Proteolytic activation of leishporin: evidence that Leishmania amazonensis and Leishmania guyanensis have distinct inactive forms

Crude extracts of Leishmania amazonensis, but not of L. guyanensis, are lytic to erythrocytes and nucleated cells, including macrophages. L. amazonensis-mediated lysis is caused by a membrane-associated pore-forming protein, named a-leishporin. Here we show that L. amazonensis, but not L. guyanensis, promastigote extracts increase their hemolytic activity when kept at 4 degrees C for a few days or at 37 degrees C for a few hours. We show that the activation in the extracts is mediated by a cytosolic serine-protease. Although L. guyanensis extracts are hemolytically inactive and unable to generate hemolytic activity, their membrane fraction becomes hemolytic in the presence of the cytosolic fraction of L. amazonensis, also by the action of a serine-protease. This suggests that L. guyanensis contains a potential lytic molecule, named here g-leishporin. The cytosolic fraction of L. guyanensis is unable to activate either a- or g-leishporin, indicating that this species does not possess the protease(s) that activate(s) the cytolysin. Trypsin, chymotrypsin, collagenase, Pronase and proteinase K, are also effective in activating a-leishporin but not g-leishporin. This suggests that the inactive forms of a-leishporin and g-leishporin are distinct in structure and/or are activated by different mechanisms. We are considering two hypotheses for the activation of leishporins: (1) proteolysis of an inactive precursor and (2) dissociation and/or proteolytic degradation of an inhibitory oligopeptide. The present data and preliminary results argue for the second hypothesis. We speculate that leishporin could be activated in the protease-rich, low pH, and dissociating environment of parasitophorous vacuole contributing for the release of the parasites from the macrophage.     

A simple and reproducible method to obtain large numbers of axenic amastigotes of different<Emphasis Type="Italic"> Leishmania</Emphasis> species

This work describes a simple method to yield large amounts of Leishmania amastigote-like forms in axenic cultures using promastigotes as the starting population. The method described induced extracellular amastigote transformation of Leishmania amazonensis (97%), Leishmania braziliensis (98%) and Leishmania chagasi (90%). The rounded parasites obtained in axenic cultures were morphologically similar, even at the ultrastructural level, to intracellular amastigotes. Moreover, the axenic amastigotes remained viable as measured by their ability to revert back to promastigotes and to infect BALB/c mice. L. amazonensis and L. braziliensis promastigotes and axenic amastigotes differed in terms of their Western blot profiles. A 46 kDa protein was recognized by specific antibodies only in axenic and lesion-derived L. amazonensis amastigotes and not in promastigotes.     

In vitro responses of human peripheral blood mononuclear cells to whole-cell, particulate and soluble extracts of Leishmania promastigotes

Whole-cell and soluble extracts of Leishmania promastigotes have both been used as skin test antigens and have also been tested as vaccine candidates. However, the differences in antigenicity between soluble and particulate Leishmania fractions are not known. We evaluated in vitro responses of PBMC from 30 American tegumentary leishmaniasis (ATL) patients and seven noninfected donors to different antigen preparations from Leishmania promastigotes, namely Leishmania amazonensis and L. braziliensis whole-cell extracts, as well as soluble and particulate fractions of L. amazonensis. All Leishmania antigen preparations stimulated significantly higher proliferation and interferon (IFN)-gamma production (but not interleukin (IL)-10 production) in PBMC from the leishmaniasis patients than in cells from the control subjects. The L. braziliensis whole-cell extract stimulated significantly higher cell proliferation and IFN-gamma production than the L. amazonensis whole-cell extract in the group of patients but not in the control group. This result can be explained by the fact that the patients were infected with L. braziliensis. Again in the group of patients, the PBMC proliferative responses as well as the levels of IFN-gamma and IL-10 stimulated by L. amazonensis whole-cell extract were significantly greater than those elicited by the L. amazonensis soluble fraction but were not significantly different from those elicited by the L. amazonensis particulate fraction. We found a higher antigenicity of the particulate fraction as compared to the soluble fraction, what suggests that the antigens present in the particulate fraction account for most of the antigenicity of whole-cell Leishmania promastigote antigen extracts.     

