Epithelial transglutaminase 2 is needed for T cell interleukin-17 production and subsequent pulmonary inflammation and fibrosis in bleomycin-treated mice

To explore the role of TG2 in the inflammation and fibrosis caused by noninfectious lung injury, we used a mouse BLM-induced pulmonary fibrosis model. WT C57BL/6 (B6) and TG2^−/− mice were exposed to BLM by intratracheal instillation and sacrificed 21 d later. Histopathological examination of the lungs of BLM-treated TG2^−/− mice revealed markedly reduced inflammatory cell infiltration (Fig. 1 A) and collagen deposition (Fig. 1 B) compared with WT animals. A Sircol assay revealed significantly decreased total lung collagen content in BLM-treated TG2^−/− mice compared with WT controls (Fig. 1 C). To evaluate the factors responsible for initiating pulmonary inflammation, we measured the levels of cytokines and chemokines in bronchoalveolar lavage fluid (BALF). Notably, IL-6 was detected (by ELISA) in BALF from WT mice, but not from TG2^−/− mice, 3 d after BLM treatment (Fig. 1 D). BALF levels of both MCP-1 and MCP-3 were higher in WT mice compared with TG2^−/− animals (Fig. 1, E and F). Semiquantitative RT-PCR analyses further showed that BLM treatment significantly increased the expression of key cytokines and chemokines on day 3 in the lungs of WT mice, but not in TG2^−/− mice (Fig. S1). Increased production of cytokines and chemokines in WT mice compared with TG2^−/− mice were mirrored in the inflammatory cell recruitment in BALF. Flow cytometric analysis of the cells present in BALF revealed that BLM-treated WT mice exhibited gradual increases in the numbers of total inflammatory cells (Fig. 1 G), lymphocytes (Fig. 1 H), macrophages (Fig. 1 I), and neutrophils (Fig. 1 J) over a 10-d period, whereas significantly fewer cells were recruited to the lungs of the TG2^−/− mice (Fig. 1, G–J). Although the initial recruitment of neutrophils 1 d after BLM instillation was comparable in WT and TG2^−/− mice (Fig. 1 J), secondary neutrophilic infiltration was observed only in the WT mice (Fig. 1 J).

IL-6 is a key stimulator of Th17 cell differentiation (Weaver et al., 2007; McGeachy and Cua, 2008; Korn et al., 2009), and the role of Th17 cells in the immune response that follows pulmonary infection is well established (Aujla et al., 2007; Korn et al., 2009). We tested whether TG2-mediated IL-6 secretion contributed to Th17 differentiation in lungs damaged by noninfectious injury (caused by BLM treatment). Numbers of CD4⁺ T cells in BALF, which constitute 40–70% of the total BALF T cell population, increased at days 5, 7, and 10 (unpublished data). WT mice displayed ∼6-fold higher percentages and numbers of CD4⁺ IL-17⁺ T cells (hereafter referred to as IL-17 instead of IL-17A) at days 7, 10, and 13 after BLM treatment than did TG2^−/− mice (Fig. 2, A and C). Thirty to forty percent of IL-17⁺ cells also expressed IL-17F, whereas all IL-17F⁺ cells coexpressed IL-17 (Fig. S2). Numbers of CD4⁺ IFN-γ⁺ T cells in WT mice were approximately twofold higher than in TG2^−/− mice (Fig. 2, B and C). Moreover, IL-17 levels in BALF correlated with CD4⁺ IL-17⁺ T cell numbers and peaked 13 d after BLM treatment (Fig. 2 D).

To evaluate the essential role of IL-17 in inflammatory cell recruitment after exposure to BLM, we treated mice between days 0 and 10 with IL-17RA-Fc, which can block both IL-17 and IL-17F (Kuestner et al., 2007). IL-17RA-Fc treatment decreased BALF levels of IL-17 (Fig. 2 D) and numbers of CD4⁺ IL-17⁺ T cells (Fig. 2 E), as well as numbers of total inflammatory cells (Fig. 2 F) and neutrophils (Fig. 2 G). We next evaluated the effect of IL-17RA-Fc on the development of fibrosis. BLM-exposed mice treated with IL-17RA-Fc between days 0 and 10 exhibited reduced fibrosis compared with isotype control-treated animals (days 0 to 21) and a similar or greater reduction than animals treated with IL-17RA-Fc between days 0 and 21 (Fig. 2 H). This suggests that the production of IL-17 and/or IL-17F after exposure to BLM is important in the pathogenesis of lung injury, predominantly during the initial inflammatory phase. This hypothesis was confirmed by the finding that mice treated with IL-17RA-Fc between days 11 and 21 displayed similar or higher levels of fibrosis to/than those treated with isotype control between days 0 and 21 (Fig. 2 H). Thus, IL-17RA-Fc was more effective at blocking fibrosis during the first half of the treatment period.

The critical role of IL-6 during BLM-induced Th17-mediated pulmonary inflammation and fibrosis was confirmed by blocking with anti-IL-6 antibodies. Treatment between days 0 and 10 decreased the numbers of both total inflammatory cells and CD4⁺ T cells by half compared with isotype controls (Fig. 3, A and B). The reduction in CD4⁺ IL-17⁺ T cell numbers by anti-IL-6 antibody treatment was more marked: less than 15% of those in isotype control-treated mice (Fig. 3 C). This reduced inflammation was reflected in the decreased pulmonary fibrosis of these mice at 21 d after BLM instillation compared with isotype controls (Fig. 3, D and E).

