Surface molecules of pathogens play an important role in stimulating host immune responses. Elucidation of the signaling pathways activated by critical surface molecules in host cells provides insight into the molecular pathogenesis resulting from bacteria-host interactions. MspTL is the most abundant outer membrane protein of Treponema lecithinolyticum, which is associated with periodontitis, and induces expression of a variety of proinflammatory factors. Although bacteria and bacterial components like LPS and flagellin are known to induce IFN-β, induction by bacterial surface proteins has not been reported. In the present study, we investigated MspTL-mediated activation of signaling pathways stimulating up-regulation of IFN-β and IFN-stimulated genes in a human monocytic cell line, THP-1 cells, and primary cultured human gingival fibroblasts. MspTL treatment of the cells induced IFN-β and the IFN-stimulated genes IFN-γ-inducible protein-10 (IP-10) and RANTES. A neutralizing anti-IFN-β Ab significantly reduced the expression of IP-10 and RANTES, as well as STAT-1 activation, which was also induced by MspTL. Experiments using specific small interfering RNA showed that MspTL activated TANK-binding kinase 1 (TBK1), but not inducible IκB kinase (IKKi). MspTL also induced dimerization of IFN regulatory factor-3 (IRF-3) and translocation into the nucleus. The lipid rapid-disrupting agents methyl-β-cyclodextrin, nystatin, and filipin inhibited the MspTL internalization and cellular responses, demonstrating that lipid raft activation was a prerequisite for MspTL cellular signaling. Our results demonstrate that MspTL, the major outer protein of T. lecithinolyticum, induced IFN-β expression and subsequent up-regulation of IP-10 and RANTES via TBK1/IRF-3/STAT-1 signaling secondary to lipid raft activation.
Periodontitis is one of the most prevalent oral diseases. It is a polymicrobial infection characterized by gingival inflammation and alveolar bone resorption that often leads to tooth loss. Oral Treponema species are present in high numbers in periodontitis patients and correlate with disease severity (1, 2, 3). They are highly heterogeneous and multiple species can be found in a single patient or single diseased site. Treponema lecithinolyticum was recently isolated and has been reported to be associated with aggressive periodontitis and root canal infections (2, 4, 5). The bacterium induces matrix metalloproteinase-2 activation in human gingival fibroblasts (HGFs)3 and periodontal ligament (PDL) cells (6), as well as osteoclastogenesis in cocultures composed of mouse calvaria and bone marrow cells (7). These observations suggest bacterial involvement in the destruction of soft and hard tissues in the periodontium. To elucidate the underlying molecular pathogenesis, we recently analyzed the function of the major outer membrane (OM) protein of T. lecithinolyticum, 62-kDa MspTL, in THP-1 and PDL cells (8, 9). MspTL induced the expression of proinflammatory factors, including IL-1β, TNF-α, IL-6, IL-8, PGE2, and ICAM-1 in an NF-κB-dependent fashion.
IFNs are the first known members of the cytokine family of proteins to induce an antiviral state, playing a pivotal role in innate antiviral immune responses. They also have various other functions including regulation of cell growth and cell motility through induction of IFN-stimulated genes (ISGs). IFNs consist of type I, II, and III IFNs, which interact with type-specific receptor complexes (10). Type I IFNs, such as IFN-α and IFN-β, are induced by viral infection and their major roles involve antiviral and immunomodulatory activities. Bacteria and bacterial components can also induce type I IFNs. Pathogen-associated molecular patterns (PAMPs), such as LPS, flagellin, and unmethylated CpG DNA, are well-known bacterial ligands for IFN-β induction via TLR4, TLR5, and TLR9, respectively (11, 12, 13, 14, 15). Recent research demonstrated that facultative intracellular bacteria such as Chlamydia trachomatis, Listeria monocytogenes, and Mycobacterium tuberculosis also induce IFN-β (16, 17, 18). In contrast to its beneficial effects following viral infection, IFN-β appears to have both beneficial and detrimental effects in relationship to intracellular bacterial infections. Some signaling molecules involved in the induction of type I IFNs are ubiquitous components of response, but other molecules are distinct according to the infectious microorganisms or their molecular Ags. While the known bacterial PAMPs use TLRs, intracellular bacteria do not use these molecules as major receptors (19).
Understanding the molecular basis of host response to bacterial infection is crucial for elucidating molecular pathogenesis and mechanisms of controlling bacterial infections. The OM of Gram-negative bacteria is the outmost barrier to the host cells. Molecules in the OM elicit diverse activities such as adhesion, cytotoxicity, antigenicity, and other cell-stimulating activities. LPS has been one of the most intensively studied OM molecules, and it belongs to well-known PAMPs. However, in our preliminary study, phenol-water extracts of T. lecithinolyticum, which contained LPS-like molecules, did not show any cell-stimulating activity (20). Since multiple bacterial species are known to be associated with periodontitis, functional elucidation of representative major OM proteins from various periodontopathogens is required to understand the unique and common elements of molecular pathogenesis. These data may provide effective strategies for controlling multicomplex microbial infections.
Our preliminary study using a gene microarray showed that MspTL from T. lecithinolyticum induced many ISGs. In this study, we investigated the signaling pathway by which T. lecithinolyticum MspTL induces IFN-β and ISGs in a monocytic cell line, THP-1 cells, and in primary cultured human HGFs. MspTL induced the expression of IFN-β via TANK-binding kinase 1 (TBK1) and IFN regulatory factor-3 (IRF-3). MspTL-induced IFN-β resulted in STAT-1 phosphorylation and the induction of ISGs such as IFN-γ-inducible protein-10 (IP-10) and RANTES. Using reporter cell lines, a neutralizing IFN-β Ab, specific small interfering RNA (siRNA), lipid raft inhibitors, and inhibitors of endosomal maturation, we determined that MspTL-mediated induction of IFN-β requires lipid raft activation, but is independent of TLR2, TLR4, MyD88, and endosomal acidification.
