INTRODUCTION: THE IMPORTANCE OF IRON
Iron is the fourth most common element in the Earth's crust. It is present in the vast majority of living organisms and is the most abundant transition metal in the human body (Halliwell and Gutteridge, Reference Halliwell and Gutteridge2007). The biological utility of iron stems from its redox chemistry, which allows it to catalyse multiple types of electron transfer reactions. Iron generally exists in 2 oxidation states, Fe (II) and Fe (III), but may also be found as Fe (IV), Fe (V) or Fe (VI). The very ability of iron readily to undergo oxidation/reduction cycles also leads to its inherent toxicity as a catalyst for the production of reactive chemical species, such as the hydroxyl radical (OH•) via the Fenton reaction:
This dangerous property requires that organisms which utilize iron maintain very tight control over its transport, metabolism and storage.
The balance between the need for iron and its toxicity is especially apparent during an infection (Schaible and Kaufmann, Reference Schaible and Kaufmann2004; Radtke and O'Riordan, Reference Radtke and O'Riordan2006). Here, the host must fulfil its own iron requirements, whilst blocking the needs of the pathogen. In addition, the host may utilize iron-mediated radical production as part of its own immune armament, but must minimize harm to itself. Mammals have evolved an elaborate control system to sequester iron away from invading pathogens, whilst pathogens in turn have evolved multiple and varied methods to obtain it (Schaible and Kaufmann, Reference Schaible and Kaufmann2004). The parasitic trypanosomatids of mammals face different challenges in their quest for iron, as they inhabit different niches within the host. In the following sections, we review current knowledge of the ways that Trypanosoma brucei, Trypanosoma cruzi and Leishmania obtain iron from their mammalian host, and describe the multiple roles that iron plays within these parasites. We discuss the effect of infection on host iron metabolism and the role of iron homeostasis in pathogenesis. Finally, we assess the therapeutic possibilities of interfering with parasite iron metabolism.
IRON UPTAKE AND TRANSPORT
Iron uptake and transport in the host
In mammals, most iron is found in erythrocytes bound to haemoglobin. The majority of this is effectively recycled by macrophages during erythrophagocytosis. However, there is inevitably some loss and iron must be replenished from the diet, either as free iron or as haem. Dietary iron is usually in the form of Fe3+ and this must be reduced prior to transport across the duodenal epithelium. In the bloodstream, iron is transported as Fe3+ bound to the carrier glycoprotein transferrin. Mammalian cells express transferrin receptors (TfR1 and TfR2), which are internalised on binding of holotransferrin. The receptor-transferrin complex is trafficked into the endosomal system. As the sorting endosome matures, it becomes acidified by the vacuolar ATPase, and Fe3+ is released from transferrin as the pH drops. The transferrin is then recycled to the cell surface along with its receptor, where it is released to re-enter the plasma. Fe3+ is essentially insoluble at physiological temperature and pH, so transport across the endosomal membrane and into the cytoplasm requires that it be reduced to Fe2+ by a ferric reductase. Fe2+ is then transported by one of a number of divalent cation transporters (e.g. DMT1, also known as NRAMP2/Slc11a2) into the cytoplasm.
When intracellular iron levels are high, cytoplasmic Fe3+ is sequestered in the storage protein ferritin and transferrin uptake is reduced. Ferritin and transferrin receptor levels are inversely regulated by the iron regulatory proteins IRP1 (cytoplasmic aconitase) and IRP2 to ensure that cellular iron overload or starvation is avoided (Wallander et al. Reference Wallander, Leibold and Eisenstein2006). IRP1 acts as a sensor of iron levels through its iron-sulphur cluster, which is required for its aconitase activity. When intracellular iron levels decrease, the FeS cluster disassembles, causing a conformational change and exposing the RNA binding domain (Dupuy et al. Reference Dupuy, Volbeda, Carpentier, Darnault, Moulis and Fontecilla-Camps2006; Walden et al. Reference Walden, Selezneva, Dupuy, Volbeda, Fontecilla-Camps, Theil and Volz2006). This allows IRP1 to increase expression of the transferrin receptor by stabilization of the corresponding mRNA. Simultaneously, it represses translation from ferritin mRNA ensuring that iron taken in by the cell is available for use. IRP2 is regulated by iron-dependent ubiquitination and proteasomal degradation (Salahudeen et al. Reference Salahudeen, Thompson, Ruiz, Ma, Kinch, Li, Grishin and Bruick2009; Vashisht et al. Reference Vashisht, Zumbrennen, Huang, Powers, Durazo, Sun, Bhaskaran, Persson, Uhlen, Sangfelt, Spruck, Leibold and Wohlschlegel2009). In the bloodstream, iron levels are controlled principally by macrophages. Macrophages recycle iron during phagocytosis and breakdown of senescent erythrocytes. The plasma membrane of macrophages (and duodenal enterocytes) contains the protein ferroportin. Ferroportin exports Fe2+ from the macrophage cytoplasm into the plasma. Ceruloplasmin then oxidizes the Fe2+ to Fe3+, thus ensuring that the concentration gradient for Fe2+ is maintained in the plasma direction, and that recycled iron can be taken up by circulating transferrin. Systemic iron metabolism in mammals is controlled by the peptide hormone hepcidin (see below).
Iron transport in trypanosomatids
Trypanosoma brucei
T. brucei differs from other human pathogenic trypanosomatids in that its lifestyle is exclusively extracellular, living as a trypomastigote in the bloodstream and tissue fluids of the mammalian host. As such, its major source of iron is transferrin (Schell et al. Reference Schell, Borowy and Overath1991a). T. brucei expresses a unique transferrin receptor which is encoded by two of the variant surface glycoprotein (VSG) expression-site associated genes (ESAGs) (Fig. 1a). Proteins encoded by ESAG6 and ESAG7 form a heterodimeric receptor that is attached to the membrane by a glycosylphosphatidylinositol (GPI) lipid anchor (Schell et al. Reference Schell, Evers, Preis, Ziegelbauer, Kiefer, Lottspeich, Cornelissen and Overath1991b; Ligtenberg et al. Reference Ligtenberg, Bitter, Kieft, Steverding, Janssen, Calafat and Borst1994; Salmon et al. Reference Salmon, Geuskens, Hanocq, Hanocq-Quertier, Nolan, Ruben and Pays1994; Steverding et al. Reference Steverding, Stierhof, Chaudhri, Ligtenberg, Schell, Beck-Sickinger and Overath1994). These two proteins appear to be adaptations of the VSG, which have evolved to bind host transferrin (Salmon et al. Reference Salmon, Hanocq-Quertier, Paturiaux-Hanocq, Pays, Tebabi, Nolan, Michel and Pays1997). Their subcellular location is restricted to the flagellar pocket, the only part of the trypanosome surface that has endocytic activity. The ESAG6/7 transferrin receptor is expressed exclusively in the mammalian stages of the life cycle. When it is expressed ectopically in the insect stage, its distribution changes to include the whole of the plasma membrane. This implies that another factor is required to maintain flagellar pocket localization (Ligtenberg et al. Reference Ligtenberg, Bitter, Kieft, Steverding, Janssen, Calafat and Borst1994).
(a) Map of the bloodstream-form variant surface glycoprotein expression site in Trypanosoma brucei. The expression-site associated genes (ESAGs) are illustrated along with their known functions. The promoter is indicated by the black flag. The ESAG6 and 7 genes encoding the transferrin receptor are located proximal to the promoter, whilst the VSG gene is always the most distal. (b) Mechanisms for changing transferrin receptor expression in adapted T. brucei. The transferrin receptor encoded by the ESAG6/7 pair from the 221 expression site (ES) (grey boxes 7 and 6) is expressed in medium containing bovine transferrin. Each silent ES contains variant copies (green boxes). The transcribed ES is indicated by the blue arrow. When trypanosomes are switched to canine transferrin, they can adapt in several ways. (1) In situ transcriptional switch. The old expression site is silenced (red X) and a new one activated, resulting in the expression of new ESAG6/7 variants. (2) Post-transcriptional upregulation of the previously expressed ESAG6/7 to increase the amount of surface transferrin receptor by increasing ESAG6/7 mRNA levels. (3) Deletion (indicated by brackets) of the expressed ESAG7 gene and transcription of a previously silent ESAG7 (dotted blue arrow). The new ESAG7 protein forms a heterodimer with the previously expressed ESAG6. (4) Gene conversion of the existing ESAG7 to replace it with a new ESAG7, again resulting in a different heterodimeric receptor. (5) Deletion of the expressed ESAG7 coupled with gene conversion of the expressed ESAG6. In this case both ESAGs are different to the previously expressed pair, but are transcribed from different ESs. These mechanisms are based on those described by van Luenen et al. (Reference van Luenen, Kieft, Mussmann, Engstler, ter Riet and Borst2005). For simplicity only 2 expression sites are illustrated, there may be up to 20 in a single trypanosome.
