The Latent Form of TGFβ₁ is Induced by TNFα Through an ERK Specific Pathway and is Activated by Asbestos-Derived Reactive Oxygen Species In Vitro and In Vivo

Abstract

Tumor necrosis factor-α (TNFα) and transforming growth factor-β₁ (TGFβ₁) are potent peptide growth factors that are likely to play important roles in the development of interstitial pulmonary fibrosis (IPF). Previously we showed that TNFα and TGFβ₁ are up-regulated in macrophages, epithelial and mesenchymal cells early after exposure to chrysotile asbestos, particularly at sites of fiber deposition in vivo. We also showed that TNFα receptor knockout mice are resistant to asbestos-induced fibrosis. Importantly, vectors that over-express TNFα cause inflammation and fibrogenesis along with increased TGFβ₁ production in C57Bl/6 mice. Recently we reported that TNFα activates the extracellular regulated kinase pathway in fibroblasts leading to a 200–400% increase in TGFβ₁ mRNA and protein. The mechanism of TNFα induction of TGFβ₁ expression appears to be complex, involving both transcriptional and post-transcriptional mechanisms. In asbestos-exposed animals, this TGFβ₁ is produced on alveolar surfaces in a latent form (controlled by binding of a latent associated peptide [LAP]) that must be activated for the TGFβ₁ to bind to its receptors and induce its multiple biological effects. Thus, we recently reported that, in vitro, reactive oxygen species (ROS) derived from chrysotile and crocidolite asbestos activate TGFβ₁ by oxidation of the LAP. Now, in preliminary findings, we have shown that over-expression of latent TGFβ₁ prior to asbestos exposure of fibrogenic-resistant TNFα receptor knockout mice produces asbestos lesions with the same severity as seen in normal C57/Bl6 mice. This finding plus the demonstration of increased amounts of TGFβ₁, increased Smad activation and amelioration of the developing disease by treating the mice with an anti-oxidant all support the concept that, in vivo, latent TGFβ₁ is activated by asbestos-generated oxygen radicals and consequently mediates at least a component of the consequent fibrogenesis. Taken together, these findings support the postulate that TNFα controls fibrogenesis by regulating TGFβ₁ expression and that one mechanism through which ROS induce lung fibrosis is by activating latent TGFβ ₁.

Keywords :

• Asbestos
• transforming growth factor-β1
• tumor necrosis factor-α
• reactive oxygen species
• lung fibrosis

INTRODUCTION

Interstitial pulmonary fibrosis (IPF) is a progressive and usually fatal disease. IPF can be caused by multiple factors including immune mechanisms, infectious agents and toxic gases and particles including asbestos (Morris and Brody, 1998). Symptoms of the disease may not appear for many years and a large proportion of cases remain idiopathic (Katzenstein and Myers, 1998). IPF is characterized by increased extracellular matrix production, varying degrees and types of inflammation, and proliferation of a variety of cells as the disease develops. IPF afflicts millions of individuals worldwide, yet there are no effective therapies available for any of the forms of IPF (Selman et al., 2004). A better understanding of the molecular mechanisms involved in initiation and progression of the disease may lead to new therapies.

Asbestosis displays the hallmarks of IPF (Brody, 1997; Mossman and Churg, 1998). We have been using brief exposure of rodents to inhaled chrysotile asbestos fibers as a model of IPF to study early molecular events important in initiating and sustaining the fibrogenic response. Our current work is focused on two potent peptide growth factors that appear to play key roles in the development of IPF. These are transforming growth factor-β₁ (TGFβ₁) and tumor necrosis factor-α (TNFα).

TGFβ₁ is a potent inducer of extracellular matrix proteins, and up-regulation of TGFβ₁ expression is a consistent feature of most fibrotic diseases. Indeed, patients with IPF have increased levels of TGFβ₁ in their lungs and the levels correlate with severity of lung disease (Bartram and Speer, 2004). We showed that inhalation exposure of rats to asbestos produces increased expression of TGFβ₁ and fibronectin at the sites of fiber deposition (Perdue and Brody, 1994) that correlates with elevated levels of extracellular matrix (ECM) as measured by morphometry (Chang et al., 1988; Coin et al., 1996). Importantly, over-expression of TGFβ₁ in the lung via a recombinant adenovirus expressing constitutively active TGFβ₁ is sufficient to cause the histopathological changes of pulmonary fibrosis in rats (Sime et al., 1997) and mice (Warshamana et al., 2002a and 2002b), while inhibitors of TGFβ₁ activity such as neutralizing antibodies to TGFβ₁ (Giri et al., 1993), TGFβ₁ soluble receptor (Wang et al., 2002) and Smad 7 protein expression block the development of bleomycin-induced fibrogenesis in mice (Nakao et al., 1999). Thus, it is clear that the induction of TGFβ₁ expression is an important step in the development of fibrosis. However, TGFβ₁ is synthesized and secreted in a latent form as a high molecular weight complex with the latent associated peptide (LAP). Latent TGFβ₁ must be converted to the mature, active form to bind to its receptors and induce its multiple biological effects.

