Abstract
TLRs interact with a growing list of pathogen-derived products and these interactions drive the activation of innate and adaptive immune responses. Dendritic cells (DC) play a key role in these events expressing a heterogeneous repertoire of TLRs. We have previously demonstrated the production of type I IFNs in DC following bacterial infections and TLR triggering. In this study, we sought to characterize the transcriptome specifically induced in human DC by IFN-β production stimulated upon LPS treatment. To this aim, by using cDNA microarrays, we compared the transcriptome of DC following LPS treatment in the absence or presence of neutralizing anti-type I IFN Abs. Interestingly, we found that the expression of TLR7 was induced during LPS-induced maturation of DC in a type I IFN-dependent manner. The induction of TLR7 in maturing DC was mainly a consequence of the transcriptional activity of IRF-1, whose binding site was located within TLR7 promoter. Moreover, we also demonstrated that “priming” of immature DC, that usually express TLR8 but not TLR7, with exogenous IFN-β induced a functionally active TLR7. In fact, treatment with the TLR7-specific ligand 3M-001 up-regulated the expression of CD83, CD86, and CD38 in IFN-β-primed DC but not in immature DC. Therefore, a robust enhancement in proinflammatory as well as regulatory cytokines was observed. These data suggest that TLR4-mediated type I IFN release activates specific transcription programs in DC amplifying the expression of pathogen sensors to correctly and combinatorially respond to a bacterial as well as viral infection.
Dendritic cells (DC)3 constitute a heterogeneous population of APCs that are critical for bridging innate and adaptive immune responses. In their steady state, immature DC (iDC) continuously take up and process foreign as well as self-Ags. Upon virus infection or microbial recognition, iDC undergo a complex process of maturation resulting in the migration from tissues to secondary lymphoid organs and up-regulation of MHC and costimulatory molecules that are essential for T cell priming (1, 2).
DC express the broadest repertoire of pattern recognition receptors (PRRs), through which they discriminate between invading pathogens recognizing a plethora of conserved motifs associated with different classes of microbes. TLRs, one class of PRRs, are type I transmembrane proteins consisting of at least 10 members in human and 11 in mouse. The pattern of TLR expression in different DC types and their interaction with their specific ligands regulate cytokine production and direct the type of immune response (3, 4); freshly isolated human plasmacytoid DC (pDC) express TLR7 and TLR9, whereas human myeloid DC (mDC) express TLR1, TLR2, TLR3, TLR4, TLR6, and TLR8 (5, 6, 7).
Each member of the TLR family recognizes a specific set of bacterial- or viral-derived molecules (8). TLR2 recognizes peptidoglycan and lipopeptides, TLR3 is triggered by dsRNA, recognition of Gram− bacteria-derived LPS is mediated by TLR4 (together with its soluble coreceptor MD-2), flagellin is recognized by TLR5, and the TLR9 ligand is known to be unmethylated bacterial DNA (CpG). Moreover, TLR7 and TLR8 recognize ssRNA and are also activated by infections with ssRNA viruses, including influenza virus and vesicular stomatitis virus (9, 10). In addition to ssRNA, the synthetic imidazoquinoline, imiquimod, a low m.w. immune response modifier, activates TLR7 in both humans and mice while its derivative resiquimod (R-848) activates TLR7 and TLR8 in humans, but only TLR7 in mice (11, 12, 13).
Every TLR triggers a specific cellular activation program due to the association with distinct adaptor molecules alone or in combination with the Toll-IL-1R (TIR) domain; these include MyD88, MyD88 adapter-like (also called TIRAP), TIR-domain containing adapter inducing IFN-β (TRIF, also called TICAM1), and TRIF-related adapter molecule (TRAM, also called TICAM2). With the exception of TLR3, which signals solely by TRIF, all TLRs recruit MyD88, which triggers signaling pathways leading to nuclear translocation of NF-κB (for review, see Ref. 8). The MyD88-independent or TRIF-mediated signaling is used by TLR3 and also by TLR4. This pathway mainly permits the phosphorylation-induced activation of IFN regulatory factor (IRF)-3 by two IκB kinase-related kinases, IKKε and TBK1 (14, 15). These events result in the assembly of the enhanceosome, a multiprotein complex consisting of ATF2/c-Jun, IRF-3, IRF-7, and NF-κB, all of which are required for the transcription of the IFN-β gene (16). IFN-β signals in an autocrine/paracrine manner via its heterodimeric receptor complex, formed by the two chains IFN-αβ-R1 and IFN-αβ-R2. Each receptor subunit engages a member of the JAK family, tyrosine kinase 2 or JAK1, that can be activated by autophosphorylation upon ligand-induced rearrangement (17). These events result in the tyrosine phosphorylation and heterodimerization of STAT-1 and STAT-2 together with IRF-9 (also called p48) to form the IFN-stimulated gene factor 3 (ISGF3) complex. ISGF3 initiates transcription of multiple IFN-stimulated genes (ISGs) by binding IFN-stimulated response elements on their promoters (17).
