Malaria is a serious health problem in developing countries. With the complete sequencing of the genomes of the parasite and of the mosquito vector, malaria research has entered the post-genome era. In this report we summarize the results and new research avenues presented at a recent meeting held with the aim of developing interdisciplinary approaches to combat this disease.
Malaria is one of the most severe infectious diseases worldwide. It kills 2–3 million young children each year. An effective vaccine is not yet in sight, and resistant parasite strains limit the use of affordable drugs in many endemic areas. The sequencing of the genomes of Plasmodium falciparum and of the mosquito vector Anopheles gambiae provides the basis for novel molecular approaches to the development of new malaria control strategies. To facilitate an interdisciplinary and concerted effort, a first conference on ‘Molecular Approaches to Malaria’ was held in Lorne, Australia, in February 2000, attracting more than 250 scientists from all over the world. Following on from this initial success, a second conference (MAM2004) was held in February 2004 at the same venue attracting 358 scientists, clinicians and epidemiologists, including many from malaria-afflicted countries in Africa, South America and Oceania.
A declared goal of the meeting was to bridge gaps between different disciplines in order to facilitate the development of new tools to combat malaria. In a presentation by David Roos (University of Pennsylvania, Philadelphia, USA) who spoke on behalf of the Plasmodium Genome Database Collaborative, it became clear that malaria research has benefited and will continue to benefit from publicly accessible genome, transcriptome and proteome databases which form the basis for modern high-throughput technologies such as microarray analyses. In the following, we summarize the major topics covered at the meeting, ranging from basic research on the molecular and cellular biology of malaria in animal models and in vitro systems through to applied aspects, including advances in the understanding of the molecular mechanisms underlying pathogenesis and the identification and characterization of drug targets.
Host cell invasion
In its human host, P. falciparum is an obligatorily intracellular parasite that first invades liver cells and then goes on to invade differentiated erythrocytes. Invasion is an active process that starts with the attachment of the extracellular merozoite to the surface of the erythrocyte, followed by an active reorientation that brings the apical end of the parasite cell into juxtaposition with the host cell membrane. The parasite then induces an invagination of the erythrocyte membrane and penetrates the erythrocyte. Inside the erythrocyte, the Plasmodium parasite remains within a parasitophorous vacuole, the membrane of which is largely devoid of host cell membrane proteins but which contains lipids derived from the erythrocyte plasma membrane. The biogenesis and the biochemical composition of the vacuole differ markedly from those of the phagocytic compartments in nucleated cells that are inhabited by other pathogens. This distinction has important conceptual implications for our understanding of the vacuole as a parasite–host interface. The identification and functional analyses of the molecules that are involved in each step of the invasion process were covered by several presentations.
There are multiple pathways by which the parasite can invade the erythrocyte (Barnwell and Galinsky, 1998). One of these pathways relies on the presence of sialic acid on the host cell plasma membrane. The erythrocyte binding antigen-175 (EBA175), a ligand that binds to sialic acid on glycophorin A, is expressed in all parasite lines examined so far. A more recently discovered, different family of ligands, called P. falciparum reticulocyte binding proteins (PfRBPs) or P. falciparum reticulocyte binding protein homologues (PfRh), includes some members that bind to erythrocytes and facilitate sialic acid-independent erythrocyte invasion (Rayner et al., 2001; Duraisingh et al., 2003). Unlike EBA175, these proteins are expressed differently in different parasite lines, raising the interesting possibility that they are involved in switching between sialic acid-dependent and sialic acid-independent invasion pathways. As summarized by Alan Cowman (WEHI, Melbourne, Australia), his group has shown that the disruption of genes encoding either EBA175 or PfRh1 results in a decreased sialic acid dependence for invasion. These observations provide a plausible and fascinating explanation for how malaria parasites evade an immune response that interferes with a specific invasion pathway.
Although the parasite appears to be able to switch between various invasion pathways, it is possible to induce a humoral immune response that inhibits invasion of merozoite stages both in vitro and in animal models. This immune response can be elicited by immunization with the apical membrane protein 1 (AMA1) which is therefore considered an important vaccine candidate (Stowers et al., 2002; Casey et al., 2004). Further evidence for the significant role of AMA1 in the invasion process comes from observations that peptides identified in phage display libraries and which bind to AMA1 are able to block invasion (Keizer et al., 2003). AMA1 is an integral membrane protein and has an ectodomain consisting of three subdomains stabilized by intramolecular disulphide bridges. The correct conformation of the ectodomain is essential for the induction of protective antibodies. Raymond Norton (WEHI, Melbourne, Australia) described a detailed structural analysis of two of the three subdomains as a prerequisite for the production of recombinant AMA1 for immunization studies. This presentation elegantly illustrated the significance of structural information for the design of effective vaccine molecules.
