Binding of live conidia of Aspergillus fumigatus activates in vitro-generated human Langerhans cells via a lectin of galactomannan specificity

Abstract

Aspergillus fumigatus is the most common aetiological fungus responsible for human pulmonary aspergilloses. This study investigated the primary contact between Langerhans cells (LC), corresponding to dendritic cells present in pulmonary mucosa and live conidia of A. fumigatus. LC play a key role in antigen presentation for initiation of the primary T cell response. In vitro-generated LC (iLC) were differentiated from cultured human cord blood CD34⁺ cells and incubated at 4°C or 37°C with fluorescein-isothiocyanate (FITC)-stained conidia or control latex beads. In vitro, conidia were shown by microscopy and cytometry to adhere to iLC in a dose- and time-dependent manner. This adhesion was not limited to iLC because interstitial dendritic and other cells also fluoresced in the presence of conidia-FITC. A lectin other than mannose receptor-type lectin was demonstrated to be responsible of conidial binding. Inhibition of binding was observed with heterologous galactomannan and EDTA, indicating a C-lectin-like receptor with galactomannan structure specificity. After binding only a few conidia were internalized in acidic vesicles, as indicated by the cessation of conidial fluorescence. Conidial binding was followed by activation and maturation of iLC, suggesting that LC present in the lung may play a role in cellular host defence against aspergilloses.

INTRODUCTION

Aspergillus fumigatus is now the most prevalent airborne fungal pathogen and is responsible for approximately 90% of human cases of aspergillosis, causing severe and usually fatal invasive infections in immunocompromised hosts [1]. The conidia of A. fumigatus that are released into the atmosphere have a diameter small enough (2–3 µm) to reach the human lung alveoli. Most of the conidia are probably excluded from the lungs through the ciliary action of the mucous epithelium, but some are internalized by epithelial and endothelial cells and serve as putative foci of infection [2]. Conidia may also interact with surfactant proteins, basement membrane laminin, fibrinogen or phagocytic cells [3]. These phagocytic cells, such as pulmonary alveolar macrophages for conidia or neutrophils for hyphae, have long been recognized as being critical in the host defence against A. fumigatus infections. The key role of cytokines in the immune response against Aspergillus has also been reported [4–6].

Other cells such as Langerhans cells (LC) are essential in immune response because they are dendritic cells (DC) derived from bone-marrow progenitors and induce the primary immune response [7]. Langerhans cells are localized in suprabasal layers of the skin and in mucosa, in particular in pulmonary mucosa. They represent a subset of DC that is characterized by the presence of cytoplasmic organelles termed Birbeck granules associated with Lag antigen and Langerin (CD207), a new molecule that has been described recently [8]. LC play a key role in antigen presentation for initiation of the primary T cell response. In vivo, LC take up antigens and migrate subsequently via lymphatic or blood vessels towards the T cell-rich zone of secondary lymphoid tissues, and subsequently are called interdigitating cells. During migration, they undergo an additional maturation step and can present the processed antigen to naive T cells [7]. These cells may be at the origin of the Th1-type cytokine response induced by conidia of A. fumigatus [6]. However, only a few recent data have been reported on the role of dendritic cells against Aspergillus [9,10].

As LC are present in the mucosa, it seemed interesting to study the potential role of LC in the host defence against A fumigatus. Because the isolation of LC from mucous membranes is a tedious process, during which only a small number of cells are obtained, we chose to use in vitro-generated human LC (iLC) from cord blood CD34⁺ progenitors [11], as these cells are known to have the characteristics of LC.

In this study, we investigated the primary contact between human LC generated in vitro from cord blood CD34⁺ progenitors (iLC) and live fluorescein-isothiocyanate (FITC)-stained conidia of A. fumigatus. In a system devoid of contaminating cells, we found that conidia adhere to iLC using a lectin-like receptor other than the mannose receptor and that only small amounts of conidia were phagocytosed and found in acidic pH vesicles. However, this binding was sufficient to induce activation of iLC, as shown by changes in the expression of various markers.

