are dimorphic protozoa responsible for numeroushuman parasitic diseases. These parasites exist as flagellatedpromastigotes within sandfly vectors and as intracellularamastigotes residing in macrophages of infected mammals(Chang, 1983). Once inside macrophages, Leishmania
establish themselves within membrane-bound compartmentsknown as the parasitophorous vacuoles (PV). Only PV of rator mouse macrophages infected with L. mexicana or L. ama-zonensis have been well characterized. These Leishmaniaspecies, which are the causal agents of New World cutaneousleishmaniasis, induce in the host cell the formation of a smallnumber of huge PV each containing numerous parasites.
2137
Leishmania donovani
amastigotes, the etiological agents ofvisceral leishmaniasis, are obligate intracellular parasitesresiding in membrane-bound compartments ofmacrophages called parasitophorous vacuoles (PV). Thestudy of these organelles is of paramount importance tounderstanding how these parasites resist the microbicidalmechanisms of macrophages and how they escape theimmune response of their hosts. Confocal microscopy ofmouse bone marrow-derived macrophages infected with L.donovani amastigotes and stained for various prelysoso-mal/lysosomal markers and for major histocompatibilitycomplex (MHC) molecules was used to define PV withrespect to the endocytic compartments of the host cells andto address the issue of their potential role in antigen pro-cessing and presentation. Forty-eight hours after infection,many PV contained cathepsins B, D, H and L and they wereall surrounded by a membrane enriched for the lysosomalglycoprotein lgp120/lamp1 but apparently devoid of thecation-independent mannose 6-phosphate receptor, amembrane protein generally absent from the lysosomes.These data suggested that PV acquire within 48 hours thecharacteristics of a lysosomal compartment. However, bothmacrosialin and the GTP-binding protein rab7p (specificmarkers of the prelysosomal compartment) were found tobe highly expressed in/on PV membrane. Thus, at thisstage, PV appear to exhibit both lysosomal and prelysoso-mal features. Infected macrophages activated with IFN-
γbefore or after infection showed PV strongly stained forMHC class II molecules but not for MHC class I molecules.
This suggests that, if infected macrophages can act asantigen-presenting cells for class I-restricted CD8+ T lym-phocytes, Leishmania antigens must exit the PV. MHCclass II molecules reached the PV progressively, indicatingthat they were not plasma membrane-bound moleculestrapped during internalization of the parasites. The redis-tribution of class II observed in infected cells did not altertheir quantitative expression on the plasma membrane atleast during the first 48 hours following the phagocytosis ofthe parasites. The invariant chains, which are transientlyassociated with class II molecules during their intracellu-lar transport and which mask their peptide-binding sites,did not reach PV or were rapidly degraded in these sites,suggesting that PV-associated class II are able to bindpeptides. This last assumption is strengthened by the factthat class II located in PV could bind conformational anti-bodies that preferentially recognize class II with tightlyassociated peptides. Together, these data showed that L.donovani amastigotes live in hydrolytic compartments ofthe host cells where the processing of parasite antigenscould eventually occur as well as the binding of parasitepeptides to MHC class II molecules, both events beingrequired before the presentation of antigens to specificCD4+ T lymphocytes.
Key words: macrophage, Leishmania donovani, phagosome,phagolysosome, MHC class I molecule, MHC class II molecule,antigen presentation, endosomal/lysosomal marker, confocalmicroscopy, parasitophorous vacuole
SUMMARY
Leishmania donovani-infected macrophages: characterization of the
parasitophorous vacuole and potential role of this organelle in antigen
presentation
Thierry Lang1,*, Raymond Hellio2, Paul M. Kaye3 and Jean-Claude Antoine1
1Unité d’Immunophysiologie Cellulaire, Département de Physiopathologie Expérimentale, Institut Pasteur, 25 rue du Dr Roux,75724 Paris Cedex 15, France2Unité de Biologie des Membranes, Département de Biologie Moléculaire, Institut Pasteur, Paris, France 3Department of Medical Parasitology, London School of Hygiene and Tropical Medicine, London, UK
Recent data indicate that 24 to 72 hours post-infection, thesePV maintain an acidic pH, contain various apparently fullyactive lysosomal hydrolases and are limited by a membraneexhibiting both lysosomal and prelysosomal features(Alexander and Vickerman, 1975; Antoine et al., 1987, 1990,1991; Prina et al., 1990; Russell et al., 1992; Lang et al., 1994).Indeed, at this stage, the PV membrane contains lysosomal gly-coproteins (lgp110/lamp-2, lgp120/lamp-1), late endosomal/prelysosomal markers like macrosialin (Rabinowitz et al.,1992) and the small GTP-binding protein rab7p (Chavrier etal., 1990), but are devoid of the cation-independent mannose6-phosphate receptor (CI-MPR), a glycoprotein generallyexcluded from the terminal lysosomes (Brown et al., 1986;Griffiths et al., 1988; Geuze et al., 1988; Croze et al., 1989).The presence of major histocompatibility complex (MHC)class II molecules in the PV of infected macrophages stimu-lated by gamma interferon (IFN-γ) or from lesion biopsies wasalso documented (Antoine et al., 1991; Russell et al., 1992).Taken together, these results suggest that the PV housing L.mexicana or L. amazonensis are acidic and hydrolytic siteswhere the processing, including denaturation and limited pro-teolysis of parasite antigens, could occur. Additionally, thebinding of immunogenic parasite peptides to MHC class IImolecules, an event required before the recognition of antigens(Ag) by specific CD4+ T lymphocytes (for a review seeBrodsky and Guagliardi, 1991; Neefjes and Ploegh, 1992),may also occur in the PV.
In contrast, until now, little was known about the functionand composition of PV housing other Leishmania species likeL. major and L. donovani, whereas paradoxically theseparasites have been the object of the most intensive biochem-ical and immunological studies in the field. We thus undertooka systematic study of the PV containing L. donovani, the aeti-ological agent of visceral leishmaniasis, using mousemacrophages infected for 48 hours with the amastigote stageof this parasite. Earlier studies have shown that these PV arestrongly acidic (Chang, 1980) and fused with compartmentsthat appeared to be lysosomes (Chang and Dwyer, 1976, 1978),but the membrane composition and contents of these organelleshave not been analysed nor has their relationship with theantigen processing and presentation processes been deter-mined. This last issue is of course of fundamental interest indetermining the role played by infected macrophages in theinduction and/or orientation of the immune responsesdeveloped during L. donovani visceral disease. PV propertiesquite different from those described above for PV containingLeishmania of the mexicana complex would be likely, sincemacrophages infected with L. donovani contain numerous PV,each one with, generally, a single parasite tightly surroundedby the PV membrane (Chang and Dwyer, 1978). However,unexpectedly, we found in the present study that PV contain-ing L. donovani displayed the same luminal and membraneprotein markers as PV housing L. amazonensis. So, in spite ofa very different morphology and with the reservation that onlya limited set of membrane and soluble proteins was studied,both types of PV appear rather similar in composition.
MATERIALS AND METHODS
Animals and parasitesFemale Balb/c and C57BL/6 mice aged between 6 and 12 weeks were
used as sources of macrophages. Amastigotes of the Leishmaniadonovani strain LV9 (MHOM/ET/67/Hu3: LV 9) were isolated fromspleens of infected hamsters by homogenization of the organsfollowed by Percoll gradient fractionation as described elsewhere(Channon et al., 1984).
Immunological reagentsHybridoma cells producing the following monoclonal antibodies(mAb) were obtained from the American Type Culture Collection(Rockville, MD). (i) M5/114, a rat anti-mouse I-Ab,d,q and I-Ed,k
IgG2b (Bhattacharya et al., 1981); (ii) Y-3P, a mouse anti-I-Ab IgG2a(Janeway et al., 1984); (iii) 34-4-20S, a mouse anti-H-2Dd IgG2a(Ozato et al., 1982). The mAb K9.18, a mouse anti H-2Kd IgG2a, wasobtained from U. Hammerling (Sloan-Kettering Institute, NY) and theA10.3.2 mAb, a mouse anti-serotonine IgG2a, from J.-L. Guesdon(Institut Pasteur, Paris, France). Hybridoma cells secreting the 14-4-4S Ab, a mouse anti-I-Ed IgG2a (Ozato et al., 1980), were kindlysupplied by O. Léo, B. Vray and N. Plasman (Université Libre deBruxelles, Belgium). Hybridoma cells secreting the H35-17.2 Ab, arat anti-mouse CD8 IgG2b, were kindly given by M. Pierres (Centred’Immunologie de Marseille Luminy, France; Goldstein et al., 1982).Hybridoma cells secreting the In-1 mAb, a rat anti-mouse invariantchain (Ii) IgG2b (Koch et al., 1982), and hybridoma cells secretingthe FA/11 Ab, a rat anti-mouse macrosialin IgG2a (Smith and Koch,1987), were kindly provided by N. Koch (Institut für Immunologieund Genetik, Heidelberg, Germany) and G. Koch (MRC laboratoryof Molecular Biology, Cambridge, UK), respectively. The M5/114and H35-17.2 Ab were purified by adsorption chromatography asdescribed previously (Antoine et al., 1991) from ascites prepared innude mice. The A10.3.2 mAb was purified from hybridoma culturesupernatants by HPLC on MonoQ (Pharmacia, Uppsala, Sweden).Hybridoma supernatants were used as sources of the other mAbwithout further purification.
