Complement receptor 3-dependent engagement by Candida glabrata β-glucan modulates dendritic cells to induce regulatory T-cell expansion

Candida glabrata is an important pathogen causing invasive infection associated with a high mortality rate. One mechanism that causes the failure of Candida eradication is an increase in regulatory T cells (Treg), which play a major role in immune suppression and promoting Candida pathogenicity. To date, how C. glabrata induces a Treg response remains unclear. Dendritic cells (DCs) recognition of fungi provides the fundamental signal determining the fate of the T-cell response. This study investigated the interplay between C. glabrata and DCs and its effect on Treg induction. We found that C. glabrata β-glucan was a major component that interacted with DCs and consequently mediated the Treg response. Blocking the binding of C. glabrata β-glucan to dectin-1 and complement receptor 3 (CR3) showed that CR3 activation in DCs was crucial for the induction of Treg. Furthermore, a ligand–receptor binding assay showed the preferential binding of C. glabrata β-glucan to CR3. Our data suggest that C. glabrata β-glucan potentially mediates the Treg response, probably through CR3-dependent activation in DCs. This study contributes new insights into immune modulation by C. glabrata that may lead to a better design of novel immunotherapeutic strategies for invasive C. glabrata infection.


Introduction
An opportunistic yeast, Candida glabrata, is the second most frequent cause of invasive candidiasis, which is often nosocomial [1,2].The rapid progression of invasive candidiasis caused by C. glabrata is associated with substantial morbidity and mortality, perhaps owing to its inherent low susceptibility to anti-fungal medications, azoles, polyenes and echinocandins [2,3].Although C. glabrata does not have the morphological flexibility that is a crucial virulence determinant of C. albicans, C. glabrata is a successful human pathogen because of its immune evasion ability [4,5].
A mechanism causing the failure of host protective immunity and an increase in Candida pathogenicity in systemic candidiasis is an increased response of regulatory T cells (Treg) [6].Treg are a reciprocal T-cell subset of the antifungal T helper 17 (Th17) [7].β-glucan is the most abundant polysaccharide of the inner skeleton of the Candida cell wall, and it plays a pivotal role in the induction of the host defence against Candida infection [8].Recognition of Candida β-glucan by dectin-1 is crucial for the induction of antifungal Th17 immunity [9,10].Treg expansion is mediated by the recognition of C. albicans via Toll-like receptor (TLR)-2-Toll/IL-1R domain-containing adaptor-inducing IFN (TRIF) signalling in antigen-presenting cells [11,12].Dectin-1 is also important for the eradication of systemic C. glabrata infection [13,14]; however, the induction of Treg by C. glabrata remains to be determined.
DCs are the most potent antigen-presenting cells that are central to immune activation and immune tolerance [22].DCs abundantly express pattern-recognition receptors (PRRs), such as C-type lectin receptors and Toll-like receptors, which play pivotal roles in sensing invading fungi.The interaction between PRRs and fungal-derived pathogen-associated molecular patterns (PAMPs) enables the functional versatility of DCs, which orchestrates T-cell fate and function [23].The induction of Treg also requires the signal transduction carried by DCs [24].At present, little is known regarding the interplay between C. glabrata and DCs.Therefore, this study aimed to investigate the role of C. glabrata β-glucan in the immune modulation of DCs via its specific receptors, dectin-1 and CR3.We also examined the effect of C. glabrata β-glucan-DC interactions on T-cell responses.A better understanding of the host immune response to C. glabrata might help develop new or improved approaches to control and treat invasive fungal infection.

