Soluble Branched β-(1,4)Glucans from Acetobacter Species Show Strong Activities to Induce Interleukin-12 in Vitro and Inhibit T-helper 2 Cellular Response with Immunoglobulin E Production in Vivo*

An Extracellular Polysaccharide, Ac-1, Produced By Acetobacter Polysaccharogenes Is Composed Of β-(1,4)GluCan With Branches Of Glucosyl Residues. We Found That Ac-1 Showed A Strong Activity To Induce Production Of Interleukin-12 P40 And Tumor Necrosis Factor-α By MacroPhage Cell Lines In Vitro. Cellulase Treatment Completely Abolished The Activity Of Ac-1 To Induce Tumor Necrosis Factor-α Production By Macrophages, Whereas Treatment Of Ac-1 With Polymyxin B Or Proteinase Did Not Affect The Activity. Results Of Experiments Using Toll-Like Receptor (Tlr) 4-Deficient Mice And Tlr4-Transfected Human Cell Line Indicated That Tlr4 Is Involved In Pattern RecogniTion Of Ac-1. In Vivo Administration Of Ac-1 Significantly Reduced The Serum Levels Of Ovalbumin (Ova)-Specific Ige And Interleukin-4 Production By T Cells In Response To Ova In Mice Immunized With Ova. Ac-1, A Soluble Branched β-(1,4)Glucan May Be Useful In Prevention And Treatment Of Allergic Disorders With Ige Production.

Toll was first identified as a protein that controls dorsoventral pattern formation in the early stage of development of Drosophila (1) and was shown to participate in antimicrobial immune responses (2). Over the years, 10 mammalian Toll homologues, called Toll-like receptor (TLRs), 1 have been identified and shown to play important roles in the recognition of various microbial components (3). TLR4, one of the identified TLRs, has been reported to function as a receptor for lipopolysaccharide (LPS), an integral component of the outer membrane of Gram-negative bacteria (4 -7). TLR2 reportedly specializes in the recognition of lipoprotein from diverse species of bacteria, including Mycobacterium tuberculosis, Mycoplasma fermentans, Treponema pallidum, and Borrelia burgdorferi (8 -15). TLR6 in combination with TLR2 recognizes zymosan and peptidoglycan (16). TLR9 and TLR5 have been shown to recognize bacterially derived CpG DNA (17) and flagellin (18), respectively. TLR3 and TLR7 have been reported to recognize double-stranded RNA and imidazoquinolines, respectively (19,20). Thus, TLRs have been identified as ancient receptors that confer specificity to the host innate immune system allowing the recognition of pathogen-associated molecular patterns.
Allergic asthma is a chronic inflammatory disease associated with a predominant T-helper 2 (Th2) cellular response, IgE synthesis, airway infiltration by eosinophils, and bronchial hyperreactivity (34,35). Naive CD4 ϩ T cells initially stimulated with an allergen in the presence of IL-4 tend to develop into CD4 ϩ T cells that secrete IL-4, IL-5, IL-6, and IL-13 for IgE isotype switching. Th1 cells, into which naive CD4 ϩ T cells preferentially differentiate in the presence of IL-12, IL-18, and interferon-␥ (IFN-␥), secrete IFN-␥ and TNF-␣ not only for induction of cell-mediated immunity but also for inhibition of Th2 responses (36 -38). Therefore, cytokines involved in Th1biased response are thought to regulate Th2-mediated allergic response.
In the present study, we found that a bacteria cellulose AC-1 derived from Acetobacter species was a potent inducer of IL-12 p40 and TNF-␣ production by macrophages in vitro. AC-1, a ␤-glucan composed of ␤- (1,4)glucan with branches of glucosyl residues, has a molecular mass of 1 ϫ 10 6 daltons. Polymyxin B-and proteinase-treated AC-1 stimulated spleen-adherent cells to produce TNF-␣, whereas such cytokine production was not detected in cellulase-treated AC-1. Results of experiments using TLR4-deficient mice and TLR4-transfected human cell line indicated that TLR4 is involved in pattern recognition of AC-1. When oral administration of AC-1 was begun immediately after ovalbumin (OVA) immunization, significant decreases in the serum levels of OVA-specific IgE and IgG 1 accompanied by augmented IFN-␥ production occurred. These results suggest that AC-1, a potent IL-12 and TNF-␣ inducer, suppresses allergic inflammation with IgE production, thus offering an approach for the treatment of allergic disorders.

