TSLP plays an important role in the DC-mediated activation of Th2 inflammatory responses. Despite the expression in vivo and potential importance of epithelial cell-derived TSLP in allergic diseases including asthma, no studies have tested the effect of cytokines and TLR ligands on the production of TSLP in human airway epithelial cells. In the present study we determined the TSLP expression profiles of normal human bronchial epithelial cells after stimulation with cytokines and TLR ligands. This study provides the first demonstration that the Th2 cytokines IL-4 and IL-13 and rhinovirus infection induce TSLP expression in airway epithelial cells. We have also shown that IL-4 synergistically enhanced dsRNA- and rhinovirus-dependent TSLP production in airway epithelial cells.
IL-4 and IL-13 are mainly produced by Th2 cells and basophils and each have heterodimeric receptors (43
). The IL-4Rα-chain is shared by receptors for IL-4 and IL-13 and is essential for the activation of the central transcription factor STAT6 (43
). This suggests that STAT6 may be one transcription factor that can induce TSLP expression. As shown in , IL-4 and IL-13 alone were not able to induce detectable TSLP protein production in NHBE, although the mRNA for TSLP was significantly up-regulated by each. However, a combination of IL-4 or IL-13 with the inflammatory cytokine TNF-α led to detectable expression of the TSLP protein (). In a recent report, Bogiatzi et al. have shown that TSLP is induced by the combination of TNF-α and Th2 cytokines but not by the individual cytokines in skin keratinocytes (44
). These results suggest that the activation of NF-κB may support the STAT6-dependent production of TSLP. Previous studies have demonstrated a synergy between NF-κB and STAT6 in the induction of CCL11 (eotaxin-1) in response to TNF-α and IL-4/IL-13 in airway epithelial cells mediated by a composite response element in the eotaxin promoter (32
). In the TSLP promoter we found three putative NF-κB binding sites at −734 bp, −929 bp, and −4552 bp and nine putative STAT6 binding sites at −514 bp, −541 bp, −2249 bp, −2356 bp, −3703 bp, −3774 bp, −3829 bp, and −4299 bp within −5000 bp upstream (32
). We also found one modified consensus site for STAT6 (position is −4557 bp) that overlapped with the 5′-end of a putative NF-κB binding region −4552 bp). Future studies will be required to identify the functional sites among these.
Airway epithelial cells have been shown to express functional TLR, notably TLR2-6 (12
). Although TLRs share signal transduction pathways for activation of the transcription factors NF-κB and IRF-3 (47
), only dsRNA, a TLR3 ligand and mimic of viral RNA and viral replication, induced TSLP expression in normal human bronchial epithelial cells (). The failure of other TLR ligands tested to induce the expression of TSLP may reflect low levels of receptor expression or adaptor proteins involved in the respective responses.
While this manuscript was under revision, Lee and Ziegler reported that TSLP is not induced by dsRNA in NHBE (48
). Our data clearly show that TSLP mRNA was expressed only transiently after activation because it was not detected at 1 h, was highly elevated at 3 h, and then returned to baseline by 12 h after stimulation with dsRNA ( and ). The fact that Lee and Ziegler selected 2 and 20 h after stimulation may explain why elevations of TSLP mRNA by dsRNA were not detected in their studies (48
). Two separate groups have also recently shown that TSLP is induced by dsRNA in small airway epithelial cells and oral epithelial cells (49
). These results suggest that dsRNA-dependent TSLP production is reproducible in human epithelial cells.
