The HTLV-I p30 Interferes with TLR4 Signaling and Modulates the Release of Pro- and Anti-inflammatory Cytokines from Human Macrophages*

Whereas adaptive immunity has been extensively studied, very little is known about the innate immunity of the host to HTLV-I infection. HTLV-I-infected ATL patients have pronounced immunodeficiency associated with frequent opportunistic infections, and in these patients, concurrent infections with bacteria and/or parasites are known to increase risks of progression to ATL. The Toll-like receptor-4 (TLR4) activation in response to bacterial infection is essential for dendritic cell maturation and links the innate and adaptive immune responses. Recent reports indicate that TLR4 is targeted by viruses such as RSV, HCV, and MMTV. Here we report that HTLV-I has also evolved a protein that interferes with TLR4 signaling; p30 interacts with and inhibits the DNA binding and transcription activity of PU.1 resulting in the down-regulation of the TLR4 expression from the cell surface. Expression of p30 hampers the release of pro-inflammatory cytokines MCP-1, TNF-α, and IL-8 and stimulates release of anti-inflammatory IL-10 following stimulation of TLR4 in human macrophage. Finally, we found that p30 increases phosphorylation and inactivation of GSK3-β a key step for IL-10 production. Our study suggests a novel function of p30, which may instigate immune tolerance by reducing activation of adaptive immunity in ATL patients.

expansion of infected cells, high proviral loads, and virus-specific immune responses, including increased pro-inflammatory cytokine production and HTLV-I-specific CTL (4,5). Studies have demonstrated that virus-expressing cells are rapidly and continuously eliminated by the immune system in HAM/TSP patients (6,7). It has also been shown that HTLV-I infected dendritic cells from HAM/TSP patients exhibit an enhanced capacity to stimulate antigen-specific CD4ϩ and CD8ϩ T-cell activation, which leads to autoimmunity (8). In sharp contrast ATL pathogenesis is characterized by monoclonal expansion of infected cells, very low proviral loads, absence of pro-inflammatory cytokines (9), and very low frequency of virus-specific CTL (10,11). HTLV-I-infected monocytes produce dysfunctional dendritic cells because of improper differentiation, which do not stimulate autologous T-cells (12). HTLV-I-infected ATL patients have pronounced immunodeficiency associated with frequent opportunistic infections by various pathogens, including Pneumocystis carinii, Toxoplasma gondii, Cryptococcus neoformans, Candida albicans, Mycobacterium avium, and Aspergillus, Cytomegalovirus, and Strongyloides (13,14). HTLV-I regulatory protein p30 has been shown to suppress virus expression at both transcriptional and post-transcriptional levels. Hence, it is anticipated, though not yet demonstrated, that p30 expression and or its functions are increased in ATL patients. The host immune system detects and responds to microbial infection mainly through a family of pattern recognition receptors called Toll-like receptors (TLRs). Signaling by these receptors induces antimicrobial genes, inflammatory cytokines, and dendritic cell maturation, which are necessary for initiating an adaptive immunity (15). Among the 10 TLRs identified in humans, TLR4 is the major lipopolysaccharide (LPS) receptor and elicits innate immune response against Gram-negative bacteria. TLR4 is expressed by B cells, myeloid dendritic cells, monocytes, macrophages, granulocytes, and T-cells. Several recent reports indicate that TLR4 is targeted by viruses such as the respiratory syncytial virus (RSV) (16), hepatitis C virus (HCV) (17), and mouse mammary tumor virus (MMTV) (18). Previous studies have established that TLR4 expression from its promoter is mainly regulated by the transcription factor PU.1 (19).
PU.1 is a member of the ets family of transcription factors with a restricted expression to B lymphocytes, macrophages and all hematopoietic lineages except T-cell lines and mature T-lymphocytes. PU.1 levels increase during granulocytic/ monocytic differentiation of immature hematopoietic progenitor cells and decline during erythroid differentiation (20). The human PU.1 protein consists of glutamine-rich and acidic residues toward the N-terminal necessary for transactivation, and DNA binding (ets) domain at the C-terminal (21). PU.1 regulates the expression of lymphoid as well as myeloid genes. Among some of the human myeloid genes regulated by PU.1 are M-CSF receptor, G-CSF receptor, GM-CSF receptor, IL-1␤, macrophage inflammatory protein 1␣ (MIP-1␣), and tumor necrosis factor ␣ (TNF-␣). Although CD4ϩ T lymphocytes represent the primary target for HTLV-I infection, the virus can infect other cell types including CD8ϩ T lymphocytes, dendritic cells, B lymphocytes, and central nervous system (CNS) astrocytes, and monocyte lineage cells such as tissue macrophages (22)(23)(24)(25). In vitro infection of HTLV-I in human monocytes, macrophages and microglial cells has also been demonstrated (26). Of note spliced viral mRNAs encoding for p40tax, p27rex, p12, and p30 were also identified in human macrophages infected with HTLV-I (25).
