|Home | About | Journals | Submit | Contact Us | Français|
Tumor necrosis factor (TNF), an important mediator of inflammatory and innate immune responses, can be regulated by binding to soluble TNF receptors. The type 1, 55-kDa TNF receptor (TNFR1), the key receptor for TNF signaling, is released to the extracellular space by two mechanisms, the inducible cleavage and shedding of 34-kDa soluble TNFR1 ectodomains (sTNFR1) and the constitutive release of full-length 55-kDa TNFR1 within exosome-like vesicles. The aim of this study was to identify and characterize Toll-like receptor (TLR) signaling pathways that mediate TNFR1 release to the extracellular space. We for the first time demonstrate that poly (I:C), a synthetic double-strand RNA (dsRNA) analog that signals via TLR3, induces sTNFR1 shedding from human airway epithelial (NCI-H292) cells, whereas ligands for other microbial pattern recognition receptors, including TLR4, TLR7 and NOD2, do not. Furthermore, poly (I:C) selectively induces the cleavage of 34-kDa soluble TNFR1 ectodomains, but does not enhance the release of full-length 55-kDa TNFR1 within exosome-like vesicles. RNA interference experiments demonstrated that poly (I:C)-induced sTNFR1 shedding is mediated via activation of TLR3-TRIF-RIP1 signaling, with subsequent activation of two downstream pathways. One pathway involves the Duox2-mediated generation of reactive oxygen species (ROS), while the other pathway is via the caspase-mediated activation of apoptosis. Thus, the ability of dsRNA to induce the cleavage and shedding of the 34-kDa sTNFR1 from human bronchial epithelial cells represents a novel mechanism by which innate immune responses to viral infections are modulated.
Tumor necrosis factor (TNF) is an important pro-inflammatory cytokine that plays a key role in regulating inflammation, innate immunity, host defense, and cellular proliferation, differentiation, and apoptosis(1–6). TNF has also been implicated in the pathogenesis of several human diseases including septic shock, AIDS, rheumatoid arthritis, diabetes, multiple sclerosis, inflammatory bowel disease, cancer and cardiovascular disorders, as well as inflammatory lung diseases, such as asthma, chronic obstructive pulmonary disease and sarcoidosis (7–15). Furthermore, anti-TNF therapies, such as soluble receptors and neutralizing antibodies, have entered clinical practice for the treatment of inflammatory disorders, such as rheumatoid arthritis and other inflammatory arthritides. TNF, which is produced mainly by activated macrophages or monocytes, exerts its effects by binding to two cell surface receptors, the 55-kDa TNFR1 (CD120a) and the 75-kDa TNFR2 (CD120b) (8, 16). Binding of TNF to TNFR1, the key receptor for TNF signaling, leads to the association of TNFR1 with TNF receptor-associated death domain (TRADD), receptor-interacting protein (RIP), and TNFR-associated factor2 (TRAF2), with resultant activation of MAP kinase and NF-κB signaling pathways (8, 17–20). Alternatively, recruitment of Fas-associated death domain (FADD) by TRADD results in caspase activation and subsequent apoptosis (19).
Soluble tumor necrosis factor receptors (sTNFR) modulate TNF bioactivity by binding to and sequestering TNF. The type 1 TNF receptor (TNFR1) is generated by either the inducible cleavage and shedding of soluble 34-kDa TNFR1 ectodomains or the constitutive release of full-length 55-kDa TNFR1 within exosome-like vesicles (21–25). TNF-α-converting enzyme (TACE, ADAM17), a member of the ADAM (a disintegrin and metalloprotease) family of zinc metalloproteases, has been identified as a TNFR1 sheddase (26, 27). In addition, a complex comprised of aminopeptidase regulator of TNFR1 shedding (ARTS-1) and nucleobindin 2 (NUCB2) has been identified as TNFR1-binding proteins that regulate both the inducible cleavage of TNFR1 ectodomains and the constitutive release of TNFR1 exosome-like vesicles to the extracellular compartment (28, 29). The constitutive release of TNFR1 exosome-like vesicles is further regulated by BIG2, a brefeldin A-inhibited guanine nucleotide-exchange protein that activates class I ADP-ribosylation factors (30).