It has been reported that dsRNA and IL-4 synergistically enhanced the production of the eosinophil-recruiting chemokines CCL11 and CCL26 (eotaxin-3) in airway smooth muscle cells and airway epithelial cells, respectively (51
). Enhancement of CCL26 by the combination of IL-4 and dsRNA was found to be a secondary effect of the induction of the IL-4 receptor by dsRNA in airway epithelial cells (52
). In the present study of TSLP expression, dsRNA and IL-4 stimulated NHBE rapidly (). In addition, enhancement of dsRNA-dependent TSLP expression by IL-4 was significantly inhibited by siRNA against STAT6 in NHBE (). These data suggest that dsRNA and IL-4 activate different transcription factors that are capable of binding the promoter of TSLP and activating its expression in airway epithelial cells, probably NF-κB, IRF-3 (dsRNA), and STAT6 (IL-4). Future studies will be required to identify the complement of transcription factors that activate the promoter of TSLP.
dsRNA and RNA viruses are recognized by the endosomal receptor TLR3 and also by the cytoplasmic RNA helicases RIG-I and MDA5 and the cytoplasmic serine-threonine kinase PKR (40
). Our data suggest that the induction of TSLP by dsRNA was mainly mediated by TLR3 (). As shown in , the induction of TSLP by dsRNA was weakly but not significantly suppressed by siRNA against MDA5 (27% inhibition) and RIG-I (23% inhibition). Future studies will be required to determine whether dsRNA-dependent TSLP production is partially regulated by MDA5 or RIG-I in airway epithelial cells.
The signal transduction pathway of TLR3 has been well studied (47
). Ligand-activated TLR3 recruits a Toll/IL-1 receptor (TIR) domain-containing adapter inducing IFN-β (TRIF) via its intracellular TIR domain. TRIF recruits a TGF-β-activated kinase-1 and a TANK-binding kinase 1 that phosphorylate IκB and IRF-3 respectively (47
). The phosphorylation of IκB represents a signal for polyubiquitination followed by degradation by the 26 S proteosome, and this enables the translocation of NF-κB into the nucleus and then induces proinflammatory genes such as TNF-α, IL-6, and IL-12 (47
). In contrast, phosphorylated IRF-3 dimerizes and translocates into the nucleus and induces IFN-β production. Our data clearly showed that the induction of TSLP by dsRNA was dependent on both NF-κB and IRF-3 (). Usually, IRF-3 critically regulates IFN-β production and then secondarily induces multiple IFN-regulated genes. In recent studies we found that induction of B cell-activating factor of TNF family (BAFF) by dsRNA occurred via an autocrine loop involving IFN-β(31
). In the case of TSLP, the induction by dsRNA was not suppressed by siRNA against the IFN receptor IFNAR2 and IFN-β did not strongly induce TSLP production in NHBE (, , and ), suggesting that dsRNA-induced TSLP production is independent of autocrine stimulation by IFN-β.
IRF-3 promotes transcription of the IFN-β gene together with other transcription factors such as NF-κB. Although the mechanisms of type I IFN induction by IRF-3 are well established, IFN independent IRF-3 activation is not fully understood. Recent studies suggest that IRF-3 directly induces the transcriptional suppresser Hes1 independently of IFN-β, which in turn inhibits the expression of the nuclear receptor RXRA (55
). Inhibition of RXRA by the activation of IRF-3 can reduce the expression of RXRA target genes including the metabolic enzymes CYP3A4 and UGT1A6 (55
). The studies of Li et al. (56
) indicate that keratinocyte-selective ablation of RXRA and retinoid×receptor βinduces TSLP expression in epidermal keratinocytes, which results in an atopic dermatitis-like syndrome. We therefore investigated whether IRF-3-dependent TSLP expression by dsRNA resulted via the induction of Hes1 and the inhibition of RXRA. Hes1 was not induced by dsRNA in NHBE and RXRA was only weakly suppressed (data not shown). Together, these results suggest that TSLP is induced by dsRNA in airway epithelial cells and that this response occurs via a pathway dependent on NF-κB and IRF-3 but independent of IFN-β and RXRA.
