Chronic asthma may be accompanied by an enhanced rate of decline in lung function irrespective of anti-inflammatory treatment. These clinical observations have been linked to structural changes in the asthmatic lung termed airway remodeling [30–32]. The pathogenesis of airway remodeling has been previously attributed to reactivation of the epithelial-mesenchymal trophic unit in which increased levels of TGFβ1 contribute to a state where hypo-proliferative but activated epithelial cells induce activation of fibroblasts to myofibroblasts [33–35]. Although TGFβ1 functions as a master switch in tissue repair and wound healing, there is substantial evidence that disordered expression of TGFβ1 may lead to fibrosis [15, 36, 37]. Clinical studies indeed confirm evidence for epithelial shedding and damage in the asthmatic airway, along with elevated levels of TGFβ1 in asthmatic bronchoalveolar lavage fluid and airway tissue [21, 24]. While not all studies have found elevated TGFβ1 levels in the airways of asthmatic subjects [38, 39], the bulk of evidence suggests that chronic asthmatic inflammation is accompanied by increased activity of TGFβ1 in the airways [21–23].
By virtue of their synthetic and contractile phenotype, myofibroblasts are considered to be a key cell type responsible for the excessive extracellular membrane protein deposition and increase in smooth muscle mass associated with remodeled airways [36, 40]. The origin of the lung myofibroblast, however, is still unclear. An unknown percentage of lung myofibroblasts derive from activation of tissue fibroblasts or homing of blood-borne fibrocytes [11, 41]. In addition, there is emerging evidence in kidney fibrosis and IPF that TGFβ1-driven EMT of tubular interstitial epithelial cells and alveolar epithelial cells may represent a significant source of tissue myofibroblasts [14, 15, 42, 43]. TGFβ1 has previously been shown to induce EMT in the alveolar-type cancer cell line, A549 . In addition, in vivo studies have suggested that EMT may occur in IPF as well as in alveolar and bronchial epithelial cells during bleomycin-induced pulmonary fibrosis [15, 44, 45].
Despite the accumulating evidence that EMT contributes to fibrotic remodeling in several organs including the lungs, there is little evidence that EMT occurs in bronchial epithelial cells and no evidence that it plays a role in the airway remodeling that accompanies chronic asthma. We hypothesized that exposure of normal bronchial epithelial cells to chronic TGFβ1 stimulation would cause them to undergo EMT, potentially representing another source of myofibroblasts involved in airway remodeling in asthma. Here we report that BEAS-2B as well as primary normal human bronchial epithelial cells show evidence of EMT upon prolonged in vitro stimulation with TGFβ1.
TGFβ1-induced downregulation of the epithelial cell specific adherence junction protein E-cadherin at both the mRNA and protein levels was the earliest effect we observed, reaching near-maximal effect within 24 hours of stimulation in BEAS-2B cells. The loss of cell-cell contact has been shown to be a crucial first event in the remodeling process in the kidney [17, 46]. Masszi  et al. further reported that the disruption of cell-cell contact is a critical regulator for TGFβ1 induced EMT in kidney cells. They suggest a two-hit mechanism in which both TGFβ1 stimulation as well as initial epithelial injury are required for the induction of EMT. This correlates with the observation that in the asthmatic airway the integrity of the epithelial layer is disrupted, which might therefore facilitate the fibrogenic action of TGFβ1. Further it has been demonstrated that β-catenin, released from the cytosolic portion of E-cadherin, can function as a transcription factor in concert with the lymphoid enhancing factor 1 (LEF1) and induces EMT in epithelial cell lines [47–49].
Myofibroblasts release a variety of ECM proteins contributing to the thickening of the lamina reticularis, a key feature in the remodeling process of the lung. We found that TGFβ1 stimulates increased expression of extracellular matrix proteins (fibronectin, collagen I and tenascin C) in BEAS-2B cells. Stable expression of a myofibroblast phenotype in renal epithelial cells has been shown to depend on both TGFβ1 and adherence signals [13, 46, 50, 51]. In this regard, TGFβ1 induced expression of fibronectin and integrins appeared to be necessary for the subsequent induction of the expression of αSMA in renal cells [51, 52].
Because EMT results in an increase in cell migration and invasiveness, we assessed the migratory and invasive capacity of TGFβ1 exposed BEAS-2B cells. We observed both increased migration and enhanced invasiveness in BEAS-2B cells subjected to chronic exposure to TGFβ1, similar to the results reported by Borthwick et al  of epithelial cells undergoing EMT in the context of obliterative bronchiolitis (OB) following lung transplantation. The acquisition of a more motile phenotype of bronchial epithelial cells undergoing EMT might facilitate invasion of the sub-epithelial layer with enhanced contribution to the deposition of excess matrix proteins. Accompanying the increased migratory and invasive phenotype, we also observed elevated production and secretion of MMP-9 and MMP-2. MMP-2 and MMP-9 not only promote a motile cell phenotype through matrix degradation but can also activate latent TGFβ1. Induction of MMP-2 expression has been reported to be an important step in kidney fibrosis by disrupting the basement membrane thereby facilitating the migration of epithelial derived myofibroblasts into the interstitium . Further, asthmatic patients have increased immunoreactivity for MMP-9 in their airway epithelium and submucosa [53, 54]. Overexpression of MMP-proteins could further contribute to airway remodeling in asthma by feeding into the cycle of excess production and turn-over of matrix proteins. A correlation between fibrosis in asthma and MMP-9 expression has recently been demonstrated in a mouse model of chronic asthma . MMP-9 knockout mice showed a modest reduction in fibrosis, although no effect on mucus production or smooth muscle thickness was observed, suggesting a restricted role of MMP-9 in airway remodeling.
