In their mammalian hosts, Leishmania are obligate intracellular parasites that mainly reside in macrophages. They are also phagocytosed by dendritic cells (DCs), which play decisive roles in the induction and shaping of T cell-dependent immune responses. Little is known about the role of DCs in the Leishmania life cycle. Here, we examined the ability of mouse bone marrow-derived DCs to serve as hosts for L. amazonensis. Both infective stages of Leishmania (metacyclic promastigotes and amastigotes) could be phagocytosed by DCs, regardless of whether they had previously been experimentally opsonized with either the complement C3 component or specific antibodies. Parasites could survive and even multiply in these cells for at least 72 hours, within parasitophorous vacuoles displaying phagolysosomal characteristics and MHC class II and H-2M molecules. We then studied the degree of maturation reached by infected DCs according to the parasite stage internalised and the type of opsonin used. The cell surface expression of CD24, CD40, CD54, CD80, CD86, OX40L and MHC class II molecules was barely altered following infection with unopsonized promastigotes or amastigotes from nude mice or with C3-coated promastigotes. Even 69 hours post-phagocytosis, a large proportion of infected DCs remained phenotypically immature. In contrast, internalisation of antibody-opsonized promastigotes or amastigotes induced DCs to mature rapidly, as shown by the over-expression of costimulatory, adhesion and MHC class II molecules. Thus, in the absence of specific antibodies (e.g. shortly after infecting naive mammals), infected DCs may remain immature or semi-mature, meaning that they are unable to elicit an efficient anti-Leishmania T cell response. Absence of DC maturation or delayed/incomplete DC maturation could thus be beneficial for the parasites, allowing their establishment and amplification before the onset of immune responses.
Leishmania are protozoan parasites with a digenetic life cycle involving blood-feeding phlebotomine insect vectors and mammals. Some Leishmania species are also the aetiological agents of major human diseases known as leishmaniases. In insects, Leishmania adopt a flagellated promastigote (Pm) form and initially multiply in the lumen of the abdominal midgut. The resulting metacyclic Pm (MPm), which are infective for mammals, accumulate in the anterior parts of the digestive tract from where they can be injected into the dermis of mammals during a blood meal. Parasites surviving in this environment are rapidly phagocytosed, mainly by resident macrophages (for a review, see Peters and Killick-Kendrick, 1987). Once inside macrophages, MPm slowly differentiate into amastigotes (Am) (Courret et al., 2001) within parasitophorous vacuoles (PVs) displaying the properties of phagolysosomes (Antoine et al., 1998; Courret et al., 2002). Although macrophages are generally considered to be the main or only host cells allowing Leishmania to survive and to multiply, other cells can harbour intact/viable parasites at least transiently and may play important roles in the parasitism or in the control of parasite growth. For example, L. major parasites have been observed in polymorphonuclear leukocytes, fibroblasts and dendritic cells (DCs) from BALB/c and C57BL/6 mice at different stages of infection [(Beil et al., 1992); Y. Belkaid and G. Milon, unpublished data cited by Tacchini-Cottier et al. (Tacchini-Cottier et al., 2000) Bogdan et al., 2000; Moll et al., 1995].
DCs form a family of leukocytes that play critical roles in the innate and adaptive immune systems as, (i) they are specialised in the recognition of pathogens through pattern recognition receptors, a process that can lead to their activation; (ii) they ensure the transport of the antigens they capture from the peripheral tissues to the draining secondary lymphoid organs and (iii) they are apparently the only antigen-presenting cells able to prime naive specific T lymphocytes (Banchereau et al., 2000). However, different differentiation stages and lineages are involved in the induction of different types of T cell-dependent immune response or T cell tolerance. For example, immature, non-activated antigen-loaded DCs induce anergy of specific T cells or the development of regulatory T cells that prevent the activation of T effector cells. At this stage, the plasma membranes of DCs display few MHC class II molecules and no, or very few, co-stimulatory molecules. In contrast, mature, activated, antigen-containing DCs, which display high levels of MHC class II and costimulatory molecules on their cell surfaces, are potent inducers of T cell immunity. Interestingly, a third differentiation stage, called semi-mature, was recently identified. Semi-mature DCs have a very similar surface phenotype to mature DCs, but unlike the latter, they do not produce IL-12 (Jonuleit et al., 2001; Steinman and Nussenzweig, 2002; Lutz and Schuler, 2002).
Recent in vivo studies have shown that DCs play different roles during the overall infectious/pathogenic processes in mice infected with Leishmania. For example, shortly after parasite inoculation, skin DCs can take up Leishmania and/or Leishmania antigens before migrating to the draining lymph nodes where they can stimulate Leishmania-specific T lymphocytes (for a review, see Moll, 2000). It must be stressed, however, that this was observed after injecting huge numbers of Pm (106 or more parasites), whereas in natural infection conditions Leishmania-carrying insect vectors generally inject 10-1000 MPm into their vertebrate hosts during their blood meal (Warburg and Schlein, 1986) (M. E. Rogers and P. A. Bates, communication at the Worldleish 2 Congress, Crete, Greece, 2001). At very late stages of infection, mouse DCs harbouring intact Leishmania or containing parasite antigens have also been identified in the lymph nodes draining the sites of parasite inoculation. These cells, but not the lymph node macrophages, which also contain Leishmania parasites/antigens, are able to reactivate Leishmania-specific T cells (Moll et al., 1995). Thus, Leishmania-containing DCs are probably involved in the maintenance of T cell memory/activity.