MATERIALS AND METHODS

Preparation of conidia and labelling with FITC

A. fumigatus was isolated from a bronchial aspiration of a patient with an aspergilloma. Conidia were stored frozen at −30°C until inoculation on malt agar then incubated at 37°C for 3 days. Conidia were harvested by washing the slant culture with 0·5 ml of a 0·05% Tween 80 solution per culture tube and scraping the conidia gently from the mycelium with a glass pipette. After addition of 4 ml sterile 0·9% NaCl solution, the suspension was filtered through 25-µm pore-size nylon filters in order to separate conidia from contaminating mycelium. The conidial suspension was centrifuged and prepared for immediate labelling.

Live conidia were suspended in 0·1 m carbonate buffer pH 9 at 1 × 10⁸ per ml then incubated with FITC (F-4274; Sigma Chemical Co., St Louis, MO, USA) at a final concentration of 0·16 mg/ml. After overnight incubation at 4°C [12], conidial solutions were diluted and washed twice in phosphate buffer saline (PBS; pH 7·2). The conidial pellets were then counted, diluted in PBS to the desired concentration and stored at 4°C in the dark.

As control, microbeads with a size comparable to conidia were coupled to FITC (polystyrene latex beads 3 µm diameter, LB-30; Sigma).

Purification and culture of cord blood CD34⁺ haematopoietic cells and culture of other human cells

Cord blood was collected according to institutional guidelines during normal full-term deliveries. Mononuclear cells were isolated by flotation on Lymphoprep (Nycomed Pharma AS, Oslo, Norway), and then depleted of adherent cells by overnight culture on a plastic surface. CD34⁺ cells were purified by immunomagnetic selection with mini-MACS (Miltenyi-Biotec GmbH, Bergisch Gladbach, Germany). This procedure regularly gave a suspension containing more than 95% of CD34⁺ cells. Isolated progenitors were cultured in RPMI-1640 (Gibco-BRL Laboratories, Grand Island, NY, USA) containing Glutamax-1, 25 mm HEPES and supplemented with 5% heat-inactivated fetal calf serum (FCS Myoclone super plus; Gibco-BRL Life Technologies, Cergy Pontoise, France), 5 × 10⁻⁵m 2-mercaptoethanol (M-6250; Sigma), antibiotics (penicillin–streptomycin–amphotericin B, A-5955; Sigma), 200 U/ml of recombinant granulocyte-macrophage colony-stimulating factor (GM-CSF, specific activity 2 × 10⁶ U/mg, kindly provided by Schering Plough Laboratories, Dardilly, France) and 50 U/ml recombinant human tumour necrosis factor-α (TNF-α, specific activity 2 × 10⁷ U/mg; Genzyme Corp., Boston, MA, USA). At day 6, the TNF-α in the culture medium was replaced by 0·5 ng/ml transforming growth factor-β 1 (TGF-β1; R&D, Abingdon, UK) and cells were used 48 h later.

Interstitial-like DC defined as CD14⁺ and CD1a^– cells at day 6 of differentiation were isolated by immunoselection, as described previously [13,14], and cultured until use with GM-CSF and TNF-α. Control cultures showed more than 90% of cells CD14⁺, CD1a⁺ and Factor XIIIa⁺ at day 12 [11].

Cell lines were cultured in RPMI-1640 supplemented with 10% heat-inactivated FCS. T lymphocytes and monocytes were isolated from the peripheral blood of healthy volunteers. Mononuclear cells were isolated by flotation on Lymphoprep. After overnight culture on a plastic surface, adherent cells represented a monocytes-rich fraction. T lymphocytes were isolated from non-adherent cells by rosetting with sheep red blood cells, as described previously [15].