Leishmania (L.) amazonensis peptidase activities inside the living cells and in their lysates

In this study we investigated the peptidase activity in Leishmania (L.) amazonensis live amastigote by confocal microscopy using peptidyl-MCA as substrates, the hydrolysis of which releases the MCA fluorophore inside the cells. Cell pre-treatment with peptidase inhibitors indicated the presence of cysteine and serine peptidases. It was noteworthy that Leishmania amastigotes incorporate only substrates (Z-FR-MCA, Z-RR-MCA) or inhibitors (E64, TLCK) containing positively charged groups. The peptidase activities in the supernatants of amastigotes and promastigotes lysates were also evaluated with the same peptidyl-MCA substrates and inhibitors in the pH range 4.5-9.0. The effects of temperature and different salts were also included in this study. The hydrolytic activities of supernatants on Z-FR-MCA clearly indicate the presence of different cysteine peptidases that adapted to work in different environment conditions. Intact Leishmania cells incorporated Z-RR-MCA, the hydrolysis of which was inhibited only by TLCK indicating the presence of at least one serine peptidase. The pH profile of Z-RR-MCA hydrolysis by amastigotes and promastigotes lysate supernatants, and the hydrolysis time course of the FRET peptide Abz-AGRRRAQ-EDDnp at RA bond, followed by removal of the two C-termini R to yield Abz-AGR-OH that is a unique characteristic of oligopeptidase B, indicate its presence in the parasite.     

Substrate-dependent pH optima of gp63 purified from seven strains of Leishmania.

The major surface glycoprotein of Leishmania, gp63, is a membrane-bound metalloprotease. Contradictory data supporting a neutral or acidic nature of this enzyme have been presented. Seven strains of Old and New World Leishmania, including Leishmania donovani complex (Leishmania infantum and L. donovani), Leishmania major, Leishmania tropica and Leishmania mexicana amazonensis were used for the purification and comparative study of gp63. The protein was extracted from promastigotes by phase separation in Triton X-114 and purified by anion exchange chromatography. In agreement with previous reports, all purified gp63 were found to be structurally and immunologically related. Both membrane-bound gp63, on the surface of promastigotes, and the purified proteases had optimal activity at neutral to alkaline pH on azocasein, whereas their activity was optimal at acidic to neutral pH against 125I-insulin B-chain. The IC50 concentrations of 1,10-phenathroline against the two substrates, at the optimal pH, were comparable, suggesting that both activities measured were associated with gp63 rather than another contaminating enzyme. This was further supported by the comparable enrichment values, estimated from the specific activity of the enzyme during purification, using both assays. These results explain the earlier apparent discrepancies and suggest that the optimum pH of gp63 is substrate-dependent and not related to species differences or to the different purification procedures applied.     

Glycosphingolipid antigens of Leishmania (Leishmania) amazonensis amastigotes identified by use of a monoclonal antibody.

Monoclonal antibodies directed against Leishmania (Leishmania) amazonensis amastigotes were produced. One monoclonal antibody (1C3) selected by indirect immunofluorescence reacted with both amastigotes and promastigotes of L. (L.) amazonensis. Glycolipid extraction from L. (L.) amazonensis amastigotes and separation by high-performance thin-layer chromatography followed by immunoblotting demonstrated that 1C3 reacts with two glycosphingolipids which migrate chromatographically similarly to ceramide-N-acetylneuraminic acid (GM1) and ceramide-N-tetrose-di-acetylneuraminic acid (GD1a). The antibody did not react with glycosphingolipids from L. (L.) amazonensis promastigotes. Immunoprecipitation of 125I- and 35S-methionine-labeled promastigotes demonstrated that 1C3 recognizes gp63 from L. (L.) amazonensis promastigotes. Biosynthetic incorporation of labeled lipids by L. (L.) amazonensis amastigotes indicated that the glycosphingolipids reactive with 1C3 contain oleic acid in their structures. Surface labeling with galactose oxidase and sodium boro[3H]hydride indicated that galactose is present in 1C3-reactive antigens, strongly suggesting that these glycosphingolipids are localized on the surface of L. (L.) amazonensis amastigotes. Inhibition experiments of macrophage infection implicated the 1C3-reactive glycosphingolipids from L. (L.) amazonensis amastigotes in Leishmania invasion. The role of gp63 in promastigote-macrophage attachment was also demonstrated by inhibition experiments performed with 1C3, consistent with data from the literature.     