In vitro experiments have demonstrated that DCs can phagocytose Leishmania. It is unclear whether both parasite stages (Pm and Am) can be efficiently internalised (Blank et al., 1993; von Stebut et al., 1998; von Stebut et al., 2000; Marovich et al., 2000; Bennett et al., 2001) and whether Leishmania can survive within these cells (Konecny et al., 1999; Qi et al., 2001). It is also not clear whether parasites induce, delay or modify the DC maturation process. Most studies indicate that L. major (Am and/or Pm stages) activates mouse DCs (Flohé et al., 1998; von Stebut et al., 1998; von Stebut et al., 2000; Konecny et al., 1999; Henri et al., 2002; McDowell et al., 2002). Other studies have shown that the phagocytosis of certain Leishmania species does not lead to DC maturation (Bennett et al., 2001; McDowell et al., 2002). Further studies are needed to determine whether these discrepancies are due to the type of DCs used, to the Leishmania species, or the origin of the parasites.
In this study, we examined the ability of mouse bone marrow-derived DCs to internalise the different stages of L. amazonensis and we assessed the degree of maturation reached by infected DCs according to the parasite stage phagocytosed. Additionally, we studied the effect of several parasite-bound opsonins, such as complement components and specific antibodies (Abs), on the DC-Leishmania interactions. We looked at the effects of opsonins because, in physiological conditions, at least some of them probably rapidly interact with the cell surface of Pm after their inoculation into the dermis of mammals or with that of Am when they leave their host cells (Schmunis and Herman, 1970; Pearson and Steigbigel, 1980; Navin et al., 1989; Pearson and Roberts, 1990; Guy and Belosevic, 1993; Peters et al., 1995; Nunes and Ramalho-Pinto, 1996; Hodgkinson and Soong, 1997; Mosser and Brittingham, 1997; Domínguez and Toraño, 1999; Domínguez et al., 2002). We show that DCs internalise both the metacyclic and Am stages of L. amazonensis, but that the degree of maturation reached by the infected DCs, as measured by the level of expression of costimulatory and adhesion molecules at the cell surface, varies according to the type of opsonins covering the parasites. Furthermore, we also found that internalised parasites can survive for at least several days in DCs, within organelles that have been partially characterised.
Materials and Methods
Female Swiss nu/nu, DBA/2J and BALB/c mice, aged between 8- and 12-weeks old, were obtained from Charles River (St Germain-surl'Arbresle, France) and Janvier (Le Genest-St-Isle, France).
Parasites and bacteria
Amastigotes of L. amazonensis, strain LV79 (MPRO/BR/1972/M1841), were purified from footpads of infected nude or BALB/c mice (Antoine et al., 1989). Hereafter, these parasites will be referred to as Am nude and Am B/c, respectively. Stationary phase Pm (SPm) and MPm were obtained from lesion-derived Am cultured at 26°C (Courret et al., 1999). Metacyclic forms were purified by negative selection using the monoclonal antibody (mAb) 3A1 (Courret et al., 1999) or by a modified version of the density gradient centrifugation method originally described by Späth and Beverley (Späth and Beverley, 2001). Hereafter, these parasites will be called MPm 3A1 and MPm Ficoll, respectively. Mycobacterium bovis BCG (bacille Calmette Guérin vaccine) was grown in Sauton medium, recovered as previously described (Gheorghiu et al., 1988) and stocked at -80°C until use.
