[beta]1-Integrins Mediate Enhancement of Airway Smooth Muscle Proliferation by Collagen and Fibronectin

Chronic asthma is characterized by poorly reversible airway obstruction and airway hyperresponsiveness (AHR) that is associated with intense persistent inflammation and structural remodeling of the airway wall (1, 2). In some instances, asthma severity is proportional to the extent of remodeling (3). Two prominent features of the remodeling process thought to contribute to the development of AHR include accumulation of airway smooth muscle (ASM), possibly from hyperplastic or hypertrophie changes (4, 5), and alterations in the amount and composition of extracellular matrix (ECM) proteins (6-10).

ECM proteins regulate diverse cellular functions, including migration, survival, maintenance of differentiated status, and enhanced growth (11). Histologic studies of asthmatic airways have revealed abnormalities in both the quantity and composition of ECM proteins. Collagen types I, III, and V; fibronectin, tenascin-C; hyaluronan; versican; and laminin a2/ß2 chains are increased, whereas other major ECM components, including collagen type IV and elastin, are decreased (6-10). Increased type I collagen, hyaluronan, and versican have been found localized within and surrounding ASM bundles from individuals with asthma (10, 12). In the bronchoalveolar lavage (BAL) fluid of individuals with asthma, increased levels of fibronectin, hyaluronate, and laminin products are found (reflecting increased ECM turnover) that correlate with asthma severity (13). Likewise, vitronectin levels are increased in BAL fluid in other inflammatory lung diseases (14). Collectively, these observations support marked ECM changes in the airway wall in asthma and lung inflammation. However, little is known of the effect on airway cell function of this enrichment of the airway wall with collagen and fibronectin, and the critical mediators and cellular mechanisms that bring about changes in ASM content in vivo are undefined.

Previously, we have reported that both fibronectin and type I collagen enhance responses of cultured human ASM cells to mitogens, such as thrombin and platelet-derived growth factor (PDGF)-BB (15). Bonacci and coworkers (16) report similar findings with fibroblast growth factor (FGF)-2-dependent proliferation of bovine tracheal smooth muscle cultured on type I collagen, and Freyer and coworkers (17) have shown that several ECM substrates, including type I collagen and fibronectin, increase survival of human ASM cells. Moreover, ASM cells cultured from individuals with asthma or grown on homologous ECM from asthmatic ASM cells proliferate more rapidly and secrete increased ECM proteins, including type I collagen (18, 19). Together, these observations affirm that ASM in situ is likely subjected to the influences of signaling through ECM/integrin interactions, possibly involving autocrine mechanisms, and have led to the suggestion that alterations in the airway wall ECM microenvironment in remodeling in asthma may favor enhanced growth responses of ASM (15-17). Integrins, which comprise aß heterodimers, are the principal receptors mediating multiple cell responses to ECM substrates (20). However, there is no direct information on the relationship between enhanced growth and expression of ß-integrins by human ASM.

To explore a role for ß1-integrins in the enhancement of PDGF-BB-dependent proliferation by selected ECM substrates known to be increased in the airways of people with asthma, we examined the effect of PDGF-BB on expression of a- and ß-integrin subunits on ASM cells grown on plastic or on type I collagen or fibronectin. In addition, using integrin subunit-specific and helerodimer-specific function-blocking antibodies, we sought to identify integrins that were required for enhancement of PDGF-BB-dependent proliferation by these ECM substrates. Some of the results of these studies have been previously reported in the form of abstracts (21, 22).

METHODS

Isolation and Culture of Human ASM Cells

Human ASM cells were obtained in accordance with procedures approved by the Guy's and St. Thomas' Hospitals' Research Ethics Committee (Study #01/047) from the lobar or main bronchus of 28 patients without asthma (mean age 60

Surface Coating with ECM Proteins

Lyophilized human plasma fibronectin (Sigma-Aldrich, Poole, UK), rat tail monomeric type I collagen (ICN Chemicals, Thame, UK), and human fibrillar collagen types I and III (Chemicon, Chandler's Ford, UK) were reconstituted in sterile phosphate-buffered saline (PBS). ECM proteins (0.1-10 µg/ml) diluted in PBS were adsorbed to tissue culture plastic ware overnight at 37°C as previously described (15). Excess, unbound ECM protein was removed by aspiration and washing with PBS. Unoccupied protein-binding sites were blocked by incubation with 0.1% bovine serum albumin for 30 minutes.