Here, we report that HTLV-I p30 protein targets the TLR4 signaling pathway. We found that virus-encoded protein p30 binds to and inhibits PU.1 DNA binding activity and PU.1-dependent transcription, leading to the down-regulation of TLR4 cell surface expression. As a result LPS-stimulated macrophages expressing p30 have a marked decrease in pro-inflammatory and an increase in anti-inflammatory cytokines released. This strategy may help the virus to evade the host innate immune responses and could make innate immune cells tolerant to opportunistic infections in ATL patients.

EXPERIMENTAL PROCEDURES
Cell Lines-The human 293T and COS-7 cell lines were maintained in Dulbecco modified Eagle's medium (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum (Hyclone, Logan, UT). The human monocytic cell line, THP-1 was maintained in RPMI (Invitrogen), 10% fetal bovine serum, and was passaged every 3 days.
Yeast Two-hybrid Screening-Full-length p30 cDNA sequence was cloned in-frame to LexA in between EcoRI and XhoI site of the pHybLex/Zeo vector (Invitrogen). This bait plasmid construct was transformed into L40 yeast strain (Invitrogen) and the bait-LexA fusion protein expression was confirmed by Western blot. To test for nonspecific activation of the bait, ␤-galactosidase filter assay was performed. Yeast twohybrid interactor hunt was done using a normal adult human spleen cDNA library in pYESTrp2 plasmid (Invitrogen) as prey and with our already tested bait. The potent interactors were selected on several YC(-WHUK) selective plates and confirmed by ␤-galactosidase filter assay. The interactor cDNA was retrieved from yeast by PCR amplification, cloned into pCR2.1TA vector (Invitrogen), and sequenced. Data base searches were performed using BLASTX.
Expression and Purification of hPU.1 Deletion Mutants-The full-length hPU.1 and the truncated mutants of hPU.1 were generated by PCR and cloned in-frame into BamHI/EcoRI site of pGEX-2T expression vector (Amersham Biosciences). The respective GST-deletion constructs were transformed into Escherichia coli BL21 cells. Expression of the fusion proteins was then induced for 3 h with 1 mM isopropyl-␤-Dthiogalactopyranoside at 37°C. The cells were sonicated in 1ϫ PBS, centrifuged, and the supernatants for respective constructs were mixed with a slurry of glutathione-Sepharose 4B (Amersham Biosciences) in PBS and incubated at room temperature. After centrifugation, the respective pellets were washed three times with 5 volumes of 1ϫ PBS with Complete and eluted with 10 mM reduced glutathione (Amersham Biosciences) in elution buffer (50 mM Tris-HCl, pH 8.4). The purified proteins were resolved in 10% SDS-PAGE and detected by Coomassie staining. PU.1 deletion mutants with Myc tag were constructed by cloning the truncated ⌬E and ⌬A fragments of PU.1 into the SalI/NotI sites of pCMV/myc/cyto vector. The expression of these mutants was confirmed by Western blot with Myc antibody (A-14, Santa Cruz Biotechnology).
In Vitro and in Vivo Binding Assay-p30-HA was in vitro transcribed and translated by using TNT Quick-coupled Transcription/Translation kit (Promega). For in vitro binding, the in vitro translated p30-HA was mixed with anti-GST antibody (Amersham Biosciences) and 100 ng of purified hPU.1-GST or truncated hPU.1-GST mutants in binding buffer (50 mM Tris-Cl, pH 7.6; 50 mM NaCl; 0.5 mM EDTA; 5 mM MgCl 2 , 0.1% Triton X-100, and 5% glycerol) containing 2.5 mg/ml bovine serum albumin and Complete protease inhibitor. After overnight binding at 4°C, protein G-agarose slurry was added, and the mixture incubated for 2 h with rotation at 4°C. The immunoprecipitated complex was washed two times for 15 min with the binding buffer without bovine serum albumin. The beads were boiled in loading dye, resolved on 15% SDS-PAGE, and detected by Western immunoblot using horseradish peroxidase-conjugated monoclonal HA antibody 3F10 (Roche Applied Science). For in vivo binding the Myctagged PU.1 deletion constructs were transfected in 293T cells with or without p30-HA, and the lysates in 1ϫ radioimmune precipitation assay buffer were co-immunoprecipitated with HA (12CA5) antibody followed by Western blot with anti-Myc (A-14) antibody.