Antibodies for fluorescence microscopy and flow cytometry
The following biotinylated mAbs were purchased from Pharmingen (San Diego, CA): 2D7 (anti-CD11a/LFA-1 α-chain), M1/70 (anti-CD11b/CR3 α-chain), HL3 (anti-CD11c/p150, 95 α-chain), M1/69 (anti-CD24/HSA), 3/23 (anti-CD40), 3E2 (anti-CD54/ICAM-1), 16-10A1 (anti-CD80/B7-1), GL1 (anti-CD86/B7-2), 2G9 (anti-I-Ad/IEd), RM134L (anti-OX40L). Biotin-labelled IgGs, used as isotype controls, were obtained from Pharmingen and Caltag Laboratories (San Francisco, CA). The anti-lamp-1 (CD107a) mAb 1D4B and the anti-lamp-2 (CD107b) mAb ABL-93 were from Pharmingen. The anti-CD68/macrosialin mAb FA/11, the mAb MOMA-2, rabbit Abs directed against H-2M, the mAb In-1 specific for invariant chains, rabbit anti-dinitrophenol Abs, rabbit IgGs specific to cathepsin B, H, L or D, the rabbit immune serum (IS) specific for rab7p and rabbit anti-hen egg ovalbumin Abs used as controls were obtained as previously described (Antoine et al., 1990; Prina et al., 1990; Lang et al., 2000; Courret et al., 2002). A goat IgG fraction specific to complement C3 was purchased from ICN Biomedicals (Aurora, Ohio). The mAb 3A1, specific to the LPG of L. amazonensis log phase Pm, and the unlabelled and biotinylated mAb 2A3-26, specific to L. amazonensis Am, were prepared as described earlier (Courret et al., 1999; Antoine et al., 1999). For some experiments, the mAb 2A3-26 was conjugated to the fluorophore Alexa Fluor 488 using a protein labelling kit (Molecular Probes). This Ab recognises an epitope specifically expressed on the surface of L. amazonensis and L. mexicana Am, but it is important to note that this epitope is already present on virtually 100% of intracellular parasites 18-24 hours after Pm internalisation (Courret et al., 2001). As an isotype control of the mAb 2A3-26, the anti-serotonin mAb G21-10 was used either unlabelled or coupled to Alexa Fluor 488. An Am-specific IS was prepared from L. amazonensis-infected BALB/c mice. A Pm-specific IS was made by immunising BALB/c mice with Pm membranes emulsified with incomplete Freund's adjuvant (Difco Laboratories, Detroit, MI). Membranes were obtained by osmotic shock lysis of L. amazonensis SPm. Before inoculation, the membranes were washed twice by centrifugation at 100,000 g in 10-fold diluted Dulbecco's PBS. Both Pm- and Am-specific IS were decomplemented by heating at 56°C for 30 minutes. Biotinylated primary Abs associated with cell preparations were detected by use of phycoerythrin-conjugated streptavidin (DAKO, Glostrup, Denmark), Alexa Fluor 488-streptavidin (Molecular Probes, Eugene, OR), Texas Red-labelled streptavidin (Pierce, Rockford, IL) or ExtrAvidin coupled to fluorescein isothiocyanate (Sigma). Anti-Ig F(ab′)2 fragments or Abs coupled to fluorescein isothiocyanate, Texas Red (Pharmingen) or Alexa Fluor 488 (Molecular probes) were used to detect cell-associated unlabelled primary Abs.
Parasite treatment with DBA/2J mouse serum or anti-Leishmania immune sera
To opsonize MPm Ficoll with complement, parasites were incubated for 10 minutes at room temperature with 5% DBA/2J mouse serum in Dulbecco's PBS and then thoroughly washed. Binding of component C3 was confirmed by staining the parasites with specific goat IgGs and a fluorochrome conjugate. Am nude (Igs are not detectable at the surface of these parasites) and MPm Ficoll were opsonized with specific Abs by incubating them for 1 hour at 4°C with a decomplemented IS prepared from Leishmania-infected BALB/c mice or from BALB/c mice immunised with Pm membranes, respectively. After thorough washing, parasites were checked for the presence of bound Abs by staining with fluorochrome-labelled donkey anti-mouse Ig F(ab′)2 fragments.
DCs were differentiated from bone marrow cells of 6- to 8-week-old BALB/C mice according to a method adapted from Méderlé et al. (Méderlé et al., 2002). Briefly, bone marrow cells were seeded at 2×106 cells per 100 mm diameter bacteriological grade Petri dish (Falcon, Becton Dickinson Labware, Franklin Lakes, NJ) in 10 ml of Iscove's modified Dulbecco's medium (IMDM; BioWhittaker Europe, Verviers, Belgium) supplemented with 10% heat-inactivated foetal calf serum (FCS; Dutscher, Brumath, France), 1.5% supernatant from a J558 cell line producing murine GM-CSF (Zal et al., 1994), 50 U/ml penicillin, 50 μg/ml streptomycin, 50 μM 2-mercaptoethanol and 2 mM glutamine (complete IMDM). Cultures were incubated at 37°C in a humidified atmosphere with 7% CO2. On day 3, 10 ml of complete IMDM was added. On day 6, suspended cells and loosely adherent cells were harvested using prewarmed 1% EDTA in Dulbecco's PBS without Ca2+ and Mg2+ (Biochrom AG, Berlin, Germany). Recovered cells were further cultured in the same conditions as above, except that complete IMDM was supplemented with 10% of the primary culture supernatant. On day 10, cells were harvested with EDTA as above and distributed in hydrophobic 6-well plates (Evergreen Scientific, Los Angeles, CA) at a concentration of 9×105 cells/well in 3 ml complete IMDM. On day 13, 2 ml of medium was removed from each well and replaced by 2 ml of fresh medium.
Infection of DCs
On day 14, DCs were infected with L. amazonensis Pm or Am or with BCG at micro-organism-DC ratios of 4:1 and 10:1, respectively. The samples were gently pipetted to dissociate DC clusters and to promote contact between micro-organisms and DCs. Infected cells and uninfected cells (controls run in parallel) were then placed at 34°C for 21 to 69 hours. Five hours after adding BCG, non-internalised bacteria were removed by gentle washing with complete IMDM. Cultures harvested 48 or 69 hours after adding micro-organisms were supplemented with 2 ml of fresh IMDM at 24 hours or at both 24 and 48 hours.
Detection of acidic compartments within DCs
Twenty-six hours after infection, DCs were incubated for 90 minutes at 34°C with 50 μM 3-(2,4-dinitroanilino)-3′ amino-N-methyldipropylamine (DAMP) as described previously (Antoine et al., 1990). Cells were then collected by EDTA treatment, washed and centrifuged on poly-L-lysine (mol. wt.=400,000; Sigma Chemical Co., St Louis, MO)-coated glass coverslips, before being stained as described previously (Antoine et al., 1990).
