Mannan derivatives induce phenotypic and functional maturation of mouse dendritic cells

Abstract

Mannan, a polysaccharide isolated from yeast binds to C-type lectins of the mannose receptor family, expressed by antigen-presenting cells (APCs) including dendritic cells (DCs) and macrophages. As these receptors mediate endocytosis, they have been targeted with ligands to deliver antigens into APCs to initiate immune responses. Immunization with tumour antigen MUC1 conjugated to oxidized mannan (OM) or reduced mannan (RM) induced differential immune responses in mice, and only mice immunized with OM-MUC1 elicited strong MUC1-specific cytotoxic T lymphocyte responses and protected mice from a MUC1 tumour challenge. In this study, the adjuvant effect of mannan and its derivatives including OM and RM, in comparison to lipopolysaccharide, on DCs were investigated. Mannan, OM and RM were capable of stimulating mouse bone marrow-derived DC in vitro, eliciting enhanced allogeneic T-cell proliferation and enhancing OTI/OTII peptide-specific T-cell responses. Injection of mice with mannan, OM and RM induced a mature phenotype of lymph node and splenic DCs. Analysis by reverse transcription—polymerase chain reaction indicated that Manna, OM and RM also stimulated up-regulation of inflammatory cytokines including interleukin-1β and tumour necrosis factor-α, and differential T helper 1 (Th1)/Th2 cytokines. Subsequent experiments demonstrated that activation of DCs was Toll-like receptor-4-dependent. The data presented here, together with evidence reported previously on OM and RM in induction of immune responses in vivo, suggest that OM and RM exert a dual capacity to target antigen to APCs as well as mature DCs.

Introduction

Mannan is a polymannose recognized by C-type lectin receptors, such as the mannose receptor (MR) (CD206) and DC-SIGN (CD209). Because of their presence on antigen-presenting cells (APCs) including dendritic cells (DCs) and macrophages, these receptors have been characterized for their uptake of mannan, fucose or glucose-containing compounds. Binding of the MR to mannan, for instance, induces endocytosis, followed by its delivery into endosomal pathway.¹ Early studies on mannosylated antigens indicated that the presence of mannose residues on antigens greatly enhanced antigen-uptake and major histocompatibility complex (MHC) class II-restricted antigen presentation by DCs.^2–4 Conjugation of mannan to antigens also enhances its uptake and presentation and modification of mannan can further divert the antigen presentation pathway to elicit differential immune responses specific to antigens.^5–27 In our previous studies, mannan, in an oxidized or reduced form conjugated to the tumour antigen mucin1 (MUC1) was shown to induce MUC1-specific T helper 1 (Th1) and Th2 responses, respectively, in immunized mice.^9,19,21 The Th1/cytotoxic T lymphocyte (CTL) responses induced by immunization with oxidized mannan (OM) subsequently lead to rejection of MUC1 expressing tumours, while the Th2/antibody responses induced by reduced mannan (RM) showed no protection.^6,11,12 As a result of the encouraging preclinical studies, MUC1 fusion protein conjugated to OM (MFP) has been used in clinical trials and is now in phase-II clinical studies.^28,29

DCs, the most powerful APCs, are crucially involved in immune induction. A mature phenotype of DCs which involves up-regulation of costimulatory molecules (e.g. CD40, CD80 and CD86) and pro-inflammatory cytokines (e.g. interleukin (IL)-12, IL-4) is required for optimal Th1 or Th2 cell priming. Although antigen-specific immune responses can be largely induced by codelivery with OM and RM because of enhanced antigen uptake mediated primarily by the MR,^6,10 the adjuvant effect of mannan on DCs is largely unresolved. First, it is not known whether mannan alone induces maturation and/or cytokine production of DCs. Mannan is a polysaccharide present on mycobacterial and yeast cell walls. In view of its cytokine function, mannan derived from Saccharomyces cerevisiae and Candida albicans at 10–100 µg/ml has been shown to induce production of tumour necrosis factor-α (TNF-α) by human monocytes in a CD14- and Toll-like receptor-4 (TLR4)-dependent manner.³⁰ Notably, TLR4 recognizes a variety of pathogen-associated molecular patterns (PAMPs) and is present in monocytes and various subsets of DCs. TLR4 ligands, such as lipolysaccharide (LPS), induce a mature phenotype in DCs.³¹ Similarly, lipomannan, a mannan derivative from mycobacteria, can also induce maturation and inflammatory responses in macrophages through TLR2 signalling and the mannan-core is required for this activity.^32,33 Although its corresponding receptor is not known, acemannan purified from Aloe vera induces phenotypic maturation of immature bone-marrow-derived dendritic cells (BMDCs) with up-regulation of costimulatory molecules including CD40, CD80 and CD86. Functional analysis indicates those DCs stimulate mixed lymphocyte reaction (MLR) and IL-12 production.³⁴ Second, it is not certain whether DC maturation is induced by ligation of C-type lectins with mannan. Cross-linking of MR with the monoclonal antibody PAM-1 has been shown to induce maturation and production of Th2 and regulatory T-cell promoting chemokines in DCs, which in turn induce T-cell anergy. This immunosuppressive effect is also seen with lipoarabinomannan.³⁵ The binding of mannan to the C-type lectin receptors including MR, however, has not been evaluated in this manner. Third, whether mannan binds to any potential receptor other than C-type lectins present on DCs is still in question. TLR4 recognizes mannan and mannan-associated PAMPs, but this has not been shown at the DC level.³⁰ Finally, as OM and RM induce differential immune responses in immunized mice,³⁶ whether in vitro they induce maturation and preferential cytokine responses in DCs in parallel with mannan is not known. Previously, we noted that in peritoneal exudate macrophages, oxidized MFP was shown to stimulate production of IL-12 in a level significantly higher than reduced MFP, although both oxidized and reduced MFPs induce TNF-α and IL-10 production.⁶ Similarly, in bovine monocyte-derived macrophages, peptide conjugated with OM stimulated considerably higher IL-12 production than peptide conjugated to RM.¹⁸

