 cm² flask. For subsequent passages we used 75-ml flasks. Hall (14) and others (15) have extensively characterized the phenotype of these cells: all primary cell cultures from each donor were examined using anti-smooth muscle -actin antibody (1:100 dilution) to confirm the presence of smooth muscle type cells using standard immunocytochemical techniques. All cell cultures used for the experiments described in this paper showed > 95% of cells staining for smooth muscle  actin. Cells were used between passages 4 and 10 and triplicate experiments were performed using cells from 3 different donors.

Assessment of Apoptosis by Propidium Iodide Counting

We quantified the rates of apoptosis in established monolayers grown in 96-well plates by serum-starving confluent cells for 48 h before culturing them for 18 h at 37°C in 5% CO₂ in the presence of additional antibodies, peptides, and/or serum as detailed below. In the case of the growth factor experiments 60% confluent cells were cultured for 48 h in the presence of growth factors or serum. The undisturbed cultures were then fixed for 1 at room temperature by adding formaldehyde solution to a final concentration of 4%. The medium was removed and phosphate-buffered saline (PBS) containing 40 µg/ml propidium iodide (PI) added to visualize cells using inverted fluorescent microscopy (Nikon Diaphot 300 with epifluorescent capabilities; Nikon, Kingston-upon-Thames, UK). Small rounded cells with pyknotic nuclei were counted and expressed as a percentage of the total cells (16). Three fields were examined per well and triplicate wells were counted and averaged within each experiment. Thus, a minimum of 500 cells were counted for each condition in every experiment. Each experiment was performed a minimum of three times and results are shown as mean ± SEM. The observer was blinded to the conditions in each well until after counts were completed.

To investigate the role of inhibitors on integrin-mediated signaling we quantified apoptosis 18 h after seeding, employing a method adapted from Aplin, Howe, and Frisch (12, 17, 18). For these studies cells were harvested after 48 h serum starvation using 0.5% trypsin/2mM EDTA. After adding 1 mg/ml soy trypsin inhibitor, the cells were pelleted, washed in 1% bovine serum albumin in PBS (PBA), resuspended in serum-free medium containing the antagonist, peptide, or antibody, and allowed to incubate for 20 min at 37°C with gentle agitation. They were then plated at a density of 10⁴/well in 96-well plates and returned to the incubator for 18 h before being fixed, stained, and counted in a blinded fashion as detailed above.

Assessment of Apoptosis by Transferase-Mediated Deoxyuridine Triphosphate Nick-End Labeling-Based Assay

After 48 h serum-starvation cells were harvested, treated with trypsin inhibitor, washed, and exposed to various antagonists as described above, but plated at a density of 1.6 × 10 ⁵ cells per well on a 96-well plate. The wells were then analyzed with a transferase-mediated deoxyuridine triphosphate nick-end labeling (TUNEL) assay. Each condition was examined in triplicate using the colorimetric-based kit Titer Tacs (R&D Systems, Abingdon, UK) according to the manufacturer's instructions. Positive controls were created using the kit's nuclease enzyme.

Coating of Culture Plates with ECM Factors

Unless specifically indicated, experiments were conducted in uncoated wells and cells were seeded onto tissue culture plastic. For some experiments, however, ECM substrata were prepared by allowing 10 µg/ml solution of fibronectin, collagens I, IV, V, vitronectin, laminin, elastin, or 100 µM RGD to adhere to culture plates overnight (19), before blocking with 1% PBA for 2 h. To inhibit matrix deposition, wells were coated with 12% solution of Poly-2-Hydroyethyl Methacrylate (HEMA) and air-dried. All plates were rinsed twice with PBS and sterilized by UV light prior to use.

Assessment of Integrin Expression by Fluorescence Activated Cell Sorting

Cells were trypsinized, washed in 1% PBA, and incubated with 2 µg primary antibody per 10⁵ cells for 30 min in the dark at room temperature. Following a rinse in PBA, the cells were incubated with 2 µg/10 ⁵ cells secondary antibody for another 30 min before two further PBA washes. Samples were resuspended in fluorescence activated cell sorting (FACS) Fix (0.5% Formaldehyde in Isoton sheath fluid; Becton Dickinson, Cowley, UK) and acquired and analyzed using a Beckman Coulter (High Wycombe, UK) Epics XL flow cytometer equipped with Expo 30 software.

