TLRs play a critical role in early innate immune response to virus infection. TLR3 together with TLR7 and TLR8 constitute a powerful system to detect genetic material of RNA viruses. TLR3 has been shown to bind viral dsRNA whereas TLR7 and TLR8 are receptors for viral single-stranded RNA. In this report we show that TLR7 or TLR8 are not expressed in human epithelial A549 cells or in HUVECs. Accordingly, A549 cells and HUVECs were unresponsive to TLR7/8 ligand R848. TLR3 was expressed at a higher level in HUVECs than in A549 cells. The TLR3 ligand poly(I:C) up-regulated IFN-β, IL-28, IL-29, STAT1, and TLR3 expression in HUVECs but not in A549 cells. An enhanced TLR3 expression by transfection or by IFN-α stimulation conferred poly(I:C) responsiveness in A549 cells. Similarly, IFN-α pretreatment strongly enhanced poly(I:C)-induced activation of IFN-β, IL-28, and IL-29 genes also in HUVECs. In conclusion, our results suggest that IFN-α-induced up-regulation of TLR3 expression is involved in dsRNA activated antiviral response in human epithelial and endothelial cells.
Toll-like receptors play a crucial role in regulating the activation of innate immune response to invading microbes (1). These evolutionarily well-conserved receptors recognize structural motifs of microbes, called pathogen-associated microbial patterns (2). Pathogen-associated microbial patterns include bacterial cell wall components of both Gram-positive and Gram-negative bacteria, as well as flagellin, bacterial and viral DNA and viral dsRNA and ssRNA (1). Ten human TLRs have been characterized so far. They are predominantly expressed by cells involved in immune functions, like in splenic macrophages and peripheral blood leukocytes, as well as in tissues exposed to the external environment such as lungs and the gastrointestinal tract (1). TLRs are essential elements in host defense against most microbes by activating innate immune response which is a prerequisite for the induction of the adaptive immunity (3).
Many RNA viruses express dsRNA at some point during their replication cycle, and dsRNA is involved in the activation of innate immunity during virus infection. TLR3 is a receptor for viral dsRNA and binding of dsRNA to its receptor activates IFN-α production (4). However, viruses have evolved strategies to sequester dsRNA to avoid the activation of antiviral pathways (5). In addition to dsRNA, also ssRNA is a molecular pattern recognized by TLRs. It has been shown that viral ssRNA is recognized by murine TLR7 and human TLR8 which activates innate immune response (6, 7, 8). At least ssRNA from HIV and influenza virus is recognized by TLR7/8. Therefore, TLR7/8 together with TLR3 constitute a powerful system to detect the genetic material of RNA viruses.
The key cytokine that regulates innate immune responses against viruses is IFN-α (9). IFN-α has antiviral and immunoregulatory functions. IFN-α serves as an important link between innate and adaptive immunity (10). It activates dendritic cells (DCs),3 NK cells, and macrophages. In synergy with IL-18, IFN-α enhances IFN-γ production in NK and T cells (11, 12, 13). IFN-γ in turn stimulates Ag presentation in DCs and macrophages. This results in the activation of adaptive immunity and the elimination of virus-infected cells by CTLs. Recently, a new family of type I IFN-like cytokines has been characterized (14, 15). These novel cytokines include IL-28A and B and IL-29 (also known as IFN-λ2/3 and IFN-λ1, respectively). Like IFN-α and IFN-β, IL-28 and IL-29 have antiviral activities (14, 15) but their immunoregulatory functions have not been characterized so far.
In addition to immune cells, epithelial and endothelial cells are important primary targets of virus infection. Both epithelial and vascular endothelial cells are highly susceptible to viruses that cause a systemic infection such as measles virus, and the spread of infection occurs via endothelial cells. Lung epithelial cells are primary targets of influenza virus infection, and they produce immunoregulatory cytokines, including IFN-β, in response to the infection (16). In this report we show that epithelial and endothelial cells do not express TLR7 or TLR8, which are the receptors for viral ssRNA. In contrast, TLR3 which is a receptor for dsRNA is expressed by these cells. Furthermore, TLR3 protein expression is strongly up-regulated by IFN-α resulting in an enhanced response to viral RNA.
Materials and Methods
A549 human lung carcinoma cells (American Type Culture Collection; ATCC CCL185) were maintained in Eagle’s MEM supplemented with 0.6 μg/ml penicillin, 60 μg/ml streptomycin, 2 mM l-glutamine, and 10% heat-inactivated FCS (Integro). HUVECs (ATCC CRL1730) and HEK293 cells (ATCC CRL1573) were maintained in RPMI 1640 medium (Sigma-Aldrich) with supplements described above.
