A Single Nucleotide Polymorphism in 3′-Untranslated Region Contributes to the Regulation of Toll-like Receptor 4 Translation*

Background: Genetic variation of SNP rs11536889 in 3′-UTR of TLR4 is implicated in certain diseases, including periodontitis, gastric atrophy, and prostate cancer. Results: The G allele of rs11536889 inhibited translation, but not transcription, of TLR4. Conclusion: Genetic variation of rs11536889 regulates TLR4 expression. Significance: Polymorphism in rs11536889 could be an excellent genetic marker for the diseases caused by TLR4-ligands. We have previously shown that a single nucleotide polymorphism rs11536889 in the 3′-untranslated region (UTR) of TLR4 was associated with periodontitis. In this study the effects of this single nucleotide polymorphism on Toll-like receptor (TLR) 4 expression were investigated. Monocytes from subjects with the C/C genotype expressed higher levels of TLR4 on their surfaces than those from subjects with the other genotypes. Peripheral blood mononuclear cells (PBMCs) from the C/C and G/C subjects secreted higher levels of IL-8 in response to lipopolysaccharide (LPS), a TLR4 ligand, than the cells from the G/G subjects. However, there was no significant difference in TLR4 mRNA levels in PBMCs from the subjects with each genotype. After stimulation with tripalmitoylated CSK4 (Pam3CSK4), TLR4 mRNA levels increased in PBMCs from both the C/C and G/G subjects, whereas TLR4 protein levels increased in PBMCs from the C/C but not G/G subjects. Transient transfection of a series of chimeric luciferase constructs revealed that a fragment of 3′-UTR containing rs11536889 G allele, but not C allele, suppressed luciferase activity induced by LPS or IL-6. Two microRNAs, hsa-miR-1236 and hsa-miR-642a, were predicted to bind to rs11536889 G allele. Inhibition of these microRNAs reversed the suppressed luciferase activity. These microRNA inhibitors also up-regulated endogenous TLR4 protein on THP-1 cells (the G/G genotype) after LPS stimulation. Furthermore, mutant microRNAs that bind to the C allele inhibited the luciferase activity of the construct containing the C allele. These results indicate that genetic variation of rs11536889 contributes to translational regulation of TLR4, possibly by binding to microRNAs.

Toll-like receptor 4 (TLR4) 2 is a pattern recognition receptor that plays an essential role in the recognition of Gram-negative bacterial lipopolysaccharides (LPS) (1). The binding of LPS to myeloid differentiation factor 2-TLR4 complex induces the formation of an M-shaped receptor multimer and leads to the assembly of the MyD88-IRAK4-IRAK2 complex (2). Formation of these Myddosome complexes leads to nuclear factor-B and mitogen-activated protein kinase activation, which induces proinflammatory cytokines (3,4). TLR4 also recruits TIR-containing adaptor-inducing interferon-␤ (TRIF) and TRIF-related adapter molecule in endosomes, which activates transcription factor interferon regulatory factor 3 and induces genes encoding type I interferons (5). TLR4 is reported to recognize exogenous ligands other than LPS, such as the fusion protein of the respiratory syncytial virus and the envelope protein of the mouse mammary tumor virus (6). In addition, endogenous molecules such as heat-shock proteins, ␤-defensin-2, high mobility group box 1 protein, and fibronectin are reported to interact with TLR4 (7).
Inherited defects in TLR4 signaling have been described in humans and are associated with greater susceptibility to bacterial infections (8). Two missense mutations in TLR4 (D299G and T399I) have been reported to be associated with endotoxin hyporesponsiveness to inhaled LPS (9). This report was followed by a series of studies investigating the potential impact of these Single nucleotide polymorphisms (SNPs) on the incidence and course of infectious diseases (10), such as septic shock with Gram-negative bacterial infection (11) and severe respiratory syncytial virus bronchiolitis in infants (12). Several papers reported the importance of these SNPs for incidences of cardiovascular diseases as well, which possibly reflects the inflammatory nature of atherosclerosis (13). Despite the accumulating reports concerning the susceptibility to such diseases, the recent studies using primary cells isolated from individuals * This work was supported Grant-in-aid for Scientific Research (C) 20592431 bearing the mutation have indicated that the D299G/T399I haplotype has little or no effect on responsiveness to LPS (14).