Several stimuli including phorbol ester, interleukin-1β, proteasome inhibitors and Staphylococcus aureus protein A can induce the proteolytic cleavage and shedding of soluble TNFR1 (27, 31–33). The goal of this study was to assess whether ligands for Toll-like receptors (TLR) can induce shedding of sTNFR1 as a mechanism by which innate immune responses can be modulated. The TLRs, comprising more than 13 family members in mammals, are pattern-recognition receptors (PRRs) that recognize restricted sets of microbial ligands (34). TLRs are comprised of extracellular leucine-rich repeats that mediate TLR ligand binding, a transmembrane domain, and an intracellular Toll/IL-1 receptor (TIR) domain that mediates signaling. Five essential TIR-domain-containing cytosolic adapters, including myeloid differentiation primary response protein 88 (MyD88), TIR domain-containing adapter protein (TIRAP or Mal), TIR domain-containing adapter inducing interferon-β (Trif or TICAM1), Trif-related adaptor molecule (TRAM or TICAM2) and sterile-alpha and Armadillo motif protein (SARM), have been identified as key TLR signaling proteins (34–37). Upon ligand binding, TLRs activate mitogen-activated protein kinase (MAPK) and NF-κB signaling pathways, with resultant induction of inflammatory cytokines and type I interferon (IFN), dendritic cell maturation and natural killer cell activation (38–40). Therefore, TLRs play an important role in innate immune responses, as well as the resultant development of adaptive immunity.
Here, we sought to identify and characterize Toll-like receptor (TLR) signaling pathways that mediate TNFR1 release to the extracellular space. We show that poly (I:C), a viral double-stranded RNA (dsRNA) analog, selectively induces the cleavage and shedding of 34-kDa soluble TNFR1 ectodomains, but does not enhance the release of full-length 55-kDa TNFR1 within exosome-like vesicles from human airway epithelial cells. The poly (I:C)-mediated increases in sTNFR1 shedding are mediated via TLR3, whereas ligands for other toll-like receptors, including TLR4, TLR7 and NOD2, do not. Furthermore, we show that poly (I:C)-induced sTNFR1 release is mediated via two TLR3-TRIF-RIP1-dependent pathways. One pathway involves the Duox2-mediated generation of reactive oxygen species (ROS), while the second pathway is via caspase-mediated activation of apoptosis. We conclude that viral dsRNA-induced shedding of 34-kDa sTNFR1 ectodomains from human bronchial epithelial cells represents a novel mechanism by which innate immune responses to viral infections are modulated.
Polyinosinic-polycytidilic acid (poly (I:C)), muramyl dipeptide (MDP), imiquimod (R837) and ultrapure LPS (E. coli 0111:B4) were from InvivoGen (San Diego, CA). Penicillin/streptomycin was from Invitrogen (Carlsbad, CA). N-acetyl-L-cysteine (NAC), diphenyleneiodonium chloride (DPI), SB203580, PD98059 and cycloheximide (CHX, InSolution™) were from EMD Biosciences (San Diego, CA). The general caspase inhibitor z-Val-Ala-Asp(OMe)-fluoromethyl ketone (z-VAD-fmk) and the neutralizing human transforming growth factor-α (TGF-α) antibody were from R&D Systems (Minneapolis, MN). Mouse anti-human TNFR1 and β-tubulin antibodies, the goat anti-TACE antibody and the purified glutathione-S-transferase (GST) protein were from Santa Cruz Biotechnology (Santa Cruz, CA). The antibody against β-actin was purchased from Sigma (St. Louis, MO). Antibodies to MEK1/2, ERK1/2, JNK, AKT, phospho-AKT (Ser473), epidermal growth factor receptor (EGFR), phospho-EGFR (Tyr1173), RIP and poly-ADP ribose polymerase (PARP) were from Cell Signaling Technology (Danvers, MA). The anti-GST antibody was purchased from Pierce (Rockford, IL), while the recombinant human EGF was from Promega (Madison, WI).
