During an innate immune response, macrophages recognize viruses by their pattern recognition receptors. In this study, we have studied the role of membrane-associated TLRs and cytoplasmic retinoic acid inducible gene-I (RIG-I)-like receptors (RLR) in regulation of IFN-β, IL-29, IL-1β, and IL-18 production and caspases 1 and 3 activation in human macrophages. We provide evidence that TLRs are mainly involved in transcriptional up-regulation of IL-1β gene expression, whereas cytosolic dsRNA recognition pathway stimulates powerful IFN-β and IL-29 gene transcription. However, robust IL-1β secretion occurred only if two TLRs were triggered simultaneously or if a single TLR was activated in conjunction with the RLR pathway. Markedly, TLR activation did not stimulate IL-18 processing or secretion. In contrast, triggering of cytosolic RNA recognition pathway with poly(I:C) transfection or influenza A virus infection resulted in caspase-1- and -3-mediated proteolytic processing of pro-IL-18 and secretion of biologically active IL-18. Furthermore, caspase 3-dependent processing of pro-IL-18 was also observed in human HaCaT keratinocytes, and forced expression of RIG-I and its downstream effector, mitochondrial antiviral signaling protein, activated proteolytic processing of pro-IL-18, caspase-3, and apoptosis in these cells. The present results indicate that in addition to robust IFN-β, IL-29, IL-1β, and IL-18 generation, RIG-I/mitochondrial antiviral signaling protein pathway activates caspase-3, suggesting a role for these RIG-I-like receptors beyond the innate cytokine response, hence, in the induction of apoptosis of the virus-infected cell.
Germline-encoded pattern recognition receptors (PRRs)3 of the innate immunity are essential for immediate host’s defense against pathogenic microbes. PRRs detect pathogen-associated molecular patterns (PAMPs), signal to the host the presence of an infection, and activate the immune system against the intruder (1, 2). The transition from the first pathogenic encounter to longer-term adaptive immune reaction is mediated through activated macrophages and dendritic cells (DCs) (3). They present pathogen-derived Ags to the effector cells of the adaptive immunity and secrete soluble antiviral cytokines, like type I (IFN-αβ) and type III IFNs (IL-28 and IL-29) as well as proinflammatory cytokines in response to PRR engagement. IFN-αβ activates NK and CTLs that recognize and eliminate virus-infected cells and secrete IFN-γ upon activation.
In mammals, membrane-associated TLRs are the best-characterized family of PRRs. The first TLR identified, TLR4, is activated by LPS, an outer cell wall component of the Gram-negative bacteria. Three of them, TLR3 and TLR7/8 recognize viral dsRNA and ssRNA, respectively (4). TLR3 and TLR7/8 reside mainly in acidified endosomal compartments. After receptor engagement, cytosolic signaling pathways are activated and transcription of antimicrobial genes is initiated (5).
The second important group of PRRs are cytoplasmic retinoid acid inducible gene-I (RIG-I)-like receptors (RLRs), including RIG-I and melanoma differentiation-associated gene-5 (MDA-5) (1). They are focused on detecting viral genomic RNA or its replication intermediates. At first, these RNA helicases were described to bind dsRNA (6, 7). Subsequently, RIG-I was also implicated to recognize viral ssRNA bearing 5′ phosphates (8, 9) found in the genome of many ssRNA viruses, such as influenza A virus, or siRNAs transcripted by phage-polymerases in vitro. In contrast, MDA-5 was essential for IFN-αβ production in response to cytosolic dsRNA analog polyinosic-polycytidylic acid (poly(I:C)) and picornavirus infection (10, 11). Similarly to TLRs depicted above, the activity of RIG-I and MDA-5 induce the expression of genes important for antiviral host defense. Other possible functions of RLRs are not known. Overall, they are more general sensors for viral infection than nucleic acid-sensing TLRs (12, 13), which are mainly expressed in macrophages and DCs (14, 15).
In conjunction with IFNs, proinflammatory cytokines are essential for efficient microbial clearance. Two of them, IL-1β and IL-18, are primarily produced by macrophages in acute phase of both bacterial and viral infections (16, 17). In general, the activity and regulation of these proinflammatory cytokines is very similar. Both IL-1β and IL-18 are produced as inactive precursors that are processed into biologically active forms by caspase-1 (18). Recent reports show that during bacterial infections, neuronal apoptosis inhibitory protein/MHC class II transcription activator/incompatibility locus protein from Podospora anserina/telomerase-associated protein (NACHT)-leucine-rich repeat and pyrin-domain containing proteins (NALPs) that belong to a third class of PRRs, namely nucleotide-binding oligomerization domain-like receptors (1, 19), activate caspase-1 and secretion of IL-1β and IL-18. Unlike IL-1β, IL-18 is also a substrate for caspase-3. However, caspase-3 processes IL-18 into its biologically inactive equivalent that is not secreted. The viral signals that activate caspase-1- and NALP-containing cytoplasmic multiprotein complexes, termed as inflammasomes, are under intensive research. In macrophages, the activity of TLRs, cytoplasmic helicases RIG-I and MDA-5, and inflammasomes form firm basis for efficient detection of microbes, and subsequent release of IFN-αβ, IL-1β, and IL-18, which serve as soluble mediators between innate and adaptive immune responses.
