Microbicidal molecules

Macrophages produce two species of highly reactive molecule that are destructive to Leishmania: nitric oxide (NO) (Liew et al. 1990) and oxygen radicals (reactive oxygen species, ROS) (Murray, 1982). Production of ROS by macrophages is detrimental to parasite survival in vitro (Buchmuller-Rouiller & Mauel, 1981, 1986) and ROS-deficient mice show an initial, increased susceptibility to L. major infection (Murray & Nathan, 1999). Leishmania therefore inhibit ROS production by their host macrophages (Buchmuller-Rouiller & Mauel, 1987; Olivier, Baimbridge & Reiner, 1992; Olivier, Brownsey & Reiner, 1992). This appears to be due to abnormal Protein Kinase C (PKC) activity (Olivier, Brownsey & Reiner, 1992), resulting from interaction between the macrophage and parasite surface molecules lipophosphoglycan (LPG) and/or gp63 (Sorensen, Hey & Kharazmi, 1994; Descoteaux & Turco, 1999).

NO is produced following the induction of the inducible Nitric Oxide Synthase (iNOS, NOS2) gene, and seems to be the more important for elimination of Leishmania since gene knock-out of iNOS in a Leishmania-resistant strain of mouse causes the mouse to become susceptible to L. major infection (Wei et al. 1995; Murray & Nathan, 1999). Furthermore, macrophages derived from iNOS-deficient mice are unable to eliminate L. major in vitro (Wei et al. 1995). In comparison, mice defective in ROS production eventually control the infection (Murray & Nathan, 1999). Macrophages infected with Leishmania promastigotes, or treated with purified Leishmania surface molecules LPG (Proudfoot et al. 1996) or glycoinositol phospholipid (GIPL) (Proudfoot, O'Donnell & Liew, 1995) are unable to induce iNOS and generate NO in response to subsequent treatment with IFNγ and/or LPS. However, LPG appears to increase NO production when it is administered to macrophages simultaneously with IFNγ (Proudfoot, O'Donnell & Liew, 1995; Proudfoot et al. 1996). This, together with our recent data showing that promastigotes are unable to inhibit NO production by macrophages that have been previously stimulated with IFNγ (Whitcombe and Olivier, unpublished data), is consistent with a model in which exposure of macrophages to IFNγ, presumably produced by neighbouring NK cells and T lymphocytes, is critical for converting them to a leishmanicidal state. Inactivating the response to IFNγ is therefore vital for the survival of Leishmania.

Inhibition of cytokine secretion

Leishmania infection inhibits production of cytokines, primarily those involved in the inflammatory response, e.g. the IL-1 family and TNFα, or activation of T lymphocytes, e.g. IL-12. Secretion of IL-1 in response to LPS is inhibited in macrophages following either infection with L. donovani or L. major (Reiner, 1987), or incubation with LPG (Frankenburg et al. 1990). Interestingly, while the inhibition of IL-1β appears to be due to a transcriptional repressor acting on the gene's promoter (Hatzigeorgiou et al. 1996), IL-1α mRNA, but not protein, is increased following infection with L. major, indicating a post-transcriptional repression (Hawn et al. 2002). Furthermore, increased IL-1α mRNA was not observed in macrophages derived from Myd88-deficient mice, or RAW267.4 macrophage-like cells transfected with dominant-negative Myd88 (Hawn et al. 2002). Since Myd88 is an important intermediate in Toll-Like Receptor (TLR)-dependent signalling, this implies a role for TLRs in manipulation of macrophage activity.

Expression of TNFα in response to LPS is also inhibited in L. donovani-infected macrophages (Descoteaux & Matlashewski, 1989). This may be indirect, resulting from an induction of the immunosuppressive IL-10 following infection (Bhattacharyya et al. 2001a).