Cytokines and TLR ligands
Human leukocyte IFN-α was provided by the Finnish Red Cross Blood Transfusion Service and was used at 100 IU/ml. IFN-β (Schering-Plough) was used at 100 IU/ml. TLR3 ligand poly(I:C) and TLR7/8 ligand R848 was purchased from Sigma-Aldrich and InVivoGen, respectively.
Cloning of TLR3 transfection vector and preparation of anti-TLR3 Abs
TLR3 cDNA, kindly provided by Dr. Kastelein (DNAX, Palo Alto, CA), was TA− cloned to pcDNA3.1/CT-GFP-TOPO vector (Invitrogen) followed by cloning to pcDNA3.1-FLAG vector (Sigma-Aldrich). The pcDNA3.1-TLR3-FLAG construct was transfected to HEK293 cells and TLR3-Flag expression was confirmed with immunoblotting with anti-FLAG mAb M2 (Sigma-Aldrich). Cell lysates prepared from TLR3-Flag-transfected cells were used to analyze rabbit polyclonal anti-TLR3 Abs raised against the TLR3 peptide SIQKIKNNPFVKQKNLIT (anti-TLR3-P Ab) or guinea pig anti-TLR3 Abs raised against an intracellular part of the TLR3 (anti-TLR3-IC Ab). Intracellular TLR3 was expressed in Escherichia coli B strain BL21(DE3) as a glutathione S-transferase fusion protein and purified by a preparative SDS-PAGE (Prep-Cell; Bio-Rad). pcDNA3.1-TLR3-FLAG construct was transfected to A549 and HEK293 cells using FuGENETM 6 transfection reagent (Roche Molecular Biochemicals).
Western blotting and immunoprecipitation
To analyze TLR3 expression levels in A549 cells or in HUVECs, the cells were left unstimulated or stimulated as indicated, after which the cells were collected and lysed in PBS containing 1% Nonidet P-40 (BDH Laboratories), 0.1% SDS, a complete protease inhibitor mixture (Roche), and 10 mM EDTA for 30 min on ice. After this the cell lysates were centrifuged for 20 min at 20,000 × g at +4°C. STAT1 expression was studied directly from cell lysates whereas endogenous TLR3 expression was analyzed by immunoprecipitation followed by Western blotting. Primary rabbit anti-TLR3 Abs (anti-TLR3-P Ab) were added to cell lysates and incubated for 20 h at +4°C. After this protein-A-Sepharose beads were added (Pharmacia) and after 2 h of continuous mixing, the beads were washed several times with cell lysis buffer. The immunoprecipitates were suspended into SDS-PAGE sample buffer, boiled, and run in SDS-PAGE gels under reducing or nonreducing conditions and transferred onto nitrocellulose membranes. To control equal loading, the membranes were stained with Ponceau-S after which they were immunoblotted with guinea pig anti-TLR3 Abs (TLR3-IC Ab). Proteins were visualized by the SuperSignal West Pico Chemiluminescent system (Pierce). Mouse anti-STAT1 mAb was purchased from Santa Cruz Biotechnology. Peroxidase-conjugated goat anti-guinea pig Ab and peroxidase-conjugated rabbit anti-mouse Ab were purchased from Jackson ImmunoResearch Laboratories. Anti-STAT5 Abs were from Santa Cruz Biotechnology.
RNA isolation and Northern blot analysis
Total cellular RNA was isolated by RNeasy kit (Qiagen) according to the manufacturer’s instructions. Samples containing equal amounts of RNA (10 μg) were size fractioned on 1% formaldehyde-agarose gels, transferred to nylon membranes (Hybond: Amersham, Buckinghamshire, U.K.) and hybridized with the IFN-β (16), IL-28, and IL-29 probes. Probes for IL-28 and IL-29 were cloned from total cellular RNA obtained from Sendai virus-infected macrophages by RT-PCR using oligonucleotides GTCTCCACAGGATCCGCAGGCCTT and CAGCCAGGGGGATCCTTTTTTGGG (IL-28), GAAGGCCAGGGATCCCTTGGAAGA and GTGTCAGGTGGATCCAGGGTGGGT (IL-29). The probes for TLR3, TLR7, and TLR8 have been previously described (17). Ethidium bromide staining of rRNA bands was used to ensure equal RNA loading. The probes were labeled with [α-32P]dATP (3000 Ci/mmol; Amersham), using a random primed DNA labeling kit (Boehringer Mannheim). The membranes were hybridized (Ultrahyb; Ambion, Austin, TX) and washed in 1× SSC/0.1% SDS and exposed to Kodak AR X-omat films at −70°C using intensifying screens.