We tried to determine whether D299G and T399I mutations in TLR4 were associated with periodontitis because TLR4 is considered to be involved in the pathogenesis of periodontal diseases (15). However, two missense mutations in TLR4 (D299G and T399I) are very rare in the Japanese population (16). We searched for SNPs in TLR4 because it is possible that other SNPs are associated with periodontitis in the Japanese population. We found four SNPs in the exons of TLR4, although none of them resulted in amino acid substitutions. One SNP in the 3Ј-untranslated region (UTR) of TLR4, National Center for Biotechnology Information ID rs11536889, was associated with moderate to severe periodontitis in Japanese subjects. The C/C genotype of rs11536889 was significantly higher in both the moderate and the severe periodontitis group than in the control group (16). Another large study analyzing eight TLR4 SNPs in 1383 prostate cancer patients and 780 age-matched controls in Sweden revealed that the G/C or C/C genotype of rs11536889 increased prostate cancer risk by 4.9% (17). Later, Hishida et al. (18) found that the G/C or C/C genotype was associated with severe gastric atrophy in Helicobacter pylori seropositive Japanese subjects. Zhou et al. (19) found that the C/C genotype was significantly associated with hepatitis type B virus recurrence after liver transplantation. Miedema et al. (20) found that this SNP was associated with an increased risk of developing chemotherapy-induced neutropenia in children with acute lymphoblastic leukemia. These findings suggest that the genetic variation of rs11536889 may have influence on human inflammatory and/or malignant diseases. Because rs11536889 is located in the 3Ј-UTR of TLR4, it should not have a direct influence on the conformation of the TLR4 protein molecule. However, because SNPs in introns and/or UTRs could influence transcription and/or translation (21)(22)(23)(24), the genetic variation of rs11536889 may have a direct effect on mRNA stability or translation efficiency.
The aim of this study was to determine whether the rs11536889 polymorphism is associated with expression or function of TLR4. The expression levels of TLR4 mRNA and protein in peripheral blood from C/C subjects were compared with those from G/C and G/G subjects. The regulatory effects of TLR4 3Ј-UTR on gene expression was measured by a luciferase reporter assay using chimeric constructs containing a small fragment of 3Ј-UTRs containing rs11536889 in THP-1 cells. MicroRNA (miRNA) are ϳ22 nucleotide non-coding RNAs that guide the RNA-induced silencing complex to the 3Ј-UTR of mRNA targets, leading to the inhibition of translation of the target mRNA (25). The possible effects of miRNA were examined using miRNA inhibitors and miRNA mimics. We show here the regulation of TLR4 expression by rs11536889 polymorphism.

EXPERIMENTAL PROCEDURES
Subjects-Twelve C/C subjects and their age-and sexmatched G/C and G/G controls whose genotype were analyzed in our previous study (16) were enrolled in the flow cytometry study. Another 100 subjects who visited Nagasaki University Hospital were screened for rs11536889 genotypes; C/C subjects and their controls were enrolled in the enzyme-linked immunosorbent assay (ELISA) and quantitative reverse transcription PCR (qRT-PCR) study. All of the subjects were Japanese. Individuals with pregnancy, diabetes mellitus, malignancy, immunodeficiencies, or infectious diseases such as acquired immune deficiency syndrome and adult T cell leukemia were excluded. The gender and mean age of the subjects are shown in Table 1. There was no significant difference in age and male/female ratio between each genotype group. Written informed consent was obtained from all of the participants in this study. The experimental protocol was approved by the Ethics Committee in Nagasaki University.
Genotyping-Buccal mucosal cells were taken from 100 newly enrolled subjects in this study. DNA was extracted with a DNeasy Blood and Tissue kit (Qiagen, Valencia, CA). The TLR4 rs11536889 polymorphism was genotyped by PCR with confronting two-pair primers according to the procedure by Hishida et al. (18). The sequences of the primers were described previously (18). Amplification conditions were 10 min of initial denaturation at 95°C followed by 35 cycles of 1 min at 95°C, 1 min at 65°C, and 1 min at 72°C, then a 5-min final extension at 72°C. The amplified DNA was visualized on a 2% agarose gel with ethidium bromide staining. The amplified DNA was 184 bp for G allele, 256 bp for C allele, and 397 bp for common band.