Each bloodstream-form expression site (ES) contains copies of ESAG6 and 7, which have a short hypervariable region. It has been proposed by P. Borst that this could reflect variability in the affinity of the different receptors for transferrins from different mammalian hosts (Bitter et al. Reference Bitter, Gerrits, Kieft and Borst1998; van Luenen et al. Reference van Luenen, Kieft, Mussmann, Engstler, ter Riet and Borst2005). In light of this hypothesis, it is notable that the closely related parasite Trypanosoma equiperdum, which has a more restricted host range, exhibits much less diversity in its ESAG6 genes than T. brucei (Isobe et al. Reference Isobe, Holmes and Rudenko2003; Witola et al. Reference Witola, Sarataphan, Inoue, Ohashi and Onuma2005). Strong evidence in favour of the host-range hypothesis comes from recent studies showing the extent to which trypanosomes will go to ensure their transferrin supply (van Luenen et al. Reference van Luenen, Kieft, Mussmann, Engstler, ter Riet and Borst2005). When trypanosomes expressing the transferrin receptor from the 221 ES are removed from bovine serum and grown in canine serum, they experience a cessation in growth after 2 days as their iron level is depleted. Within a week, growth resumes at the normal rate. Many of the outgrowing trypanosomes have switched to a new VSG expression site with different copies of the ESAGs; however, a number retain expression from the 221 locus. These cells are referred to as adaptors and have undergone different changes, all of which result in alterations in the expression of ESAG6 and/or 7 (Fig. 1b) In the simplest case, the levels of ESAG6 and 7 proteins are highly upregulated, so even with low affinity transferrin binding, the cell's iron needs are supplied (Fig. 1b, panel 2). In other cases, there is DNA rearrangement at the ESAG6/7 locus of the 221 ES. In some adaptors, ESAG7 is deleted from the ES and transcribed from a ‘silent’ ES elsewhere, the new heterodimeric receptor having a higher affinity for dog serum (Fig. 1b, panel 3). On some occasions, the ESAG7 gene in the transcribed ES is replaced by a different copy from another ES, again creating a novel heterodimeric receptor (Fig. 1b, panel 4). Some cells do both, deleting ESAG7 and replacing the copy of ESAG6 with a different one from another ES (Fig. 1b, panel 5). In all 3 cases, the receptor expressed by the adaptors is different from that expressed by the parental strain. These changes only occur if no calf serum is added to the medium indicating that selection is a consequence of the need to take up enough canine transferrin.
Some investigators have questioned the relevance of host transferrin to ESAG variability (Salmon et al. Reference Salmon, Paturiaux-Hanocq, Poelvoorde, Vanhamme and Pays2005). Others have suggested that the difference in affinity only comes into play once the host has started to make significant antibody responses to the ESAG6/7 heterodimer (Steverding, Reference Steverding2003, Reference Steverding2006). Specific antibodies may compete with transferrin for the receptor and so a high-affinity receptor is required to maintain the intracellular iron levels required for trypanosome replication. A trypanosome has been estimated to take up ~85 000 Fe3+ ions per generation but only requires ~40 000, and it is therefore assumed that only 50% of the receptors need be occupied at any given time (Steverding, Reference Steverding1998). Calculations based on the relative K ds of the various receptor dimers for different transferrins appear to support the hypothesis that switching expression sites is only of benefit once the anti-ESAG6/7 antibody levels rise. However, most studies on transferrin uptake in T. brucei have not taken into account that in chronic trypanosomiasis, especially in cattle, there is profound anaemia (d'Ieteren et al. Reference d'Ieteren, Authie, Wissocq and Murray1998; Naessens, Reference Naessens2006) in which host transferrin levels, particularly holotransferrin, are decreased. In such a situation, antibodies against the ESAG6/7 transferrin receptor may have a far more significant effect on trypanosomal iron uptake since they are competing with decreased levels of transferrin, particularly if the expressed transferrin receptor has a low affinity for the transferrin of that host (Stijlemans et al. Reference Stijlemans, Vankrunkelsven, Brys, Magez and De Baetselier2008). It is possible that this chronic and progressive host anaemia is one reason why bloodstream-form T. brucei have evolved to require relatively little iron (Steverding, Reference Steverding1998). The finding that strong selection for changes in ESAG6/7 expression does occur under differential serum conditions is strong evidence for a selective role of host transferrin (Mussmann et al. Reference Mussmann, Janssen, Calafat, Engstler, Ansorge, Clayton and Borst2003, Reference Mussmann, Engstler, Gerrits, Kieft, Toaldo, Onderwater, Koerten, van Luenen and Borst2004; van Luenen et al. Reference van Luenen, Kieft, Mussmann, Engstler, ter Riet and Borst2005), but does not rule out the contribution that the other ESAGs may also play in active ES selection in a given host.
The trypanosomal transferrin receptor can also be upregulated when parasites are depleted of iron using chelators (Fast et al. Reference Fast, Kremp, Boshart and Steverding1999). This upregulation occurs post-transcriptionally, but is not dependent on the iron regulatory activity of aconitase, in contrast to human transferrin receptor expression (see above), since upregulation still occurs in a cytosolic aconitase null mutant (Fast et al. Reference Fast, Kremp, Boshart and Steverding1999). The rapid upregulation of ESAG6/7 transferrin receptor expression can also be induced by moving trypanosomes expressing the transferrin receptor from the 221 ES from bovine to dog serum. Since the transferrin receptor encoded in the 221 ES has poor binding capacity for dog transferrin, this process mimics iron starvation (Mussmann et al. Reference Mussmann, Engstler, Gerrits, Kieft, Toaldo, Onderwater, Koerten, van Luenen and Borst2004), providing strong evidence that transferrin receptor expression can be profoundly influenced qualitatively (sequence) and quantitatively (expression level) by the host in which the trypanosome finds itself, and by the iron status of that host. This upregulation of ESAG6/7 expression results in mislocalization of some of the protein onto the cell surface, suggesting that the factors retaining the protein in the flagellar pocket are saturable. Cell-surface localized transferrin receptor is non-functional and is likely to consist of ESAG6 homodimers (Schwartz et al. Reference Schwartz, Peck, Tazeh and Bangs2005). An alternative explanation for apparent retention in the flagellar pocket has proposed that protein dimers with 2 GPI anchors (e.g. VSG) are targeted to the cell surface while those with 1 GPI anchor (e.g. ESAG6/7) are targeted to the flagellar pocket and cycled through the endosomal system. They are ultimately turned over by lysosomal degradation or shedding (Schwartz et al. Reference Schwartz, Peck, Tazeh and Bangs2005; Schwartz and Bangs, Reference Schwartz and Bangs2007).
After transferrin is bound to the ESAG6/7 transferrin receptor, it is endocytosed (Fig. 2), a process that requires the small GTPase TbRab5A and clathrin (Morgan et al. Reference Morgan, Allen, Jeffries, Hollinshead and Field2001; Pal et al. Reference Pal, Hall, Jeffries and Field2003). RNA-mediated knockdown suggests that the phosphatidylinositol-3 kinase TbVPS34 also plays a role in transferrin trafficking, possibly downstream of TbRab5A (Hall et al. Reference Hall, Gabernet-Castello, Voak, Goulding, Natesan and Field2006). The endosome acidifies and the iron is released from the transferrin, to be used by the parasite. Unlike the mammalian transferrin receptor, which remains tightly bound to its cargo throughout, ESAG6/7 loses affinity for apotransferrin at low pH. The apotransferrin is then degraded by the cathepsin B-like protein, TbcatB in the lysosome (Maier and Steverding, Reference Maier and Steverding1996; O'Brien et al. Reference O'Brien, Mackey, Fetter, Choe, O'Donoghue, Zhou, Craik, Caffrey and McKerrow2008), while the receptor is recycled back to the flagellar pocket via TbRab11 positive vesicles (Steverding et al. Reference Steverding, Stierhof, Fuchs, Tauber and Overath1995; Jeffries et al. Reference Jeffries, Morgan and Field2001). The degraded transferrin fragments are recycled to the surface and exocytosed, also by a Rab11-dependent mechanism (Pal et al. Reference Pal, Hall, Jeffries and Field2003; Hall et al. Reference Hall, Smith, Langer, Jacobs, Goulding and Field2005). How the iron gets from the endolysosomal system to the cytoplasm has not yet been resolved. By analogy with other systems, this is likely to involve a ferric reductase/divalent metal ion transporter couple (dotted area, Fig. 2). The T. brucei genome contains 2 putative ferric reductases, a cytochrome b 561-type (Tb927.6.3320) and an NADPH-dependent flavoprotein (Tb11.02.1990), and either or both could fulfil this role. There are also multiple putative divalent cation transporters (e.g. Tb11.01.0725 and Tb11.01.0760), although none has yet been implicated in iron transport. The intracellular transport and metabolism of iron in bloodstream-form African trypanosomes is an area that requires further investigation.