Herein, we demonstrate the adjuvanticity of mannan, OM and RM on in vitro BMDCs, as well as in vivo splenic and lymph node DCs. Phenotypic maturation of DCs was determined by elevated surface expression of costimulatory molecules including CD40, CD80 and CD86. Functional maturation of DCs was confirmed by allogeneic MLR and peptide-specific T-cell assays. Furthermore, we demonstrate the maturation of BMDCs caused by mannan is TLR4 dependent. Finally, patterns of cytokine mRNA expression of BMDCs appear variable in response to mannan, OM, RM stimulation.

Materials and methods

Animals

C57BL/6, C3H/HeN, C3H/HeJ, BALB/c, OTI and OTII mice, aged 6–10 weeks, used throughout this study were purchased from Walter and Eliza Hall Institute (Melbourne, Australia). C57BL/6 and C3H/HeN mice were used as wild type mice. C3H/HeJ mice were donors of BMDCs unresponsive to TLR4 stimulation because of an inactivating point mutation in Lpsd/Ran gene.³⁷ BALB/c, OTI and OTII mice were donors of allogeneic and peptide-specific T cells.

Preparation of stimulants

Mannan derived from S. cerevisiae was purchased from Sigma (St. Louis, MO). Mannan in its oxidized or reduced form was prepared as previously described.¹² To caution against endotoxin contamination, sterile buffers were used throughout preparation. Briefly, mannan was oxidized into a poly aldehyde compound by treating 14 mg of mannan in 1 ml of 0·1 m phosphate buffer (pH 6·0) with 100 µl of 0·1 m sodium periodate for 1 hr at 4°. Following a 30 min incubation with 10 µl of ethanediol at 4° to quench excessive periodate, the mixture was passed through a PD-10 column (Sephadex G-25M column, Pharmacia Biotech, Upsaala, Sweden) equilibrated in 0·1 m bicarbonate buffer (pH 9·0) and the OM fraction was collected in 2 ml. Addition of 1 mg sodium borohydride into OM solution reduces aldehydes to amines and alcohols, resulting in formation of RM. Mannan in its original form was also prepared in pH 9·0 bicarbonate buffer. LPS derived from Escherichia coli (L3137, Sigma, Castle Hill, Australia) and poly(I:C) (Sigma) were prepared in phosphate-buffered saline (mPBS). In this study, pH 9·0 buffer (referred to as pH 9) was used as the background control to stimulants (mannan, OM and RM) and LPS or poly(I:C) as the positive control.

Generation and stimulation of BMDCs

Bone marrow cells from femurs and tibias of C57BL/6, C3H/HeN, C3H/HeJ mice were collected and treated with ACK lysis buffer (0·15 m NH₄Cl, 1 mm KHCO₃ and 0·1 mm disodium ethylenediaminetetra-acetic acid) to lyse erythrocytes. Cells were washed and cultured at 2 × 10⁶ cells/3 ml in six-well plates with complete RPMI media (2% HEPES buffer, 0·1 mm 2-mercaptoethanol, 100 U/ml penicillin, 100 µg/ml streptomycin, 2 mm glutamine and 10% fetal calf serum) supplemented with 10 ng/ml granulocyte—macrophage colony-stimulating factor (GM-CSF) and IL-4 (BD-PharMingen, San Diego, CA). By this culture method, the purity of BMDCs was consistently >80% CD11c⁺. At day 6, stimulants (pH 9, mannan, OM, RM and LPS) were added into DC cultures at specified doses and cells were harvested at specific time-points as mentioned below.

T-cell purification

Splenocytes were collected through cell strainers, washed and incubated in ACK lysis buffer at 37° for 5 min to lyse erythrocytes. Splenocytes were washed, counted and incubated with antibody cocktail containing rat anti-mouse Gr-1 (RB6-8C5), anti-B220 (RA3-6B2), anti-CD11b (M1/70.15), anti-erythrocyte (TER-119) and anti-MHC-class II (M5/114) at 4° for 30 min. Cells were washed and unwanted cells were depleted with two rounds of bead separation. In each round, cells were incubated with BioMag goat anti-rat magnetic beads (Qiagen, Hilden, Germany) (eight beads per cell) at 4° for 25 min. Cells bound to the beads were removed by magnetic attraction. The purity of T cells was >94%.