Nuclear extracts and nuclear protein/DNA-binding reactions were performed as described previously (18, 19). Consensus NF-κB-binding oligonucleotides (5′-GATCAGTTGAGGGGACTTTCCCAGCC-3′) were purchased from DNA Technology A/S. The probes were labeled by Klenow fill-in reaction and the binding reaction was done at room temperature for 30 min. The samples were analyzed by electrophoresis in 6% nondenaturing low-ionic strength polyacrylamide gels in 0.25× TBE. The gels were dried and visualized by autoradiography.
TLR3, TLR7, and TLR8 expression and TLR-stimulated response in A549 cells and HUVECs
TLR3, TLR7, and TLR8 are pattern recognition receptors that detect viral RNA (4, 6, 7, 8). We were interested in seeing whether epithelial or endothelial cells express these receptors that would enable direct sensing of viral RNA. We have previously shown that IFN-α can up-regulate TLR3 and TLR7 gene expression in human macrophages (17). Therefore, we also studied the effects of IFN-α and IFN-β on TLR expression in epithelial A549 cells and in HUVECs. A549 cells and HUVECs were left untreated or stimulated with IFN-α or IFN-β for 6 h, after which total cellular RNA was isolated and analyzed by Northern blotting. RNA from IFN-α-stimulated macrophages was included as a positive control. TLR3 mRNA was expressed at a low level in HUVECs whereas in A549 cells TLR3 expression was undetectable. IFN-α and IFN-β strongly up-regulated TLR3 mRNA expression in both cell types (Fig. 1). Although TLR7 and TLR8 were expressed in high levels in IFN-α-stimulated human primary macrophages, no detectable TLR7 or TLR8 expression was seen in untreated or cytokine stimulated A549 cells or in HUVECs (Fig. 1).
Because TLR3, TLR7, and TLR8 were weakly or not at all expressed in nonstimulated A549 cells and in HUVECs, we determined whether these cells were responsive to TLR3 and TLR7/8 ligands. Poly(I:C) is a synthetic analog of dsRNA. Therefore, it can be used to trigger TLR3 signaling. Similarly, the imiquimod derivative R848 specifically activates the TLR7/8 pathway (20, 21). A549 cells and HUVECs were stimulated with poly(I:C) or R848 for 1 h, after which nuclear extracts were prepared and NF-κB activation was analyzed by EMSA. R848 was not able to stimulate NF-κB DNA binding in A549 cells or in HUVECs (Fig. 2 A). In contrast, poly(I:C) activated NF-κB DNA binding in both cell types suggesting that TLR3 signaling pathway functions in A549 cells and in HUVECs. The main NF-κB species induced was a p50/p65 dimer (data not shown).
Endogenous IFN-α/IFN-β production leads to secondary cellular responses, e.g., activation of STAT1 protein expression that has been shown to be under direct control of type I IFNs (22). To study whether poly(I:C) or R848 can activate IFN-α/IFN-β production in A549 cells or in HUVECs we analyzed the ability of these TLR ligands to induce STAT1 protein expression by Western blotting. R848 was not able to activate STAT1 expression in A549 cells or in HUVECs (Fig. 2 B). The induction of STAT1 protein expression by poly(I:C) in HUVECs indicates that these cells can produce IFN-α/IFN-β in response to TLR3 stimulation. However, in contrast to HUVECs, poly(I:C) had little effect on STAT1 protein expression in A549 cells. To confirm that STAT1 induction pathway is intact in A549 cells, the cells were also stimulated with IFN-α and STAT1 protein expression was analyzed. IFN-α clearly enhanced STAT1 protein expression in A549 cells and in HUVECs. To control equal loading, the membranes were stained with anti-STAT5 Abs.