In this report, we show that engaged TLRs and cytoplasmic PRRs differentially induce expression, processing, and secretion of IL-1β and IL-18, and caspase-1 and -3 activation. In addition, our results demonstrate that the RIG-I/mitochondrial antiviral signaling protein (MAVS) signaling pathway is linked to activation of caspase-3 and induction of apoptosis in human macrophages and keratinocytes.
Materials and Methods
Differentiation of macrophages from PBMCs
Monocytes from healthy blood donors (Finnish Red Cross Blood Transfusion Service) were purified as previously described (20). Monocytes were differentiated into macrophages by maintenance in Macrophage-serum-free medium (Invitrogen Life Technologies) supplemented with 10 ng/ml GM-CSF (Biosource International) for 7 days.
Generation of activated T cells
T cells were further purified from nonadherent PBMCs with nylon wool columns as previously depicted (21). T cells were activated with plate-bound anti-CD3 and anti-CD28 Abs (R&D Systems; both at 1 μg/ml), and cultured in RPMI 1640 supplemented with 10% FCS and IL-2 (10 ng/ml) (Biosource) for 6 days. Growing T cells were further expanded in RPMI 1640 supplemented with 10% FCS and IL-2, and used 2 wk after isolation. At this point, >99% of the cells were CD3 positive. Before conducting experiments, T cells were transferred into RPMI 1640 medium specified above but without IL-2 for 16 h.
Poly(I:C) (Sigma-Aldrich), LPS (Escherichia coli 0111:B4, Sigma-Aldrich), and R848 (Invivogen) were used at 30 μg/ml, 1 μg/ml, and 2 μM, respectively, unless mentioned otherwise. To introduce poly(I:C) into cytoplasm, Lipofectamine 2000 (Invitrogen Life Technologies) was used according to the manufacturer’s instructions. IFN-β (Betaferon) was from Schering Plough and used at 100 IU/ml.
Macrophages were activated with TLR ligands and cytosolic poly(I:C) or infected with influenza A virus in Macrophage-serum-free medium without GM-CSF. Each macrophage sample represents a pool of separately stimulated cells from three different blood donors, and the results are representatives of three independent, but similarly performed experiments.
To analyze the biological activity of macrophage-derived IL-18, IFN-γ expression level was measured from human primary T cells by real-time PCR. T cells were incubated in cell culture media collected from untreated or treated macrophages with or without neutralizing Abs against IL-18 for 4 h. The monoclonal IL-18 Abs were purchased from MBL companies.
For transient transfections of expression vector constructs, HaCaT cell (American Type Culture Collection) were cultured in DMEM supplemented with 10% FCS, 2 mM l-glutamate, and antibiotics (all from Invitrogen Life Technologies). TLR3 expression plasmid was purchased from Invivogen. The RIG-I and ΔRIG-I plasmids have been presented previously (22, 23). The expression vectors were transfected into HaCaT cells using FuGENE 6 transfection reagent (Roche Diagnostics) as specified by the manufacturer.
Influenza A strain Udorn/72 (H3N2) were grown in 11-day-old embryonated eggs as previously described (20). With the virus dose used, >90% of macrophages were infected (data not shown).
Retroviral infection of HaCaT cells by MAVS cDNA
MAVS cDNA was amplified (5′-prime: CCG TTT GCT GAA GAC AAG ACC TAT AAG TAT and 3′-prime: CTA GTG CAG ACG CCG GTA CAG CAC CA) using PFU polymerase (Stratagene), and cloned into modified retroviral pMSCVpuro vector (Invitrogen) where the puromycin gene had been replaced by human CD8 cDNA. N-terminal HA-epitope tag sequence was inserted into the MAVS cDNA. The parental and MAVS-encoding vectors were cotransfected with pCL retroviral packaging vector and pVSVG-envelope protein into human embryonic kidney 293T cells using standard calcium phosphate precipitation method to produce retroviruses (24). After 48 h of transfection, growing HaCaT cells were combined with culture supernatant containing retrovirus and polybrene (8 μg/ml), and centrifuged at 2500 rpm at room temperature for 45 min. A second infection was done 24 h later, after which the cells were used for experiments. The infection level was verified by flow cytometry and with PE-conjugated anti-CD8-Ab (BD Biosciences). The average infectivity was 78% for parental vector and 40% for vector encoding MAVS.