TG2 is a ubiquitous enzyme expressed in both parenchymal cells and inflammatory cells. Innate immune cells, such as macrophages, are known to secrete inflammatory cytokines (including IL-6) that initiate acute inflammation after tissue injury. Moreover, TG2 activity in inflammatory cells, but not endothelial cells, was crucial to the pathogenesis of atherosclerosis induced by a high-fat diet in LDL receptor-deficient mice (Boisvert et al., 2006). However, in our study, BALF IL-6 levels peaked 3 d after BLM treatment (Fig. 1 D) and did not correlate with numbers of infiltrating inflammatory cells (Fig. 1, G, H, and J), suggesting that innate immune cells may not be the primary source of IL-6 in our model. We hypothesized that lung epithelial cells or fibroblasts may instead produce IL-6 in a TG2-dependent manner, in response to BLM-induced tissue injury. To test this, we evaluated IL-6 expression in the lungs of WT B6 mice 2–3 d after BLM treatment. Basal IL-6 expression levels in the lung were low and increased by BLM treatment. Dense immunoreactive foci were localized in the pulmonary epithelial cells, particularly in the round-shaped type II epithelial cells that expressed prosurfactant protein-C (pro-SP-C; Fig. 4 A). TG2-dependent IL-6 production from pulmonary epithelial cells was evaluated using purified epithelial cells. Sort-purified primary mouse lung epithelial cells (MLECs; purity > 90%; Fig. 4 B, left) from WT and TG2^−/− mice were treated with BLM. Treatment with BLM for 48 h increased secretion of IL-6 by BLM-treated primary MLECs from WT, but not TG2^−/−, mice (Fig. 4 B, right). IL-6 levels in the culture supernatants of BLM-treated primary mouse lung fibroblasts (MLFs) were not greater than in PBS-treated controls of either mouse strain (Fig. 4 B, right). This selective responsiveness of epithelial cells was corroborated by the fact that BLM stimulated TG2 activity in primary MLECs, but not primary MLFs, as evaluated by Western blotting, ELISA, and fluorescence microscopy in conjunction with biotinylated pentylamine incorporation (Fig. S3, A–C). TG2-mediated IL-6 production from MLECs was confirmed by the addition of the TG2 inhibitor cysteamine (CyM). IL-6 production from WT MLECs was inhibited by CyM in a dose-dependent manner (Fig. 4 C).

Next, the role of NF-κB activation in TG2-induced IL-6 secretion by pulmonary epithelial cells was evaluated. BLM treatment increased NF-κB reporter activity in primary MLECs from WT, but not TG2^−/−, mice (Fig. S4 A). NF-κB activity was not altered in primary MLFs from either mouse strain (Fig. S4 B). Levels of phospho-p65 in the lungs of WT mice increased 3 d after exposure to BLM and were higher than those in TG2^−/− mice at the same time point (Fig. S4 C). Moreover, the secretion of IL-6 by primary MLECs from WT mice was suppressed by treatment with the NF-κB inhibitor Bay-11-7082 (Fig. 4 D). These data suggest that BLM induced secretion of IL-6 by lung epithelial cells through activation of a TG2-dependent NF-κB signaling pathway.

Because TG2 is also expressed in T cells, the failure to trigger a Th17 response in TG2^−/− mice may reflect a defect in the differentiation of TG2^−/− T cells. To assess this, we induced the Th17 differentiation of lymphocytes isolated from WT and TG2^−/− mice by addition of IL-6 and TGF-β to the growth medium. We detected no difference between the percentages of CD4⁺ T cells expressing IL-17 in WT and TG2^−/− mice (Fig. S5 A). Moreover, TGF-β–treated T cells from TG2^−/− mice displayed similar in vitro differentiation into regulatory T cells as did T cells from WT mice (Fig. S5 B). These findings show that defects in Th17 differentiation in TG2^−/− mice are not T cell intrinsic. The importance of T cell-extrinsic factors for Th17 differentiation after BLM-induced injury was confirmed by applying culture supernatant from BLM-stimulated primary MLECs from WT and TG2^−/− mice to activated CD4⁺ T cells from WT mice. Incubation with culture supernatants from WT MLECs, but not from TG2^−/− MLECs, increased the Th17 differentiation of CD4⁺ T cells. These effects were abrogated by the addition of anti–IL-6 or anti–TGF-β antibodies to the culture medium (Fig. 4 E).

The TG2-dependency of IL-6 production in pulmonary epithelial cells and subsequent Th17 differentiation were the characteristic features of BLM-induced lung injury. In contrast, exposure to LPS, an infection-associated stimulus, induced secretion of IL-6 by primary MLECs and macrophages from both WT and TG2^−/− mice (Fig. 4 F), and culture supernatants from LPS-stimulated primary MLECs both from WT and TG2^−/− mice induced Th17 differentiation of CD4⁺ T cells (unpublished data). Activation of epithelial TGF-β, which contributes to Th17 differentiation, after treatment with either BLM or LPS, was also independent of TG2 expression (Fig. S6 and data not shown). These results indicate that epithelial TG2 activity is crucial for IL-6 secretion and Th17 responses in noninfectious BLM-induced injury.