Materials and Methods
Reagents and immunochemicals
Pyrrolidine dithiocarbamate (PDTC), methyl-β-cyclodextrin (MβCD), nystatin, filipin, bafilomycin, chloroquine, polymyxin B, and Escherichia coli LPS (0127:B8) were purchased from Sigma-Aldrich. Helenalin was purchased from BIOMOL Research Laboratories. CpG-ODN (CpG-oligodeoxynucleotide) 2006 and palmitoyl-3-cysteine-serine-lysine-4 (Pam3CSK4) were from InvivoGen. Abs were purchased from the following sources: rabbit anti-human IFN-β Ab (31410-1) and HRP-conjugated anti-rabbit IgG were from R&D Systems; FITC-conjugated mouse anti-human CD25 and mouse anti-human TLR4 mAb (HTA 125) were from BD Biosciences; anti-STAT-1 Ab, anti-phospho-STAT-1 (Tyr701) Ab, anti-STAT-3 Ab, and anti-phospho-STAT-3 (Tyr705) Ab were from Cell Signaling Technology; mouse anti-IRF-3 mAb and goat polyclonal Abs to lamin B were from Santa Cruz Biotechnology; polyclonal rabbit anti-IRF-7 Ab was from Abcam; mouse anti-human TLR2 mAb (TL2.1) was from eBioscience; and goat polyclonal Abs to β-actin were from Sigma-Aldrich. Small interfering RNAs (siRNA for inducible IκB kinase (IKKi), TBK1, MyD88) and control A siRNA were from Santa Cruz Biotechnology. MspTL was labeled with the fluorochromes such as Alexa 488 (green fluorescence) and Alexa 568 (red fluorescence) from Invitrogen, according to the manufacturer’s instructions. ELISA kits to determine level of IFN-β, IP-10, and IL-8 were purchased from R&D Systems. All of the other reagents and chemicals used were of analytical grade.
Cell culture and recombinant MspTL preparation
For experiments examining host response, we used THP-1 cells (American Type Culture Collection TIB-202), a human monocytic cell line, and primary cultured HGFs prepared as described previously (6). THP-1 cells were maintained in RPMI 1640 medium supplemented with l-glutamine (Invitrogen), 10% heat-inactivated FBS, and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin sulfate). Primary cultured HGFs were maintained in DMEM supplemented with 10% FBS and antibiotics. Recombinant MspTL preparation and endotoxin decontamination were performed as described previously (8). Endotoxin activity in the MspTL preparation was determined by the Limulus amebocyte lysate (LAL) assay using a LAL endochrome kit (Charles River Endosafe) according to the manufacturer’s protocol. For a mock control, E. coli transformed with pQE-30 without the insert DNA was cultured and cell extracts were prepared as described previously for the control of endotoxin contamination during MspTL preparation (8).
Treatment of host cells with MspTL and T. lecithinolyticum
THP-1 cells were seeded and cultured in 35-mm culture dishes at a concentration of 1 × 106 cells/ml. HGFs were seeded in 35-mm culture dishes at a concentration of 2 × 105 cells/ml and cultured until they reached confluence. After changing to serum-free medium, the cells were stimulated with MspTL (0.1–10 μg/ml), T. lecithinolyticum at different multiplicities of infection (MOIs), and LPS (10 μg/ml) for 2, 4, 8, or 12 h. The cells were harvested and used for RNA isolation to perform real-time RT-PCR for IFN-α, IFN-β, IFN-γ, IP-10, RANTES, and IL-8. Conditioned culture medium was collected and stored at −70°C for the measurement of IFN-β, IP-10, and IL-8 by ELISA. To see the neutralizing effect of anti-MspTL, THP-1 cells were treated with MspTL or T. lecithinolyticum, which was preincubated with the Abs for 30 min. Nontreated cells were used as a negative control.
In some experiments, THP-1 cells and HGFs were pretreated with the NF-κB activation inhibitors PDTC (40 μM) and helenalin (10 μM), the lipid raft inhibitors MβCD (1 mM), nystatin (10 μg/ml), and filipin (1 μg/ml) for 1 h before addition of MspTL. To elucidate the effect of IFN-β induced by MspTL treatment on cytokine expression, THP-1 cells were treated with MspTL in the presence of a rabbit anti-human IFN-β Ab. Rabbit IgG was used as an isotype Ab control. To test whether MspTL directly induced IFN-β or indirectly via proinflammatory mediators released by stimulated host cells, THP-1 cells were treated with the protein synthesis inhibitor cycloheximide (10 μM) for 1 h before MspTL stimulation. To examine the effects of MspTL on endosomal maturation, THP-1 cells were pretreated with bafilomycin (30 nM) and chloroquine (25 μM), inhibitors of endosomal maturation, for 1 h before MspTL treatment. CpG-ODN 2006 (100 nM) was used as a positive immunostimulant to observe endosomal maturation.
Since we used MspTL that was cloned and expressed in E. coli, we rigorously examined the cell-stimulating activity of this MspTL to confirm that the activity was not a result of E. coli endotoxin contamination. THP-1 cells were treated with heat-inactivated MspTL (95°C for 30 min) or mock extract, and the cellular activity was analyzed using ICAM-1 expression as described previously (8). THP-1 cells were also treated with MspTL (10 μg/ml) or E. coli LPS (10 μg/ml) in the presence of an endotoxin inhibitor, polymyxin B (50 μg/ml), for 12 h and then assessed for ICAM-1 expression. Finally, LPS and TLR2 ligand contamination was also evaluated using NF-κB reporter cell lines, CHO/CD14/TLR4 and CHO/CD14/TLR2 cells, as described previously (21). These cells have the gene-encoding membrane CD25 with the human E-selectin promoter, which contains NF-κB binding sites. These cells were plated in 6-well plates and cultured in Ham’s F-12 medium (Invitrogen) supplemented with 10% defined FBS (HyClone), 1 mg/ml G418 (Calbiochem), and 400 U/ml hygromycin B (Calbiochem) at 37°C and 5% CO2. When the cells were 80% confluent, MspTL, E. coli LPS as a TRL4 ligand, or Pam3CSK4 as a TLR2 ligand were added for 18 h at 37°C in a CO2 incubator. The cells were then washed once with PBS, detached with 2 mM EDTA in PBS, and incubated with FITC-conjugated mouse anti-human CD25 at 4°C for 30 min. After washing with PBS, 15,000 cells were analyzed on a FACSCalibur flow cytometer (BD Biosciences). Isotype-mached IgG was used as a control for nonspecific binding.