IFNs and poly(I:C) up-regulate TLR3 protein expression in A549 cells and in HUVECs
Because HUVECs and A549 cells responded differentially to poly(I:C) we wanted to characterize TLR3 protein expression in these cells. For this purpose we prepared polyclonal Abs against the intracellular domain of the TLR3 (anti-TLR3-IC Ab) and a selected peptide (SIQKIKNNPFVKQKNLIT) of the extracellular part of the TLR3 (anti-TLR3-P Ab). We characterized polyclonal anti-TLR3 Abs by transfecting the TLR3-Flag construct to HEK293 cells for transient expression of TLR3. The HEK293 cell lysates were immunoblotted under reducing (R) and nonreducing (NR) conditions with polyclonal Abs (Fig. 3). The expression was confirmed by immunoblotting with anti-Flag Ab (Fig. 3). Anti-TLR3-P Abs readily bound to TLR3-Flag run under nonreducing conditions suggesting that they are suitable for immunoprecipitation experiments. Endogenous TLR3 expression was not detectable in HEK293 and A549 cells or in HUVECs with direct anti-TLR3 Ab Western blotting (data not shown). Therefore, we used immunoprecipitation with anti-TLR3-P Ab followed by immunoblotting with anti-TLR3-IC Ab to detect TLR3 expression. Stimulation of A549 cells and HUVECs with IFN-α, IFN-β, and IFN-γ for 20 h resulted in up-regulation of endogenous TLR3 protein expression (Fig. 4). Poly(I:C) clearly up-regulated TLR3 expression in HUVECs. Interestingly, as in the case of STAT1 (Fig. 2 B), poly(I:C) did not induce TLR3 protein expression in A549 cells although a low basal expression of the receptor was seen in these cells.
IFN-α enhances poly(I:C)-induced NF-κB activation
Next we studied whether IFN-α-induced up-regulation of TLR3 expression would result in enhanced TLR3 signaling. For this, we analyzed the effect of IFN-α pretreatment on poly(I:C)-induced activation of NF-κB. A549 cells and HUVECs were pretreated with IFN-α for 20 h after which they were stimulated with different concentrations of poly(I:C) for 1 h. After this nuclear extracts were prepared and analyzed with EMSA. Poly(I:C) activated NF-κB DNA binding in both cell types (Fig. 5). However, IFN-α pretreatment of the cells strongly enhanced NF-κB DNA binding in A549 cells and in HUVECs (Fig. 5).
IFN-α enhances poly(I:C)-induced IFN-β, IL-28, and IL-29 gene expression in A549 cells and in HUVECs
Because IFN-α clearly up-regulated TLR3 expression in A549 cells and in HUVECs we asked whether IFN-α pretreatment is required for poly(I:C)-induced antiviral response. A549 cells and HUVECs were untreated or pretreated with IFN-α for 20 h, after which the cells were stimulated with poly(I:C) for 3 h. Total cellular RNA was prepared and IFN-β, IL-28, and IL-29 mRNA expression was analyzed by Northern blotting. In HUVECs poly(I:C)-stimulated IFN-β, IL-28, and IL-29 mRNA expression was low, but IFN-α pretreatment strongly enhanced the expression of these cytokine genes (Fig. 6). In contrast to HUVECs, IFN-α pretreatment of A549 cells was absolutely required for poly(I:C)-induced IFN-β, IL-28, and IL-29 mRNA expression (Fig. 6). These results suggest that epithelial and endothelial cells are able to produce high levels of antiviral cytokines only if they are exposed to IFN-α before TLR3 ligand stimulation.
Transfection of TLR3 to A549 cells restores poly(I:C) responsiveness
As described above we found out that IFN-α is required for poly(I:C)-induced antiviral response in A549 cells and this response is associated with IFN-α-induced up-regulation of TLR3 protein expression. To assess whether the expression of TLR3 could confer poly(I:C) responsiveness to A549 cells, the cells were transfected with the human TLR3 expression vector. The cells were left unstimulated or stimulated with poly(I:C) for 20 h, and endogenous STAT1 protein expression as well as IL-28 and IL-29 mRNA expression were studied. Expression of TLR3-Flag was confirmed by immunoblotting with anti-TLR3-IC Ab. As shown in Fig. 7,A, STAT1 protein expression is highly elevated in TLR3-transfected A549 cells in response to poly(I:C) stimulation (Fig. 7,A). Similarly, IL-28 and IL-29 mRNA expression was strongly induced by poly(I:C) only in TLR3-transfected A549 cells (Fig. 7 B). Our results show that TLR3 is required for poly(I:C)-induced antiviral response in epithelial cells and that IFN-α is needed for the maximal sensitization of epithelial cells to TLR3 triggering.
TLR7 and TLR8 are pattern recognition receptors for viral ssRNA (6, 7, 8), whereas dsRNA is recognized by TLR3 (4). Therefore TLR3, TLR7, and TLR8 are the main receptors for detection of viral genetic material. The TLR7/8 ligand R848 induces IFN-α or IL-12 production in human myeloid or plasmacytoid DCs, respectively (23). This demonstrates, that TLR7 and TLR8 mediate their effects in a cell type-specific manner. In this report we show that human epithelial and endothelial cells are unresponsive to TLR7/8 stimulation. Neither TLR7 nor TLR8 genes were expressed in A549 cells or in HUVECs apparently explaining the unresponsiveness of these cells to TLR7/8 stimulation. In contrast to TLR7/8 ligand stimulation, both epithelial A549 cells and HUVECs were responsive to TLR3 ligand poly(I:C). However, IFN-α was essential for poly(I:C)-induced activation of IFN and IFN-like IL-28 and IL-29 genes. IFN-α enhanced TLR3 protein expression in both cell types resulting in activation of IFN-β, IL-28, and IL-29 gene expression in response to poly(I:C) stimulation. Our results suggest that in the studied epithelial and endothelial cells TLR3 is the main receptor that detects viral exogenous genetic material. Whether this applies to other sublineages of endothelial or epithelial cells remains to be studied.