In this study, we hypothesize that TG2 expressed by pulmonary epithelial cells responded to BLM treatment and induced Th17 differentiation, leading to inflammatory cascade and fibrosis. We questioned how the tissue microenvironment produced by activated pulmonary epithelial cells could lead to Th17 differentiation. One possibility is that the local microenvironment in the lung has been translated into tissue dendritic cells. Educated dendritic cells migrate into the draining lymph nodes and activate and differentiate naive T cells; differentiated Th17 cells then migrate to the inflamed lung. Another possibility is that the tissue microenvironment in the lung directly induces T cell activation and differentiation. To evaluate these possibilities, we intratracheally administered BLM, followed by application of FTY720. FTY720, a superagonist of sphingosine-1-phosphate receptor 1, inhibits the egress of T cells from, but not migration of tissue dendritic cells from inflamed tissue to, lymph nodes (Worbs et al., 2006). If BLM-induced T cell activation and differentiation occurs in the draining lymph nodes, then differentiated Th17 cells would be expected to accumulate there and would not migrate into the lung in the presence of FTY720. FTY720 treatment between days 3 and 7 (analyzed on day 7), between days 3 and 8 (analyzed on day 8), and between days 5 and 8 (analyzed on day 8) did not show accumulation of differentiated Th17 cells in the draining lymph nodes compared with PBS-treated controls (Fig. 5 A). FTY720 treatment also inhibited pulmonary Th17 differentiation, suggesting a lack of migration of unactivated T cells into the lung after FTY720 treatment (Fig. 5 B). Time-scheduled treatment of FTY720 suggested that T cell migration and Th17 differentiation in the lung occur primarily between days 5 and 8 after BLM instillation. A reduction in the numbers and percentage of CD4⁺IL-17⁺ T cells in the lung according to the serial shortening of the treatment duration of FTY720 did not lead to proportionate accumulation of these cells in the draining lymph nodes (Fig. 5 C).

We next evaluated T cell proliferation and Th17 differentiation in the lung and draining lymph nodes during this period. CFSE-labeled Thy1.1⁺CD4⁺ T cells were adoptively transferred into B6 (Thy1.2) mice 1 d after BLM instillation. Recipient mice received 300 µg of depleting anti-Thy1.2 antibody on days −1 and 2. Donor CD4⁺ T cells in the lung showed proliferation and Th17 differentiation from day 5 after BLM instillation. A sequential increase in the percentages and numbers of donor Thy1.1⁺CD4⁺ T cells and CD4⁺IL-17⁺ cells in the lung was evident from days 5–8 after BLM instillation (Fig. 5 D). In contrast, CD4⁺ T cell proliferation in the draining lymph nodes and spleens of BLM-administered mice was comparable to that in PBS-treated control animals, reflecting homeostatic proliferation of these cells in the lymphoid organs of T cell–depleted hosts (Fig. 5 E and not depicted). Notably, homeostatic proliferation did not induce Th17 differentiation in either the lung or the draining lymph nodes (Fig. 5 E). Transfer of sort-purified Thy1.1⁺CD4⁺CD45RB^hi (naive phenotype) cells also revealed comparable donor CD4⁺ T cell proliferation and Th17 differentiation in the lung (Fig. 5 F). The effect of FTY720 on naive phenotype CD4⁺ T cell recruitment into the lung was evaluated by treating with FTY720 during the whole experimental period. Increased numbers of donor CD4⁺ and CD4⁺IL-17⁺ T cells after BLM treatment were blocked by FTY720 treatment (Fig. 5 G), indicating that migration of unactivated T cells into the lung was blocked by FTY720 treatment. In the draining lymph nodes, FTY720 treatment, again, did not lead to the accumulation of donor CD4⁺ and CD4⁺IL-17⁺ T cells after naive phenotype CD4⁺ T cell transfer and BLM treatment (Fig. 5 G).

To further confirm the role of the lung epithelium and epithelial TG2 in Th17 responses in vivo, BM-chimeric mice were exposed to BLM, and numbers of cells in BALF were analyzed 10 d later. WT mice reconstituted with TG2^−/− BM (BM^TG2−/−→WT) exhibited a similar percentage of CD4⁺ IL-17⁺ cells to WT control chimeras (BM^WT→WT), whereas TG2^−/− mice reconstituted with WT BM (BM^WT→TG2^−/−) displayed reduced percentages of CD4⁺ IL-17⁺ cells that were comparable to those in TG2^−/− control chimeras (BM^TG2−/−→TG2^−/−; Fig. 6 A). Th17 differentiation in the chimeras correlated with BALF IL-6 levels (Fig. 6 B). Histopathological analysis 21 d after BLM treatment revealed fewer inflammatory cells and reduced collagen deposition in the lungs of chimeras in which TG2^−/− mice were the recipients (BM^WT→TG2^−/− and BM^TG2−/−→TG2^−/−), compared with those in which WT mice were the recipients (BM^WT→WT and BM^TG2−/−→WT; Fig. 6 C). Similarly, fibrosis, expressed as percentage of fibrotic tissue (as assessed by Masson’s trichrome staining; Fig. 6 D), and collagen content (Fig. 6 E) were reduced in the TG2^−/− host chimeras (BM^WT→TG2^−/− and BM^TG2−/−→TG2^−/−). Collectively, these results indicate that nonhematopoietic (i.e., epithelial) TG2 is important for generation of the local Th17 response and subsequent fibrosis that follows BLM-induced tissue injury.