Flow Cytometric Analysis of Freshly Isolated Peripheral Blood Monocytes-Peripheral blood mononuclear cells (PBMC) were isolated from 12 C/C subjects, and their controls with the G/C and G/G genotypes, and separated by gradient centrifugation of heparinized blood using endotoxin-free Ficoll-Paque (GE-Healthcare) according to manufacturer's protocol. The cells were double-stained with anti-CD14-PE (clone UCHM-1, Sigma) and anti-TLR2-FITC (clone TL2.1, eBioscience, San Diego, CA), anti-TLR4-FITC, (clone HTA125, Imgenex, San Diego, CA), or isotype-matched control-FITC (clone MOPC141, Sigma) monoclonal antibodies. After washing in 1% fetal bovine serum (FBS) (Moregate Biotech, Bulimba, Australia) in phosphate-buffered saline, the cells were resuspended in 2% formaldehyde for 30 min and washed again. The cells were resuspended in the solution consisting of 10% dimethyl sulfoxide (DMSO) (Sigma) and 90% FBS and stored at Ϫ80°C for batch analysis. On the day of analysis, all cells were thawed and washed in 1% FBS, and expression levels of TLR2 and TLR4 on the surface of CD14 ϩ cells were analyzed by flow cytometry (FACScan, BD Biosciences). Data were analyzed using FlowJo software (Tree Star, Ashland, OR).
Cytokine Production Assays-PBMCs were isolated from 19 C/C subjects and their controls with G/C and G/G genotypes as described above. The cells were resuspended in RPMI 1640 medium (Invitrogen) supplemented with 10% FBS. PBMCs were plated at a density of 1 ϫ 10 5 cells per well in a 96-well dish and stimulated with 100 ng/ml Pam 3 CSK 4 (InvivoGen, San Diego, CA) or 1 ng/ml ultrapure Escherichia coli LPS (Invivo-Gen). After incubation for 20 h, cell-free supernatants were harvested and analyzed for IL-8, IL-6, and TNF-␣ production using a commercial ELISA kit (R&D, Minneapolis, MN). Relative Quantification of TLR4 mRNA Levels-PBMCs were isolated from the C/C, G/C, and G/G subjects as described above (n ϭ 19 in each genotype group). Total RNA was extracted with RNeasy Mini kit (Qiagen) with on-column DNase treatment according to manufacturers' instructions and reverse-transcribed for 50 min at 42°C using 1 unit/l avian myeloblastosis virus reverse transcriptase (Promega, Madison, WI) and 60 ng/ml random primer (Invitrogen). The cDNA was treated with RNase H (Invitrogen) and purified with QIAprep Spin Miniprepkit (Qiagen). We failed to obtain cDNA from one G/C subject and one G/G subject because of the limitation of blood volumes withdrawn from these subjects; subsequent analysis was performed without these samples. The sequences of the primers and probes for TLR4 were described previously (26). The sequences of the primers and probes for the internal control, glyceraldehyde-3-phosphate dehydrogenase were 5Ј-CAT GTT CGT CAT GGG TGT GA-3Ј (forward), 5Ј-CAG GGG TGC TAA GCA GTT GG-3Ј (reverse), and 5Ј-FAM-CAT CAG CAA TGC CTC CTG CAC CA-TAMRA-3Ј (probe). These were designed using Primer 3 software and analyzed for homology to other known sequences using the Basic Local Alignment Search Tool (BLAST). To avoid the amplification of genomic DNA and to ensure that the PCR signal was generated from cDNA, primers were placed within different exons close to intron/exon boundaries. The amplification reactions were performed using Premix Ex Taq (Takara, Shiga, Japan) in a Mx3000 system (Stratagene, La Jolla, CA) with 20 l of final reaction mixture containing 2ϫ Premix EX Taq (10 l), 10 M (0.4 l) forward and reverse primers and probe (final concentration was 0.2 M), and cDNA (2 l). The thermal profile was as follows: 30 s at 95°C and 40 cycles of 5 s at 95°C and 20 s at 60°C. Relative quantification of TLR4 mRNA was performed according to the ⌬⌬Ct method (27). THP-1 is a human acute monocytic leukemia cell line that expresses endogenous TLR4. We used this cell line to analyze the effects of miRNAs on TLR4 expression. These cells were stimulated with ultrapure E. coli LPS (1 g/ml) and subjected to the analysis of TLR4 mRNA expression as described above.
miRNA Analysis-miRNA were isolated from PBMCs and THP-1 cells using the miRNeasy kit and RNeasy MinElute Cleanup kit (Qiagen) according to the manufacturer's instructions. qRT-PCR assays were performed using miScript 2 RT kit, miScript SYBR Green PCR kit, and miScript Primer Assay. The forward primers for hsa-miR1236, hsa-miR642a, and RNU6B were obtained from Qiagen. The thermal profile was as follows: 15 min at 95°C, 40 cycles of 15 s at 94°C, 30 s at 55°C, and 30 s at 70°C. Relative quantification of hsa-miR1236 and hsa-miR642a was performed according to the ⌬⌬Ct method. An aliquot of these reactions at 38 cycles was electrophoresed on an ethidium bromide-stained 2% agarose gel, and specific amplification of the PCR products was confirmed.