High-dose C. glabrata infection mediates regulatory T-cell expansion and high interleukin-10 production
To examine the alteration in immune responses to systemic C. glabrata infection, mice were intravenously inoculated with various doses of C. glabrata and a splenic T-cell population, and serum cytokines were determined at day 7 post-infection (figure 1 and electronic supplementary material, figure S1).Total splenic cell numbers were increased in accordance with the increased doses of C. glabrata, while the proportion of splenic CD3 + CD4 + T cells was notably increased in medium-dose C. glabrata-infected mice (10 × 10 6 cells) (figure 1a,b).However, the proportion of CD3 + CD4 + T cells was diminished in high-dose C. glabrata-infected mice (30 × 10 6 cells) (figure 1b).Interestingly, the splenic Foxp3 + Treg population was significantly increased in a dose-dependent fashion.This population was the most abundant in high-dose C. glabrata-infected mice, which was consistent with the decreased CD3 + CD4 + T cells in this high-dose infected group (figure 1c and electronic supplementary material, figure S1a).
Serum interleukin (IL)-17 concentrations were significantly increased upon C. glabrata infection, but IL-17 concentrations were not different among the different doses of C. glabrata (figure 1d).Elevated serum interferon (IFN)-γ and IL-10 concentrations were associated with the increased doses of C. glabrata (figure 1e,f).Notably, serum IL-10 concentrations in high-dose C. glabrata-infected mice (figure 1f) were substantially elevated in tandem with the increase in Foxp3 + Treg (figure 1c).
Splenocytes from all groups were restimulated ex vivo with immobilized anti-CD3 to determine T-cell specific responses (figure 1g,k and electronic supplementary material, S1b).The proportion of restimulated CD3 + CD4 + T cells was not changed in the low-dose and medium-dose groups, but that of CD3 + CD4 + T cells was reduced in splenocytes from high-dose C. glabratainfected mice (figure 1g).Foxp3 + Treg were apparently increased in the restimulated splenocytes from high-dose C. glabrata infection (figure 1h and electronic supplementary material, figure S1b).Additionally, IFN-γ, IL-17 and IL-10 concentrations were detected in the culture supernatant of the restimulated splenocytes.All cytokines were increased in a dose-dependent fashion (figure 1i-k).However, IL-10 concentrations were markedly augmented in splenocytes from high-dose C. glabrata-infected mice (figure 1k).Taken together, these results suggest that systemic C. glabrata infection induces the immunosuppressive response by promoting Treg expansion and IL-10 production.

Candida glabrata β-glucan induces antigen-specific Treg expansion and augmented IL-10 production
Mannans and β-glucans are major carbohydrate constituents in the Candida cell wall and they play critical roles in immune modulation [8,25].Therefore, we examined whether these components from C. glabrata are involved in eliciting the Treg response.C. glabrata mannan or β-glucan was mixed with an immunogenic antigen, oval chicken albumin (OVA), and was subcutaneously administered to mice on days 0 and 7.In this experiment, complete Freund's adjuvant (CFA) was used as the positive control because this adjuvant potentially activates the immune response.On day 14, lymph node (LN) cells from immunized mice were restimulated ex vivo with OVA and a T-cell population, and the cytokine profile was analysed (figure 2).CFA effectively enhanced antigen-specific CD4 T-cell expansion, while C. glabrata mannan and β-glucan did not (figure 2a).In agreement with the increased number of CD3 + CD4 + T cells, the number of antigen-specific Foxp3 + and Foxp3 + CD25 + Treg was diminished in the CFA immunized group (figure 2b,c, electronic supplementary material, figure S2).In contrast, C. glabrata β-glucan, but not mannan, potentially induced Foxp3 + and Foxp3 + CD25 + Treg expansion (figure 2b,c).
IFN-γ, IL-17 and IL-10 production was also detected in the supernatant of the antigen-restimulated LN cells (figure 2d-f).In accordance with T-cell responses (figure 2a,b), LN cells from C. glabrata mannan-immunized mice did not produce any cytokines, while LN cells from C. glabrata β-glucan-immunized mice markedly produced IL-10, but only slightly produced IFN-γ and IL-17.These data suggest that C. glabrata β-glucan plays a crucial role in Treg induction and IL-10 production in systemic C. glabrata infection.

DC responses are mainly modulated by C. glabrata β-glucan
DCs play a central role in T-cell differentiation, and the interaction between DCs and mannans and β-glucans in the Candida cell wall is important to modulate the host immune responses [26].Therefore, we compared the responses of bone marrow-derived DCs (BMDCs) to β-glucan, mannan and heat-killed C. glabrata yeasts that could not produce all secretory molecules, and allowed the cell wall components (mainly composed of mannan and β-glucan) to interact with DCs.DC maturation and cytokine production were determined at 24 h post-stimulation by flow cytometry and an enzyme-linked immunosorbent assay (ELISA), respectively (figure 3 and electronic supplementary material, figure S2).Heat-killed yeasts and β-glucan strongly induced the expression of DC maturation markers (CD40, CD80, CD86 and MHC class II) to similar levels, while mannan slightly activated DCs (figure 3a-d and electronic supplementary material, figure S2a,b).β-Glucan and, to a lesser extent, heat-killed yeasts potentially activated BMDCs to produce all cytokines, including IL-10, while mannan failed to stimulate cytokine production in BMDCs (figure 3e-k).Additionally, these different DC responses induced by heat-killed yeasts, mannan and β-glucan were not associated with DC viability because a cell viability test showed no difference among all stimulations (electronic supplementary material, figure S2c).Therefore, C. glabrata β-glucan, but not mannan, may be the main component that modulates DC responses.4k).Furthermore, all doses of both Candida β-glucans did not affect DC viability (electronic supplementary material, figure S3c).These data suggest that DCs respond differently to β-glucan from two different Candida species, and C. glabrata β-glucan has a greater ability to promote IL-10 production in DCs.