EXPERIMENTAL PROCEDURES
Animals-C3H/HeN mice, C3H/HeJ mice and BALB/c mice (SLC, Shizuoka, Japan) were used in the experiments at 6 -10 weeks of age. Mutant mice (F2 interbred from 129/Ola x C57BL/6) with a deficiency in TLR2 were generated by gene targeting and kindly provided by S. Akira (Osaka University, Japan) (4). Age-and sex-matched groups of TLR2deficient (TLR2 Ϫ/Ϫ ) mice and their littermate (TLR2 ϩ/Ϫ ) mice were used for the experiments. These mice were bred in our Institute under specific pathogen-free conditions.
Purification of AC-The polysaccharide AC series of the Acetobacter species used in this study were prepared and purified by the method reported in a previous paper (30). Briefly, five strains of polysaccharideproducing Acetobacter species were cultivated in a shaking flask at 30°C for 5 days, and the cells were removed by both centrifugation (10,000xg) of diluted broth and filtration with celite. The cell-free polysaccharides were precipitated by the addition of isopropyl alcohol, and the precipitate was dissolved in water. Then 5% aqueous cetyltrimethylammonium bromide (CTAB) solution was added until no more precipitate was formed. The insoluble acidic polysaccharide-cetyltrimethylammonium bromide complex was collected by centrifugation and redissolved in 20% sodium chloride solution. After dialysis against running water, the polysaccharide was precipitated with ethanol and dissolved in water. The acidic polysaccharide thus obtained was dialyzed against distilled water and lyophilized. We named these polysaccharides AC series as follows: AC-1, Acetobacter polysaccharogenes MT-11-2; AC-2, A. polysaccharogenes 1007; AC-3, A. polysaccharogenes 1011; AC-4, Acetobacter xylinum MH-1597; AC-5, A. xylinum 1053). The endotoxin content of AC-1 (100 g/ml) was estimated by using an endotoxin-specific chromogenic Limulus test (Wako) to be 32 pg/ml. This level of endotoxin does not affect TNF-␣ production in a culture supernatant of peritoneal adherent cells of C3H/HeN mice. In some experiments, AC-1 (100 g/ml) was treated with 10 g/ml of polymyxin B (Wako) at 37°C for 2 h for deactivation of LPS activity or treated with actinases E (Kakenshouyaku) 37°C for 24 h and then DEAE anion chromatography for removal of protein. AC-1 (5 mg) was incubated with 1 mg of ␤-(1,4)glucanase (Wako) in 0.05 M acetate buffer (2.5 ml) at pH 5.0 and 37°C for 24 h. LPS was treated in the same manner.
Analysis with High-performance Anion-exchange Chromatography Coupled with a Pulsed Amperometric Detection-The polysaccharides from Acetobacter species were subjected to acid hydrolysis using trifluoroacetic acid and high-pH anion-exchange chromatography with pulsed amperometric detection in an effort to identify optimum hydrolysis conditions for composition analysis of their carbohydrate components. The 10-mg sample was mixed with a final concentration 2 N trifluoroacetic acid containing 10 g/ml fucose as an internal standard in a glass tube sealed with a screw cap. After 18 h at 100°C, solutions were dried under a stream of nitrogen; the residues were then redissolved in water and the hydrolyzate solutions transferred to autosampler vials. The hydrolyzates were stable in water at Ϫ20°C for at least 1 month. Samples were analyzed by high-performance anionexchange chromatography on a Dionex 500 system (Dionex Corp.) supplied with a pulsed amperometric detector. The system was equipped with a CarboPac column (packed silica appropriate for mono-, di-, tri-, and oligosaccharide analysis). Sodium hydroxide solution (NaOH, 250 mM in water) was used as the eluant. Analyses were performed in the isocratic mode (% water/% NaOH ϭ 48:52). Flow rate was set to 0.6 ml/min. To minimize carbonate formation in the system, which leads to a dramatic reduction of the retention times, a small amount of Ba(CH 3 COO) 2 (4 mM) was added to the alkaline eluant. Data were collected and analyzed on computers equipped with the Dionex Peak-Net software.