The pathway for iron uptake in bloodstream-form Trypanosoma brucei. (1) Transferrin (Tf) binds to the ESAG6/7 heterodimer in the flagellar pocket. (2) The receptor-Tf complex is taken up into a sorting endosome under control of the small GTPase Rab 5A. The exact role of Vps34 remains to be clarified but it is required for trafficking of transferrin. (3) The endosome is acidified by the vacuolar ATPase and acquires hydrolytic enzymes. Iron is released as the pH drops and is enzymatically reduced to Fe2+ by a ferric reductase (Fe-RED). The Fe2+ is transported into the cytoplasm, via a divalent cation transporter (DCT). (4) The Tf is degraded by the protease TbcatB. (5) The Tf fragments are released into the bloodstream in a Rab11-dependent process. (6) The receptor is recycled back to the surface via Rab11 positive recycling endosomes and, (7) released onto the flagellar pocket membrane. The mechanism of iron reduction and transport into the cytoplasm have not yet been characterized and are shown in the dotted box for this reason.
In procyclic-form trypanosomes, rather more is known about the fate of intracellular iron, but less about its uptake. The rate of endocytosis in procyclics is greatly reduced compared to that of bloodstream-forms, even though procyclics require a greater amount of iron (Morgan et al. Reference Morgan, Allen, Jeffries, Hollinshead and Field2001). A major difference between the two stages is the metabolic activation of the mitochondrion in the procyclic stage. The change from bloodstream to procyclic-forms involves the production of an active respiratory chain terminating in a standard cyanide-sensitive cytochrome c oxidase. This is accompanied by the expression of TCA cycle enzymes including fumarase. These biochemical changes involve the production of FeS cluster proteins such as the Rieske protein of cytochrome c reductase (complex III). FeS cluster proteins are enzymes in which iron atoms are co-ordinated with inorganic sulphur, in addition to cysteine thiol groups. Two of the major players in FeS cluster assembly have been characterized in T. brucei. These are the cysteine desulphurase TbIscS2, and the metallochaperone, TbIscU. The first is required to remove sulphur from cysteine for incorporation into FeS clusters, and the second provides a surface on which the FeS clusters are assembled, prior to incorporation into their apoprotein. RNAi-mediated depletion of either protein is detrimental to procyclic trypanosomes and they are required for both mitochondrial and cytosolic FeS cluster protein production (Smid et al. Reference Smid, Horakova, Vilimova, Hrdy, Cammack, Horvath, Lukes and Tachezy2006). As TbIscS2 and TbIscU are restricted to the mitochondrion, this suggests that all trypanosome FeS clusters are assembled in the mitochondrion and exported to the cytosol as required. ‘Knockdown’ of FeS cluster assembly also causes a switch in the metabolism of the mitochondrion toward a more bloodstream-like form, with increased utilization of the alternative oxidase instead of the respiratory chain, and decreased mitochondrial production of ATP and acetate. These changes are also observed when RNAi is used to deplete components of the respiratory chain, including the Rieske protein (Horvath et al. Reference Horvath, Horakova, Dunajcikova, Verner, Pravdova, Slapetova, Cuninkova and Lukes2005). The switch to a more bloodstream-like mitochondrion provides a clear link between the iron-utilization machinery and life-cycle stage metabolic transitions in T. brucei.
TbIscS2 and TbIscU provide the sulphur and the scaffold respectively for FeS cluster assembly. The iron is donated by the highly conserved mitochondrial protein frataxin. T. brucei contains a frataxin homologue (Tb927.3.1000), depletion of which leads to a rapid decrease in the activities of several FeS-dependent enzymes including aconitase, fumarase and succinate dehydrogenase, while the activity of non FeS-dependent proteins remains unaffected (Long et al. Reference Long, Jirku, Mach, Ginger, Sutak, Richardson, Tachezy and Lukes2008b). In contrast to other systems, RNAi-mediated knockdown of frataxin is not accompanied by an increase in free iron, suggesting that trypanosomes may have the facility to store or remove iron before it builds up to deleterious levels. There is, however, an increase in reactive oxygen species in frataxin-depleted cells (Long et al. Reference Long, Jirku, Mach, Ginger, Sutak, Richardson, Tachezy and Lukes2008b). Aspects of frataxin function appear to be well-conserved across a billion years of evolution, as the human frataxin gene can rescue the T. brucei RNAi cell line, and the trypanosomes also process the human protein in the same way as human cells. Interestingly, mutation of the signal-sequence cleavage site to abrogate processing indicates that the human frataxin does not require cleavage to carry out its function in a trypanosome mitochondrion. This contrasts with the situation in human mitochondria, where processing is necessary (Condo et al. Reference Condo, Ventura, Malisan, Rufini, Tomassini and Testi2007; Long et al. Reference Long, Jirku, Ayala and Lukes2008a). T. brucei remains the only trypanosomatid in which FeS cluster biogenesis has been characterized.
Leishmania spp
Iron uptake has been relatively little studied in most species of Leishmania. The exceptions are L. amazonensis and L. chagasi/infantum Footnote 1, although the whole pathway has not been elucidated in either. In the mammalian host, Leishmania reside in an acidified phagolysosomal compartment in macrophages. In this compartment, they can take iron directly from host-endocytosed transferrin, without the need for their own transferrin endocytosis system, since the iron is released in the low pH milieu. All they need is a ferric reductase and a divalent cation transporter (Fig. 3). Two such activities have been identified (see below).
Binding of transferrin and lactoferrin by promastigotes of L. chagasi has been shown, although the receptors responsible have not been characterized at the gene level (Wilson et al. Reference Wilson, Vorhies, Andersen and Britigan1994, Reference Wilson, Lewis, Miller, McCormick and Britigan2002). A 70 kDa protein binds both, in a non-specific manner (Voyiatzaki and Soteriadou, Reference Voyiatzaki and Soteriadou1992; Wilson et al. Reference Wilson, Lewis, Miller, McCormick and Britigan2002) but, as mentioned above, a transferrin receptor is not required for acquisition of iron within the phagolysosome. The ferric reductase activity identified in Leishmania derives electrons from NADPH rather than ascorbate (Wilson et al. Reference Wilson, Lewis, Miller, McCormick and Britigan2002). This suggests that the Cytb 561 ferric reductase (LinJ30_V3.2050) is not involved since this protein typically uses ascorbate as an electron donor. Instead, it is likely to be the flavoprotein reductase encoded by LinJ30_V3.1630, which uses NADPH. In L. amazonensis, Leishmania Iron Transporter 1 (LIT1), a member of the ZIP (Zrt/IRT-like proteins) divalent cation transporter family has been characterized and shown to be an iron transporter (Huynh et al. Reference Huynh, Sacks and Andrews2006). This protein is localized to the plasma membrane reflecting that Leishmania amastigotes can obtain iron directly from the lumen of the phagolysosome, unlike the extracellular T. brucei (Fig. 3). As with other members of the ZIP family, LIT1 has 8 predicted transmembrane domains. This transporter is required for intracellular replication and correct parasitophorous vacuole formation, as demonstrated by LIT1 null mutants (Huynh et al. Reference Huynh, Sacks and Andrews2006). Phagocytosed LIT1 null mutants differentiate into intracellular amastigotes as normal, and vacuole maturation is not inhibited since the vacuolar membranes still acquire the lyososmal marker Lamp1. However, the development of parasites in the vacuole is severely compromised. In L. amazonensis infections, the parasitophorous vacuole expands as the amastigotes replicate, but in the LIT1 mutants the vacuole remains small and there is no replication of amastigotes. Instead, they appear to degenerate within the vacuole. This inability to multiply within the host was also apparent in vivo as there is no lesion development in infected mice, although parasites persist within the skin (Huynh et al. Reference Huynh, Sacks and Andrews2006).