T-cell proliferation

To set up allogeneic MLR, stimulated non-adherent BMDCs were washed, irradiated (3000 cGy), and added as stimulatory cells (0, 625 and 1250 cells) into 96-well U-bottom plates containing 5 × 10⁴ allogeneic BALB/c T cells in quadruplets with a final volume of 200 µl per well. Following 96 hr of coculture, cells were pulsed with 1 µCi of ³H-thymidine (Amersham, Amersham, UK) each well for the final 16 hr of culture. Thereafter, cells were harvested and radioactivity measured by the Packard TopCount scintillation counter (PerkinElmer, Boston, MA) in counts per minute (c.p.m.). To evaluate peptide-specific T-cell proliferative responses, stimulated BMDCs were pulsed with titrated H-2K^b restricted SIINFEKL peptide (0·000001–0·1 µg/ml) or I-A^b restricted ovalbumin (OVA)_323−339 ISQAVHAAHAEINEAGR peptide (0·001–10 µg/ml) for 3 hr. Irradiated BMDCs were seeded at 2 × 10³ cells into the 96-well plates containing 2 × 10⁴ OTI or OTII T cells in triplicates with a final volume of 200 µl per well. Control wells, which contained T cells alone or T cells with stimulated DCs in the absence of peptide, were invariably included in all experiments performed. In these wells, T cells did not proliferate. Proliferation of T cells was monitored between day 2 and day 5 by ³H-thymidine uptake as described above. Peak proliferation of OTI on day 2 and OTII on day 3 is shown.

Injection of mice

To evaluate local and systemic DC activation, each stimulant was injected with 26 gauge needles intradermally into rear footpads or intravenously at bases of tails, respectively, at specified doses into four mice. After 18 hr, popliteal lymph nodes (LNs) were collected for analysis of the local DCs, while spleens were collected for the systemic DC analysis.

Isolation of lymph DCs (LNDCs) and splenic DC-enriched cells

DC were isolated from draining LNs as previously described.³⁸ LNs and spleens isolated from two to four mice were digested with collagenase D (Roche, Basel, Switzerland) and DNase I (Boehringer, Mannheim, Germany) at room temperature for 20 min and then treated with 0·1 m EDTA (pH 7·2) to dissociate T cells from DCs. LN cells were counted and incubated with CD11c⁺ MAC beads (10⁶ cells/1 µl beads) at 8–12° for 25 min. After being washed, they were resuspended in 4° mPBS containing 0·2 mm EDTA. These cells were then passed through autoMACS columns (Miltenyi Biotech, Germany) following manufacturer's instructions, and CD11c⁺ DCs (>95% purity) were collected. Washed splenic cells were treated with the following procedures at 0–4° in Ca²⁺- and Mg²⁺-free media in order to obtain the DC-enriched fraction. In brief, these cells were resuspended in the premade 1·077 g/cm³ isosmotic Nycodenz (Nycomed Pharma AS, Oslo, Norway) medium and centrifuged at 1700 g for 15 min with low or no break. The DC-enriched low-density leucocytes (3–5% of total cells) were collected and resuspended in divalent metal free balanced salt solution (BSS) that contains EDTA.

Phenotypic analysis of DCs

To evaluate phenotypic maturation of BMDCs and splenic DCs, 2–5 × 10⁵ cells were stained with anti-mouse CD11c⁺ conjugated to APC (Pharmingen) in combination with in-house rat anti-mouse CD40, CD80, CD86 or MHC-class II conjugated to fluoroscein isothiocyanate (FITC). The isotype controls, rat anti-mouse immunoglobulin G2a (IgG2a) for APC and FITC, were included. The DC population was gated based on negative propidium iodide staining and positive expression of CD11c. Their activation status was determined by mean fluorescence intensity (MFI) of FITC. To assess the phenotype of LNDC subpopulations, 1–7 × 10⁴ LNDCs stained with anti-CD11c⁺ APC, anti-CD8-phycoerythrin, anti-DEC205-PerCP (Pharmingen) in combination of each activation marker conjugated with FITC. DCs were first gated on the basis of CD11c^high expression to exclude CD11c^low plasmacytoid DCs. Based on CD8 and DEC205 expression, CD11c^high DCs could be further divided into three subpopulations: CD8⁻ DEC205⁻, CD8⁺ DEC205⁻ and CD8^low DEC205⁺ DCs.³⁹ Similarly, the activation status of each LNDC subpopulation was determined by mean fluorescence of FITC.

Test of LPS as a contaminant in mannan, OM and RM

To ensure mannan, OM and RM preparations free from endotoxin contamination, OM (800 µg/ml) was representatively tested in parallel with known LPS concentration samples by EML (Victoria, Australia). The result showed a concentration <0·06 EU/ml (equal to <0·01 ng/ml) of LPS was present in the OM sample. We titrated LPS (0·001–1000 ng/ml) to test whether BMDCs could be matured at this dosage of LPS after 18 hr of incubation.

Evaluation of cytokine mRNA expression via RT—PCR

BMDCs (2 × 10⁶) collected within 1–5 hr after stimulation with pH 9, mannan, OM, RM and LPS were snap-frozen in liquid nitrogen for 15 s. Frozen cells were stored at −20° for not more than 1 week if not used immediately. mRNA was extracted using a GeneElute Direct mRNA Miniprep Kit (Sigma) following the manufacturer's instructions. Briefly, cells were disrupted by sodium dodecyl sulphate/proteinase K digestion to release RNA and eliminate RNase. Thereafter, caution was taken to avoid any airborne or contactable RNase contamination. Polyadenylated mRNA was captured by hybridization with oligo dT30 covalently linked to 1 µm polystyrene beads. mRNA-bead complexes were washed on a microcentrifuge spin filter, followed by elution of mRNA into 10 mm Tris-HCL (pH 7·5) at 70°. mRNA was then coprecipitated with glycogen (Sigma) in ethanol and potassium acetate at −20° overnight. The precipitant was pelleted, dried and dissolved in 10 µl of RNase-free water. Oligo(dT) primers were added for hybridization with mRNA at 70° for 5 min. The mRNA sample was then mixed with dNTPs, mannan-murine leukemia virus (MLV) reverse transcriptase and mannan-MLV RT reaction buffer and left at 40° for 50 min for synthesis of the first strand cDNA. The cDNA product was kept in −70° if not used immediately.