Candida glabrata β-glucan mediates Treg expansion via CR3 on DCs
After demonstrating the in vivo effects of C. glabrata β-glucan on Treg expansion (figure 2), we next examined whether the engagement of C. glabrata β-glucan to its receptors, dectin-1 and CR3, in DCs affects Treg differentiation.In vitro DC-T-cell co-cultures were performed in the presence of a soluble anti-CD3 mAb.In this experimental setting, the Fc portion of anti-CD3 mAb binds to the Fc receptor on DCs, while the Fab portion binds to CD3 on T cells.With this interaction, T-cell receptor (TCR) signalling is activated [31,32].In parallel, the co-signals and cytokines were derived from BMDCs that were stimulated with C. albicans and C. glabrata β-glucan in the presence or absence of a dectin-1 antagonist or anti-CD11b mAb (figure 7).Forty-eight hours after the co-cultures, CD3 + CD4 + and CD3 + CD4 + Foxp3 + T cells were identified using flow cytometry (figure 7a ).Although the proportion of CD3 + CD4 + T cells was not different in all conditions (figure 7a), the Foxp3 + Treg population was notably increased when BMDCs were stimulated with C. glabrata β-glucan (figure 7b).Intriguingly, the Foxp3 + Treg population was increased when dectin-1 on BMDCs was blocked, but this population was markedly decreased upon CR3 blockade (figure 7b).An immunosuppressive cytokine, IL-10 and an anti-fungal cytokine, IL-17, in the supernatant of the DC-T-cell co-cultures were also determined.Consistent with the Treg responses (figure 7b), IL-10 concentrations were increased upon dectin-1 blockade and diminished when CR3 on BMDCs was blocked (figure 7c).In contrast, concentrations of IL-17, representing Th17, which is a reciprocal development pathway of Treg [33], were decreased when dectin-1 on BMDCs was blocked and it was increased upon CR3 blockade (figure 7d).Blockades of both receptors in BMDCs stimulated with C. albicans and C. glabrata β-glucan produced similar Foxp3 + Treg and cytokine responses (figure 7a-d).
To observe the impact of dectin-1 and CR3 ligation in DCs on activated T cells along with Treg, splenic T cells were isolated and co-cultured with BMDCs in the presence of soluble anti-CD3 as described above.At 48 h after the co-culture, T-cell proliferation was determined by flow cytometric analysis (electronic supplementary material, figure S10).CD3 + CD4 + CD25 + Foxp3 − activated T cells and CD3 + CD4 + Foxp3 + Treg were gated, and then CFSE low cells (proliferated cells) were analysed (electronic supplementary material, figure S10a).Consistently, T cells co-cultured with C. glabrata β-glucan-stimulated DCs showed the lower number of CD3 + CD4 + CD25 + Foxp3 − activated T cells but had the increased number of CD3 + CD4 + Foxp3 + Treg.Blockade of CR3, but not dectin-1, in C. glabrata β-glucan-stimulated DCs increased the number of CD3 + CD4 + CD25 + Foxp3 − activated T cells and decreased the number of CD3 + CD4 + Foxp3 + Treg (electronic supplementary material, figure S10b,c).These results suggested that the alteration in CD3 + CD4 + CD25 + Foxp3 − activated T cells may be resulted from the suppressive activity of the Treg mediated by C. glabrata β-glucan-stimulated DCs in a CR3-dependent manner.
To evaluate the effect of dectin-1 and CR3 ligation in DCs on activated T cells and Treg, splenic T cells were isolated and co-cultured with BMDCs in the presence of soluble anti-CD3.T-cell proliferation was assessed after 48 h of co-culture by flow cytometric analysis.Activated T cells (CD3 + CD4 + CD25 + Foxp3 − ) and Treg (CD3 + CD4 + Foxp3 + ) were identified and analysed for CFSE low cells (proliferated cells) (electronic supplementary material, figure S10a).T cells co-cultured with C. glabrata β-glucanstimulated DCs displayed a lower number of activated T cells but an increased number of Treg.Blockade of CR3, but not dectin-1, in C. glabrata β-glucan-stimulated DCs augmented the number of activated T cells and reduced the number of Treg (electronic supplementary material, figure S10b,c).These findings suggest that the decrease in activated T cells may be owing to the suppressive activity of Treg mediated by C. glabrata β-glucan-stimulated DCs in a CR3-dependent manner.