It has been demonstrated that L. amazonensis re-route transferring bearing endosomes to fuse with the parasitophorous vacuole, a modified phagosome, thus ensuring a constant supply of iron to replicating amastigotes (Borges et al. Reference Borges, Vannier-Santos and de Souza1998). The mechanism of interference with the host endocytic pathway has not been elucidated. There is also evidence that Leishmania may act to increase the uptake of holotransferrin by the host cell. Infection with L. donovani was shown to stabilize transferrin receptor mRNA and to promote uptake of radio-labelled iron from holotransferrin. L. donovani depletes the intracellular labile iron pool leading to the activation of iron regulatory proteins, which then decrease ferritin expression and upregulate transferrin receptor 1 (Das et al. Reference Das, Biswas, Solanki and Mukhopadhyay2009).
Leishmania also require a source of porphyrins since they lack the biosynthetic capacity for their production. Recent work has identified haemoglobin as a source of iron for intracellular amastigotes (Carvalho et al. Reference Carvalho, Cruz, Santarem, Castro, Costa and Tomas2009). L. infantum axenic amastigotes were unable to grow on iron-depleted medium, but growth could be restored by addition of haemoglobin. A 46 kDa flagellar pocket protein of L. donovani had previously been identified as the haemoglobin receptor (Krishnamurthy et al. Reference Krishnamurthy, Vikram, Singh, Patel, Agarwal, Mukhopadhyay, Basu and Mukhopadhyay2005). Carvalho et al. (Reference Carvalho, Cruz, Santarem, Castro, Costa and Tomas2009) showed that antibodies against this receptor could block the utilization of haemoglobin. The exploitation of haem as an iron source by amastigotes may explain why the LIT1 ferrous iron transporter null mutants were able to persist in infected mice (Huynh et al. Reference Huynh, Sacks and Andrews2006). In this respect, it is notable that the major replication sites of visceralising Leishmania species are also sites of macrophage erythrocyte recycling. Erythrocyte haemoglobin could therefore provide intracellular amastigotes with a ready source of haem iron. Leishmania, unlike the trypanosomes, possess an orthologue of ferrochelatase (LmjF17.1460), which is functional, since protoporphyrin IX can replace haemin in culture medium for L. amazonensis (Chang and Chang, Reference Chang and Chang1985).
It is clear that Leishmania are capable of obtaining iron from a number of host sources. The presence of the iron transport machinery on the surface of the Leishmania amastigote, together with the demonstration that genetic disruption of iron uptake completely abrogates infection, implicates this machinery as an Achilles' heel of the parasite suitable for targeting with chemotherapy.
Trypanosoma cruzi
Almost nothing is known about iron uptake or iron sources in T. cruzi. Given the wide range of hosts and host cells infected, it is possible that T. cruzi utilizes several different iron sources, such as myoglobin in cardiac muscle cells – a source of both iron and porphyrins. The cytoplasmic habitat of T. cruzi also gives it access to the host's iron storage protein, ferritin. This could be particularly relevant to amastigotes in hepcidin-responsive macrophages (see below). Insect-stage epimastigotes have a requirement for both haem and non-haem iron in vitro (Lalonde and Holbein, Reference Lalonde and Holbein1984).
Transferrin binds to and is taken up by T. cruzi amastigotes in vitro (Lima and Villalta, Reference Lima and Villalta1990). However, the physiological relevance of this interaction is unclear since amastigotes replicate in the cytoplasm of host cells, a niche in which transferrin is conspicuously absent. Consistent with this, very little intracellular staining has been detected in amastigotes using gold-labelled transferrin (Soares and de Souza, Reference Soares and de Souza1991). This does not exclude the possibility that trypomastigotes could pick up transferrin in the bloodstream, or that epimastigotes can acquire iron from transferrin in the bloodmeal. Both life-cycle stages have been shown to bind transferrin, and in epimastigotes it accumulates in the reservosome, a late endosome/lysosome–like organelle (Soares and de Souza, Reference Soares and de Souza1991; Soares et al. Reference Soares, Souto-Padron and De Souza1992). Despite these preliminary studies, the uptake and usage of iron by T. cruzi, particularly the intracellular amastigote, is poorly understood. Given the large number of people infected, or at danger from this parasite, and the dearth of drugs available, a deeper understanding of parasite iron requirements could be fundamental to new chemotherapeutic strategies.
How do trypanosomatids utilize scavenged haem: questions to be answered
One problem that arises from the use of haem by all the pathogenic trypanosomatids is how the haem ring is broken to release iron (Chang and Chang, Reference Chang and Chang1985; Lara et al. Reference Lara, Sant'anna, Lemos, Laranja, Coelho, Reis Salles, Michel, Oliveira, Cunha, Salmon and Paes2007). There is no readily identifiable orthologue of haem oxygenase in the genome sequences of any of the trypanosomatids, although a putative activity has been measured in promastigotes of L. donovani (Srivastava et al. Reference Srivastava, Sharma, Kamboj, Rastogi and Pandey1997). Additionally, trypanosomes also lack ferrochelatase (present in Leishmania), and so, theoretically, cannot insert iron into scavenged porphyrins. This raises the possibility that trypanosomatids accomplish breakage of the porphyrin ring and insertion of iron into protoporphyrin by a different mechanism from other organisms. However, the simplest explanation for the use of haem, given the apparent lack of both haem oxygenase and ferrochelatase, is that trypanosomes simply incorporate scavenged haem directly into apoproteins without going through intermediate steps. This would then require that the iron used by non-haem proteins be obtained independently from mammalian or insect sources. It is implicit from this that one source of iron would not be able to complement the loss of the other.
THE ROLE OF IRON IN TRYPANOSOMATIDS
Iron has multiple roles in most organisms and trypanosomatids are no exception, with many activities common to both parasite and host. These include the reduction of ribonucleotides for DNA synthesis and cytochrome-based oxidative respiration. Ribonucleotide reduction utilizes a classical eukaryotic enzyme (Dormeyer et al. Reference Dormeyer, Schoneck, Dittmar and Krauth-Siegel1997; Hofer et al. Reference Hofer, Schmidt, Graslund and Thelander1997), with the major difference in trypanosomatids being that the reducing power is provided by the kinetoplastid-specific thiol trypanothione in a reaction catalysed by tryparedoxin (Dormeyer et al. Reference Dormeyer, Reckenfelderbaumer, Ludemann and Krauth-Siegel2001). Below, we discuss several examples of iron-dependent processes in trypanosomatids, which are distinct from the mammalian host. These could have potential as targets for therapeutic intervention.
Antioxidant defences
Fe-dependent superoxide dismutases (Fe-SODs)
Unlike their mammalian hosts, which contain Cu/Zn and Mn-dependent SODs, trypanosomatids express 4 different Fe-dependent superoxide dismutases (Dufernez et al. Reference Dufernez, Yernaux, Gerbod, Noel, Chauvenet, Wintjens, Edgcomb, Capron, Opperdoes and Viscogliosi2006; Wilkinson et al. Reference Wilkinson, Prathalingam, Taylor, Ahmed, Horn and Kelly2006). Fe-dependent SODs are restricted to bacteria and some protozoa. These enzymes help to protect the cell from oxidative stress by catalysing the dismutation of superoxide radicals into H2O2 and O2, the H2O2 then being metabolized by peroxidases. Iron removes an electron to oxidize 1 molecule of superoxide (step 1), and is then reoxidized, by reaction with a second superoxide molecule in the presence of protons (step 2).
1. Fe(III)-SOD+O2•−→Fe2+-SOD+O2
2. Fe2+-SOD+O2•−+2H+→Fe(III)-SOD+H2O2
The net result of these two reactions is regeneration of Fe(III)-SOD and the production of 1 molecule each of O2 and H2O2 from 2 superoxide radicals (Halliwell and Gutteridge, Reference Halliwell and Gutteridge2007).