The NCI-H292 human airway epithelial cell line was from the American Type Culture Collection (Manassas, VA). NCI-H292 cells were cultured in RPMI-1640 with L-glutamine (Mediatech, Manassas, VA) supplemented with 10% fetal bovine serum (Cambrex, Walkersville, MD) and 1x penicillin/streptomycin (100 U/mL penicillin, 100 μg/mL streptomycin, Invitrogen) and grown in a humidified incubator with 5% CO2 at 37°C. The primary human small airway epithelial cells (SAEC) were cultured in SABM medium supplemented with SAGM SingleQuots (Lonza, Walkersville, MD).
Confluent cultures of NCI-H292 cells were pre-incubated with DMSO or the indicated inhibitors for 30 minutes prior to the addition of poly (I:C) (25 μg/ml) for 2 h at 37 °C. For siRNA experiments, cells were transfected with siRNA duplexes for 3 days prior to stimulation with poly (I:C) for 2 h. Conditioned media were collected and cleared of cells and debris by centrifugation (10 min, 14,000 × g, 4 °C). The quantity of sTNFR1, IL-8 and TGF-α in culture medium was determined using a Quantikine sandwich ELISA kit (R&D Systems, Minneapolis, MN). To examine the effects of MEK and AKT on sTNFR1 shedding, NCI-H292 cells were grown to 50% to 70% confluence prior to transfection with dominant negative (DN) mutants of MEK K97R (a generous gift from Dr K.L. Guan at the University of Michigan), Akt1 (Millipore, Billerica, MA) or control vector using Fugene 6 reagent (Roche, Indianapolis, IN). Twenty-five hours later, the cells were treated with or without poly (I:C) for 2 hours and cell lysates were assessed for the expression of DN plasmids by Western blotting.
NCI-H292 cells were lysed in cold buffer containing 1% Triton X-100, 1% n-octyl-beta-D-glucopyranoside, 50 mM Tris, pH 7.5, 150 mM NaCl (Sigma, St. Louis, MO) and fresh Complete™ protease inhibitor (Roche Applied Science, Indianapolis, IN). For detection of phosphorylated proteins, cells were washed in 1X TBS buffer and lysed in buffer containing phosSTOP (Roche Applied Science). After centrifugation of cell lysates (5 min, 14,000 × g, 4 °C) to remove cellular debris, protein concentrations were determined using a BCA kit (Pierce, Rockford, IL). Cellular proteins were separated by SDS-PAGE using 4–12% Bis-Tris Nupage gels (Invitrogen, Carlsbad, CA), electroblotted onto nitrocellulose membranes and reacted with appropriate antibodies as previously described (28). Following stimulation with poly (I:C), cell culture medium was spiked with GST (1 μg) to serve as a control for protein loading and precipitated with 20% trichloroacetic acid prior to Western blotting. For repeated blotting, membranes were stripped using the Re-Blot Western blot recycling kit (Chemicon International, Temecula, CA) and re-probed with indicated antibodies. Signals were detected by chemiluminescence (Super Signal, Pierce, Rockford, IL).
NCI-H292 cells were cultured in six-well plates overnight and transfected with siRNA (5 to 100 nM) for 6 h using LipofectamineTm RNAiMAX (Invitrogen). Cells were cultured in complete medium for 72 h before treatment with poly (I:C). The ON-TARGETplus siRNA duplexes SMARTpools specific for TLR3, TRIF, RIP1 and TACE, the ON-TARGETplus siRNA for Duox1 (target sequence: GCUAUCACGUGCUUUCAGA) and Duox2 (target sequence: GAGGAUAAGUCCCGUCUAA), and the non-targeting control siRNA duplexes were from Dharmacon (Lafayette, CO).