Differential microbicidal effects of human histone proteins H2A and H2B on Leishmania promastigotes and amastigotes

Recent studies have shown that histone proteins can act as antimicrobial peptides in host defense against extracellular bacteria, fungi, and Leishmania promastigotes. In this study, we used human recombinant histone proteins to further study their leishmaniacidal effects and the underlying mechanisms. We found that the histones H2A and H2B (but not H1(0)) could directly and efficiently kill promastigotes of Leishmania amazonensis, L. major, L. braziliensis, and L. mexicana in a treatment dose-dependent manner. Scanning electron microscopy revealed surface disruption of histone-treated promastigotes. More importantly, the preexposure of promastigotes to histone proteins markedly decreased the infectivity of promastigotes to murine macrophages (Mφs) in vitro. However, axenic and lesion-derived amastigotes of L. amazonensis and L. mexicana were relatively resistant to histone treatment, which correlated with the low levels of intracellular H2A in treated amastigotes. To understand the mechanisms underlying these differential responses, we investigated the role of promastigote surface molecules in histone-mediated killing. Compared with the corresponding controls, transgenic L. amazonensis promastigotes expressing lower levels of surface gp63 proteins were more susceptible to histone H2A, while L. major and L. mexicana promastigotes with targeted deletion of the lipophosphoglycan 2 (lpg2) gene (but not the lpg1 gene) were more resistant to histone H2A. We discuss the influence of promastigote major surface molecules in the leishmaniacidal effect of histone proteins. This study provides new information on host innate immunity to different developmental stages of Leishmania parasites.     

Leishmania interaction with human monocytes and neutrophils: modulation of the chemotactic response

Infection of host cells with Leishmania, which are obligate parasites of mononuclear phagocytes, most probably involved chemotaxis of host cells towards the parasite. We have examined the chemotactic properties of a sonicate derived from L. mexicana amazonensis promastigotes for normal human peripheral blood monocytes and neutrophils. L. m. amazonensis sonicate exhibited chemotactic activity for monocytes and neutrophils. Treatment at 65 degrees C for 30 min, enhanced the activity for neutrophils but not for monocytes, while treatment at 100 degrees C for 60 min abolished the activity. Additional studies showed that the sonicate generated chemotactic activity in serum, presumably by activating the alternative complement pathway to produce C5a, for monocytes and neutrophils. Incubation of monocytes and neutrophils with the sonicate inhibited the chemotactic activity of these cells towards various chemoattractants. When the sonicate was heat-treated the inhibitory effect was lost, except when sonicate was used as a chemoattractant. These results indicate the presence of specific receptors for factor(s) from L. m. amazonensis promastigotes on human monocytes and neutrophils.     

Role of interleukin-1beta in activating the CD11c(high) CD45RB- dendritic cell subset and priming Leishmania amazonensis-specific CD4+ T cells in vitro and in vivo

Cutaneous leishmaniasis associated with Leishmania amazonensis infection is characterized by uncontrolled parasite replication and profound immunosuppression; however, the underlying mechanisms remain largely unclear. One possibility is that the L. amazonensis parasite modulates antigen-presenting cells, favoring the generation of pathogenic Th cells that are capable of recruiting leukocytes but insufficient to fully activate their microbicidal activities. To test this possibility, we infected bone marrow-derived dendritic cells (DCs) of C57BL/6 mice with L. amazonensis or Leishmania major promastigotes and assessed the activation of DC subsets and their capacity in priming CD4(+) T cells in vitro. In comparison to L. major controls, L. amazonensis-infected DCs secreted lower levels of interleukin-1alpha (IL-1alpha) and IL-1beta, were less potent in activating the IL-12p40-producing CD11c(high) CD45RB(-) CD83(+) CD40(+) DC subset, and preferentially activated CD4(+) T cells with a IFN-gamma(low) IL-10(high) IL-17(high) phenotype. Although the addition of IL-1beta at the time of infection markedly enhanced DC activation and T-cell priming, it did not skew the cytokine profile of DCs and pathogenic Th cells, as local injection of IL-1beta following L. amazonensis infection accelerated Th cell activation and disease progression. This study suggests that intrinsic defects at the level of DC activation are responsible for the susceptible phenotype in L. amazonensis-infected hosts and that this parasite may have evolved unique mechanisms to interfere with innate and adaptive immunity.     