RNA preparation and real-time RT-PCR
Total RNA from THP-1 cells and HGFs treated with various stimulators and/or inhibitors was isolated using the TRIzol reagent (Invitrogen). cDNAs were synthesized by mixing RNA (1 μg) and Maxime RT PreMix (iNtRON) in a 20-μl reaction volume and incubating the mixture at 45°C for 1 h. After heating at 95°C for 5 min, cDNAs (2 μl) were mixed with 10 μl of SYBR Premix Ex Taq (Takara Bio) and primer pairs (4 pmol each) in a 20-μl reaction volume, followed by PCR for 40 cycles (95°C denaturation for 15 s, 60°C annealing for 15 s, and 72°C extension for 33 s) in an ABI PRISM 7500 Fast real-time PCR system (Applied Biosystems). The PCR products were subjected to a melting curve analysis to verify a single amplification product. PCR without RT was performed as a negative control. The housekeeping gene encoding GAPDH was used as a reference to normalize expression levels and to quantify changes in gene expressions between nontreated control and MspTL-treated cells or between MspTL- and inhibitor-treated cells. The sequences of the primers for real-time RT-PCR were as follows: 5′-AGG AAT AAC ATC TGG TCC AAC A-3′ and 5′-TGA GCT TGA CAA AGT GGT CG-3′ for the IFN-α gene; 5′-GAC TAT TGT TGA GAA CCT CCT-3′ and 5′-TCG GAG GTA ACC TGT AAG TC-3′ for the IFN-β gene; 5′-CTG ACT TGA ATG TCC AAC GCA-3′ and 5′-TTC AAA TAT TGC AGG CAG GAC A-3′ for the IFN-γ gene; 5′-TTC TTA GTG GAT GTT CTG ACC-3′ and 5′-GTG TTT GGA ATT GTA TGT AGG T-3′ for the IP-10 gene; 5′-GGA AAT CTT CGC ACC TCA AGG-3′ and 5′-CGC GTG CTT GGT CAG GTG GT-3′ for the RANTES gene; 5′-GTG AAG GTG CAG TTT TGC CA-3′ and 5′-TCT CCA CAA CCC TCT GCA C-3′ for the IL-8 gene; 5′-GTG GTG GAC CTG ACC TGC-3′ and 5′-TGA GCT TGA CAA AGT GGT CG-3′ for the GAPDH gene.
RNA interference assay
siRNA transfections targeting endogenous IKKi, TBK1, and MyD88 were conducted in THP-1 cells (1 × 106 cells/well in 3 ml) seeded in 6 well-culture plates and grown in RPMI 1640 without serum and antibiotics. siRNA (0.6 μM) for IKKi, TBK1, or MyD88 was mixed with 10 μl of the cationic liposome HilyMax (Dojindo Laboratories) in 270 μl of medium for 15 min, and then added to the cells and incubated in a CO2 incubator at 37°C for 2 h. The culture medium was replaced with fresh RPMI 1640 medium supplemented 10% FBS and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin sulfate) and incubated for an additional 24 h at 37°C. To address possible nonspecific reactions of siRNA, the cells were transfected with control A siRNA, a nontargeting 20- to 25-nucleotide siRNA.
To verify a reduction of endogenous mRNA expression of IKKi, MyD88, and TBK1 in transfected cells, total RNA was isolated from transfected THP-1 cells using TRIzol (Invitrogen). RNA (1 μg) was mixed with a Maxime RT PreMix kit (iNtRON) and incubated at 45°C for 1 h. cDNAs were subjected to PCR for IKKi, MyD88, and TBK1 in 50-μl mixtures containing the specific primers: 5′-GCA GGA GCT AAT GTT TCG GG-3′ and 5′-CTG GAG AGG ACC TCC GCT A-3′ for the IKKi gene; 5′-TAA GAA GGA CCA GCA GAG C-3′ and 5′-CAT GTA GTC CAG CAA CAG C-3′ for the MyD88 gene; 5′-TTG AAG AGG AGA CAA CAA CAA GA-3′ and 5′-CAT TCC ACC CAC CAC ATC T-3′ for the TBK1 gene; 5′-CAC TGA CAC GTT GGG AGT GG-3′ and 5′-CAT GGA GAA GGC TGG GGC TC-3′ for the GAPDH gene.
To see the signaling molecules involved in IFN-β expression induced by MspTL, THP-1 cells were treated with MspTL (10 μg/ml) or IFN-γ (500 pg/ml) for various time periods, harvested, and washed with PBS. For whole-cell preparation, the cell pellets were incubated on ice for 10 min with RIPA buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% Nonidet P-40, 0.25% SDS) containing protease inhibitor cocktail (Roche Applied Science) and 1 mM PMSF. The lysates were separated by 10% SDS-PAGE and transferred to a polyvinylidene difluoride membrane. The membrane was blocked with 5% BSA for 1 h and incubated with anti-phospho-STAT-3 (Tyr705) Ab overnight at 4°C, followed by reaction with HRP-conjugated anti-rabbit IgG for 1 h. Bound Abs were detected with an ECL system (Amersham Biosciences). The membrane was stripped and reprobed with Abs in order of anti-phospho-STAT-1 (Tyr701) Ab, anti-STAT-3 Ab, anti-STAT-1 Ab, and β-actin. To examine IRF-3 and IRF-7 activation by MspTL, lysates of THP-1 cells treated with MspTL or LPS were subjected to native PAGE (7% gel) to detect monomeric and dimerized forms of IRF-3 and IRF-7 as follows: The gel was pre-run with running buffer (anode buffer: 25 mM Tris, 192 mM glycine (pH 8.3); cathode buffer: 25 mM Tris, 192 mM glycine, 1% sodium deoxycholate (pH 8.3)) at 40 mA for 30 min. The samples in sample buffer (60 mM Tris-HCl (pH 6.8), 25% glycerol, 0.1% bromphenol blue, 1% sodium deoxycholate) were applied to the gel and separated at 25 mA. The proteins were then transferred to polyvinylidene difluoride membranes using Tris-glycine transfer buffer at 110 V for 60 min at 4°C. Membranes were blocked for 1 h in 5% BSA solution and reacted with a mouse anti-IRF-3 mAb or a polyclonal rabbit anti-IRF-7 Ab. Bound Abs were detected as described above using an ECL system.