A rabbit immune serum made against bovine liver CI-MPR andcross-reacting with mouse CI-MPR (Griffiths et al., 1988) was kindlyprovided by B. Hoflack (EMBL, Heidelberg, Germany). A rabbit anti-rat lgp120 immune serum, cross-reacting with mouse lgp120 (Lewiset al., 1985), was obtained from I. Mellman (Yale University, NewHaven, CT). Rabbit IgG specific to rat cathepsins B, D, H and L wereobtained from B. Wiederanders and H. Kirchke (1986; Martin LutherUniversity, Halle-Wittenberg, Germany). Affinity-purified rabbitanti-rab7p Ab were a kind gift from P. Chavrier (Centre d’Immunolo-gie de Marseille Luminy, France; Chavrier et al., 1990). Rabbit anti-hen egg ovalbumin (OVA) antibodies were purified as describedelsewhere (Prina et al., 1990). Rat IgG were isolated from serum byDEAE-cellulose chromatography as described previously (Antoine etal., 1991).
Donkey anti-rabbit IgG (H+L) Ab conjugated to fluorescein isoth-iocyanate (FITC) were obtained from Amersham (Les Ulis, France).FITC-labelled goat anti-rat IgG (H+L) F(ab′)2 fragments werepurchased from Caltag (San Francisco, CA) and FITC-conjugatedgoat anti-mouse IgG (H+L) F(ab′)2 fragments, Texas Red-conjugatedgoat anti-mouse IgG (H+L) Ab (rat Ig-adsorbed) from Jackson Lab-oratories (West Grave, PA).
Cell culturesBone marrow cells were flushed from femurs and tibias of mice andsuspended in RPMI 1640 medium (Seromed, Berlin, Germany) sup-plemented with 10% heat-inactivated foetal calf serum (Dutscher,Brumath, France), 20% L-929 cell-conditioned medium as a sourceof macrophage colony-stimulating factor (M-CSF) and 50 µg/ml ofgentamicin (Seromed). For light and electron microscopy, cells weredeposited in 24-well tissue culture plates (Costar, Cambridge, MA)containing 12 mm diameter round glass coverslips (2×105 cells/well)and in 35 mm tissue culture dishes (Corning Glass Works, Corning,NY; 2×106 cells/dish), respectively. For quantification of plasmamembrane-associated class II molecules by flow cytometry (FCM)
T. Lang and others
2139
Leishmania donovani-infected macrophages
analysis, 100 mm bacteriological Petri dishes (Greiner, Nürtingen,Germany) were seeded with 1×107 bone marrow cells. After 3 daysat 37˚C in a 5% CO2-air humidified atmosphere, fresh medium wasadded and two days later, adherent macrophages were washed withDulbecco’s phosphate-buffered saline (PBS). Macrophages were thenincubated in culture medium devoid of M-CSF for 24 hours beforeinfection.
Macrophage infection, treatment with rIFN-γ andassessment of the viability of L. donovani amastigotesMacrophages were infected by coculture with L. donovani amastig-otes at a parasite to macrophage ratio of 8:1 for 2 hours. Beforeinfection, macrophage numbers were determined on cultures run inparallel by counting nuclei as described previously (Antoine et al.,1991). After infection, cells were washed at least six times to removeuningested parasites. In some cultures, murine rIFN-γ produced byEscherichia coli (3.2 fg endotoxin/µg) was added at the concentrationof 100 units/ml either 24 hours before infection or just after infection.In these cultures, medium was supplemented every day with freshrIFN-γ (100 units/ml). Uninfected cells run in parallel were similarlytreated. Percentages of infected macrophages and numbers ofamastigotes per macrophage were determined on Giemsa-stainedpreparations after counting about 500 cells per sample.
Viability of internalized amastigotes was assessed by using a col-orimetric MTT (3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazoliumbromide) (Sigma Chemical Co., St Louis, MO) assay after selectivelysis of macrophages (Kiderlen and Kaye, 1990).
Quantification of plasma membrane-associated MHCclass II moleculesQuantification of MHC class II molecules expressed on themacrophage plasma membrane was performed on live cells using aFCM analysis. Uninfected and 48 hour-infected macrophages werewashed with PBS and detached from the dishes by treatment for 20minutes at 37˚C with Ca2+-and Mg2+-free PBS containing 2 mg/mlglucose. Macrophages were then incubated for 30 minutes with 1%bovine serum albumin (BSA)(IBF, Villeneuve-la-Garenne, France)and 0.1% sodium azide in cold PBS to block nonspecific binding sites.Cells were then successively incubated for 30 minutes at 4˚C with asaturating concentration (1 µg/ml) of M5/114 mAb or with the sameconcentration of normal rat IgG and, after washings with cold PBS-BSA, with a saturating concentration (10 µg/ml) of FITC-labelledgoat anti-rat Ig F(ab′)2 fragments. Cells were then postfixed with 1%paraformaldehyde. For each sample, 5000 cells were examined on aFACScan fluorescence-activated cell sorter (Becton-Dickinson,Mountain View, CA). The data were collected and analysed on aHewlett Packard (HP 217) computer using LYSIS software.
Immunodetection at the light microscope levelMacrophages were fixed for 1 hour at room temperature with 4%paraformaldehyde in 0.1 M sodium cacodylate/HCl buffer, pH 7.4,quenched with 50 mM NH4Cl in PBS for 10 minutes and permeabi-lized for 30 minutes with 0.01% saponin (Sigma) in PBS 0.1% BSA.Cells were then sequentially incubated for 1 hour at room tempera-ture with either a mAb (FA/11, M5/114, Y-3P, 14-4-4S, In-1, K9.18,34-4-20S), rabbit Ab (anti-rab7p or anti-cathepsin B, D, H or L) oran immune serum (anti-CI-MPR or anti-lgp120) and then with anadequate FITC-conjugate. For double-immunofluorescencelabellings, cell preparations were first incubated with both primary Abbelonging to different species and then with both species-specificsecondary Ab conjugated to different fluorochromes. Immunologicalreagents were diluted in PBS containing 0.1% BSA and 0.01%saponin. Each incubation was followed by 3 washings with PBS,0.01% saponin. Finally, cells preparations were incubated for 5minutes in propidium iodide (1 µg/ml in PBS), washed twice in PBSand water, and mounted in Mowiol (Calbiochem, San Diego, CA).Macrophages were examined under a conventional Zeiss fluorescence
microscope (Oberkochen, Germany) or under a Leitz confocal laserscanning microscope (Wild Leitz Instruments, Heidelberg, Germany).
As controls, coverslips were incubated either with an irrelevantisotype-matched mAb, or with rabbit anti-OVA Ab, or else withnormal rabbit serum according to the primary Ab or immune serumused in the specific tests and then with an appropriate FITC or TRconjugate. A very low background was observed with the irrelevantmAb (H35-17.2 or A10.3.2) or purified polyclonal Ab (anti-OVA).For controls made with non-immune serum the background washigher but randomly distributed.
Electron microscopyMacrophages were fixed for 2 hours with a mixture of 4%paraformaldehyde and 2.5% glutaraldehyde (Sigma) in 0.1 M sodiumcacodylate/HCl buffer, pH 7.4. Cells were washed overnight in PBS,post-fixed for 1 hour at room temperature with 1% osmium tetraox-ide in the same buffer; after which they were scraped off with a rubberpoliceman and concentrated in agar. The dehydration in gradedethanol and Epon embedding have been described elsewhere (Lang etal., 1988). Thin sections were stained with lead citrate and uranylacetate, and examined under a Jeol 100 CX transmission electronmicroscope (Tokyo, Japan) at 80 kV.