We examined whether glucocorticoids can inhibit IL-4- and dsRNA-dependent TSLP production, because glucocorticoids are widely used in the therapeutic management of inflammatory airway diseases. In vivo studies in human subjects have demonstrated that glucocorticoids inhibit the production of IL-4 and IL-5 in cells from bronchoalveolar lavage in allergic asthmatics and reduce the number of IL-4- and IL-13-expressing cells in nasal tissue from patients with allergic rhinitis (57
). As shown in , the potent topical glucocorticoid FP partially but significantly inhibited the induction of both TSLP mRNA and TSLP protein by dsRNA and IL-4 in NHBE. This suggests that glucocorticoids may inhibit both the production of Th2 cytokines from cells in the airways and the TSLP response of airway epithelial cells to these cytokines. By such effects, glucocorticoids could blunt the differentiation of Th2 cells in allergic airway disease. Future clinical studies will be required to investigate whether glucocorticoids suppress TSLP production in vivo.
The mechanisms of the anti-inflammatory effects of glucocorticoids have been extensively studied (42
). The activated glucocorticoid receptor dimerizes and interacts with transcription factors such as NF-κB and AP-1 and represses the genes activated by these transcription factors. Glucocorticoids also exert posttranscriptional control on gene expression by decreasing the stability of mRNA for inflammatory genes. Our previous studies have suggested that NF-κB is not a major target of glucocorticoids in human airway epithelial cells (36
). Possible mechanisms of the effect of glucocorticoids on the expression of TSLP may include the influence of destabilizing sequences known as AU-rich elements in the 3′ untranslated region of the transcripts. In fact, the 3′ untranslated of TSLP contains seven AUUUA motifs. Future studies will be required to determine the mechanism of inhibition of TSLP by glucocorticoids in airway epithelial cells.
Although asthma is characterized by Th2-type inflammation, the normal CD4+
T cell response to viral infection is thought to be predominantly of the Th1 type. It has been suggested that in the lower airways of asthmatics with a pre-existing Th2-type allergic microenvironment the responses to viral infection may be skewed toward inappropriate and potentially harmful Th2 responses (5
). Indeed, respiratory syncytial virus infection exacerbated the Th2 cytokine response and lung pathologic lesions in a mouse model of allergic asthma, whereas respiratory syncytial virus infection of nonallergic mice did not induce Th2 cytokine response significantly (61
). Respiratory syncytial virus infection enhanced the pulmonary Th2 cytokine response only when mice were inoculated after sensitization to OVA (62
). Th2 cytokines such as IL-4, IL-5, IL-10, and IL-13 enhanced the expression of ICAM-1, which serves as the rhinovirus entry receptor and has been proposed to be involved in respiratory syncytial virus entry (63
). Basal levels of ICAM-1 expression were increased in the nasal epithelial cells from atopic subjects and the airway epithelial cells from asthmatic subjects (65
). These data suggest that susceptibility to virus infection is increased in a Th2 environment. Our studies suggest that virus infection may further enhance the development of Th2 responses, creating a vicious cycle. We found that the Th2 cytokine IL-4 and the virus product dsRNA synergistically enhanced TSLP production from epithelial cells (). Importantly, we also found that the rhinovirus, which is the most common known trigger of asthma exacerbations, induced TSLP expression and production in the presence of IL-4 in airway epithelial cells (). It is possible that the TSLP production triggered by the rhinovirus could be higher in the asthmatic airway compared with NHBE due to the elevation of ICAM-1 (the receptor for rhinovirus) and the reduction of the antiviral response in epithelial cells from patients with asthma (65
). TSLP-activated DC express OX40L and promote the differentiation of TNF-α-producing inflammatory Th2 cells from naive T cells (24
). In addition, TLR3 activation strongly induced eosinophil-recruiting chemokines in the presence of Th2 cytokines (51
). Together, these findings suggest that viral infection in a Th2 environment may amplify allergic reactions locally. The effect of in vivo infection with respiratory viruses such as rhinovirus on the production of TSLP in airway epithelial cells from patients with asthma is worthy of investigation.
In summary, we report in this study that the expression of TSLP was induced in airway epithelial cells by stimulation with TLR3 ligand and Th2 cytokines and was synergistically induced by the combination of both stimuli; this response was suppressed by glucocorticoid treatment. Our findings indicate that respiratory viral infection and the recruitment of Th2 cytokine-producing cells may amplify Th2 inflammation via the production of TSLP in the asthmatic airway.