Chronic inflammation induces TGF-β activation, which, in turn, dampens the inflammatory response (Serhan and Savill, 2005; Li et al., 2006). We found the bioactivity of TGF-β in the BALF of BLM-exposed WT mice to be significantly higher than that in similarly treated TG2^−/− mice (Fig. S7 A). This increase displayed similar kinetics to that of inflammatory cell infiltration (Fig. 1, G–J), peaking at day 10 after BLM exposure (Fig. S7 A). Consistent with these findings, the bioactivity of TGF-β in the BALF of WT mice reconstituted with TG2^−/− BM (BM^TG2−/−→WT) was markedly higher than that in TG2^−/− mice reconstituted with WT BM (BM^WT→TG2^−/−; Fig. S7 B), a response that mirrors inflammatory cell recruitment in these chimeras (Fig. 6, A and B). Thus, epithelial TG2 induced an inflammatory cascade, comprising Th17 induction and TGF-β activation.

We next evaluated the role of fibroblast-derived TG2 in the effector phase of fibrogenesis. TGF-β treatment increased the expression and cross-linking of the ECM protein fibronectin in primary MLFs from WT mice, but not from TG2^−/− animals (Fig. S8, A and B). Furthermore, TGF-β–mediated increases in TG2 activity, fibronectin expression, and cross-linking, as well as α-SMA expression in IMR90 human fibroblasts, were inhibited by the TG2 inhibitor CyM (Fig. S9, A and B). These results indicate that TG2 is a downstream effector of TGF-β responses in fibroblasts during wound healing/fibrosis.

We have shown that TG2 is very important in both the initial inflammatory phase and subsequent fibrosis phase in our BLM-induced lung fibrosis model. Because clinical cases of pulmonary fibrosis are characterized by both inflammation and wound healing/fibrosis, and the inhibition of either process alone using antiinflammatory or antifibrotic drugs does not adequately reduce disease progression (Rogliani et al., 2008; Kim and Meyer, 2008; du Bois, 2010), the distinct roles of TG2 in both processes are intriguing and warrant further evaluation. We assessed the effects of inhibiting TG2 during the early and late phases on the development of lung fibrosis. BLM-exposed mice treated with CyM between days 0 and 10 exhibited markedly reduced leukocyte infiltration and fibrosis, comparable to mice treated with CyM between days 0 and 21, as evaluated by histopathological analyses (Fig. 7 A), measurements of fibrotic areas (Fig. 7 B), and total lung collagen content (Fig. 7 C). Mice treated with CyM between days 11 and 21 also exhibited significantly reduced fibrosis relative to animals treated with vehicle between days 0 and 21 (Fig. 7, A–C). We also observed significantly reduced fibrosis in mice treated with CyM between days 15 and 28 compared with animals treated with vehicle alone between days 15 and 28 in a 28-d BLM-induced lung fibrosis model (Fig. S10).

A causal role for TG2 in pathological increases in collagen deposition in the kidney and liver has been proposed; specifically, TG2 is considered to be important in the development of fibrosis in these organs (Shweke et al., 2008; Elli et al., 2009). Originally described as an enzyme that irreversibly cross-links ECM proteins, TG2 has recently been shown to play an active role in inflammation (Kim, 2006). Here, we used a BLM-induced pulmonary fibrosis model to demonstrate that activation of epithelial TG2 induces a Th17 inflammatory response in the lung that causes excessive wound healing and fibrosis through activation of fibroblast TG2. We have shown that TG2 is a critical regulator of inflammatory and repair responses and links noninfectious tissue injury and fibrogenesis.

In this study, we identified a critical role for epithelial cell–derived TG2 in inducing the secretion of IL-6, and consequent Th17 inflammation, after BLM-induced lung damage. During protective immunity against pathogens and autoimmune tissue destruction, IL-6 secreted by innate immune cells is thought to activate Th17 differentiation (Weaver et al., 2007; McGeachy and Cua, 2008; Korn et al., 2009). However, we found that WT lung epithelial cells treated with BLM produced IL-6, and that soluble factors secreted from BLM-exposed lung epithelial cells induced Th17 differentiation. These effects were blocked by anti–IL-6 and anti–TGF-β antibodies, indicating that IL-6 and TGF-β are the primary cytokines involved in Th17 differentiation after BLM-induced lung injury. Treatment with a blocking anti–IL-6 antibody in vivo also confirmed the critical role of IL-6 in Th17-mediated pulmonary inflammation and fibrosis.

In contrast, culture supernatants from BLM-stimulated lung epithelial cells from TG2^−/− mice contained only low levels of IL-6 and did not support Th17 differentiation. The critical involvement of epithelial TG2 in Th17 responses in vivo was further confirmed in our BM chimera study. As anticipated, WT recipients produced greater amounts of IL-6 than TG2^−/− recipients. In WT mice reconstituted with WT or TG2^−/− BM cells, 30–50% of CD4⁺ T cells produced IL-17, whereas very few CD4⁺ IL-17⁺ cells were present in TG2^−/− recipients. Although this disparity may stem from differences in the amount of IL-6 secreted by epithelial cells, IL-6 production is not always TG2-dependent. In an LPS-induced septic shock model, both WT and TG2^−/− mice rapidly secreted large amounts of IL-6 and IFN-γ (Falasca et al., 2008). In the present study, LPS induced the secretion of IL-6 from both lung epithelial cells and lung macrophages in a TG2-independent manner. Bacterial molecules activated epithelial and inflammatory cells directly through TLRs in a process that seemed to be TG2 independent.