To see the translocation of activated IRF-3 and IRF-7 into the nucleus, THP-1 cells treated with MspTL (10 μg/ml) or LPS (10 μg/ml) were harvested, washed with PBS, and lysed with nuclei isolation buffer (10 mM HEPES (pH 7.9), 10 mM KCl, 10 mM MgCl2, 1 mM DTT, and 0.5 mM PMSF) on ice for 15 min. After adding Nonidet P-40 to a final concentration of 0.5% and vortexing for 10 s, the lysates were centrifuged at 6500 × g for 20 s at 4°C. Nuclear pellets were treated with nuclei lysis buffer (20 mM HEPES (pH 7.9), 1 mM EDTA, 1 mM DTT, 1 mM PMSF, 400 mM NaCl, and 25% glycerol) on ice for 30 min. After centrifuging at 10,000 × g at 4°C for 15 min, the lysates were separated by SDS-PAGE (10% gel) and transferred to polyvinylidene difluoride membranes. The membranes were incubated in blocking buffer (5% BSA in 0.1% PBS-Tween 20) for 1 h and reacted with an anti-human IRF-3 mAb or an anti-human IRF-7 mAb for 1 h, followed by reaction with HRP-conjugated secondary Abs for 1 h. Bound Abs were detected using an ECL system (Amersham Biosciences). Sample protein concentrations were measured using the BCA protein assay kit (Pierce). Goat polyclonal Abs to lamin B and β-actin were used as loading controls for nuclear extracts and whole-cell lysates, respectively.
Formation of lipid raft aggregates in THP-1 cells by MspTL
THP-1 cells (1 × 105 cells) were treated with MspTL (10 μg/ml) labeled with Alexa 568 or LPS (10 μg/ml) for 1 h. The cells were fixed with 4% paraformaldehyde in PBS and were then treated with FITC-labeled cholera toxin B chain, which binds ganglioside GM1, as a raft marker. The cells were observed using a confocal laser microscope (LSM5; Zeiss).
Effects of lipid raft inhibitors on cell proliferation
To examine the effects of lipid raft inhibitors on cell proliferation, THP-1 cells treated with the inhibitors in experimental conditions were subjected to the MTT test (R&D systems) according to the manufacturer’s protocol, whereby viable cells form intracellular formazan crystals as a consequence of their metabolic activity.
MspTL binding and internalization assay
MspTL binding to and internalization into host cells was analyzed by flow cytometry. THP-1 cells (1 × 106 cells) were treated with Alexa 488-labeled MspTL for 1–2 h. The binding and internalization of the labeled MspTL was detected with a FACSCalibur flow cytometer after washing the cells with cold PBS. Total fluorescence was regarded as the sum of binding and internalization of MspTL. Since trypan blue is known to quench extracellular fluorescence emission of the cells (22), the measured fluorescence of the cells after incubation with 0.4% trypan blue at room temperature for 5 min was regarded as that of internalized MspTL. Alexa 488-labeled BSA was used as a control for nonspecific binding. In some experiments, the cells were pretreated with lipid raft inhibitors for 1 h as described above and then treated with Alexa 488-labeled MspTL for 1 h. To see whether MspTL mediates the internalization of T. lecithinolyticum whole cells, the bacteria were labeled with 10 μM CFSE as described previously (9). CFSE-labeled bacteria were diluted with PBS to the cell number of 1 × 108 (corresponding to 2 μg MspTL), incubated with anti-MspTL Ab for 30 min, and then added to THP-1 cells (1 × 106 cells) for 4 h. After washing the cells with PBS, the binding and internalization of T. lecithinolyticum was detected with a FACSCalibur flow cytometer as described above. An isotype-matched Ab was used as a negative control.
Statistically significant differences between nontreated and stimulator-treated cells were analyzed by ANOVA, and differences between stimulator- and inhibitor-treated cells were analyzed by Student’s t test. A p value of <0.05 was considered statistically significant.
MspTL induces IFN-β in THP-1 cells and human HGFs
In our previous study, we showed that MspTL stimulated PDL cells and HGFs to induce IL-1β, TNF-α, IL-6, IL-8, PGE2, and ICAM-1 expression, all of which are regulated by NF-κB (8, 9). In the preliminary studies, microarray analyses for MspTL global transcriptional regulation indicated that MspTL induced numerous ISGs in THP-1 cells. These results prompted us to study IFN regulation by MspTL. MspTL induced IFN-β mRNA in both THP-1 cells and primary cultured HGFs, but did not induce IFN-α or IFN-γ mRNA after up to 12 h of exposure (data not shown). In THP-1 cells, IFN-β mRNA was induced as early as 2 h after MspTL treatment (Fig. 1,A), whereas this induction was slightly delayed in HGFs and had a reduced expression fold change (Fig. 1,E). A dose-dependent increase in IFN-β mRNA by MspTL was observed, and anti-MspTL Ab inhibited IFN-β mRNA induced by MspTL (Fig. 1,C). Up-regulation of IFN-β levels was also detected in the culture supernatants of THP-1 cells treated with MspTL (Fig. 1,B). With T. lecithinolyticum whole cells, significant release of IFN-β was observed at a MOI of 100 (Fig. 1 D). The MOI of 100 corresponded to 2 μg of MspTL or 1 × 108 bacteria. The detection limit of IFN-β levels in the assay was <10 pg/ml. Irrelevant recombinant proteins expressed in E. coli were used as a negative control and did not induce IFN-β (data not shown).
To confirm that the cell-stimulating activity of MspTL was not caused by endotoxin contamination, we examined the ability of MspTL to induce ICAM-1 or IL-8 in various contexts. First, heat-inactivated MspTL did not stimulate THP-1 cells to express ICAM-1, and the mock extract did not induce ICAM-1 either (data not shown). Second, polymyxin B, an endotoxin inhibitor, did not affect IFN-β induction by MspTL, whereas it significantly inhibited ICAM-1 induction by E. coli LPS (data not shown). Third, MspTL did not induce the NF-κB reporter in Chinese hamster ovary (CHO) cells (CHO/CD14/TLR4 and CHO/CD14/TLR2) to express membrane CD25 through TLR4- and TLR2-dependent NF-κB activation. However, LPS, a ligand of TLR4, and the bacterial lipopeptide Pam3CSK4, a ligand for TLR2, induced NF-κB-regulated CD25 expression when analyzed by flow cytometry (Fig. 2). Fourth, neutralizing Abs for TLR2 and TLR4 did not affect IL-8 production induced by MspTL in THP-1 cells (data not shown). These results indicate that MspTL was not contaminated with endotoxin or TLR2 ligands. Concomitantly, these results indicate that MspTL does not utilize TLR2 and TLR4 as cellular receptors to transduce signals in host cells.