Kinetics of TLR4 mRNA, TLR4 Protein, and miRNAs-PBMCs were isolated from the G/G and C/C subjects (n ϭ 4 in each group) and resuspended in RPMI 1640 medium supplemented with 10% FBS. The cells were plated at a density of 1 ϫ 10 6 cells per well in a 24-well dish. THP-1 cells in logarithmic growth were resuspended in RPMI 1640 medium supplemented with 10% FBS. The cells were plated at a density of 6 ϫ 10 5 cells/well in a 24-well dish and treated with 100 ng/ml phorbol 12-myristate 13-acetate (Sigma) for 36 h to induce differentiation into macrophage-like cells. PBMCs and THP-1 cells were stimulated with 100 ng/ml Pam 3 CSK 4 and 1 g/ml ultrapure E. coli LPS, respectively. After incubation, a portion of PBMCs and THP-1cells underwent qRT-PCR analysis to determine the relative expression of TLR4 mRNA as described above. Another portion of PBMCs was stained with anti-TLR4 mouse IgG (clone HTA125) and a goat anti-mouse IgG-PE (Beckman Coulter, Fullerton, CA), washed, resuspended in 2% formaldehyde, washed again, resuspended in the solution consisting of 10% DMSO and 90% FBS, and stored at Ϫ80°C for batch analysis. On the day of analysis, cells from each subject were thawed, and surface expression levels of TLR4 of PBMCs were analyzed by flow cytometry (FACSCanto II, BD Biosciences). Data were analyzed using FlowJo software. For the miRNA analysis, PBMCs from the G/G subjects (n ϭ 3) and differentiated THP-1 cells were stimulated with 0.1-1 g/ml ultrapure E. coli LPS. After incubation, miRNA were extracted, and levels of hsa-miR1236 and hsa-miR642a were quantified as described above.
Plasmid Construction and Luciferase Assays-A 4.3-kbp genomic fragment of the human TLR4 promoter (Ϫ4121 to ϩ190) was amplified from human genomic DNA using the Expand High Fidelity PCR system (Roche Applied Science) with primers 5Ј-CTC CAT GGC ACA TTC TGC AGT AAA CTT GGA GGC-3Ј (sense) and 5Ј-CAC GCA GGA GAG GAA GGC CAT GGC TG-3Ј (antisense). The thermal cycling condition was 2 min at 94°C, 30 cycles of 30 s at 94°C, 30 s at 62°C, 4 min at 68°C, and 7 min at 68°C. The PCR fragment was inserted into the plasmid vector pCR2.1-TOPO (TOPO Cloning kit; Invitrogen) for subcloning. The 4.3-kb genomic PCR fragment of the TLR4 promoter was subcloned into the NcoI restriction site of pGL3-Basic (Promega) and sequenced (Ϫ4121LUC). Deletion of this construct (Ϫ742 to ϩ190) was generated by digestion with HindIII and subsequent religation of the remaining plasmid (Ϫ742LUC). Two types of 219-bp genomic fragments (G/G and C/C) of the 3Ј-UTR of TLR4 at SNP rs11536889 were amplified from two different human genomic DNAs with the Expand High Fidelity PCR system using the primer sets (5Ј-TGG GAT CCC TCC CCT GTA CCC TTC-3Ј (sense) and 5Ј-CTG GAT CCG TTT CTG AGG AGG CTG GAT G-3Ј (antisense)). They were digested with BamHI and inserted into the BamHI restriction site of Ϫ4121LUC (Ϫ4121LUCϩ3ЈGG and Ϫ4121LUCϩ3ЈCC) and Ϫ742LUC (Ϫ742LUCϩ3ЈGG and Ϫ742LUCϩ3ЈCC). All constructs were sequenced to verify fidelity. Exponentially growing THP-1 cells were used for transfection assays. The cells were resuspended in ␣-minimum essential medium, fetal calf serum, and penicillin and streptomycin (Invitrogen). Twenty-four hours after plating, the cells at 40ϳ60% confluence were transfected using a Lipofectamine reagent (Invitrogen). The transfection mixture included 1 g of the luciferase (LUC) construct and 2 g of pSV-␤-galactosidase control vector (Promega) as an internal control. In the miRNA inhibition assay, 375 ng of anti-hsa-miR1236, anti-hsa-miR642a, or AllStars Negative Controls (Qiagen) were co-transfected. In the miRNA transfection assay, mutant hsa-miR1236 mimics (mut-miR-1236) and mutant hsa-miR642a mimics (mut-miR-642a) that were predicted to interact with rs11536889 C allele were synthesized (Fig. 9A). The mutant miRNA mimics (37.5 ng) or AllStars Negative Controls (375 ng) were co-transfected. Two days post-transfection, the cells were deprived of serum for 12 h, and 1 g/ml E. coli LPS (L2630, E. coli O111:B4, Sigma) or 1 ng/ml human recombinant IL-6 (Wako Pure Chemical Industries, Tokyo, Japan) was added 3 h before harvesting. The luciferase activity was measured according to the supplier's protocol using a luminescence reader Accuflex Lumi 400.