In conclusion, TG2-dependent epithelial IL-6 production is a defining characteristic of noninfectious epithelial injury. TG2 serves as a stress sensor in epithelial cells (Ientile et al., 2007) and transduces epithelial damage into IL-6 secretion, which provides the signal for inflammatory Th17 differentiation after noninfectious injury. Epithelium-derived IL-6 may also modulate the local balance between Th17 and regulatory T cells, and thus influence immune homeostasis (Ivanov et al., 2007). The role of TG2 in this context requires further investigation, especially under noninfectious conditions. Th17 cells have been implicated in various autoimmune diseases, including multiple sclerosis (Kebir et al., 2007) and psoriasis (Di Cesare et al., 2009), and it will be interesting to determine whether the activation of TG2 promotes Th17 differentiation in such diseases.

T cell activation and Th17 differentiation in the lung after BLM instillation requires further comment. In this study, we obtained an unexpected finding that CD4⁺ T cell proliferation and Th17 differentiation occurred primarily in the lung, not in the draining lymph nodes, after BLM-mediated pulmonary epithelial injury. A long-standing immunological paradigm is the central sensitization of naive T cells in the lymphoid organs, which reflects the characteristic migratory pathway of naive T cells confined to these sites (Dailey, 1998). For this reason, CD4⁺ T cell activation and differentiation in the peripheral tissue is intriguing, and recent studies showing that naive T cells also migrate to peripheral tissues, including the lung (Cose et al., 2006; Staton et al., 2006), further confirm our data; adoptive transfer of naive phenotype CD4⁺ donor T cells revealed CD4⁺ T cell proliferation and Th17 differentiation in the lung after BLM instillation, and FTY720 treatment did not lead to accumulation of donor CD4⁺ and CD4⁺IL-17⁺ T cells in the draining lymph nodes. Immune activation and amplification in peripheral tissues is considered to be significant for perpetuating chronic inflammation; i.e., ectopic lymphoid structures in the joints in rheumatoid arthritis (Manzo and Pitzalis, 2007). Moreover, migration to and activation of naive tumor-specific CD8⁺ T cells inside the tumor tissue has warranted proposal of a new pathway, where naive T cells can be activated by interaction with tissue-resident cells (Thompson et al., 2010). In this study, naive CD8⁺ T cells were activated both by tumor cells themselves and by cross-presenting antigen-presenting cells within the tumor (Thompson et al., 2010). The nature of the tissue-resident cells in the lungs of BLM-administered mice that provide such signals has not been determined. As is the case in tumor tissue, pulmonary epithelial cells themselves or cross-presenting tissue-resident antigen-presenting cells may perform this function. Naive CD4⁺ T cell migration, activation, and differentiation in the inflamed lung may thus serve as a novel pathway of peripheral activation of naive T cells after epithelial tissue injury. The characteristic tissue microenvironment of the lung, or induced lymphoid-like structures in this tissue, may be responsible for this. The importance of other mucosal tissues for activation of naive T cells warrants further investigation. Such alternate T cell activation pathways may provide new methods for regulation of immune responses.

Th17 lymphocytes mediate lung inflammation by regulating neutrophil recruitment (Aujla et al., 2007; Weaver et al., 2007; McGeachy and Cua, 2008; Korn et al., 2009). Here, we demonstrated that induction of a Th17 response is critical during the initial inflammatory cascade that follows intratracheal BLM instillation, consistent with recent works (Sonnenberg et al., 2010; Wilson et al., 2010). Recruitment of inflammatory cells, most notably neutrophils, paralleled this Th17 response, whereas blocking IL-17 signaling through the administration of IL-17RA-Fc between days 0 and 10 greatly reduced inflammatory cell/neutrophil recruitment. IL-17RA-Fc treatment also reduced lung fibrosis measured at day 21, consistent with previous studies that neutrophil activation and elastase production are critical to the pathogenesis of BLM-induced lung fibrosis (Jones et al., 1998; Chua et al., 2007). However, treatment with IL-17RA-Fc during the late phase (days 11 to 21) increased fibrosis by 1.4-fold compared with isotype control-treated mice. Initial amplification of the inflammatory response and subsequent negative regulatory effects of IL-17 were previously reported in allergic asthma (Schnyder-Candrian et al., 2006). IL-17 secreted by γδ T cells has been shown to provide a help signal during pulmonary epithelial repair (Braun et al., 2008). This pathway may contribute to the responses we observed in this study. Indeed, an essential role for IL-17A in lung inflammation, but not fibrosis, has also been reported in another noninfectious injury model of experimental silicosis (Lo Re et al., 2010).