Cell Preparation-All cell lines were grown in tissue culture flasks at 37°C in 5% CO 2 , 95% air and passaged every 2 or 3 days to maintain logarithmic growth. Two mouse macrophage cell lines, J744.1 and RAW264.7, and a human embryonic kidney cell line, HEK 293 (human embryonic kidney 293), were obtained from the Institute of Physical and Chemical Research Cell Bank (Tsukuba, Japan) and maintained in Dulbecco's minimum essential medium with 10% fetal bovine serum (JRH Bioscience). Adherent cells from the peritoneal cavity of C3H/ HeN, C3H/HeJ, TLR2 ϩ/Ϫ , or TLR2 Ϫ/Ϫ mice were used as a source of mouse macrophages. Briefly, peritoneal exudate cells (PEC) were suspended in RPMI 1640 containing 10% fetal bovine serum were cultured on plastic plates for 2 h at 37°C. After nonadherent cells had been removed, fresh complete medium was added to the adherent cells with or without AC-1 or control reagents in the presence of IFN-␥ (30 units/ml).
Plasmids-Murine TLR tagged with p3xFLAG at the carboxyl terminus was generated by PCR and ligated into the expression plasmid CMV14 (Sigma). All TLR plasmids used in transfections were purified using an Endo-free plasmid kit (Qiagen). A mouse CD14 expression plasmid, pcDNA3.1(ϩ)-mCD14, and a NF-B-responsive reporter, pGL3-NFB-Luc, have been described previously (39). A carboxyl-terminal FLAG-tagged each TLRs expression plasmid and mouse MD2 expression plasmid were generated by inserting the whole mTLR and MD-2 coding region cDNA into the p3xFLAG-CMV14 vector.
Luciferase Assay-HEK 293 cells were seeded in 6-well plates at a density of 1 ϫ 10 6 cells/well 1 day before transfection. HEK 293 cells were transiently transfected with 0.1 g of pGL3-NF-B/Luc (a luciferase reporter construct containing the consensus NF-B binding sequence), 0.4 g of pSV-␤-galactosidase as an internal control (Promega, Madison, WI), and either 1.0 g of mTLRs/FLAG using LipofectAMINE TM (Invitrogen). At 48 h after transfection, the HEK 293 cells were stimulated with AC-1 for 8 h. The cells were harvested, washed, and lysed in 200 l of lysis buffer, and the luciferase activity was measured using a luminometer (Wallac 1420, PerkinElmer Life Sciences) with a Dual-Luciferase Reporter Assay System (Toyo Ink, Tokyo, Japan) or Steady-Glo Luciferase Reporter Assay System (Promega) according to the manufacturer's instructions. Background luciferase activity was subtracted.
Cytokine Measurement-J744.1 or RAW264.7 cells (5 ϫ 10 5 cells/ml) were stimulated with the AC series, and the levels of IL-12 p40 and TNF-␣ production were determined by using an enzyme-linked immunosorbent assay (ELISA) 24 h after incubation. Adherent cells of PEC from C3H/HeN and C3H/HeJ or TLR2 ϩ/Ϫ and TLR2 Ϫ/Ϫ mice were incubated for 24 h with AC-1, LPS (Escherichia coli O55, Sigma), synthetic lipoprotein (palmitoyl-Cys(R,S)-2,3-di(plamitoyloxy)-propyl-Ala-Gly-OH, Bachem AG, Bubendorf, Switzerland), or synthetic lipid A (Ono-4007, Osaka Japan) in the presence or absence of IFN-␥, and the culture supernatant was then collected. IL-12 p40 and TNF-␣ levels in the culture supernatant were determined by ELISA. ELISAs for IL-12 p40 and TNF-␣ were performed using commercially available kits from Genzyme (Cambridge, MA). Immunization and Challenge-Mice were immunized intraperitoneally with 50 g of chicken OVA (Gradee, Sigma) absorbed in 100 l of the alum on day 0. Twelve days later, the mice were injected intraperitoneally with 50 g of OVA in 100 l of alum. The mice were challenged intragastrically with a suspension of 100 l of AC-1 (10 mg/ml) or PBS on days 1, 4, 9, and 14. Spleens and sera were obtained on day 29 and stored at Ϫ20°C until analysis.