Luciferase Assays and Western Blots-Luciferase assays to detect the effect of p30 on PU.1 expression was carried out in COS-7 cells with a trimerized PU.1 responsive-luciferase element, pTKPU.1 ϫ 3 Luc, a gift by Dr. T Oikawa. COS-7 cells were seeded at a density of 1 ϫ 10 6 cells/60-mm dish and transfected using Effectene reagent (Qiagen). The amounts of different constructs transfected are as indicated in Fig. 4. 36-h posttransfection, cells were lysed in 1ϫ luciferase lysis buffer (Promega) and analyzed using the Luciferase Reporter Assay system (Promega) according to the manufacturer's protocol. Luciferase activity was normalized with protein concentration. Data represent results obtained from two independent experiments. To study the effect of p30 on TLR4 expression, transient transfection of human monocytic THP-1 cells was carried out using Effectene (Qiagen) with TLR4-P-Luc reporter construct (gift of Dr. M Rehli). THP-1 cells were diluted with RPMI-10% fetal bovine serum the day before transfection. Next day the cells were centrifuged, washed two times with 1ϫ PBS and seeded at a density of 5 ϫ 10 6 cells/60-mm dish. Transfection was carried out at a ratio of 8 (enhancer):25 (effectene) for 1 g of transfected DNA. The cells were lysed 36-h post-transfection and analyzed as described above. Data (Fig. 5B) represent results from two independent experiments. To show that similar effects could be seen with physiological expression of p30, luciferase assay was carried out with TLR4-P-Luc reporter construct in THP-1 cells with wild type (pHTLV-X1MT) and p30/p12-deleted (pHTLV-⌬PSX) HTLV-I molecular clones. Western blots were carried out with equal amounts of protein extracts to confirm the expression of transfected constructs. As a negative control, we used pGL3-MRE (myb-responsive element) luciferase plasmid (27) and repeated the same experiment with a dose-dependent increase of p30. PU.1 was detected with T-21 antibody, c-Myb with C-19, p30 with 3F10, actin with C-11 (Santa Cruz Biotechnology), and ␤-tubulin with H-235 (Santa Cruz Biotechnology). To show that the activation of TLR4 promoter was specific, as a negative control (interferon regulatory factor-1) IRF-1 was cloned between SalI-NotI sites of pCMV/myc/cyto and the IRF-1 Myc-tagged expression plasmid was used in increasing doses along with a TLR4-luciferase promoter reporter construct. Western blots to detect phospho-GSK3 (Ser 9 ) and phosphoglycogen synthase (Ser 641 ) were carried out using phosphospecific antibodies from Cell Signaling.
Biotin-labeled DNA Pull-down Assay-To examine the effects of p30 on the DNA binding activity of PU.1 and to elucidate if increasing amounts of p300 restore this activity, 40 ng of biotinylated trimerized PU.1 responsive element (PU.1 ϫ 3) was added to 100 l of nuclear extract from 293T cells transfected with either 5 g of PU.1 and increasing amounts (2, 4, 6 g, respectively) of p30-HA expression plasmids (Fig. 4C) or 2 g of PU.1, 5 g of p30-HA and increasing amounts of p300 (1, 3, 5 g, respectively)( Fig. 4D) in a total volume of 200 l of binding buffer (25 mM HEPES pH 7.9, 5 mM KCl, 0.5 mM MgCl 2 , 0.5 mM EDTA, 1 mg/ml bovine serum albumin, 10% (v/v) glycerol, and 0.25 mM dithiothreitol). The nuclear mixtures with biotin-labeled probe were preincubated on ice for 2 h with gentle agitation. 60 l of 50% slurry of streptavidin magnetic beads (Roche) was added to the nuclear lysates and kept on ice for 1 h. The beads were collected by a magnet, washed two times with binding buffer and resuspended in 2ϫ SDS-PAGE loading buffer. The samples were heated for 5 min at 95°C and loaded on 12% SDS-polyacrylamide gel, blotted, and detected for PU.1 using T-21 antibody (Fig. 4, C and D).