cDNA was amplified using messageScreen mouse cytokine multiplex PCR kits (Biosource International, Camarillo, CA), for evaluation of mouse DC inflammatory cytokines (IL-6, IL-1β, GM-CSF and TNF-α) and Th1/Th2 cytokines (interferon-γ (IFN-γ), IL-12p40, IL-10, IL-4 and IL-13). PCR reactions were prepared in 25 µl containing 12·5 µl of 2× mPCR buffer, 2·5 µl of 10× dinucleotide triphosphates, 2·5 µl of 10× mPCR primer mixture (containing primers for amplifying cytokine and housekeeping enzyme reduced glyceraldehyde phosphate dehydrogenase cDNAs), 0·5 µl of Taq polymerase, 4·5 µl of ddH₂O and 2·5 µl of cDNA, 10× positive control or ddH₂O as blank control. The PCR thermocycle profile was optimized as follows: the initial denaturing step at 95° for 30 s, five cycles of the denaturing step at 94° for 30 s and the annealing step for 90 s, 35 cycles of the denaturing step at 94° for 30 s and the annealing step at 68° for 60 s and the final 70° for 10 min. The PCR product was stored at 4°. Ten µl of each PCR product and DNA molecular weight marker was mixed with 2 µl of 6× gel loading buffer (0·25% bromophenol blue, 30% of glycerol) and 10 µl of mixture was loaded onto 1·3% agarose gel containing 0·5 µg/ml ethidium bromide. The electrophoresis was run at 100 V for 90 min and the gel was examined with UV transilluminator.

Statistical analysis

All data are shown as the mean ± standard error of the mean (SEM). The data generated in this study were analysed by Student's t-test. Significance of difference was determined by the P-value (≤0·05).

Results

Mannan, OM and RM induce a mature phenotype of BMDCs in vitro

BMDCs do not express CD4 or CD8, resembling blood monocyte-derived DCs (MoDCs).⁴⁰ In a dose-dependent manner, mannan, OM and RM all up-regulated the expression of CD40, CD80 and CD86 of BMDCs (Fig. 1a), while pH 9 as the background control did not show up-regulation of these markers equivalent to non-pulsed DCs (data not shown). While mannan and its derivatives had little stimulatory effect at 50 µg/ml, their abilities to mature DCs were evident at 800 µg/ml (Fig. 1a, b). LPS at 1 µg/ml was used as a control. In a time course experiment (3–48 hr), mannan, OM and RM activated CD40, CD80 and CD86 expression of DCs. Up-regulation of maturation markers became apparent at 15 hr, with CD86 being most up-regulated. While mannan and its derivatives mature DCs in intermediate levels, BMDCs stimulated with LPS expressed the highest levels of maturation markers (Fig. 2). Additionally, mannan, OM and RM did not stimulate up-regulation of MHCnclass II molecules, since, in our experience, C57/BL6 BMDCs cultured in the GM-CSF and IL-4 system already express high levels of MHC class II. LPS as a strong DC maturation agent could only moderately up-regulate MHC class II expression (data not shown).

Figure 1.

(a) Mannan (M), OM, RM and LPS induce up-regulation of CD40, CD80 and CD86 expression of BMDCs in a dose-dependent manner. Stimulants, including pH 9 buffer (background control), 50, 400 and 800 µg/ml of mannan, OM and RM, and 1 µg/ml LPS (positive control), were added into BMDC cultures in duplicates. MFI of CD40, CD80 and CD86 expressed on CD11c⁺ populations was measured 18 hr after stimulation by flow cytometry. (b) Up-regulation of CD40, CD80 and CD86 induced by mannan (M), OM and RM (800 µg/ml) in comparison to pH 9 buffer and LPS is demonstrated by representative histograms of a replicate. The shaded area represents DCs stained with isotype control. The data shown is representative of three different experiments. *P = 0·05 versus pH 9-stimulated DCs.

Figure 2.

Kinetic analysis of DC activation by mannan, OM and RM. BMDCs were stimulated with pH 9 buffer, mannan (800 µg/ml), OM (800 µg/ml), RM (800 µg/ml) and LPS (1 µg/ml) in duplicate wells. Cells were collected at 3 hr, 6 hr, 15 hr, 24 hr and 48 hr and analysed for CD40, CD80 and CD86 expression by MFI. The data shown is representative of two experiments. *P = 0·05 versus pH 9-stimulated DCs.