Phenotypic analysis of DC

Cells were washed with PBS containing 1% bovine serum albumin (BSA), 0·6% acid citrate–dextrose and 0·02% sodium azide (PBS/BSA/ACD/azide); 1 × 10⁵ cells were then incubated with 10 µl labelled MoAb for 30 min at 4°C. The MoAb used were CD83 (HB15A, IgG2 k) and isotypic controls (IgG1, IgG2a and IgG2b) from Immunotech (Marseilles, France), CD86 (B-T7, IgG1) from Diaclone (Besançon, France), HLA-DR (L243, IgG2a) and CCR6 (2H4, IgM) from Becton Dickinson, CCR7 (53103·111, IgG2B) from R&D. CD14 (TÜK4, IgG2κ) and CD1a (NA1/34, IgG2a) from Dako (Trappes, France), E-cadherin (HECD-1, IgG1) from Takara (Japan) and Factor XIIIa (polyclonal) from Behring. Anti-Langerin MoAb (CD207, clone DCGM4) and antimannose receptor (clone DCGM1) kindly provided by Dr Sem Saeland (Schering-Plough Research Laboratories, Dardilly, France). Using unlabelled antibodies, we added a second step including GAMIg-FITC (Zymed Laboratories, San Francisco, CA, USA). After washing in PBS/BSA/ACD/azide, cells were fixed with 1% formaldehyde in PBS/BSA/ACD/azide. Analysis was performed on a FACScan flow cytometer (Becton Dickinson, Mountain View, CA, USA). Cell size was estimated by light diffraction at small angles (FSC).

Binding assay of conidia-FITC on DC

Cell suspensions were washed and cultured in culture medium without amphotericin B for 24 h before use. Aliquots of 100 µl of cell suspension containing 0·5 × 10⁶ cells were distributed in culture tubes (Falcon 35 2054) and 100 µl conidia-FITC were then added. Incubation was performed at 37°C or 4°C under various duration and concentration conditions. After incubation, the cells were fixed by addition of 400 µl ice-cold PBS containing 1% FCS, 0·01% Na₃N, 0·6% acid citrate–dextrose and 1% formaldehyde. As control, a set of tubes was fixed immediately after addition of conidia-FITC.

Inhibition of conidia-FITC uptake was studied using glucose (Sigma, G7528), galactose (Sigma, G0625), fucose (Sigma, F8150) mannose (Sigma, M8296), N-acetyl-D-glucosamine (Sigma, A8625), 4-O- D-galactopyranosyl D-galactopyranose (Sigma, G9662), 6-O-β D-galactopyranosyl D-galactose (Sigma, G5643), mannan from Saccharomyces cerevisiae (Sigma, M7504), galactomannan polysaccharide from seeds of Ceratonia siliqua (Sigma, G0753), dextran (Sigma, D4751) and EDTA (Life Technologies, Gaithersburg, MD, USA, 15575–017). Cells were incubated with inhibitors for 1 h at 37°C before addition of conidia-FITC. After an additional step of 1 h of incubation with conidia, cells were fixed and fluorescence analysis was performed on a FACScan flow cytometer (Becton Dickinson).

Intracellular localization of conidia-FITC was revealed by addition of 50 µm monensin (Sigma, M5273) after FACScan analysis. After 30 min of incubation at 4°C, cells were analysed again by FACScan flow cytometry.

Inhibition of conidium-FITC binding was studied by incubation of LC with an excess of anti-CD207, DCGM1, anti-CD1a or an isotype control for 30 min at 8°C before contact with conidia-FITC. The binding of MoAb was assessed by addition of GAMIg-RPE-Cy5 (DAKO, Trappes, France).

Binding of dextran-FITC

Samples of 0·5 × 10⁶ cells were washed in RPMI/FCS, centrifuged and 20 µl dextran-FITC (Dx-FITC, MW 40 000, ref D-1845; Molecular Probes, Eugene, OR, USA) at 100 µg/ml were added to the cell pellet. Incubation was performed at 37°C or 4°C for 15 min then cells were washed three times with ice-cold PBS containing 1% FCS, 0·01% Na₃N and 0·6% acid citrate–dextrose, and then fixed with 1%PBS-formaldehyde.

Inhibition of Dx-FITC uptake was studied using the same reagents as those used for inhibition of conidial binding; cells were also preincubated with reagents for 1 h at 37°C.

Fluorescence analysis was performed on a FACScan flow cytometer (Becton Dickinson).