The 4 enzymes present in trypanosomatids are differentially compartmentalized. In T. brucei, SOD A and SOD C are found in the mitochondrion whilst SOD B1 and B2 are glycosomal, with some SOD B1 present in the cytoplasm. Phylogenetic analyses have indicated that the trypanosomatid Fe-SODs belong to 2 distinct clades, with SOD A and SOD C in one group and the SOD Bs in the other. Both sets of SODs appear to have been acquired by lateral gene transfer from prokaryotes (Dufernez et al. Reference Dufernez, Yernaux, Gerbod, Noel, Chauvenet, Wintjens, Edgcomb, Capron, Opperdoes and Viscogliosi2006). RNAi-mediated knockdown of the Fe-SODs in T. brucei demonstrates that SOD B1/B2 expression is essential in bloodstream forms. However, the RNAi construct could not discriminate between the two isoforms due to their extreme nucleotide sequence conservation (Wilkinson et al. Reference Wilkinson, Prathalingam, Taylor, Ahmed, Horn and Kelly2006). The ability to generate SOD B1 and SOD B2 individual null mutants with no specific defects in growth or virulence indicates that the two isoforms can complement each other, but that at least one must be expressed for survival (Prathalingham et al. Reference Prathalingham, Wilkinson, Horn and Kelly2007). However, there was a difference between the two null mutants in susceptibility to trypanocidal agents. SOD B1 mutants were more susceptible to benznidazole and nifurtimox than the wild type, whereas SOD B2 mutants had comparable sensitivity (Prathalingham et al. Reference Prathalingham, Wilkinson, Horn and Kelly2007). This argues that one-electron reduction of the drugs takes place in the cytoplasm where SOD B1 is at a higher level, rather than the glycosome. The two mitochondrial isoforms SOD A and SOD C are dispensible for the bloodstream form; however, RNAi-mediated knockdown of SOD A renders cells more susceptible to the superoxide generator paraquat (Wilkinson et al. Reference Wilkinson, Prathalingam, Taylor, Ahmed, Horn and Kelly2006). SOD A is expressed in the less metabolically active mitochondrion of the bloodstream form, probably to protect the kinetoplast DNA, and by implication, turnover of paraquat occurs in this organelle. In contrast, knockdown of SOD C produces no discernible effects, either on growth or susceptibility to oxidizing agents and the role of SOD C remains to be resolved.
In T. cruzi, SOD A has been implicated in the control of programmed cell death. Inducible overexpression of SOD A protects the parasite from serum-mediated death (Piacenza et al. Reference Piacenza, Irigoin, Alvarez, Peluffo, Taylor, Kelly, Wilkinson and Radi2007), demonstrating a direct role for superoxide radicals in the cell death signalling programme. TcSOD A appears to be concentrated around the kinetoplast suggesting a role in protecting mitochondrial DNA from oxidative damage (Taylor and Kelly, Reference Taylor and Kelly2006). A similar location has been noted for the mitochondrial peroxiredoxin (Wilkinson et al. Reference Wilkinson, Temperton, Mondragon and Kelly2000). It is tempting to speculate that both enzymes, together with others, constitute an antioxidant complex guarding kDNA from oxidative damage.
Increased expression of the T. cruzi SOD B1 isoform results in greater susceptibility to the superoxide generator gentian violet (Temperton et al. Reference Temperton, Wilkinson, Meyer and Kelly1998). This seemingly paradoxical outcome has also been observed in bacteria and mammalian cells, where SOD activity has been upregulated. A possible explanation is that higher SOD activity results in increased H2O2 levels. In situations where the ability to metabolize H2O2 is limited, this can result in the production of OH• radicals by the Fenton reaction. Of more interest, the SOD B1 overexpressers were also more susceptible to the trypanocidal drug benznidazole, but not to the nitrofuran drugs, in contrast to T. brucei (Temperton et al. Reference Temperton, Wilkinson, Meyer and Kelly1998). This suggests that at least part of the trypanocidal mechanism of benznidazole is mediated through increased oxidative stress in the cytoplasm or glycosome. The glycosomal SODs B1 and B2 are developmentally regulated in Leishmania, with SOD B1 being more highly expressed in amastigotes and SOD B2 in promastigotes (Plewes et al. Reference Plewes, Barr and Gedamu2003). A single allele deletion of SOD B1 was found to confer a lower survival rate in macrophages or when exposed to paraquat, suggesting that glycosomal SOD activity is important in Leishmania amastigotes (Plewes et al. Reference Plewes, Barr and Gedamu2003).
It is clear that the Fe-dependent SODs, particularly the glycosomal isoforms, play important roles in pathogenic trypanosomatids and that they or their iron co-factor are suitable targets for drug development. Inhibitors of SOD could then be used in combination with other trypanocidal drugs, such as benznidazole, to potentiate their effects, thus minimising the effective dose, and decreasing toxic side effects.
Ascorbate-dependent peroxidase
Trypanosomatids lack the classical catalase and selenium-dependent glutathione peroxidases which can rapidly metabolize H2O2. Instead, they express a battery of different peroxidases including tryparedoxin-dependent peroxiredoxins, and non-selenium dependent glutathione peroxidases (Flohe et al. Reference Flohe, Hecht and Steinert1999; Irigoin et al. Reference Irigoin, Cibils, Comini, Wilkinson, Flohe and Radi2008; Wilkinson and Kelly, Reference Wilkinson and Kelly2003). In addition, T. cruzi and Leishmania possess an ascorbate-dependent peroxidase (APX), which can metabolize hydrogen peroxide but not organic peroxides (Wilkinson et al. Reference Wilkinson, Obado, Mauricio and Kelly2002a; Adak and Datta, Reference Adak and Datta2005). Ascorbate peroxidases are more usually plant enzymes and different isoforms are specific to different compartments of the plant cell, including both stromal and thylakoid compartments of the chloroplast. APX is a haemoprotein belonging to the catalase-peroxidase superfamily, and is absent from African trypanosomes.
The catalytic mechanism of APX includes 3 steps. The first involves oxidation of the haem iron to the ferryl (Fe(IV)) state by the H2O2 molecule to form compound I (underlined below), coupled with the formation of a cationic radical on the porphyrin ring. Compound I therefore has an oxidation state of +5.
1. Fe3+-porphyrin+H2O2→Fe4+=O porphyrin•++H2O
Compound I then reacts sequentially with 2 molecules of ascorbate (Asc) to produce the ascorbyl free radical (AFR, monodehydroascorbate) and regenerate the resting enzyme (oxidation state +3).
2. Fe4+=O porphyrin•++Asc→Fe4+=O+AFR•
3. Fe4+=O+Asc→Fe3+-porphyrin +AFR•
The net reaction is: H2O2+2 Asc→H2O+2 AFR•
Ascorbate is only required to return the enzyme to its resting state. In the absence of ascorbate, the porphyrin π-cation radical transfers to cysteine or tryptophan residues in the protein chain and the enzyme can become irreversibly oxidized (Hiner et al. Reference Hiner, Martinez, Arnao, Acosta, Turner, Lloyd Raven and Rodriguez-Lopez2001; Kitajima et al. Reference Kitajima, Kurioka, Yoshimoto, Shindo, Kanaori, Tajima and Oda2008). The AFR is reduced back to Asc by the action of a number of small molecule reductants (e.g. trypanothione, (Krauth-Siegel and Ludemann, Reference Krauth-Siegel and Ludemann1996)) or redox active enzymes. It is not clear whether any other reductant can fulfil the role of ascorbate in the above reaction.
Although Leishmania and T. cruzi proteins are clearly derived from the same ancestral gene (62% amino acid identity between L. major and T. cruzi APXs), they now appear to have different biological roles. The Leishmania enzyme is mitochondrial and has a classical mitochondrial targeting sequence, whereas the T. cruzi protein has a much longer N-terminal extension, which seems to be related to plastid transit peptides, and is resident in the endoplasmic reticulum (Wilkinson et al. Reference Wilkinson, Obado, Mauricio and Kelly2002a). Why the proteins should be differentially targeted in 2 related organisms is not obvious. Leishmania APX appears to have a role in protection of mitochondrial membrane lipids from oxidative stress and its expression is induced on exposure to H2O2 (Dolai et al. Reference Dolai, Yadav, Pal and Adak2008). In the ER, H2O2 is a constitutive by-product of the Ero1/protein disulphide isomerase redox cycle and it is likely that the T. cruzi APX plays an analogous role to Leishmania APX, protecting membrane lipids in the ER membrane from peroxidation. T. cruzi also expresses another peroxidase in the ER, glutathione-dependent peroxidase II (GPX II), which cannot metabolize H2O2, but does metabolize phospholipid hydroperoxides (Wilkinson et al. Reference Wilkinson, Taylor, Touitha, Mauricio, Meyer and Kelly2002b). Thus, it is likely that these two proteins act in concert to protect the ER membrane from the consequences of oxidative stress.