In addition to IFN-α, IFN-β and IFN-γ up-regulated TLR3 protein expression in A549 cells and in HUVECs. Poly(I:C) readily enhanced TLR3 and STAT1 protein expression in HUVECs. In contrast, poly(I:C) stimulation did not enhance TLR3 or STAT1 expression in A549 cells. In accordance with these results, low induction of IFN-β, IL-28, and IL-29 genes was seen in HUVECs but not in A549 cells after poly(I:C) stimulation. TLR3 protein was expressed at a lower level in A549 cells compared with HUVECs. Enhanced TLR3 expression by IFN-α stimulation or by TLR3 gene transfection conferred poly(I:C) responsiveness to A549 cells. STAT1 protein expression and IFN-β, IL-28, and IL-29 mRNA expression was clearly induced by poly(I:C) only in IFN-α-stimulated or TLR3-transfected A549 cells. Our results support the idea that TLR3 signaling efficiency depends on the amount of the receptor on the cell surface, and that up-regulation of TLR3 expression by IFN-α/IFN-β or IFN-γ is necessary for dsRNA-induced IFN gene activation. Interestingly, it has been previously shown that IFN-α sensitizes HUVECs to dsRNA-induced apoptosis that is associated with enhanced TLR3 expression (24). Therefore, it would be of considerable interest to study the role of TLR3 in dsRNA-induced apoptosis.
Our result suggest that IFN-α-induced up-regulation of TLR3 expression enhances the antiviral response in epithelial and endothelial cells. During viral infection macrophages and plasmacytoid DCs produce high levels of IFN-α, which could then induce TLR3 expression in endothelial and epithelial cells. Viral dsRNA is first formed intracellularly during viral infection. However, during later stages of virus infection dsRNA may be released from dying cells. Therefore, it is likely to be present also extracellularly which may enhance IFN-α/β production via TLR3 that resides on the plasma membrane. However, at present it is not clear whether TLR3 acts on the plasma membrane or at endosomal compartment where TLR7 and/or TLR8 have been located (25) and where they interact with viral ssRNA. It has been suggested that the cellular localization of TLR3 is cell type specific (26). Further studies are needed to identify the cellular site where dsRNA encounters TLR3 during virus infection.
In our experimental setting, IFN-α-pretreated and poly(I:C)-stimulated A549 epithelial cells and HUVECs expressed IFN-β, IL-28, and IL-29 genes at high levels. IFN-β production by these cell types has been previously described (16, 27). In contrast, the finding that IL-28 and IL-29 gene expression can be induced in epithelial and endothelial cells is a novel one. Like IFN-α/IFN-β, IL-28 and IL-29 have antiviral activity. IL-28 and IL-29 have their own receptor complex and they activate the Jak-STAT signaling pathway to mediate their biological effects (14, 15). Therefore IL-28/IL-29 and their receptor pathway may have important functions in host cell antiviral resistance. Both macrophages and DCs are responsive to IL-28 and IL-29 and these cytokines can induce an antiviral state in the cells (Ref. 14 and data not shown). Alveolar macrophages migrate along the lung epithelium and both macrophages and DCs are in contact with endothelial cells. Therefore, IL-28 and/or IL-29 produced by epithelial or endothelial cells may have important functions in activating APCs or in inducing an antiviral state in these cells.
We have seen that IFN-α is an essential factor in the activation of many secondary response genes after TLR stimulation (28, 29, 30, 31, 32). In this report we provide evidence that IFN-α-induced up-regulation of TLR3 protein expression is necessary for dsRNA activated antiviral response in human epithelial and endothelial cells. In addition, IL-28 and IL-29 produced by these cells are likely important mediators of antiviral immunity and resistance.
The authors have no financial conflict of interest.
We thank Hanna Valtonen, Mari Aaltonen, and Teija Westerlund for expert technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This study was supported by the Medical Research Council of the Academy of Finland, the Sigrid Juselius Foundation, and the Finnish Cancer Foundation.
Abbreviation used in this paper: DC, dendritic cell.