miRNA Inhibition Assay-THP-1 cells were plated at a density of 8 ϫ 10 5 cells per well in a 24-well dish and treated with 100 ng/ml phorbol 12-myristate 13-acetate for 24 h. The cells were transfected with 375 ng of anti-hsa-miR1236, anti-hsa-miR642a, AllStars Hs Cell Death Control siRNA (Qiagen), or AllStars Negative Controls using a HiPerFect Transfection reagent (Qiagen) according to the manufacturer's protocol. 12 hours post-transfection, the cells were stimulated with 1 g/ml ultrapure E. coli LPS. After 24 h of incubation, the cells were analyzed for expression of cell surface TLR4 as described above under "Kinetics of TLR4 mRNA, TLR4 Protein, and miRNAs." Statistical Calculation-Comparisons between the two groups were performed by the Mann-Whitney test unless otherwise indicated. For the luciferase assay, comparisons between the two groups were performed by the unpaired Student's t test. Comparisons between multiple groups were performed by the Kruskal-Wallis H-test with the Newman-Keuls-test.

RESULTS
PBMCs were isolated from subjects with the G/G, G/C, or C/C genotype of rs11536889 and stained with PE-labeled anti-CD14 and FITC-labeled anti-TLR2, FITC-labeled anti-TLR4, or FITC-labeled isotype-matched control monoclonal antibody. Monocytes were identified based on their CD14 positivity, and TLR2 or TLR4 expression was analyzed by flow cytometry. Representative results of the flow cytometric analyses of each genotype are shown in Fig. 1A. Monocytes from C/C subjects expressed significantly higher levels of TLR4 (p Ͻ 0.01) than those from G/G and G/C subjects, whereas no significant difference was found in the staining with anti-TLR2 or control antibodies (Fig. 1B). When PBMCs were stimulated with LPS, a TLR4 ligand, the cells from C/C and G/C subjects secreted significantly higher levels of IL-8 than the cells from G/G subjects ( Fig. 2A). PBMCs from C/C subjects secreted more than 2-fold higher levels of IL-6 and TNF-␣ than the cells from the subjects with the other genotypes, but this was not statistically significant (Fig. 2, B and C). These results suggested that the level of TLR4 protein expression was associated with the response to LPS. When PBMCs were stimulated with Pam 3 CSK 4 , a TLR2 ligand, no significant difference was found in IL-8, IL-6, and TNF-␣ levels produced by cells from C/C, G/C, and G/G subjects. To determine whether these differences in TLR4 protein expression were due to the regulation of transcriptional activity by rs11536889 polymorphism, mRNA expression for TLR4 in PBMCs from subjects with each genotype were examined. However, there was no significant difference in the mRNA expression for TLR4 (Fig. 3). Then we stimulated PBMCs from the G/G and C/C subjects with Pam 3 CSK 4 and compared the kinetics of mRNA and protein expression for TLR4 (Fig. 4A). TLR4 mRNA expression in PBMCs from the G/G and C/C subjects increased 2.5 Ϯ 0.9-and 2.1 Ϯ 0.6-fold, respectively, 2 h after the stimulation. Both decreased nearly to the base-line levels 3 h after stimulation. There was no significant difference between TLR4 mRNA expression in PBMCs from the G/G subjects and that from the C/C subjects. TLR4 protein expression of PBMCs from the C/C subjects increased 2.2 Ϯ 1.6-, 1.7 Ϯ 0.5and 1.5 Ϯ 0.3-fold at 10, 20, and 30 h after stimulation, respectively. In contrast, the levels of TLR4 protein on PBMCs from the G/G subjects did not increase by 30 h after stimulation. Fold induction of TLR4 protein on PBMCs from the C/C subjects was significantly greater than that on PBMCs from the G/G subjects at 10, 20, and 30 h after stimulation (Fig. 4B).