Type I IFNs (several IFN-α, IFN-β, IFN-κ, and IFN-ω) are cytokines that play a central role in host resistance to viral or microbial infections and are often considered important components linking innate and adaptive immunity. We and others (18, 19, 20, 21) have previously demonstrated the production of type I IFNs in DC following bacterial infections or TLR triggering. This autocrine IFN may have critical effects on the biology of DC. Indeed, type I IFNs promote the differentiation of human blood monocytes into DC with potent T cell stimulatory activities and contribute to DC maturation (22, 23).
We have shown, in the past, that the LPS- and poly I:C-induced maturation of human DC induces mostly the production of IFN-β, among all type I IFN subtypes. We also addressed the different responsiveness of these cells to these cytokines at the different stages of maturation (21, 24). In this study, we sought to determine the contribution of IFN-β to LPS-induced DC maturation. To this aim, we compared the transcriptome of DC following 24 h LPS treatment in the presence or absence of anti-type I IFN Abs, to neutralize the action of IFN. We found that the LPS-treated DC induce TLR7 in a type I IFN-dependent manner via IRF-1. Similarly, TLR7 induction was observed following IFN-β treatment of immature DC (iDC). Moreover, we show that IFN-β-pretreated iDC are able to respond to specific TLR7 ligand by expressing surface markers of activation and maturation and producing proinflammatory and regulatory cytokines, whereas the iDC fail to do so.
Our results, therefore, highlight a new status of alert to pathogen infections induced by IFN-β production in DC during LPS-induced maturation.
Materials and Methods
Generation of DC
DC were prepared as previously described (21). Briefly, PBMC were isolated from freshly collected buffy coats obtained from healthy voluntary blood donors (Blood Bank of University “La Sapienza,” Rome, Italy) by density gradient centrifugation using Lympholyte-H (Cedarlane Laboratories). Monocytes were purified by positive sorting using anti-CD14-conjugated magnetic microbeads (Miltenyi Biotec). The recovered cells were > 99% CD14+ + and 95% CD14−. At day 5, DC were starved from IL-4 and GM-CSF for 20 h before their stimulation.
Abs and other reagents
mAbs specific for CD1a, CD14, CD86, CD83, CD38, IgG1 (BD Pharmingen) were used as direct conjugates to FITC or PE. Sheep antiserum raised against human leukocyte IFN, a gift from G. Uzé (Centre National de la Recherche Scientifique, Unité Mixte de Recherche 5124, Montpellier, France), was used at a 1/100 dilution (25). LPS from Escherichia coli 0111:B4 (Sigma-Aldrich) was used at a concentration of 1 μg/ml. IFN-β (Avonex; Biogen) was used at 200 pM. TLR7 (3M-001) and TLR8 (3M-002) ligands (3M Pharmaceutical) were used at 25 and 5 μM, respectively (13). These optimal doses of TLR7 and TLR8 ligands to induce DC maturation were evaluated by titrating different concentrations. Poly I:C (Amersham Biosciences) was used at 50 μg/ml. Pam3CSK4 (Alexis) was used at 2 μg/ml. H3N2 influenza A (Flu) strain A/Beijing/353/89 virus was grown in 11-day-old embryonated chicken eggs and used at a concentration of 12.8 hemagglutination U/ml (gift of Dr. I. Julkunen, National Public Health Institute, Helsinki, Finland).
RNA isolation and real-time RT-PCR quantification
RNA was extracted with RNeasy Mini kit (Qiagen) according to the manufacturer’s instructions that includes a DNase I treatment. Reverse transcriptions were primed with oligo(dT) and performed using the murine leukemia virus reverse transcriptase (Invitrogen Life Technologies). Quantitative PCR assays were done at least in triplicates using the Platinum TaqDNA Polymerase (Invitrogen Life Technologies) and the SYBR Green I (BioWhittaker Molecular Applications) on a LightCycler (Roche Diagnostics). All quantification data are presented as a ratio to the GAPDH level. The SEs (95% confidence limits) were calculated using the Student t test. Quantification standard curves were obtained using dilutions (4-log range) of the PCR products in 10 μg/ml sonicated salmon sperm DNA. The sequences of the primer pairs used for the quantification of GAPDH have been previously described (24). The primers used for mRNA quantification were: 2′-5′-oligoadenylate synthetase 2 (OAS2)_forward (for): 5′-AACTGCTTCCGACAATCAAC-3′; OAS2_reverse (rev): 5′-CCTCCTTCTCCCTCCAAAA-3′; SOCS1_for: 5′-AACTGCTTTTTCGCCCTTA-3′; SOCS1_rev: 5′-GCCACGTAGTGCTCCA-3′; Viperin_for: 5′-CTTTGTGCTGCCCCTTGAGGAA-3′; Viperin_rev: 5′-CTCTCCCGGATCAGGCTTCCA-3′; TLR7_for: 5′-TTACCTGGATGGAAACCAGCTACT-3′; TLR7_rev: 5′-TCAAGGCTGAGAAGCTGTAAGCTA-3′; CD38_for: 5′-GGCCTGGGTGATACATGGTGGA-3′; CD38_rev: 5′-ACAGCGACTGGCTCAGATCTCA-3′.