In contrast to these in vitro studies, we have shown an elevated presence of IL-1β and TNFα, as well as various chemokines, following infection with L. donovani, and particularly L. major, in the mouse air-pouch model (Matte & Olivier, 2002). It is unclear at this stage whether these are produced by infected macrophages, or by other members of the heterogeneous leukocyte population that is recruited by both strains of Leishmania (Matte & Olivier, 2002), and whether this represents a host defence mechanism, incompletely inhibited by the parasite, and/or benefits the Leishmania, for example by increasing the number of macrophages present, that can subsequently be infected.

IL-12 plays a critical role in the regulation of cellular immune responses. It is essential for T lymphocyte activation and subsequent IFNγ secretion, which in turn results in macrophage activation and production of microbicidal molecules. As such, several studies have documented its suppression by Leishmania. L. major promastigotes (Carrera et al. 1996), L. mexicana amastigotes (Weinheber et al. 1998) and the phosphoglycan portion of LPG (Piedrafita et al. 1999) have all been shown to cause a marked inhibition of macrophage IL-12 production. This inhibition has been also reported in L. major-infected mice (Belkaid et al. 1998). The intracellular signalling mechanisms responsible are not well understood at this time, but it appears to involve elevated intracellular Ca²⁺ following interaction with the macrophage complement receptor CR3 and/or Fcγ receptor (Marth & Kelsall, 1997; Sutterwala et al. 1997), rather than an NF-κB dependent mechanism (Piedrafita et al. 1999).

Major histocompatibility complex expression

Development of an effective immune response against a pathogen such as Leishmania requires antigen presentation by specialized phagocytes to T lymphocytes. Whereas dendritic cells are now recognized to be the more important antigen presenting cell type, macrophages also present antigens through the MHC class II system, resulting in recognition by activated T lymphocytes and a reciprocal augmentation in microbicidal activities. MHC class I-mediated antigen presentation, leading to CD8+T cell activation, is also protective against Leishmania in some situations (Uzonna, Joyce & Scott, 2004), but, unlike MHC class II, is not essential for the complete clearance of Leishmania from the host (Locksley et al. 1993; Huber et al. 1998).

Some studies have shown that L. donovani infection inhibits induction of MHC class I and II by IFNγ (Reiner, Ng & McMaster, 1987; Reiner et al. 1988) and that inhibition of MHC class II mRNA transcription is independent of cyclic AMP (Kwan et al. 1992). In contrast, macrophages infected with L. amazonensis express normal level of MHC class II (Prina et al. 1993; Lang et al. 1994a), but suffer defects in antigen loading (Prina et al. 1993). This controversy may reveal a mechanistic difference between Leishmania species. Defective antigen loading has also been observed in L. major infected macrophages (Fruth, Solioz & Louis, 1993). Furthermore, some workers have suggested that MHC class II molecules may be sequestered in the phagolysosome (Lang et al. 1994b), and even endocytosed and degraded by L. amazonensis amastigotes (De Souza Leao et al. 1995). Similarly, Kima and collaborators showed that amastigote antigens were also sequestered, and therefore unavailable for presentation by MHC class II molecules (Kima et al. 1996). Another study linked presentation of the LACK antigen of L. major and L. amazonensis to the developmental stage, and thus infectiousness, of the parasite: log and stationary phase promastigotes permitted transient presentation, whereas mature (metacyclic) permitted very little and amastigotes, no detectable presentation at all (Courret et al. 1999). Collectively, these studies show that MHC class II-mediated antigen presentation is inhibited during Leishmania infection, but there remains some controversy as to the mechanism(s) involved. Future studies will therefore be of much importance.

Antigen presentation depends upon cellular communication through co-stimulatory molecules such as B7/CD28 and CD40/CD40L. It has been demonstrated that expression of B7-1 on L. donovani-infected macrophages is not induced by LPS stimulation (Kaye et al. 1994) and that this inactivation process is prostaglandin-dependent (Saha et al. 1995). Indeed, inhibition of CD40/CD40L binding is important for preventing iNOS production (Soong et al. 1996) and macrophage microbicidal activities (Kamanaka et al. 1996; Soong et al. 1996), and indeed cure of leishmaniasis depends upon a functional CD40/CD40L interaction (Campbell et al. 1996; Heinzel, Rerko & Hujer, 1998). Recent findings suggest that p38-dependent signalling triggered by CD40 is altered in infected macrophages, and that this may contribute to reduced iNOS expression (Awasthi et al. 2003). Another study suggests that L. chagasi infection inhibits macrophage ICAM-1 expression in response to LPS with and without IFNγ (De Almeida, Cardoso & Barral-Netto, 2003).