Roles of free GPIs in amastigotes of Leishmania

Glycosylated phosphatidylinositols (GPIs) are abundant cell surface molecules of the Leishmania. Amastigote-specific GPIs AmGPI-Y and AmGPI-Z, both ethanolamine (EtN)-containing glycolipids, were identified in Leishmania amazonensis. A paucity of GPI-anchored proteins in amastigotes of L. amazonensis made the kinetoplastid suitable for evaluating the importance of free (i.e. unconjugated to protein or polysaccharide) GPIs. A strain deficient in both AmGPI-Y and AmGPI-Z was produced by stable transfection of wild-type Leishmania with a GPI-phospholipase C gene. Phosphatidylinositol deficiency was not detected in the transfectants. GPI-deficient promastigotes infected murine macrophages in vitro and differentiated into amastigotes whose growth was arrested within the host cells. Cytostasis of amastigotes was also observed during axenic culture of GPI-deficient parasites. In a hamster model of leishmaniasis, GPI-deficient promastigotes produced smaller lesions with 20-fold fewer amastigotes than infections with control parasites. Together, these observations indicate that EtN-GPIs may be essential for amastigote viability, replication, and/or virulence. Implicit in these observations is the notion that drugs targeted against the GPI biosynthetic pathway might be of value in the management of human leishmaniasis.     

Phosphomonoesterases of Leishmania mexicana mexicana and other flagellates

Amastigotes and log-phase promastigotes of Leishmania mexicana mexicana contained distinct acid phosphatase, 3'-nucleotidase and 5'-nucleotidase activities, distinguishable by their response to pH and inhibitors. Both tartrate-sensitive and tartrate-resistant acid phosphatase were present in the two forms, amastigotes possessed less tartrate-resistant acid phosphatase than promastigotes. A tartrate-sensitive acid phosphatase was secreted into the medium in large amounts during the growth in vitro of L. m. mexicana promastigotes. The 5'-nucleotidase activity of both parasite forms was inhibited by ammonium molybdate, sodium tartrate and, to less extent, by sodium fluoride whereas 3'-nucleotidase was inhibited by EDTA. All three activities were shown to be present on the external surface of both amastigotes and promastigotes. The three phosphomonoesterase activities were also detected in extracts of L. m. amazonensis, L. donovani, L. tarentolae, Crithidia fasciculata, Herpetomonas muscarum muscarum, H.m. ingenoplastis and Trichomitus batrachorum whereas 5'-nucleotidase was not detected in Trypanosoma brucei brucei extract and 3'-nucleotidase was absent from extracts of Trichomonas vaginalis and Tritrichomonas foetus.     

Macrophage damage by Leishmania amazonensis cytolysin: evidence of pore formation on cell membrane

We have previously shown that both promastigotes and amastigotes of Leishmania amazonensis contain a lytic protein that damages erythrocytes and nucleated cells, including macrophages (F. S. M. Noronha, F. J. Ramalho-Pinto, and M. F. Horta, Infect. Immun. 64:3975-3982, 1996). Using the patch-clamp technique, we show here that cell damage by parasite extracts is mediated by the formation of nonselective pores on the target membrane. This demonstrates that L. amazonensis cytolysin is a pore-forming protein (PFP), here named leishporin. We show that the diameters of the pores formed by parasite extracts are heterogeneous, varying from approximately 1.6 to >6.1 nm according to cytolysin concentration or time. We also show that pore formation involves the binding of the PFP to the target cell membrane, a temperature-independent event that is necessary but not sufficient to lyse cells. This is followed by a temperature-dependent step that triggers lysis, probably the insertion and the polymerization of protein subunits in the lipid bilayer. We provide evidence that suggests that polymerization of single subunits must occur for pore formation. We show, in addition, that L. amazonensis expresses molecules antigenically homologous to other PFPs.     

Presentation of the Leishmania antigen LACK by infected macrophages is dependent upon the virulence of the phagocytosed parasites