Morphology and structural feature of C. glabrata β-glucans
The morphological appearance and structure of Candida β-glucans are implicated in receptor binding activity [36][37][38][39].Therefore, we observed the morphology and structure of C. glabrata β-glucan (figures 8 and 9, and electronic supplementary material, figure S8).The morphology of C. glabrata and C. albicans β-glucan was characterized by scanning electron microscopy.C. glabrata β-glucan formed a small spherical shape of approximately 2 µM in size, while C. albicans β-glucan formed a large oval shape of approximately 5 µM in size.Furthermore, C. glabrata β-glucan was a coarse and rough particle, but C. albicans β-glucan appeared as a dense fine particle (figure 8).
Characterization of the C. glabrata β-glucan structure was then carried out using nuclear magnetic resonance (NMR) spectroscopy (figure 9 and electronic supplementary material, figure S8).NMR characterization of carbohydrates [40][41][42], especially polysaccharides [39,43] can be challenging owing to various factors, such as poor solubility in deuterated solvents [39], the complexity of the carbohydrate structure [44][45][46], and signal overlap between protons of hydroxyl groups and the pyranose/furanose region [40,41].Initially, C. glabrata β-glucan showed poor solubility in polar deuterated NMR solvents, such as dimethyl sulfoxide (DMSO)-d 6 , D 2 O and CD 3 OD.After several trials, adding deuterated trifluoroacetic acid (TFA-d 1 ) to a solution of C. glabrata β-glucan in DMSO-d 6 enhanced the solubility compared with using DMSO-d 6 alone.Protonation and exchange of the deuterium ion from TFA-d 1 with the hydroxyl proton additionally simplified the coupling signals between the hydroxyl and pyranose ring protons, thus, facilitating the interpretation of the 1 H NMR spectra.