Assessment of Morphology by Confocal Microscopy

Cells were prepared as above and plated onto ECM-coated 22-mm microscope glass slides, which resided in the wells of 6-well plates. After 18 h of culture, the cells were fixed with 4% Formaldehyde for 30 min at room temperature, rinsed with 1% PBA, and permeabilized (20mM HEPES, pH 7.6, 300 mM sucrose, 50 mM NaCl, 3 mM MgCl₂, 0.5% Triton X-100) at -20°C for 5 min. Following a further rinse in PBA, the cells were stained with 10 µg/ml fluorescein isothiocyanate (FITC)-labeled Phalloidin at 4°C for 20 min, rinsed 3 times, stained with PI as above, washed another 3 times, and mounted onto microscope slides with 1,4 Diazabicyclo[2 · 2 · 2] octane (DABCO) 2 mg/ml in PBS and glycerol as described by Johnson (20). Representative areas of the samples were viewed using a Zeiss (Welwyn Garden City, UK) LSM510 confocal microscope and a Zeiss Axiovert 100 M inverted fluorescence microscope with a Planapochromat 63× 1.4 oil objective. The Argon 488 nm laser line was used to scan the green fluorescence of the FITC-conjugated phalloidin while the Helium Neon 543 laser line was used to image the PI stain. Images are displayed as projections of some or all of the slices using the software provided (LSM 510).

Materials

Chemicals used were analytical grade or higher. 96-well plates were sourced from Nunc (Loughborough, UK) and other plastic culture ware from Costar (High Wycombe, UK). The Thromboxane A₂ mimetic U46619 and all ECM proteins except for elastin were from Calbiochem (Nottingham, UK), elastin from Elastin Products (Pacific, MO) and PBS and soy trypsin inhibitor from Gibco (Paisley, UK). All other compounds were bought from Sigma (Poole, UK). The growth factors were a gift from Claire Ditchfield (Nottingham).

The functionally blocking anti-integrin antibodies used were mouse anti-₁ (Upstate Biotechnology, Buckingham, UK), mouse anti-₃ (Clone C3), mouse anti-₄ (Clone HP2/1), mouse anti-₅ (Clone SAM1), rat anti-₆ (Clone Go H3), rat anti-_v (Clone 69.6.5), mouse anti-₂ (Clone 7E4), and anti-₃ (Clone SZ21), all from Immunotech (Luton, UK), and mouse anti-₁ (Serotec, Kidlington, UK). As a positive control, we employed mouse and rat anti-₂ microglobulin (Clone B1G6, Immunotech; and Clone YTH470.5, Serotec, respectively) and as a negative control, mouse and rat anti-IgG (Clone 679.1Mc7 and purified, both from Immunotech). For FACS we used the following secondary antibodies: Phycoerythrin-conjugated goat anti-mouse and FITC-conjugated rabbit anti-rat (both DAKO, Ely, UK). For immunostaining we used mouse anti--smooth muscle actin (Sigma).

Statistical Analysis

The EC₅₀ for cycloheximide was defined in each individual experiment and used to calculate mean values. Each data point in individual experiments was calculated from the mean of triplicate determinations.

Statistical analysis of data was performed using the t test for single comparisons. Where multiple comparisons were made within the same experiment, ANOVA was performed, followed by Bonferroni's or Dunnett's tests where appropriate. All values in the text represent means ± SEM of n separate experiments.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Background Rates of Apoptosis and Effects of Cytokines

To establish a baseline, we assessed the background rates of apoptosis in human airway myofibroblasts in culture under routine serum-containing and serum-free conditions. Levels of apoptosis were low both in serum (5.0 ± 1.2%, n = 3) and following serum starvation for 66 h (8.9 ± 0.8%, n = 3, P = 0.07). Because of reports detailing the positive effect of growth factors on airway myofibroblast proliferation, we set out to examine the effect of these potential survival factors on apoptosis (Figure 1A). Addition of the known mitogens at the concentrations shown did not affect rates of apoptosis.