We have previously demonstrated that murine macrophages (Mphi) infected with Leishmania promastigotes, in contrast to Mphi infected with the amastigote stage of these parasites, are able to present the Leishmania antigen LACK (Leishmania homologue of receptors for activated C kinase) to specific, I-Ad-restricted T cell hybrids and to the T cell clone 9.1-2. These T cells react with the LACK (158-173) peptide, which is immunodominant in BALB/c mice. Here, we show that the level of stimulation of the LACK-specific T cell hybridoma OD12 by promastigote-infected Mphi is clearly dependent upon the differentiation state of the internalized parasites. Thus, shortly after infection with log-phase or stationary-phase promastigotes of L. major or of L. amazonensis, Mphi strongly activated OD12. The activity was transient and rapidly lost. However, under the same conditions, activation of OD12 by Mphi infected with metacyclic promastigotes of L. major or of L. amazonensis was barely detectable. At the extreme, Mphi infected with amastigotes were incapable to stimulate OD12. Thus, the presentation of LACK by infected Mphi correlates with the degree of virulence of the phagocytosed parasites, the less virulent being the best for the generation/expression of LACK (158-173)-I-Ad complexes. While the intracellular killing of the parasites appears to be an important condition for the presentation of LACK, it is not the only requisite. The partial or total destruction of intracellular L. amazonensis amastigotes does not allow the presentation of LACK to OD12. A preferential interaction of LACK (158-173) with recycling rather than newly synthesized MHC class II molecules does not explain the transient presentation of LACK by Mphi infected with log-phase or stationary-phase promastigotes because brefeldin A strongly inhibited the presentation of LACK to OD12. Taken together, these results suggest that virulent stages of Leishmania, namely metacyclics and amastigotes, have evolved strategies to avoid or minimize their recognition by CD4+ T lymphocytes.     

Attempt to differentiate Leishmania (Leishmania) amazonensis, L. (L.) chagasi, Leishmania (Viannia) braziliensis and L. (V.) guyanensis using the SSR-PCR technique

The ability to differentiate reference strains of Leishmania (Leishmania) amazonensis, L. (L.) chagasi, Leishmania (Viannia) braziliensis and L. (V.) guyanensis was evaluated using the simple sequence repeat polymerase chain reaction (SSR-PCR) technique. This technique differentiates the Leishmania species, generating distinct DNA amplicon profiles. The SSR-PCR profiles were similar to but more reproducible than those produced by RAPD. SSR-PCR is presented as an alternative to other molecular methods for the differentiation of Leishmania species or strains.     

Identification of the promastigote surface protease in seven species of Leishmania

Twelve different strains of Leishmania, including L. major, L. donovani, L. infantum, L. tropica, L. mexicana, L. amazonensis, L. braziliensis, and L. enriettii were examined for the presence of an ectoenzyme structurally and functionally related to the promastigote surface protease found in L. major LEM 513. All strains examined possess a protease that is labelled by surface iodination of living promastigotes. The electrophoretic migrations of the labelled proteases are similar in all species showing distinct ectoprotease activity. In addition, proteases that cross-react immunologically with the polypeptide moiety of the surface protease of L. major LEM 513 were found in 10 strains. These proteases were in all cases labelled by surface radioiodination. Two of the strains, L. amazonensis and L. braziliensis, do not show a strict correlation between protease activity, surface iodination, and immunological cross-reactivity with the promastigote surface protease of L. major LEM 513, although both strains possess distinct neutral proteases with electrophoretic behavior similar to that of the enzyme of L. major. The amount of proteolytic activity detected at the surface of living cells depends on the strain tested, and correlates qualitatively with the amount of promastigote surface protease detected on zymograms. We conclude that the proteolytic activity found at the surface of Leishmania promastigotes is a common feature of the species infective for humans and that the promastigote surface protease described in this article is structurally and functionally conserved in Old and New World Leishmania.     

Heme binding to Leishmania mexicana amazonensis

Leishmania mexicana amazonensis is a pathogenic parasite whose growth, due to a biosynthetic deficiency, is dependent on a supply of exogenous heme. Utilizing [55Fe]hemin, we have demonstrated that heme binding to non-dividing cultured promastigotes of L. m. amazonensis at 4 degrees C reaches equilibrium within 6 h, is 95% dissociable by 28 h and is elevated approximately 5-fold by decreasing the pH of the binding buffer to 5.4. Metalloporphyrins substituted either at the central metal atom or in the porphyrin ring all displaced [55Fe]hemin binding to varying extents. Scatchard analysis revealed the affinity of the interaction to be 0.03 nM-1 and the number of binding sites to be 400 per promastigote. These findings are remarkably similar to those demonstrated in murine erythroleukemia cells and are characteristic of a receptor-ligand interaction. During logarithmic growth, promastigote heme binding was increased approximately 10-fold compared to stationary phase organisms. The increase was caused by a 4-fold greater number of binding sites per promastigote with no significant change in affinity. These findings demonstrate not only that L. m. amazonensis promastigotes bind heme specifically, but that the binding may be regulated by the growth phase of the parasite.     