Modulation of intracellular calcium signalling and protein kinase C

One of the first second messengers that was reported to be modulated by Leishmania parasites was calcium (Ca²⁺). L. major and L. donovani infections were shown to augment the basal intracellular Ca²⁺ concentration of phagocytes (Eilam, El On & Spira, 1985; Olivier, Baimbridge & Reiner, 1992). This elevation seems to result from augmented Ca²⁺ uptake by infected cells in response to depletion of intracellular Ca²⁺ stores (Mansfield & Olivier, 2002). The parasite surface molecule LPG also seems to ligate Ca²⁺ directly (Eilam, El On & Spira, 1985) and to be involved in the rapid elevation of intracellular Ca²⁺ concentration (Mansfield & Olivier, 2002). L. donovani-infected cells have shown altered Ca²⁺-dependent signalling, since cells no longer respond to the chemotactic peptide f-Met-Leu-Phe and produce reactive oxygen species (ROS) (Olivier, Baimbridge & Reiner, 1992). This may be due to a reduction in host inositol tri-phosphate (IP₃) concentration (Olivier, Baimbridge & Reiner, 1992), perhaps by the action of a parasite acid phosphatase (Das et al. 1986), or a Ca²⁺-dependent activation of a host IP3 phosphatase (Olivier, 1996).

Several kinases depend upon Ca²⁺ availability to be fully active, for instance classical isoforms of PKC. However, PKC activity has been shown to be reduced in L. donovani-infected macrophages, favouring parasite survival (McNeely & Turco, 1987; Descoteaux, Matlashewski & Turco, 1992; Olivier, Brownsey & Reiner, 1992). It appears that promastigote LPG can interact with PKC at its diacyl glycerol (DAG) binding sites, interfering with its interactions with Ca²⁺ and phospholipids, as well as blocking the insertion of PKC into the membrane (reviewed in Descoteaux & Turco, 1999). Of interest, the reduced sensitivity of PKC to DAG was first demonstrated biochemically in cells that have been infected with L. donovani amastigotes (Olivier, Brownsey & Reiner, 1992). Since amastigotes lack LPG, this indicates that an alternative mechanism for inhibiting PKC activity despite elevated Ca²⁺ must operate in amastigote-infected cells. Leishmania has been reported to inhibit MARCKs-Related Protein, a substrate of PKC associated with the cell cytoskeleton and involved in vacuole dispersal (Kane & Mosser, 2000). Another Leishmania surface glycolipid, GIPL, may also inactivate PKC, but the underlying mechanisms are unclear (McNeely et al. 1989). More recently, it has been demonstrated that Leishmania-mediated PKC inhibition applies only to the family members that are dependent upon Ca²⁺ for their full activation (Bhattacharyya et al. 2001b). This selective inhibition may be due to the negative action of ceramides and IL-10 produced by infected phagocytes (Bhattacharyya et al. 2001a,b; Ghosh et al. 2001).

Inhibition of JAK/STAT pathways

Given that some of the important macrophage functions suppressed by Leishmania are IFNγ-inducible (e.g. NO production, MHC class II expression), it is not surprising that Leishmania can inhibit the JAK2/STAT1 signalling cascade, which is known to be the major pathway activated by IFNγ receptor activation. In fact, infected macrophages show defective phosphorylation of JAK1, JAK2 and STAT1 on IFNγ stimulation (Nandan & Reiner, 1995; Blanchette et al. 1999). This inactivation depends on the activation of phosphotyrosine phosphatases (PTPs), in particular the PTP, SHP-1. One study has shown that inactivation of JAK2/STAT1 is caused by the negative regulation of the IFNγ receptor in infected cells (Ray et al. 2000). However, this conflicts with the observations that other IFNγ-dependent events, such as MHC class I-mediated antigen presentation, are not affected by infection (Kima, Ruddle & McMahon-Pratt, 1997).