Functional maturation induced by mannan, OM and RM was evidenced by enhanced MLR, OTI and OTII T-cell proliferation

To characterize functional maturation of BMDCs, we evaluated whether BMDCs matured by mannan, OM and RM were enhanced in their ability to stimulate proliferation of: (a) allogeneic T cells from BALB/c mice; (b) SIINFEKL peptide specific CD8 T cells from OTI mice; and (c) OVA_323−339 peptide-specific T cells from OTII mice. As shown in Fig. 3(a) mannan, OM, RM and LPS-stimulated DCs all enhanced allogeneic T-cell proliferation. In comparison with pH 9, DCs activated with mannan, OM and RM increased T-cell proliferation. BMDCs stimulated with mannan, OM and RM greatly enhanced proliferation of peptide-specific CD8 T cells, when 0·00001 µg/ml and 0·000001 µg/ml of SIINFEKL peptide concentrations were used (Fig. 3b). There were no significant differences between mannan and its derivatives. OVA_323−339-specific OTII T cells also showed an enhanced proliferation in the presence of mannan, OM and RM at 72 hr compared to pH 9 buffer control (Fig. 3c). LPS-pulsed DCs exhibited the greatest capacity in enhancing allogeneic, OTI and OTII T cell stimulation.

Figure 3.

(a) BMDCs matured by mannan (M), OM and RM stimulate allogeneic T-cell proliferation in MLR. BMDCs in cultures were activated with pH 9, 400 µg/ml of mannan, OM and RM, and 1 µg/ml of LPS. After 18 hr of stimulation, DCs were irradiated and cocultured with 5 × 10⁴ BALB/c T cells. On day 4, the culture was pulsed with ³H-thymidine and radioactivity was measured. (b) BMDCs matured by mannan, OM and RM enhance OTI T cell responses. BMDCs prestimulated with mannan, OM, RM and LPS for 18 hr were pulsed with titrated SIINFEKL peptide for 3 hr and irradiated DCs were seeded at 2 × 10³ together with 2 × 10⁴ OTI CD8 T cells. Proliferation of T cells was monitored at day 2 of culture. (c) BMDCs matured by mannan, OM and RM enhance OTII T-cell responses. Pre-stimulated DCs pulsed with OVA_323−339 were irradiated and cocultured with OTII T cells. Proliferation of CD4 T cells was measured at day 3. The data shown is representative of at least three separate experiments. *P = 0·05 versus pH 9-treated DCs.

Maturation of BMDCs by mannan, OM and RM requires TLR4 signalling

C3H/HeJ mice are defective in TLR4 signalling because of a point mutation and are unresponsive to LPS stimulation.³⁷ To test whether mannan, OM and RM required TLR4 to mature DCs, C3H/HeN and C3H/HeJ mice were used. As shown in Fig. 4, up-regulation of activation markers in wild type C3H/HeN BMDCs in response to mannan and LPS was diminished in C3H/HeJ BMDCs, suggesting the stimulatory effect of mannan was TLR4-depedent and MR/DC-SIGN/C-type lectin cross-linking was insufficient for DC maturation. In these experiments, OM and RM (800 µg/ml) were also included as stimulants and their stimulatory effects were equivalent to mannan (data not shown). This TLR4 dependence of mannan, OM and RM stimulation was not caused by LPS contamination, because OM (800 µg/ml) was representatively tested for the endotoxin level. OM contained <0·06 EU/ml (equivalent to <0·01 ng/ml) of LPS. In an experiment where titrated doses of LPS were tested, any dose below 0·1 ng/ml did not stimulate maturation of BMDCs (Fig. 5). The amount of LPS present in OM, which was representative of mannan and RM, was at two orders of magnitude below the minimal stimulatory dose.

Figure 4.

Mannan (M) and LPS stimulate C3H/HeN, but not C3H/HeJ BMDCs. Stimulants including pH 9 buffer, 800 µg/ml of mannan, 1 µg/ml of LPS and 50 µg/ml of poly (I:C) were added into C3H/HeN and C3H/HeJ BMDC cultures. MFI of CD40, CD80 and CD86 expressed on CD11c⁺ populations was measured 18 h after stimulation by flow cytometry. (a) In contrast to C3H/HeN DCs, up-regulation of costimulatory molecules, particularly CD40 and CD86, was greatly diminished in C3H/HeJ DCs in response to mannan and LPS stimulation, while up-regulation remained constant by poly(I:C). Mannan only minimally stimulated up-regulation of CD80 in C3H/HeN DCs because of its high expression in the steady state. (b) Representative histogram profiles of a replicate are shown. The shaded area represents DCs stained with isotype control. The data shown is representative of two different experiments.

Figure 5.

LPS present in OM is not sufficient to induce BMDC maturation. BMDCs were incubated with titrated LPS (0–1000 ng/ml) for 20 hr. MFI of CD40, CD80 and CD86 on CD11c⁺ populations was measured by flow cytometry. The data shown is representative of two different experiments. *P = 0·05 versus unstimulated DCs.

Mannan, OM and RM induce a mature phenotype of draining LN and splenic DCs in vivo

The observation that mannan, OM and RM were able to mature BMDCs led us to determine their ability to mature DCs in vivo. As shown in Figs 6 and 7, intradermal or intravenous injection with mannan, OM and RM induced a mature phenotype of draining LNDCs and splenic DCs, respectively, in mice.

Figure 6.