In addition to the dominant role of IL-17, other cytokines such as IFN-γ, IL-1β, IL-13, TGF-β, IL-10, and IL-4 have also been implicated in the BLM-induced pulmonary fibrosis model (Wilson et al., 2010). IFN-γ was reported to have a critical role during BLM-induced pulmonary fibrosis (Chen et al., 2001), and we also found that numbers of CD4⁺ IFN-γ⁺ T cells in WT mice were approximately twofold higher than in TG2^−/− mice. However, there was a dramatic reduction in CD4⁺ IL-17⁺ T cells compared with CD4⁺ IFN-γ⁺ T cells in TG2^−/− mice. Nonetheless, our data showed that the reduction of pulmonary fibrosis in IFN-γ^−/− mice was not dramatic compared with TG2^−/− mice (unpublished data), and blocking the IL-17 signaling pathway through IL-17RA-Fc treatment almost completely abrogated lung inflammation, leading us to consider that IL-17 rather than IFN-γ plays a predominant role during the initial inflammation after BLM instillation. A recent work showing the predominant role of IL-17A in the BLM-induced lung fibrosis model (Wilson et al., 2010) also supports our contention. These authors showed that IL-17A–driven fibrosis is facilitated by IFN-γ, and that the reduced fibrotic response in IFN-γ^−/− mice was associated with decreased IL-17A expression and reduced circulating neutrophils (Wilson et al., 2010), which is preferable for the dominant role of IL-17 in the BLM-induced lung fibrosis model.

In the BLM-treated lung, amplification of the initial inflammatory response as a result of Th17 polarization results in activation of TGF-β, which would be expected to dampen local inflammation and facilitate the shift to fibrogenesis (Serhan and Savill, 2005; Li et al., 2006). TGF-β acts as a fibrogenic switch that increases ECM synthesis and reduces the activity of ECM-degrading enzymes (Serhan and Savill, 2005; Li et al., 2006). It also induces expression of proteins that stabilize the ECM, including plasminogen activator inhibitor-1 and TG2 (Verderio et al., 2004). When intracellular TG2 is activated, it localizes to the marginal zone and is secreted (Gaudry et al., 1999). Extracellular TG2 acts as an ECM remodeler that increases the cross-linking of ECM molecules (Akimov and Belkin, 2001; Verderio et al., 2004). Intriguingly, intracellular TG2 activity has been suggested to be critical in renal fibrosis after subtotal nephrectomy (Johnson et al., 1997). In the present study, we found that intracellular TG2 activity in fibroblasts mediated the majority of the fibrogenic effects of TGF-β, including the synthesis of fibronectin and α-smooth muscle actin. The fibrogenic effects of TGF-β were decreased by the TG inhibitor CyM and were not observed in primary MLFs from TG2^−/− mice. Thus, we suggest that during the fibrogenesis of lung fibroblasts, activated TGF-β exerts its effects in a primarily TG2-dependent manner.

The direct involvement of TG2 in the effector fibrogenic pathway downstream of TGF-β indicates that inhibition of TG2 may directly reduce fibrosis regardless of the presence or absence of active inflammation, a conclusion that is highly relevant to IPF, in which overt inflammation is not common (Kim and Meyer, 2008; Rogliani et al., 2008; du Bois, 2010). CyM treatment during the late phase alone in our BLM-induced lung fibrosis model significantly reduced lung fibrosis and collagen deposition. The much greater effect of CyM in mice treated during the early phase suggests that the inflammatory response is more dominant in BLM-treated mice than in human IPF patients. Thus, TG2 inhibition may modulate two essential pulmonary fibrosis pathways simultaneously and TG2 itself may represent an ideal molecular target in IPF, and inhibitors specific for TG2 may be useful tools in the treatment of pulmonary fibrotic diseases including IPF. But despite all this, the BLM-induced lung fibrosis model is an inflammation-oriented fibrosis model that may not be an ideal representation of human IPF, and inflammation and subsequent fibrosis processes may continue even in the late phase of the disease model. Therefore, it would be better to use a more fibrosis-oriented model to demonstrate the critical role of TG2 in inflammation-independent pulmonary fibrogenesis.

B6 mice were obtained from The Jackson Laboratory. TG2^−/− mice (De Laurenzi and Melino, 2001) were backcrossed to B6 mice for 12 generations (N12). 8–12-wk-old male mice were used for experiments, and all mice, including WT B6, were bred and maintained at the animal facility of Seoul National University College of Medicine. All animal experiments were performed with the approval of the Institutional Animal Care and Use Committee at Seoul National University.

Pulmonary fibrosis was induced by intratracheal instillation with BLM (1.5 mg/kg, Nippon Kayaku, Japan) in 50 µl PBS as described previously (Jones et al., 1998). A proportion of the mice were injected intraperitoneally with CyM (40 mg/kg/day; Sigma-Aldrich), recombinant murine IL-17RA-Fc (100 µg/mouse/day; R&D Systems) or control human IgG1 (100 µg/mouse/day; R&D Systems) suspended in PBS. To block IL-6 signaling, mice were injected intraperitoneally with blocking anti–IL-6 mAb (100 µg/mouse; eBioscience) or control Rat IgG1 (100 µg/mouse; eBioscience) every 3 d, starting 1 d before receiving BLM treatment). To evaluate Th17 differentiation in the lung, BLM-exposed mice were treated with PBS or FTY720 (2 mg/kg/day, i.p.) during the periods indicated.

Recipient mice were lethally irradiated (950 rad whole-body irradiation in two split doses) and injected intravenously with T cell–depleted BM cells (3 × 10⁶/mouse), as described previously (Surh et al., 1997). Reconstituted mice were used in the experiments 10 wk after BM cell transfer.