Total RNA was isolated using the RNeasy Mini Kit and QIAshredder (Qiagen, Valencia, CA). cDNA templates were prepared with high capacity cDNA RT kit (Applied Biosystems, Foster City, CA). Quantitative real-time RT-PCR was performed for 40 cycles using 7500 real time PCR system, TaqManR Gene Expression Assays and PCR master mix (Applied Biosystems, Foster City, CA). Gene expression was quantified relative to expression of 18S rRNA using the control sample as calibrator to calculate the difference in Ct values (ΔΔCt) and presented as relative mRNA expression.
Confluent NCI-H292 cells were loaded with 10 μM CM-H2DCFDA (Invitrogen, Carlsbad, CA) in culture medium (RPMI 1640). After 30 min, cells were trypsinized, washed twice in HBSS and dispersed in 50 uL aliquots into 96-well plates at a concentration of 106 cells/ml. Cells were pre-treated with either N-acetyl-L-cysteine (NAC, 2 mM) or diphenyleneiodonium chloride (DPI, 3 μM) for 5 min. Cells were stimulated by poly (I:C) and intracellular hydrogen peroxide production was measured using a fluorescent plate reader (excitation and emission wavelengths, 485 and 538 nm, respectively) (Fluoroskan, Labsystems). Increases in fluorescence at 5 minutes were calculated and background fluorescence of the cell-free medium was subtracted.
Data are presented as mean ± SEM. A two-tailed Student’s t test or one-way ANOVA was used to determine statistically significant differences. A P value less than 0.05 was considered significant.
To examine the role of TLRs in sTNFR1 shedding, the NCI-H292 human airway epithelial cell line was stimulated with ligands for TLR3, TLR4, TLR7, and NOD2 (nucleotide-binding oligomerization domain containing 2) for 2 h. The synthetic dsRNA analog poly (I:C) induced a significant increase in sTNFR1 release in a dose-dependent manner, whereas other TLR ligands did not (Fig. 1a and 1b). The ability of poly (I:C), but not other TLR ligands, to induce sTNFR1 was confirmed using primary cultures of human small airway epithelial cells (Fig. 1c). Similarly, poly (I:C), but not other TLR ligands, induced IL-8 release from NCI-H292 cells (Supplemental Figure 1). Western blotting demonstrated that poly (I:C) induced the release of a 34-kDa TNFR1 into the culture medium, which is consistent with a cleaved, soluble TNFR1 ectodomain, while a full-length, 55-kDa TNFR1 indicative of an exosome-like vesicle was not detected (Fig. 1d and 1e). Poly (I:C)-induced sTNFR1 was not inhibited by cycloheximide (CHX), which shows that TNFR1 shedding does not require de novo protein synthesis (Fig. 1f). These data demonstrate that poly (I:C) induces the shedding of soluble TNFR1 from human airway epithelial cells.
Since TLR3 is the major receptor for double-stranded viral RNA, we assessed its role in poly (I:C)-induced sTNFR1 shedding (41). As shown in Fig. 2a, RNA interference (RNAi) was utilized to specifically reduce TLR3 mRNA levels, whereas mRNA levels of TNFR1 and TACE were not affected. The RNAi-mediated knockdown of TLR3 mRNA expression significantly abrogated poly (I:C)-induced sTNFR1 release by 86% (Fig. 2b). The RNAi-mediated knockdown of TRIF, which is the critical adaptor protein for TLR3, also significantly reduced poly (I:C)-mediated sTNFR1 release by 90% (Fig. 2c and 2d). These data demonstrate that signaling via TLR3 and TRIF are essential for poly (I:C)-induced sTNFR1 release. Receptor interacting protein 1 (RIP1) is an important adaptor signaling protein that is downstream of TRIF. The RNAi-mediated knockdown of RIP1 significantly reduced poly (I:C)-induced sTNFR1 shedding by 45% (Fig. 2e and 2f). Taken together, these data demonstrate that poly (I:C)-induced sTNFR1 shedding from NCI-H292 human airway epithelial cells is mediated by a TLR3-TRIF-dependent pathway that involves the downstream participation of RIP1.