RNA isolation and RT-PCR
Total cellular RNA was isolated by RNeasy kit (Qiagen) according to the manufacturer’s instructions. Total of 0.5 μg of RNA was reverse transcribed into cDNA by a high capacity cDNA reverse transcription kit (Applied Biosystems) in a 25 μl reaction mixture containing optimized reverse transcription buffer, random primers, deoxyNTP mixture, and MultiScribe reverse transcriptase. The conditions for cDNA synthesis were as follows: annealing at 25°C for 20 min and synthesis at 37°C for 120 min.
Quantitative real-time PCR
For quantitative real-time PCR, TaqMan analysis was done in a 96-well optical reaction plate in ABI Prism 7700 Sequence Detector (Applied Biosystems). Amplification was conducted in 25 μl of reaction mixture containing 20 ng of the cDNA, TaqMan universal PCR master mix (Applied Biosystems), and primers and probes provided as Predeveloped TaqMan assay reagents by Applied Biosystems. For each sample, PCR amplification of the endogenous 18S rRNA was determined to allow normalization between the samples, according to the manufacturer’s instructions (Applied Biosystems). The thermocycling conditions consisted of an initial step of 50°C for 2 min followed by 95°C for 10 min, and 40 cycles of denaturation at 95°C for 15 s, and annealing and extension steps at 60°C for 1 min. Real-time PCR was performed at least in duplicate for each cDNA product. No template control (NTC), in which molecular grade water was used instead of template, was included in each assay.
Real-time PCR data analysis
The real-time PCR data were developed by using Sequence Detector System version 1.9.1 software (Applied Biosystems). The cycle threshold value (CT) of a sample, that is the cycle number at which measured fluorescence was greater than the manually fixed threshold fluorescence in the amplification plot, was selected according to the manufacturer’s guidelines. Relative units were calculated by a comparative CT method. First, ΔCT, the difference between the CT value of the target amplicon and 18S rRNA, was determined for each sample. Second, a calibrator ΔCT value of 40, obtained from NTC, was subtracted from average CTs of 18S rRNAs to get ΔΔCT. Finally, the amount of target normalized to an endogenous control, which was relative to the NTC-calibrator, was calculated by the equation 2-ΔΔCT. Standard curve and statistics were generated in GraphPad Prism 4 software.
Western blot analysis
Cells were suspended into a buffer containing 10 mM Tris (pH 7.4), 150 mM NaCl, and 25% ethylene glycol supplemented with complete mini protease inhibitor mixture (Roche Diagnostics) and homogenized by ultrasound sonication. Total of 10 μg of proteins were separated on 15% SDS-PAGE at 200 V, and transferred on Immobilon-P transfer membranes (Millipore). Blots were stained with Ponceau red to confirm the equal loading and transfer of the protein samples. Membranes were blocked in PBS containing 5% nonfat milk and stained with Abs against IL-1β, IL-18, caspase-1 p20, and caspase-3 p17/19 at 4°C for 18 h. After this, membranes were incubated at room temperature for 1 h with appropriate HRP-conjugated secondary Abs (Dako Cytomation). Proteins were visualized by ECL system (PerkinElmer). Polyclonal anti-IL-1β, anti-IL-18, and anti-RIG-I Abs have been described (20, 22, 25). Polyclonal Abs against caspase-1 p20 and caspase-3 p17/19 were purchased from Sigma-Aldrich and Cell Signaling Technology (Danvers, MA), respectively.
Human IL-1β Eli-pair and IL-18 ELISA were purchased from Diaclone (Besançor Cedex) and Medical and Biological Laboratories, respectively.