Conventional epifluorescence and laser scanning confocal microscopy
At different times after infection, DCs were collected by EDTA treatment. After centrifugation, cell pellets were resuspended in Dulbecco's PBS without Ca2+ and Mg2+. DCs were gently centrifuged on poly-L-lysine-coated glass coverslips and incubated at 34°C for 30 minutes, before fixation with paraformaldehyde and permeabilisation with saponin as described previously (Lang et al., 1994). Simple or double labellings with primary Abs and fluorescent conjugates were done according to routine procedures (Lang et al., 1994). Cell preparations were mounted in Mowiol (Calbiochem, San Diego, CA) before observation under an Axiophot Zeiss epifluorescence microscope or under a LSM 510 Zeiss confocal microscope. Confocal microscopy images were acquired and analysed using the LSM 510 software (version 3.1).
DCs were recovered at different times after infection as described above. Cell pellets were resuspended in cold Dulbecco's PBS with 2% FCS and 0.05% sodium azide (PBS-FCS-Az) and transferred to a round-bottomed 96-well plate (Costar, Corning, NY) at a concentration of 2 to 5×105 cells/well. All subsequent steps were done on ice and with cold reagents. Cells were centrifuged (300 g, 5 minutes) and then incubated in PBS-FCS-Az supplemented with 10% donkey serum for 20 minutes. After centrifugation, cells were incubated for 30 minutes in PBS-FCS-Az containing the primary biotinylated Abs. After three washing steps, they were incubated with PBS-FCS-Az containing phycoerythrin-conjugated streptavidin for 30 minutes. Cells were then washed again and treated with the CytoFix/CytoPerm reagent (Pharmingen) for 30 minutes. They were then either treated with the 2A3-26 mAb followed by fluorescein isothiocyanate-conjugated donkey anti-mouse Ig F(ab′)2 fragments or with the Alexa Fluor 488-conjugated 2A3-26 mAb to allow the detection of intracellular Leishmania. All washing steps were done using the Perm/Wash buffer (Pharmingen) supplemented with 10% donkey serum and all Abs were diluted in this same buffer. In all staining experiments, appropriate isotype controls (irrelevant rat, mouse or hamster mAbs) were incubated with cells at concentrations identical to those of the corresponding primary Abs. Flow cytometry results were acquired using a FACScan machine (Becton Dickinson, Mountain View, CA) and data analysed using the Cell Quest™ 3.1 software (Becton Dickinson).
Bone marrow-derived DCs that had not been incubated with any micro-organism were used as references for immature cells and bone marrow-derived DCs that had been incubated with BCG were used as references for mature cells. These cells were compared with DCs put in contact with Leishmania. As expected, flow cytometry showed that immature DCs (iDCs) expressed only tiny amounts of the costimulatory molecules CD80, CD86 and CD40. These DCs were also positive for the CD24 (Heat Stable antigen) and the CD54 (ICAM-1) molecules, and displayed heterogeneous levels of MHC class II molecules. Finally, the co-expression of CD11c, CD11b and CD11a molecules indicated that they exhibited a myeloid phenotype (Fig. S1, http://jcs.biologists.org/supplemental).
Live BCG was used as a positive control for optimal induction of DC maturation because, like Leishmania, these micro-organisms can live in intracellular compartments of phagocytic cells and are endowed with natural adjuvant activity. Upon stimulation of iDCs for 21 hours with live BCG, the expression level of costimulatory, adhesion and MHC class II molecules was considerably increased in all cells (Fig. S1, http://jcs.biologists.org/supplemental), indicating that they had started maturing. Hereafter, these cells are referred to mature DCs (mDCs).
To define further the phenotype of unstimulated DCs and BCG-treated DCs, we examined the localisation of MHC class II and H-2M molecules by confocal microscopy (Fig. 1). In unstimulated DCs, both molecules were co-located within vesicular organelles and moderate MHC class II molecule staining was detected on the plasma membrane (Fig. 1A). The class II- and H-2M-containing compartments, also known as MIIC (Geuze, 1998), were both acidic and positive for lamp-1 molecules, which are markers of late endosomes and lysosomes (data not shown). Upon contact with BCG, MHC class II molecules were segregated from H-2M molecules and redirected to the cell surface. Conversely, H-2M molecules migrated towards the centre of the cells (Fig. 1B) along with lamp molecules (data not shown). These distribution patterns are typical of iDCs and mDCs, respectively (Pierre et al., 1997).
iDCs are able to internalise different developmental stages of Leishmania
We first examined the ability of bone marrow-derived DCs to internalise the Pm and Am stages of L. amazonensis, which initiate and maintain parasitism in the mammalian hosts, respectively. After 21 hours in co-culture at a parasite:DC ratio of 4:1, we determined the percentages of infected DCs using the 2A3-26 mAb. In these conditions, the flow cytometry and microscopy analyses of stained DCs gave similar results. A typical result is shown in the inset of Fig. 2. iDCs could phagocytose efficiently both forms of the parasite (Fig. 2A). However, when we compare the results obtained with parasites without bound Abs, namely Am nude, SPm or MPm Ficoll, it is clear that the infection rate was lower with Am than with Pm.