Discussion
Studies have shown the crucial role of cell wall β-glucan of C. albicans in the induction of anti-fungal immunity [8,48].In this study, we found that the cell wall β-glucans of C. glabrata had different effects on DC immunity and subsequent T-cell responses.The role of cell wall β-glucan in the yeast form of Candida was investigated because C. glabrata is a nondimorphic yeast.Candida cell walls are mainly constructed by an outer mannan layer and inner β-glucan skeleton [8].C. glabrata mannan was found to induce DC and macrophage activation and cytokine production [49], but C. glabrata mannan in our study lacked the ability to induce immune responses (figures 2 and 3).These differences may have resulted from the inter-strain variations of the mannan structure [50].However, β-glucan is exposed in the cell wall of the yeast form.Therefore, β-glucan in the inner layer can interact with the host PRRs [51].Accordingly, C. glabrata heat-killed yeast and β-glucan showed similar DC responses (figure 3).
Dectin-1 is a receptor that recognizes Candida β-glucans and it has a major role in inducing anti-fungal immunity [10].The genetic ablation of dection-1 in mice leads to an increased susceptibility to systemic infections caused by various Candida species [13,52].Dectin-1-dependent responses to cell wall β-glucans differ among Candida species [13,39].We found that the interaction between dectin-1 on BMDCs and the distinct Candida β-glucan led to a variation in DC responses (figures 5 and 6, electronic supplementary material, figures S6 and S7).CR3 also recognizes yeast-derived β-glucan and plays a pivotal role in the host defence mechanism [16,20].In this study, we found that the ligation between β-glucans from distinct Candida species and CR3 also had different effects on immune responses (figures 5-7).
Although there is no evidence showing the binding characteristics between β-glucan and CR3, several lines of evidence suggest that the variation in β-glucan structures and contents is a main factor that affects receptor recognition and immune responses [13,36,53].NMR analysis of other and our previous publications has shown that β-glucan of the C. albicans yeast form is composed of β-(1,3)-glucan with β-(1,6)-branching [37,39,54].In this study, we found that β-glucan of C. glabrata yeast was also composed of β-(1,3)-glucan with β-(1,6)-branching, which is consistent with a previous report [47] (figure 9 and electronic supplementary material, figure S8).However, the composition of β-(1,3)-glucan and β-(1,6)-glucan of both Candida species was different.β-(1,3)-glucan of C. glabrata was homogeneous and uniform (figure 9 and electronic supplementary material, figure S8).Additionally, a high-field NMR analysis showed that C. glabrata β-glucan contained a larger amount of β- (1,6)-glucan than C. albicans [37,39,47,54].Curdlan consists entirely of (1→3)-β-d-glucosidic linkages [55].Blockade of dectin-1 and CR3 affected BMDC response to curdlan differently than Candida β-glucan.(electronic supplementary material, figures S7 and S8).The data have confirmed that β-glucan structures and contents are significant in the response of DC.The different particle sizes of β-glucans may be partly responsible for the variation in immune responses [38].As shown by a morphological analysis, C. albicans and C. glabrata β-glucan differed in size and shape (figure 8).The particulate or immobilized form of β-glucan is required for inducing signal transduction via dectin-1, while the soluble and particulate forms of fungal β-glucan can activate CR3 [18,19,56].In this study, we used the particulate form of Candida β-glucans to recapitulate the architecture of β-glucan in the cell wall.The activation of human neutrophils by particulate yeast β-glucan is dependent on CR3 [18].We also found that ligation of the particulate form of C. albicans and C. glabrata β-glucans to CR3 activated DC responses (figures 5 and 6).Additionally, particulate Candida β-glucans interacted with dectin-1 and CR3 in a different manner, which resulting in the differential immune responses (figures 5-7).These differences can be simply explained by the distinct signal transduction of dectin-1-dependent and CR3-dependent pathways.The ligation of β-glucans to dectin-1 mainly mediates downstream signalling via the Syk/CARD9 axis, while the ligation of β-glucans to CR3 induces a cellular response via the Src/ERK and Syk/PI3K pathways [28,[57][58][59].Treg, which play a critical role in immune suppression and enhance Candida pathogenicity, are greatly increased in disseminated invasive candidiasis caused by C. albicans [6].The induction of Treg by C. albicans requires Toll-like receptor-2 and TRIF signalling in antigen-presenting cells [11,12].In this study, we found that C. glabrata was also capable of Treg induction, but via the engagement of its β-glucan to CR3 on DCs (figures 1, 2 and 7).Although C. glabrataβ-glucan activated dectin-1 and CR3 (figures 5 and 6), a binding assay suggested that C. glabrataβ-glucan preferentially bound to CR3 (figure 10).The signal transduction via dectin-1 on DCs appeared to have less influence on Treg induction, while the signal from CR3 was probably dominant, which consequently led to Treg induction (figure 7b).One mechanism that may be involved in this Treg induction is via the PD-1/PD-L1 pathway [34,35].This possibility is supported by our finding that ligation of C. glabrata β-glucan to CR3, but not dectin-1, induced high PD-L1 expression, and this expression was inhibited by CR3 blockade (figure 7e,f).The C. glabrataβ-glucan-mediated immunomodulatory function of CR3 may result from the activation of CD11b [28].A recent study showed that the immunomodulatory effect of soluble β-glucan depended on the stimulation of CR3, and this stimulation may be owing to the ligation of soluble β-glucan to the lectin site located on CD11b of CR3 [19,60].Additionally, CD11b signalling can negatively regulate inflammatory signalling in immune cells [59,61] and the activation of CD11b in DCs can suppress the DC response and inhibit T-cell activation [62,63].The interference of CD11b ligation in this study indicated immunomodulatory activity of C. glabrata β-glucan (figure 7).
The high IL-10 production induced by C. glabrata β-glucan may be attributed to the RORγt + Foxp3 + Tr17.Previous studies have shown that RORγt + Foxp3 + Tr17 are significant producers of IL-10 [64].The development of Tr17 is dependent on IL-6/stat3 signalling [65].Stimulation of BMDCs with C. glabrata β-glucan increased the production of IL-6 (figure 4g).However, in our study, the blockade of CR3 enhanced IL-6 production (figure 6d) while reducing the number of Foxp3 + Treg (figure 7b).Another line of evidence suggests that RORγt repressed IL-10 production [66], and loss of IL-10 led to increased RORγt expression [67].Therefore, further study of this RORγt + Foxp3 + Tr17 subpopulation in response to C. glabrata infection is necessary for the development of therapeutic applications.
In this study, we used GM-CSF and IL-4-derived BMDCs to examine DCs in vitro.This culture system is cost-effective, and primary BMDCs have shown better results than DC cell lines.However, GM-CSF-derived BMDCs consist of a mixed population of DCs and macrophages, as reported by Helft et al. [68].Including IL-4 in GM-CSF-derived DC cultures has been shown to reduce the number of macrophages [68], indicating that DCs are likely the primary cell population in GM-CSF and IL-4-derived BMDCs in our system.Nevertheless, further research exploring DC responses to Candida β-glucan in alternative in vitro DC culture systems, such as FLT3L-derived BMDCs, may provide additional benefits and insights into DC-mediated anti-fungal immunity [69].In conclusion, our findings suggest the underlying mechanisms responsible for inducing Treg by systemic C. glabrata infection.In this study, a single strain of each Candida species was selected for investigating the immune response related to inter-species variation of cell wall β-glucan.Nonetheless, intra-species diversity of cell wall β-glucan should be considered and further investigated to confirm the role of C. glabrata β-glucan.A better understanding of immune modulation by C. glabrata may lead to important contributions to the development of new immunotherapeutic strategies for treating invasive C. glabrata infection.