RESULTS
Features of the in vitro model of infectionAn in vitro system was developed for studying host-parasitecellular interactions between L. donovani amastigotes isolatedfrom the spleen of infected hamsters and mouse bone marrow-derived macrophages cultured in presence or absence of rIFN-γ. Macrophages were infected with freshly isolated L. donovaniamastigotes at a parasite-to-host cell ratio of 8:1. Followingthis protocol, 70 to 80% of the population was infected 2 hoursafter adding parasites and an average of 3 to 4 parasites percell were phagocytosed. The percentage of infected cells didnot vary with time after infection and the number of amastig-otes/macrophage slightly increased and reached 5.5 at 48 hourspost-infection. Evolution of these parameters was similar in thepresence or absence of IFN-γ (data not shown).
Viability of the parasites present in macrophages was assessedby a method allowing the selective lysis of the host cells whileleaving the liberated amastigotes completely unaffected in termsof their ability to transform into promastigotes and subsequentproliferation (Kiderlen and Kaye, 1990). Forty-eight hours afterinfection, the number of viable parasites/well, estimated by aMTT assay, was unaffected by the presence of rIFN-γ (A570nmof 0.380±0.087 and 0.345±0.060 in the presence and absence ofrIFN-γ, respectively). Thus, at a dose of 100 units/ml, rIFN-γ didnot activate macrophages for leishmanicidal activity.
Forty-eight hours post-infection, the ultrastructure ofinfected macrophages showed three different types of para-sitophorous vacuoles as described by Chang and Dwyer(1978). The first type contained a single amastigote sur-rounded by a tight-fitting membrane leaving no vacuolar space(Fig. 1A). The second type, smaller in size, also containedapparently only one amastigote partially bound to the PVmembrane, the remaining portion of the parasite being free inthe lumen of the PV (Fig. 1B). The last type, larger in size,housed at least two or three amatigotes attached to the PVmembrane (Fig. 1C). These various aspects of PV trulyreflected heterogeneity in PV size and were not due to differ-
2140 T. Lang and others
Fig. 1. Macrophages from Balb/c mice were incubated for 2 hours with L. donovani amastigotes, after which they were treated with rIFN-γ(100 units/ml). Forty-eight hours after infection, they were processed for electron microscopy. Different types of vacuoles can be observed. In(A) the PV membrane is tightly apposed to the parasite. In (B) the amastigote partially adheres to the PV membrane (arrowheads) with theremaining portion being free in vacuolar space filled with electron-dense material. In (C) three amastigotes are present in the same PV, inwhich they are attached to the membrane. fp, flagellar pocket; PM, plasma membrane; n, nucleus of the macrophage. Bars, 0.25 µm.
2141Leishmania donovani-infected macrophages
ences in the plane of section, since they were also observedby light microscopy.
Characterization of the PVLocalization of different organelle markers in 48 hour-infected macrophagesIn order to characterize more precisely the PV containing L.donovani amastigotes, it was of interest to define this com-partment with respect to endosomal and lysosomal markers.This was accomplished by immunolabelling of infectedmacrophages with immune sera or purified Ab directed againstthe following markers: (1) the lysosomal hydrolases, cathep-sins B, D, H and L (Wiederanders and Kirschke, 1986); (2) thelysosomal membrane glycoprotein lgp120. This protein hasbeen detected mainly in lysosomes (Lewis et al., 1985) but alsoin late endosomes/prelysosomes (Griffiths et al., 1988; Geuzeet al., 1988; Griffiths, 1992); (4) CI-MPR. This membraneprotein is often used to distinguish between late endosomes/prelysosomes and lysosomes, and has been described as amarker of the trans-Golgi network and prelysosomal compart-ments (Brown et al., 1986; Griffiths et al., 1988; Geuze et al.,1988; Croze et al., 1989; Griffiths, 1992); (5) macrosialin, amacrophage-specific marker found at low levels on the plasmamembrane, absent from the terminal lysosomes but highlyexpressed in the tubular prelysosomal/lysosomal compartment(Rabinowitz et al., 1992); (6) rab7p, a GTP-binding proteinthat is restricted to the cytosolic side of late endosomal/prelysosomal membranes (Chavrier et al., 1990). Stainingpatterns of all these markers were analysed by confocalmicroscopy in macrophages infected for 48 hours andcompared with those observed in uninfected macrophages.
In uninfected cells, cathepsins B, D, H and L were locatedmainly in perinuclear vesicles and in many granules dispersedfar from the perinuclear region as illustrated in Fig. 2A-C. Insome cells, labellings were also associated with tubularelements similar to those previously described by others andourselves (Swanson et al., 1987; Knapp and Swanson, 1990;Tassin et al., 1990). The staining pattern of the various cathep-sins was rather similar but was stronger and more distinct forcathepsins D, H and L. In infected macrophages stained forcathepsin B, D, H or L, not all PV were labelled but positiveones were strongly stained (Fig. 2D,E,F). A depletion ofpositive vesicles and tubules was also observed in infectedcells, especially in cells stained for cathepsin D. Interestingly,the distribution within PV of cathepsin D on the one hand andof cathepsins B, H and L on the other hand appeared different.Cathepsin D staining was often all located on the periphery ofPV (Fig. 2D) whereas stainings of the other cathepsins lookedlike Phrygian caps partially enveloping amastigotes (Fig.2E,F). The latter distribution was reminiscent of that obtainedfor endocytic markers entering PV (Rabinovitch et al., 1985).These results indicate that cathepsins B, H and L were verylikely located in the PV lumen whereas at least part ofcathepsin D could be bound to the PV membrane (Diment etal., 1988).
lgp120 was localized in small vesicles distributed through-out the cytoplasm as well as in larger perinuclear vesicles ofuninfected macrophages (Fig. 2G) and was frequently associ-ated with tubular elements (not shown). In infected cells, allPV exhibited lgp120 staining and the fluorescence of these
organelles was restricted to their periphery (Fig. 2H). This wasparticularly evident for the large PV containing two or moreparasites. The presence of cathepsins and lgp120 in PV led usto consider these organelles as lysosome-like and our resultswere consistent with the involment of the lysosomal membraneand lysosomal content in PV genesis. However, althoughlysosomes were shown to be greatly enriched in lgp, someresults indicate that lgp120 and cathepsins are also localized inendosomes and prelysosomal compartments (Diment et al.,1988; Geuze et al., 1988; Griffiths et al., 1988). In this context,it was thus important to study the distribution in infectedmacrophages of prelysosomal markers described as beingabsent from the terminal lysosomal compartment like CI-MPR,macrosialin and rab7p.
As we described previously (Tassin et al., 1990; Antoine etal., 1991), CI-MPR was mainly located in a few vesiclesclustered on one side of the nucleus of uninfected macrophagesand/or in vesicles dispersed in the cytoplasm (Fig. 3A). Thestaining pattern of CI-MPR was very similar in L. donovani-infected macrophages and no labelling was detected in PV(Fig. 3B). Similar findings were obtained whethermacrophages had been treated or not with rIFNγ. Even after alonger time of infection (8 days), the PV remained negative forthis marker, which is in contrast to what has been describedfor PV housing L. mexicana in late in vitro infections (Russellet al., 1992). Thus, the absence of CI-MPR and the presenceof lysosomal markers also agreed with the idea that PV arelysosomal in nature. To confirm or to invalidate this point, cellswere stained for other endosomal/prelysosomal markers.