RESULTS

Conidia bind to iLC in a dose- and time-dependent manner

Conidial binding was studied using iLC taken between days 8 and 10 of differentiation and sequential dilutions of conidia labelled with FITC (conidia-FITC) incubated for 60 min at 37°C. As a control, the same experiment was performed at 4°C (Fig. 1a). Conidial binding was greater at 37°C than at 4°C, suggesting an active mechanism (Fig. 1c). The percentage of cell binding varied within the experiments from 30 to 70% over 1 h at 37°C, but was dose-dependent at 37°C; calculation of the amounts of markers per cell indicated higher binding efficiency with higher concentrations of conidia (Fig. 1b).

Fig. 1.

Incubation of FITC-labelled A. fumigatus conidia or FITC-latex beads with iLC at 37°C and 4°C. The experiments were performed with different conidia : cell ratios or with different incubation times, and results are expressed either as a percentage of labelled cells (a, c, respectively) or as mean fluorescence intensity (b, d). Incubations at 6 and 24 h were performed with the addition of amphotericin B to the medium to avoid conidial germination. Data represent a typical experiment from a set of three.

At a ratio of 10 conidia per iLC, binding was time-dependent with a plateau over 1 h of incubation at 37°C (Fig. 1c). The mean fluorescence intensity (MFI) confirmed that saturation of iLC required 60 min (Fig. 1d). Control microbeads of comparable volume stained by FITC did not bind significantly to iLC even after 24 h of contact (Fig. 1c, d). Microscopic observation confirmed that microspheres did not bind significantly, as the percentage of iLC associated with microbeads was below 5% and that most of the cells did not bind more than one microsphere (Fig. 2a). In contrast, conidia bound rapidly to iLC and conidia were bound to dendrites (Fig. 2b, c) and to the cell body (Fig. 2d, e). The number of conidia per cell increased progressively with duration of incubation; furthermore, within 15 min iLC formed aggregates of cells and conidia, with the number of cells involved in aggregates increasing with time (Fig. 2f, g). This phenomenon did not interfere with cytofluorometric measurements, as aggregates of iLC were dissociated easily after agitation without apparent loss of bound conidia.

Fig. 2.

Microscopic examination of iLC after incubation with FITC-labelled latex beads (a) or FITC-labelled A. fumigatus conidia (b, c, d, e, f, g) at 37°C for 1 h (b, c, d, e) or 6 h (f and g).

Conidia were rarely internalized in acidic vesicles after incubation with iLC

The decrease in the percentage of labelled cells and MFI after long periods of incubation (Fig. 1c, d) suggested the possibility of cell internalization of conidia with quenching of FITC or receptor modulation.

After 1 or 2 h of incubation with conidia-FITC, cell suspensions were treated with monensine in order to neutralize the intraendosomal and intralysozomal pH and consequently suppress FITC acid-pH quenching. As shown in Fig. 3, treatment with monensine enhanced to only a small extend the number of fluorescent cells and MFI, indicating that only a small proportion of conidia-FITC were located in acidic vesicles.

Fig. 3.

Effect of monensin on the percentage of labelled iLC. After a first acquisition step on FACS (white areas), monensin was added to the tubes and a new acquisition step was performed 30 min later (grey areas). Data represent the result of one experiment from a set of three.

Microscopic observation of the corresponding suspensions using an objective (×40) that can be used in both phase contrast and epifluorescence confirmed that most conidia-FITC remained associated with the cell membrane. iLC with intracytoplasmic conidia were observed only rarely; even after incubation periods up to 24 h, the percentage of iLC containing at least one conidia remained below 5%. Membrane-associated conidia were fluorescent, whereas internalized conidia were no longer fluorescent suggesting phagocytosis via an acidic vesicle (data not shown). These observations indicate that iLC presented limited potential of phagocytosis for live conidia.