Extracellular and cytosolic dsRNA differentially activate IFN-β, IL-29, IL-1β, and IL-18 mRNA expression in human primary macrophages
Virus infection typically involves the presence of viral replication intermediates, such as dsRNA, that are detected by PRRs of the innate immunity. To study the effect of extracellullar and cytoplasmic dsRNA on cytokine response in human primary macrophages, we introduced synthetic dsRNA analog, poly(I:C), directly into cell culture medium, or transfected it with lipofectamine into cytosol (t-poly(I:C)), respectively. After 6 h of stimulation, the cells were collected, total cellular RNA was prepared and cDNA synthesized, and IFN-β, IL-29 (also known as IFN-λ1), IL-1β, and IL-18 mRNA expression was studied by real-time PCR. Extracellular poly(I:C) strongly and dose-dependently activated IL-1β mRNA expression, while cytosolic poly(I:C) had a modest enhancing effect on IL-1β gene expression (Fig. 1,A). Extracellularly given poly(I:C) weakly up-regulated basal IL-18 mRNA expression, whereas cytosolic poly(I:C) had no effect on IL-18 mRNA production (Fig. 1,B). In contrast to IL-1β gene transcription, IFN-β and IL-29 mRNA expression was strongly activated by t-poly(I:C) (Fig. 1, C and D).
To study the kinetics of cytokine gene expression, macrophages were stimulated with extracellular and cytosolic poly(I:C) for different time periods after which total cellular RNA was isolated and IFN-β, IL-29, and IL-1β mRNA expression was studied by real-time PCR. In accordance with results in Fig. 1, extracellular poly(I:C) had a weak enhancing effect on IFN-β mRNA expression at all time point tested, a maximal induction seen at 3 h after extracellular poly(I:C) stimulation (Fig. 2, B and C). In contrast, cytosolic poly(I:C) strongly activated IFN-β and IL-29 gene expression already at 3 h after stimulation. The opposite was seen with IL-1β gene expression: extracellular poly(I:C) strongly activated IL-1β mRNA expression whereas cytosolic poly(I:C) had a modest enhancing effect on IL-1β gene expression at 3 h after stimulation (Fig. 2 A). Our results suggest that extracellular and cytoplasmic dsRNA recognition pathways induce strong expression of proinflammatory and antiviral cytokines, respectively, in human macrophages.
Cytoplasmic dsRNA recognition pathway activates pro-IL-18 processing
Sensing of extracellular and cytoplasmic dsRNA by PRRs clearly activate different pattern of cytokine gene expression. Therefore, we assessed the effect of dsRNA on cytokine expression at the protein level. Both IL-1β and IL-18 are synthesized as inactive 31 kDa (p31) and 24 kDa (p24) precursors, respectively, which remain in the cytoplasm. The cysteine protease caspase-1 is responsible for the production of biologically active IL-1β p17 and IL-18 p18 in macrophages. Macrophages were treated with extracellular or cytoplasmic poly(I:C) for different time periods, after which protein lysates were prepared and analyzed by Western blotting. Macrophages activated with extracellular poly(I:C) expressed high levels of pro-IL-1β already at 3 h which was accompanied by proteolytic processing of pro-IL-1β (Fig. 3,A). However, no biologically active form of IL-1β was detected suggesting that extracellular poly(I:C) stimulation did not activate caspase-1. Consistently to Fig. 1 and 2, cytosolic poly(I:C) was a weak inducer of pro-IL-1β expression (Fig. 3,B). Instead, it induced proteolysis of pro-IL-18 to IL-18 p15 and p16 fragments, and their appearance was most prominent at 18 h (Fig. 3 B). This suggested that caspase-3, which processes pro-IL-18 into biologically inactive p15 and p16 polypeptide chains, was activated. Surprisingly, pro-IL-18 proteolysis was absent in human macrophages treated with extracellular poly(I:C).
Extracellular and cytosolic dsRNA differentially activate caspases 1 and 3
Caspase-1 is expressed as an inactive p45 proenzyme that becomes biologically active after proteolytic cleavage. The functional enzyme comprises of p10 and p20 subunits, which assemble into a heterotetramer (26). To study caspase-1 activation in response to extracellular and cytosolic poly(I:C), we performed Western blot analysis with anti-caspase-1 p20 specific Abs. Cytosolic poly(I:C) strongly induced caspase-1 p20 which was most distinct at 18 h after poly(I:C) transfection (Fig. 3,D). In contrast, extracellular poly(I:C) had no notable effect on caspase-1 p20 expression at any time-points studied (Fig. 3,C). Instead, extracellular poly(I:C) induced formation of a slower migrating form of caspase-1 p20, labeled as caspase-1 p22 in Fig. 3 C. Caspase-1 p22 has also been reported in human monocytic THP-1 cells (27).