Latent TGFβ₁ can be activated by a number of mechanisms including proteolytic cleavage (Harpel et al., 1992), interaction with the integrin α Vβ 6 (Munger et al., 1999) or exposure to reactive oxygen species (Ehrhart et al., 1996). The factors mediating the expression and activation of TGFβ ₁ in asbestos lesions are not all understood. It is our hypothesis that TNFα induces expression of TGFβ₁ in cells of the lung (Sullivan et al., 2005) and that asbestos-induced reactive oxygen species activate latent TGFβ₁ leading to increased expression of ECM and ultimately the scar that typifies asbestosis and IPF (Pociask et al., 2004).

TNFα has been shown to induce the expression of TGFβ₁ in a variety of cell types including microglial cells (Chao et al., 1995), mature adipocytes (Samad et al., 1999), human proximal tubular cells (Phillips et al., 1996), and rat pulmonary artery endothelial cells (Phan et al., 1992) in vitro. We previously showed that TNFα rapidly up-regulates TGFβ₁ and collagen mRNA expression in both primary mesenchymal and epithelial cells of the lung (Warshamana et al., 2001).

Importantly, Sime and colleagues showed that adenovirus-mediated expression of TNFα in lungs of normal rats induces TGFβ₁ production and interstitial fibrogenesis (Sime et al., 1998). In another study, administration of anti-TNFα antibody reduced bleomycin-induced lung injury and decreased expression of TGFβ₁ in the treated mice (Phan and Kunkel, 1992). Importantly, we showed that animals genetically engineered to have both TNFα receptors knocked out are resistant to the fibrogenic effects of asbestos (Liu and Brody, 2001), silica and bleomycin (Ortiz et al., 1999). There is also accumulating evidence from patients with IPF that TNFα is an important mediator of fibrogenesis. TNFα expression is increased in the lungs of these patients, and agents such as pirfenidone and etanercept with anti-TNFα properties have shown promise in treatment of some patients (Selman et al., 2004). Furthermore, TNFα polymorphisms indicative of high levels of TNFα expression correlate with an increased risk of developing IPF (Whyte et al., 2000; Maier et al., 2001; Yucesoy et al., 2001).

TNFα Induces Expression of TGFβ1 Through the ERK Pathway

In a recent study (Sullivan et al., 2005), we showed that treatment of fibroblasts isolated from the lungs of adult C57 mice with recombinant murine TNF α (5 ng/ml) resulted in increased levels of TGFβ 1 mRNA within 2 hr, peaking at 200–400% of untreated controls after 12 hr, and persisting for 24 hr. The increase in TGFβ1 mRNA correlated with an increase in secretion of latent TGFβ1 as measured by ELISA (R&D Systems, Minneapolis, MN). We also found that cell lines derived from mouse embryo fibroblasts (Swiss 3T3) (Sullivan et al., 2005), mouse Type II alveolar epithelial cells (C10), human alveolar epithelial (A549) and human lung fibroblasts (HFL1) up-regulated their TGFβ1 expression in response to TNFα (Figure 1).

FIG. 1 TNFα induces expression of TGFβ₁ mRNA in lung fibroblasts and alveolar epithelial cells. Confluent monolayers of primary mouse lung fibroblasts (MLF), Swiss 3T3 fibroblasts, mouse alveolar epithelial cells (C10), human lung fibroblasts (HFL1) or human alveolar epithelial cells (A549) were serum-starved for 48 hr and then treated with TNFα (10 ng/ml) or media alone (untreated control). At the indicated times, total RNA was isolated and TGFβ₁ mRNA was measured by quantitative real-time RT-PCR. The average CT calculated for TGFβ₁ mRNA in unstimulated cells was 25.5, except for A549 cells that were 20.3. Relative quantitation was determined using the comparative C_T method with data normalized to 36B4 riboprotein mRNA and calibrated to the average Δ C_T of the corresponding untreated control. Data are represented as the mean ± SEM, relative units, with an n = 3.

