INTRODUCTION

TOP ABSTRACT INTRODUCTION STRESSES IN THE AIRWAY... CELLULAR RESPONSES TO... COOPERATIVE RESPONSE TO... CONCLUSIONS REFERENCES

Keywords: airway remodeling; asthma; mechanotransduction

A growing body of evidence demonstrates that physical forces elicit a number of biologically relevant signals in the human body. Although the best-characterized responses to mechanical stimuli are in the cardiovascular (1, 2) and musculoskeletal (3) systems, reviews have highlighted our rapidly evolving understanding of how mechanical stimuli participate in the regulation of pulmonary physiology and pathophysiology (4). The airway wall exists in a mechanically dynamic environment, with minute amounts of circumferential and longitudinal expansion and contraction accompanying normal breathing movements (7). In asthma, the activation of smooth muscle exposes the airway wall to dramatic deformations and large stresses (8).

One of the most consistent findings in cells exposed to a variety of mechanical forces is the stimulation of growth (1, 4, 5, 9), and modulation of the extracellular matrix (4, 5, 10). Remodeling of the airway wall is characterized by airway wall thickening, subepithelial fibrosis, increased myocyte muscle mass, myofibroblast hyperplasia, and mucous metaplasia (reviewed in References [11] and [12]). Hence, it is plausible that the mechanical stresses that accompany bronchoconstriction may participate in the thickening and remodeling of the airway wall in subjects with asthma.

	STRESSES IN THE AIRWAY WALL

TOP ABSTRACT INTRODUCTION STRESSES IN THE AIRWAY... CELLULAR RESPONSES TO... COOPERATIVE RESPONSE TO... CONCLUSIONS REFERENCES

Activation of airway smooth muscle results most obviously in increased active loading of the smooth muscle cells that surround the airways. Consequent to this constriction, the cellular constituents internal to the smooth muscle layer, mostly epithelial cells, and a small number of fibroblasts or myofibroblasts, are exposed to a passive mechanical stress that is predominantly compressive (8). This compressive stress can be estimated by modeling the airway as a thick-walled cylinder under external pressure, as in Ressler and coworkers (13):  = (R/t + 1)P_sm, where  is the mean circumferential wall stress, R is the inside radius of the cylinder, t is the cylinder thickness, and P_sm is the external pressure due to smooth muscle constriction. Assuming maximally activated airway smooth muscle can generate an intralumenal pressure of 30 cm H₂O (14), the circumferential compressive stress for a thick-walled cylinder such as the airway is ~ 45 cm H₂O.

Because of the folding or buckling of the airway wall into a rosette pattern under smooth muscle loading, the actual stresses in the airway wall are much more complex. Mucosal folding has been observed for many years (15), and has led to a number of investigations of the consequences of the mucosal folding pattern on airway mechanics. Lambert has demonstrated that the mucosal folding pattern exerts a dominant effect on lumenal obstruction for a given degree of smooth muscle constriction (16), and Wiggs and coworkers (8), Seow and coworkers (17), and Lambert and coworkers (18) have all proposed explanations as to the factors determining the mucosal folding pattern. Although the mechanism(s) responsible for mucosal folding and the exact geometry of the folded airway remain open questions, the result of this folding is that the cells lining the inside of the airway experience a highly nonuniform stress field, with areas of compressive, tensile, and shear stress. The characterization of this stress field is highly complex, and is therefore best approached by using numerical methods. Wiggs and coworkers (8) have utilized the finite element method to model the airway wall as a two-layer cylinder, with a thin, stiff inner layer representing the epithelium and the subepithelial collagen layer, and a thicker, less stiff layer representing the interstitium that separates the epithelium from the smooth muscle layer. Constriction of the outer perimeter of the model airway wall resulted in buckling of the structure, similar to in situ observations. The pattern of buckling was most strongly dependent on the thickness of the inner layer, suggesting that subepithelial fibrosis may strongly influence the airway folding pattern, and thus determine the degree of lumenal obstruction. Large compressive stresses (> 10 times the applied stress) developed throughout the thin inner layer of the model, and extended into the thick interstitial layer under crevices of lumenal folds. In contrast, somewhat smaller tensile stresses (about two times the applied stress) developed at the apex of folds in the thin inner layer, and in the infolded areas of the underlying interstitial layer. Although the model relies on basic assumptions of airway wall geometry and constituent mechanical properties, it demonstrates the complexity of the loading pattern in the constricted airway wall, and the potential for smooth muscle loading to be amplified at the inner airway surface by the folding pattern and relative mechanical properties of the inner and outer wall compartments.