Homologues of LMPK, a mitogen-activated protein kinase from Leishmania mexicana, in different Leishmania species

LMPK, a mitogen-activated protein (MAP) kinase homologue of Leishmania mexicana, is essential for the proliferation of the amastigote, the mammalian stage of the protozoan parasite. This has been demonstrated using deletion mutant promastigotes, the insect stage of the parasite: first, in vitro after differentiation to amastigotes, which subsequently lost their potential to proliferate; second, by infection of peritoneal macrophages, which were able to cope with the infection and cleared the parasites; third, by infection of BALB/c mice, which showed no lesion development. The lmpk deletion mutant promastigotes are a potential live vaccine because they infect macrophages, transform to amastigotes and deliver amastigote antigens to raise an immune response without causing the disease. In addition, inhibition of LMPK in a wild-type infection is likely to resolve the disease and as such, is an ideal target for drug development against leishmaniasis. Here we investigated the presence and copy number of lmpk homologues in Leishmania amazonensis, L. major, L. tropica, L. aethiopica, L. donovani, L. infantum, and L. braziliensis and discuss the results with regard to drug development and vaccination using kinase deletion mutants.     

Freeze-fracture and cytochemistry study of the interaction between Leishmania mexicana amazonensis and macrophages

The process of interaction between macrophages and promastigote and amastigote forms of Leishmania mexicana amazonensis was analyzed using freeze fracture and cytochemistry. The promastigotes inside endocytic vacuoles of macrophages presented an altered distribution of intramembranous particles and a wavy aspect of the plasma membrane. However, amastigotes did not show such alterations. The membrane alterations are probably caused by intracellular cell lysis of the promastigotes by the macrophages. An accumulation of intramembranous particles was seen in the plasma membrane of amastigote forms in the area of adhesion to the macrophages. The parasitophorous vacuole membrane had intramembranous particles randomly distributed. The enzyme activity of Mg++-ATPase, 5'-nucleotidase and NAD(P)H-oxidase was cytochemically detected, at the ultrastructural level, in normal mouse peritoneal macrophages and in macrophages infected with Leishmania mexicana amazonensis. Mg++-ATPase and 5'-nucleotidase are uniformly distributed throughout the macrophage's plasma membrane but were not detected in the membrane lining endocytic vacuoles containing ingested parasites (parasitophorous vacuole). NAD(P)H-oxidase activity was seen in those portions of the macrophage's plasma membrane which enter in direct contact with parasites and also in association with the membrane of the parasitophorous vacuole. The amount of reaction product, indicative of NAD(P)H-oxidase activity, was larger in macrophages which interacted with the promastigote than in those which interacted with the amastigote form of L. mexicana amazonensis. Concanavalin A binding sites and anionic sites of the macrophage's surface, labeled before the interaction, are not interiorized together with the parasites, however, are observed in endocytic vacuoles which do not contain parasites.     

Enhanced replication of Leishmania amazonensis amastigotes in gamma interferon-stimulated murine macrophages: implications for the pathogenesis of cutaneous leishmaniasis

During Leishmania major infection in mice, gamma interferon (IFN-gamma) plays an essential role in controlling parasite growth and disease progression. In studies designed to ascertain the role of IFN-gamma in Leishmania amazonensis infection, we were surprised to find that IFN-gamma could promote L. amazonensis amastigote replication in macrophages (Mphis), although it activated Mphis to kill promastigotes. The replication-promoting effect of IFN-gamma on amastigotes was independent of the source and genetic background of Mphis, was apparently not affected by surface opsonization of amastigotes, was not mediated by interleukin-10 or transforming growth factor beta, and was observed at different temperatures. Consistent with the different fates of promastigotes and amastigotes in IFN-gamma-stimulated Mphis, L. amazonensis-specific Th1 transfer helped recipient mice control L. amazonensis infection established by promastigotes but not L. amazonensis infection established by amastigotes. On the other hand, IFN-gamma could stimulate Mphis to limit amastigote replication when it was coupled with lipopolysaccharides but not when it was coupled with tumor necrosis factor alpha. Thus, IFN-gamma may play a bidirectional role at the level of parasite-Mphi interactions; when it is optimally coupled with other factors, it has a protective effect against infection, and in the absence of such synergy it promotes amastigote growth. These results reveal a quite unexpected aspect of the L. amazonensis parasite and have important implications for understanding the pathogenesis of the disease and for developing vaccines and immunotherapies.