cDNA and antisense RNA (aRNA) synthesis
Five micrograms of total RNA were used for each analyzed condition (control, 4 h IFN-β, 24 h LPS, 24 h LPS plus anti-IFNs). As internal reference for microarray analysis, a pool of RNA from iDC collected from five healthy donors was used. First and second strand cDNA synthesis and the synthesis of aRNA were performed using the Amino-Allyl MessageAmp aRNA Amplification kit (Agilent). The quality of aRNA was tested by using Agilent Bioanalyzer 2100. The reaction allows the incorporation of amino-allyl-modified NTP that are subsequently labeled with the N-hydroxysuccinimide (NHS) ester-labeled Cy3/5 dye (Amersham Biosciences). The reference aRNA and the aRNAs from the different samples were labeled with Cy5 and Cy3, respectively. The labeled aRNAs (750 ng) were fragmented and prepared for the hybridization using the Agilent In Situ Hybridization kit Plus (Agilent).
Microarray analysis
Microarrays were purchased from Agilent. Agilent’s Human 1A Oligo Microarray (V2) contains 20,173 oligonucleotide probes (60-mer) spanning conserved exons across the transcripts of the targeted full-length genes. Microarrays were hybridized at 60°C in the Agilent Oligo Microarray Hybridization chamber (22 K format) for 17 h under constant rotation. Following hybridization, slides were washed once for 1 min in 6× standard saline phosphate/EDTA (SSPE)-0.005% N-lauroylsarcosine at room temperature, then washed for 1 min with 0.06× SSPE-0.005% N-lauroylsarcosine. Microarrays were stabilized and dried with the Stabilization and Drying Solution (Agilent). Using a ScanArray lite scanner and ScanArray Express software (Packard Biochip Technologies; PerkinElmer) microarrays were scanned at 532 and 633 nm. The results and images were quantified using Quantarray software 2.1 (Packard Biochip Technologies; PerkinElmer). Raw data and sample information were then entered into Genespring (Sylicon Genetics). The LOWESS normalization was performed because in experiments where two fluorescent dyes (red and green) are used, intensity-dependent variation in dye bias may introduce spurious variations in the collected data. LOWESS normalization merges two-color data, applying a smoothing adjustment that removes such variation. Moreover, the Student t test was applied to compare the different conditions.
Cell extracts
Nuclear cell extracts were prepared as previously described (26). Briefly, cell pellets (5 × 106) were resuspended in 1 ml of buffer A (0.5% Nonidet P-40, 10 mM EDTA, 10 mM EGTA, 10 mM KCl, 10 mM HEPES (pH 7.9)) to which 1 mM DTT, 0.5 mM PMSF, 10 μg/ml leupeptin, 1 μg/ml pepstatin, 1 μg/ml antipain were freshly added) and incubated on ice for 10 min. Nuclei were sedimented by centrifuging the lysates at 1,200 × g for 10 min. The nuclear pellets were resuspended in 30–40 μl buffer C (1 mM EDTA, 1 mM EGTA, 0.4 M NaCl, 20 mM HEPES (pH 7.9), 5 mM MgCl2, 25% glycerol, with fresh addition as above) and incubated for 10 min on ice with occasional mixing. The suspensions were clarified by centrifuging at 15,000 × g for 10 min. The supernatants were recovered as nuclear extracts and were rapidly frozen on crushed dry ice and stored at −80°C. Whole cell extracts were prepared as previously described (24). Briefly, cells (107) were lysed in 30–50 μl of ice-cold whole cell extraction buffer (20 mM HEPES (pH 7.9), 50 mM NaCl, 0.5% Nonidet P-40, 1 mM DTT, 10 mM EDTA, 2 mM EGTA, 10 μg/ml leupeptin, 100 mM NaF, 0.5 mM PMSF, 10 mM sodium orthovanadate, and sodium molybdate). The lysate was incubated for 30 min on a shaker at 4°C and insoluble debris was removed by centrifugation (13,000 × g at 4°C for 10 min) and the lysate was stored at −80°C.
Flow cytometry analysis
Approximately 1–2 × 105 cells were aliquoted into tubes and washed once in PBS containing 2% FCS. The cells were incubated with mAbs at 4°C for 30 min. The DC were then washed and fixed with 2% paraformaldehyde before analysis on a FACScan using CellQuest software (BD Biosciences). The change in expression of the cell surface molecules was determined by subtracting the median fluorescent intensity (MFI) of the isotype control Ab from the MFI from the specific Ab.
Cytokine and chemokine determinations
Western blot analysis
Western blot was performed as previously described (24). Briefly, 30 μg of whole cell extracts were separated by 10% SDS-PAGE gel and blotted onto nitrocellulose membranes. Blots were incubated with the indicated Abs: IRF-7 Ab (Santa Cruz Biotechnology), TLR7 Ab (Novus Biologicals), or β-actin Ab (Sigma-Aldrich) and reacted with anti-rabbit HRP-coupled secondary Ab using an ECL system (Amersham Biosciences).