	CELLULAR RESPONSES TO MECHANICAL STRESS

TOP ABSTRACT INTRODUCTION STRESSES IN THE AIRWAY... CELLULAR RESPONSES TO... COOPERATIVE RESPONSE TO... CONCLUSIONS REFERENCES

Given the highly dynamic mechanical environment of the airway wall, there have been surprisingly few reports examining the responses of airway cells to mechanical stimuli. Although the stress field in the airway interstitium is primarily compressive (8), little is known about the behavior of lung fibroblasts under compressive loading. Lung fibroblasts have been cultured on elastic membranes and exposed to various patterns and durations of cyclic stretching, resulting in direct regulation of gene expression (19, 20), and increased cell proliferation mediated by autocrine or paracrine factors (21). Although these results are most relevant to airway distention during breathing and mechanical ventilation, stretch-mediated growth and gene expression in fibroblasts may also be important in areas of the constricted airway that experience tensile stress due to buckling (8). Additional studies are required to establish the response of fibroblasts to compressive stress, especially in the context of the three-dimensional interstitial matrix in which fibroblasts are typically embedded.

The effects of bronchoconstriction on smooth muscle physiology are much more difficult to simulate and interpret. In the airway, smooth muscle stress is actively generated. At this time, only passive (externally generated) loading of smooth muscle cells has been used to study cellular responses to mechanical stimuli in vitro. To study the effects of passive loading, airway smooth muscle cells have been cultured on collagen-coated silastic membranes and exposed to long-term cyclic uniaxial stretch (22, 23). Cyclic stretch increased smooth muscle proliferation and total cellular protein count (22), and increased the proportion of contractile proteins such as myosin, myosin light chain kinase, and desmin relative to unstretched controls (23), suggesting that stretch promoted a contractile rather than synthetic phenotype. Passive stretch of smooth muscle cells is most relevant to the deformations that accompany breathing and mechanical ventilation. However, the results demonstrate the responsiveness of airway smooth muscle to deformation, and indicate that constriction of airway smooth muscle may serve as a stimulus directly relevant to airway remodeling.

Modeling of the airway wall stress field under simulated constriction indicates that the airway epithelial lining of the lung will experience large, predominantly compressive stresses (8). Airway epithelial cells are far more abundant than fibroblasts, possess the capacity to modulate the inflammatory environment of the airway wall, and produce factors that influence the recruitment, proliferation, and activity of fibroblasts and smooth muscle cells (24). For these reasons airway epithelial cells are likely to be a major responder to mechanical forces in the airway wall, and a dominant factor in controlling the airway wall response to applied loads.

Early reports indicated that epithelial cells isolated from rabbit trachea expressed stretch-activated channels (25), and that mechanical stimulation led to intra- and intercellular Ca²⁺ signaling and increased ciliary beat frequency (26, 27). Stretch of mucosal explants from rabbit trachea increased inositol 1,4,5-trisphosphate (28), providing a potential mechanism leading to Ca²⁺ signaling.

Detachment of human bronchial epithelial cells from their underlying substrate induced increased expression of interleukin 8 (IL-8), in a manner that appeared to depend on the deformations that accompany detachment (29). Furthermore, hyperosmotic stress leading to cell shrinkage also increased IL-8 expression in human bronchial epithelial cells, presumably through a volume-sensing mechanism (30). Both of these studies suggest a link between cell deformation and IL-8 expression, and raise the possibility that mechanical stimuli have a proinflammatory effect.

More direct evidence of a link between deformation and functional changes comes from monolayer cultures of primary cat airway epithelial cells, and a human airway epithelial cell line, both of which have been grown on elastic substrates and subjected to cyclic distention. Both cyclic elongation and cyclic compression significantly slowed repair of wounded monolayers in both cell types (31). Cyclic stretch downregulated synthesis of prostaglandin E₂, prostaglandin I₂, and thromboxane A₂ in both cell types by inactivating cyclo-oxygenase (7). The deformations used in these studies were chosen to mimic those associated with breathing and mechanical ventilation, and the dramatic reduction in prostanoid synthesis elicited by this stimulus provided the first clear evidence that mechanical stimuli could produce changes in molecules with widespread effects throughout the airway wall.