MAP kinase and Akt signaling pathways have been reported to be activated downstream of TLR3. As shown in Figure 3a, treatment of NCI-H292 cells with siRNA targeting RIP1 did not inhibit phosphorylation of MEK1/2 or ERK1/2. In contrast, the RNAi-mediated knockdown of RIP1 reduced the poly (I:C)-mediated phosphorylation of Akt (Fig. 3b). Transient expression of either a dominant-negative MEK mutant or a dominant-negative Akt1 mutant (Fig. 3c) did not inhibit sTNFR1 shedding (Fig. 3d and 3e). Consistent with these findings, neither the ERK inhibitor, PD98059, nor the p38 inhibitor, SB203580, inhibited poly (I:C)-mediated sTNFR1 shedding (Fig. 3f). Lastly, poly (I:C) stimulation did not induce phosphorylation of JNK, which is consistent with the conclusion poly (I:C)-induced sTNFR1 shedding occurs independently of JNK signaling pathways (Supplemental Fig. 2). These data demonstrate that poly (I:C)-induced TNFR1 shedding is not mediated via MEK-, ERK-, p38- or Akt-dependent pathways.
TLR signaling can induce the production of reactive oxygen species (ROS), which can function as downstream intracellular signaling molecules (42, 43). Furthermore, ROS can activate TACE-mediated TNFR1 cleavage and shedding (26, 27). Consequently, experiments were performed to assess the role of ROS signaling in poly (I:C)-induced sTNFR1 shedding. As shown in Figure 4a, poly (I:C) induced oxidation of CM-H2DCFDA indicative of ROS production in NCI-H292 cells, which could be inhibited by pre-treatment with the antioxidant, N-acetyl-L-cysteine (NAC), or the NADPH oxidase/flavoprotein inhibitor, diphenyleneiodonium chloride (DPI). Having shown that poly (I:C) induces ROS generation, we next demonstrated that NAC or DPI reduced poly (I:C)-induced TNFR1 shedding by more than 40% (Fig. 4b), which demonstrates a role for ROS in this process. We surmised that if ROS-mediated TNFR1 shedding is downstream of poly (I:C)-induced TLR3-TRIF-RIP1 signaling, then there should be no additive effect of combined ROS inhibition and RNAi-mediated RIP1 knockdown as compared to ROS inhibition alone. Consistent with this, treatment of NCI-H292 cells with NAC and siRNA targeting RIP1 did not show additive suppression of poly (I:C)-induced TNFR1 shedding when compared to cells treated with NAC and control siRNA (Fig. 4c). Dual oxidase 1 (Duox1) and Dual oxidase 2 (Duox2), members of NADPH oxidase (Nox) family, are highly expressed in airway epithelial cells and therefore are candidates to mediate poly (I:C)-induced ROS generation (44, 45). As shown in Fig. 4d to 4g, the RNAi-mediated knockdown of Duox2, but not Duox1, inhibited poly (I:C)-induced TNFR1 shedding by 30%. Furthermore, the RNAi-mediated knockdown of both Duox1 and Duox2 did not have an additive effect on inhibition of sTNFR1 release as compared to the RNAi-mediated knockdown of Duox2 alone (Fig. 4h). These data are consistent with the conclusion that Duox2-mediated ROS production contributes to sTNFR1 shedding downstream of poly (I:C)-mediated activation of a TLR3-TRIF-RIP1 signaling pathway.
Synthetic dsRNA can induce cancer cell apoptosis in a TLR3-dependent manner, while activation of apoptosis initiates sTNFR1 shedding (46). Therefore, we assessed whether poly (I:C)-induced sTNFR1 release from NCI-H292 cells is dependent upon caspase activation. As shown in Fig. 5a, the 89-kDa cleavage form of poly-ADP ribose polymerase (PARP), an early marker of caspase-dependent apoptosis, was detected in poly (I:C) treated cells but not in cells treated with medium alone, which confirms that poly (I:C) can activate cellular apoptosis in NCI-H292 cells. z-VAD-fmk, a pan-caspase inhibitor, inhibited PARP cleavage (Fig. 5a) and reduced sTNFR1 release by approximately 27% (Fig. 5b). As shown in Fig. 5c, poly (I:C)-induced PARP cleavage was reduced in cells treated with siRNA targeting RIP1. Furthermore, the combination of NAC and z-VAD-fmk had an additive inhibitory effect on sTNFR1 release as compared to z-VAD-fmk alone (Fig. 5d), which suggests that separate pathways involving ROS generation and cellular apoptosis mediate poly (I:C)-induced sTNFR1 shedding. Taken together, these data are consistent with a role for RIP1-dependent caspase activation in poly (I:C)-mediated sTNFR1 release.