View larger version (28K): [in this window] [in a new window]   View larger version (22K): [in this window] [in a new window]   Figure 1.   Effect of growth factors on apoptosis. Cells were grown on tissue culture plastic; apoptotic cells were assessed by PI staining and are expressed as % ± SEM of all cells counted. N = 3. SFM, serum-free medium.(A) Growth factors were added to 60% confluent serum-starved cells at the concentrations shown. Cells were analyzed 48 h after addition of growth factors (B) Serum-starved cells were incubated with 50 µM CHX and seeded in the presence or absence of growth factors as shown, and samples were analyzed 18 h after plating. *P < 0.01 compared with `SFM' by ANOVA.

Given this low background rate of apoptosis, even following prolonged serum starvation, we surmised that airway myofibroblasts fail to undergo apoptosis because they receive a strong survival signal. To investigate the potential nature of this survival signal, we seeded cells in the presence of cycloheximide to inhibit new protein synthesis and were able to demonstrate a marked increase in rates of apoptosis (54.5 ± 4.7%, n = 3) compared with cells seeded in the absence of cycloheximide (2.7 ± 0.3%, n = 3, EC₅₀ = 21.2 ± 1.9 µM). This effect could not be reversed by the addition of growth factors, but was diminished by the addition of serum (19.9 ± 3.2%, P < 0.01 versus CHX, n = 3) (Figure 1B). These data suggest that a strong anti-apoptotic signal other than that received through the mitogens studied is responsible for survival signaling for airway myofibroblasts. Adding cycloheximide to confluent monolayers of airway myofibroblasts fails to increase apoptosis (2.5 ± 0.2% versus 1.4 ± 0.8%, P > 0.05, n = 3), implying that the necessary components for survival signaling are already present in adherent monolayer culture (data not shown).

Effect of Matrix Factors on Survival

To study the effect of the individual matrix components on cell survival, we seeded the airway myofibroblasts onto the individual ECM substrates: apoptosis was uniformly low under these conditions (range of apoptosis 3.4 ± 0.3% to 8.4 ± 4.4%, n = 3), except in the cells seeded on elastin (apoptosis 31.5 ± 8.7%, n = 3), which, however, does not bind cells via integrins (Figure 2A). We hypothesized that, similar to other adherent cells, myofibroblasts are dependent on the ECM for a survival signal (21). We therefore inhibited matrix deposition by seeding cells into wells coated with HEMA, which resulted in 100% apoptosis. This effect is not due to toxicity because adding HEMA to confluent adherent monolayers did not increase cell death (n = 3; results not shown). To confirm that it is indeed the matrix factors that provide the critical survival signal, we again inhibited the cells' own protein synthesis by adding cycloheximide but provided them with a substratum of preformed matrix factors. Fibronectin, collagens I and IV, and laminin all produced marked inhibition of apoptosis in airway myofibroblasts seeded in the presence of cycloheximide (Figure 2B), suggesting that these matrix components provide a significant survival signal for airway myofibroblasts. To substantiate that these cells were indeed dying by apoptosis, we confirmed cell membrane integrity using the trypan blue exclusion test (trypan blue negative cells > 90% for all conditions 4 h after seeding) and confirmed apoptosis by a second assay, namely TUNEL (Figure 2C).

View larger version (25K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   View larger version (14K): [in this window] [in a new window]   Figure 2.   Effect of ECM proteins on background rates of apoptosis. Serum-starved cells were plated into wells pre-coated with ECM factors or HEMA as shown and analyzed after 18 h (n = 3). *P < 0.01 compared with `uncoated' by ANOVA (A). Apoptotic cells were analyzed by PI counting and expressed as % ± SEM of all cells counted. (B) Cells were incubated with 50 µM CHX prior to seeding and analyzed as above. (C) Samples were seeded serum-free ± pre-incubated with 50 µM CHX. DNA fragmentation was quantified using Titer Tacs kit. For Unlabeled cells: the TdT labeling enzyme was excluded. Positive control was created by addition of the kit's endonuclease enzyme to SFM cells.