We then assessed the effect of several opsonins (complement C3 component, Leishmania-specific Abs) on parasite uptake. When Am nude pre-incubated with an anti-Am IS or Am from BALB/c mice (Am B/c), naturally coated with anti-Leishmania Abs, were incubated with DCs for 21 hours, the percentage of infected DCs increased. The number of infected DCs also increased, although to a lesser extent, when MPm Ficoll added to DCs were incubated with a specific anti-Pm IS before infection. Likewise, MPm purified by a negative selection using the 3A1 mAb (MPm 3A1), which had a small amount of the mAb on their cell surface (N. Courret and J.-C. A., unpublished results), were more efficiently phagocytosed than MPm Ficoll or SPm (Fig. 2A). The number of intracellular parasites was slightly higher 21 hours after infection with Ab-coated parasites than with uncoated ones (Fig. 3). In contrast, opsonization with the complement C3 component did not promote the entry of Pm or affect the percentage of DCs containing intracellular parasites (data not shown).
It is important to state that, whatever the experimental conditions used, virtually all DC-associated parasites detected with the 2A3-26 mAb were surrounded by host membrane (see below), proving their intracellular location.
Leishmania survive and multiply within DCs
To determine whether parasites could survive and multiply inside DCs, we compared the percentages of infected DCs and the number of parasites per DC 21 and 69 hours after adding the parasites to DCs. During the period examined, the percentages of infected DCs remained stable or decreased very slightly, probably due to weak multiplication of the host cells (data not shown). The mean number of parasites per infected DC increased more than 2-fold when Am were used to initiate infection and about 1.5-fold when Pm were used. Similar data were obtained with parasites without bound Abs and Ab-coated parasites (Fig. 3). In both cases, infection could be maintained for at least 5 days (data not shown). Thus, in our culture conditions, parasites can multiply in DCs. These results also indicate that the mechanism of transformation/differentiation from a Pm form to an Am form is not impaired within DCs.
Intracellular Leishmania are located within phagolysosomes
The previous findings indicated that DCs, like macrophages, can serve as host cells for Leishmania Am. This suggested that compartments harbouring parasites in DCs are similar to macrophage PVs, which are highly permissive to parasite growth. To test this point, we partially characterised the organelles harbouring Leishmania in DCs. For the sake of clarity, these organelles will also be named PVs hereafter. The phenotype of the DC-associated PVs, which so far has been poorly investigated (Flohé et al., 1997; Henri et al., 2002), was determined by confocal immunofluorescence microscopy using Abs directed against molecules of the host cell endocytic organelles. Furthermore, the pH of their lumen was qualitatively assessed using the weak base DAMP. Twenty-one hours after phagocytosis of Am nude or Pm (SPm or MPm Ficoll), Leishmania were located in membrane-delimited compartments, some of which contained DAMP in their lumen (Fig. 4A-C). These data indicate that at least a part of these organelles is at acidic pH. The proportion of DAMP+-PVs was higher after infection with Pm (about 65% versus 35% after infection with Am), possibly because PVs are more tight-fitting at this stage of infection with this parasite form, a morphology that allows a better retention of soluble components like DAMP after fixation of the cells (Fig. 5). However, regardless of whether DCs were infected with Am or Pm, 90-100% of PVs were positive for the late endosomal/lysosomal glycoproteins lamp-1 and lamp-2 (Fig. 4D-F, Fig. 5). We also detected macrosialin, a member of the lysosomal-associated membrane protein family, and MOMA-2 antigen in the PV membrane (data not shown). It was hard to detect cysteine and aspartyl proteases like cathepsins B, H, L and D within PVs, although some were clearly positive for these proteases and particularly for cathepsin H (data not shown). This is different of what has been described for PVs of bone marrow-derived macrophages, which are clearly positive for the abovementioned proteases (Courret et al., 2002). This might be because the expression levels of all these proteases are lower within DCs than within bone marrow-derived macrophages. Only 35-65% of PVs in DCs expressed rab7p, a molecule involved in the fusion of the late endosomes/lysosomes (Fig. 5). In comparison, at this time point, 80-100% of PVs present in macrophages are rab7p+ (Courret et al., 2002). Perhaps in connection with this last point, following infection with Am, the enlargement of the PVs was much less in DCs than in macrophages during the first 48 hours post-phagocytosis. These observations are consistent with the idea that PV biogenesis is different in DCs and macrophages, or simply that the PV formation is delayed in DCs.
Several molecules involved in the MHC class II molecule-restricted antigen presentation process were also present in the PV membrane. Thus, at 21 hours post-phagocytosis of Am nude or MPm Ficoll, 90-100% of PVs contained MHC class II and H-2M molecules (Fig. 4G-I and Fig. 5) and a small proportion of them contained invariant chains Ii (data not shown). As already extensively documented for parasites located in macrophages stimulated with IFN-γ (De Souza Leao et al., 1995; Antoine et al., 1999; Courret et al., 2001), parasites in DCs also contained both MHC class II and H-2M molecules in internal structures, very likely the parasite lysosomal compartments called megasomes (Fig. 4G-I, arrowheads), indicating that these molecules are internalised by the parasites.