Measurement of Cytokine Production of Spleen Cells-Spleen cells were incubated on a nylon wool column at 37°C in 5% CO 2 for 60 min. The cell population eluted from the column contained Ͼ90% T cells as determined by flow cytometric analysis with anti-CD3⑀ monoclonal antibody. T cells (5 ϫ 10 5 ) and mitomycin C-treated naive spleen cells (5 ϫ 10 5 ) were cultured in 96-well cell culture plates (Falcon, BD Biosciences) with 200 g of OVA. After 48 h of culture, the cultured supernatants were collected, and the amounts of secreted IL-4 and IFN-␥ in the supernatants were determined by ELISA. Commercial ELISA kits were used to measure the levels of IL-4 and IFN-␥ (Genzyme Diagnostics) according to the manufacturer's instructions.
Statistical Analysis-The statistical significance of the data was determined by Student's t test. A p value of less than 0.05 was taken as significant.

Structure of the Polysaccharide AC Series-Five strains of
Acetobacter species that produce a new type of extracellular soluble polysaccharide were isolated, and the composition of their carbohydrate components was analyzed using high-performance anion-exchange chromatography with a pulsed amperometric detector. A representative polysaccharide, AC-1, isolated from the culture filtrate of A. polysaccharogenes MT-11-2(1005) is composed of D-glucose, D-galactose, D-mannose, and D-glucuronic acid in the molar ratio of 3.0:1.0:1.1:1.5. The composition of the carbohydrate components in each sample is shown in Table I. NMR spectroscopy indicated that the dominant D-glycosidic linkages must be in ␤-configuration (30). To obtain information on the mode of glycosidic linkages, both native and carboxyl-reduced polysaccharides were methylated, and the partially methylated sugars in the acid hydrolysate were analyzed by gas liquid chromatography. The identities and proportions of the cleavage fragments of the native and carboxyl-reduced polysaccharides indicate that polysaccharide AC-1 has a highly branched structure with a repeating unit of 11 sugar residues. It contains a backbone chain of ␤-(1,4)linked D-glucose, residues and two of the four D-glucose residues are branched at the O-3 position. There are two kinds of side chains; one is terminated with D-glucose residues and the other with D-glucuronic acid residues, as indicated by the increase in approximately 1 mol of tetra-O-methyl-D-glucose in the methylated, carboxyl-reduced polysaccharide. In addition, the polysaccharide contains (1,6)-linked D-glucose, (1,6)-linked D-galactose, and (1,2)-linked D-mannose residues. They are most probably located in the side chains, as revealed by the fragmentation analysis (31). Similar analysis was performed on AC-4 derived from A. xylinum MH-1597. The structural feature of polysaccharide AC-4 resembles that of polysaccharide AC-1 except for the sugar arrangement in the side chains. Thus, AC is a ␤-glucan composed of ␤-(1,4)glucan with branches of glucosyl residues (Fig. 1).
Bacterial Cellulose Derived from Acetobacter Species Stimulated Mouse Macrophages to Produce IL-12 p40 and TNF-␣-We screened soluble ␤-(1,4)glucans derived from Acetobacter species by measuring the levels of cytokine production in the mouse macrophage cell lines J774.1 and RAW264.7. Among the various preparations, AC-1, -2, and -3, all of which are derived from A. polysaccharogenes, stimulated J774.1 to produce IL-12 p40 ( Fig. 2A). AC-1, derived from Acetobacter polysaccharogenes MT-11-2, at a final concentration of 100 g/ml induced the maximal level of TNF-␣ production in both mouse macrophage cell lines among the preparations (Fig. 2, B and C). RAW264.7 cells did not produce IL-12 p40 in response to AC-1 or LPS (data not shown). AC-1 was used in the following experiments.