Integrin Subunits on Airway Myofibroblasts

Having defined the probable role for matrix in survival signaling in airway myofibroblasts, we set out to examine the important integrin receptors mediating this response. Flow cytometry confirmed that ₅, _v, and ₁ integrin subunits are universally expressed and that 40.9 ± 18.4% of cells were 6-positive. Less than a third of cells have detectable surface ₁, ₃, and ₄ (Figure 3). We recognize that our results may underestimate expression levels of some integrins, as it is conceivable that the trypsin used to remove the cells from the culture flasks may have cleaved off antibody recognition sites. In an attempt to minimize this potential problem, exposure times of trypsin were kept as short as possible throughout.

View larger version (15K): [in this window] [in a new window]   Figure 3.   Integrin expression by FACS. Cells were labeled with anti-integrin antibodies as shown. neg. control, mouse and rat anti-IgG antibody; pos. control, mouse and rat anti-2 microglobulin antibody. Samples labeled with only the secondary antibodies were analyzed as further controls: expression levels were similar to those seen with `neg. control' (results not included in figure for clarity). Results shown are mean from three separate donors.

Integrins Involved in Survival Signal Transduction

As expected, only universally expressed integrins were essential for human airway smooth muscle cell survival. Blocking binding of matrix factors to ₅ increased apoptosis on all ECM substrata (P < 0.01 versus control antibody, n = 3). A functionally blocking antibody to ₁ had a similar effect, but this increase in apoptosis only reached statistical significance when cells were seeded on collagens IV and V, laminin, and fibronectin (P < 0.01 versus control antibody, n = 3) (Figure 4). This increase in cell death, caused by incubating cells with the ₅ antibody prior to seeding, was concentration dependent. In keeping with an integrin-mediated ECM survival signal, pre-incubating the cells with RGD peptides, but not their negative control Arg-Ala-Asp (RAD), leads to increased cell death in a concentration-dependent manner, although the magnitude of the effect was somewhat smaller than that seen with the functionally blocking antibodies. At concentrations of 100 µg/ml, the blocking peptides Arg-Gly-Asp-Ser (RGDS) and Arg-Gly-Asp-Thr-Pro (RGDTP) increased cell death from a control level of 5.22 ± 0.4% to 11.0 ± 0.9% and 11.9 ± 1.2%, respectively (P < 0.01, n = 3). Their addition to confluent monolayers has no effect (Figure 5). To confirm the role of RGD in transducing a survival signal, we seeded cells that had been incubated with 50 µM cycloheximide into wells coated with 100 µM RGDTP peptides. RGDTP rescued cells from a baseline of 54.8 ± 1.0% to 32.9 ± 1.9% apoptotic cells (P < 0.0001, n = 3; data not shown). This result is in keeping with an ₅₁-mediated survival signal, though ₃₁, _v1, _IIbIIIa, _v3, _v5, and _v6 are also known to be RGD sensitive.

View larger version (15K): [in this window] [in a new window]   Figure 4.   Effect on apoptosis. Cells incubated with functionally blocking antibodies prior to seeding onto various ECM proteins as shown. Serum-starved cells were incubated with anti-integrin antibodies at 50 µg/ml and plated onto wells coated with ECM factors as shown. C, control antibody 2 microglobulin at same concentration. * denotes significant increase in apoptosis at P < 0.01 by ANOVA (n = 3) of condition versus C for individual matrix factor.

View larger version (17K): [in this window] [in a new window]   Figure 5.   Effect of incubating cells with RGD peptides prior to seeding. Serum-starved cells were incubated with peptides containing the RGD sequence known to block integrin binding site or with the negative control peptide containing RAD sequence at concentrations shown prior to seeding. After 18 h cells were fixed and stained with PI; apoptotic cells are expressed as % ± SEM of all cells counted. * denotes significant increase in apoptosis above corresponding control (P < 0.01 by ANOVA, n = 3).