After infection of DCs for 21 hours with Ab-opsonized parasites (Am or MPm), Leishmania were localised in PVs exhibiting features similar to those of the PVs that developed after phagocytosis of parasites devoid of Abs (data not shown). However, the level of associated MHC class II molecules was much lower (Fig. 4K), whereas H-2M molecules were still strongly expressed in these organelles (Fig. 4J).
Similar results were obtained after 69 hours of infection (data not shown). Thus, PVs exhibit phagolysosomal properties suitable for the development of L. amazonensis Am, as well as MIIC characteristics.
General patterns of MHC class II and H-2M molecules in Leishmania-infected DCs
Interestingly, 21 hours after the infection of DCs with parasites devoid of Abs (Am nude or MPm Ficoll), the general distributions of class II and H-2M molecules were very similar to those observed in iDCs (Fig. 1A). In other words, a strong co-localisation of H-2M and intracellular class II molecules was observed in numerous organelles including PVs, and a moderate expression of class II molecules was observed at the cell surface (Fig. 4G-I). This suggested that at this stage of infection, parasitised DCs had not yet started maturing or that this process was not yet detectable by this method. In contrast, the distributions of class II and H-2M molecules in cells infected for 21 hours with Ab-coated parasites (Am or Pm) indicated that they had matured in these conditions. Thus, the expression of class II molecules at the cell surface was increased and the number of class II+- and H-2M+-vesicles was strongly diminished. In fact, in these cells, most of the H-2M molecules were located in PVs (Fig. 4J-L).
Flow cytometry study of the differentiation stage reached by Leishmania-infected DCs
Data concerning the distributions of MHC class II and H-2M molecules in infected DCs suggested that parasites without bound Abs are unable to induce the maturation of their host cells, at least during a 21-hour period, whereas Ab-opsonized parasites strongly induce this process. To test this assumption, we used flow cytometry to study the expression level of various molecules known to be up-regulated during maturation (Fig. S1, http://jcs.biologists.org/supplemental). For each molecule examined, we first analysed the entire DC populations. Unlike that of BCG, the uptake of Leishmania parasites (either Am or MPm) devoid of Abs by DCs did not obviously up-regulate CD40, CD54, CD80, CD86 or OX40L after a 21-hour co-culture period (Fig. 6). The surface expression of CD24 was unexpectedly decreased in these conditions (data not shown). Similar data were obtained after 69 hours of contact between DCs and parasites not coated with Abs, whereas for DCs infected with BCG for 69 hours, the expression levels of these molecules further increased (data not shown). Thus, in the absence of added opsonins, Am or MPm do not induce or only poorly induce DC maturation. However, the maturation of DCs infected with such parasites was strongly induced by LPS or BCG, indicating that the parasites do not inhibit the activation of DCs (data not shown).
In contrast, the infection of DCs with Ab-coated Leishmania induced a strong maturation, as revealed by the net increase in the surface expression of CD40, CD54, CD80, CD86, OX40L and MHC class II molecules and a moderate increase in that of CD24 (Fig. 6 and data not shown). These effects were clearly detectable in DCs incubated for 21 hours with (i) Am nude or Mpm Ficoll pre-incubated before infection with Leishmania-specific IS, (ii) MPm 3A1, or (iii) Am B/c, and were more pronounced after 69 hours (data not shown). However, not all opsonins induced such a strong DC maturation. Indeed, DCs infected with MPm Ficoll that had been coated with complement C3 component, a condition that should better correspond to that of natural primary infection, did not up-regulate activation markers more than DCs infected with unopsonized MPm Ficoll (data not shown).
The abovementioned flow cytometry analyses were conducted on the total DC populations co-cultured with the parasites, which included both uninfected and infected DCs. To enable us to distinguish between the behaviour of the uninfected and infected DCs present in the same populations, cells were co-cultured with parasites for different periods before labelling the activation markers listed above. Cells were then fixed and permeabilised before staining intracellular parasites with the mAb 2A3-26. Only the results with CD86 are given, but very similar data were obtained for the other cell surface molecules examined (i.e. CD40, CD54, CD80 and OX40L). iDCs and mDCs (incubated with BCG) were used as negative and positive controls, respectively. After 21 hours of contact between DCs and parasites not incubated with Abs (Am or MPm), the percentages of Leishmania-infected DCs expressing high levels of CD86 (noted in the upper right area of the plots shown in Fig. 7) were only slightly higher than those of CD86high cells present in iDC populations. The percentages of CD86high-infected DCs slightly increased at 69 hours post-infection, whereas those of CD86high cells present in iDC populations remained low. It is also noteworthy that between 21 and 69 hours of infection with Am or MPm devoid of Abs, the mean fluorescence intensities (MFI) of the CD86high Leishmania-containing DCs only moderately increased from 333 to 508 units and from 393 to 514 units, respectively, whereas the values of MFI for BCG-activated DCs reached 750 and 1195 units after 21 and 69 hours of contact, respectively. In contrast, the CD86 expression level of the uninfected DCs present in DC populations incubated with any form of the parasite did not vary over the 69-hour culture period (see percentages in the upper left areas of the plots).