Like caspase-1, caspase-3 is a latent p32 zymogen, which cannot exert its enzymatic activities unless it is proteolytically processed. The mature, bioactive, enzyme consists of p12 and p17 subunits. The Western blot analysis performed with caspase-3 p17/19 specific Abs clearly indicated that extracellular poly(I:C) was not able to activate caspase-3 (Fig. 3,C). In contrast, cytoplasmic poly(I:C) strongly activated caspase-3 (Fig. 3,D). The first appearance of caspase-3 p17/19 monomers took place already at 3 h after cytosolic poly(I:C) treatment which was in good agreement with the kinetics of IL-18 p15/16 formation seen in Fig. 3 B. An unknown protein staining was observed with anti-caspase-3 Ab (marked with arrow) in macrophages stimulated with cytoplasmic dsRNA. In conclusion, the results strongly illustrate that cytoplasmic dsRNA recognition pathway activates caspase-3.
TLRs and cytosolic dsRNA recognition pathway differentially regulate processing and secretion of IL-1β and IL-18
Besides difference in caspase-1 and -3 activation, our data also sharply indicated the difference between extracellular and cytosolic dsRNA–induced processing of IL-1β and IL-18. Microbes express several PAMPs that may simultaneously activate different membrane-associated or cytoplasmic PRRs. The cytokine response may quantitatively or qualitatively alter depending on the engaged PRRs and/or their cooperation (28). To address whether different PRRs synergistically induce pro-IL-1β and pro-IL-18 processing and IL-1β and IL-18 secretion, macrophages were stimulated with extracellular and cytosolic poly(I:C), LPS, and R848 (a TLR7/8 ligand), or with their combination. All TLR agonists as well as cytosolic poly(I:C) strongly activated the synthesis of pro-IL-1β which was absent in untreated macrophages (Fig. 4,A). In accordance with previous results, IL-1β was secreted only after synergistic trigger of two TLR ligands (29). Furthermore, we also observed that the TLR pathway synergized with cytosolic dsRNA recognition pathway to induce IL-1β secretion (Fig. 4,B). In contrast, pro-IL-18 processing and IL-18 secretion occurred only in macrophages that had been activated with cytosolic poly(I:C) (Fig. 4, A and B). Surprisingly, none of the TLR agonists tested alone or in cooperation were able to induce pro-IL-18 processing or secretion of IL-18. Our results strongly suggest that cytoplasmic dsRNA recognition machinery, but not TLRs, is the trigger for pro-IL-18 processing and subsequent IL-18 secretion in human primary macrophages.
TLR pathway is not involved in activation of caspases 1 and 3
Caspase-1 and -3 activation was also studied in response to simultaneous PRR activation. Consistently, TLR ligands: extracellular poly(I:C), LPS, and R848 induced formation of caspase-1 p22 which is not involved in processing of pro-IL-1β and pro-IL-18 to their biologically active forms. Furthermore, TLR engagement did not result in the appearance of caspase-3 p17/19 (Fig. 4 A). In contrast, cytosolic poly(I:C) activated caspase-1 p20 and caspase-3 p17/19. The TLR pathway had no effect on cytosolic poly(I:C)-induced caspase-1 and/or caspase-3 activation.
Effect of TLR pathway on influenza A virus infection-induced caspases 1 and 3 activation and IL-1β and IL-18 secretion
As shown in Fig. 4, cytoplasmic dsRNA activated caspase-1 and -3 and secretion of the mature IL-18. We have previously reported that influenza A virus infection of human macrophages results in caspase-dependent IL-18 secretion (20, 21, 25). To study the effect of TLR ligation on virus-induced processing and secretion of IL-1β and IL-18, macrophages were stimulated with extracellular poly(I:C), LPS, or R848 with or without influenza A virus infection for 18 h. In contrast to TLR ligands, only a weak induction of pro-IL-1β was seen in influenza A virus-infected macrophages (Fig. 5,A). However, TLR ligands and influenza A virus infection had a tremendous synergistic effect on IL-1β secretion. Notably and similarly to cytoplasmic poly(I:C), influenza A virus infection alone was able to activate secretion of IL-18 (Fig. 5 B).
Activation of caspase-1 and -3 was also studied in response to simultaneous coactivation with TLRs and influenza A virus infection. Like cytoplasmic poly(I:C), influenza A virus infection alone resulted in the appearance of caspase-1 p20 subunit and activation of caspase-3 (Fig. 5,A). Consistent with results shown in Fig. 3,C and 4 A, individual TLRs enhanced the formation of caspase-1 p22 and they were insufficient to promote caspase-3 activation.