MspTL induces IP-10 and RANTES via IFN-β
To investigate whether MspTL-induced IFN-β acts in an autocrine/paracrine fashion to stimulate expression of ISGs, we analyzed the regulation of the representative ISGs IP-10 and RANTES by MspTL. IP-10 and RANTES are known to be primarily regulated by type I and type II IFNs. As shown in Fig. 3, MspTL significantly up-regulated mRNA expression of IP-10 and RANTES in both THP-1 cells (Fig. 3, A and B) and HGFs (Fig. 3, D and E). Levels of IP-10 and RANTES mRNA in THP-1 cells peaked at 8 h. The up-regulation of IP-10 in culture supernatants from THP-1 cells treated by MspTL was confirmed at the protein level by ELISA (Fig. 3,C). To verify whether IP-10 and RANTES were up-regulated via IFN-β induced by MspTL or directly by MspTL, we analyzed their expression following treatment with a neutralizing anti-IFN-β Ab. THP-1 cells and HGFs were treated with MspTL in the presence of anti-IFN-β Ab before stimulation. As shown in Fig. 4, anti-IFN-β Ab significantly reduced the expression of IP-10 and RANTES mRNA induced by MspTL (Fig. 4, B, C, E, and F) in both cell types, whereas IFN-β mRNA expression was not affected (Fig. 4, A and D). An isotype-matched Ab did not affect cytokine expression (data not shown). IP-10 levels in THP-1 cell culture supernatants were also significantly reduced by treatment with the anti-IFN-β Ab (Fig. 4,G). Additionally, the protein synthesis inhibitor cycloheximide did not affect IFN-β induction by MspTL, whereas it suppressed IP-10 induced by MspTL (Fig. 4, H and I). Taken together, these results indicate that MspTL directly induces IFN-β, which in turn stimulates IP-10 and RANTES. LPS-induced IP-10 and RANTES were also significantly decreased by treatment with the anti-IFN-β Ab in both cell types.
MspTL induces IRF-3 dimerization and translocation into the nucleus
Since the expression of type I IFNs, including IFN-β, is regulated by the activation of IRF-3 and/or IRF-7 (23), we tested whether MspTL was capable of activating these transcription factors. In most cell types, IRF-3 and IRF-7 normally reside in the cytoplasm in an inactivated state and become activated upon phosphorylation and subsequent dimerization. The activated transcription factors are then translocated into the nucleus. As shown by Western blot (Fig. 5,A), IRF-3 dimer formation was observed in whole-cell lysates of THP-1 cells separated on native gel as early as 10 min after MspTL treatment. Increased amounts of IRF-3 in nuclear extracts in both THP-1 cells (Fig. 5,C) and HGFs (Fig. 5,D) were also detected as early as 30 min after MspTL treatment, thus validating the MspTL-stimulated phosphorylation and dimerization of IRF-3. In contrast, phosphorylation and dimerization of IRF-7 was not observed following MspTL treatment in both cell types (Fig. 5 B–D). The protein membranes were stripped and reprobed for the nuclear protein lamin B and the cytoplasmic protein β-actin as loading controls. Equivalent amounts of lamin B were detected in nuclear extracts, but β-actin was not detected, suggesting that there was no contamination with cytoplasmic proteins in the nuclear extracts.
NF-κB activation is not involved in IFN-β induction by MspTL
Since type I IFN promoters have NF-κB binding sites as well as IRF-3 binding sites, we used NF-κB inhibitors to test whether the induction of IFN-β and ISGs by MspTL was also associated with the NF-κB signaling pathway. As shown in Fig. 6, IFN-β and IP-10 expression induced by MspTL in THP-1 cells was not affected by pretreatment with the NF-κB inhibitors helenalin and PDTC at either the mRNA (Fig. 6, A and B) or protein level (Fig. 6, D and E). In contrast, IL-8 expression induced by MspTL was significantly reduced by the inhibitors (Fig. 6, C and F). These results indicate that MspTL stimulates cytokine production through at least two distinct pathways: NF-κB-dependent pathways and NF-κB-independent pathways. We also found that IFN-β and IP-10 induced by LPS were regulated in a similar fashion as was observed for MspTL (Fig. 6 A–C). In our previous study, MspTL-induced ICAM-1, IL-6, and IL-8 were shown to be inhibited by TPCK, another NF-κB inhibitor (8).
TBK1 is involved in IFN-β induction by MspTL
We showed that cell treatment with MspTL activated IRF-3 and induced IFN-β. Activation of IRF-3 and subsequent expression of IFN-β are known to be associated with the IKK-related kinases, TBK1 and IKKi (24, 25). We next used siRNA technology to examine the possible involvement of these factors in IFN-β induction by MspTL. Transfection of THP-1 cells with siRNA specific for IKKi or TBK1 resulted in remarkable reduction in the appropriate mRNAs as assessed by RT-PCR, while scrambled siRNA (nontargeting siRNA) used as a control did not affect specific mRNA expression (Fig. 7,A). Knockdown of TBK1 resulted in a significant reduction in IFN-β and IP-10 expression after a 4-h treatment with MspTL, but IKKi knockdown did not affect IFN-β and IP-10 expression (Fig. 7 B). TBK1-specific siRNA also reduced IRF-3 translocation into the nucleus by MspTL, validating the role of TBK1 in MspTL signaling to IRF-3 (data not shown). THP-1 cells transfected with TBK1- or IKKi-specific siRNA and then stimulated with LPS showed significant reductions in IFN-β and IP-10 expression. These results indicate that TBK1 is involved in cell responses to both MspTL and LPS, whereas IKKi is only involved in LPS signaling.
MyD88 is an intracellular adaptor molecule for several TLRs to induce type I IFN signaling as well as NF-κB signaling by ligand binding. Although we observed that TLR2 and TLR4 were not important for MspTL signaling as described above, we used MyD88-specific siRNA to reduce expression of MyD88 to examine the possible involvement of other intracellular TLRs (TLR7, TLR8, and TLR9) in cellular responses to MspTL. MyD88 is essential to signal transduction through these receptors. As shown in Fig. 7 B, knockdown of MyD88 did not affect induction of IFN-β, IP-10, or IL-8 in response to MspTL treatment. LPS-stimulated IFN-β expression is known to be dependent on TLR4, but independent of MyD88. In our experiments, knockdown of MyD88 resulted in a significant reduction of IL-8 induced by LPS, but did not affect IFN-β and IP-10 expression. These results validate the finding of MyD88-independent signaling by MspTL.