Flow Cytometry for TLR4 Cell Surface Expression-Human monocytic THP-1 cells were transfected with p30 (0.5 g) using Nucleofector Kit V (Amaxa Biosystems) according to the manufacturer's protocol. For this study, untransfected and p30 transfected THP-1 cells were unstimulated or stimulated with 5 M phorbol 12-myristate 13-acetate (PMA) (CalBiochem) for 3 h after 15 h post-transfection. Cells were harvested by centrifugation at 600 rpm, washed two times with 1ϫ PBS and incubated with anti-human TLR4 monoclonal antibody HTA125 (BD Biosciences) or isotype control at a final concentration of 10 g/ml for 1 h on ice. Cells were then washed twice and incubated with FITC-conjugated mouse IgG (BD Biosciences) at a final concentration of 5 g/ml for 1 h on ice. Cells were then washed twice and analyzed by a flow cytometer equipped with the manufacturer's software (FACS Diva, Becton Dickinson) for data acquisition and analysis. As a negative control, the surface expression of CD14 was analyzed as described above with CD14-FITCconjugated antibody by flow cytometry in THP-1-untransfected or p30-transfected cells by Nucleofector kit V. To demonstrate physiological relevance of p30Ј effects, the TLR4 surface expression was measured as described above by flow cytometry in THP-1 cells by nucleofection of wild type (pHTLV-X1MT) and p30/p12-deleted (pHTLV-⌬PSX) HTLV-I molecular clones. Real-time RT-PCR-Total RNA was extracted from untransfected and p30 transfected THP-1 cells by TRIzol (Invitrogen) and treated with DNaseI (Invitrogen). The total RNA was reverse-transcribed, and the resulting cDNA was analyzed by real-time PCR using human TLR4 and GAPDH primer sets (Superarray BioScience). The authenticity of the PCR products was verified by melting curve analysis.
Cytokine Profile Analysis-THP-1 cells (2 ϫ 10 6 ) were plated into 24-well dishes and were transfected with p30 by Nucleofector kit. After 12 h of post-transfection, cells were stimulated with LPS (Sigma). Supernatants were collected at intervals of 3 h and were analyzed for the levels of tumor necrosis factor-␣ (TNF-␣), IL-8, MCP-1 (monocyte chemoattractant protein-1), and IL-10 using the Beadlyte TM human multicytokine detection system 3 from Upstate Cell Signaling Solutions (Lake Placid, NY) as per the manufacturer's instructions. Briefly, the multicytokine 3 standard was resuspended in assay buffer and then serially diluted from 2500 to 15.6 pg/ml. 50 l of standard or sample was added to each well of a 96-well plate with 25 l of the bead solution and was incubated overnight at 4°C. The Beadlyte TM reporter solution was added to each well and incubated at room temperature for 1.5 h. Beadlyte TM streptavidin-phycoerythrin was diluted 1:25 in assay buffer and was added to each well and incubated at room temperature for 30 min before the addition of the Beadlyte TM stop solution. The plate was then analyzed on the Luminex LabMAP TM system (Luminex Corp., Austin, TX) and analyzed using Beadview multiplex data analysis software (Upstate).
Statistical Data-Statistical analyses were performed by using Student's t tests or one-way variance analysis (independent group analysis). Statistical significance was considered to be p Ͻ 0.05.

Interaction between p30 and Transcription Factor PU.1 in a Yeast
Two-hybrid Assay-To identify binding partners of the HTLV-1 p30, we performed a yeast two-hybrid screen using the full-length p30 protein as bait. L40 yeast cells were cotransformed by a plasmid encoding a chimera between p30 and the Lex A transcription factor DNA binding domain, along with a normal spleen-derived cDNA library fused in-frame to the Gal-4 activation domain (Invitrogen). The fusion protein LexA-p30 was expressed in L40 yeast strain as demonstrated by Western blot analysis (Fig. 1A). Following screening on selective media (Fig. 1B), ␤-galactosidase-positive clones were retrieved by PCR, cloned, and sequenced. One of the isolated clones was found to correspond to a fragment of the PU.1 transcription factor.
This clone was selected for further analysis on the basis of its hematopoietic restricted expression and also because of the fact that p30 has been reported to act as a transcriptional and posttranscriptional regulator of gene expression. We next determined if HTLV-1 p30, without being fused to Lex A, could associate with PU.1 in mammalian cells. 293T cells, which do not express endogenous PU.1, were transfected with an untagged PU.1 expression vector with or without an HA-tagged p30 vector. In vivo interaction was detected by co-immunoprecipitation using an anti-HA antibody followed by Western blot to detect untagged PU.1. As expected, HA antibody immunoprecipitated p30 only from transfected cells (Fig. 1C, lanes 3 and  4). PU.1 was readily immunoprecipitated and detected by Western blot only when both p30 and PU.1 were coexpressed (Fig. 1C, lane 4). Levels of p30 and PU.1 were analyzed by Western blot (Fig. 1C, lower panels). To ensure that interactions were specific we coexpressed c-Myb and p30 in 293T cells. Under the same conditions in which p30 interacted with PU.1, no binding was detected between p30 and c-Myb (Fig. 1D). These results indicate that p30 and PU.1 specifically interact in transfected cells and that p30 does not affect the expression level of PU.1 when expression of the latter is driven by a CMV promoter.