Macrophages stimulated with cytosolic dsRNA or infected with influenza A virus produce cytokines that activate IFN-γ gene expression in human T cells
IFN-α has been shown to induce IFN-γ production from human T cells in synergy with IL-18 (21). Our present and published data show, that macrophages stimulated with cytosolic poly(I:C), or infected with influenza A virus produce both IFN-αβ and IL-18. Therefore, we determined the IFN-γ inducing activity of the cell culture supernatants collected from macrophages stimulated with extracellular and cytosolic poly(I:C), or infected with influenza A virus. The supernatants were subjected to human primary T cells in the presence or absence of IL-18 neutralizing Abs. After 4 h of stimulation, T cells were harvested and their IFN-γ mRNA expression was analyzed by real-time PCR. Results displayed in Fig. 6 show that supernatants from extracellular poly(I:C)-induced macrophages were not able to enhance IFN-γ gene expression in T cells. However, supernatants from cytosolic poly(I:C)-stimulated or influenza A virus-infected macrophages strongly activated IFN-γ gene expression. Anti-IL-18 Abs completely inhibited IFN-γ inducing activity of macrophage supernatants showing that influenza A virus infection and cytosolic dsRNA can induce production of biologically active IL-18. Our results show that cytokines produced by cytosolic poly(I:C)- or influenza A virus-activated macrophages enhance the adaptive Th cell type 1 immune response.
RIG-I/MAVS-signaling pathway activates caspase-3 and apoptosis in HaCaT keratinocytes
The antiviral role of the RIG-I/MAVS signaling pathway is commonly accepted in influenza A virus infection (10, 22, 30). Both cytosolic dsRNA and influenza A virus activated caspase-3 and subsequent generation of IL-18 p15/p16 fragments in macrophages. For further studies, we used human HaCaT keratinocytes. Similarly to human macrophages, also HaCaT cells constitutively express pro-IL-18 (31, 32). In contrast to macrophages, we did not detect secretion of biologically active form of IL-18 in HaCaT cells (data not shown). However, both cytoplasmic poly(I:C) and influenza A virus infection resulted in the formation of IL-18 p15/16 and activation of caspase-3 in HaCaT cells (Fig. 7,A). Moreover, these treatments also enhanced expression of endogenous RIG-I, which is positively regulated by IFN-αβ. Type I IFNs are also well known for their proapoptotic effects. To define whether IFN-αβ is directly related to caspase-3 activation as well as pro-IL-18 processing, HaCaT cells were left untreated or pretreated for 18 h with IFN-β after which they were activated for different time periods with cytoplasmic poly(I:C). IFN-β pretreatment clearly promoted RIG-I expression but had a very modest enhancing effect on pro-IL-18 processing or caspase-3 activation. Moreover, IFN-β pretreatment did not significantly alter the formation of IL-18 p15/16 and activation of caspase-3 after cytoplasmic poly(I:C) stimulation at any time point studied (Fig. 7 B).
To define the role of RIG-I and its immediate downstream signaling partner MAVS in pro-IL-18 processing and activation of caspase-3, we performed functional studies in HaCaT keratinocytes which, unlike primary macrophages, are permissive to gene transfection experiments. Forced expression of full-length RIG-I and its constitutively active form ΔRIG-I, containing only its two caspase recruitment domains, resulted in robust appearance of IL-18 p15/p16 (Fig. 7,C). Furthermore, RIG-I and ΔRIG-I strongly activated caspase-3. In contrast, forced expression of TLR3 had no effect on pro-IL-18 processing or caspase-3 activation. To analyze the effect of MAVS on caspase-3 activation and apoptosis in HaCaT cells, we performed retroviral gene transfer experiments. The retroviral transfection of parental vector had no effect on RIG-I protein production, but retroviral transfer of MAVS clearly induced RIG-I protein expression, proteolysis of pro-IL-18 into IL-18 p15/p16 polypeptide chains, and caspase-3 activation (Fig. 7 D). In addition, retroviral expression of MAVS was detrimental to HaCaT cells: apoptotic cell shrinkage, rounding, and detachment from cell culture plates were seen at 24 h. In conclusion, our results show that the RIG-I/MAVS signaling pathway is tightly interwoven into antiviral and apoptotic events both in human primary macrophages and keratinocytes.
Pathogens are a highly diverse group of microbes. Their detection is therefore a major challenge for the immune system. The host defense against infections is initiated by the innate immune system operating on the basis of recognition of general pathogen features (2). Pathogens are identified through PRRs that detect PAMPs, signal the presence and type of infection, and activate the optimal host defense. Microbes express various PAMPs that may simultaneously engage different membrane-associated or cytosolic PRRs. Therefore, the innate immune response, including antiviral and proinflammatory cytokine production, is likely to be regulated in a synergistic or cooperative manner by the particular PRR pathways. Napolitani and coworkers have shown that combinations of TLR agonists trigger synergistically expression of several proinflammatory cytokines in human DCs leading to enhanced Th cell type 1-polarizing capacity (29). In this report, we have examined the role of membrane-associated TLRs and cytosolic RNA-recognition pathway in the activation of caspases 1 and 3, and subsequent generation of IL-1β and IL-18 in human macrophages and keratinocytes.