Conidium-FITC binding capacity is not restricted to iLC

We used iLC generated from CD34⁺ haematopoietic cells in the presence of granulocyte-macrophage colony-stimulating factor (GM-CSF) and TNF-α for 6 days then cultured with GM-CSF and TGF-β. At day 8, cells had a phenotype of immature LC (CD1a⁺, CD83^low, CD86^low, E-cadherin⁺ and CD207⁺), as shown in Table 1. For pure LC and DC suspensions, at day 5, cells were separated according to their CD1a status [16]. After a further step of culture in medium supplemented with GM-CSF and TGF-β for CD1a⁺ cells or medium supplemented with GM-CSF and TNF-α for CD1a^– cells, at day 8, one population had the phenotype of enriched LC and the other a phenotype of interstitial DC (Table 1). Both cell types bound conidia, as shown in Fig. 4. HaCat, a cell line derived from keratinocytes, and B cell lines derived from Burkitt's lymphoma such as Jiyoye and RPMI were very efficient in binding conidia. Cord blood monocytes and THP1, a monocyte-derived cell line, also bound conidia but to a lesser extent, whereas T lymphocytes did not.

Table 1.

Phenotypes of iLC at day 10 of differentiation, and LC and DC obtained after cell fractionation at day 5 and cultured until day 10 in appropriate culture medium. Results represent the mean of three to five experiments; s.d. is presented in brackets

	iLC	LC	DC
CD1a	78·4 (8·3)	77·7 (12)	46·9 (9·6)
CD83	7·12 (3·0)	12·4 (7·4)	26·1 (6·1)
CD86	6·05 (5·2)	14·2 (7·4)	15·6 (3·1)
E-cadherin	58·1 (5·3)	45·7 (14)	7·78 (3·2)
CD207	53·6 (9·4)	54·8 (13)	6·90 (2·0)

Fig. 4.

Incubation of FITC-labelled A. fumigatus conidia for 1 h with different human cells. In vitro-generated Langerhans cells (iLC), in vitro-generated interstitial DC (iDC); HaCaT, a cell line derived from keratinocytes; BL60, Jiyoye and RPMI: B cell lines derived from Burkitt's lymphoma; cord blood T lymphocytes or monocytes; THP1, a monocyte-derived cell line. Open bars correspond to incubation at 37°C without conidia-FITC, black bars to incubation at 37°C with conidia-FITC and grey bars to incubation at 4°C with conidia-FITC. Another experiment was performed, with comparable results.

The binding structure on the cell membrane is a lectin with galactomannan specificity

As A. fumigatus is known to have a cell wall rich in polysaccharides, particularly in galactomannans, the binding between conidia and iLC was analysed by using sugar inhibition experiments. Common sugar molecules such as galactose, glucose, mannose, N-acetylglucosamine, which are constituents of conidia [17,18], and fucose, implicated in binding with the mannose receptor, had no effect on conidial binding. However, they inhibited partly the binding and internalization of dextran (Dx)-FITC which was used as a ligand for mannose receptor detection, with mannose being the more efficient (Fig. 5). Disaccharides, that were used to test whether a particular link between sugar residues was involved, had no effect. Moreover, polysaccharides such as yeast mannan and control dextran were more efficient in inhibiting Dx-FITC binding, despite the smaller amount of competitor added, but had no effect on conidial binding. Conidial binding was inhibited only by plant galactomannan, which is the least efficient polysaccharide for inhibition of Dx-FITC binding. Altogether, these data indicated that the cell receptor used for conidium-FITC binding was a lectin of galactomannan specificity, different to the well-known mannose receptor. Treatment of cells with EDTA inhibited Dx-FITC and conidium-FITC binding, suggesting that both cell receptors were C-type lectins.

Fig. 5.

Comparative inhibition of binding of conidia-FITC and Dx-FITC to iLC in the presence of various saccharides. Saccharides or EDTA were added to the medium for 1 h at 37°C before addition of conidia. With complex sugars such as mannan, galactomannan and dextran, serial dilution were tested and the presented data correspond to the higher concentration assayed. Mannan was expressed as µg/ml because the molecular mass of the molecule cannot be determined. Results are representative of at least two experiments.