Material and methods
An extended description of the methods can be found in the electronic supplementary material.

Animals
Six-to seven-week-old female C57BL/6 mice were purchased from Nomura Siam International Co. (Bangkok, Thailand).

Cultivation of Candida yeasts
Candida glabrata (ATCC2001) and C. albicans (SC5314) were selected because these reference strains are used for quality control and antifungal drug susceptibility testing.Additionally, the cell wall glucans of both Candida strains have been well characterized [37,39,54].All Candida yeasts were cultured in yeast peptone dextrose (YPD) agar (HiMedia Laboratories, Mumbai, India) at 30°C for 48 h.The yeast cells (5 × 10 6 cells) from the colonies of each strain were subcultured in 15 ml of YPD broth at 30°C with shaking at 180 rpm for 8 h.Subsequently, Candida cultures were diluted to an OD 600 of 0.1 and grown in 1.2 l of YPD broth medium at 30°C with shaking at 150 rpm for 13 h [39].Under these culture conditions, all Candida species grow as budding yeast-like cells [70][71][72].The morphology of all Candida yeasts was confirmed using bright field microscopy (Olympus BX50, Tokyo, Japan).

Preparation of heat-killed yeasts
Heat-killed yeasts were prepared as previously described [43].Candida yeast cells were harvested and washed twice with sterile Dulbecco's phosphate-buffered saline (DPBS; GIBCO, Thermo Fisher Scientific, NY, USA).The yeasts were then diluted to 1 × 10 8 cells ml −1 in sterile DPBS and were inactivated by heating at 100°C for 10 min.The heat-killed yeast cells were collected by centrifugation and resuspended in RPMI 1640 (GIBCO, Thermo Fisher Scientific).

Candida mannan extraction
Cell wall mannans from Candida were extracted as previously described [43,73].Briefly, 100 g of Candida yeast cell pellets were resuspended in 250 ml of citrate buffer (0.02 M, pH 7.0), and the yeast suspension was autoclaved at 121°C for 90 min.After cooling down, the supernatant was separated and collected by centrifugation, and the residual sediment was re-extracted with the same procedures.The supernatants from the first and second autoclaves were combined.Subsequently, an equal volume of Fehling's solution was added, and the mixture was stirred at 4°C for overnight.The sediment of mannan-copper complex was collected by centrifugation and then dissolved in 6-8 ml of 3 N HCl.The solution was then added drop-wise into 100 ml of methanol-acetic acid (ratio: 8 : 1, v/v) to remove the copper.The mannan precipitate was collected by centrifugation and washed with HCl and methanol-acetic acid repeatedly until the precipitate became colourless.The mannan precipitate was dissolved in sterile water and further dialysed in sterile water.The concentration of mannan was determined using the phenol-sulfuric acid method [74].Mannan was kept in the lyophilized form.All yeast culture and mannan extraction procedures were performed using endotoxin-free water and containers.