We first examined macrosialin labelling. In uninfectedmacrophages it was found associated with the periphery oflarge vesicles as well as with small vesicles and tubularelements (Fig. 3C). After infection, rather similar macrosialin
Fig. 2. Immunolabelling of uninfected and L. donovani-infectedmacrophages from Balb/c mice with Ab directed against differentlysosomal markers. Macrophages were infected and treated or notwith rIFN-γ as described in the legend to Fig. 1. At 48 hours post-infection, fixed and permeabilized cells were incubated with eitherone of the four anti-cathepsin Ab or with the anti-lgp120 immuneserum and, after washings, with an appropriate FITC-conjugate(green staining). Uninfected macrophages cultured in parallel weresimilarly treated. Labelling of DNA with propidium iodide (redstaining) allowed us to visualize the host cell nuclei (n) and parasitenuclei (stars). Optical sections (A,C,F,H: 0.8 µm thickness; B,D,E,G:0.3 µm thickness) of labelled cells obtained with a confocal laserscanning microscope are shown. (A,B,C) Uninfected macrophagesstained for cathepsin D,H and L, respectively. Cathepsins are locatedin perinuclear vesicles, in many granules spread in the cytoplasm andin tubules (A). (D,E,F) Immunolabelling of infected macrophageswith anti-cathepsin D, anti-cathepsin H and anti-cathepsin L Ab,respectively. PV are strongly stained (arrows).(G,H) Immunolabelling of lgp120 in uninfected (G) and infected (H)macrophages. In (G) two micrographs of the same cell are showncorresponding, respectively, to the first (upper micrograph) and tothe third (lower micrograph) optical sections from the basal level ofthe cell. In (H) four serial sections of the same cell are shown, thefirst one (upper micrograph) corresponding to the basal level of thecell. lgp120 is detected in large perinuclear vesicles as well as insmaller dispersed vesicles of uninfected macrophages (G) and on theperiphery of PV (arrows) in infected macrophages (H). Bars, 5 µm.
2142
labelling was observed. In addition, PV were strongly andhomogeneously stained at their periphery (Fig. 3F).
Finally, the distribution of rab7p was compared in unin-
fected and infected macrophages. In uninfected macrophages,the labelling was concentrated in juxtanuclear vesicles. Somepositive vesicles were also detected at the cell periphery (Fig.
T. Lang and others
Fig. 2. For legend see p. 2141
2143Leishmania donovani-infected macrophages
3D). A pattern quite similar was observed in infectedmacrophages. PV were also specifically labelled with anti-rab7p Ab (Fig. 3E).
These results thus show the association of both lysosomaland prelysosomal markers with PV and support a centralposition for these compartments.
Fig. 3. Localization of endosomal/prelysosomal markers in uninfected (A,C,D) and in L. donovani-infected macrophages (B,E,F) treated withrIFN-γ as described in the legend to Fig. 1. Forty-eight hours after adding parasites, macrophages from Balb/c mice were fixed, permeabilizedand sequentially incubated with immune serum or mAb and with an appropriate FITC-conjugate (green staining). Cells were then incubatedwith propidium iodide (red staining) in order to visualize parasite nuclei (stars) and macrophage nuclei (n). Uninfected macrophages cultured inparallel were similarly processed. Cell preparations were analysed by confocal microscopy. Optical sections 0.8 µm thick (A,B,D,E,F) or 0.3µm thick (C) are shown. (A,B) Immunolabelling of CI-MPR in an uninfected macrophage (A) and in an infected macrophage (B). CI-MPR isfound in perinuclear vesicles and in vesicles dispersed in the cytoplasm. Arrows in (B) point to PV that appear devoid of CI-MPR. (C)Uninfected macrophage stained for macrosialin. Heavily stained vesicles are detected in this cell. (F) Two serial optical sections of an infectedmacrophage stained for macrosialin, the upper micrograph corresponding to the basal level of the cell. Note the presence of this protein on theperiphery of PV (arrows). (D,E) Two serial sections of an uninfected macrophage (D) and of an infected macrophage (E) stained for rab7p. Themicrographs located on the left-hand sides of the panels correspond to the basal level of the cells. In the uninfected macrophage, the labelling isassociated with numerous vesicles close to the nucleus and with fewer and smaller vesicles far from the nucleus. In the infected macrophage,large juxtanuclear positive vesicles are still detected and PV are stained on their periphery (arrows). Bars, 5 µm.
2144
Kinetics of macrosialin, lgp120 and cathepsin Dappearance in PVPV very rapidly acquired the prelysosomal and lysosomalmembrane proteins macrosialin and lgp120 as shown in Fig. 4.At 2 hours post-infection, 70% and 90% of the infectedmacrophages already had macrosialin and lgp120, respec-tively, in their PV, and from 6 hours post-infection all infectedmacrophages had PV positive for both proteins. As theseproteins were not detected on the plasma membrane of themacrophages their presence in PV membrane very likelyresults from the rapid fusion of late endocytic compartmentswith the nascent PV. In about 40% of the infectedmacrophages, PV also contained cathepsin D early afterinfection. However, infected macrophages with cathepsin D-positive PV did not exceed 50% even at 48 hours post-infection. The absence of this enzyme in PV of about 50% ofcells could be real. Alternatively, membrane proteins likelgp120 and macrosialin could be rendered insoluble by thefixative much better than could soluble proteins or partlysoluble proteins like cathepsin D (Antoine et al., 1987; Prinaet al., 1990).
Expression and localization of MHC molecules inmacrophages infected for 48 hoursWe next examined the expression and subcellular localizationof MHC molecules in L. donovani-infected macrophages tosee whether they could potentially act as antigen-presentingcells (APC) for Leishmania-specific T lymphocytes and todefine the position of PV with regard to the Ag presentationprocesses.
Quantification and distribution of MHC class II moleculesrIFN-γ induced a strong expression of MHC class II moleculesin both uninfected and infected macrophages as shown by thepercentage increase in cells reacting with the M5/114 mAb(which recognizes both I-A and I-E MHC class II molecules),measured after such a treatment (Fig. 5). In these experiments,conditions for the detection of both intracellular and cellsurface class II molecules or for the detection of only cellsurface class II molecules gave similar percentages of positivecells, whatever the state, uninfected or infected, of themacrophages (Fig. 5). In addition, quantification of plasmamembrane-associated MHC class II molecules showed that, at48 hours post-infection, rIFN-γ-treated macrophages expressedon their cell surface similar levels of class II to similarly treateduninfected macrophages (Fig. 6). Collectively, these resultsindicated that L. donovani infection does not impede the quan-titative expression of MHC class II molecules on the cellsurface.
As already described, MHC class II molecules were detectedon the plasma membrane of uninfected macrophages as wellas in intracellular compartments of the endocytic pathway(Harding et al., 1990; Lang and Antoine, 1991, Harding andGeuze, 1992; Lang et al., 1994). After infection, intracellularMHC class II molecules exhibited an important redistributionand, at 48 hours post-infection, a large part of the intracellularstaining was now associated with all the periphery of most PVas shown in Fig. 7A-C (arrows). In 20 to 30% of positive PV,however, staining appeared strictly located at only part of theirmembrane (Fig. 7A, arrowheads). Interestingly, PV-associatedclass II could be labelled with: (i) the M5/114 mAb, which rec-ognizes all types of molecules, whatever their conformation(Fig. 7A); (ii) the anti-I-Aαb Y-3P mAb (C57BL/6) (Fig. 7B)or the anti-I-Eαd 14-4-4S mAb (Balb/c) (Fig. 7C), which havebeen described as being much more reactive with the compact
T. Lang and others
Fig. 4. Kinetics of macrosialin (
h), lgp120 (.), cathepsin D (n) andMHC class II molecule (d) appearance in PV. Macrophages fromBalb/c mice were pre-incubated with rIFN-γ for 24 hours. Then, cellswere put into contact with amastigotes for 2 hours, washed andreincubated in fresh medium containing rIFN-γ. At various timeslater, cells were fixed, permeabilized and labelled with the Ab and anappropriate FITC-conjugate. Cell preparations were examined undera conventional fluorescence microscope and the percentages ofmacrophages with macrosialin, lgp120, cathepsin D or MHC class II-positive PV were determined after counting about 500 cells. For eachmarker, results were normalized, taking as 100% the total number ofstained infected macrophages counted. Whatever the time afterinfection, all infected macrophages were positive for macrosialin,lgp120 and cathepsin D, but only 70 to 80% of infected macrophageswere positive for class II molecules.
Fig. 5. Percentage of uninfected and infected macrophagesexpressing MHC class II molecules. Macrophages fom Balb/c micewere treated or not with rIFN-γ as described in the legend to Fig. 1.At 48 hours post-infection, cells were fixed, permeabilized (whiteand black bars) or not (hatched and cross-hatched bars) with saponinand incubated successively with the M5/114 Ab and an appropriateFITC-conjugate. Uninfected cells run in parallel were similarlyprocessed and used as positive controls. Each bar is the mean ± 1 s.d.of three experiments. In each experiment, the percentage of positivecells was estimated after counting about 500 cells.
2145Leishmania donovani-infected macrophages
conformers of MHC class II carrying tightly associatedpeptides (Germain and Hendrix, 1991). These last findings alsoshowed that both I-A and I-E molecules could reach PV.