Trypanosome alternative oxidase (TAO)
Bloodstream-form African trypanosomes are dependent on glycolysis since the normal mitochondrial respiratory chain is absent. Instead, oxidation of ubiquinol is carried out by the so-called alternative oxidase (TAO), resulting in the net transfer of electrons from glycerol-3-phosphate to molecular oxygen to produce water. Unlike the conventional respiratory chain, the transfer of electrons to TAO does not result in the generation of ATP and it does not create a proton gradient (Clarkson et al. Reference Clarkson, Bienen, Pollakis and Grady1989). TAO is not present in the mammalian host (or in T. cruzi and Leishmania, although there is a second TAO-like sequence in all trypanosomatids examined (Chaudhuri et al. Reference Chaudhuri, Ott and Hill2006)). In contrast to the other respiratory oxidases, TAO is composed of a single protein of 37·5 kDa and has 2 iron-binding sites constituted by conserved glutamate and histidine residues on helices 1,3,4 and 5 (Chaudhuri et al. Reference Chaudhuri, Ott and Hill2006). The current model for the alternative oxidase active site suggests that TAO is a member of the diiron carboxylate protein family (Andersson and Nordlund, Reference Andersson and Nordlund1999). These proteins are characterized by the motif E-Xn-E-X-X-H-Xn-E-Xn-E-X-X-H, where E is glutamate and H is histidine, with X being any amino acid; n signifies a number of residues. The 4 carboxylates of the glutamates and imidazole nitrogens of the histidine co-ordinate 2 iron atoms in the catalytic centre, hence the name diiron carboxylate protein (Berthold and Stenmark, Reference Berthold and Stenmark2003). Although the T. brucei TAO, like the other alternative oxidases, has 2 strongly hydrophobic segments, it is thought to be positioned along the matrix face of the inner mitochondrial membrane rather than spanning the membrane (Andersson and Nordlund, Reference Andersson and Nordlund1999; Chaudhuri et al. Reference Chaudhuri, Ott and Hill2006). Site-directed mutagenesis of the putative iron-binding ligands H165, E214, E266 and H269 in TAO has demonstrated their importance to activity. Any one of these mutations abolishes the ability of the TAO gene to complement an E. coli haem-deficient mutant (Ajayi et al. Reference Ajayi, Chaudhuri and Hill2002). Iron-dependence of TAO has also been established by iron chelation with o-phenanthroline, resulting in strong inhibition which was reversed by addition of iron, but not other metals such as copper (Ajayi et al. Reference Ajayi, Chaudhuri and Hill2002).
In T. brucei, TAO is developmentally regulated with the mRNA level dropping rapidly on differentiation from bloodstream to procyclic-forms (Chaudhuri et al. Reference Chaudhuri, Sharan and Hill2002). This differential mRNA accumulation appears to be due to a change in the half-life of the mRNA. This is mediated by a labile protein factor since the transcript is stabilized by cycloheximide treatment. During the transition to metacyclic trypomastigotes, which are pre-adapted for infection of the mammalian host, TAO activity is again upregulated and the standard cyanide-sensitive respiratory chain downregulated (Bienen et al. Reference Bienen, Webster and Fish1991). Although the level of TAO expressed in procyclics is lower than in bloodstream-forms, it also plays an important role in this stage, as inhibition of the cyanide sensitive respiratory chain is not lethal unless the TAO is inhibited simultaneously (Coustou et al. Reference Coustou, Besteiro, Biran, Diolez, Bouchaud, Voisin, Michels, Canioni, Baltz and Bringaud2003).
Due to the reliance of bloodstream-form African trypanosomes on glycolysis, coupled to the glycerol-3-phosphate oxidation required to regenerate NAD+, TAO has been proposed to be a candidate drug target (Yabu et al. Reference Yabu, Yoshida, Suzuki, Nihei, Kawai, Minagawa, Hosokawa, Nagai, Kita and Ohta2003; Chaudhuri et al. Reference Chaudhuri, Ott and Hill2006). However, computer modelling of metabolic flux and RNAi-mediated depletion of TAO suggests, in vitro, that the TAO protein would have to be inhibited by more than 95% for there to be a significant effect on trypanosome growth (Helfert et al. Reference Helfert, Estevez, Bakker, Michels and Clayton2001). In T. brucei, TAO inhibition is only lethal when coupled with administration of glycerol to block glycerol-phosphate dehydrogenase or with repeated administration of the inhibitor (Yabu et al. Reference Yabu, Minagawa, Kita, Nagai, Honma, Sakajo, Koide, Ohta and Yoshimoto1998, Reference Yabu, Yoshida, Suzuki, Nihei, Kawai, Minagawa, Hosokawa, Nagai, Kita and Ohta2003). Nevertheless, the demonstration that inhibitor treatment can cure mice suggests that the TAO may be a viable drug target. In T. vivax-infected mice, a single dose of ascofuranone (50 mg kg−1) has been shown to elicit cure without glycerol (Yabu et al. Reference Yabu, Suzuki, Nihei, Minagawa, Hosokawa, Nagai, Kita and Ohta2006).
J-base biosynthesis
A unique feature of trypanosomatids is the presence of a hypermodified base within their DNA. This base, β-D-hydroxymethyldeoxyuridine, more usually known as J, is a derivative of thymidine and is synthesized within the DNA strand in a two-step modification process. The first step in J-base synthesis involves the hydroxylation of thymidine residues in DNA to provide a hydroxyl group to which the glucose moiety is attached (Fig. 4). The hydroxylase which carries this out, is a member of the Fe2+- and 2-oxoglutarate-dependent dioxygenases, although the similarity to other members of this family is not immediately obvious (Yu et al. Reference Yu, Genest, ter Riet, Sweeney, DiPaolo, Kieft, Christodoulou, Perrakis, Simmons, Hausinger, van Luenen, Rigden, Sabatini and Borst2007). These enzymes use molecular oxygen to donate the oxygen atom required for the hydroxyl group. During this process, the iron is probably oxidized to an oxyferryl (Fe(IV)=O) state as in the APX reaction, except that in this case, the oxygen atom is derived from molecular oxygen rather than H2O2 (Schofield and Zhang, Reference Schofield and Zhang1999).
Iron uptake in Leishmania amastigotes. Transferrin is taken up by the infected mammalian cell via its transferrin receptor (TfR). The Tf:TfR complex is endocytosed and the endosome fuses with the phagosome. Fe3+ is liberated from the transferrin due to the low pH. A parasite cell surface NADPH-dependent ferric reductase reduces Fe3+ to Fe2+. The Fe2+ is then transported into the parasite by the divalent ion transporter LIT1. The amastigote may also acquire iron from the cytoplasmic labile iron pool by an undefined mechanism. Simultaneously, the infected cell tries to deplete iron from the phagosome through the concerted action of the Dcytb ferric reductase and the NRAMP1 divalent cation transporter. K: kinetoplast.
Hydroxylation of thymidine residues in DNA. Thymidine hydroxylases JBP1 and JBP2 use oxidative decarboxylation of 2-oxoglutarate in the presence of molecular oxygen to add a hydroxyl group to the methyl carbon of the thymidine base. The reaction is catalysed by Fe2+. R1 and R2 represent the 5′ and 3′ends of this strand of the DNA duplex.
There are 2 thymidine hydroxylases in trypanosomatids, referred to as J-binding protein 1 (JBP1) and J-binding protein 2 (JBP2). Both contain homologous thymidine hydroxylase domains and mutation of the putative iron-binding residues in Leishmania JBP1 (H189, D191 and H239) renders the protein unable to rescue J-base biosynthesis in JBP1 null mutant T. brucei, as does mutation of the conserved arginine residue required for 2-oxoglutarate binding (R255) (Yu et al. Reference Yu, Genest, ter Riet, Sweeney, DiPaolo, Kieft, Christodoulou, Perrakis, Simmons, Hausinger, van Luenen, Rigden, Sabatini and Borst2007). These mutations do not affect the DNA binding of the enzyme and so the inference is that they affect the catalytic capability, as would be expected if they were required for iron incorporation. Similar mutagenesis studies on iron and 2-oxoglutarate binding residues, also confirm that JBP2 is a member of this family (Cliffe et al. Reference Cliffe, Kieft, Southern, Birkeland, Marshall, Sweeney and Sabatini2009; Vainio et al. Reference Vainio, Genest, ter Riet, van Luenen and Borst2009). In neither case has the hydroxylase activity been reconstituted in vitro. Whilst J-base appears to be dispensible for T. brucei (and is absent in the procyclic stage), JBP1 is essential in Leishmania (Genest et al. Reference Genest, ter Riet, Dumas, Papadopoulou, van Luenen and Borst2005). The situation in T. cruzi is unclear.