Cell Proliferation and Attachment: Application of Blocking Antibodies

Human ASM cells in Dulbecco's modified Eagle's medium containing 0.5% FBS were seeded (10,000 cells/cm^sup 2^) onto plastic or ECM substrate (0.1-10 µg/ml) precoated plastic ware. For proliferation studies, cells were left undisturbed for 8 hours at 37°C. Thereafter, the medium was replaced with FBS-free Dulbecco's modified Eagle's medium, supplemented with insulin (1 µM), transferrin (5 µg/ml), ascorbate (100 µM), and bovine serum albumin (0.1%). Growth was initiated after 72 hours by replacement with fresh FBS-free Dulbecco's modified Eagle's medium containing PDGF-BB (5 ng/ml) and subsequently every 48 hours until confluence. In some experiments, cells were pretreated for 30 minutes with integrin function-blocking monoclonal antibodies and corresponding isotype-matched control antibodies before and during stimulation with PDGF-BB. Integrin function-blocking anti-human mouse monoclonal antibodies (Chemicon, Chandler's Ford, UK, unless stated otherwise) were anti-a1 (clone FB12), anti-a2 (clone P1E6), anti-a3 (clone P1B5), anti-a1 (clone P1H4), anti-a5 (clone P1D6), anti-av (clone L230, gift from Dr. John F. Marshall [23]), anti-ß1 (clone 6S6), anti-avß3 (clone LM609), and anti-ß4 (clone ASC-3). Mouse purified or isotype-matched IgG (Chemicon) was used as a nonimmune control.

Proliferation of human ASM was assessed on Day 9 by milochondrial-dependent reduction of MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) to formazan, described in detail elsewhere (15). Cell attachment was assessed at 60 minutes by methylene blue dye uptake using the protocol described by Oliver and coworkers (24).

Flow Cytometric Labeling of Surface Integrins

Cell surface integrins were localized by binding of the previously mentioned anti-human integrin subunit monoclonal antibodies. Near-confluent unstimulated or PDGF-BB-stimulated cells on plastic or ECM substrate-coated plastic ware were harvested using trypsin/ EDTA, washed twice in PBS (200 × g for 5 minutes), resuspended, and fixed with 200 µl 4% formaldehyde (methanol-free, EM-grade, Polysciences, Warrington, PA) for 30 minute on ice in round-bottom fluorescence-activated cell sorter tubes. After centrifugation, fixed cells (30,000/tube) were resuspended in PBS containing 3% FBS for 30 minutes to prevent nonspecific antibody binding. Resuspended cells on ice were then incubated with anti-a-or anti-ß-integrin subunit antibodies (1 µg/ml for 30 minutes), washed in PBS, and then incubated for 30 minutes in the dark with an anti-mouse, phycoerythrin (PE)-conjugated secondary antibody (Sigma-Aldrich, Poole, UK). Numbers of PE-labeled cells were determined on a FACS Calibur flow cytometer (Becton Dickinson, Oxford, UK) using CellQuest Pro software (Becton Dickinson) (15). After analysis, cells were confirmed microscopically to be intact.

Statistical Analysis

Data are expressed as mean ± SEM of duplicate or triplicate observations obtained from cells cultured from n patient donors. Data were compared using one- or two-way analysis of variance (ANOVA), where appropriate, followed by Bonferroni's t test post hoc to evaluate statistical differences between treatment groups (SigmaStat; SPSS Inc., Chicago, IL). A p value of less than 0.05 was considered significant.