In a recent study, we investigated the mechanisms by which infection with L. donovani inactivates macrophage STAT1α. We observed that the reduced activity of STAT1α in the nucleus of infected cells is caused by rapid and sustained STAT1α protein degradation. This phenomenon seems to be specific to STAT1α, since STAT3 levels were unchanged. Using both PTP inhibitors and SHP-1 deficient macrophages, we found that this was a PTP independent process. Instead, we found that the phenomenon was reversed by inhibitors of both the proteasome and PKC. Taken together, these results argue for a direct role of the proteasome pathway in the specific proteolysis of STAT1α in macrophages infected with Leishmania, representing a new mechanism whereby pathogens can subvert microbicidal actions.

Inhibition of MAP kinase cascades

Infection of naïve macrophages by L. donovani has been shown not to activate the most important mitogen activated proteins kinase (MAPK) family members: Extracellular-Regulated Kinase (ERK1/2), p38, and Jun N-terminal Kinase (JNK) (Prive & Descoteaux, 2000), consistent with a ‘silent’ model of entry. A number of other studies, however, go further and show that MAPKs are actively repressed, and cannot be activated when infected macrophages are stimulated with a variety of agonists. In one study, ceramide-mediated inactivation of ERK1/2, and a resulting inhibition of the transcription factors AP-1 and NF-κB, was proposed to explain the absence of NO generation by L. donovani-infected cells (Ghosh et al. 2002). Most studies, however, have focused on suppression of MAPK cascades by dephosphorylation. In one case, L. amazonensis amastigotes were shown to rapidly alter phosphorylation of ERK1 in response to LPS, and this was linked to the activity of a parasite Protein Tyrosine Phosphatase (PTP) (Martiny et al. 1999). Another suggested that a host PTP was responsible for the inactivation of ERK1/2, and resulted in the repression of Elk-1 and c-fos transcription factor expression (Nandan, Lo & Reiner, 1999). This aspect is discussed in greater detail below.

Other MAP kinase family members are also inactivated by infection. For example, the p38 MAPK has shown abnormal activity in L. major-infected cells that are stimulated with a CD40 activating antibody, which mimics the macrophage-T lymphocyte interaction (Awasthi et al. 2003). Consistent with previous work, infected macrophages in this study cannot express iNOS: the authors therefore suggest that this may, in part, result from p38 inactivation. This is supported by the observation that treatment of macrophages with anisomycin, an activator of p38, reduces the survival of L. donovani (Junghae & Raynes, 2002). Such conclusions must be treated with caution, however, since the use of any compounds that contribute to macrophage activation might be expected to hinder parasite survival.

The parasite surface molecule LPG has been implicated in the inactivation of MAPKs, since phagocytosis of LPG deficient L. donovani promastigotes caused MAPK activation, without the requirement for subsequent macrophage stimulation (Prive & Descoteaux, 2000). This does not necessarily mean that LPG itself directly causes inhibition but that, in the absence of LPG, the interaction of Leishmania with its host cell, and thus the stimulation of host cell receptors, is modified. We have observed that LPG-deficient L. donovani still induce the PTP SHP-1 (see below), with resulting inactivation of JAK2 and MAPK signalling pathways and downstream transcription factors (Olivier et al., unpublished data).