P30 Interacts with the ets Domain of PU.1-To identify which domain of PU.1 is involved in binding p30 we constructed several GST fusion constructs deleted for the transactivation, the DNA binding or both domains ( Fig. 2A). These vectors were expressed in BL-21 bacteria (Fig. 2B) and purified to near homogeneity as shown by SDS-PAGE and Coomassie staining (Fig. 2C). In vitro binding assays were performed using the different GST-PU.1 purified proteins along with rabbit reticulocyte in vitro translated p30HA protein and binding of p30 was detected by Western blot using anti-HA antibody (Fig. 2D). As expected from results presented in Fig. 1, p30 interacted with the full-length GST-PU.1 but not to GST control (Fig. 2D). In fact, this result indicated that the ets domain of PU.1 is required and is sufficient for in vitro binding of p30. Comparable input of in vitro translated p30 and of each GST-PU.1 fusion protein used in each binding reaction was confirmed by Western blot using HA and GST antibody, respectively (Fig. 2, E and F). To further confirm the binding site in vivo, Myc-tagged PU.1 truncated mutants were constructed (Fig. 3A) and coexpressed along with p30 in 293T cells. Immunoprecipitation assays confirmed that p30 interacts with the ets domain of PU.1 (Fig. 3B). Levels of expression of both PU.1 truncated mutants were not affected by p30 (Fig. 3C).   ence of HTLV-I p30, we used a PU.1-responsive reporter construct containing 3 tandem PU.1 binding sites (PU.1 ϫ 3-TK-LUC). When cotransfected with a PU.1 expression vector (pcDNA.PU.1) into COS-7 cells, activation of this reporter construct was readily observed (Fig. 4A); however, this activation decreased when p30 expression vector was added (Fig. 4A). No repression of basal promoter activity by p30 was observed in the absence of PU.1 (Data not shown). The amount of p30 expressed was monitored by Western blotting with an anti-HA antibody (Fig. 4A). The decrease of PU.1 transactivation was not because of a decrease in PU.1 protein expression from the CMV promoter as Western blot of extracts from p30-transfected cells showed PU.1 levels similar to those found in the absence of p30 (Fig. 4A).
Previous studies have reported that p30 interaction with the coactivator p300 competes with the viral transactivator Tax protein for recruitment of p300 onto the transcription complex of the viral LTR thereby resulting in lower virus expression, which can be rescued by increasing amounts of p300 (28,29). Because PU.1 is also known to interact with coactivators CBP/p300 (30,31) and this interaction is required for transcriptional activity of PU.1, we tested whether p30 may affect PU.1-dependent transcription. COS-7 cells were transfected with PU.1 ϫ 3-TK-LUC, a PU.1-specific reporter construct along with PU.1, p30, and increasing amounts of CBP or p300 expression vectors. Under these experimental conditions, p30 was able to inhibit PU.1-mediated transcription in a dose-dependent manner (Fig. 4B) and p300 but not CBP was able to rescue p30-mediated repression (Fig. 4B). Neither p30 nor PU.1 expression was significantly affected in these assays as shown by Western blot analysis (Fig. 4B). Because previous studies have demonstrated that PU.1 ets domain is involved in DNA binding activity (32,33), and this domain interacted with p30, we tested whether such interaction may impair ability of PU.1 to bind DNA. The binding assays were performed using biotinylated DNA probe containing 3ϫ PU.1 consensus binding sites incubated with nuclear protein extracts from cells transfected with PU.1 in the absence or presence of an increasing amounts of p30 as described under "Experimental Procedures." (Fig. 4C). Results from these experiments demonstrated that p30 inhibits the PU.1 DNA binding activity in a dose-dependent manner (Fig. 4C) thereby preventing PU.1-dependent transcription. Consistent with luciferase results presented above increasing amounts of p300 efficiently competes with p30 and restore DNA binding activity of PU.1 (Fig. 4D).
P30-mediated Inhibition of PU.1 in Human Macrophages-As mentioned above HTLV-I has been found to infect and replicate in human macrophages in vitro as well as in vivo. To confirm that p30 represses PU.1 transcription in cells expressing endogenous PU.1, the human macrophage cell line THP-1 was transfected with a PU.1 reporter construct along with increasing amounts of p30. Consistent with results presented above, p30 was able to inhibit endogenous PU.1-mediated transcription in human macrophages in a dose-dependent manner (Fig. 5A). Under these experimental conditions a decreased expression of endogenous PU.1 protein can be observed. This is in contrast with PU.1 expressed from a CMV-driven promoter (Figs. 3 and 4) and is explained by the fact that PU.1 autoregulates expression from its own promoter, therefore decreased PU.1 transcriptional activity will effect in a decrease in endogenous PU.1 expression (34).