Iron and anti-kinetoplastid drugs
There have been few studies directly addressing the role of iron in drug-mediated killing of kinetoplastid parasites. Indeed, for most clinically relevant anti-kinetoplastid drugs, the exact mechanism of action remains unknown and many of their effects may be pleiotropic. However, a recent study has implicated iron in drug activity (particularly metalloids) in Leishmania (Mehta and Shaha, Reference Mehta and Shaha2006). Using deferoxamine (DFO), Mehta and Shaha showed that iron depletion inhibits changes in the mitochondrial membrane potential and ATPase activity produced by antimonial or arsenical drugs. This inhibition can be reversed if the DFO is saturated with iron prior to incubation with the parasites, showing that the effect is dependent on iron chelation. In addition, it has been demonstrated that DFO can reduce the level of cell death induced by treatment with SbIII or AsIII, whereas addition of iron causes a slight exacerbation of cell death. The changes in mitochondrial membrane potential and activity prior to drug-induced death mirrors the mechanism shown in T. cruzi for complement-induced programmed cell death of epimastigotes, which is mediated by mitochondrial O2•−(Piacenza et al. Reference Piacenza, Irigoin, Alvarez, Peluffo, Taylor, Kelly, Wilkinson and Radi2007). Thus, the drug-induced death of Leishmania may involve activation of their cell death programme at the mitochondrial level. Exacerbation of this process by iron is likely to reflect increased oxidative stress resulting from OH• generated by the superoxide-driven Fenton reaction.
Further studies of the role of iron in the mechanism of drug action/resistance in trypanosomatids are clearly warranted. This is particularly the case with drugs that may trigger oxidative stress, such as Ornidyl (difluoromethylornithine). Ornidyl blocks polyamine biosynthesis and therefore prevents the production of trypanothione (N1,N8-bisglutathionylspermidine) (Fairlamb et al. Reference Fairlamb, Henderson, Bacchi and Cerami1987). Almost all the antioxidant defences in trypanosomatids derive their reducing power ultimately from trypanothione. Therefore, prolonged inhibition of spermidine synthesis will result in increased oxidative stress by depletion of trypanothione, hence the level of iron available to the Fenton reaction may be key to the trypanocidal activity. As iron has also been implicated in the action of benznidazole (via SOD B1), it would seem important to dissect this interaction for future drug design (Temperton et al. Reference Temperton, Wilkinson, Meyer and Kelly1998; Prathalingham et al. Reference Prathalingham, Wilkinson, Horn and Kelly2007; Francisco et al. Reference Francisco, de Abreu Vieira, Arantes, Pedrosa, Martins, Silva, Veloso, de Lana, Bahia, Tafuri and Carneiro2008).
IRON IN PATHOGENESIS AND IMMUNITY
The peptide hormone hepcidin is released from the liver during infection. One of its roles is to prevent iron release from macrophages into the bloodstream. This is achieved by destabilization of the ferroportin molecule which is internalized and degraded, leading to a block in iron export and an increase in ferritin-bound iron in macrophages (Nemeth et al. Reference Nemeth, Tuttle, Powelson, Vaughn, Donovan, Ward, Ganz and Kaplan2004, Reference Nemeth, Preza, Jung, Kaplan, Waring and Ganz2006). In addition, hepcidin promotes degradation of ferroportin on duodenal enterocytes, thereby reducing dietary iron intake. It has also been demonstrated that macrophages can produce hepcidin themselves under the control of Toll-like receptor 4 signalling, creating an autocrine feedback loop resulting in iron sequestration (Peyssonnaux et al. Reference Peyssonnaux, Zinkernagel, Datta, Lauth, Johnson and Nizet2006). In this way, plasma iron levels are decreased, and bloodstream pathogens prevented from accessing iron. An unfortunate effect of this innate immune strategy is to create a situation of functional anaemia in the host. This condition, known as ‘anaemia of chronic infection’ illustrates the capacity of iron to be both beneficial and dangerous, often simultaneously.
Trypanosoma brucei
For a bloodstream parasite such as T. brucei, this anaemia can create problems as it is dependent on its host for iron. In animal models, anaemia during T. brucei infection is type I cytokine driven and appears to be specifically targeted at removal of iron from the plasma during the chronic phase, thus reducing the pool of iron available to the circulating trypomastigotes (Stijlemans et al. Reference Stijlemans, Vankrunkelsven, Brys, Magez and De Baetselier2008). Using C57Bl/6 mice, it is apparent that anaemia in experimental trypanosomiasis occurs in 2 distinct waves, corresponding to the acute and chronic phases of the infection. During the acute phase, there is a rapid drop in the number of erythrocytes, which then recovers, although not to uninfected levels. This recovery is followed by a continuous decrease as the chronic phase takes over. Transferrin mRNA levels increase rapidly during the acute phase, possibly in response to the removal and degradation of transferrin by the multiplying trypanosomes, but then drop back and fall below the baseline level during the chronic phase. Ferroportin mRNA levels also decrease during the chronic phase. At the same time, ferritin mRNA levels increase and remain high, as does the mRNA for divalent metal transporter 1. These findings show that the infected host responds by increasing the sequestration of iron within the macrophage cytoplasm and decreasing export into the plasma. This sequestration leads to decreased erythropoiesis as iron is unavailable for incorporation into haemoglobin. It may be this cytokine driven anaemia during chronic infection that promotes the selection of different transferrin receptors on the trypanosome. A higher affinity receptor may be required when serum holotransferrin levels drop, particularly as the trypanosome receptor cannot distinguish holo from apotransferrin.
Leishmania – the role of NRAMP1
For many years it has been established that ‘natural resistance associated macrophage protein 1’ (NRAMP1, now known as Slc11a1Footnote 2) is involved in resistance to several intracellular pathogens, including Salmonella typhimurium, Mycobacterium bovis and Leishmania (Vidal et al. Reference Vidal, Tremblay, Govoni, Gauthier, Sebastiani, Malo, Skamene, Olivier, Jothy and Gros1995). Mice strains carrying a single mutation in transmembrane helix 4 (G169D) are more susceptible to these pathogens than are strains carrying wild type NRAMP1 alleles. Resistance can be restored by reintroducing a wild type copy of the gene, whilst NRAMP1 null mutant mice are susceptible to these pathogens (Govoni et al. Reference Govoni, Vidal, Gauthier, Skamene, Malo and Gros1996). NRAMP1 is a late endosomal/lysosomal protein expressed in macrophages and granulocytes. It functions as a divalent cation/proton symporter which transports metal ions down a proton gradient across the phagosomal membrane (Biggs et al. Reference Biggs, Baker, Botham, Dhital, Barton and Perry2001) and is related to the major divalent metal ion transporter DMT1 (NRAMP2/Scl11a2). The net effect of NRAMP1 activity in acidified phagosomes is to deplete iron (and some other transition metals) from the lumen of the phagosome, thus denying pathogens that reside within this compartment access to the metal. Iron chelators can mimic the effect of NRAMP1 in null mutant macrophages infected with Salmonella. This suggests that NRAMP1 mediates its antimicrobial effect through its iron transport properties, although the mechanism by which this controls such diverse pathogens has yet to be established (Jabado et al. Reference Jabado, Cuellar-Mata, Grinstein and Gros2003). In Leishmania, the LIT1 iron transporter gene is upregulated more rapidly when parasites infect wild type cells than NRAMP1 null cells, suggesting that NRAMP1+ phagosomes are depleted in iron (Huynh et al. Reference Huynh, Sacks and Andrews2006). In this context, it is interesting that cells infected with L. amazonensis mutants defective in the LIT1 iron transporter do not develop the expected parasitophorous vacuole. Since the parasites cannot transport iron from the phagosome, but still express a ferric reductase, it is probable that they enhance the natural activity of NRAMP1 by providing reduced iron at a greater rate than in cells infected with wild type Leishmania (Huynh et al. Reference Huynh, Sacks and Andrews2006). The mechanisms identified in mice may also be present in humans (Bucheton et al. Reference Bucheton, Abel, Kheir, Mirgani, El-Safi, Chevillard and Dessein2003; Mohamed et al. Reference Mohamed, Ibrahim, Miller, White, Cordell, Howson, Peacock, Khalil, El Hassan and Blackwell2004; Blackwell et al. Reference Blackwell, Fakiola, Ibrahim, Jamieson, Jeronimo, Miller, Mishra, Mohamed, Peacock, Raju, Sundar and Wilson2009). Genetic polymorphism studies in human populations have shown a linkage between mutations in the NRAMP1 promoter region and susceptibility to visceral leishmaniasis. In the new era of genomics and deep sequencing, there is scope for much wider studies looking at the role of iron metabolism in susceptibility to VL.