WT B6 mice received an intratracheal instillation of BLM (1.5 mg/kg). 1 d later, mice were injected with 5 × 10⁶ purified and CFSE-labeled CD4⁺ T cells from B6.Thy1.1 lymph nodes via the tail vein. To deplete endogenous T cells, anti-Thy1.2 antibody (30-H12, 300 µg/mouse) was administered i.p. to recipient mice on days −1 and 2 of BLM treatment. For the transfer of naive phenotype CD4⁺ T cells, cells from B6.Thy1.1 lymph nodes were labeled with APC-conjugated anti-CD4 and PE-conjugated anti-CD45RB. CD4⁺CD45RB^high cells were purified using a FACS Aria cell sorter (BD) and subsequently labeled with CFSE.

Lung tissues were fixed in 4% paraformaldehyde, processed, and embedded in paraffin. Sections were stained with H&E for histopathological analysis or with Masson’s trichrome for the evaluation of collagen content and distribution. The extent of fibrosis was measured using the ProgRes CF system (Jenoptik). To investigate IL-6 expression in epithelial cells, lung tissue from BLM-exposed mice was stained with anti–IL-6 (R&D Systems) and anti–pro-surfactant protein-C (pro-SP-C; Millipore). Alexa Fluor 488–conjugated donkey anti–goat IgG (Invitrogen) and Alexa Fluor 546–conjugated anti–rabbit IgG (Invitrogen) antibodies were used for visualization. Image acquisition and processing was performing using a confocal fluorescence microscope (Olympus) and FV10-ASW 2.0 Viewer (Olympus).

Bronchoalveolar lavage was performed with five 1.0-ml aliquots of PBS through a tracheal cannula. To evaluate cytokine production, BALF cells were harvested and restimulated with 50 ng/ml PMA and 1 µg/ml ionomycin (Sigma-Aldrich) for 4 h. For intracellular staining, brefeldin A (BD) was added during the final 2 h of stimulation. Cells were fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, and incubated with anti-CD4, anti-CD8, anti-IL-17, and anti–IFN-γ antibodies (eBioscience). Intracellular cytokine levels were assayed by flow cytometry.

The total collagen content of the left lung was measured by the Sircol collagen dye binding assay, according to the manufacturer’s instructions (Biocolor Ltd.).

BALF cells were incubated with mAbs to mouse CD4, CD8α, B220, CD11c, CD11b, Gr1, and F4/80 that were conjugated to fluorescein, phycoerythrin, PerCP-Cy5.5, or allophycocyanin (eBioscience). Cells were analyzed using a FACSCalibur (BD) and FlowJo software (Tree Star).

The lungs were perfused with PBS and digested with collagenase D (400 µg/ml; Invitrogen), dispase (50 µg/ml; Invitrogen), and DNase I (0.5 U/ml; Sigma-Aldrich) for 1 h at 37°C. Cells were suspended in 30% Percoll, layered onto the top of a 70% Percoll gradient and centrifuged (800 g, 20 min, 4°C). The top layer, an enriched population of epithelial cells and fibroblasts, was retained. Cells were labeled with FITC-conjugated anti-Ly5.2 (eBioscience) and anti–pro-SP-C (Millipore) mAbs, followed by PE-conjugated anti–rabbit IgG. Pro-SP-C–expressing epithelial cells were subsequently purified using a FACSAria cell sorter (BD) and epithelial cell purity after sorting was >90%. Cells were maintained in Keratinocyte Basal Medium (KBM; Lonza). For isolation of fibroblast-like cells, cells were harvested after Percoll gradient centrifugation, and labeled with biotinylated anti-Ly5.2 antibody, followed by streptavidin microbeads (Miltenyi Biotech). The negative fraction was collected using an LS magnetic column. Cells were allowed to adhere to a culture dish for 1 h at 37°C and nonadherent cells were removed. Adherent fibroblast-like cells (MLFs) were maintained in DME containing 10% FBS. The second passage of these cells was used for experiments.

Intracellular TG2 activity was evaluated by assay of 5-(biotinamido) pentylamine (BP) incorporation into cellular proteins. Primary MLECs and MLFs were incubated with 1 mM BP for 1 h before harvesting. The cell homogenates were coated onto microtiter plates, and BP-labeled proteins were assayed by ELISA. Primary MLECs of WT B6 mice were cultured on chamber slides, treated with BLM (5 µg/ml), and incubated with 0.1 mM BP for the final hour of the culture period. TG2 activity was assessed visually using Alexa Fluor 488–conjugated streptavidin (DAKO).

Lymphocytes isolated from lymph nodes were stimulated with anti-CD3 and anti-CD28 (1 µg/ml each; BD) in the presence of IL-6 (10 ng/ml; eBioscience) and TGF-β1 (5 ng/ml; R&D Systems) for 3 d, or anti-CD3 (1 µg/ml) in the presence of TGF-β1 (0.2–5.0 ng/ml) for 3 d. Cells were harvested and restimulated with PMA (50 ng/ml) and ionomycin (1 µg/ml) for 4 h. For intracellular staining, brefeldin A (1 µg/ml, Sigma-Aldrich) was added during the final 2 h of stimulation. Cells were then fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, and incubated with anti-CD4, anti-CD8, anti-IL-17, anti-IFN-γ, or anti-Foxp3 mAbs (eBioscience).