Kinetics of MHC class II molecule appearance in PVThe acquisition of MHC class II molecules by PV was deter-mined in macrophages treated with rIFN-γ before adding theparasites, and stained for class II at various times afterinfection. As shown in Fig. 4, class II molecules reached PVapparently much more slowly than macrosialin and lgp120.Thus, at 2 hours post-infection less than 20% of class II-positive macrophages had class II in their PV and it took about2 days for all positive macrophages also to exhibit positive PV,whereas during this time the percentage of infectedmacrophages and of macrophages expressing MHC class IImolecules remained similar (data not shown). In macrophagesinduced for the expression of MHC class II after infection,most PV also contained class II molecules at 48 hours post-infection (data not shown). These results, together with theresults presented in Fig. 4, clearly demonstrate that a large partof the PV-associated MHC class II molecules, if not all thesemolecules, were not plasma membrane-bound moleculestrapped during internalization of the parasites.
Distribution of the invariant chains Ii in infectedmacrophagesThe distribution of the Ii chains, which transiently associatewith the MHC class II molecules during their intracellulartransport, was also analysed in order to determine whether Iiaccompanied MHC class II molecules during their transloca-tion into PV. These molecules serve at least two functions inAg presentation: (i) they prevent class II molecules frombinding peptides in the rough endoplasmic reticulum (RER);
and (ii) they contain in their cytoplasmic domain a topogenicsignal that allows the targeting of the (class II-Ii) complexes toendocytic compartments. In the latter, Ii is progressivelydegraded from the luminal COOH terminus, a process that isrequired for MHC class II molecules to acquire the ability tobind peptide (for a review see Brodsky and Guagliardi, 1991;Neefjes and Ploegh, 1992; Teyton and Peterson, 1992). Usingthe anti-Ii In-1 mAb, which recognizes a cytoplasmic epitope,Ii could be localized in the perinuclear region and in a reticularnetwork dispersed throughout the cytoplasm of almost allmacrophages stimulated with rIFN-γ (Fig. 7D,E). The samepattern was observed in uninfected and infected macrophagesand we could not detect any labelling in PV (Fig. 7E). Double-labellings of infected macrophages with In-1 and the anti-classII Y-3P mAb confirmed this last finding. Indeed, PV-associ-ated class II molecules detected with Y-3P could not be co-stained with In-1 (Fig. 7F). Unfortunately, the P4H5 mAb thatrecognizes a luminal epitope of Ii, although working well inimmunoprecipitation experiments, did not stain macrophageswhatever fixation procedure was used (fixation as described inMaterials and Methods, fixation with 4% paraformaldehyde for15 minutes at room temperature or fixation with a mixture ofmethanol/acetone (50/50, v/v) for 20 minutes at −20˚C). Nev-ertheless, as the membrane segment and the cytoplasmicdomain of class II-associated Ii are the last to disappear duringthe degradation process of this polypeptide (Pieters et al.,1991), results obtained with In-1 very likely indicate that PV-associated MHC class II molecules are devoid of Ii chains.Altogether, these data argue for a complete and rapid prote-olytic cleavage of Ii in PV or taking place upstream of PV.
Distribution of MHC class I moleculesFinally, we compared the distribution of MHC class Imolecules in uninfected vs L. donovani-infected macrophages.In these studies, the two mAb K9.18 and 34-4-20 S recogniz-ing the H-2Kd and H-2Dd molecules, respectively, were used.Both gave a signal on uninfected and infected macrophagesunstimulated with rIFN-γ but the MHC class I moleculeexpression was markedly increased after treatment with rIFN-γ. In both conditions, class I molecules were localized on theplasma membrane of uninfected and infected macrophages(Fig. 7G) but PV were unlabelled (Fig. 7H).
DISCUSSION
PV as prelysosomal/lysosomal compartmentsThe early work of Chang and Dwyer (1976, 1978) and Chang(1980) demonstrated that PV containing L. donovani maintaina strongly acidic pH and are capable of fusing with compart-ments that appeared to be lysosomes. In the present study, weprovide evidence that, rapidly after infection, these organellesare enriched for the lysosomal glycoprotein lgp120 and containvarious lysosomal proteases (cathepsins B, D, H and L) in theirlumen. Collectively, these data strongly suggest that veryquickly after phagocytosis the PV acquires the characteristicsof a lysosomal compartment, in terms of content andmembrane. Very similar conclusions were also drawn fromstudies on phagosomes containing either degradable Bacillussubtilis (Lang et al., 1988) or heat-killed Listeria monocyto-
0 0
00100 101 102 103 104 100 101 102 103 104
100 101 102 103 104 100 101 102 103 104
Num
ber
of c
ells
Fluorescence intensity
B
C
D
Control −ΙFN-γ
M5/114 −ΙFN-γ
Control +ΙFN-γ
M5/114 +ΙFN-γ
A
Fig. 6. Flow cytometric analysis of MHC class II moleculeexpression on the plasma membrane of uninfected (dotted lines) andinfected (continuous lines) macrophages, untreated (A,B) or treated(C, D) with rIFN-γ. At 48 hours of infection, live cells weresuccessively incubated at 4˚C with normal rat IgG (control) (A,C) orM5/114 mAb (B,D) and with an appropriate FITC-conjugate.Uninfected cells were similarly processed in parallel. Data arerepresentative of 3 experiments.
2146
genes (Harding and Geuze, 1992). Others features of L.donovani-containing PV are however not easily reconcilablewith the previous proposal: namely, the presence in and ontheir membrane of macrosialin and rab7p, respectively, previ-ously identified as markers of the late endosomes/prelyso-somes (Chavrier et al., 1990; Rabinowitz et al., 1992). Clearly,additional studies will be required to elucidate completely PVbiogenesis and at this point only tentative explanations can beproposed: (1) that rapidly after phagocytosis and after theirfusion with prelysosomes, PV housing L. donovani are sur-rounded by a membrane of the prelysosomal type; or (2) thatPV are formed by the coalescence of prelysosomal andlysosomal compartments. Previous data suggest that these twocompartments are in dynamic equilibrium and exchangecontent and membrane through transient fusions or connec-tions (Tassin et al., 1990; Griffiths, 1992; Storrie and Deng,
1992; Rabinowitz et al., 1992). The presence in these struc-tures of non-degradable organisms like Leishmania amastig-otes could possibly modify this equilibrium, leading to theformation of a stable mixed organelle. This mixed structurewould be maintained as long as the parasites were alive.Results obtained with macrophages infected with L. amazo-nensis support the second explanation. PV housing this Leish-mania species exhibit a similar if not identical composition tothat of PV containing L. donovani (Antoine et al., 1987, 1990,1991; Prina et al., 1990; Lang et al., 1994) but their enlarge-ment is accompanied by a strong depletion of most of the hostcell organelles containing acid phosphatase and arylsulphatase,very likely prelysosomes and lysosomes (Barbieri et al., 1985),suggesting that both types of organelles are involved in PVformation. Whatever the explanation, the apparent absence ofCI-MPR in PV (this study; Antoine et al., 1991; Russell et al.,
T. Lang and others
2147Leishmania donovani-infected macrophages
1992) is difficult to explain because it has been repeatedlydescribed as being absent from the terminal lysosomes ofvarious cell types including macrophages (Brown et al., 1986;Griffiths et al., 1988; Geuze et al., 1988; Croze et al., 1989;Rabinowitz et al., 1992; Harding and Geuze, 1992) and as amarker of late endosomes/prelysosomes (Griffiths et al., 1988;Griffiths, 1992). CI-MPR could be located in specializedportions of the prelysosomal compartment not involved in theformation of PV membrane (Rabinowitz et al., 1992) or elsein other types of vesicular compartments. It is interesting tonote in this respect that, like our present results, the recentstudies of Racoosin and Swanson (1993) dealing with the fateof macropinosomes in mouse bone marrow-derivedmacrophages are compatible with the idea that rab7p and CI-MPR are located in distinct compartments in this cell type.