Injection with mannan (M), OM and RM induces a mature phenotype of LNDCs. Mice were injected intradermally with 350 µg of mannan, OM and RM, 50 µl of pH 9 buffer and 5 µg of LPS. Popliteal LNDCs were harvested 18 hr after injection and analysed for their surface expression of CD40, CD80 and CD86. (a) The CD11c^high population was gated in the dot plot (CD11c versus non-stain FITC). Within this population, three subpopulations including CD8⁻ DEC205⁻, CD8⁺ DEC205⁻ and CD8^low DEC205⁺ were identified. (b) Expression levels of CD40, CD80 and CD86 of total (CD11c^high) and each subpopulation in different injection groups are shown. (c) Expression of CD40, CD80 and CD86 is demonstrated by representative histogram profiles of total DCs pooled from two immunized mice from each injection group. The shaded area represents DCs stained with isotype control. The data shown is representative of two separate experiments. *P = 0·05 versus pH 9-stimulated DCs.

Figure 7.

Intravenous injection with mannan (M), OM and RM stimulates a mature phenotype of splenic DCs. Mice were injected intravenously with 100 µl of pH 9, 4·3 mg of mannan, OM, and RM, and 20 µg of LPS. Splenic CD11c^high cells were analysed for their surface expression of CD40, CD86 and CD86 18 hr after injection. The data shown is representative of three experiments. *P = 0·05 versus pH 9-stimulated DCs.

In agreement with previous studies, we identified three major subsets in the popliteal LNs: CD8⁻ DEC205⁻, CD8⁺ DEC205⁻ and CD8^low DEC205⁺ DCs (Fig. 6a).³⁹ DC activation was analysed based on these three subsets. In mice injected with pH 9 as the background control, CD8^low DEC205⁺ subpopulation appeared to be the major subpopulation (>50% of total DCs, data not shown) and demonstrated a more mature phenotype by expressing higher CD40, CD80 and CD86 levels, while the other two subpopulations were immature (Fig. 6b). This observation co-ordinated with previous findings.³⁹ Interestingly, immature DCs, especially CD8⁺ DEC205⁻ DCs, responded rapidly to mannan, OM and RM stimulation by considerably up-regulating CD40 and CD80, and less so with CD86. They were the only subset that up-regulated MHC class II expression in response to all stimulants (data not shown). Immature CD8⁻ DEC205⁻ DCs also responded to mannan, OM and RM stimulation by up-regulating mainly CD80. CD8^low DEC205⁺ DCs, known as the migratory subpopulation, responded to mannan, OM and RM stimulation and exhibited a much more mature phenotype than that of the pH 9 control group (Fig. 6b). In addition, up-regulation of CD40, CD80 and CD86 of total LNDCs pooled from each injection group in response to mannan, OM and RM injections was observed (Fig. 6c).

In addition to intradermal injection, mice were also injected intravenously with stimulants. It was found that the splenic DC population was more homogeneous, when CD40, CD80 and CD86 expression of splenic DCs of pH 9-injected mice was examined. Previous studies have demonstrated three conventional splenic DC subpopulations (CD4⁺ CD8⁻, CD4⁻ CD8⁻ and CD4⁻ CD8⁺ DCs) express intermediate levels of CD40, CD80 and CD86 with a high turn-over rate, and they could undergo further maturation in response to stimulation with bacterial adjuvants such as LPS.⁴¹ In our studies, mice injected with mannan, OM, RM and LPS all demonstrated up-regulation of CD40, CD80 and CD86 maturation markers (Fig. 7). While LPS particularly stimulated up-regulation of CD86, mannan, OM and RM appeared to up-regulate more CD40 and CD80 expression.

BMDCs produce similar inflammatory cytokines, but differential Th1/Th2 cytokines, in response to mannan, OM, RM and LPS stimulation

Previous studies demonstrated mannan was able to induce TNF-α production, and OM and RM conjugated to MUC1-antigen stimulated production of Th1 (IL-12) and Th2 (IL-10) cytokines, respectively, by macrophages.^6,18,30 We questioned whether mannan, OM and RM were able to stimulate cytokine production in BMDCs, and if they could, whether the cytokine patterns were different. In our experience, cytokine mRNAs, especially those of Th1/Th2, in BMDCs, could disappear in 6 hr; therefore, we monitored the first 5 hr after adding stimulants. As shown in Fig. 8(a) mannan, OM, RM and LPS all stimulated robust mRNA up-regulation of innate inflammatory cytokines including IL-6, IL-1β, GM-CSF and TNF-α. Although IL-6 and GM-CSF mRNAs appeared in pH 9-stimulated BMDCs, we confirmed that they were in fact intrinsically produced by cultured BMDCs (data not shown). Hereby, especially IL-1β and TNF-α mRNAs were rapidly up-regulated in the first hour, suggesting these cytokines were required for innate counteraction against the stimulants, which all possessed a pathogen-derived structure. mannan, OM, RM and LPS also selectively stimulated Th1/Th2 cytokine mRNA expression, while BMDCs treated with pH 9 did not produce any of these cytokines (Fig. 8b). mannan appeared to be a stimulator for IFN-γ, IL-12p40 and IL-4. OM primarily stimulated IFN-γ and IL-12p40 and less IL-4 production. RM stimulated production of IL-10, less IFN-γ and IL-12p40 compared to OM, IL-4 and IL-13. LPS derived from E. coli stimulated primarily IL-12p40, as well as IFN-γ and IL-10, but little IL-4 production. RM stimulation of mRNA cytokine production exhibited a delayed kinetics by occurring at the fourth hr, comparing to the third hr for other stimulants. RM stimulation also resulted in abundance of IL-13 production not seen in other stimulants (Fig. 8b). It is therefore clear that OM and RM induced differential cytokine mRNA levels in BMDCs.