To determine the effects of rs11536889 polymorphism on gene expression, transient transfection analyses in THP-1 cells were performed using a series of chimeric constructs containing the human TLR4 promoter ligated to a luciferase reporter and TLR4 3Ј-UTR. Luciferase activity of pGL3-Basic construct was increased ϳ1.6-fold by the insertion of TLR4 promoter Ϫ4121 to ϩ190 bp (Ϫ4121LUC) or Ϫ742 to ϩ190 bp (Ϫ742LUC) (Fig. 5). Luciferase activity of the TLR4 promoter construct was further increased by stimulation with 1 g/ml LPS or 1 ng/ml IL-6. The insertion of the 219-bp fragment of 3Ј-UTR containing rs11536889 G allele (Ϫ4121LUCϩ3ЈGG and Ϫ742LUCϩ3ЈGG) inhibited this up-regulation. The insertion of a 219-bp fragment of 3Ј-UTR containing the rs11536889 C allele (Ϫ4121LUCϩ3ЈCC and Ϫ742LUCϩ3ЈCC) did not inhibit the up-regulation.
To identify a possible causal effect of the G to C nucleotide transversion of SNP rs11536889, we analyzed the secondary structure of the mRNA (Vienna RNA Package 1.8.4). The exchange had no effect on the mRNA folding; the minimum free energies of the optimal secondary structure of G and C alleles were Ϫ1813.12 and Ϫ1813.86 kcal/mol, respectively, making a marginal difference. A prediction for miRNA targets (the microRNA data resource) identified hsa-miR-1236 and hsa-miR-642a as candidates for miRNAs that bind to the region containing rs11536889 G allele (Fig. 6A). The altered nucleotide is located within the potential binding sites of hsa-miR-1236 and hsa-miR-642a and introduces a mismatch at the 6th and 13th positions, respectively. Both hsa-miR-1236 and hsa-miR-642a were expressed in THP-1 cells (Fig. 6B). The level of hsa-miR-1236 was increased 10 h after LPS stimulation, whereas that of TLR4 mRNA was increased 2 h but decreased nearly to the base-line levels 3 h after LPS stimulation. No upregulation of hsa-miR-642a was induced by LPS stimulation (Fig. 6C). hsa-miR-1236 and hsa-miR-642a were also detected in PBMCs; however, the kinetics after LPS stimulation varied widely between the individuals (Fig. 6D). When luciferase constructs were co-transfected with the inhibitors for hsa-miR-642a and/or hsa-miR-1236, the repression of luciferase activity of Ϫ4121LUCϩ3ЈGG was reversed (Fig. 7, A and B). Importantly, the inhibitors for the miRNAs did not affect the luciferase activities of Ϫ4121LUCϩ3ЈCC, indicating that the effect of these inhibitors was specific for the rs11536889 G allele.
Because THP-1 cells express endogenous TLR4, we used these cells to examine the effects of miR-642a and miR-1236 on TLR4 expression. The cells were genotyped for rs11536889 by PCR with confronting two-pair primers and were shown to be G/G (data not shown). There was no mismatch between the TLR4 sequence of THP-1 cells and that of transient transfection constructs of the G allele. These cells were transfected with the miRNA inhibitors or control siRNA. In each transfection, transfection efficiency was confirmed using AllStars Hs Cell Death Control siRNA. When the cells were stimulated with LPS, the levels of TLR4 of the cells transfected with anti-miR-642a, anti-miR-1236, and the combination of both was statistically higher than that of the control, indicating that both miR-642a and miR-1236 contributed to the suppression of TLR4 on THP-1 cells (Fig. 8B). When the cells were left unstimulated, TLR4 expression on the cells transfected with anti-miR-642a and/or anti-miR-1236 was slightly increased, but no statistical difference was found (Fig. 8A).