RESULTS

Attachment and Proliferation of Human ASM Cells on ECM Substrates

Initial experiments centered on the kinetics of human ASM cell attachment to several ECM substrates that are elevated in the airways of individuals with asthma. Adherence to monomeric or fibrillar type I collagen, fibronectin, fibrillar type III collagen, vitronectin, or tenascin-C (all 10 µg/ml) occurred within 10 minutes of seeding and, by 4 hours, 90% of all cells were attached (Figure 1). At time points exceeding 2 hours, attachment to ECM substrates did not differ from that of plastic (p > 0.05 by two-way ANOVA, n = 4). Maximal attachment to all substrates occurred at 8 hours. Accordingly, in subsequent studies to examine modulation of PDGF-BB-dependent proliferation by these ECM substrates, cells were allowed 8 hours to attach.

After stimulation with PDGF-BB (5 ng/ml), human ASM cells in plates precoated with monomeric but not fibrillar type I collagen were significantly more proliferative than PDGF-BB-stimulated cells on plastic (p < 0.01-0.001, n = 6; Figure 2A). A similar concentration-dependent increase in PDGF-BB-dependent proliferative capacity occurred with fibronectin (p < 0.01-0.001, n = 6; Figure 2B), confirming previous findings (15). Enhancement of PDGF-BB-stimulated proliferation by vitronectin was variable and less marked (p < 0.01 by two-way ANOVA, compared with collagen type I or fibronectin), reaching significance (p < 0.05 compared with plastic, n = 6) at 30 µg/ml only (Figure 2B). The effect of coating with monomeric type I collagen and fibronectin (3-30 µg/ml) in combination was additive (p < 0.01 by two-way ANOVA, n = 4) on potentiation of PDGF-BB-dependent proliferation compared with either substrate alone, although synergy may have occurred at 1 µg/ml-the lowest concentration examined (see Figure El in the online supplement). In contrast, PDGF-BB-stimulated cells on type III collagen or tenascin-C were no more proliferative than responses on plastic (p > 0.05, n = 6; Figure 2C). In all experiments, proliferation of human ASM cells cultured on these ECM substrates (1-30 µg/ml) in the absence of PDGF-BB did not differ from unstimulated cells on plastic (p > 0.05, n = 6 for each ECM) (not shown).

Regulation of Human ASM Cell Integrin Expression by PDGF-BB

Previous reports suggest PDGF-BB can upregulate integrin expression on various cell types (25), which could account for the potentiation of PDGF-BB-dependent proliferation of human ASM cells by monomeric type I collagen or fibronectin. Accordingly, expression of multiple integrins was compared in unstimulated or PDGF-BB-stimulated human ASM cells cultured on plastic, monomeric type I collagen, or fibronectin. Flow cytometry confirmed that a5- and ß1-integrin subunits were universally expressed and that approximately 50 to 70% of unstimulated cells on plastic had detectable surface a1, a2, a3, and av integrin subunits and the avß3 heterodimer, and less than 30% expressed the a4 integrin subunit (Figures 3A and 3B). With the exception of the a3 subunit, which was increased by about 20% (p < 0.05, n = 3), PDGF-BB had no effect on numbers of cells on plastic expressing these integrins (p > 0.05, n = 3) (Figure 3B). Likewise, culture on 10 µg/ml monomeric type I collagen or fibronectin did not alter this profile irrespective of treatment with PDGF-BB (p > 0.05, n = 3) (Figures 3C and 3D). Again, the proportion of cells on monomeric type I collagen expressing a3 integrin subunits was increased by approximately 20% after PDGF-BB stimulation (p < 0.05, Figure 3C). Similarly, culture alone on fibronectin increased numbers of a3 integrin-expressing cells by 20 to 25% (p < 0.05 by two-way ANOVA, n = 3). Mean fluorescence intensity of labeling for all integrins examined was unchanged (p > 0.05 by two-way ANOVA; not shown).