In silico analysis of human TLR7 promoter
The analysis of the human promoter of the TLR7 gene (accession NM_016562.3) was done by using the transcription factor binding site (TFBS) prediction program MatInspector (online at www.genomatix.de/matinspector.html). MatInspector uses informations of MatInd (that generates position weight matrices) to scan nucleotide sequences for matches to this pattern by calculating a matrix similarity score that reaches 1 only if the test sequence corresponds to the most conserved nucleotide at each position of the matrix. The analyzed sequence was a region spanning 1000 bp upstream the transcription starting site (TSS) (27).
EMSA
EMSA was conducted as previously described (26). Briefly, to measure the association of DNA-binding proteins with different DNA sequences, synthetic double-stranded oligonucleotides were end-labeled with [γ-32P]ATP by T4 polynucleotide kinase. For the analysis of IRF-1 complexes, nuclear cell lysates (15 μg) were used in EMSA experiments. For supershift analysis, 1 μg of anti-IRF-1 Ab (Santa Cruz Biotechnology) was added to the reaction. The oligonucleotide used was as follows: IRF-1 site-TLR7 5′-CTATAAAAACGAAAGAAATTTGGT-3′. For competition analysis, a 100-fold molar excess of cold IRF-1 site-TLR7 and unrelated oligonucleotides was added to the binding reaction.
Plasmid and transfection assays
Rc-CMV empty vector and Rc-CMV IRF-1 have been previously described (28). All DNA plasmids were prepared using Endofree Plasmid Maxi kits (Qiagen). DC were transfected using Nucleofector technology according to the manufacturer’s instruction (Amaxa Nucleofector) as previously described (29). Twenty hours after transfection, total RNA was extracted and analyzed for TLR7 expression by real-time RT-PCR.
Statistical analysis
Statistical analysis was calculated using a two-tailed for paired data Student’s t test. A p value <0.05 was considered statistically significant.
Results
Defining type I IFN-dependent response in monocyte-derived DC following TLR4 triggering
Before analyzing the IFN-β-related functional and phenotypical modifications occurring in DC following LPS treatment, initial studies were designed to test the efficacy of type I IFN neutralization in our experimental model. To this aim, we monitored IRF-7 expression by Western blotting because it is known that IRF-7 expression occurs following type I IFN release during most viral and bacterial infections (30, 31). Whole lysates were prepared from control iDC, or DC that had been treated for 24 h with IFN-β or LPS in presence or absence of anti-IFN Abs. These time points were judged to provide the best snapshot of IFN-mediated changes in LPS-matured DC and were used for all subsequent experiments. As expected, IRF-7 protein expression was strongly up-regulated after IFN-β and LPS treatments (Fig. 1⇓A). Moreover, the presence of anti-IFN Abs completely inhibited the LPS-induced IRF-7 expression, proving the requirement of IFN-β production for its induction and the ability of these neutralizing Abs to dramatically inhibit the autocrine/paracrine type I IFN-mediated intracellular signaling. No reduction of IRF-7 expression was observed by the addition of sheep serum to LPS-treated cultures (data not shown). Equal loading of all samples was verified by immunoblotting with an anti-β-actin Ab (Fig. 1⇓A, lower panel).
Type I IFN-dependent functional modifications in DC during LPS-induced maturation. A, iDC were left untreated (Ctr), stimulated for 24 h with IFN-β (200 pM), with LPS (1 μg/ml) in presence or absence of anti-IFN Abs. Whole cell extracts were prepared and 30 μg of lysates subjected to immunoblot analysis using anti-IRF-7 or anti-β-actin specific Abs. This is a representative experiment performed with DC obtained from other two donors. B, Supernatants were harvested from control cells (Ctr), LPS-treated DC with or without anti-type I IFN Abs or sheep antiserum (serum) as control for 24 h. The level of CXCL-10 was measured by ELISA. The results represent the means ± SE of three different experiments; p < 0.05 LPS vs LPS + anti-IFN Abs. C, Surface expression of CD38 was evaluated by FACS in iDC (Ctr) or LPS-matured DC in presence or absence of anti-IFN Abs. A total of 5000 cells were tested per sample. A representative FACS profile is shown. The analysis was repeated three additional times, using DC from a total of four different blood donors. α, anti.
We next monitored the production of the IFN-inducible CX chemokine ligand 10 (CXCL10) by ELISA. Matured DC produce CXCL10 and this has been shown to be IFN-regulated (32, 33, 34). DC were matured with LPS, with or without anti-IFN Abs, for 24 h and supernatants were then harvested. CXCL10 release, observed following LPS treatment, was significantly decreased in DC matured with LPS in presence of anti-IFN Abs (Fig. 1⇑B). To further confirm that the activity of these neutralizing Abs was specific, DC were treated with LPS in presence of sheep control serum; in this condition, no decrease of CXCL10 production was observed (Fig. 1⇑B).
Furthermore, we investigated whether the profile of CD38, a surface molecule known to be induced also by type I IFN (35, 36) was affected by anti-IFN Abs. Surface expression of CD38 was examined using flow cytometry analysis and as expected a marked reduction in the LPS-induced increase in CD38 expression was observed (Fig. 1⇑C).