Therefore (i) even after 69 hours of infection with Am nude or MPm Ficoll, about 70-80% of Leishmania-containing DCs are immature; (ii) Leishmania must directly interact with DCs to induce a weak up-regulation of CD86 in a minority of cells, regardless of the parasite stage used to start infection; (iii) the uninfected DCs present in DC populations incubated with parasites do not mature.
Similar analyses were performed on DC populations incubated with Ab-coated Am or MPm. After 21 hours in these conditions, 71% of Leishmania-containing DCs and 30-50% of the uninfected DCs present in the same populations expressed high levels of CD86 (Fig. 7). After 69 hours of infection, the percentages of CD86high-infected DCs remained elevated (above 60%). Between 21 and 69 hours, the CD86-specific MFI strongly increased for DCs infected initially with Ab-coated Am (from 408 to 817) or Ab-coated MPm (from 477 to 872).
Thus, the interactions between DCs and Ab-coated Am or MPm result in the strong expression of CD86 in both uninfected and Leishmania-containing DCs. These results are quite different to those obtained for DCs infected with parasites devoid of bound Abs.
We demonstrated that DCs derived from mouse bone marrow precursors cultured in the presence of GM-CSF are able to internalise both infective stages of Leishmania parasites (Am and MPm) in the absence of added opsonins. However, at the same parasite to DC ratio, infection with MPm is more efficient than infection with Am. This could be due to (i) differences in the probability of mobile (Pm) and non-mobile (Am) parasite forms encountering DCs, which grow in clusters when immature; (ii) a difference in the receptor-ligand pairs involved in the attachment and internalisation of Am and MPm by DCs. Numerous glycoconjugates that are differentially expressed on the parasite cell surface according to developmental stage [e.g. lipophosphoglycan, leishmanolysin/gp63, low molecular weight glycoinositolphospholipids and a membrane proteophosphoglycan (Ilgoutz and McConville, 2001)], could be involved in the binding of parasites to DCs, but their role remains to be documented. On the DC side, certain receptors belonging to the mannose receptor family (e.g. DC-SIGN and DEC-205) or the integrin family [e.g. CD11b/CD18 (CR3) and CD11c/CD18] may recognise some of the abovementioned parasite glycoconjugates. However, so far, only a role for CR3 in the uptake of L. major Am by mouse Langerhans' cells (Blank et al., 1993) and for DC-SIGN in the binding of axenic L. pifanoï Am to human monocyte-derived DCs (Colmenares et al., 2002b) have been demonstrated. In any case, our results appear different from others showing that Am are much more efficiently internalised by DCs than Pm (Blank et al., 1993; von Stebut et al., 1998; von Stebut et al., 2000). However, these other studies were done using another type of DC (Langerhans' cells or foetal skin-derived DCs) and another Leishmania species (L. major).
Various groups have demonstrated the presence of Abs on the surface of lesion-derived Am (Pearson and Roberts, 1990; Guy and Belosevic, 1993; Peters et al., 1995; Hodgkinson and Soong, 1997) and shown that these Abs are involved in the uptake of parasites by host cells (Peters et al., 1995). Natural Abs have also been detected in the serum of various animal species. This is especially true for some natural IgMs, which are capable of binding the Leishmania Pm stage, but the involvement of these molecules in the internalisation of parasites by host cells has never been illustrated (Schmunis and Herman, 1970; Pearson and Steigbigel, 1980; Navin et al., 1989). In contrast, numerous groups have shown that complement components, notably C3b and C3bi present on the surface of Pm incubated with serum, are mediators of the phagocytosis of this parasite stage (for a review, see Mosser and Brittingham, 1997). Hence, we further addressed the potential concern that opsonization of Am and MPm with complement components or Abs may enhance their infectivity for DCs. Our results indicate that opsonization with Abs significantly or moderately enhances the percentages of infected DCs when Am or MPm are used to initiate infections, respectively. In contrast, opsonization of Pm with complement did not enhance the uptake of parasites by DCs.
We also provide evidence that once inside DCs, Leishmania can survive and even multiply for at least a few days. As they are able to sustain parasite survival and growth, DCs may serve as hosts for Leishmania, consistent with the view that DCs play a role in persistent infections and in the maintenance of a Leishmania-specific immune response within mammalian hosts (Moll et al., 1995).
The multiplication of intracellular parasites indicates that they encountered a favourable environment. Henri and colleagues (Henri et al., 2002) have reported that in mouse splenic DCs, L. major parasites reside in phagosomes that are closely related to late endocytic compartments. Our results indicate that the DC compartments housing parasites are rather similar to the PVs that form in macrophages (Antoine et al., 1998). Indeed, they are acidic and their membrane shares many features with those of late endosomes and lysosomes. However, PVs harbouring rab7p are less numerous in DCs than in macrophages during the period studied and the progressive enlargement of the PVs is less marked in DCs than in macrophages. This could reflect a restriction of the fusion processes occurring in DCs compared to what happens in infected macrophages and may have consequences in the parasite antigen presentation capacities of infected DCs. In this respect, various molecules involved in the antigen presentation process have been detected within PVs. These include MHC class II and H-2M molecules, but also to a lesser extent invariant chains and acidic proteases like cathepsins. Interestingly, parasites present in both macrophages and DCs can internalise MHC class II and H-2M molecules but it is not clear whether the parasites use this process as a means to thwart the antigen presentation process.