Detection of intracellular viruses by PRRs is primarily based on sensing viral nucleic acids (4). Most DNA and RNA viruses generate dsRNA at some point during their replication cycle (33). TLR3 was the first PRR characterized to be involved in dsRNA-induced cytokine response (34). However, gene-knock-out studies suggest that induction of IFN-αβ synthesis in response to viral infection is not dependent on activation of TLR3 alone (35). In contrast, RNA helicases RIG-I and MDA-5 recognize the presence of cytoplasmic viral RNA, and their activation results in robust IFN-αβ expression. In line with this knowledge, transfected RNA duplex poly(I:C), which mimics cytosolic dsRNA, activated strongly IFN-β and IL-29 gene expression, whereas extracellular poly(I:C) had little effect on their transcription in human macrophages (Fig. 1 and 2). Our experiments clearly show that triggering of TLR3 pathway with extracellular dsRNA results in high IL-1β gene transcription that is nearly absent from macrophages activated by cytosolic dsRNA recognition pathway. It has been shown that influenza A virus- and West Nile virus-infected TLR3-deficient mice have impaired production of proinflammatory cytokines IL-6 and TNF-α (36, 37). These studies and our results emphasize the importance of TLR3 in the activation of proinflammatory cytokine response in viral infections.
The synthesis and secretion of biologically active IL-1β and IL-18 in response to extracellular and cytoplasmic dsRNA have not been systematically characterized. Activated macrophages are the primary source of these proinflammatory cytokines in acute-phase of both bacterial and viral infections. The initial translation products are inactive precursors that require cleavage before they are biologically active and ready for secretion (18). It is well-reported that TLR agonists can act both as transcriptional/translational and processing-inducing stimuli for IL-1β, but cannot alone provide signal for IL-1β secretion (16). In accordance, we observed that only upon simultaneous activation of two PRRs: dual TLR ligation, engagement of TLR together with cytoplasmic dsRNA pathway, or TLR activation plus influenza A virus infection resulted in IL-1β precursor processing and efficient secretion of the cytokine (Fig. 4 and 5). Noticeably, influenza A virus-induced modest processing and secretion of IL-1β was greatly enhanced by activation of TLR3, TLR4, or TLR7/8 (Fig. 5). Our data suggests that the main function of TLR pathway is to prime macrophages for actual microbial infection by activating transcription of proinflammatory cytokines like IL-1β. However, robust secretion of IL-1β takes place only if macrophages have simultaneously received signals from different cellular compartments or at least from two separate receptor systems.
It is well established that both endolysosomal TLR7/8 and cytosolic RIG-I pathways activate IFN-αβ production in response to influenza A virus infection (10, 22, 30, 38, 39). However, our results show that TLR7/8 pathway was not able to activate maturation or secretion of IL-18 in human macrophages. In contrast, influenza A virus infection or activation of cytosolic RNA recognition pathway with transfected poly(I:C) strongly activated IL-18 processing and secretion (Fig. 4 and 5). Moreover, both stimuli activated the secretion of the biologically active form of IL-18 in conjunction with IFN-αβ, as judged by IFN-γ production in T cells in the presence or absence of IL-18 neutralizing Abs (Fig. 6). To conclude, our results suggest that viral cytoplasmic RNA is the main trigger to activate IL-18 maturation and secretion in macrophages during viral infection.
A marked difference was also observed in caspase-1 activation after TLR triggering and cytosolic dsRNA stimulation. Both cytosolic poly(I:C) and influenza A virus infection resulted in strong appearance of caspase-1 p20 subunit, while the appearance of caspase-1 p22 was detected only after TLR engagement (Fig. 4 A and 5A). The amino-terminally extended form of caspase-1 p20, caspase-22, was initially described in a study investigating autocatalysis of caspase-1 in human monocytic THP-1 cells (27). Our results strongly suggest that caspase-1 p22 is not able to process pro-IL-1β or pro-IL-18 into its biologically active form. In conclusion, we provide evidence that induction of caspase-1 p20 is strongly associated with pro-IL-1β and pro-IL-18 processing and secretion of respective mature cytokines, and that these events take place in macrophages stimulated with the cytosolic dsRNA or influenza A virus infection, but interestingly not in cells activated by single TLR agonists.