Analysis of IFN-β-induced gene expression in LPS-matured DC
To evaluate the extent to which IFN-β contributes to the LPS-induced gene expression profile, we next performed microarray analysis in DC following 4 h of treatment with IFN-β or 24 h of treatment with LPS in the presence or absence of anti-IFN Abs. The RNA was collected from four different healthy donors and similar gene expression profiles were found between donors. aRNAs were hybridized to Agilent’s Human 1A Oligo Microarray (version 2) that contains 22 K oligonucleotide probes and the data were analyzed with Gene Spring Software. Because one of the principal goals of our study was to elucidate a specific list of genes regulated by type I IFN production during LPS treatment, we began our analysis by profiling the changes in gene expression that occur in DC in response to 4 h IFN-β treatment compared with that observed after 24 h stimulation with LPS. A total of 218 genes were found to be modulated by IFN-β and 315 genes by LPS. We focused our attention on the genes that were up-regulated by the two different treatments and which were down-modulated in the presence of anti-IFN Abs. The expression profile of 23 genes followed this criteria and 5 of these were validated by real-time RT-PCR (see below). A hierarchical clustering analysis of this set of genes is shown in Fig. 2⇓A. Some of these genes are part of the so-called “IFN signature,” a set of genes known to be regulated by type I IFN signaling (37), such as the OAS 1 and 2 (38), suppressor of cytokine signaling (SOCS) 1 (39, 40), protein kinase R (PKR) (41) and Viperin (42). Others were ISGs with unknown function but which have been shown to be induced via IFN production including IFN-induced protein 35 and IFN-induced protein with tetratricopeptide repeats 5 (37). We validated the results obtained by microarray analysis by quantifying the mRNA level of OAS2, SOCS1, and Viperin by real-time RT-PCR (Fig. 2⇓B). Moreover, the CD38 profile of expression, observed by FACS analysis (see Fig. 1⇑C), was confirmed also at the mRNA level by microarray (Fig. 2⇓A) and by real-time RT-PCR (Fig. 2⇓B).
Specific transcriptional profile induced by type I IFN release along LPS-induced DC maturation. Total RNA was isolated from DC either following 4 h IFN-β or 24 h LPS treatment in presence or absence of anti-IFN Abs. A, Hierarchical clustering of gene expression data by four different healthy donors is shown. Each row represents a separate gene and each column a different condition, as indicated. The normalized expression index for every transcript sequence (rows) in each sample (columns) is indicated by a color code. Red, yellow and green squares indicate that expression of the gene is greater than, equal to or less than the mean level of expression across the different donors. The scale extends from fluorescent ratios of 0.25 to 5.0. B, mRNA expression of OAS2, SOCS1, Viperin, CD38, and GAPDH was analyzed by real-time RT-PCR. Data were then normalized with GAPDH level of expression. Results are means ± SE of triplicate values. This is a representative experiment of four different analyzed donors. α, anti.
Interestingly, IFN-β and LPS treatments also induced the expression of a subset of genes which have not previously been reported to be IFN-inducible in DC. Importantly, two of these were PRRs including nucleotide-binding oligomerization domain 2 (NOD2) and TLR7 (Fig. 2⇑A). Whereas NOD2 (also called CARD15) plays a central role in the immune response to intracellular bacteria by recognizing muramyl dipeptide (MPD), a derivative of peptidoglycan (43), TLR7 recognizes ssRNA and is activated by infections with ssRNA viruses (9, 10). The TLR7 gene has been reported to be induced by type I IFN in monocytes/macrophages or in B cells (44, 45), but its inducibility by type I IFN has never been characterized in the context of DC maturation.
The expression of defensin-β1 (DEFB1) was also found to be IFN regulated. DEFB1 is an antimicrobial peptide implicated in the resistance of epithelial surfaces to microbial colonization and it has already been described that IFN-γ treatment or TLR3 and 4 triggering can drive its expression (46, 47).
Type I IFN-dependent regulation of TLR7 expression in LPS-treated DC
We focused our attention on the modulation of TLR7 expression to understand its regulation in mDC. Indeed, in humans the expression of TLR7, together with TLR9, is known to be restricted only to the endosomes of the pDC but not of the mDC, whereas TLR8 expression is constitutively expressed in mDC, albeit at low levels (5, 13). We confirmed the results obtained by microarray analysis measuring TLR7 mRNA level by real-time RT-PCR either in iDC or IFN-β- and LPS-treated DC in presence or absence of anti-IFN Abs (Fig. 3⇓A). Whole cell lysates were prepared from DC following 24 h of treatment with IFN-β or with LPS in the presence or absence of anti-IFN Abs. TLR7 protein expression was quantified by Western blotting (Fig. 3⇓B). Consistent with mRNA expression data, TLR7 protein levels were also induced following stimulation with LPS and were inhibited when anti-IFN Abs were added to the cultures, further proving a clear dependence on IFN production and signaling in the regulation of TLR7 by LPS. Equal loading of all samples was verified by immunoblotting with an anti-β-actin Ab (Fig. 3⇓B, lower panel).
TLR7 expression by LPS-induced IFN-β production. A, TLR7 gene expression was quantified by real-time RT-PCR in total RNA extracted from iDC (Ctr) or DC treated with IFN-β, LPS, or LPS + anti-IFN Abs as described in the legend of Fig. 2⇑. B, Whole cell lysates collected from cells treated for 24 h as described in (A) were analyzed by immunoblotting using a specific anti-TLR7 Ab. This analysis was repeated two additional times using DC from a total of three different blood donors. α, anti.