As hsa-miR-1236 and hsa-miR-642a were predicted to interact only with the rs11536889 G allele, we synthesized mutant miRNAs, mut-miR-1236 and mut-miR-642a, that were predicted to interact only with the rs11536889 C allele (Fig. 9A). When luciferase constructs were transfected simultaneously with mut-miR-642a, mut-miR-1236, or the combination of both, the luciferase activity of Ϫ4121LUCϩ3ЈCC was inhibited FIGURE 1. Expression of TLR4 on the surface of monocytes from subjects classified by rs11536889 genotype. PBMCs from the G/G, G/C, and C/C subjects were stained with PE-labeled anti-CD14 and FITC-labeled anti-TLR2, FITC-labeled anti-TLR4, or FITC-labeled isotype-matched control monoclonal antibodies. CD14 ϩ cells were gated, and TLR4 or TLR2 expression was analyzed by flow cytometry. Representative results of the cells from the subjects with each genotype are shown (A). The mean fluorescence intensities are expressed with S.D. (B). *, p Ͻ 0.05. (Fig. 9, B and C). Importantly, the inhibitors for the miRNAs did not change the luciferase activity of Ϫ4121LUCϩ3ЈGG, indicating that the effect of inhibitors was specific for the rs11536889 C allele.

DISCUSSION
The peripheral blood monocytes from the C/C subjects expressed significantly higher levels of TLR4 protein on their surfaces than those from the G/G and G/C subjects in this study. When PBMCs were stimulated with Pam 3 CSK 4 , the levels of TLR4 protein on their surface significantly increased in the C/C subjects, whereas no increase was detected in the G/G subjects. This suggested that TLR4 protein expression was down-regulated in the G/G subjects or up-regulated in the C/C subjects. Induction of luciferase activity in THP-1 cells transfected with a TLR4 promoter construct was inhibited by insertion of a 219-bp fragment of 3Ј-UTR containing the rs11536889 G allele but not the C allele. Considering that no other mismatches were found in the nucleotide sequences between these two 219-bp fragments (data not shown), TLR4 protein expression seems to be down-regulated by a small region containing the rs11536889 G allele in 3Ј-UTR. One study demonstrated that the expression of TLR4 on human monocytes was not affected by LPS stimulation (28), but another study showed that LPS up-regulated the expression of TLR4 on human monocytes (29). The distinct behaviors of TLR4 in these studies might be due to differences in the donor genotype.
In contrast to the clear difference between TLR4 protein expression on the monocytes from the G/G, G/C, and C/C subjects, there was no significant difference in TLR4 mRNA expression levels in PBMCs in this study. TLR4 mRNA levels in PBMCs from both the G/G and C/C subjects were elevated 2 h after stimulation with Pam 3 CSK 4 . However, the TLR4 protein levels on PBMCs increased only in the C/C subjects but not in the G/G subjects. This evidence suggested that the genetic variation at rs11536889 affected translation rather than transcription of TLR4.  . Expression of TLR4 mRNA in PBMCs from subjects classified by rs11536889 genotype. PBMCs were isolated from the G/G, G/C, and C/C subjects, and total RNA was extracted. After reverse transcription, mRNA levels for TLR4 were determined by qRT-PCR using a specific TaqMan probe. The relative mRNA levels for TLR4 are expressed with S.D. Anti-miR-642a and anti-miR-1236 up-regulated TLR4 expression on the surface of THP-1 cells stimulated with LPS. This result suggests that TLR4 expression of the subjects was, at least in part, inhibited by miR-642a and miR-1236. These miRNA inhibitors also up-regulated the luciferase activity of Ϫ4121LUCϩ3ЈGG but did not affect the luciferase activity of Ϫ4121LUCϩ3ЈCC. Conversely, mut-miR-642a and mut-miR-1236 suppressed the luciferase activity of Ϫ4121LUCϩ3ЈCC but did not affect the luciferase activity of Ϫ4121LUCϩ3ЈGG. These results strongly suggest that binding of the miRNAs is specific to each genotype. miRNAs usually base pairs to target mRNA with imperfect complementation, resulting in translational inhibition, whereas perfect base-pairing induces target mRNA degradation (30). The imperfect complementation between miR-642a and TLR4 mRNA (mirSVR score (31): Ϫ0.3230) and between miR-1236 and TLR4 mRNA (mirSVR score: Ϫ0.2059) may lead to translational inhibition. Although both miR-642a and miR-1236 were expressed in PBMCs and THP-1 cells, the expression of these miRNAs may be different in other cell types (32,33). The inhibitory effects of these miR-NAs on TLR4 expression in different tissues remain to be elucidated.