Next, an experiment was carried out to determine whether AC-1 induces production of IL-12 p40 and TNF-␣ by primary culture of macrophages from naive mice. As shown in Fig. 3A and B, AC-1 induced significantly high levels of IL-12 p40 and TNF-␣ production by peritoneal adherent cells from BALB/c mice. To further purify the polysaccharides from AC-1, we treated AC-1 with protease and subjected them to ion exchange chromatography. Protease-treated products of AC-1 exhibited almost the same level of activity to induce TNF-␣ production as that of nontreated AC-1 (Fig. 3, A and B). Thus, the protein moiety of AC-1 is not required for the activity of AC-1. Because Acetobacter species are Gram-negative bacteria containing LPS, we examined the effect of treatment of AC-1 with polymyxin B, which neutralizes LPS activities, on the production of TNF-␣ by macrophages. As shown in Fig. 3C, polymyxin B inhibited the activity of LPS (100 ng/ml) but not the activity of AC-1. Thus, the possibility that induction of the activity of AC-1 is due to contamination by LPS is ruled out.
Backbone Chain of ␤-(1,4)Glucan Is Important for AC-1 Activity-As shown in Fig. 1, AC-1 has a branched structure containing a backbone chain of ␤-(1,4)-linked D-glucose, two of every four glucose residues being substituted at the O-3 positions to form two kinds of branches. To determine the involvement of ␤-(1,4)glucose linkage of AC-1 in stimulation of macrophages, AC-1 was treated with ␤-(1,4)endoglucanase, which degrades polysaccharides possessing ␤-(1,4)glucan backbones. As shown in Fig. 4A, the ability of AC-1 to induce production of TNF-␣ by macrophages was abolished by cellulase treatment, whereas the ability of LPS to induce production of TNF-␣ by macrophages was not affected by the same treatment. We further examined the effect of laminarin, a ␤-(1,3)glucan antagonist, which partially inhibits binding of ␤-(1,3)glucan to receptors (40), on the activity of AC-1. When RAW264.7 cells were stimulated with AC-1 in the presence of laminarin, TNF-␣ production was not inhibited by any dose of laminarin (Fig. 4B). These results suggest that ␤-(1,4)glucan is required for the stimulatory activity of AC-1. TLR4 Is Involved in TNF-␣ Production by PEC-adherent Cells Stimulated with AC-1-TLRs have been identified as ancient receptors that confer specificity to the host innate immune system, enabling the recognition of pathogen-associated molecular patterns. We examined the involvement of TLR in the recognition of AC-1 by using TLR-deficient mice. The peritoneal adherent cells from C3H/HeN mice with normal TLR4 (TLR4 ϩ/ϩ ) were cultured with AC-1, and TNF-␣ levels in the supernatant were determined by ELISA. The results indicated that the adherent cells produced TNF-␣ in response to in vitro stimulation with AC-1 in a dose-dependent manner ( Fig.  5A and B). We then examined TNF-␣ production by the peritoneal adherent cells of C3H/HeJ mice with mutated TLR4 (TLR4 Ϫ/Ϫ ) to determine whether TLR4 is involved in TNF-␣ production induced by AC-1. As shown in Fig. 5A, the cells from TLR4 Ϫ/Ϫ mice responded to lipoprotein, a ligand for TLR2, but did not respond to synthetic lipid A or AC-1 to produce a high level of TNF-␣, indicating that TLR4 is involved in AC-1-mediated TNF-␣ production by the adherent cells. We next examined the production of TNF-␣ in the supernatants of the peritoneal adherent cells from TLR2 ϩ/Ϫ or TLR2 Ϫ/Ϫ mice cocultured with AC-1. As shown in Fig. 5B, the cells from TLR2 Ϫ/Ϫ mice did not respond to lipoprotein to produce TNF-␣, whereas those from TLR2 Ϫ/Ϫ mice produced TNF-␣ in response to AC-1 as well as lipid A, although to a lesser degree compared with those from TLR2 ϩ/Ϫ mice. These data suggest that TLR2 is not involved mainly in AC-1-mediated cellular activation.