The mitogen activated protein kinases (MAPKs), extracellular-stimulus regulated kinase 1 and 2 (ERK1/ERK2), c-jun N-terminal kinase (JNK), and p38 have been implicated in regulation of TGFβ₁ expression (Grewal et al., 1999; Hamaguchi et al., 1999; Isono et al., 2000; Yue and Mulder, 2000) and TNFα is a potent activator of MAPKs (Wajant et al., 2003; Sullivan et al., 2005). To determine if MAPK are required for TNFα induction of TGFβ₁, we used a panel of well-characterized pharmacological inhibitors of individual MAPKs. Both PD98059 and U0126, specific inhibitors of ERK1/ERK2 activation, significantly blocked TNFα -induced expression of TGFβ₁ mRNA and protein in MLFs while inhibitors of JNK or p38 had little effect (Sullivan et al., 2005).

These results suggest a critical role for ERK1/ERK2 activation in TNFα -induced expression of TGFβ₁. The specificity of the response to TNFα was demonstrated by treating fibroblasts isolated from lungs of mice that are genetically engineered to lack both TNFα receptor 1 and receptor 2 (Jackson Laboratory, Bar Harbor, ME) and showing that the ERK pathway was not activated and TGFβ₁ expression was not increased in these cells (Sullivan et al., 2005). To determine if these results are biologically relevant, we over-expressed TNFα in the lungs of mice via a recombinant adenovirus carrying a cDNA for murine TNFα (AdTNF) and then examined ERK1/ERK2 activation and TGFβ₁ expression by immunohistochemistry. We found significant increases in phosphorylated-ERK1/ERK2 (phosphorylation is indicative of activation) and TGFβ₁ protein in the lungs of AdTNF-treated animals when compared to control adenovirus-treated animals. These results suggest that TNFα activates the ERK1/ERK2 pathway and up-regulates TGFβ₁ in vivo.

TNFα Increases Transcription of the TGFβ₁ Gene and Stabilizes TGFβ₁ mRNA

An increase in the steady-state levels of an mRNA could be the result of increased transcription and/or mRNA stabilization. Treatment of Swiss 3T3 fibroblasts with actinomycin D (ActD), an inhibitor of transcription, prior to addition of TNFα partially blocked the TNFα induction of TGFβ ₁ suggesting that de novo transcription is required for maximal TNFα induction of TGFβ₁ mRNA (Figure 2). Nuclear runon transcription assays confirmed an increase in TGFβ₁ transcription in response to TNFα (unpublished data). Chromatin immunoprecipitation assays revealed that the increase in TGFβ₁ transcription was associated with recruitment of RNA polymerase II to the TGFβ₁ gene (unpublished data). We also observed significant stabilization of the TGFβ₁ mRNA in response to TNFα (Sullivan et al., 2005). In untreated cells, the half-life of TGFβ₁ was ∼ 15 hr, whereas in TNFα -treated cells no degradation of the mRNA was detected after 24 hr exposure (Sullivan et al., 2005). This combination of transcriptional and post-transcriptional events leads to a 3–4-fold increase in steady-state TGFβ₁ mRNA 6–12 hr after exposure to TNFα. The mechanisms involved in TNFα -induced TGFβ₁ transcription and mRNA stabilization are currently under investigation.

FIG. 2 Inhibition of transcription partially blocks TNFα induction of TGFβ₁ mRNA. Quiescent Swiss 3T3 fibroblasts were pre-treated with actinomycin D (5 μ M) or vehicle (DMSO) as control for 30 min followed by addition of 10 ng/ml TNFα for another 6 hr. Expression of TGFβ₁ mRNA was determined by quantitative real-time RT-PCR as described in Figure 1. Data are represented as the mean ± SEM, relative units, with an n = 3. An unpaired Student's t-test was used to assess the difference between the indicated groups.

Asbestos-Induced Reactive Oxygen Species Activate Latent TGFβ₁

Deposition of asbestos fibers at the bronchiolar-alveolar duct bifurcations is associated with the induction of reactive oxygen species (ROS) and the release of TGFβ₁. TGFβ₁ is produced on alveolar surfaces in a latent form that must be activated to have a biological effect. We recently reported that ROS derived from chrysotile and crocidolite asbestos can activate TGFβ₁ in vitro by oxidation of the latent associated peptide (Pociask et al., 2004). We also showed that over-expression of active TGFβ₁ was sufficient to induce fibrosis in fibrogenic-resistant TNFα receptor knockout mice (Liu et al., 2001). To investigate the role of asbestos-induced ROS in the activation of TGFβ₁ in vivo, TNFα receptor knockout mice were transduced with an adenovirus vector that over-expresses full-length latent TGFβ₁ (adLTGF-β) or an empty vector control virus (MG3), and 7 d later exposed to chrysotile asbestos.