PBMCs from the C/C and G/C subjects secreted significantly higher levels of IL-8 in response to LPS than PBMCs from the G/G subjects. This could be due to the higher expression of TLR4 in the C/C subjects. In accordance, Bihl et al. (34) showed that increased expression of TLR4 resulted in a more sensitive response to LPS. PBMCs from the C/C subjects secreted  Sequence alignments of 3Ј-UTR of human TLR4 (rs11536889 G allele), hsa-miR-1236, and hsa-miR-642a are shown. The position of SNP rs11536889 is underlined (A). Differentiated THP-1 cells were stimulated with 1 g/ml LPS. After the indicated time of incubation, the cells were harvested; total RNA and miRNA were extracted. Reverse transcription and cDNA amplification were performed as described under "Experimental Procedures." An aliquot of these reactions was electrophoresed on an ethidium bromide-stained 2% agarose gel (B). Levels of TLR4 mRNA, hsa-miR-1236, and hsa-miR-642a were measured by qRT-PCR (C). -Fold induction is expressed with S.D. *, p Ͻ 0.05 (versus the unstimulated controls). Freshly isolated PBMCs were stimulated with 100 ng/ml LPS. After the indicated time of incubation, the cells were harvested and analyzed for hsa-miR-1236 or hsa-miR-642a expression levels as described under "Experimental Procedures" (D). -Fold induction is expressed with S.D. *, p Ͻ 0.05 (versus the unstimulated controls).
slightly higher levels of IL-6 and TNF-␣ than cells from the subjects with the other genotypes; however, there was no significant difference even when the number of subjects in each group was increased from 12 to 19. It is known that myeloid differentiation factor 2 as well as CD14 is essential for the recognition of LPS (35). In addition, the intracellular signaling events after LPS binding to TLR4 could affect the magnitude of cytokine responses (5). The inter-individual variation of these molecules could affect sensitivity to LPS and may result in the ambiguous differences in IL-6 and TNF-␣ levels. Although PBMCs from the G/C subjects secreted higher levels of IL-8 than those from G/G subjects, no significant difference was seen in TLR4 expression on the surface of peripheral blood monocytes. The causal relationship between G/C genotype and TLR4 expression and function needs to be investigated further.
In our previous study the genetic variation of rs11536889 was associated with periodontitis. The C/C genotype was observed more frequently in the periodontitis group than in the control group. The up-regulation of TLR4 expression in the C/C subjects may prompt the recognition of Gram-negative bacterial LPS in periodontal pockets, accelerate the production of proinflammatory cytokines, and lead to the greater destruction of periodontal tissue. Kornman et al. (36) found that a variant in IL-1B associated with high levels of IL-1␤ production is a strong indicator of increased susceptibility to severe periodontitis; this is yet another example of how the genetic background prone to accelerated inflammation is linked to periodontitis. The genetic variation of rs11536889 is also associated with severe gastric atrophy in H. pylori seropositive Japanese (18), hepatitis type B virus recurrence after liver transplantation (19), and increased prostate cancer risk (19). The associations may be related to accelerated inflammation caused by LPS from H. pylori in gastric mucosa, hepatitis type B virus-related proteins in the liver, and Gram-negative bacteria in the prostate, respectively. A prospective study consisting of the G/G and C/C subjects to reveal a more precise predispositional effect of this SNP on these diseases remains to be performed.
To the best of our knowledge, this is the first report to reveal the biological significance of the genetic variation of rs11536889. This SNP regulates the expression of TLR4 and has some influence on the response to LPS. These findings may provide a novel approach for therapeutic interventions against diseases caused by harmful TLR4-mediated responses. FIGURE 7. Effects of microRNA inhibitors on luciferase constructs in THP-1 cells. THP-1 cells were transfected with a series of chimeric luciferase constructs in the presence or absence of miRNA inhibitors. The transfectants were left unstimulated or stimulated with 1 g/ml LPS (A) or 1 ng/ml IL-6 (B) for 3 h; luciferase activity was examined. The results of the assay obtained from three separate transfections with constructs pGL3-Basic, Ϫ4121LUC, Ϫ4121LUCϩ3ЈGG, and Ϫ4121LUCϩ3ЈCC, with or without miRNA inhibitor have been combined. The values are expressed with S.E. *, p Ͻ 0.05; **, p Ͻ 0.01 (versus the unstimulated controls).