P30-mediated Inhibition of PU.1 Decreases Cell Surface Expression of the TLR4-Expression of human TLR4, a major component of innate immune response against Gram-negative bacteria, is mainly regulated by the transcription factor PU.1. Therefore, results outlined above would predict a repressive effect of p30 on TLR4 expression. To verify this hypothesis a TLR4 promoter luciferase reporter construct was transfected in THP-1 cells. In agreement with the results presented above, p30 expression efficiently repressed TLR4 promoter activity, in a dose-dependent manner (Fig. 5B). We also confirmed our finding by measuring endogenous TLR4 mRNA expression by quantitative real time RT-PCR. In these assays p30 expression reproducibly resulted in significant ( p ϭ 0.007) 3.5-4-fold down-regulation of TLR4 mRNA expression in THP-1 cells (Fig. 5C). The specific effect of p30 on the TLR4 promoter was demonstrated by transfection of a Myb responsive luciferase construct in THP-1 cells, which express abundant levels of endogenous c-Myb protein. In this experiment, the dose increase amount of p30 had no effect on Myb-mediated transactivation showing that the effect observed on the TLR4 are specific, and p30 does not generally interfere with transcription ( Fig. 5D). To further demonstrate that the inhibition on the TLR4 luciferase vector was caused by p30 expression and not transfection artifacts, we cloned IRF-1, a transcription factor involved in immunity but not in TLR4 expression or activity. In contrast to p30-mediated inhibition, transfection of a dose increase of IRF-1 had no significant effect on the TLR4 promoter reporter construct (Fig. 5E). Together these results demonstrate a specific effect of HTLV-I p30 on the TLR4 promoter.
Because TLR4 signaling is transduced by interactions of the ligands with the TLR4 receptor expressed on the cell surface, we further investigated surface expression of TLR4 by FACS analysis in mock-or p30-transfected THP-1 cells using nucleofactor according to the manufacturer's instructions. Under our experimental conditions, nearly 70% of THP-1 cells were transfected as monitored by transfection of a GFP-expressing vector (data not shown). Previous studies have shown that expression of TLR4 is very low on THP-1 cells but drastically increase following phorbol myristyl acetate (PMA) stimulation (35). In the absence of stimulation by PMA, p30 expression resulted in a small decrease in TLR4 surface expression (5%) in THP-1 cells (Fig. 6A). In contrast, p30 had a more significant effect following PMA stimulation resulting in nearly 70% down-regulation of TLR4 expression from the THP-1 cell surface (Fig. 6A). The presence of two distinct populations in PMA-stimulated THP-1 cells detected by FACS, based on their level of TLR4 expression (medium and high) has been previously reported (35). To ensure that the down-regulation of TLR4 from the cell surface was specifically caused by p30 expression and not experimental procedure, we measured CD14 surface marker expression under the same experimental conditions. In contrast to TLR4 sharp down-regulation, CD14 expression in THP-1 cells stimulated with PMA was only marginally affected by p30 expression (Fig. 6B).
p30 Down-regulates TLR4 under Physiological Conditions when Expressed in the Context of an HTLV-I Molecular Clone-Because the data presented above relied on high levels of p30 expression from a CMVdriven promoter, we decided to investigate whether these effects could be reproduced when p30 was expressed at more physiological levels in the context of an infectious HTLV-I molecular clone. HTLV-I wild-type molecular clone (pHTLV-X1MT) and p30/p12-deleted pHTLV-deltaPSX (kindly provided by D. Derse) (36) were used for these assays. Whereas the wild-type HTLV-I molecular clone suppressed TLR4 promoter activity, pHTLV-deltaPSX deleted for p30/p12 had no effect (Fig. 7A). Although the pHTLV-deltaPSX is deleted for both p30 and p12, the effect observed on the TLR4 promoter (Fig.  7A) can be attributed to p30 because cotransfection of pHTLV-deltaPSX along with a p12 expression vector had no effect (Fig. 7B). To further confirm the effect on endogenous TLR4 surface expression, THP-1 cells were nucleofected with pHTLV or pHTLV-deltaPSX and TLR4 expression was quantified by FACS analysis. Under these experimental conditions, physiological expression of p30 from an HTLV-I molecular clone resulted in 30% down-regulation of TLR4 expression (Fig. 7C) as opposed to 70% when p30 was expressed from a CMV promoter (Fig. 6A). In fact, pHTLV-deltaPSX lacking p30 and p12 had no effect on TLR4 but this could be rescued by coexpression of p30 thereby confirming that the effect is dependent on p30 and not p12. This further excludes the possibility that another viral gene missing from the pHTLV-deltaPSX may be involved (Fig. 7C). Together these data demonstrate that physiological expression of p30 from a molecular clone is sufficient to repress TLR4 expression in a biological setting.