Candida β-glucan extraction
The preparation of Candida β-glucan microparticles, including extraction, depyrogenation and sterilization, was performed by following the protocol provided by Prof. David L. Williams (East Tennessee State University, Johnson City, TN, USA) [37,54].Briefly, Candida yeasts were resuspended in 0.1 N NaOH and boiled at 100°C for 15 min.After cooling down, the residues were collected by centrifugation.The residue was resuspended in 0.1 N NaOH and the same procedure was repeated two more times (a total of three extractions).The harvested residues were boiled at 100°C in 2NH 3 PO 4 for 15 min (three extractions) and then boiled at 100°C in acidic ethanol (1%v/v of H 3 PO 4 in absolute ethanol) for 15 min (three extractions).The resulting residues were neutralized to a pH of 7.0.The β-glucan microparticles were washed three times with endotoxin-free water.Isolated β-glucan microparticles were depyrogenated in 250 mM NaOH and subsequently neutralized to a pH of 7.0 in 250 mM H 3 PO 4 .The β-glucans were washed in endotoxin-free water (three times) and sterilized by autoclaving at 121°C under 15 psi of pressure for 30 min.The β-glucans were stored in a lyophilized form until later use.
All yeast culture and β-glucan extraction procedures were performed using endotoxin-free water and containers.The three batches of extracted glucans were pooled and used for all experiments.The concentrations of β-glucans in all experiments are used as dry weight per volume (w/v).

Structure analysis by NMR spectroscopy
Chemical structure characterization was conducted using a JEOL JNM-ECZ500R/S1 spectrometer (JEOL, Peabody, MA, USA) operating at 500 MHz for 1 H NMR and 126 MHz for 13 C NMR.To perform C. glabrata β-glucan analysis, 50 mg of β-glucan was dissolved in 1.57 ml of DMSO-d 6 and 0.08 ml of TFA-d 1 .Approximately 0.5 ml of the solution was then withdrawn to perform the NMR experiment at room temperature.NMR chemical shifts were referenced to the residual DMSO-d 6 proton at δ 2.54 ppm and carbon at δ 40.45 ppm.

Systemic C. glabrata infection and ex vivo re-stimulation
Mice were intraperitoneally administered dexamethasone (0.1 mg g −1 of body weight; Dexton-Vet, T. P. Drug Laboratories (1969), Bangkok, Thailand) 3 consecutive days before infection and at 5 days post-infection [39,75].On day 0, the mice were intravenously inoculated with 5 × 10 6 , 10 × 10 6 and 30 × 10 6 of C. glabrata yeast cells in 100 µl of DPBS.In the control group, the mice received 100 µl of DPBS.On day 7 post-infection, serum and spleens were harvested from all control and infected mice.
In the ex vivo re-stimulation assay, 200 µl of 10 µg ml −1 of purified anti-mouse CD3 antibody (clone 145-2C11; BioLegend) was coated on a 24-well plate at 4°C overnight.The anti-CD3 solution was then removed, and the coated wells were rinsed with culture medium twice.The splenocytes were then seeded at a concentration of 4 × 10 6 cells per well in 1 ml of complete RPMI medium containing 55 µM 2-mercaptoethanol (2ME; GIBCO).Forty-eight hours after the stimulation, culture supernatants were collected for cytokine quantification, and the cells were harvested and stained for CD3, CD4, CD25, Foxp3 and IL-10.