TNF-α converting enzyme (TACE, ADAM17) has been identified as a TNFR1 sheddase that can be activated by EGFR signaling pathways in both a TGF-α-dependent and TGF-α–independent manner (27, 47). Therefore, we assessed the roles of TACE and TGF-αmediated EGFR activation in poly (I:C)-mediated sTNFR1 release. As shown in Fig. 6a to 6c, the RNAi-mediated knockdown of TACE mRNA expression reduced poly (I:C)-induced sTNFR1 release by 50%. In contrast, poly (I:C)-induced TGF-α release was completely inhibited by the RNAi-mediated TACE knockdown. Since poly (I:C) stimulation also induced TGF-α release, we next assessed whether TGF-α-induced EGFR activation mediates poly (I:C)-induced sTNFR1 shedding. As shown in Figure 6d, treatment of cells with an anti-TGF-α neutralizing antibody did not suppress sTNFR1 release. In addition, poly (I:C) did not induce EGFR phosphorylation, whereas EGF did (Fig. 6e). This suggests that the small quantity of TGF-α released in response to poly (I:C) stimulation was not sufficient to induce phosphorylation and activation of the EGFR. Taken together, these data demonstrate a role for TACE in poly (I:C)-induced sTNFR1 shedding that does not require TGF-α release or EGFR activation.
Soluble TNF receptors play an important role in regulating inflammatory events by binding to TNF. Consistent with this, knock-in mice expressing a mutated non-sheddable TNFR1 develop an immune hyperreactivity phenotype manifested by spontaneous hepatitis, enhanced susceptibility to endotoxic shock, exacerbated TNF-dependent arthritis, and experimental autoimmune encephalomyelitis (48). A variety of mediators have been identified that can induce sTNFR1 shedding, including bacterial products, such as Staphylococcus aureus protein A (27, 49). It is not known, however, whether microbial pathogens can induce sTNFR1 shedding via activation of pattern recognition receptors as a mechanism by which the innate immune system responds to infection. Here, we hypothesized that Toll-like receptor (TLR) ligands and downstream signaling pathways induce the release of either soluble 34-kDa sTNFR1 ectodomains or full-length 55-kDa TNFR1 within exosome-like vesicles to the extracellular space. We show that poly (I:C), a viral dsRNA analog, selectively induces cleavage and shedding of soluble 34-kDa sTNFR1 from human airway epithelial cells, but does not enhance the release of full-length 55-kDa TNFR1 within exosome-like vesicles. Poly (I:C)-induced sTNFR1 shedding is mediated via a TLR3, whereas ligands that activate other toll-like receptors, such as TLR4, TLR7, and NOD2, do not. Furthermore, we show that poly (I:C)-induced sTNFR1 release is mediated by at least two TLR3-TRIF-RIP1-dependent pathways. One pathway involves the Duox2-mediated generation of ROS, whereas the other pathway is via caspase-dependent apoptosis. This demonstrates that sTNFR1 shedding is part of the airway epithelial innate immune response to viral infection that can attenuate excessive TNF-mediated inflammation.