We next examined whether the activation of other TLRs expressed in DC could modulate TLR7 gene expression as we had observed following LPS stimulation. DC were stimulated for 24 h with synthetic bacterial lipopeptide (Pam3cysk4) and with synthetic dsRNA (poly I:C) which engage TLR2 and 3, respectively. RNAs were prepared and TLR7 mRNA levels were analyzed by quantitative PCR (Fig. 4⇓A). Poly I:C and LPS treatments both strongly induced TLR7 expression, in contrast to Pam3Cysk4 which did not stimulate TLR7 expression consistent with a failure of TLR2 ligands to induce type I IFNs (48).
Transcriptional regulation of the TLR7 gene. A, Total RNA was extracted from nonstimulated (Ctr), Pam3cysk4 (2 μg/ml), poly I:C (pI:C) (50 μg/ml), and LPS (1 μg/ml) treated DC upon 24 h treatment. Real-time RT-PCR was performed to measure TLR7 mRNA level. One representative of three different experiments is shown. B, Schematic map of ∼600 bp of the human TLR7 promoter. TATAAA is the consensus sequence for the TATA box. C, DC were treated with IFN-β or LPS with or without anti-IFN Abs for 4 h. Nuclear extracts (15 μg) were prepared and analyzed by EMSA using a specific radiolabeled oligonucleotide corresponding to the IRF-1 binding site present within TLR7 promoter. Supershift assay (SS) was performed using an anti-IRF-1 Ab, where indicated. This experiment was repeated additional two times. D, Nuclear extracts from LPS-treated DC (4 h) were used for competition studies where 100× excess of cold oligonucleotides containing the putative IRF-1 consensus sequence within the TLR-7 promoter and a sequence unrelated (unr) to IRF-1 were added to the binding reaction. E, Total RNA was also extracted from DC transfected with vectors expressing IRF-1 or with an empty vector (EV). TLR7 expression was evaluated by quantitative PCR. This is a representative experiment performed additional two times. α, anti.
This differential activation of the TLR7 gene prompted us to examine its promoter region for TFBS. Using the TFBS prediction program MatInspector (27), a region spanning 1000 bp upstream and 500 bp downstream of the TSS was analyzed. Interestingly, TFBS for IRF-1, STAT1, and NF-κB were detected within 600 bp of the TSS (Fig. 4⇑B). To probe the involvement of each of these factors to the regulation of TLR7 gene induction, oligonucleotides corresponding to the binding sites for IRF-1, STAT1, and NF-κB from the TLR7 promoter were designed. Nuclear extracts were prepared following stimulation of DC with IFN-β and LPS in the presence or absence of anti-IFN Abs and analyzed by EMSA for the specific binding of each factor to the respective site. No STAT1 binding and only a slight signal in the case of NF-κB were detected (data not shown). Conversely, a clear induction of IRF-1 DNA-binding activity was observed in LPS- and IFN-β-treated cells and this binding activity was completely lost when anti-IFN Abs were added (Fig. 4⇑C). The identity of the complexes was confirmed by supershift experiments using Abs raised against IRF-1. Moreover, complex formation was also inhibited by the addition of an excess of unlabeled oligonucleotide containing the putative IRF-1 consensus sequences within TLR7 promoter when nuclear extracts from IFN-β-stimulated (data not shown) or LPS-treated DC were used (Fig. 4⇑D).
We also examined the effect of IRF-1 on TLR7 mRNA expression in DC transiently transfected using nucleofection with a vector expressing IRF-1 or with an empty vector. As shown in Fig. 4⇑E, the expression of TLR7 was clearly increased by the presence of IRF-1 compared with DC transfected with the empty vector. All together, these results indicate a key role for IRF-1 in the regulation of TLR7 along LPS-induced DC maturation.
Maturation of IFN-β-pretreated DC by a specific TLR7 agonist
Although TLR7 and TLR8 are phylogenetically and structurally related, their selective activation leads to completely different immune responses. It has been shown that the small molecule imidazoquinoline analogues 3M-001 and 3M-002 are selective ligands for TLR7 and TLR8, respectively (13). 3M-001 directly activates pDC and, to a lesser extent, monocytes; conversely, 3M-002 activates monocytes and mDC. Because our findings support a model whereby TLR7 gene and protein expression are induced following type I IFN release, we examined whether a 24 h pretreatment of DC with IFN-β would enable these cells to respond to TLR7 ligands. Thus, IFN-β-pretreated DC were stimulated for 24 h with TLR7 agonist and expression of CD83, CD86, and CD38 surface markers were evaluated by flow cytometry analysis to assess the activation status of the cells (Table I⇓). We observed that a prestimulation of DC with IFN-β provoked the maturation of the 3M-001-treated DC, phenomenon not observed in the absence of IFN-β treatment. As expected, TLR8 triggering alone induced an up-regulation of the considered markers and IFN-β by itself caused an increase of CD38 and CD86.