P30 Expression Reduces Pro-inflammatory Cytokine Production from LPS-stimulated THP-1 Cells-Interaction of LPS with TLR4 activates downstream signaling pathways and mediates the release of pro-inflammatory cytokines such as MCP-1, FIGURE 7. p30 down-regulates TLR4 under physiological condition when expressed in the context of an HTLV-I molecular clone. A, luciferase assay was carried out with 0.5 g of TLR4-P-Luc reporter construct in THP-1 cells with 0.5 g each of wild type (wtHTLV) and p30/p12-deleted (HTLVp30del) HTLV-I molecular clones respectively. Data represents results from two independent experiments. B, luciferase assay with 0.5 g of TLR4-promoter luciferase reporter and 0.5 g of p30/p12-deleted (HTLVp30del) molecular clone with or without 0.5 g of pMH-p12HA vector. C, flow cytometry of cell surface expression of TLR4 in THP-1 cells after nucleofection with (0.5 g each) of HTLV-I wild type and p30-deleted molecular clones. The percentage of TLR4 down-regulation/increase is indicated in the panel.
TNF-␣ and IL-8 (37). Because p30 reduces the cell surface expression of TLR4, we investigated whether this effect may be associated with variation on cytokine release following stimulation of TLR4 by LPS. As expected pro-inflammatory cytokines are produced in very low amounts by THP-1 cells in the absence of LPS stimulation. Our studies showed no significant effect of p30 under these conditions (data not shown). However, the presence of p30 strongly and significantly ( p Ͻ 0.05) decreased the release of pro-inflammatory cytokines MCP-1, TNF-␣, and IL-8 after stimulation with LPS (Fig. 8, A-C).
P30 Inhibits GSK-3␤ Activity and Increases Anti-inflammatory Cytokine IL-10 Release by THP-1 Cells Following Stimulation-Recent studies showed that inhibition of GSK3-␤ potently suppresses the production of pro-inflammatory cytokines while concurrently augmenting production of anti-inflammatory cytokine, IL-10 in response to multiple agonists of TLR signaling pathways (38). We found that p30 expression resulted in increased IL-10 production by THP-1 cells following LPS stimulation (Fig. 9A). This finding was of interest because 80% of ATL patients have high levels of IL-10 in serum (39).
These results prompted us to investigate whether p30 may modulate GSK3 kinase activity. Phosphorylation of the constitutively active GSK3-␤ on position Ser 9 results in its inhibition (40) and an increased IL-10 secretion in response to TLR stimulation (38). In agreement with a recent study, we found increased GSK3-␤ Ser 9 phosphorylation in response to treatment of THP-1 cells with a specific inhibitor of GSK3 ␤ (SB216763) or following LPS stimulation (Fig. 9B). Importantly, p30 expression also resulted in an increased GSK3-␤ Ser 9 phosphorylation suggesting that p30 inhibits GSK3 kinase activity (Fig.  9B). In support of this finding, we observed a similar decrease in levels of phosphorylated glycogen synthase on Ser 641 , a direct target site for GSK3-␤, in samples from cells treated with either a specific inhibitor of GSK3-␤ or cells expressing p30 (Fig. 9B).

DISCUSSION
This study reports for the first time an effect of the HTLV-I-encoded protein on the Toll-like receptor 4 signaling pathway and suggests that HTLV-I subverts components of the host innate immunity. We found that HTLV-I accessory protein p30 interacts with the transcription factor PU.1. We further delineated the binding domain of PU.1 by in vivo and in vitro binding experiments showing that p30 interacted with the ets domain of PU.1. As a result of this interaction DNA binding activity of PU.1 was greatly impaired and PU.1-dependent transcription inhibited in a dose-dependent manner by p30. Consistent with the fact that both PU.1 and p30 interact with the coactivator p300, exogenous expression of the latter restored both DNA binding and transcriptional activity of PU.1. Previous studies have demonstrated that TLR4 promoter is mainly controlled by PU.1 and TLR4 signaling is targeted by viruses such as RSV, HCV, and MMTV. We found that p30 expression in human macrophage like THP-1 cells resulted in inhibition of PU.1-dependent transcription and decreased expression of TLR4 mRNA by real-time quantitative RT-PCR along with decreased  . p30 inhibits GSK-3␤ and augments anti-inflammatory IL-10 production in response to TLR4 stimulation by LPS. A, augmentation of IL-10 production by p30 in response to LPS stimulation in THP-1 cells. IL-10 production in p30-transfected and untransfected cells following 6 h of LPS stimulation assayed from supernatants by Luminex. B, THP-1 cells were transfected with (0.5 g) or without p30-HA by nucleofection and then either stimulated with 1 g/ml LPS or treated with 10 M GSK-3␤ inhibitor (SB216763) for 1 h, as indicated. Following incubation, cells were lysed in 1ϫ radioimmune precipitation assay buffer and resolved by a 4 -15% SDS-PAGE gradient gel. The blots were detected using phospho-GSK-3␤ (Ser 9 ) and phosphoglycogen synthase (Ser 641 ) antibody. ␤-Actin (C-11) was used as loading control. expression of TLR4 expression from the cell surface by FACS analysis. Importantly, inhibition of TLR4 by p30 was confirmed in a biological relevant system with physiological expression of p30 from an infectious molecular clone. We further demonstrated that, following LPS stimulation in THP-1 cells, p30 expression usurps TLR4 signaling and concurrently decreases the release of pro-inflammatory cytokines TNF-␣, MCP-1, IL-8 and increases the release of the anti-inflammatory cytokine IL-10. These findings are relevant to ATL pathogenesis because it has been reported that 80% of ATL have elevated serum IL-10 levels that correlated with disease progression. Moreover, pretreatment with IL-10 protects tumor cells from lysis by tumorspecific CTLs (42). In addition, the combination of IL-10 and TNF-␣ stimulate AP1 and NF-B DNA binding both of which are constitutively activated in ATL patient samples (43,44).