L. donovani-infected macrophages can expressMHC class II molecules after stimulation with rIFN-γThe resolution or the uncontrolled development of diseasecaused by Leishmania results, at least partially, from theinduction of T lymphocyte-mediated immunity and, morespecifically, subsets of CD4+ T cells have been shown to playa major role in the outcome of infection. Interaction of parasiteAg with APC expressing MHC class II molecules is requiredfor the induction of Leishmania-specific CD4+T cells (see, forreviews: Liew, 1989; Müller et al., 1989; Locksley et al., 1991;Locksley and Louis, 1992) but the type(s) of APC involved inthese diseases has (have) not yet been specified. It is interest-ing to note in this context that activated macrophages contain-ing L. donovani or L. major promastigotes can present parasiteAg as well as third-party Ag targeted to the PV using geneti-cally transformed parasites (Lang and Kaye, 1991; Kaye et al.,1993). However, under the conditions used in these studies,parasites were rapidly killed after their phagocytosis. So,whether truly infected macrophages can act as APC remains tobe determined. To approach this question, we have examinedin the present work the ability of Leishmania-infectedmacrophages to express MHC class II molecules. We wereunable to demonstrate any effect of infection either on the per-centage of macrophages stained for MHC class II molecules oron the quantitative expression of these molecules on the cellsurface. Similar findings were reported from the earlier studyof Kaye et al. (1988). Likewise, L. amazonensis infection ofmouse macrophages does not alter the induction of MHC classII by IFN-γ or their quantitative expression on the plasma
membrane of infected cells, at least during the first 48 hoursfollowing the phagocytosis of the parasites (Antoine et al.,1991; Prina et al., 1993; Lang et al., 1994). In contrast, a partialunresponsiveness to IFN-γ in the expresssion of class IImolecules has been shown by Reiner et al. (1987, 1988) in L.donovani-infected macrophages. The origin of this discrepancyis not known but an inability of Leishmania-infectedmacrophages to express MHC class II molecules when appro-priately stimulated does not seem a general way for theparasites to escape the immune responses.
PV of L. donovani-infected macrophages stimulatedwith rIFN-γ contain MHC class II moleculesMany questions remain regarding the intracellular routesfollowed by Ag and MHC class II molecules for class II-restricted Ag presentation to CD4+ T lymphocytes. Immuno-cytochemical data on non-phagocytic and phagocytic cellsindicate that MHC class II molecules are localized in variouscompartments of the endocytic pathway, including earlyendosomes, late endosomes/prelysosomes and lysosomes(Guagliardi et al., 1990; Brodsky and Guagliardi, 1991; Peterset al., 1991; Lang and Antoine, 1991; Neefjes and Ploegh, 1992;Harding and Geuze, 1992). The most recent studies favour theprelysosomal/lysosomal compartments as the most attractivecandidates for the formation of (Ag-MHC class II) complexes.The present data also agree with this proposal. Indeed, we foundthat in macrophages infected with L. donovani and activated byrIFN-γ: (i) PV that exhibit features of the prelysosomal/lysosomal compartments are strongly stained for MHC class IImolecules; (ii) PV-associated class II molecules are apparentlydevoid of Ii and are thus potentially able to bind peptides; (iii)finally, PV-associated class II molecules can bind conforma-tional Ab that preferentially recognize class II molecules withtightly associated peptides. Whether class II located in PV arefunctionally involved in the presentation of L. donovani Agremains however to be determined.
Kinetics experiments showed that MHC class II moleculesreach PV in a few hours but no so quickly however as lgp120and macrosialin, which is quite intriguing, since in uninfectedmacrophages lgp120 and class II molecules apparently co-localize in a large number of organelles (data not shown). Dif-ferences in the turnover of lgp120, macrosialin and class IImolecules located in PV and variations in the class II turnoverwith time after infection could explain these findings.
Within PV, MHC class II molecules were distributed either
Fig. 7. Localization of MHC class II molecules, of invariant chains and of MHC class I molecules in infected macrophages from Balb/c mice(A,C,D,E,G,H) and from C57Bl/6 mice (B,F). Macrophages were infected and cultured in the presence of rIFN-γ as described in the legend toFig. 1. At 2 days of infection, cells were fixed, permeabilized, labelled with either M5/114, Y-3P, 14-4-4S, In-1 or 34-4-20S mAb and then withan appropriate FITC-conjugate (green staining, except in F). Parasite nuclei (stars) and macrophage nuclei (n) were visualized by propidiumiodide staining (red staining, except in F). (A,B,C) Infected macrophages stained for MHC class II molecules with M5/114 (A), Y-3P (B), 14-4-4S (C) mAb. In (A) and (B,C) optical sections 0.8 µm and 0.3 µm thickness, respectively, are shown. MHC class II molecules are localized in asmall number of vesicles and at the periphery of all PV (arrows). In one PV, class II staining is restricted to part of this organelle and thusappears as a half-circle (A, arrowheads). (D,E) Infected macrophage stained with the anti-Ii In-1 mAb. (D) and (E) represent a 3-dimensionalreconstruction and an optical section (0.3 µm thickness), respectively, of the same infected cell. The Ii staining surrounds PV but no labellingassociated with the PV membrane can be detected. PV appear as holes in the cytoplasm (E, arrowheads). (F) Infected macrophage double-stained with the Y-3P and In-1 mAb. Texas Red-labelled goat anti-mouse Ig Ab and FITC-labelled goat anti-rat Ig F(ab′)2 were used to detectY-3P and In-1, respectively. No Ii staining can be detected at the level of the PV; MHC class II-specific staining is associated with the plasmamembrane (large arrows) and with PV (small arrows). (G,H) Macrophages stained for the class I molecule H-2Dd with the 34-4-20S mAb.(G) Three-dimensional reconstruction of an uninfected macrophage (left cell) and of an infected macrophage (right cell). Strong plasmamembrane labelling of both cells is observed (large arrows). (H) Optical section (0.3 µm thickness) of the infected macrophage shown in (G).No labelling can be detected in PV (arrowheads). Bars, 5 µm.
2148
homogeneously or in a polarized fashion. In some largevacuoles containing several parasites, the polarized class IIstaining appeared associated with the binding sites of amastig-otes to the PV membrane. Such a distribution of the MHC classII molecules was already observed in PV of Balb/c andC57BL/6 macrophages housing L. amazonensis (Antoine et al.,1991). However, whereas this phenomenon seems relativelyuncommon in L. donovani-containing PV, it occurs in a largemajority of PV containing L. amazonensis (Lang et al., 1994).Its significance is not yet known but a plausible hypothesiscould be that MHC class II molecules bind to plasmamembrane components of amastigotes, and we are currentlyattempting to test this theory.
PV housing L. donovani lack significant amounts ofMHC class I moleculesA large amount of data now suggest that CD8+ T cells thatrecognize processed Ag bound to MHC class I molecules makea significant contribution to anti-leishmania immunity (Mülleret al., 1991; Hill et al., 1989) including protective immunitydirected against L. donovani (Stern et al., 1988; Kaye et al.,1992). It is generally thought that Ag presented via the MHCclass I molecules are degraded in the cytosol or in the biosyn-thetic pathway of the APC. Cytosolic peptides are then trans-ported into the endoplasmic reticulum where they associatewith the MHC class I molecules before being driven to the cellmembrane (for a review, see Monaco, 1992). How Leish-mania Ag bind to class I molecules is thus an intriguingquestion, since in contrast to the situation with intracellularpathogens like Listeria monocytogenes, which induce theformation of CD8+ T lymphocytes but which leave theendocytic compartments soon after phagocytosis and reach thecytosol of infected cells (Kaufmann et al., 1986; Brunt et al.,1990), Leishmania remain within vacuolar compartments.
The access of MHC class I molecules to phagocytic com-partments could explain this apparently new pathway of Agpresentation. However, whereas plasma membrane of infectedmacrophages was labelled with both anti-H-2Kd and anti-H-2Dd mAb, no labelling could be detected in PV whatever theAb used. It thus seems that if infected macrophages can act asAPC for Leishmania-specific CD8+ T cells, Leishmania Agand MHC class I molecules most likely do not meet in PV,with the reservation that the threshold of sensitivity of ourimmunocytochemical procedure would not allow the detectionof very small amounts of functional class I molecules in theseorganelles. Alternative explanations remain, however, moreplausible, like the leakage towards the cytosol of parasite Agpresent in PV or the trapping of parasite Ag within vesiclesshuttling between PV and compartments of the biosyntheticpathway.