BALF or cell culture supernatants were analyzed for mouse IL-6, IL-17, MCP-1, MCP-3, and TGF-β1 (R&D systems) levels by sandwich ELISA (Lee et al., 1999).

Primary mouse MLECs or MLFs were co-transfected with p3κB-Luc and pRL-TK reporter constructs (Promega) for 24 h and then treated with BLM at various concentrations. Luciferase activity was assayed using a kit (Promega) with a Victor3 plate reader (Perkin Elmer) and normalized against renilla luciferase activity. To evaluate bioactivity of TGF-β, NIH-3T3 cells were co-transfected with pGL3-PAI1TRE-Luc and pRL-TK reporter constructs for 24 h, and then incubated with BALF obtained from BLM-treated mice or media supernatants for various lengths of time. Luciferase activity was assayed using a kit and normalized against renilla luciferase activity.

Total RNA was isolated from lung tissues using TRIzol (Invitrogen). cDNA generated from 1 µg of total RNA by Superscript II (Invitrogen) was analyzed by Q-PCR using Assay-on-Demand TaqMan probes with primers for TG2, TGF-β1, and GAPDH (Applied Biosystems) and an ABIPRISM 7900HT Sequence Detector (Perkin Elmer). All data were normalized to the respective 18s rRNA expression. The following primer pairs were used for semiquantitative PCR: IL-6, 5′-GACAAAGCCAGAGTCCTTCAGAG-3′/5′-CTAGGTTTGCCGAGTAGATCTC-3′; MCP-1, 5′-AGGTCCCTGTCATGCTTCTG-3′/5′-TCTGGACCCATTCCTTCTTG-3′; TNF, 5′-AAGGGATGAGAAGTTCCCAA-3′/5′-AGAACCTGGGAGTAGACAA-3′; MIP-1α, 5′-ATGAAGGTCTCCACCACTGC-3′/5′-CCCAGGTCTCTTTGGAGTCA-3′; MIP-2, 5′-AGTGAACTGCGCTGTCAATG-3′/5′-TTCAGGGTCAAGGCAAACTT-3′; and IP-10, 5′-AAGTGCTGCCGTCATTTTCT-3′/5′-GTGGCAATGATCTCAACACG-3′. PCR products were analyzed by resolution on a 2% agarose gel followed by densitometric analysis (Labworks 4.6; UVP BioImaging Systems).

Cells were harvested in a lysis solution containing 50 mM Tris/HCl (pH 7.6), 1% NP-40, 150 mM NaCl, 2 mM EDTA, and 100 µM PMSF in the presence of a protease inhibitor cocktail (Roche) and phosphatase inhibitor (Sigma-Aldrich). After incubation on ice for 30 min, cellular debris was removed by centrifugation (10 min, 4°C). Protein concentrations were assayed using the BCA protein assay reagent (Thermo Fisher Scientific). Total protein was resolved by SDS-PAGE, and then transferred to polyvinylidene difluoride membranes. After blocking with 5% skim milk, membranes were probed with the appropriate antibody. Blots were developed with an enhanced chemiluminescence Western blotting detection system (GE Healthcare). Antibodies used were as follows: anti-TG2 (Neomarker), anti-p65 (Abcam), anti–phospho-p65 (Cell Signaling technology), anti–α-smooth muscle actin (DAKO), anti-actin (Sigma-Aldrich), and anti-fibronectin (Santa Cruz Biotechnology, Inc.).

Statistical significance was analyzed using the Student’s t test. A p-value of <0.05 was taken to indicate statistical significance.

Fig. S1 shows semiquantitative RT-PCR analysis of cytokine and chemokine expression in the lungs of BLM-exposed mice. Fig. S2 shows that 30–40% of IL-17⁺CD4⁺ T cells coexpress IL-17F. Fig. S3 shows TG2 expression and activity in primary MLECs and primary MLFs after BLM treatment. Fig. S4 describes NF-κB reporter activity in primary MLECs after BLM treatment. Fig. S5 shows in vitro differentiation of Th17 and regulatory T cells. Fig. S6 shows TGF-β bioactivity from primary macrophages and primary MLECs stimulated with LPS and BLM. Fig. S7 shows the bioactivity of TGF-β in the BALF of BLM-exposed mice. Fig. S8 shows the expression of fibronectin and TG2 in primary MLF after TGF-β treatment. Fig. S9 shows increased TG2 activity and α-SMA expression in human fibroblasts after TGF-β treatment. Fig. S10 shows the effect of CyM treatment between days 15 and 28 in a 28 d BLM-induced lung fibrosis model.

We thank Drs. H.R. Kim and E. Choi for critical comments on the manuscript.

This work was supported by grants from MarineBio Technology Project, Ministry of Land, Transport, and Maritime Affairs (D-S. Lee) and Korea Healthcare Technology R&D Project, Ministry for Health, Welfare & Family Affairs (D-S. Lee; nos. A10019010110000100 and A08402210210000400).

The authors have declared that no conflict of interest exists.

Author contributions: K. Oh conducted most of the experiments. H.B. Park and O.J. Byoun performed the chimera study. D.M. Shin and E.M. Jeong performed some biochemical experiments in the supplemental data. Y.W. Kim, Y.S. Kim, and G. Melino interpreted the experimental results. D.S. Lee and K. Oh planned experimental approaches and analyzed and interpreted the experimental results. D.S. Lee, K. Oh, and I.G. Kim directed the project and wrote the paper.