Trypanosoma cruzi
Acute T. cruzi infection in mouse models involves anaemia which can be reversed or blocked by administration of the trypanocidal drug nifurtimox (Marcondes et al. Reference Marcondes, Borelli, Yoshida and Russo2000). In addition, anaemia is an indicator of the reactivation of Chagas disease in heart-transplant recipients (Theodoropoulos et al. Reference Theodoropoulos, Silva and Bestetti2009). In the mouse model, the mechanism for the anaemia was postulated to be a decrease in the lifespan of erythrocytes. However, recent studies have shown that resistance to T. cruzi infection involves a key inducible regulator of haematopoeisis, LRG-47, a member of the p47-GTPase family of interferon-induced proteins. These GTPases are thought to play a major role in defence against intracellular pathogens, particularly phagosomal pathogens such as Salmonella and Mycobacteria (MacMicking, Reference MacMicking2005). They are also involved in the regulation of autophagy as an innate defence mechanism against intracellular pathogens (Gutierrez et al. Reference Gutierrez, Master, Singh, Taylor, Colombo and Deretic2004).
LRG47 null mutant mice develop severe anaemia when infected with T. cruzi, coupled with a general failure of haematopoiesis. These mice are susceptible to trypanosome infection, dying within 19 days, whereas wild-type mice survive for more than 30 days. In wild-type mice, expression of LRG-47 is induced on infection, in response to IFN-γ. The infected null mutants develop a profound anaemia by day 15, accompanied by alterations in splenic architecture, thrombocytopenia, lymphopenia and a dramatic depletion of bone marrow cellularity (Santiago et al. Reference Santiago, Feng, Bafica, Roffe, Arantes, Cheever, Taylor, Vieira, Aliberti, Gazzinelli and Sher2005). As well as these haematopoietic defects, macrophages derived from the null mutants also have a decreased ability to kill intracellular amastigotes (Santiago et al. Reference Santiago, Feng, Bafica, Roffe, Arantes, Cheever, Taylor, Vieira, Aliberti, Gazzinelli and Sher2005). The profound anaemia, and atrophy of spleen and bone marrow seen during infection of these null mice, suggest that LRG-47 plays a critical role in the control of haematopoiesis in infected animals and LRG-47 has recently been shown to be required for the correct response of haematopoietic stem cells to chemical or pathogenic insult (Feng et al. Reference Feng, Weksberg, Taylor, Sher and Goodell2008). This suggests that the anaemia seen in some models of T. cruzi infection may be due to interference with the induction of LRG-47, leading to subsequent profound haematopietic defects.
The role of the macrophage iron-withholding response in either acute or chronic T. cruzi infection has not been studied. Given that T. cruzi replicates in the cytoplasm of macrophages it could be expected that such a response would benefit the parasite by allowing access to iron, unlike the situation with T. brucei and Leishmania.
IRON METABOLISM AS A DRUG TARGET
Iron is vital for all trypanosomatid parasites and plays a significant role in pathogenesis and immune control of these organisms. Depletion of this essential nutrient rapidly decreases the rate of DNA synthesis, increases oxidative stress levels via loss of SOD and APX activity, blocks J-base synthesis and stops electron transfer to the alternative oxidase, leading inexorably to death of the parasite. Iron chelation has been tested against all 3 groups of human pathogenic trypanosomatids. While iron chelation often has significant effects on parasites in vitro, during infection it is difficult to separate direct effects on the parasite from indirect effects mediated through the immune response. This may be particularly true for the intracellular trypanosomatids, Leishmania and T. cruzi.
In T. brucei, which one would expect to be especially sensitive as a bloodstream parasite, the iron chelator deferoxamine (DFO) has been shown, in vitro, to decrease growth rate, DNA synthesis and oxygen consumption, and has an IC50 of approximately 3·3 μm (Breidbach et al. Reference Breidbach, Scory, Krauth-Siegel and Steverding2002; Merschjohann and Steverding, Reference Merschjohann and Steverding2006). These chelators act by iron sequestration, preventing the incorporation of iron into newly synthesized enzymes, rather than stripping iron from already active proteins. Decreases in growth rate and oxygen consumption are almost entirely accounted for by inhibition of the alternative oxidase, rather than by a cumulative effect on all iron-dependent proteins. Fe-SOD activity is not affected by iron chelation, possibly due to the protein having a longer half-life. Surprisingly, there are no published studies of the effects of iron chelation on T. brucei infection in animal models.
DFO has been tested in animal models of L. major and T. cruzi infection. Using a cutaneous infection model (L. major in BALB/c mice), intraperitoneal injections of DFO have only a modest effect on lesion development. However, this may simply reflect the short half-life of DFO in plasma (5–10 min), with insufficient concentrations being achieved in the footpad. Surprisingly, when mice are given iron supplementation, in the form of intraperitoneal injections of iron-dextran, lesion development is significantly retarded. Mice injected with 8 mg iron dextran show no lesions up to 18 weeks after challenge (Bisti et al. Reference Bisti, Konidou, Papageorgiou, Milon, Boelaert and Soteriadou2000). The effect of iron supplementation appears to be due to its effect on the immune response of the host, rather than to any direct effect on the parasite. (Bisti et al. Reference Bisti, Konidou, Papageorgiou, Milon, Boelaert and Soteriadou2000; Bisti and Soteriadou, Reference Bisti and Soteriadou2006). In this model, iron overloading leads to an increased and sustained infiltration of neutrophils and a strong oxidative burst during the initial phase of infection. In the later stages, there is activation of NF-κB and an increased number of IFN-γ positive CD4+ T-cells are recruited to the draining lymph node. Thus, in this system iron probably mediates its effect via reactive oxygen species signalling through NF-κB, leading to a sustained TH1 response against the parasites (Bisti et al. Reference Bisti, Konidou, Boelaert, Lebastard and Soteriadou2006; Bisti and Soteriadou, Reference Bisti and Soteriadou2006).
Experiments on iron depletion or overload in T. cruzi-infected mice have been carried out using DFO. In the case of the former the treated mice were also maintained on an iron-deficient diet to ensure their iron depleted status (Lalonde and Holbein, Reference Lalonde and Holbein1984). In these experiments, it was clear that T. cruzi infection is more severe and produces higher mortality in iron-overloaded mice than in iron-deficient mice. In a particularly susceptible mouse strain, C3H, lethality is reduced from 100% to 45% by DFO treatment. The time to death is also extended from 36·5 to 43·7 days for those that succumbed. A second study in which Swiss male mice were injected with DFO daily, also showed reduced parasite growth and mortality during the acute phase (Arantes et al. Reference Arantes, Pedrosa, Martins, Veloso, de Lana, Bahia, Tafuri and Carneiro2007). When the trypanocidal drug benznidazole is coupled with DFO, in the Swiss male mouse/T. cruzi Y strain model, the result is more effective than benznidazole alone, but only when the DFO treatment is started prior to infection. Since these animals are not maintained on an iron-deficient diet, pre-treatment was probably necessary to deplete iron levels sufficiently to make a difference to the parasite (Francisco et al. Reference Francisco, de Abreu Vieira, Arantes, Pedrosa, Martins, Silva, Veloso, de Lana, Bahia, Tafuri and Carneiro2008). Nevertheless, as suggested by the interaction between SOD overexpression and benznidazole susceptibility (Temperton et al. Reference Temperton, Wilkinson, Meyer and Kelly1998), modification of iron metabolism may potentiate the effects of other drugs against these parasites.
CONCLUDING REMARKS
Iron is essential for trypanosomatids, and the mechanism of acquisition and use are potential targets for new therapeutic approaches. These can be aimed at stimulating the host to kill the parasite or directly at the parasite itself. The last 30 years have seen a rapid increase in our knowledge of the biology of these parasites, but there are still many gaps in our understanding of the role of iron in the life cycle and pathogenesis of these organisms. This is especially the case with the American trypanosome T. cruzi, particularly the intracellular amastigote. We hope this review will go some way to stimulating further research into iron and trypanosomatids.
ACKNOWLEDGEMENTS
We thank Shane R. Wilkinson and Belinda S. Hall (Queen Mary, University of London) for critical reading of the manuscript, and Anne Koerber for help with the figures. We acknowledge the financial support of The Wellcome Trust (Grant No. 084175/Z/07/Z).