TLRs are essential for both innate and adaptive immune responses by serving as highly conserved pattern-recognition receptors that bind a variety of endogenous and exogenous stimuli (34). TLR3 is the major receptor for double-stranded viral RNA (41). Upon TLR3 ligation, TRIF, the critical adaptor protein for TLR3, activates downstream signaling through receptor interacting protein 1 (RIP1) (34, 50–52). Having shown that poly (I:C) induces sTNFR1 shedding via a TLR3-TRIF-dependent pathway, we investigated the relevant downstream signaling pathways. Binding of TLR3 to its ligands can activate downstream MAPK signaling pathways (34, 52). Moreover, ERK can activate TACE, while binding of Staphylococcus aureus-derived protein A to EGFR leads to TACE phosphorylation in a c-Src-ERK1/2-dependent manner (27). Here, we show that the poly (I:C)-induced phosphorylation of MEK1/2, ERK1/2 and p38 in NCI-H292 cells is not dependent upon RIP1. Furthermore, expression of a dominant-negative MEK mutant or pharmacological inhibition of either ERK or p38 did not attenuate poly (I:C)-induced sTNFR1 shedding. These data are consistent with the conclusion that poly (I:C)-induced sTNFR1 shedding is partially mediated by RIP1, but independent of MEK, ERK and p38.
RIP1, which mediates TLR3 signaling downstream of TRIF, can modulate p27 levels via activation of a phosphoinositide 3-kinase (PI3K)-AKT pathway (53). In addition, dsRNA-mediated activation of interferon regulatory factor 3 (IRF3) and IFN-β is dependent upon PI3K activity (54). Consistent with RIP1-mediated activation of AKT, we show that the RNAi-mediated knockdown of RIP1 expression substantially reduced poly (I:C)-induced AKT phosphorylation. However, over-expression of a dominant-negative Akt1 mutant did not affect sTNFR1 shedding. These data indicate that poly (I:C)-induced sTNFR1 release is not dependent upon AKT signaling.
Next, we considered that poly (I:C)-mediated sTNFR1 release might also be mediated via the generation of reactive oxygen species (ROS), which can be induced in response to activation of various TLRs (42, 43, 55, 56). ROS signaling has also been shown to activate TACE and thereby induce TNFR1 ectodomain cleavage and shedding (26, 27, 57). Here, we show that poly (I:C) induces ROS production by NCI-H292 cells. The antioxidant, N-acetyl-L-cysteine (NAC), and the NADPH oxidase inhibitor, diphenyleneiodonium chloride (DPI), suppressed poly (I:C)-induced TNFR1 shedding by greater than 40%, which is consistent with a role for ROS signaling pathways in poly (I:C)-induced sTNFR1 shedding. Furthermore, poly (I:C)-mediated ROS generation appeared to be downstream of RIP1, consistent with activation of a TLR3-TRIF-RIP1 signaling pathway. NADPH oxidases (Nox) are the major source for generating ROS in airway epithelial cells; the Nox family is comprised of five NADPH oxidases, as well as Duox1 and Duox2, which are highly expressed by airway epithelial cells(44, 45). We used RNAi to knock-down Duox2 expression, which inhibited TNFR1 shedding by 30%, whereas knockdown of Duox1 had no effect. Our findings contrast with those of previous studies, which showed that Duox1 is required for TLR ligand-induced production of IL-8 and VEGF (58). Taken together, our data are consistent with the conclusion that Duox2-mediated ROS signaling participates in poly(I:C)-mediated sTNFR1 shedding downstream of RIP1. Interestingly, previous work showed that Duox2 mRNA and protein levels are induced in airway epithelial cells exposed to interferon-γ, poly (I:C), or rhinovirus, suggesting this oxidase functions specifically in antiviral responses (59, 60).