Recognition of molecular targets by intracellular TLRs occurs in the endosome, and endosomal acidification/maturation are required for proper recognition. To clarify the involvement of endosomes in the interaction between MspTL and host cells, THP-1 cells were treated with bafilomycin and chloroquine, two inhibitors of endosomal acidification. As shown in Fig. 7 C, cell treatment with bafilomycin or chloroquine did not affect MspTL induction of IFN-β, IP-10, and IL-8 expression. However, induction of IFN-β by CpG DNA, a TLR9 ligand, was significantly inhibited by these endosomal inhibitors. These results imply that intracellular pathogen recognition receptors requiring endosomal activation are not involved in MspTL-medicated signaling involving either the NF-κB or the IFN signaling pathway.
MspTL activates STAT-1 via IFN-β
IFN-β acts in an autocrine or paracrine fashion to stimulate ISGs via binding to the type I IFN receptor and subsequent STAT activation. Since STAT-1 is a component of the classical transcription factor complex for the type I IFN signal transduction pathway, we determined whether STAT-1 activation was involved in the up-regulation of IP-10 and RANTES in response to MspTL-induced IFN-β. Whole-cell extracts treated with MspTL or LPS were subjected to Western blotting using an anti-p-STAT-1 Ab (Tyr701). STAT-1 has two isoforms, STAT-1α (91 kDa) and the splice variant STAT-1β (84 kDa). As shown in Fig. 8, MspTL treatment of THP-1 cells resulted in phosphorylation of STAT-1α/β as early as at 2 h after initial treatment (Fig. 8), while total STAT-1 protein remained unchanged. STAT-1 phosphorylation following MspTL treatment was significantly inhibited by an anti-IFN-β Ab, indicating that STAT-1 phosphorylation resulted from the activity of IFN-β induced by MspTL. Engagement of type I IFN receptor also results in STAT-3 phosphorylation to form STAT-1/3 heterodimers or STAT-3/3 homodimers. However, STAT-3 phosphorylation by MspTL was not observed. Cell treatment with IFN-γ used as a positive control resulted in phosphorylation of both STAT-1 and STAT-3.
Lipid rafts are involved in MspTL signal transduction resulting in induction of IFN-β, IP-10, and IL-8
Although a cellular receptor for MspTL has not been identified, we tested the involvement of lipid rafts in MspTL-induced signaling. Lipid rafts are specific glycolipid- and cholesterol-enriched domains in the lipid bilayer of the cell plasma membrane. Upon stimulation, some cell-surface receptors and intracellular signaling molecules are mobilized into lipid rafts, where multiprotein complexes assemble and interact to transduce intracellular signaling (26). To investigate whether disruption of lipid rafts affected cellular response to MspTL treatment, THP-1 cells were pretreated with the lipid raft inhibitors MβCD, nystatin, and filipin for 1 h before MspTL stimulation. As shown in Fig. 9, preincubation of THP-1 cells with these lipid raft-disrupting agents significantly inhibited MspTL-stimulated expression of IFN-β, IP-10, and IL-8 at both the mRNA (Fig. 9,A–C) and protein (Fig. 9 D–F) levels. LPS signaling is known to involve lipid raft formation (27), and LPS-induced IFN-β, IP-10, and IL-8 were also inhibited by all three inhibitors. Our results indicate a physiological role for membrane lipid rafts as critical regulators of MspTL-mediated signal transduction. Lipid raft inhibitors did not affect cell viability in the range of the concentrations used in this study as determined using an MTT test (data not shown).
Lipid raft activation of THP-1 cells by MspTL was also observed by confocal microscopy. After incubation of THP-1 cells with Alexa 568-labeled MspTL for 1 h, labeled MspTL colocalized with activated raft aggregates detected by FITC-cholera toxin B chain (Fig. 9 G). LPS was used as a positive control and was also detected in raft aggregates.
Lipid rafts are involved in MspTL internalization into THP-1 cells
To determine whether lipid rafts were involved in MspTL internalization, we tested whether lipid raft inhibitors could block MspTL internalization. Alexa 488-labeled MspTL was incubated with THP-1 cells pretreated with lipid raft inhibitors, and MspTL internalization was analyzed by flow cytometry. As shown in Fig. 10,A, incubation of THP-1 cells for 1 h with Alexa 488-labeled MspTL showed a significant increase in the number of fluorescent cells. Trypan blue to quench extracellular fluorescence emission reduced the number of the fluorescent cells by ∼15%, indicating that the majority of MspTL fluorescence was due to internalized MspTL. The lipid raft inhibitors MβCD and filipin significantly reduced the number of fluorescent cells, and trypan blue quenching did not significantly decrease the number of these fluorescent cells. Interestingly, nystatin did not affect the total number of fluorescent cells, but it did significantly reduce the number of fluorescent cells after trypan blue quenching. Internalization of MspTL into paraformaldehyde-fixed cells was minimal (data not shown). Alexa 488-labeled BSA used as a negative control was not able to bind or internalize into THP-1 cells (data not shown). Taken together, these results indicate that MspTL binding to and internalization into the cells are dependent on lipid raft activation. To see the role of MspTL in the internalization of T. lecithinolyticum, THP-1 cells were incubated with CFSE-labeled whole bacteria preincubated with anti-MspTL Ab for 30 min. As shown in Fig. 10 B, the binding and internalization of T. lecithinolyticum was significantly inhibited by anti-MspTL Ab, suggesting that MspTL plays a major role in T. lecithinolyticum internalization.