In order to exclude definitively the possibility of binding of conidia-FITC by mannose receptors, iLC were incubated with saturating amounts of DCGM1, a monoclonal antibody (MoAb) that recognizes the human LC mannose receptor [19], DCGM4, a MoAb that recognizes CD207, or unrelated MoAb such as anti-CD1a or an IgG1 isotype control. After MoAb saturation, conidia-FITC, Dx-FITC or GAMIg-RPE Cy5 were added and, as shown in Fig. 6, the presence of MoAbs did not impair the binding of conidia-FITC. However, as expected, the binding of Dx-FITC was inhibited by DCGM1 saturation.

Fig. 6.

Comparative inhibition of binding of conidia-FITC and Dx-FITC to iLC in the presence of different antibodies. iLC were saturated by MoAbs before the addition of Dx-FITC, conidia-FITC or GAMIg-RPE-Cy5.

Conidial binding triggered iLC maturation

iLC were cultured for 48 h in the presence of amphotericin B at 25 µg/ml after initial contact with live conidia. Phenotypic analysis revealed maturation of iLC, as shown by the appearance of CD83 and a decrease in CD1a and CD207 expression (Fig. 7). This maturation was associated with loss of CCR6 required for epithelial localization of LC and the appearance of CCR7 required for migration towards lymph nodes. HLA-DR and CD86 overexpression suggested that iLC developed the ability to activate T lymphocytes.

Fig. 7.

Phenotypic analysis of iLC after 48 h incubation with A. fumigatus conidia. Grey area represents iLC incubated with conidia, and open area control without conidia. Another experiment was performed giving identical results.

DISCUSSION

This study described for the first time the active adhesion of live A. fumigatus conidia on in vitro-generated human LC (iLC). This adhesion was followed by rare phagocytosis and by activation of iLC enabling them to migrate and activate T lymphocytes for a specific immune response.

Conidium-FITC binding to iLC was an active phenomenon at 37°C and was specific to conidia, as no binding was observed with control latex beads of comparable size. Conidial binding is not a consequence of FITC labelling as unlabelled conidia also bound to iLC with comparable efficiency (microscopic observation, data not shown). Using human polymorphonuclear cells (PMN), Sturtevant and Latgé [20] reported a higher percentage of conidial binding at 37°C (95% within 45 min with PMN compared with 25% within 60 min with iLC). With both cell types, a negligible association was reported at 4°C in contrast to the result of Kan and Bennet [21].

Phagocytosis of conidia by iLC is a minor phenomenon

By flow cytometry, conidia ingested by iLC could not be directly detected, as microscopic observations indicated that fluorescence of ingested conidia-FITC was quenched in acidic vesicles. Modification of intracellular pH by addition of monensine allowed the detection of the hidden fluorescence corresponding to phagocytosed conidia. The data confirmed low rate of conidial ingestion.

Phagocytosis of A. fumigatus conidia has been described previously with human in-vitro-generated DC using heat-killed conidia in order to prevent fungal overgrowth and allow incubation times of 24–48 h [6]. DC were generated from peripheral blood CD34⁺ progenitors under culture conditions that yielded the equivalent of interstitial DC and not LC. In a previous review, Clemons et al. [22] described an interaction between DC and heat-killed conidia. DC were derived from bone marrow CD34⁺ progenitor cells under culture conditions that might produce some LC, but few details were given; approximately 15% of the DC phagocytosed conidia. All together these data indicated that phagocytosis was probably not the major mechanism by which human LC and DC acquire the antigenic structures required for presentation to T cells.

A recent study [9] demonstrating considerable phagocytosis of conidia in murine pulmonary DC may reflect species differences. Differential rates of conidial phagocytosis were also observed in relation to the A. fumigatus strains used in murine macrophages, or to the cell types [23]. As shown in Fig. 4, cells of various origin can bind conidia; conidial binding has also been described with human non-professional phagocyte cells such as type II pneumocytes [24] and human umbilical cord endothelial cells [25].