In vivo immunization and ex vivo re-stimulation
In the scruff of the neck, mice were subcutaneously administered C. glabrata mannan or β-glucans (50 µg mannan or β-glucans per 1 g of body weight) mixed with OVA (Grade V; Sigma Aldrich, Saint Louis, MO, USA) (30 µg per mouse) in 200 µl of DPBS on days 0 and 7.The positive control mice were subcutaneously administered 200 µl of CFA and incomplete Freund's adjuvant (IFA) mixed with OVA on days 0 and 7, respectively.The negative control mice received OVA in DPBS.
On day 14, the draining LNs (cervical, axillary and brachial LNs) were excised and digested with 300 U ml −1 of collagenase type IV (Sigma Aldrich) and 10 U ml −1 of DNase I (Sigma Aldrich) at 37°C with shaking at 1500 rpm for 30 min.The cells were washed and resuspended in complete RPMI medium.
In the ex vivo re-stimulation assay, LN cells (2 × 10 6 cells) were cultured in 48-well plates in 0.5 ml of complete RPMI medium containing 55 µM 2ME in the presence of 250 µg ml −1 OVA.Forty-eight hours after the stimulation, the culture supernatant was collected for cytokine quantification, and the cells were harvested and stained for CD3, CD4, CD25 and Foxp3.

Generation and stimulation of BMDCs
BMDCs were generated in vitro as previously described [39,73].Briefly, bone marrow cells were obtained from mouse femurs and tibias.The cells (1 × 10 6 1 ml −1 ) were seeded in 24-well plates and cultured in a complete RPMI medium supplemented with 10 ng ml −1 of recombinant murine GM-CSF (Peprotech, Rocky Hill, NJ, USA) and 10 ng ml −1 of recombinant murine IL-4 (Peprotech).The cells were incubated at 37°C under a humidified atmosphere containing 5% CO 2 for 7 days.Half a volume of the media was replaced with the fresh media containing the cytokines every 2 days.
BMDCs were stimulated with heat-killed Candida, Candida mannans or β-glucans at various concentrations, as indicated in the figure legends.Twenty-four and 48 h after the stimulation, the culture supernatants were collected for cytokine measurements, and the cells were harvested for assessing DC maturation markers.Unstimulated BMDCs were used as the negative control.
Twenty-four and 48 h after the stimulation, the culture supernatants were collected for cytokine quantification, and the cells were harvested for assessing DC maturation markers.4.12.In vitro direct co-culture of DCs and T cells BMDCs were untreated or pre-treated with 25 µg ml of a dectin-1 antagonist or 10 µg ml −1 of a purified anti-mouse/human CD11b antibody for 2 h.The cells were subsequently stimulated with 25 µg ml −1 of Candida β-glucans for 24 h.The BMDCs were collected and washed with completed RPMI medium twice.The BMDCs (1 × 10 5 cells) were co-cultured with T cells (1 × 10 6 ) isolated from spleens of intact mice using immunomagnetic beads (Pan T Cell Isolation Kit II, mouse; Miltenyi Biotec, San Diego, CA, USA) at a DC : T-cell ratio of 1 : 10 in the presence of 30 ng ml −1 of soluble anti-mouse CD3 antibody (clone 145-2C11, BioLegend) in 48-well plates [31,32,39].At 48 h, the culture supernatants were collected for cytokine quantification, and the cells were stained for CD3, CD4, CD25 and Foxp3.
To determine BMDCs and T cells, live cells were gated and the cells were acquired at 20 000 cells per sample.With regard to LN cells and splenocytes, live cells were gated and the cells were acquired at 50 000 cells per sample.Accordingly, the same electronic gate was applied to the same set of samples for flow cytometric acquisition and analyses.

Figure 10 .
Figure 10.Binding of C. albicans and C. glabrata β-glucan to dectin-1 and CR3.RAW264.7 cells were incubated with various amounts of C. albicans or C. glabrata β-glucan for 20 min.After washing, the cells were stained with fluorescent-tagged anti-dectin-1 mAbs or fluorescent-tagged anti-CD11b mAbs.The expression of dectin-1 and CR3 was determined by flow cytometry (n = 3).The percentage binding was calculated as described in the Materials and methods section.(a) β-glucan and dectin-1 binding analysis and a flow cytometric histogram of dectin-1 expression.(b) β-glucan and CR3 binding analysis and a flow cytometric histogram of CR3 expression.

4. 13 .
Dectin-1 and CR3 binding assay RAW264.7 cells were incubated with various amounts of C. albicans or C. glabrata β-glucan for 20 min, and the cells were washed twice.Subsequently, the cells were stained with fluorescent-tagged anti-dectin-1 mAbs or fluorescent-tagged anti-CD11b mAbs.The expression of dectin-1 and CR3 was determined by flow cytometry.The percentage binding was calculated by the following formula: 100 − (MFI of sample treated with β-glucan × 100 per MFI of untreated sample).