MspTL is the most abundant OM protein of T. lecithinolyticum associated with destructive periodontitis and endodontic infections, and it is a potent stimulator of various cytokine expression (8, 9). Since the cell wall of oral treponemes, including T. lecithinolyticum, is similar to that of Gram-negative bacteria, LPS or LPS-like molecules are expected to be important surface molecules for stimulating immune responses. However, phenol water-extracted LPS-like molecules from some oral treponemes act as antagonists to E. coli LPS (20, 28). In our previous experiments, phenol water extracts from T. lecithinolyticum were biologically inert (20). Therefore, in the present study, we selected MspTL as a representative molecule for investigations of the molecular basis of host immune response to the bacterium. Our preliminary microarray analyses of global gene regulation in THP-1 cells showed that MspTL stimulated the expression of various ISGs in addition to previously reported proinflammatory factors (8, 9). These results prompted us to investigate the regulation of IFN-β and ISGs by MspTL in a human monocytic cell line, THP-1 cells, and in primary cultured HGFs. MspTL induced IFN-β expression via TBK1 and IRF-3, which in turn up-regulated IP-10 and RANTES in a putative STAT-1-dependent manner. Monocytes and macrophages are major sources of inflammatory mediators, and gingival fibroblasts are the most abundant cells in periodontal connective tissue. Therefore, understanding the response of these cells to MspTL is crucial for defining host-bacteria interactions in periodontitis. MspTL stimulates cytokine production through at least two independent pathways. IL-1β and IL-8 are induced through the NF-κB pathway, as shown in our previous report (8) and in the present study. We have now demonstrated that IP-10 and RANTES are induced through the IFN-β/TBK1/IRF-3 pathway. Host cell stimulation by MspTL appears to be independent of TLRs. CHO cells transfected with TLR2 or TLR4 did not respond to MspTL. Since inhibitors of endosomal maturation did not affect MspTL induction of either IFN-β or IL-8, we can exclude the involvement of endosomal TLRs (TLR3, TLR7, TLR8, and TLR9) in MspTL-mediated host response. Expression of these TLRs by THP-1 cells was confirmed in our preliminary study.
Numerous studies have shown that bacteria and their components can induce IFN-β. Most prominently, intracellular bacteria such as C. trachomatis, L. monocytogenes, and M. tuberculosis induce IFN-β and ISGs, including IP-10 and RANTES. However, they are employing common and distinct signaling components for IFN-β induction. In macrophages, fibroblasts, and epithelial cells, C. trachomatis induces IFN-β and IP-10 through a TLR2- and TLR4-indepentent pathway involving TBK1 and IRF-3 activation as found in MspTL stimulation (17, 29). C. trachomatis infection in IFN type I receptor knockout mice did not induce IP-10, and treatment of macrophages taken from Chlamydia-infected mice with a neutralizing anti-IFN-β Ab significantly reduced IP-10 expression, suggesting a direct role for IFN-β in IP-10 induction. C. trachomatis uses MyD88 and requires endosomal maturation (17). In Chlamydia species, infection is a prerequisite for IFN-β and IP-10 induction, since UV-inactivated or heat-killed bacteria did not induce IFN-β and IP-10. L. monocytogenes also induces IFN-β in macrophages via IRF-3 and TBK1 in a manner that is independent of TLR2 and TLR4. In contrast to C. trachomatis, L. monocytogenes induces IFN-β independent of MyD88. Listeriolysin O is essential for L. monocytogenes to activate the type I IFN pathway in macrophages. Listeriolysin O is a pore-forming toxin that lyses endosomal membranes enabling L. monocytogenes to escape from the phagosome and enter the cytoplasm. However, treatment with listeriolysin O alone was unable to induce IFN-β in macrophages, indicating that the entry process of live bacterium was required (19). M. tuberculosis induced IFN-β in macrophages independent of TLR2, TLR4, and MyD88 as in L. monocytogenes (30). IFN-β induced by M. tuberculosis is required for expression of IP-10 and RANTES via a type I IFN receptor and STAT-1. Although it has been demonstrated that whole bacteria induce IFN-β, the important stimulatory bacterial components of these intracellular bacteria have not been defined.
T. lecithinolyticum is not an intracellular bacterium. However, the bacterium itself and MspTL alone were able to be internalized into cells. Treatment with lipid raft inhibitors that disrupted binding and internalization of MspTL resulted in a significant reduction of IFN-β, IP-10, and IL-8 expression induced by the protein. MβCD and filipin inhibited both binding and internalization. Since nystatin inhibited MspTL internalization without affecting the binding, it is likely that the internalization is required for both the IRF-3 and the NF-κB pathways of MspTL signaling. MβCD depletes cholesterol from the cell membrane, while nystatin and filipin, polyene compounds, bind to and sequester cholesterol, disrupting the formation and trafficking of caveolae. The distinct behavior of nystatin in interaction between MspTL and THP-1 cells could be explained by a different mode of action of the compound. At a concentration with antibiotic activity, nystatin binds to sterols and forms pores in the cytoplamic membrane, whereas filipin forms large complexes with sterols between the leaflets of the lipid bilayer of the membrane (31). For lipid raft inhibition, we used a concentration for each inhibitor that did not show discernable cytotoxicity in the cells. Lipid raft-dependent internalization of Porphyromonas gingivalis, a representative periodontopathogen, has been also demonstrated in epithelial cells and macrophages (32, 33). Lipid raft inhibitors, in particular MβCD, drastically reduced the internalization of beads coated with the OM vesicles of P. gingivalis in HeLa cells and of P. gingivalis whole bacteria in mouse macrophages. However, these inhibitors only minimally affected the binding of the membrane-coated beads and bacteria. Cholesterol depletion by MβCD suppressed the ability of P. gingivalis for intracellular persistence and cytokine induction in mouse macrophages (33) and abolished actin polymerization induced by P. gingivalis OM vesicles (32). Lipid raft-dependent internalization was also observed with Treponema denticola Msp-deduced peptide conjugate, P34BSA, which induced actin stress fiber formation through PI3K and RhoA activation, although lipid raft activation by P34BSA, but not lipid raft-dependent internalization, is linked to PI3K and RhoA activation (34).