Figure 8.

Mannan (M), OM and RM stimulate production of similar innate inflammatory cytokines, but differential Th1/Th2 cytokines in BMDCs. Stimulants including pH 9 buffer, 400 µg/ml of mannan, OM and RM, and 1 µg/ml of LPS were added into BMDC cultures. Cells were harvested at 1, 2, 3, 4 and 5 hr poststimulation, followed by mRNA extraction. After reverse transcription, cDNA of (a) the housekeeping enzyme GAPDH, inflammatory cytokines (IL-6, IL-1β, GM-CSF and TNF-α) and (b) Th1/Th2 cytokines (IL-10, IFN-γ, IL-2, IL-12p40, IL-4 and IL-13) were amplified and analysed. While all selected innate inflammatory cytokines were greatly up-regulated, differential Th1/Th2 profiles were noted, across mannan, OM, RM and LPS. The data shown is representative of at least two experiments.

Discussion

Mannan, a fungal component derived from S. cerevisiae or C. albicans, is a potential PAMP recognized by TLRs, which are present in a wide spectrum of cell types including APCs at various levels. Lipoarabinomannan and lipomannan derived from mycobacteria are recognized by TLR2 and induce differential effector responses.³³ Phospholipomannan, present in the cell surface of C. albicans, and glucuronoxylomannan, the major capsular polysaccharide of Cryptococcus neoformans, are recognized by TLR2, and TLR4, respectively.⁴² Previous studies on macrophages have identified mannan as a ligand for TLR4 that stimulates TNF-α production.³⁰ On the other hand, mannan is recognized by the C-type lectin receptors such as MR and DC-SIGN. Several of these receptors (e.g. MR) are considered as pattern recognition receptors involved in host defence and innate immunity, being able to recognize a range of carbohydrates present on the surface and cell walls of microbes, and mediate endocytic and phagocytic activities. They are abundantly expressed by macrophages and myeloid DCs.^43–47 In previous studies, cross-linking of monoclonal anti-human MR antibody PAM-1 induces maturation of human MoDCs and production of immunosuppressive cytokines including IL-10, IL-1ra and IL-1RII, leading to T-cell anergy.³⁵ In comparison to these findings, we demonstrate that binding of mannan, OM and RM results in maturation of in vitro BMDCs, which is verified by up-regulation of maturation markers including CD40, CD80 and CD86 in a dose-dependent manner. BMDCs stimulated with mannan, OM and RM virtually reaches a similar magnitude of activation levels apparent in 15 hrs. Functional maturation is also evidenced by enhanced allogeneic MLR and peptide-specific T-cell proliferation. These findings coordinate with previous studies on maturation of BMDCs by acemannan.³⁴ We show that DC maturation induced by mannan, OM and RM is not due to LPS contamination, because the level of LPS present in OM is two magnitudes lower than the stimulatory dose and OM has been repeatedly tested more than 20 times in previous clinical trials and the LPS level has been <0·06 EU/ml.^28,29 The finding that mannan, OM, RM and LPS fail to induce maturation of TLR4-defective DCs suggests mannan and its derivatives mature DCs in a TLR4-dependent manner and C-type lectin cross-linking with mannan, OM and RM is insufficient to stimulate DCs. It has been proposed that the nature and structure of a ligand may determine how the receptor is engaged. Different levels of engagement of multiple receptors and affinities may affect the intracellular signalling pathways. In this way, anti-MR antibody may bind to the MR with different affinities in comparison to mannan,³⁵ resulting in differential DC responses.

Functionally, DCs pulsed with mannan, OM and RM exhibit enhanced capacities in stimulating allogeneic T-cell proliferation. This is supported by the results obtained after stimulating OTI and OTII T cells with BMDCs pulsed with mannan, OM and RM prior to pulsing with SIINFEKL and OVA_323−339 peptides from ovalbumin. Mannan, OM and RM show once again an enhanced capacity in stimulation of antigen-specific CD8 and CD4 T-cell proliferation. Our observation in functional maturation co-ordinates with phenotypic maturation of DCs induced by mannan, OM and RM. Although it appears that the stimulatory effect of mannan, OM and RM is less than that of LPS in DCs, clinical use of LPS for induction of DC immunogenicity is undesirable, because of its toxicity (e.g. induction of endotoxic shock) and batch-to-batch variation. Vaccines based on OM conjugates are advantageous, because they also target the C-type lectin expressing DCs to facilitate antigen presentation in addition to their moderate stimulatory effect. Moreover, OM conjugates are safe in patients and are currently in phase II clinical trials.