Conversion of inactive pro-caspase-1 into enzymatically functional caspase-1 is closely intertwined to both pro-IL-1β and pro-IL-18 processing, and it is suggested to occur in cytoplasmic NALP-inflammasomes (40). However, viral signals that activate caspase-1-containing NALP-inflammasomes have not been characterized. Recently, Kanneganti and colleagues (41) reported that NALP3/Cryopyrin is required for poly(I:C)-induced IL-1β and IL-18 secretion in vivo. In contrast, danger signals released from dying cells, such as gout-associated monosodium urate crystals, also result in maturation of these proinflammatory cytokines (42). In our experiments, both cytosolic poly(I:C) and influenza A virus induced caspase-1 activation and subsequent processing and secretion of IL-1β and IL-18. Thus, it is possible that virus-activated cytosolic ssRNA and/or dsRNA recognition machinery provide danger signals, or alternatively induce their production, leading to the activation of caspase-1-containing NALP-inflammasomes. Very recent data also suggest that anti-apoptotic and chaperon-like proteins participate in the function of NALPs (43, 44). Also, these molecules can be indirectly or directly targeted by viruses and by doing so influence the maturation and secretion of IL-1β and IL-18. Clearly, further studies are necessary to delineate the roles and actions of inflammasomes in response to viral infections.
Viral infections are associated with programmed cell death of the infected cells. Apoptosis can be regarded as a part of the host’s innate immune defense, because apoptotic cell death reduces viral spread by preventing the production of progeny viruses, and viruses have evolved mechanisms that antagonize apoptosis of the host cell (45). The PRRs involved in the activation of apoptosis have remained unrecognized in viral infections. The antiviral function of RNA helicases RIG-I and MDA-5 and their adaptor protein MAVS is well established during influenza A virus infection (10, 22, 30). In the present study, we show that transient and retroviral expression of RIG-I and MAVS, respectively, activated caspase-3 and induced apoptosis in HaCaT cells (Fig. 6, C and D). These results imply that RIG-I/MAVS signaling pathway also has a role beyond the innate cytokine response, hence, in the induction of apoptosis of virus-infected cells. Interestingly, RIG-I-deficient mouse embryos die at days 12.5 to 14.0, possibly due to the aberrant occurrence of apoptosis (12). Therefore, it is highly possible that the antiviral RIG-I/MAVS pathway regulates apoptosis during development and viral infections.
RIG-I has been shown to activate IFN-αβ production in response to dsRNA and viral ssRNA bearing 5′ phosphates in several different cell-types. In addition to viral RNA molecules, siRNAs transcripted by phage-polymerases in vitro are known to activate RIG-I pathway. Marques and coworkers (46) have shown that siRNAs, which are devoid of two nucleotide 3′ overhangs and range from 21 to 27 nucleotides in length, can activate the IFN response. Our results show that the RIG-I/MAVS pathway is associated with activation of caspase-3 and apoptosis, suggesting that certain siRNA molecules may also trigger programmed cell death. Apoptosis elicited by siRNAs may complicate the use of RNAi as a method to down-regulate specific gene expression. In contrast, apoptosis-inducing activities of siRNAs may have beneficial effects on siRNA-based treatments of viral infections or cancer. It will be important to define whether IFN-αβ-inducing ability of siRNAs is associated with apoptosis of the target cells.
Cells of the innate immune system, including macrophages and DCs, sense viral presence at least with two different receptor systems: membrane-associated TLRs and cytosolic RNA helicases. Our results suggest that cytoplasmic receptors for ssRNA and/or dsRNA are key components of the innate immune response against influenza A virus due to their ability to trigger robust IFN-αβ production, activation of caspases 1 and 3, and secretion of biologically active IL-18 in human macrophages.
We thank Dr. Ilkka Julkunen for providing influenza A virus, anti-IL-18, and anti-RIG-I Abs, and Dr. Tada-atsu Imaizumi for RIG-I expression plasmid Jeremiah 33:3.
The authors have no financial conflict of interest.
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 work was supported by grants from the Research Council for Biosciences and Environment of the Academy of Finland, and the Sigrid Juselius Foundation.
Abbreviations used in this paper: PRR, pattern recognition receptor; PAMP, pathogen-associated molecular pattern; DC, dendritic cell; RIG-I, retinoid acid inducible gene-I; RLR, RIG-I-like receptor; MDA-5, melanoma differentiation-associated gene-5; poly(I:C), polyinosic-polycytidylic acid; NALP, NACHT-leucine-rich repeat and pyrin-domain containing protein; NTC, no template control; CT, cycle threshold value; MAVS, Mitochondrial antiviral signalling protein; t-Poly(I:C), transfected Poly(I:C).