Type I IFNs are known to possess the potent antiviral activity in mediating resistance to viruses and immunostimulation. However, IFN functions during bacterial infection have been poorly understood. Existing data on the role of type I IFNs demonstrate both beneficial and detrimental effects on host defense against bacteria. Depending on the specific bacterial species and host animal models, contradictory results have been demonstrated. Type I IFNs induced by L. monocytogenes decreased host immunity, possibly through the induction of host immune effector cell apoptosis (19). A hypervirulent strain (HN878) of M. tuberculosis failed to induce Th1 type immunity, possibly as a result of increased type I IFN induced by the bacteria (35). Mice infected with the strain HN878 presented with increased lung bacillary loads and death when treated intranasally with purified IFN-α/β. In chlamydial infections, controversial results have also been reported. IFN-α/β receptor knockout mice were more resistant to Chlamydia muridarum (C. trachomatis mouse pneumonitis) infection and showed less bacterial burden with decreased macrophage apoptosis (36). In contrast, IFN-β expression by C. trachomatis infection was associated with a protective host defense, inducing IP-10 expression, which was correlated with the ability to clear bacterial infection in mice (17). IFN-β has been demonstrated to be a crucial factor for host defense against common extracellular bacterial pathogens such as group B streptococci, Streptococcus pneumoniae, and encapsulated E. coli (37). Infection with these bacteria in the IFN-α/β receptor knockout mice or IFN-β knockout mice resulted in significantly increased lethality that was associated with severe bacteremia. The increased susceptibility of these mice was correlated with decreased expression of mediators of antibacterial host defense, including TNF-α, IFN-γ, and NO. Immune response to pathogens is required for an efficient host defense, but an overactive immune response can also result in tissue damage. To gain insight into the role of IFN-β induced by various pathogens, intensive studies will be needed to identify bacterial ligands and host receptors that induce IFN-β and to establish specific cytokine profiles induced by IFN-β in affected tissues.
MspTL-induced IFN-β significantly up-regulated expression of IP-10 and RANTES. IP-10 is a CXC chemokine that chemoattracts T lymphocytes and monocytes. Using a human cytokine Ab array detecting 36 cytokines, IP-10 was detected at significantly higher levels in the gingival crevicular fluid of periodontally diseased sites than in healthy sites (38). Also, cells producing increased amounts of IP-10 were observed in inflamed gingival tissue from periodontitis patients (39). Although the role of IP-10 in periodontal disease has not been elucidated, accumulating data suggest a role for IP-10 in chronic inflammatory diseases, including arthritis and atherosclerosis. IP-10 has been reported to recruit inflammatory cells into inflamed joints and cause bone destruction, leading to arthritis in a mouse model (40). IP-10 deficiency significantly reduced early lesion formation in the aorta of ApoE−/− mice through modulation of both effector and regulatory T cells (41). IP-10 was also reported to be up-regulated at sites of murine colitis and clinical inflammatory bowl disease, and an anti-IP-10 Ab attenuated chronic colitis by preventing the activation and recruitment of CXCR3+ leukocytes (42). RANTES is a CC chemokine and is involved in the activation and recruitment of neutrophils, eosinophils, monocytes, and Th1 cells to sites of infection. It is also a chemotactic factor for osteoclasts associated with bone resorption (43). RANTES was detected at significantly higher levels in gingival crevicular fluid from chronic periodontitis and generalized aggressive periodontitis, and it was positively correlated with both probing depth and clinical attachment loss (44, 45). P. gingivalis induced the secretion of RANTES from human whole blood (46). Periodontitis is among the most prevalent chronic diseases in humans, and up-regulation of IP-10 and RANTES by bacterial components may contribute to the progression of this disease. Therefore, identification of the bacterial components that induce these factors is important for elucidating molecular patterns and developing strategies for subversion of proinflammatory activities. Moreover, the direct or indirect roles of IFN-β in the induction of ISGs in periodontitis should be investigated. Host immune responses to periodontopathogens are crucial to the progression and severity of periodontitis, both of which are determined by the balance between pro- and antiinflammatory mediators.
To our knowledge, MspTL is the first major outer protein of a periodontopathogen proven to induce IFN-β. LPS and CpG DNA are the prominent established bacterial components that induce type I IFN. The signaling pathways utilized by these ligands to induce IFN-β include TLR4/TRIF for LPS and TLR9/MyD88 for CpG DNA (47). Flagellin also induces IFN-β via TLR5 in bone marrow-derived macrophages through a mechanism mediated by STAT-1 activation, but not IRF-3 (12). In P. gingivalis, live bacteria and bacterial components like FimA, a major fimbrial protein, and LPS were analyzed using cDNA microarrays to determine signaling pathways of host immune response in macrophages (48). Response was assessed with specific regard to acute infection by live bacteria and chronic infection by bacteria and bacterial breakdown. Results revealed both common and unique gene regulation and that only P. gingivalis LPS was involved in initial IFN-β induction through the TLR7-MyD88 pathway. The physiological role of IFN-β needs to be elucidated in P. gingivalis infection. The OM of oral treponemes is fragile and is easily released from the protoplasmic cylinder of the bacterium (49). Since the MspTL protein is homologous to MspA of Treponema maltophilum, which is also frequently found in periodontitis (50) and which elicits a host response similar to MspTL (8), these proteins may play a particularly important role in chronic periodontitis.
In summary, MspTL, the major surface protein of T. lecithinolyticum, activated TBK1 and IRF-3 in THP-1 cells, a monocytic cell line, and in primary cultured HGFs. Subsequent expression of IFN-β and activation of STAT-1 led to the induction of IP-10 and RANTES expression. Lipid raft activation seems to be a prerequisite for MspTL internalization and activation of the cellular signaling resulting in this induction. Efforts to identify the host proteins that participate in the recognition of MspTL and induction of IFN-β expression are ongoing and will provide insight into the cell signaling and functional roles of bacterial surface proteins in host immunomodulation.
The authors have no financial conflicts 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 the Korea Research Foundation Grant funded by the Korean Government (Ministry of Education and Human Resource Development, Basic Research Promotion Fund) (KRF-2006-531-E00072). S.-H. L., H.-K. J., and H.-R. L. are recipients of a scholarship from the BK21 Program (Craniomaxillofacial Life Science 21).
Abbreviations used in this paper: HGF, human gingival fibroblast; CHO, Chinese hamster ovary; CpG-ODN, CpG-oligodeoxynucleotide, unmethylated dsDNA; IKKi, inducible IκB kinase; IP-10, IFN-γ-inducible protein-10; IRF, IFN regulatory factor; ISG, IFN-stimulated gene; MβCD, methyl-β-cyclodextrin; MOI, multiplicity of infection; OM, outer membrane; Pam3CSK4, palmitoyl-3-cysteine-serine-lysine-4; PAMP, pathogen-associated molecular pattern; PDL, periodontal ligament cell; PDTC, pyrrolidine dithiocarbamate; siRNA, small interfering RNA; TBK1, TANK-binding kinase 1.