In conclusion, L. donovani, L. amazonensis and L. mexicanaamastigotes live in hostile compartments of the macrophagesthat, with the exception of size, exhibit very similar properties(Antoine et al., 1987, 1990, 1991; Prina et al., 1990; Russellet al., 1992; Lang et al., 1994; the present study). This impliesthat the biology of the amastigote stages of these differentLeishmania species must be similar. Additionally, after theirinfection with these Leishmania species, mouse bone marrow-derived macrophages remain responsive to IFN-γ, whichinduces especially the expression of MHC class II moleculesand increases the level of expression of MHC class I
molecules, the key molecules for the Ag presentation processes(Antoine et al., 1991; Russell et al., 1992; Lang et al., 1994;the present study). Of particular interest is the presence ofMHC class II molecules and the apparent absence of MHCclass I molecules in PV. Such events could also occur in vivo,since the expression of MHC class II molecules by infectedmacrophages has been already observed in murine infectioussites, including those of susceptible mice (McElrath et al.,1987; Russell et al., 1992; T. Lang and J.-C. Antoine, unpub-lished results), which raises the question of the escape of theparasites from the Ag presentation process. This will be ofinterest for future study.
This work was supported by the Institut Pasteur, the UNDP/WorldBank/WHO Special Programme for Research and Training inTropical Diseases (grant no. ID 890422) and by the British MedicalResearch Council. We thank M.-C. Prevost for invaluable technicalassistance. The authors also thank G. R. Adolf for the gift of IFN-γand P. Chavrier, B. Goud, J.-L. Guesdon, U. Hammerling, B. Hoflack,G. Koch, N. Koch, O. Léo, J. Mehringer, I. Mellman, M. Pierres, N.Plasman and B. Vray, who have kindly donated immune sera, anti-bodies and cell hybridomas.
REFERENCES
Alexander, J. and Vickerman K. (1975). Fusion of host cell secondarylysosomes with the parasitophorous vacuoles of Leishmania mexicana-infected macrophages. J. Protozool. 22, 502-508.
Antoine, J.-C., Jouanne, C., Ryter, A. and Zilberfarb, V. (1987).Leishmania mexicana: a cytochemical and quantitative study of lysosomalenzymes in infected rat bone marrow-derived macrophages. Exp. Parasitol.64, 485-498.
Antoine, J.-C., Prina, E., Jouanne, C. and Bongrand, P. (1990).Parasitophorous vacuoles of Leishmania amazonensis-infected macrophagesmaintain an acidic pH. Infect. Immun. 58, 779-787.
Antoine, J.-C., Jouanne, C., Lang, T., Prina, E., de Chastellier, C. andFréhel, C. (1991). Localization of major histocompatibility complex class IImolecules in phagolysosomes of murine macrophages infected withLeishmania amazonensis. Infect. Immun. 59, 764-775.
Barbieri, C. L., Brown, K. and Rabinovitch, M. (1985). Depletion ofsecondary lysosomes in mouse macrophages infected with Leishmaniamexicana amazonensis: a cytochemical study. Z. Parasitenk. 71, 159-168.
Bhattacharya, A., Dorf, M. E. and Springer, T. A. (1981). A sharedalloantigenic determinant on Ia antigens encoded by the I-A and I-Esubregions: evidence for I region gene duplication. J. Immunol. 127, 2488-2495.
Brodsky, F. M. and Guagliardi, L. E. (1991). The cell biology of antigenprocessing and presentation. Annu. Rev. Immunol. 9, 707-744.
Brown, W. J., Goodhouse, J. and Farquahr, M. G. (1986). Mannose-6-phosphate receptors for lysosomal enzymes cycle between the Golgicomplex and endosomes. J. Cell Biol. 103, 1235-1247.
Brunt, L. M., Portnoy, D. A. and Unanue, E. R. (1990). Presentation ofListeria monocytogenes to CD8+ T cells requires secretion of hemolysin andintracellular bacterial growth. J. Immunol. 145, 3540-3546.
Chang, K.-P. and Dwyer, D. M. (1976). Multiplication of a human parasite(Leishmania donovani) in phagolysosomes of hamster macrophages in vitro.Science 193, 678-680.
Chang, K.-P. and Dwyer, D. M. (1978). Leishmania donovani. Hamstermacrophage interactions in vitro: cell entry, intracellular survival, andmultiplication of amastigotes. J. Exp. Med. 147, 515-530.
Chang, K.-P. (1980). Endocytosis of Leishmania-infected macrophages,fluorometry of pinocytic rate, lysosome-phagosome fusion andintralysosomal pH. In The Host Invader Interplay (ed. H. Van den Bossche),pp. 231-234. Elsevier/North-Holland Biomedical Press, Amsterdam.
Chang, K.-P. (1983). Cellular and molecular mechanisms of intracellularsymbiosis in leishmaniasis. Int. Rev. Cytol. suppl. 14, 267-305.
Channon, J. Y., Roberts, M. B. and Blackwell, J. M. (1984). A study of thedifferential respiratory burst activity elicited by promastigotes and
T. Lang and others
2149Leishmania donovani-infected macrophages
amastigotes of Leishmania donovani in murine resident peritonealmacrophages. Immunology 58, 345-355.
Chavrier, P., Parton, R. G., Hauri, H. P., Simons, K. and Zerial, M. (1990).Localization of low molecular weight GTP binding proteins to exocytic andendocytic compartments. Cell 62, 317-329.
Croze, E., Ivanov, I. E., Kreibich, G., Adesnik, M., Sabatini, D. D. andRosenfeld, M. G. (1989). Endolyn-78, a membrane glycoprotein present inmorphologically diverse components of the endosomal and lysosomalcompartments: implications for lysosome biogenesis. J. Cell Biol. 108, 1597-1613.
Diment, S., Leech, M. S. and Stahl, P. D. (1988). Cathepsin D is membrane-associated in macrophage endosomes. J. Biol. Chem. 263, 6901-6907.
Germain, R. N. and Hendrix, L. R. (1991). MHC class II structure, occupancyand surface expression determined by post-endoplasmic reticulum antigenbinding. Nature 353, 134-139.
Geuze, H. J., Stoorvogel, W., Strous, G. J., Slot, J. W., Bleekemolen, J. E. andMellman, I. (1988). Sorting of mannose 6-phosphate receptors and lysosomalmembrane proteins in endocytic vesicles. J. Cell Biol. 106, 2491-2501.
Goldstein, P., Goridis, C., Schmitt-Verhulst, A.-M., Hayot, B., Pierres, A.,Van Agthoven, A., Kaufman, Y., Esshar, Z. and Pierres, M. (1982).Lymphoid cell surface interaction structures detected using cytolysis-inhibiting monoclonal antibodies. Immunol. Rev. 68, 5-42.
Griffiths, G., Hoflack, B., Simons, K. and Mellman, I. (1988). The mannose6-phosphate receptor and the biogenesis of lysosomes. Cell 52, 329-341.
Griffiths, G. (1992). The compartments of the endocytic pathway. InEndocytosis. From Cell Biology to Health, Disease and Therapy, NATO ASISeries, vol. H62 (ed. P. J. Courtoy), pp. 73-83. Springer-Verlag, Berlin,Heidelberg.
Guagliardi, L. E., Koppelman, B., Blum, J. S., Marks, M. S., Cresswell, P.and Brodsky, F. M. (1990). Co-localization of molecules involved inantigen processing and presentation in an early endocytic compartment.Nature 343, 133-139.
Harding, C. V., Unanue, E. R., Slot, J. W., Schwartz, A. L. and Geuze H. J.(1990). Functional and ultrastructural evidence for intracellular formation ofmajor histocompatibility complex class II-peptide complexes during antigenprocessing. Proc. Nat. Acad. Sci. USA 87, 5553-5557.
Harding, C. V. and Geuze, H. J. (1992). Class II MHC molecules are presentin macrophage lysosomes and phagolysosomes that function in thephagocytic processing of Listeria monocytogenes for presentation to T cells.J. Cell Biol. 119, 531-542.
Hill, J. O., Awwad, M. and North, R. J. (1989). Elimination of CD4+
suppressor T cells from susceptible Balb/c mice releases CD8+ Tlymphocytes to mediate protective immunity against Leishmania. J. Exp.Med. 169, 1819-1827.
Janeway, C. A. Jr, Conrad, P. J., Lerner, E. A., Babich, J., Wettstein, P.and Murphy, D. B. (1984). Monoclonal antibodies specific for Iaglycoproteins raised by immunization with activated T cells: possible role ofT cell-bound Ia antigens as targets of immunoregulatory T cells. J. Immunol.132, 662-667.
Kaufmann, S. H. E., Hug, E. and De Libero, G. (1986). Listeriamonocytogenes-reactive T lymphocyte clones with cytolytic activity againstinfected target cells. J. Exp. Med. 164, 363-368.
Kaye, P. M., Patel, N. K. and Blackwell, J. M. (1988). Acquisition of cell-mediated immunity to Leishmania. II. Lsh gene regulation of accessory cellfunction. Immunology 65, 17-22.