Attenuation of Human ASM Cell Attachment to Type I Collagen or Fibronectin by Integrin-specific Blocking Antibodies

Subsequent experiments employed function blocking monoclonal antibodies to examine a requirement for specific integrins mediating attachment and increased cell numbers on ECM substrates. Preliminary studies showed that inhibition of cell attachment to plates coated with 10 µg/ml monomeric type I collagen or fibronectin was maximal after 60 minutes with 10 µg/ml anti-ß1 integrin (Figure E2). This concentration and incubation period was used across the panel of blocking antibodies. Attachment to monomeric type I collagen or fibronectin was inhibited but not abolished by the ß1-integrin subunit-blocking antibody (p < 0.01 compared with control IgG, n = 4) (Figure 4). Similarly, blocking antibodies to a2-integrin subunits or the avß3 heterodimer reduced but did not abolish attachment to monomeric type I collagen (p < 0.05 compared with control IgG, n = 4). In contrast, attachment to fibronectin was inhibited by blocking of a5 integrin subunits (p < 0.01 compared with control IgG, n = 4). The combination of antibodies to multiple a integrin subunits (a1, a2, a3, a4, a5, and av) had no additional effect. However, inclusion of the anti-ß1 blocking antibody reduced attachment to type I collagen or fibronectin to about 15% in both cases (Figure E3).

In growth studies, pretreatment with the ß1-integrin blocking antibody (0.1-10 µg/ml) partially reduced enhancement of PDGF-BB-dependent proliferation by either monomeric type I collagen or fibronectin. Maximal attenuation of this effect occurred at 10 µg/ml antibody and was 48 ± 12% and 44 ± 10% against monomeric type I collagen or fibronectin, respectively (p < 0.05 compared with IgG, n = 5) (not shown). Potentiation of PDGF-BB-stimulated proliferation by either type I collagen or fibronectin was prevented by a2, a4, or a5 subunit-blocking antibodies in combination with anti-ß1 integrin (Figure 5). The effect was concentration-dependent, with 10 µg/ml of each antibody combination suppressing PDGF-BB-stimulated proliferation to levels found in uncoated plates (p < 0.001 compared with IgG, n = 5). In parallel experiments, combinations of a1, a3, or av blocking antibody (0.1-10 µg/ml) with the anti-ß1 integrin antibody had no effect on enhancement of PDGF-BB-dependent proliferation by either monomeric type I collagen or fibronectin (p > 0.05 compared with IgG, n = 5). Similarly, pretreatment with the avß3 heterodimer-specific antibody had no effect (Figure E4).

DISCUSSION

Airway remodeling is an established feature of asthma of varying severity in which enrichment of the airway wall with collagen and fibronectin and ASM accumulation are hallmarks. Although concentrations of various ECM substrates in contact with ASM in the asthmatic airway remain unknown, we and others have shown that type I collagen and fibronectin ECM substrates enhance ASM proliferative (15, 16) and survival (17) responses and have suggested that alterations in ECM in remodeling in asthma may favor accumulation of ASM in the intact airway wall. In the present study, we examined the effect on human ASM proliferation of several interstitial ECM substrates that are increased in the airways of individuals with asthma. We report that PDGF-BB-dependent proliferation was enhanced two- to threefold by monomeric type I collagen, fibronectin, and to a much lesser extent (1.2-fold) by vitronectin; fibrillar collagen type I, fibrillar type III collagen, and tenascin-C had no effect, suggesting enhanced proliferation by the ECM resulted from specific rather than nonspecific protein interactions. To address possible mechanisms, flow cytometry studies indicated that upregulation of cell surface integrin ECM receptors by PDGF-BB alone or in combination with monomeric collagen type I or fibronectin did not account for the enhanced proliferation. Function blocking monoclonal antibodies to individual integrin subunits or to the avß3 heterodimer revealed that attachment to monomeric type I collagen was mediated partially by a2ß1 and avß3 integrins expressed by ASM, whereas attachment to fibronectin was dependent on a5ß1. In contrast, enhancement of PDGF-BB-stimulated proliferation by either fibronectin or monomeric collagen type I involved multiple ß1 integrins including a2ß1, a4ß1, and a5ß1; combined blocking of a1ß1, a3ß1, or avß1; or blocking the avß3 heterodimer had no effect.