Induction of surface markers by TLR7 triggering in IFN-β-pretreated DCa
Furthermore, the secretion of the proinflammatory cytokines, such as TNF-α and IL-6, and of the regulatory cytokines, IL-12p70, IL-10, and IFN-α was analyzed. Supernatants were harvested from DC treated for 24 h with the TLR7 agonist 3M-001 with or without 24 h IFN-β pretreatment. 3M-002-stimulated DC were used as our internal control and a strong production of all the analyzed cytokines was observed upon TLR8 triggering with the exception of the poor release of IFN-α (Fig. 5⇓). Conversely, in the pretreated DC, a clear induction of IL-12, IL-6, TNF-α, and IL-10 production was detected following TLR7 activation compared with DC treated with 3M-001 alone. Only a scant production of IFN-α was observed upon 3M-001 stimulation in DC pretreated with IFN-β compared with the levels which accumulated in the culture medium following FLU virus infection.
Analysis of cytokine production in IFN-β pretreated DC following TLR7 triggering (A–E). Cell culture supernatants were harvested from DC treated for 24 h with the TLR7 agonist 3M-001 with or without 24 h pretreatment with IFN-β, or stimulated either with 3M-002 or IFN-β alone. The level of IL-12 (A), TNF-α (B), IL-6 (C), or IL-10 (D) were measured by CBA. IFN-α production (E) was evaluated by ELISA. Results are means ± SE from four different experiments; p < 0.05 3M-001 vs IFN-β + 3M-001.
Discussion
DC orchestrate innate and adaptive immune responses due to their ability to sense a diversity of danger signals via PRRs such as TLRs, and to initiate potent T cell responses (49, 50). Understanding how distinct signals can be activated in the context of infection in response to a wide array of invading pathogens is a very intriguing challenge.
In the past, we and other groups have demonstrated that the activation of certain TLRs can lead to the transcriptional regulation of type I IFN subtypes (18, 19, 20, 21) and this autocrine IFN may have critical effects on the biology of DC. Type I IFNs are cytokines that play a central role in host immunity, in fact they are expressed by many cell types in response to viral or microbial infections (21, 51). By binding to specific transmembrane receptors, they trigger a response that culminates in the induction of a large number of genes modulating and linking the innate and the adaptive immune responses. Because TLR4, triggered by its ligand LPS, induces IFN-β as a “signature” cytokine (21, 48, 52), in the present study we have investigated the IFN-β-related transcriptional and subsequent functional consequences that can occur in human monocyte-derived DC along LPS-induced maturation. To this end, using gene expression profiling, we identified a set of 23 genes that were induced in DC following IFN-β and LPS treatments and down-regulated by neutralizing the autocrine IFN-β produced following LPS stimulation by adding anti-IFN Abs (Fig. 2⇑). Among the induced ISGs, several of these were well known ISGs with antiviral activity. Additional genes were identified, which have not previously been characterized as ISGs. We found that different sensors for pathogen-derived molecules, such as TLR7 as well as NOD2 and DEFB1, were induced in a type IFN-dependent manner. Both NOD2 and DEFB1 are involved in the resistance to bacterial pathogens. NOD2 (also called CARD15) is a member of the phylogenetically conserved NACHT-LRR (NLR) family and plays a role in the immune response to intracellular bacteria by recognizing MPD, a peptidoglycan component (43); DEFB1 is an antimicrobial peptide implicated in the resistance of epithelial surfaces to microbial colonization and belongs to the family of defensins, microbicidal and cytotoxic peptides made by neutrophils (53). Despite the great importance of the NOD pathway and the role of defensins in the innate immune response, our interest especially focused on characterizing TLR7 gene regulation in DC, because very little is known about it. Indeed, while the TLR7 gene has been reported to be induced by type I IFN in monocytes/macrophages or in B cells (44, 45, 54), its inducibility by autocrine release of type I IFN has never been characterized in the context of DC maturation.
TLR7 and TLR8 belong to the “TLR9 subfamily” (55). The members of this subfamily of TLRs (TLR3, TLR9, TLR7, and TLR8) are evolutionary related and sense viral and bacterial nucleic acids at the endosomal subcellular compartment. This is in contrast to the other TLRs, such as TLR2 and TLR4, which are localized at the cell surface. Initially, TLR7 and 8 were shown to trigger IFN production in response to the imidazoquinolines, imiquimod and resiquimod (or R-848), low m.w. immune response modifiers with potent antiviral and antitumor properties (11). Subsequently, the immunostimulatory action of several additional guanine nucleoside analogs has been shown to be controlled exclusively via TLR7 (56). Only later, then, the natural ligands of TLR7 and 8 were identified in the virus-derived ssRNAs, specifically guanosine- and uridine-rich ssRNA oligonucleotides derived from HIV-1 (10) or ssRNA derived from wild-type influenza virus and synthetic ssRNA (polyU) (57). Moreover, only very recently, the possible role of TLR7 in recognizing “self” RNA, acting in this way as a sensor of endogenous “danger signals,” has been described. In fact, the recognition of U1 small nuclear RNA, that is part of the small nuclear ribonucleoproteins, one of the major component of the immune complexes associated with the pathogenesis of the autoimmune disease systemic lupus erythematosus, is dependent on TLR7 (58). All these findings highlight the importance of understanding the regulation of expression and the consequent activity of TLR7.