Activation of caspases, which are cysteine-dependent proteases, results in cleavage of downstream substrates and the induction of cellular apoptosis (61). Synthetic dsRNA can directly trigger TLR3-dependent apoptosis in human breast cancer cells (46), while poly (I:C) has been used as a therapeutic adjuvant for the treatment of neoplasia based upon its ability to induce apoptosis and IFN production, with resultant anti-tumor immune responses (62). Furthermore, apoptosis-inducing agents can promote sTNFR1 shedding in vascular endothelial cells as a mechanism to limit inflammation in the setting of apoptotic cell death (63). Here, we hypothesized that the poly (I:C) signaling pathway that induces sTNFR1 shedding in human airway epithelial cells may involve caspase activation and the initiation of apoptosis. Consistent with this concept, we detected cleavage of poly-ADP ribose polymerase (PARP), an early marker of caspase activation and apoptosis, in poly (I:C)-treated cells. Furthermore, z-VAD-fmk, a broad-spectrum caspase inhibitor, significantly reduced sTNFR1 shedding by 27%, which is consistent with a role for caspase activation in poly (I:C)-induced sTNFR1 shedding. We also assessed the role of RIP1 in poly (I:C)-induced caspase-mediated sTNFR1 shedding. RIP1 signaling modulates apoptosis, necrosis, and autophagy and thereby regulates cell fate (55, 64). We show that poly (I:C)-induced PARP cleavage is RIP1-dependent, which is consistent with the conclusion that caspase activation is downstream of RIP1. This shows that poly (I:C)-induced sTNFR1 shedding is partially mediated by RIP1-dependent induction of apoptosis.
Lastly, we assessed the role of TACE, which functions as a TNFR1 sheddase, in poly (I:C)-mediated sTNFR1 release (27, 47). The RNAi-mediated knockdown of TACE expression reduced poly (I:C)-induced sTNFR1 release by 50%, whereas TGF-α release was almost totally suppressed, which suggests that additional TNFR1 sheddases may participate in this process. ROS have been reported to activate TACE via a pathway involving ATP-induced TGF-α shedding and EGFR activation (47). Therefore, we investigated the roles of TGF-α and EGFR in poly (I:C)-induced sTNFR1 shedding. Although poly (I:C) induced the release of TGF-α from NCI-H292 cells, experiments utilizing a neutralizing anti-TGF-α antibody showed that TGF-α does not mediate poly (I:C)-induced sTNFR1 shedding. Similarly, poly (I:C) did not induce EGFR phosphorylation. These data show that poly (I:C)-induced sTNFR1 shedding occurs via TACE-dependent pathways that do not require TGF-α release or EGFR activation.
In summary, we have shown that the double-stranded viral RNA homologue, poly (I:C), selectively induces shedding of soluble 34-kDa sTNFR1 ectodomains from the NCI-H292 human airway epithelial cell line via a TLR3-TRIF-RIP1-dependent signaling pathway. Furthermore, poly (I:C)-induced sTNFR1 shedding involves the participation of at least two downstream pathways, one mediated by Duox2 and ROS generation and the other via caspase activation and the induction of apoptosis. However, additional pathways downstream of TLR3 and TRIF are likely involved in this process as RIP1-dependent ROS signaling and caspase activation only partially account for TLR3-TRIF-mediated sTNFR1 shedding in response to poly (I:C). In particular, additional TLR3 signaling proteins that do not involve RIP1 and might be involved include tumor necrosis factor receptor-associated factors 3 and 6 (TRAF3, TRAF6), interleukin-1 receptor-associated kinase 2 (IRAK2) and IκB kinases (52, 65–69). Furthermore, our results suggest that TLR3-mediated sTNFR1 shedding by airway epithelial cells represents a physiological response to infections caused by double-stranded RNA viruses. Thus, we have identified a mechanism by which signaling through the TLR3 pattern recognition receptor regulates the innate immune response to viral-induced airway inflammation through the shedding of 34-kDa soluble sTNFR1, which is then available to bind and potentially attenuate excessive TNF bioactivity.
This work was funded by the Division of Intramural Research, NHLBI, NIH.
We thank Drs. Joel Moss and Martha Vaughan for their very helpful discussions and review of the manuscript.
Publisher's Disclaimer: This is an author-produced version of a manuscript accepted for publication in The Journal of Immunology (The JI). The American Association of Immunologists, Inc. (AAI), publisher of The JI, holds the copyright to this manuscript. This version of the manuscript has not yet been copyedited or subjected to editorial proofreading by The JI; hence, it may differ from the final version published in The JI (online and in print). AAI (The JI) is not liable for errors or omissions in this author-produced version of the manuscript or in any version derived from it by the U.S. National Institutes of Health or any other third party. The final, citable version of record can be found at www.jimmunol.org.