Given the highly compressive nature of the stresses that accompany bronchoconstriction (8), it seems likely that compressive stress is the dominant stimulus linking mechanical stress to airway remodeling in asthma. Developing a system to produce a well-controlled and repeatable cellular compressive stress, and combining it with cells that are maximally representative of the in vivo phenotype, requires a delicate balance between mechanical and phenotypic fidelity. In terms of phenotypic accuracy, complex models of the airway wall that include multiple cell types are under investigation (32), but it appears difficult to combine such a system with well-controlled mechanical stress in a reproducible manner. In terms of reproducing the mechanical environment of the airway epithelium, expansion and contraction of epithelial cells on elastic membranes appear capable of reproducing the major forces associated with vital capacity movements, but difficulties remain in applying a homogeneous stress field to the entire cell culture surface (31). More importantly, thus far only monolayer cultures that lack some or all of the differentiated characteristics of the in vivo airway epithelium have been successfully exposed to mechanical stimuli in such systems (7, 31).

Air-liquid interface cultures of primary human and rat airway epithelial cells faithfully reproduce the mucociliary, pseudostratified characteristics of the in vivo airway epithelium (35). Because the porous substrate on which the cells are cultured is only minimally elastic, the difficulty arises in providing a representative mechanical stimulus to cultures in this setup. This difficulty was overcome by Ressler and coworkers (13), who developed a transcellular pressure stimulus (Figure 1) that generates an apical-to-basal compressive stress comparable in magnitude to that identified in modeling studies of the constricted airway wall (8). Rat tracheal epithelial cells were found to increase expression (Figure 2) of early growth response 1 (Egr-1), endothelin 1 (Et-1), and transforming growth factor  (TGF-) in a time- and magnitude-dependent manner (13).

View larger version (89K): [in this window] [in a new window]   Figure 1.   Schematic of transcellular pressure device developed by Ressler and coworkers (13). Differentiated airway epithelial cells cultured on porous culture membranes at the air-liquid interface are exposed to compressive stress by application of an apical-to-basal pressure difference. Reprinted by permission from Reference 13.

View larger version (102K): [in this window] [in a new window]   Figure 2.   Genes upregulated by compression of rat airway epithelial cells, using the system described in Figure 1. Expression of Egr-1, Et-1, TGF-1, and the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) after 0-6 h with (+) or without (-) the application of 20 cm H2O transcellular pressure. Reprinted by permission from Reference 13.

Each of these genes is relevant to airway remodeling, and provides a potential link between epithelial compression and airway remodeling. Egr-1 is an immediate-early gene, and encodes a transcription factor with putative binding sites in the promoter region of a number of genes, including platelet-derived growth factor (PDGF), fibronectin, TGF-, and tumor necrosis factor  (36, 37). Endothelin 1, in addition to being a potent bronchoconstrictor, is proinflammatory, and profibrotic, providing a mitogenic signal for smooth muscle cells and fibroblasts, a chemotactic signal for fibroblasts, and promoting collagen synthesis by fibroblasts (reviewed in Michael and Markewitz [38]). TGF- is an important regulator of lung fibrosis, stimulates expression of fibronectin and collagen genes in fibroblasts (reviewed in Bonewald [39]), and induces a myofibroblast phenotype thought to be crucial for the development of pulmonary fibrosis (40).

The mechanism of transduction in this in vitro system remains unknown. Although a hydrostatic pressure increase and substrate stretch have been ruled out (13), potential mechanisms include extrusion of the basal cells into the pores of the culture substrate, cell shrinkage and initiation of cell volume regulation, and intercellular shear stress associated with fluid flow down the pressure gradient. Whatever the mechanism, it is instructive to compare the in vitro system with the in situ constricted airway. Both computational modeling and micrographs of constricted airways indicate that large cellular deformations occur at the base and apex of lumenal folds (8). Furthermore, the highly nonuniform stress field in the in vivo constricted airway likely produces pressure gradients capable of driving water across cell membranes and generating intercellular flows, leading to both cell volume regulation and surface shear stresses. Hence, mechanisms that operate in vitro are likely present in the intact airway.

	COOPERATIVE RESPONSE TO MECHANICAL STRESS

TOP ABSTRACT INTRODUCTION STRESSES IN THE AIRWAY... CELLULAR RESPONSES TO... COOPERATIVE RESPONSE TO... CONCLUSIONS REFERENCES

Although much can be learned from studying individual cellular responses to mechanical stimuli, airway remodeling is a complex phenomenon, requiring cooperation among multiple cell types. In vitro models that incorporate multiple cell types (32, 41) will likely provide new insights into the cellular and tissue responses to mechanical stimuli, and the critical factors linking both mechanical and inflammatory stimuli to the multitude of changes characterizing the remodeling asthmatic airway.