Tyrosine dephosphorylation by SHP-1

As discussed above, inhibition of MAPK signalling pathways is due, at least in a large part, to the activity of Protein Tyrosine Phosphatases (PTPs) (Martiny et al. 1999; Nandan, Lo & Reiner, 1999). We recently obtained clear evidence that activation of the host PTP SHP-1 is responsible for the dephosphorylation and inactivation of ERK1/2, as SHP-1-deficient macrophages showed normal JAK2 and ERK1/2 activity following infection with L. donovani, and responded to IFNγ by increased NO production (Forget et al., unpublished). In a parallel study, we have demonstrated that a general tyrosine dephosphorylation of high molecular weight proteins occurred rapidly – even within 5 minutes – following the initial interaction of macrophages with L. donovani promastigotes (Blanchette et al. 1999), and amastigotes (Abu-Dayyeh & Olivier, unpublished). This was accompanied by an increase of PTP activity, in particular the activity of SHP-1 (Blanchette et al. 1999). We also observed that SHP-1 directly associated with JAK2 upon L. donovani infection and this interaction may well account for the absence of JAK2 phosphorylation, and thus activation, following IFNγ stimulation (Blanchette et al. 1999). More recently, using a bone marrow-derived macrophage cell line deficient in SHP-1, we have firmly established that this PTP is responsible for the inactivation of JAK2 and ERK1/2 MAPK. The restoration of these signalling pathways in SHP-1-deficient cells also restores the ability of infected cells to generate NO in response to IFNγ stimulation (Forget et al., unpublished). Our studies are supported by others, who have reported that SHP-1 strongly interacts with MAPKs following L. donovani infection (Nandan, Lo & Reiner, 1999). In this study, infected macrophages did not respond to PMA, an artificial PKC activator, and this was reflected by an inhibition of ERK1/2 phosphorylation, Elk-1 activation and c-fos mRNA expression. A striking difference between this latter study and ours is that the inhibition of ERK1/2 was observed 17 hours post-infection, whereas we and others observed SHP-1 activation and JAK2 inhibition within the first hour post-infection (Blanchette et al. 1999; Martiny et al. 1999; Forget et al., unpublished). Differences between cell lines and their status of differentiation may contribute towards this discrepancy. However, the rapid inhibition of host cell functions is vital for parasite survival. Recently, it has been claimed that the elongation factor EF-1α of Leishmania could be responsible for the late induction of SHP-1 observed 16 hours post-infection (Nandan et al. 2002; Nandan & Reiner, 2005). However, the authors did not prove the claim by generating an EF-1α deficient parasite. In our opinion, the rapidity of SHP-1 activation requires a receptor-mediated process, although another mechanism may contribute at later time-points.

The biological significance of these studies is emphasised by our in vivo studies, using mice deficient for SHP-1, or treated with peroxovanadium compounds, which are potent PTP inhibitors. Treatment with peroxovanadium compounds almost completely protects against development of murine models of cutaneous and visceral leishmaniasis, following inoculation with L. major and L. donovani, respectively (Matte et al. 2000). Interestingly, this protection is not observed in iNOS-deficient mice (Matte et al. 2000). This is consistent with the in vitro studies described above which suggested that the activation of SHP-1 by Leishmania is important for the prevention of NO generation. Mice of the viable motheaten strain, which are deficient for SHP-1, also do not develop swelling following infection of footpads with L. major, and have a significantly reduced parasite load (Forget et al. 2001). Histochemical analysis and in situ hybridization of the infected footpads showed that, unlike wild-type animals, iNOS gene expression and the activities of STAT1 and NF-κB, which are know to contribute to iNOS induction, was not inhibited following infection. In fact, there was an activation of these intermediates, contributing to increased iNOS expression and thus NO generation (Forget et al. 2001). This was supported by our finding that aminoguanidine, an inhibitor of iNOS, completely reversed the protection against L. major observed in SHP-1 deficient animals (Forget et al. 2001). We recently obtained similar observations concerning the role of SHP-1 in the control of visceral leishmaniasis conferred by L. donovani infection in vivo. Interestingly, we found by a pathogenomic approach that repression of genes involved in intracellular signalling and the immune response by L. donovani is almost entirely SHP-1 dependent (Olivier et al., unpublished).