In addition to type I collagen and fibronectin, other interstitial ECM components that are prominent in remodeling in asthma in the proximity of the ASM bundles include type III collagen and tenascin-C (10). Little appears known of the effects of type III collagen on smooth muscle proliferative responses. Perhaps consistent with the paucity of direct information, we found no effect of type III collagen on PDGF-BB-dependent proliferation of human ASM cells. This was in contrast to parallel studies in which type III collagen enhanced interleukin-1ß-stimulated granulocyte-macrophage colony-stimulating factor (GM-CSF) release from these cells (Peng and Hirst, unpublished observation), suggesting these cells are capable of responding to this substrate under appropriate conditions, albeit via a nonproliferative pathway. We also examined growth regulation of human ASM cells by tenascin-C. In rat-cultured pulmonary artery smooth muscle cells, it amplifies mitogenic responses to FGF-2 and is a prerequisite for epidermal growth factor-dependent smooth muscle cell proliferation (26). Its effects appear mediated primarily by the avß3 vitronectin receptor (27), which was expressed by 40 to 60% of human ASM cells (Figure 3). Its lack of effect on PDGF-BB-dependent human ASM cell proliferation may reflect a species-, tissue-, or mitogen-specific difference, possibly reflecting a requirement for other tenascin-ligating integrins such as a8ß1 (28).

The mechanism that underlies enhancement of growth factor-induced human ASM cell proliferation by ECM components is unknown. Although upregulation of integrins by PDGF occurs on various cell types (25), our flow cytometry studies failed to detect this with either PDGF-BB alone or in combination with monomeric collagen type I (excepting a3 expressing cells; see below) or fibronectin, suggesting this could not account for the amplified proliferative signal. Recent data suggest growth factors bind ECM components to form complexes that enhance subsequent growth factor activity through integrin collaboration with growth factor receptors. In the case of VEGF interacting with fibronectin to enhance endothelial cell migration, the amplified response involves sustained ERK activation that requires activation of both the Flk-1 (VEGF receptor) and a5ß1 (major receptor for fibronectin) (29). Orsini and coworkers (30) have previously demonstrated that sustained ERK activation is required for human ASM proliferation. Whether monomeric collagen type I or fibronectin induce a sustained ERK signal that underlies amplified proliferative responses remains an open question. Consistent with this possibility, a recent study suggests enhancement of PDGF-dependent human ASM migration by type I collagen requires activation of Src kinase, which is an upstream regulator of ERK (31).

Enhancement of PDGF-BB-dependent proliferation by monomeric, but not fibrillar type I collagen, suggests only the proteolytic denatured form (short monomeric peptide fragments of depolymerized collagen) rather than the triple helical fibrillar configuration is sufficient to elicit proliferative signals in ASM. A requirement for collagen to be in a monomeric form for enhancement of growth factor-stimulated proliferation may have contributed to the lack of effect found with type III collagen, which was examined in fibrillar form only. Consistent with this possibility, human vascular smooth muscle cells on polymerized type I collagen fibrils in vitro are reported to be arrested in Gl of the cell cycle through upregulation of the cyclin-dependent kinase inhibitor p27^sup Kipl^, whereas monomeric type I collagen supports smooth muscle cell proliferation (32) and may be explained by utilization of differing integrin heterodimers. Cellular adhesion on type I collagen is mediated through a2ß1 (and a1ß1) integrins; interaction with these integrins requires collagen to be in its native helical configuration (33). However, proteolytically cleaved, denatured collagen also recognizes avß3 integrin (34). Consistent with this, we observed a partial reduction in attachment of human ASM cells to monomeric type I collagen with both a2 integrin subunit and avß3 heterodimer-specific blocking antibodies.