Light microscopy and flow cytometry showed that most DCs incubated with either Am nude or MPm Ficoll remained immature. However, two-colour flow cytometry showed that some markers like CD86 and CD40 were up-regulated at the surface of a small proportion of DCs containing intracellular parasites. This phenomenon concerned a very small number of cells after 21 hours of infection, but was more obvious after 69 hours. No such increase could be demonstrated in iDCs, which have not been in contact with Leishmania, or in the uninfected DCs present in the DC populations incubated with either stage of the parasite. It is thus tempting to speculate that the entry of Leishmania, or the presence of intracellular parasites, can pass unnoticed in many DCs but is not completely silent in others. This is sustained by the fact that a weak negative modulation of CD24 expression was observed in infected DC populations compared to in iDCs. Although we currently have no evidence about the identity of the Leishmania molecules that are involved in this weak/slow DC maturation, it is possible that Leishmania express danger signals known as pathogen-associated molecular patterns (PAMPs), which may interact for instance with Toll-like-receptors at the surface of or within DCs. However, the engagement of such receptors on/in DCs generally results in a strong maturation phenotype, which does not really fit with our results.
Alternatively, Leishmania may have developed mechanisms to prevent the maturation of host DCs or leading to the incomplete maturation of these cells, at least at certain stages of infection. Leishmania may also be able to modify the kinetics of production of chemokines, cytokines and other unknown molecules involved in the maturation process to fit better their parasitism requirements. It is generally accepted that the full maturation of DCs is dependent on the production of cytokines and chemokines, which act in an autocrine or paracrine fashion (Cebon et al., 2001). The fact that after a 3-day co-culture of DCs and parasites devoid of bound Abs, only 20-30% of L. amazonensis-containing DCs and no more than 7% of the uninfected ones had reached an intermediate maturation stage suggests that cytokines like TNF-α are not produced in great quantities. Therefore, these conditions do not seem to allow the trans-activation of uninfected cells by parasite-containing DCs. On the contrary, when DCs were put in contact with BCG, almost all cells, including the DCs that do not contain BCG, were driven to a full maturation stage.
Contrary to what has been described above, DCs quickly mature after contact with Ab-coated parasites (Am or Pm). Furthermore, uninfected cells are trans-activated, very probably consecutively to the stimulation of the infective ones. This suggests that a greater amount of soluble activating molecules are produced in these conditions. These data imply that FcγRs are involved in the stimulation of Leishmania-infected DCs. In support of this conclusion, it was recently shown that FcγRs are expressed by mouse bone marrow-derived DCs and that these receptors can be involved in the triggering of the maturation process (Regnault et al., 1999; Schuurhuis et al., 2002; Tan et al., 2003). In contrast, opsonization of Pm with complement did not lead to the strong maturation of DCs, which in these conditions, behaved as DCs infected with parasites devoid of added opsonins. This finding may be linked with the fact that phagocytosis through complement receptors is relatively silent and does not elicit the release of inflammatory mediators (for a review, see Aderem and Underhill, 1999).
It is possible that the data reported in this study could explain some aspects of the development of the parasitism caused by Leishmania and/or the relationships between parasite growth and the development of the immune responses. Our results suggest that shortly after infection with a small number of Pm (102-103, as in natural infections), i.e. in the absence of Leishmania-specific Abs, either DCs are not infected because of their low phagocytic activity and competition with resident macrophages, or, if they are infected, they remain immature or reach an intermediate maturation stage not allowing the stimulation of Leishmania-specific naive CD4 T lymphocytes. In both conditions, adaptive immune responses would not be induced immediately, which may allow the amplification of the parasite population even in mammalian hosts exhibiting a resistant phenotype. In favour of this hypothesis, a `silent' phase of parasite amplification, lasting between 1 and 5 weeks and occurring before the onset of immune response/immunity, has been described in C57BL/6 and BALB/c mice infected with low doses of L. major Pm (Belkaid et al., 2000; Lang et al., 2003). At later stages of infection, when specific Abs are present, the internalisation of Ab-opsonized parasites by DCs is probably followed by the strong maturation of these host cells and thus by the amplification of the immune response. In this respect, a role for Abs and FcγRs in the development of the local cutaneous immune response induced in BALB/c mice by L. pifanoï and L. amazonensis has been documented (Kima et al., 2000; Colmenares et al., 2002a).
This work was funded by the Institut Pasteur and the Centre National de la Recherche Scientifique [regular budget to J.-C. Antoine and appel d'offres `Puces à ADN 2000-2002' (co-PI G. Milon/M. Pages/P. Glaser)]. The confocal microscope was purchased with a donation from M. and L. Pollack. We are grateful to S. Amigorena, J.-L. Guesdon, H. Kirschke, G. L. E. Koch, N. Koch, D. L. Sacks, B. Wiederanders and M. Zerial for the kind gifts of immunological reagents, and to G. Langsley for critical reading of the manuscript.
Supplemental data available online
- Accepted September 5, 2003.
- © The Company of Biologists Limited 2004