Expression of Mouse TLR4 Conferred AC-1-mediated NF-B Activation in a Human Cell Line-To confirm the involvement of TLR in AC-1-mediated cellular activation, we inserted the coding region of mTLR cDNA into a mammalian expression plasmid, CMV14, with a carboxyl-terminal FLAG tag. The FLAG-tagged mTLR was transiently expressed in a HEK 293 cell line. The expression of the FLAG-tagged mTLR was confirmed by means of Western blotting. Cells were transiently transfected with 0.1 g of pGL3-NF-B/Luc. Forty-eight hours after the transfection, cells were stimulated with AC-1, and 8 h later cells were lysed, and the luciferase activity was measured. As shown in Fig. 6A, transfection of the control vector alone did not mediate the induction of AC-1 NF-B activity, indicating that the parental HEK 293 cells were hyporesponsive to AC-1. When mTLR4 was expressed, AC-1 significantly induced NF-B activation, whereas the cells expressing other TLRs including 1, 2, 6, 9, 1 ϩ 2, or 2 ϩ 6 did not respond to AC-1. Coexpression of TLR2 with TLR6 activates NF-B when stimulated not with AC-1 but with zymosan (Fig. 6B). Polymyxin B-treated AC-1 also induced NF-B activation in TLR4-transfected HEK 293 cells, again confirming that contamination of LPS is not involved in the AC-1 activities (Fig. 6C).
Effects of Oral Administration of AC-1 on Th2 Response with IgE Production to OVA-To investigate the role of AC-1 in the modulation of immune response in vivo, we examined the effect of oral administration on antibody production in mice immunized with OVA. BALB/c mice were sensitized intraperitoneally with OVA/alum on day 0 and day 12. Mice were orally administered AC-1 or PBS once every 5 days from day Ϫ1 to day 14 (Fig. 7A). The serum level of OVA-specific IgE, IgG 1 , or IgG 2a on day 29 after immunization with OVA/alum was measured. The AC-1-treated mice had lower OVA-specific IgE and IgG 1 levels in serum than did PBS-treated mice after OVA sensitization. On the other hand, the level of OVA-specific IgG 2a was significantly higher in AC-1-treated mice than in PBS-treated mice sensitized with OVA (Fig. 7B).
We next separated T cells from the spleens of mice treated with AC-1 or PBS on day 29 after OVA immunization. As shown in Fig. 7C, CD3 ϩ T cells from PBS-treated mice sensitized with OVA produced IL-4 in response to OVA, whereas the level of IL-4 production was significantly reduced in AC-1treated mice sensitized with OVA. The level of IFN-␥ production in response to OVA was increased in the culture supernatant of spleen T cells from AC-1-treated mice sensitized with OVA. Thus, Th2 responses were suppressed, but Th1 responses were augmented by oral administration of AC-1 in these mice. DISCUSSION We have identified an active substance derived from A. polysaccharogenes that shows activity to induce production of TNF-␣ and IL-12 p40 by macrophages. AC-1 is a soluble ␤-glucan composed of ␤-(1,4)glucan with branches of glucosyl residues and has a molecular mass of 1 ϫ 10 6 daltons. Our results suggest that the ␤-(1,4)glucan backbone of AC-1 is important for the activation of macrophages. The production of TNF-␣ by peritoneal adherent cells from TLR4-deficient mice in response to AC-1 was significantly impaired, and the biological response of HEK 293 cells to AC-1 was reconstituted by transfection with murine TLR4. These results suggest that TLR4 is at least partly involved in the cellular response to AC-1 containing soluble ␤-(1,4)glucan with branches of glucosyl residues produced by Acetobacter species.
Toll was first identified as a protein that controls dorsoventral pattern formation in the early stage of development of Drosophila (1) and was shown to participate in antimicrobial immune responses (2). Recently, several mammalian Toll homologues have been identified and have been shown to play important roles in the recognition of various bacterial components (3). Among them, TLR4 has been shown to recognize LPS derived from Gram-negative bacteria, whereas TLR2 recognizes peptidoglycan and lipopeptide derived from Gram-positive bacteria and zymosan containing ␤-(1,3)glucan produced by yeast (4 -10,16). TLR9 has recently been shown to play a critical role in the recognition of bacteria-derived CpG DNA (17). Although the functions of other TLRs are not known, it has been shown that TLR6 or TLR1 can form a signaling complex with TLR2 (41). The results of the present study showed that production of TNF-␣ in response to AC-1 normally occurs in spleen adherent cells from TLR2-deficient mice. Furthermore, HEK 293 cells transfected with TLR2 alone or TLR2 together with TLR1 or TLR6 did not respond to AC-1. The activity of AC-1 was not affected by treatment with protease, thus ruling out the possibility of contamination by the biologically active protein moiety in AC-1. Furthermore, the cellular response to AC-1 is not due to LPS contamination because polymyxin B did not inhibit the activity of AC-1 but did inhibit the activity of LPS in a dose-dependent manner. Taken together, TLR4 is a receptor for pattern recognition of AC-1.