ILC bind conidia via a receptor inhibited by galactomannan

Attempts to identify the receptor responsible for conidial binding were made by Kan et al. using murine alveolar macrophages [18] or human monocytes [21], and by Bozza et al. using murine pulmonary cells [9]. The first study of Kan et al. [18] reported a receptor consistent with the mannosyl–fucosyl receptor, and the second [21] reported a β-glucan-inhibitable receptor. The study of Bozza et al. [9] indicated that internalization of A. fumigatus conidia was mediated either by the mannose receptor or by a C-type lectin of galactomannan specificity or by CR3 (to a smaller extent). In our study, we confirmed the presence of a C-type lectin of galactomannan specificity as shown by its inhibition in the presence of the same galactomannans from Ceratonia siliqua; however, we have excluded binding by a mannose receptor. Another mechanism of binding was suggested by the glycosphingolipid-mediated attachment of A. fumigatus conidia to pneumocytes [24]. Recent data published by Garlanda et al. [26] indicated that binding of the long pentraxin, PTX3, on conidia is abolished in the presence of galactomanan. This suggests an affinity of this pentraxin for galactomanan residues. However, since the kinetics of the synthesis of PTX3 by DC and monocytes requires an inflammatory background, it would appear that PTX3 is not implicated in our results.We cannot exclude the possibility that the differences observed correspond to the type of used cells which implies that various cell types use different mechanisms of binding.

Until recently, it was agreed that two lines of defence are responsible for the innate immunity of the host against A. fumigatus: pulmonary alveolar macrophages and peripheral blood phagocytes [2]. In mice susceptible to invasive aspergillosis, interleukin (IL)-4 produced by CD4⁺ T lymphocytes impaired antifungal activity of neutrophils [4]. In another study, it was reported that the interaction between neutrophils and soluble β-glucan of the yeast Saccharomyces cerevisiae, enhanced the antimicrobial functions of neutrophils [27].

Conidia binding triggers iLC activation

The functions of DC in immune response were recently studied in two yeasts, Cryptococcus neoformans [28] and Candida albicans [29]. The first study in an animal model showed that certain DC of lymph nodes were needed for induction of the protective response after immunization with soluble antigen but not with killed yeast. In the second study, immature murine myeloid DC rapidly and efficiently phagocytosed both yeasts and hyphae of the fungus. In vitro, yeasts activated DC for Th1 priming while hyphae escaped the phagosome, inhibited Th1 priming and induced the production of IL-4. Similar results were obtained in an animal model with A. fumigatus conidia, swollen conidia and hyphae [9]. Pulmonary DC were shown to transport conidia or hyphae to the draining lymph nodes and spleen, without fungal growth, suggesting that conidia and hyphae may have undergone degradation for efficient antigen presentation by DC. They produced IL-12 in response to conidia and IL-4 and IL-10 in response to hyphae. Expression of MHC class II and CD80 and CD86 molecules were all greatly increased in DC from spleen and particularly in those from lymph nodes 6 h after infection. Therefore, pulmonary DC undergo functional maturation upon migration from the airways to the local and peripheral lymphoid organs in mice with aspergillosis.

In vitro studies with human DC [6] with heat-killed A. fumigatus conidia showed an increase in mean fluorescent intensity for HLA-DR, CD80 and CD86 despite an unmodified percentage of fluorescent cells. Our results with live conidia indicated activation of iLC using a larger set of markers with an enhanced MFI, and an increased percentage of labelled cells; the latter result is related to the live status of conidia.

Our data showed that infected human iLC down-regulate CD1a and CD207 but up-regulate CD83, indicating the acquisition of a mature stage. At the same time, they down-regulate CCR6 and up-regulate CCR7. These events, shown for the first time with A. fumigatus conidia and human iLC, were expected in the occurrence of LC migration towards lymph nodes. Finally, human LC overexpress HLA-DR molecules and CD86 suggesting that infected LC may activate T lymphocytes to induce a primary immune response. In contrast it would be interesting to determine whether the fungus can to stop this response, since it is known that A. fumigatus produces molecules that can interfere with host-response, such as antiphagocytic factors [3,30]. These studies reveal novel insights into the key role of LC in the immune host-response to A. fumigatus, providing another cell intermediate in the defence against fungal infection.