Kaye, P. M., Cooke, A., Lund, T., Wattie, M. and Blackwell, J. M. (1992).Altered course of visceral leishmaniasis in mice expressing transgenic I-Emolecules. Eur. J. Immunol. 22, 357-364.
Kaye, P. M., Coburn, C., McCrossan, M. and Beverley, S. M. (1993).Antigens targeted to the Leishmania phagolysosome are processed for CD4+
T cell recognition. Eur. J. Immunol. 23, 2311-2319. Kiderlen, A. F. and Kaye, P. M. (1990). A modified colorimetric assay of
macrophage activation for intracellular cytotoxicity against Leishmaniaparasites. J. Immunol. Meth. 127, 11-18.
Knapp, P. E. and Swanson, J. A. (1990). Plasticity of the tubular lysosomalcompartment in macrophages. J. Cell Sci. 95, 433-439.
Koch, N., Koch, S. and Hammerling, G. J. (1982). Ia invariant chain detectedon lymphocyte surfaces by monoclonal antibody. Nature 299, 644-645.
Lang, T., de Chastellier, C., Ryter, A. and Thilo, L. (1988). Endocyticmembrane traffic with respect to phagosomes in macrophages infected withnon-pathogenic bacteria: phagosomal membrane acquires the samecomposition as lysosomal membrane. Eur. J. Cell Biol. 46, 39-50.
Lang, T. and Antoine, J.-C. (1991). Localization of MHC class II molecules inmurine bone marrow-derived macrophages. Immunology 72, 199-205.
Lang, T. and Kaye, P. M. (1991). Presentation of Leishmania donovanipromastigotes occurs via a brefeldin A-sensitive pathway. Eur. J. Immunol.21, 2407-2413.
Lang, T., de Chastellier, C., Fréhel, C., Hellio, R., Metezeau, P., de SouzaLeão, S. and Antoine, J.-C. (1994). Distribution of MHC class I and ofMHC class II molecules in macrophages infected with Leishmaniaamazonensis. J. Cell Sci. 107, 69-82.
Lewis, V., Green, S. A., Marsh, M., Vihko, P., Helenius, A. and Mellman, I.(1985). Glycoproteins of the lysosomal membrane. J. Cell Biol. 100, 1839-1847.
Liew, F. Y. (1989). Functional heterogeneity of CD4+ T cells in leishmaniasis.Immunol. Today 10, 40-45.
Locksley, R. M., Heinzel, F. P., Holaday, B. J., Mutha, S. S., Reiner, S. L.and Sadick, M. D. (1991). Induction of TH1 and TH2 CD4+ subsets duringmurine Leishmania major infection. Res. Immunol. 142, 28-32.
Locksley, R. M. and Louis, J. A. (1992). Immunology of leishmaniasis. Curr.Opin. Immunol. 4, 413-418.
McElrath, M. J., Kaplan, G., Nusrat, A. and Cohn, Z. A. (1987). Cutaneousleishmaniasis. The defect in T cell influx in Balb/c mice. J. Exp. Med. 165,546-559.
Monaco, J. J. (1992). A molecular model of MHC class-I-restricted antigenprocessing. Immunol. Today 13, 173-179.
Müller, I., Pedrazzini, T., Farrel, J. P. and Louis, J. (1989). T-cell responsesand immunity to experimental infection with Leishmania major. Annu. Rev.Immunol. 7, 561-578.
Müller, I., Pedrazzini, T., Kropf, P., Louis, J. and Milon, G. (1991).Establishment of resistance to Leishmania major infection in susceptileBalb/c mice requires parasite-specific CD8+ T cells. Int. Immunol. 3, 587-597.
Neefjes, J. J. and Ploegh, H. L. (1992). Intracellular transport of MHC class IImolecules Immunol. Today 13, 179-184.
Ozato, K., Mayer, N. M. and Sachs, D. H. (1980). Hybridoma cell linessecreting monoclonal antibodies to mouse H-2 and Ia antigens. J. Immunol.124, 533-540.
Ozato, K., Mayer, N. M. and Sachs, D. H. (1982). Monoclonal antibodies tomouse major histocompatibility complex antigens. IV. A series of hybridomaclones producing anti-H-2d antibodies and an examination of expression ofH-2d antigens on the surface of these cells. Transplantation 34, 113-120.
Peters, P. J., Neefjes, J. J., Oorschot, V., Ploegh, H. L. and Geuze, H. J.(1991). Segregation of MHC class II molecules from MHC class I moleculesin the Golgi complex for transport to lysosomal compartments. Nature 349,669-676.
Pieters, J., Horstmann, H., Bakke, O., Griffiths, G. and Lipp, J. (1991).Intracellular transport and localization of major histocompatibility complexclass II molecules and associated invariant chain. J. Cell Biol. 155, 1213-1223.
Prina, E., Antoine, J.-C., Wiederanders, B. and Kirschke, H. (1990).Localization and activity of various lysosomal proteases in Leishmaniaamazonensis-infected macrophages. Infect. Immun. 58, 1730-1737.
Prina, E., Jouanne, C., de Souza Leão, S., Szabo, A., Guillet, J.-G. andAntoine, J.-C. (1993). Antigen presentation capacity of murinemacrophages infected with Leishmania amazonensis amastigotes. J.Immunol. 151, 2050-2061.
Rabinovitch, M., Topper, G., Cristello, P. and Rich, A. (1985). Receptor-mediated entry of peroxidases into the parasitophorous vacuoles ofmacrophages infected with Leishmania mexicana amazonensis. J. Leuk.Biol. 37, 247-261.
Rabinowitz, S., Horstmann, H., Gordon, S. and Griffiths, G. (1992).Immunocytochemical characterization of the endocytic and phagolysosomalcompartments in peritoneal macrophages. J. Cell Biol. 116, 95-112.
Racoosin, E. L. and Swanson, J. A. (1993). Macropinosome maturation andfusion with tubular lysosomes in macrophages. J. Cell Biol. 121, 1011-1020.
Reiner, N. E., Ng, W. and McMaster, W. R. (1987). Parasite-accessory cellinteractions in murine leishmaniasis. II. Leishmania donovani suppressesmacrophage expression of class I and class II major histocompatibilitycomplex gene products. J. Immunol. 138, 1926-1932.
Reiner, N. E., Ng, W., Ma, T. and McMaster, W. R. (1988). Kinetics ofgamma-interferon binding and induction of major histocompatibilitycomplex class II mRNA in Leishmania donovani-infected macrophages.Proc. Nat. Acad. Sci. USA 85, 4330-4334.
Russell, D. G., Xu, S. M. and Chakraborty, P. (1992). Intracellulartrafficking and the parasitophorous vacuole of Leishmania mexicana-infected macrophages. J. Cell Sci. 103, 1193-1210.
Smith, M. J. and Koch, G. L. E. (1987). Differential expression of murinemacrophage surface glycoprotein antigens in intracellular membranes. J.Cell Sci. 87, 113-119.
2150
Stern, J. J., Oca, M. J., Rubin, B. Y., Anderson, S. L. and Murray, H. W.(1988). Role of L3T4+ and Lyt-2+ cells in experimental visceralleishmaniasis. J. Immunol. 140, 3971-3977.
Storrie, B. and Deng, Y. (1992). Protein exchange within the lysosome andpre-lysosome compartment: a mechanism for maintaining organellefunctionality. In Endocytosis. From Cell Biology to Health, Disease andTherapy, NATO ASI series, vol. H62 (ed. P. J. Courtoy), pp. 85-90. Springer-Verlag, Berlin, Heidelberg.
Swanson, J., Bushnell, A. and Silverstein, S. C. (1987). Tubular lysosomesmorphology and distribution within macrophages depend on the integrity ofcytoplasmic microtubules. Proc. Nat. Acad. Sci. USA 84, 1921-1925.
Tassin, M.-T., Lang, T., Antoine, J.-C., Hellio, R. and Ryter, A. (1990).Modified lysosomal compartment as carrier of slowly and non-degradabletracers in macrophages. Eur. J. Cell Biol. 52, 219-228.
Teyton, L. and Peterson, P. A. (1992). Invariant chain-A regulator of antigenpresentation. Trends Cell Biol. 2, 52-56.
Wiederanders, B. and Kirschke, H. (1986). Antibodies to rat liver cathepsins:characterization and use for the identification of enzyme precursors. Biomed.Biochim. Acta 45, 1421-1431.