HTLV-I induces a lifelong infection despite a very low antigenic variability and consequently has evolved multiple strategies to evade host immune clearance including down-regulation of MHC class I by HTLV-I p12 (45), suppression of viral expression by HTLV-I p30 (46,47), and inhibition of Tax-mediated transcription by HTLV-I HBZ (48). The results presented here suggest an additional role of p30 in targeting TLR4 signaling and inducing immunosuppressive signals. IL-10 inhibits a broad spectrum of activated macrophage and monocyte functions including the production of IL-12, nitric oxide, expression of MHC class II and co-stimulatory molecules (49), and down-regulate immunity to various pathogens including retroviruses (50).
The dendritic (DC) cells located in the blood and periphery are functionally immature in that they are effective in the uptake and processing of antigens but not in stimulating CD4ϩ or CD8ϩ lymphocytes. After exposure to inflammatory cytokines and bacterial products, DCs undergo maturation and are able to present antigens for priming and stimulating T cells. Therefore, interference of TLR4 signaling by p30 and reduced pro-inflammatory cytokines released along with increased IL-10 may impair the ability of dendritic cells to activate adaptive immunity in ATL patients and thereby explaining the limited proliferation of virus-specific CTL reported in ATL patients. In fact, HTLV-I-infected DCs obtained from the peripheral blood of ATL patients have been found to be defective in stimulating proliferation of CD4 and CD8 T-cells (51).
Consistent with the recent finding that inhibition of GSK3-␤ stimulates IL-10 production (38), we found that p30 increased the phosphorylation of GSK3-␤ on serine 9, a response known to inactivate kinase activity. In fact, we also demonstrated reduced GSK3-␤-dependent kinase activity by reduced phosphorylationofglycogensynthaseadirecttargetspecificallyphosphorylated by GSK3-␤. Taken together our results demonstrate novel functions of HTLV-I in exploiting host innate immunity by targeting the TLR4 signaling pathway, inducing decrease in pro-inflammatory cytokines production and by inhibiting GSK3␤ kinase activity leading to increase in anti-inflammatory IL-10 secretion. The mechanism by which p30 induces phosphorylation and inactivation of GSK3 remain to be uncovered and is currently under investigation. Previous studies have also shown that GSK3 inhibition selectively increases the amount of nuclear CREB (Ser 133 ) DNA binding activity and its association with the coactivators CBP/p300 without any discernible effects on the amount of nuclear NF-B p65 (RelA) associated with the coactivator (38). Therefore, p30-mediated inhibition of GSK3␤ may, in part, explain the suppressive effect of p30 on Tax-mediated viral LTR transactivation.
Finally, recent evidence indicates that graded reduction in the expression of PU.1 in CD34ϩ derived bone marrow stem cell leads to an intermediate stage of poorly differentiated preleukemic population, which, with the accumulation of additional genetic mutations results in an aggressive form of acute myeloid leukemia (AML) (41). Along these lines it has been shown that HTLV-I can infect and replicate in bone marrow derived undifferentiated CD34ϩ stem cells (52). As the origin of leukemic ATL cells is unknown; it is possible that ATL cells are derived from infection of undifferentiated cells such as CD34ϩ, which later differentiate into CD4 T-cell. Combined with our data, p30-mediated reduction of PU.1 expression, these studies allow us to speculate that p30 could play an important role in setting an appropriate cellular environment for further Tax-mediated genetic alterations and transformation which warrant further studies.