Cell Morphology

Hirst and coworkers (11) have previously noted effects of different matrix factors on growth patterns and differentiation of myofibroblasts. We went on to study the morphologic appearance and intermediate filament formation of these cells. Myofibroblasts seeded on collagen IV or fibronectin retained their normal growth pattern, cytoskeleton formation, and spreading. In comparison, cells plated onto collagens I and V and laminin appeared less polar with diminished ability to spread. In addition, cells grown on collagen V displayed decreased intermediate filament and defective cell-to-cell contact formation, whereas collagen I seemed to promote a dense cytoskeletal network. Myofibroblasts seeded onto elastin remained rounded, with only rudimentary intermediate filaments detectable (Figure 6).

View larger version (27K): [in this window] [in a new window]   Figure 6.   Airway myofibroblasts grown on different ECM proteins. Cells were grown on different matrix factors as shown. After 18 h the cells were fixed; intermediate filaments were stained with FITC-labeled phalloidin (green), and nucleic acids counterstained with PI (red). Cells depicted are taken from representative areas for each ECM factor.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Airway remodeling is a feature of chronic asthma and may be a marker of disease severity (22); hallmarks of these structural changes observed include an increase in both ECM deposition and airway SMC number. In this study we have described novel aspects of the matrix-airway SMC interaction and highlighted key survival signals to which airway myofibroblasts respond. Background rates of apoptosis, even following prolonged serum starvation are low, implying the presence of a strong survival signal. In confluent monolayers this was not dependent on continued protein synthesis, as the addition of cycloheximide did not affect apoptosis. However, much higher rates of apoptosis were observed following seeding of cells in the presence of cycloheximide, indicating the likely requirement for synthesis of survival factor(s) at the time that cell-matrix interactions are formed. We demonstrated that known airway smooth muscle mitogens and growth factors could not prevent apoptosis when protein synthesis was inhibited, but that serum could. One explanation for this is that, in addition to known mitogens, serum contains soluble matrix components, which could provide the necessary survival signal. Our data confirm that adding ECM in the form of fibronectin, collagens I and IV, and laminin, or providing a substratum of the integrin-binding peptide RGD to newly seeded cells that were unable to synthesize their own matrix factors, provided a critical survival signal. This appears to be mediated at least in part through the fibronectin receptor/₅₁ integrin, as functionally blocking antibodies directed against these integrin subunits were able to inhibit survival signaling, as did blocking a fibronectin recognition site with soluble RGD peptides.

Integrins also play a pivotal role in cell adhesion to ECM (12). One would therefore predict that either preventing the laying down of matrix (e.g., with cycloheximide or HEMA) or interfering with integrin binding (e.g., by blocking adhesion sites with antibodies or RGD peptides) would impede cell adhesion. Even though we did not formally test for adhesion, cells were attached to the substratum provided in all experiments. These attachments may have been formed by non-specific binding or adhesion to residual matrix factors present, via `unblocked' integrins or by integrin recognition sites other than RGD (the ₅₁ integrin, for instance, recognizes LDV and PHSRN sites in addition to the RGD sequence present on ECM proteins) (23). As the matrix factors tested contain several different integrin recognition sites which have a predilection for specific integrin heterodimers, this may explain some of the differences between the signaling properties of individual matrix proteins and the apparent discrepancy between adhesion and survival signaling. Therefore, even under circumstances where survival was not affected, matrix components were able to drastically affect cell morphology, including formation of the cell cytoskeleton. This is likely to not only affect the cell's ability to contract, but also to have implications for cell metabolism, migration, secretion, and proliferation. The marked plasticity of airway SMCs and their capacity to exhibit diverse contractile, proliferative, and pharmacologic properties has been widely reported (24). It is therefore probable that cells exposed to altered concentrations of matrix factors in vivo would display aberrant responses to mitogens and contractile agonists, thus contributing to the abnormal responsiveness and remodeling of the asthmatic airway.