It has been described that, in different human DC subpopulations, only pDC express TLR7, whereas mDC constitutively express TLR8 (5, 6, 7). Here, we showed that TLR7 gene expression can be activated in mDC upon TLR4 stimulation in a type I IFN-dependent manner via a mechanism which is dependent upon IRF-1. Consistent with these findings, TLR3 triggering was also able to increase TLR7 mRNA levels, likely through the release of IFN-β, whereas TLR2 stimulation failed to do so (Fig. 4⇑). These findings are consistent with a failure of TLR2 to activate IFN-β production (48).
A key question arises from these observations: can IFN-β priming of mDC render them responsive to specific TLR7 agonists? To answer this question, we pretreated immature DC with IFN-β and then stimulated with the small molecule imidazoquinoline analogues 3M-001 and 3M-002, selective ligands for TLR7 and TLR8, respectively (13). We demonstrated that the TLR7 gene is not only induced in these conditions, but is also functional. In fact, 3M-001 treatment, which selectively activates TLR7 was able to induce the expression of maturation and activation markers, like CD83, CD86 and CD38, in IFN-β-treated DC but not in iDC (Table I⇑). Therefore, a robust enhancement in proinflammatory as well as regulatory cytokines (IL-12, IL-6, and TNF-α) was observed (Fig. 5⇑). Importantly, the acquisition of TLR7 expression by mDC does not confer upon these cells the ability to produce significant amounts of type I IFN following stimulation with specific agonists (Fig. 5⇑).
The scenario detailed in our study is consistent with recent reports demonstrating that combined TLR ligation of DC triggers distinct signaling pathways resulting in an enhanced activation of these cells in a synergistic manner in terms of cytokine production and expression of costimulatory molecules (59, 60, 61). The exogenous IFN-β can mimic, in our case, the endogenous IFN-β production stimulated by TLR3 and TLR4 triggering, which has been shown in other studies to potently act in synergy with endosomal TLRs (60). This synergy can also be explained by our findings that highlight the ability of TLR3 and TLR4-activated pathways to induce, in a type I IFN-dependent manner, sensors normally absent in iDC and ensuring in this way that powerful effectors will be generated to respond combinatorially to microbial products in different cellular compartments. Another explanation supporting this hypothesis is that TLR4 can participate in the immune response to certain viruses (62, 63, 64, 65, 66, 67); the virus-induced triggering of TLR4 can mediate the transcriptional activation of certain genes encoding intracellular sensors for viruses, such as TLR7, to amplify the response to specific viral infections.
Our findings might be relevant in understanding the complex activation of the IFN-induced intracellular pathways occurring in DC during pathogen infections and might be helpful in developing new DC-based strategies against persistent viral infections.
Acknowledgments
We are grateful to Rosella Mechelli (Neurology and Center for Experimental Neurological Therapy, S. Andrea Hospital, “La Sapienza” University of Rome, Rome, Italy) and Marta Scandurra (Department of Infectious, Parasitic, and Immuno-Mediated Diseases, Istituto Superiore di Sanità, Rome, Italy) for helpful discussion and to Katherine A. Fitzgerald (University of Massachusetts Medical School, Worcester, MA) for critical reading of the manuscript. We also thank Gillez Uzé (Centre National de la Recherche Scientifique, Unité Mixte de Recherche 5124, Montpellier, France) and Ilkka Julkunen (National Public Health Institute, Helsinki, Finland) for providing reagents and Eugenio Morassi for preparing drawings.
Disclosures
M. Tomai is a current employee of 3M Pharmaceuticals.
Footnotes
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.
↵1 This work was supported by grants from the Istituto Superiore di Sanità-National Institutes of Health Program (No. 5303), from Istituto Superiore di Sanita (No. 5AD/F2) and from Fondazione Italiana Sclerosi Multipla (Cod. 2005/R/7) to E.M.C.
↵2 Address correspondence and reprint requests to Dr. Eliana M. Coccia, Department of Infectious, Parasitic, and Immuno-mediated Diseases, Istituto Superiore di Sanità, Viale Regina Elena 299-00161 Rome, Italy. E-mail address: e.coccia{at}iss.it
↵3 Abbreviations used in this paper: DC, dendritic cell; iDC, immature DC; PRR, pattern recognition receptor; pDC, plasmacytoid DC; mDC, myeloid DC; TIR, Toll-IL-1R; TRIF, TIR-domain containing adapter inducing IFN-β; IKK, IκB kinase; IRF, IFN regulatory factor; MFI, median fluorescent intensity; ISGF3, IFN-stimulated gene factor 3; ISG, IFN-stimulated gene; aRNA, antisense RNA; SOCS, suppressor of cytokine signaling; OAS, 2′-5′-oligoadenylate synthetase; CBA, cytometric bead array; TFBS, transcription factor binding site; TSS, transcription starting site; DEFB1, defensin-β1; NOD, nucleotide-binding oligomerization domain; MDP, muramyl dipeptide.
- Received December 23, 2006.
- Accepted March 9, 2007.
- Copyright © 2007 by The American Association of Immunologists