FIG. 6. Reconstitution of responses of HEK 293 cells to AC-1 after transfection with murine TLR4.
A, HEK 293 cells were cotransfected with NF-B-luciferase plasmids and ␤-galactosidase reporter plasmids together with the indicated TLR expression constructs; 48 h later transfectants were treated with AC-1 (100 g/ml) for an additional 8 h, and luciferase activity was measured. B, HEK 293 cells were co-transfected with NF-B-luciferase and ␤-galactosidase reporter plasmids together with TLR2 alone or with a complex of TLR2 and TLR6 expression constructs and stimulated with AC-1, zymosan, or lipoprotein (L.P.), and luciferase activity was measured. C, TLR4 transfectants were stimulated with AC-1 or LPS pretreated with polymyxin B (10 g/ml), and luciferase activity was measured. Three independent experiments showed similar results; data are shown from a representative experiment. *, p Ͻ 0.05. addition to ␤-(1,3)glucans. In the present study, HEK 293 cells co-transfected with TLR2 together with TLR6 cDNAs did not respond to AC-1. It has recently been reported that curdlan, a linear ␤-(1,3)glucan, stimulated the binding of macrophage to pattern-recognition receptors using MyD88 for its signal transduction other than TLR-2 and TLR-4 (29). We found that laminarin, a ␤-(1,3)glucan antagonist that inhibits the binding of ␤-(1,3)glucan to receptors, did not inhibit AC-1 activity, suggesting that pattern recognition receptors for ␤-(1,3)glucan are not involved in the stimulatory activity of AC-1. AC-1 is a ␤-glucan composed of ␤-(1,4)glucan with two branches of glucosyl residues O-␤-D-Glc-(1-6)-␤-D-Glc-(1-6)-␤-D-Glc-(1-6)-D-Gal-(1-6)-D-Gal and D-mannose and D-glucuronic. At present, the portion of AC-1 that is necessary for recognition by TLR4 is currently unknown. When AC-1 was treated with a ␤-(1,4)en-doglucanase, which degrades polysaccharides possessing ␤-(1,4)glucan backbones, the ability of AC-1 to induce production of TNF-␣ by macrophages was completely inhibited, whereas the same treatment did not affect the ability of LPS to activate macrophages. These results suggest that the ␤-(1,4)glucan backbone is essential for the stimulatory activity of AC-1.
It is notable that AC-1 modulates immune responses not only in vitro but also in vivo. In the present study, we found that oral administration of AC-1 to mice significantly inhibited the production of IgE and IgG 1 , whereas it augmented IgG 2a production following OVA immunization. The inhibition of IgE and IgG 1 production is mainly due to insufficient induction of Th2 response, whereas the augmentation of IgG 2a production is because of increased Th1 response producing IFN-␥. Naive CD4 ϩ T cells initially stimulated in the presence of IL-12 tend to develop into CD4 ϩ Th1 cells that produce IFN-␥, which are not only for induction of cell-mediated immunity against intracellular parasites and tumor but are also for inhibition of Th2 responses causing allergic diseases (36 -38). Therefore, AC-1induced cytokines involved in Th1-biased response are thought to regulate Th2-mediated allergic response. Induction of IL-12 production by macrophages/dendritic cells via TLR4 signaling may be a major mechanism by which AC-1 promotes Th1biased response to OVA. AC-1 derived from Acetobacter species may have prophylactic and therapeutic applications not only for controlling allergic diseases with dominant Th2 responses but also for preventing tumor development and infection of which protective mechanisms mainly depend on Th1 responses.
In conclusion, we have found that soluble branched ␤-(1,4)glucans from A. polysaccharogenes show strong activities to induce the production of IL-12 p40 and TNF-␣ in vitro and inhibit Th2 response with IgE production in vivo. TLR4 is required for cellular responses to AC-1. AC-1 may be useful for the prevention of allergy responses.