Alpha smooth muscle actin is characteristically expressed in myofibroblasts, enabling contractibility and an overall more invasive motile cell type. We detected upregulation of αSMA on the mRNA as well as protein level in BEAS-2B by days 3 to 4. In a recent clinical study Larsen et al  showed evidence for activated mobile fibroblasts in the BAL fluid of mild asthmatics, which upon stimulation with TGFβ1 produced more ECM proteins. These in vivo data are consistent with our hypothesis that epithelial cells undergo transition into myofibroblasts in the context of asthma.
Our results using BEAS-2B cells show that TGFβ1 clearly induces EMT in the transformed bronchial epithelial cell line. Importantly, we also observed an almost identical pattern of EMT following stimulation with TGFβ1 in primary normal human bronchial epithelial cells (NHBE). These results establish that TGFβ-induced EMT is not limited to alveolar epithelial cells but can also be induced in normal human bronchial epithelial cells in vitro. It is important to note, however, that our experiments utilized normal rather than asthmatic cells. Wound healing is part of the normal response of the epithelium to injury. In asthma the chronic cycle of injury and repair is thought to lead to the deregulation of factors involved, resulting in airway remodeling . Our observations further highlight the possible functional consequences of EMT in both physiologic wound healing as well as pathophysiologic remodeling in the airway. We also studied cells cultured under submerged conditions rather than at the air-liquid interface. A recent report by Hackett et al.  did study TGFβ1 induced EMT in both primary airway epithelial cells from normal and asthmatic donors as well as grown under submerged versus air-liquid interface conditions. They observed no differences between normal and asthmatic cells under submerged conditions. Under air-liquid interface conditions, the only significant difference they observed was that EMT was restricted to the basal cells in normal cultures but was less restricted in asthmatic cultures.
The inflammatory cytokine IL-1β is elevated in BAL fluid of symptomatic asthmatics , and there is evidence for a cross-talk between the TGFβ1 and IL-1β signaling pathways . Furthermore, overexpression of IL-1β caused emphysema and fibrosis in the airway walls in a murine model of COPD  and IL-1β has been shown to induce endothelial to mesenchymal transformation in skin . Therefore, we assessed the impact of IL-1β on TGFβ1-induced EMT in BEAS-2B cells. By itself, IL-1β induced a statistically significant decrease in E-cadherin expression and a statistically significant increase in tenascin C expression. When added together with TGFβ1, IL-1β had a significant additive effect on the changes in expression of these genes. Considering the critical role decreased E-cadherin plays in the initiation of EMT, the limited effect of IL-1β may prove to be biologically significant. Our results are compatible to the report by Kim et al.  showing synergistic effects of TGFβ1 and IL-1β on the expression of mesenchymal markers in the A549 cancer cell line, without evidence of induction of EMT by IL-1β alone.
Bronchial asthma is a chronic inflammatory disorder in which corticosteroids have become the first line of therapy. Whereas multiple studies support the benefit of corticosteroid treatment in respect to asthma symptoms and disease exacerbations [53, 63], there is considerable uncertainty concerning whether corticosteroids significantly slow airway remodeling. Several clinical studies as well as studies using murine models of allergic airway inflammation have suggested that corticosteroids reduce subepithelial fibrosis [19, 63–66]. Other clinical studies show evidence for persistently elevated levels of TGFβ1 and peribronchial fibrosis in the airway of asthmatic patients despite the reduction of inflammatory cells following treatment with corticosteroids [22, 67].
We were therefore interested to test the impact of corticosteroids in our model of EMT. Preincubation of BEAS-2B cells with dexamethasone or budesonide followed by TGFβ1 stimulation in a moderate concentration did not prevent the morphological changes or influence the reduction in E-cadherin expression. The effect on the induction of ECM proteins was variable as we observed no reduction of the TGFβ1 induced expression of fibronectin-EDA, and a slight reduction in the expression of collagen I and tenascin C. Budesonide, but not dexamethasone, inhibited TGFβ1 induced αSMA expression. We did not confirm the lack of efficacy of corticosteroids in abrogating EMT using primary airway epithelial cells, and this will be important to do in future studies.
These data suggest that corticosteroid have only a modest impact on TGFβ1-induced EMT. This finding is consistent with reports that while corticosteroid have proven to be very beneficial in treating asthmatic inflammation, their efficacy in preventing or reversing the remodeling process may be limited. New therapy strategies may need to be developed to target airway remodeling in asthma. TGFβ1 has been proposed as a target using anti-sense oligonucleotide, pan specific neutralizing antibodies as well as kinase inhibitors targeting TGFβ1 receptors. Anti-TGFβ1 and TGFβ2 antibodies have been shown to be effective in animal models of renal and ocular fibrosis and are currently in phase I/II trials in humans (reviewed in ).