Previous studies have shown that immunization with mannan and its derivatives as vaccine delivery vehicles greatly enhance antigen presentation and immune responses. Apart from enhanced antigen uptake, we suggest that the adjuvanticity of mannan, OM and RM is in part the result of indirect induction of DC maturation in vivo. We observe that injection with mannan, OM and RM can induce maturation of DCs present in draining LNs and spleens in mice. Of note, TLR4 is expressed in cultured BMDCs and MoDCs in high levels, but is in extremely low levels in mouse conventional DC subpopulations (CD11c⁺ CD8⁺, CD11c⁺ CD8⁻). Freshly isolated DCs do not respond to LPS stimulation.^48,49 The fact that LPS, as well as mannan, OM and RM, can mature in vivo DCs is most likely due to indirect mediation by TLR4 high-expressing non-DCs (e.g. endothelial cells), DC precursors (e.g. monocytes) or macrophages, a mechanism distinct from stimulation of in vitro BMDCs. The effect of these cells may strengthen the stimulatory effect of mannan, OM and RM locally, and partially explain why immature LNDCs, especially CD8⁺ DEC205⁻ subpopulation, respond robustly to mannan, OM and RM injection to a level similar to LPS, in the absence of TLR4 signalling in DCs. In contrast to intradermal injection, our initial attempts in intravenous immunizations using lower doses (1·4 mg and 2 mg) of mannan, OM and RM could only slightly drive up-regulation of activation markers in splenic DCs (data not shown). This may be due to the following reasons. First, splenic DCs already exhibit an intermediate level of a mature phenotype in the steady state.⁴¹ Second, mannan, OM and RM are likely to be accumulated in the internal organs and vastly diluted in the system after immunization. Third, cytokines produced by systemic TLR4⁺ cells (e.g. monocytes) are not concentrated enough to induce high levels of DC maturation. Under these circumstances, an increase of dosage to 4·3 mg is required to up-regulate expression of activation markers, particularly CD40 and CD80, to an apparent level of difference.

The extensive maturation of LN CD8⁺ DCs in mannan, OM and RM injected mice may imply their initial preference in Th1 cell priming. In mice, only CD8⁺, but not CD8⁻ DCs, have been shown to induce antigen-specific Th1 responses, characterized by IL-12 and IgG2a production.⁵⁰ However, mice immunized intradermally with OM and RM-conjugated antigens into footpads generate differential antigen-specific Th1 and Th2 responses, respectively.¹¹ Modifying mannan has a large impact on antigen processing pathways, which are followed by overall APCs for antigen presentation. OM mainly delivers antigens into the MHC-class I processing pathway, while RM into the MHC-class II pathway.¹⁰ As a result, it appears that the immune settings are more conducive to the response elicited by the OM-antigen than the RM-antigen in the lymph milieu after immunization, resulting in its exceptional adjuvanticity for eliciting antitumour responses. This also implies, at least in mice, that intradermal injection with OM, which targets immature LNDC subpopulations, may be superior to any other forms of systemic administration.

In our cytokine studies, we demonstrate that mannan, OM and RM stimulate production of inflammatory cytokines. The intrinsic production of cytokines including IL-6 and GM-CSF indicates that these inflammatory cytokines are required for development of bone marrow precursor cells into functional DCs. GM-CSF and IL-6 have both been implicated in cultures for generation of DCs.⁵¹ Nonetheless, IL-6, together with IL-1β and TNF-α, are greatly up-regulated in response to mannan, OM, RM and LPS. The rapid induction of these cytokines indicates the significance of innate immunity for clearing pathogen-associated structures such as mannan and LPS. Apart from inflammatory cytokines, mannan, OM and RM differentially stimulate Th1/Th2 cytokines. This differential stimulation is not surprising, because the structural change of a ligand may result in alteration of its binding affinity to its receptor, resulting in distinctive cellular responses. LPS, for example, derived from different pathogens, has different stimulatory properties because of the structural variations in the lipid moiety.^52,53 RM appears to induce cytokine production in a more delayed kinetics. It exclusively induces abundant mRNA of Th2-skewing cytokines including IL-10 and IL-13, together with IFN-γ and IL-12p40. Although a trace amount of IL-4 mRNA was detected, OM induces primarily IFN-γ and IL-12p40. These findings appear to co-ordinate with previous macrophage and immunization studies on oxidized and reduced MFP.^6,18 The cytokine pattern of mannan is similar to OM, but with more IL-4 and IL-13. Dual IL-12p40, IFN-γ as Th1, and IL-10 and weak IL-4 as Th2, cytokine production caused by stimulation of E. coli LPS is also observed. Because cytokine mRNA is produced in such kinetics by BMDCs, our studies also urge revision of current immunotherapeutic protocols, in which human ex vivo MoDCs are usually left overnight after maturation procedures prior to immunization into patients. Particularly, with OM as an adjuvant for tumour antigen, the mRNA production begins in 3 hr. Prolonged maturation may result in exhaustion of cytokine production, leading to a decreased capability in priming T-cell responses.

Our studies have provided insights of the interaction between mannan and DCs. Although the ligand-receptor binding mechanisms between mannan and C-type lectins and between mannan and TLR4 remain to be clarified, we demonstrate that mannan, OM and RM can induce a mature DC phenotype in a TLR4-dependent manner. We have clarified the kinetics of cytokine mRNA expression in BMDCs in the presence of mannan, OM and RM, which has important implications in manipulation of ex vivo DCs for immunotherapy. Moreover, the proposed mechanisms for LN immune induction with OM and RM after intradermal injection should explain why oxidized MFP can induce such strong Th1 and CTL responses.

Abbreviations

BMDC

bone-marrow-derived dendritic cell

c.p.m.

counts per minute

DC

dendritic cell

LN

lymph node

LNDC

lymph node dendritic cell, LPS, lipopolysaccharide

mannan

mannan

MFI

mean fluorescence of intensity

MLR

mixed lymphocyte reaction

MR

the mannose receptor family

OM

oxidized mannan

PAMP

pathogen-associated molecular pattern

poly(I:C)

Polyriboinosinic polyribocytidylic acid

RM

reduced mannan

SEM

standard error of the mean

TLR4

Toll-like receptor 4