The strong survival signal provided by matrix to airway myofibroblasts is a novel and interesting observation. The requirement for cell-matrix interactions by anchorage-dependent cells for cell survival was first described in epithelial cells (21), and the importance of this finding was emphasized by Meredith's report that endothelial cells need to synthesize ECM to avoid programmed cell death when they are seeded onto plastic (25). Since then it has become clear that the survival effect conveyed by different matrix factors is both cell-type- and integrin-specific. Even though several integrin heterodimers may bind to the same matrix factor (e.g., ₃1, ₄1, _51, and _v1 are all expressed on airway myofibroblasts and are all known to bind fibronectin), only specific integrins are involved in survival signaling. Our data suggest that the predominant survival signal in airway SMC is mediated through ₅₁, although the functional blocking antibodies directed against these integrin subtypes did not induce cell death to the same extent as the protein synthesis inhibitor. This may be due to technical factors (e.g., incomplete antibody binding under these experimental conditions) or redundancy in survival signal transductionpromiscuity in the matrix-integrin interaction has been described previously (23) and may include redundancy involving the matrix factors, recognition sites, or integrin heterodimers. A predominant fibronectin and ₅₁-mediated survival signal was initially described in CHO cells (26), but its effect on myofibroblasts has not been previously noted. Even though the relevance of these signaling events has not been demonstrated in vivo, their significance is unlikely to be restricted to a culture environment: aberrant integrin-mediated signaling and adhesion-independent survival has been linked to tumor onset, progression, and metastatic dissemination (23). The fact that fibronectin has been shown to be increased in asthmatic airways (2) adds weight to our hypothesis that a fibronectin-₅₁- mediated survival effect for SMC may be relevant in the pathogenesis of airway remodeling.

The integrin-mediated effect of the ECM on myofibroblast cell number may be a direct effect on the cell cycle machinery (e.g., via phosphatidyl-inositol-3-kinase, NF-B [27], p53 [28], or bcl-2 [26]), or an indirect result. Presence of one matrix factor may inhibit deposition of another, and as integrins are known to mediate growth factor receptor phosphorylation (29), they are likely to play a role in modulating tyrosine kinase-mediated signaling events. The most important conclusion of this study is that airway myo- fibroblast survival may play an important role in the regulation of airway remodeling in chronic inflammation. Whereas airway smooth muscle mass may increase following increased levels of mitogenic stimuli such as thrombin and PDGF in the airways, it is also likely that the altered and increased matrix in the airways acts synergistically to maintain this process by providing a strong survival signal for airway myofibroblasts and by facilitating growth factor signaling. This combination of marked proliferation signal with the presence of a strong survival signal results in the increase in airway smooth muscle mass, which contributes to the anatomic narrowing of the airways and irreversible airflow obstruction seen in chronic asthma.

There is currently little evidence that existing treatment regimens control this response. Despite the fact that glucocorticoids can inhibit proliferation of airway smooth muscle in vitro (32), adding beclomethasone to human airway SMC in culture does not reverse the matrix-inducing effect of asthmatic serum (10). Likewise, even though Cys-LT1 receptor antagonists have anti-proliferative actions in vitro, they have no effect on matrix production (33). Phosphodiesterase type 3 inhibitors (34) and ₂ agonists such as salbutamol (35) have antiproliferative effects on cultured airway SMC but their action has not been examined in relation to matrix production or in a clinical setting. At present, there are few data to support the idea that controlling airway inflammation with steroids results in less remodeling in vivo. Yet interference with the vicious circle of matrix production and matrix-mediated survival signaling provides a potential therapeutic avenue for novel treatments in chronic asthma.

In summary, therefore, we describe here the key survival signal in human airway myofibroblasts. It is likely that the strong survival signal provided to these cells by their extracellular environment is in part responsible for the increase in airway smooth muscle mass seen in chronic asthma. This increased cell number in the context of inflamed airways in turn probably contributes to the continuing cycle of matrix deposition and cell proliferation, thus playing a part in causing irreversible airflow obstruction, which develops in persistent asthma.