As mentioned, the flow cytometric integrin expression findings provide new information that about 40 to 60% of human ASM cells express the promiscuous vitronectin receptor, avß3, and 60-80% of cells express a2 subunits, which are required for type I collagen binding (33). These data confirm an earlier report showing similar expression patterns on human ASM of multiple ß1 integrin-associating subunits, including a1, a3, a4, and av, and is consistent with a5 and ß1 subunits being universally expressed (17). Integrin function-blocking studies suggested that enhancement of PDGF-BB-stimulated proliferation by monomeric type I collagen or fibronectin involved multiple integrins, including a2ß1 (major receptor for type T collagen), a4ß1 (major counterreceptor for vascular cell adhesion molecule-1, osteopontins, and thrombospondins, but also binds fibronectin [35]), and a5ß1 (major fibronectin receptor). Of note, although numbers of a3-expressing cells were increased by stimulation with PDGF-BB or culture on monomeric type I collagen or fibronectin, neutralization of a3ß1 (major laminin receptor with weak affinity for type I collagen and fibronectin) did not suppress ECM-enhanced proliferation, suggesting it was not required for accelerated growth under these conditions. Furthermore, the finding that, in each case, blocking antibodies abolished enhancement of proliferation by type I collagen or fibronectin irrespective of whether the target integrins were universally (a5, ß1), intermediately (a2), or poorly expressed (a4) suggests that increased cell numbers could involve transient expression of specific integrin subunits followed by selective expansion of a subpopulation of human ASM cells. This possibility is suggested particularly by the finding that enhancement of PDGF-BB-dependent proliferation by fibronectin was abolished by neutralization of a4 and ß1 when less than 20% of human ASM cells were found to express a4 subunits on fibronectin with or without PDGF-BB stimulation. However, no evidence of this expansion was found by flow cytometry in PDGF-BB-stimulated cells on fibronectin or type I collagen. Alternatively, abrogation of enhanced proliferation on these ECM substrates by a4 and ß1 neutralization could involve aberrant cross-reactivity of the anti-a4 antibody for more widely or universally expressed integrins. Again, we found no supporting evidence from the flow cytometric or attachment studies and so the question over the unexplained efficacy of the a4 and ß1 combination remains open.

Blocking of fibronectin-induced potentiation of PDGF-BB-dependent proliferation by a2 and ß1 neutralization was also unexpected because a2ß1 is the major type I collagen receptor; nor did we expect to find that a5ß1 neutralization prevented enhanced proliferation on monomeric type I collagen, because a5ß1 is a major fibronectin receptor. Evidence supports a5ß1 ligating monomeric denatured collagen substrates (36), and collagen-binding regions are present in purified fibronectin (37). Recent findings in a cell-free system show that reconstituted a2ß1 heterodimers ligate both type I collagen and fibronectin (38). Again, we found no evidence of such cross-reactivity in our attachment assays after integrin neutralization. Another explanation for differences in integrin requirements between attachment and proliferation is modification of the initial ECM coating by autocrine ECM protein secretion, particularly because the proliferation studies were performed over several days rather than minutes with attachment. Though not examined, during this period, human ASM cells secrete an ECM network of mixed composition (19), which may or may not depend on the initial coating conditions. Consistent with this possibility, Icchi and coworkers (39) report that human vascular smooth muscle cells grown on monomeric but not fibrillar type I collagen markedly upregulated autocrine expression of ECM proteins including fibronectin, type I collagen, and the major a4ß1 ligand, thrombospondin-1. It remains to be established whether autocrine type I collagen or fibronectin contributes to the cross-reactivity observed or if a similar profile of blocking exists in cells in which autocrine ECM secretion is prevented.

In conclusion, as with other cell types, the biological responsiveness of ASM reported here is associated with fibronectin and with the short monomeric peptide fragments of depolymerized type I collagen acting via multiple ß1 integrins (a2ß1, a4ß1, or a5ß1). Such fragments can result in the airways in asthma from increased matrix metalloproteinase activity during episodes of allergic inflammation (40-42), and this could directly influence ASM responses in the intact airway to peptide mitogens, possibly contributing to the remodeling process. Targeting multiple ß1 integrins may be a useful pharmacologic approach to control pathologic growth of ASM in asthma.