A novel coculture model has been developed by Zhang and coworkers (32), incorporating human bronchial epithelial cell line 16HBE14o- seeded onto three-dimensional collagen gels embedded with human bronchial myofibroblasts. This model reconstitutes the basic morphology of the airway wall, and features epithelial cells possessing tight junctions, cilia, and directional ion transport. Injury of the epithelium by application of poly-L-arginine, a surrogate for eosinophil cationic protein, resulted in increased myofibroblast proliferation. Conditioned medium from epithelium damaged by mechanical scraping also increased myofibroblast proliferation. It is not clear whether the response to cell scraping was due to release of mediators from damaged cells, mechanical stimulation of cells by the act of mechanical scraping, or viable cells undergoing a wound-healing response. Poly-L-arginine significantly increased soluble basic fibroblast growth factor (bFGF), Et-1, and PDGF, whereas mechanical scraping enhanced soluble bFGF, insulin-like growth factor I (IGF-I), PDGF, and latent TGF-2. Specific antibodies directed against these factors, either alone or in combination, were able to partially inhibit the proliferation of myofibroblasts in response to poly-L-arginine-induced epithelial damage.

Swartz and coworkers (41) pioneered an alternative approach that allows primary differentiated airway epithelial cells to be combined in coculture with fetal human fibroblasts. In this case the cells are not in physical contact, but are capable of bidirectional communication via soluble mediators. In this system, the epithelium is exposed to a transcellular compressive stress, as in the work of Ressler and coworkers (13), while the fibroblasts serve as a reporter system to demonstrate the integrated response to soluble mediators released by the epithelium in response to mechanical compression.

Fetal lung fibroblasts responded to mechanical stimulation of the epithelium by increasing collagen synthesis, especially of type III collagen, in a pressure-dependent manner. The epithelial cells contributed to the matrix-remodeling activity of the coculture system by increasing the ratio of matrix metalloproteinase 9 (MMP-9) to tissue inhibitor of metalloproteinase 1 (TIMP-1). These changes bear a striking resemblance to the remodeling asthmatic airway, which is characterized in part by subepithelial fibrosis, featuring prominent changes in collagen type III (42), and altered MMP-9/TIMP-1 ratio (45, 46).

In contrast to the findings of Zhang and coworkers (32), mechanical stimulation did not affect proliferation of fibroblasts. Culture conditions, however, exerted a dramatic effect on fibroblast proliferation. Fibroblasts cultured alone incorporated three times more [³H]thymidine than fibroblasts in coculture with epithelial cells. In contrast, conditioned medium collected from epithelial cells had the opposite effect on cocultured epithelial cells, promoting [³H]thymidine incorporation by fibroblasts compared with fibroblasts cultured alone. This is in agreement with multiple studies that demonstrate the growth-promoting effect of airway epithelial conditioned medium on fibroblasts (32, 47, 48). One explanation is that the growth inhibitory factor(s) released by epithelial cells is unstable in comparison with the stimulatory factor(s). In such a case, transfer of conditioned medium might allow degradation of inhibitory factors, tilting the balance from inhibitory to stimulatory. Alternatively, the bidirectional communication between cells in coculture could phenotypically alter the epithelial cells to produce a more inhibitory balance of growth signals than is produced by epithelial cells cultured alone. In either case, these findings demonstrate the complexity of the integrated response of cells to each other, and to external stimuli.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION STRESSES IN THE AIRWAY... CELLULAR RESPONSES TO... COOPERATIVE RESPONSE TO... CONCLUSIONS REFERENCES

Although it is generally accepted that airway remodeling is a consequence of the inflammatory milieu of the asthmatic airway, the presence of dramatic cellular stresses in the constricted airway, in combination with a responsive population of cells, raises the distinct possibility that the airway is a mechanically responsive tissue. The intriguing finding that mechanical stress applied to airway epithelial cells produces a profibrotic environment, in the absence of an inflammatory response, demonstrates the potential for mechanical stimuli to participate in airway remodeling (Figure 3). As the cellular stresses that accompany bronchoconstriction are better defined, and cell culture models improve, the physiological responses of the constituent cells of the airway wall to these stresses can be better defined. By combining multiple cell types, we can begin the process of integrating the multiple contributing cells and pathways that lead to remodeling. Finally, by combining mechanical stimuli with inflammatory mediators, or activated inflammatory cells (33, 49), we can further enhance our understanding of the interplay between mechanical and inflammatory contributions to airway remodeling.

View larger version (29K): [in this window] [in a new window]   Figure 3.   Links between inflammatory and mechanical stimuli for airway remodeling. In the direct inflammatory paradigm, mediators, cytokines, and growth factors elaborated by inflammatory cells are the signals responsible for airway wall remodeling. In the mechanotransduction paradigm, the signals for airway wall remodeling derive from the effects of mechanical stress on the airway epithelium. Note that the remodeled airway may be more susceptible to buckling (8), creating a potential positive feedback loop.