Mice over-expressing latent TGFβ₁ and exposed to asbestos had significantly higher levels of active TGFβ₁ in whole lung lysates measured by ELISA, greater amounts of Smad-2 phosphorylation assessed by immunohistochemistry and, importantly, more severe asbestos lesions than did mice receiving MG3 and asbestos or adLTGF-β alone. Latent TGFβ₁ over-expressers treated with the anti-oxidant Tempol (4-hydroxy-2,2,6,6-tetramethylpiperidine-N-oxyl) prior to asbestos exposure demonstrated less active TGFβ₁, lower levels of Smad-2 phosphorylation and consequently less severe asbestos lesions than untreated mice that were exposed to adLTGF-β and asbestos. Together these results suggest that latent TGFβ₁ can be activated by asbestos-derived ROS at sites of fiber deposition and that TGFβ ₁ expression plays a central role in the development of the fibrogenic lesions.

SUMMARY AND CONCLUSIONS

Exposure to asbestos fibers leads to a rapid increase in the expression of TNFα and TGFβ ₁, as well as other cytokines, at the sites of fiber deposition followed by synthesis and deposition of ECM. It is clear that expression of TGFβ₁ plays an important role in the pathogenesis of asbestosis and IPF. TNFα is also emerging as a key cytokine in the fibrogenic process. Our findings provide evidence that the fibrotic properties of TNFα may be mediated at least in part by its ability to up-regulate TGFβ₁ through a complex mechanism involving the ERK-specific MAPK signaling pathway. TGFβ₁ expression may be controlled at the levels of transcription, mRNA stability, mRNA translation, secretion of preformed protein, activation of the latent protein to its active form (Phillips et al., 1995, 1996, 1997; Morrisey et al., 2001) and inactivation by circulating proteins and ECM (Bartram and Speer, 2004).

We showed that TNFα increases transcription of the TGFβ₁ gene and stabilizes the transcript leading to a 2–4-fold increase in TGFβ₁ expression (Sullivan et al., 2005). Asbestos-induced ROS provide a mechanism for activating this latent TGFβ₁. Whether ROS activate latent TGFβ₁ in vivo directly, as seen in vitro (Pociask et al., 2004), or through modulation of proteases that are known to activate latent TGFβ₁ (Harpel et al., 1992) remains to be determined. Regardless, the increase in biologically active TGFβ₁ at the alveolar surface may lead to autoinduction of TGFβ₁ (Van Obberghen-Schilling et al., 1988) induction of other profibrotic factors such as CTGF (connective tissue growth factor) (Grotendorst, 1997; Duncanet al., 1999; Lasky and Brody, 2000) and synthesis and deposition of ECM (Bartram and Speer, 2004). Our current model is depicted in Figure 3.

FIG. 3 Proposed role of TNFα in asbestos-induced fibrogenesis. TNFα is expressed by a variety of cell types in the lung early after exposure to asbestos. TNFα binding to its cellular receptors activates the MEK/ERK signaling pathway leading to increased TGFβ₁ expression by mesenchymal and epithelial cells. Asbestos-induced ROS activate latent TGFβ₁ that can then bind to its cellular receptors activating signaling pathways that lead to increased production of ECM, a hallmark of fibrosis. Whether TNFα directly up-regulates TGFβ₁ or does so through a secondary mediator; elucidation of the cis- and trans-acting elements involved in TNFα up-regulation of TGFβ₁; and the precise mechanism of ROS activation of latent TGFβ₁ are currently under investigation in our laboratory.

TNFα has also been shown to antagonize TGFβ₁-induced signaling (Verrecchia et al., 2000) suggesting that these two pleiotropic cytokines may tightly regulate each other to maintain homeostasis. Clearly, the regulatory mechanisms underlying TNFα induction of TGFβ₁ expression and activation of the latent protein could be central to understanding the pathogenesis of asbestosis and IPF and may present new targets for therapeutic intervention.

Grant support is acknowledged from the National Institutes of Health (HL 60532 [ARB], ES06766 [ARB]).