The active mechanism by which Plasmodium and other apicomplexan parasites invade cells has been linked to the phenomenon of gliding motility. As described by Dominique Soldati (Imperial College, London, UK) the molecular motor that mediates gliding activity of another apicomplexan, Toxoplasma gondii, has been dissected in considerable detail. The actin-based motor consists of a myosin heavy and light chain anchored in the inner membrane complex and linked to the cytoplasmic tail of microneme proteins (MICs) via aldolase, an F-actin binding protein. The MICs are associated into complexes and are involved in binding to host cells. During invasion, the transmembrane MICs are cleaved and released from the parasite surface. Cleavage occurs within the membrane-spanning domain and is mediated by a protease, which most probably belongs to the evolutionarily conserved rhomboid family of serine proteases. Several genes encoding members of this family are found in apicomplexan genomes, including that of P. falciparum. Information obtained using the more-amenable T. gondii system provides a conceptual framework for the identification of functionally equivalent components in Plasmodium.
Invasion of the erythrocyte and establishment of the malaria parasite within its host cell is likely to involve proteases and other enzymes with hydrolytic properties. Matthew Bogyo (Stanford Medical School, Stanford, USA) presented an elegant study in which chemical probes were designed to bind covalently to cysteine residues in the active site of cysteine proteases. Binding and cross-linking occur as the result of enzyme activity. The compounds were coupled to fluorescent dyes that allowed the identification and localization of active cysteine proteases. The highest levels of these enzymes were detectable in trophozoite and schizont stages, i.e. when the parasite differentiates to eventually become an invasive merozoite. By screening of combinatorial libraries, derivatives could be isolated that inhibited specific cysteine proteases and thus allowed an analysis of the consequences of ‘lack of function’. The experimental potential of this novel approach was exemplified by the identification of a protease that appears to have an essential role during invasion. There was a lively discussion on the value of small chemical probes for the analysis of protein function, with the point being emphasized that the approach avoids difficulties that can arise when one attempts to knock out essential genes in a haploid organism.
Modification of the host erythrocyte
As the parasite grows and develops within its host erythrocyte, it induces a range of morphologically and biochemically distinct modifications of the erythrocyte. An entire session dealt with the cell biological and biochemical mechanisms underlying host cell modification. In a presentation by Lawrence Bannister (Guy's Hospital, London, UK), it became clear that a precise ultrastructural description of the infected erythrocyte is an important prerequisite for a meaningful interpretation of molecular data. A detailed electron microscopical analysis of serial sections revealed new insights into the morphology of various developmental stages of the parasite with a focus on the secretory compartments. In contrast to most other eukaryotic cells, the parasite Golgi apparatus is reduced to a single cisterna. Coated vesicles derived from the nuclear envelope/endoplasmic reticulum complex form a trafficking route to the single Golgi cisterna. Concomitant with the first nuclear divisions during schizogony, the cisterna replicates and rhoptries are formed, whereas the biogenesis of micronemes and dense granules occurs immediately before the separation of merozoites from the residual body.
It is generally believed that transport of membrane proteins from the parasite to destinations within the erythrocyte occurs via transport vesicles (Taraschi et al., 2003), but it has remained unclear whether such vesicles traverse the vacuolar space. Ultrastructural data revealed vesicles lined by two membranes budding from the parasite's endoplasmic reticulum plasma (Lawrence Bannister, Guy's Hospital, London, UK). It is plausible that such vesicles, after fusion with parasite plasma membrane, release single membrane-bound vesicles into the vacuolar space, explaining how membrane proteins traverse the vacuolar lumen. Immunolocalization studies will clarify which parasite proteins traverse this pathway to reach specific destinations within the erythrocyte, which itself lacks the complex machinery required for intracellular protein trafficking. The transport of parasite proteins that become integral components of the erythrocyte plasma membrane is of particular interest as members of the family of variant P. falciparum erythrocyte membrane proteins (PfEMP-1) mediate adherence of infected erythrocytes to endothelial cells and thus contribute to the severe clinical symptoms of human malaria (Baruch et al., 1995; Smith et al., 1995; Su et al., 1995). PfEMP-1 molecules are complexed with other proteins such as the ‘knob associated histidine rich protein’ (KAHRP) and PfEMP-3 at the cytosolic face of the erythrocyte membrane to form ultrastructurally visible protrusions called knobs (Cooke et al., 2000). PfEMP-1 proteins are also found associated with the so-called Maurer's clefts, a membrane-bound compartment that appears in the cytoplasm of infected erythrocytes and that is located in the vicinity of the host cell plasma membrane. It is therefore likely that the Maurer's clefts act as transit compartments that are involved in exposing parasite proteins on the erythrocyte plasma membrane (Kriek et al., 2003). A key protein, PfSBP1, is believed to play an important role in tethering the clefts to the cytoskeleton of the erythrocyte plasma membrane (Blisnick et al., 2000). Brian Cooke (Monash University, Melbourne, Australia) reported on a targeted disruption of the PfSBP1 gene resulting in viable parasites that replicate normally. However, erythrocytes infected with the transgenic parasites show markedly reduced copy numbers of PfEMP-1 on the surface and a reduction of cytoadherence under flow conditions. While these observations underscore a role of PfSBP1 in the translocation of PfEMP-1 onto the erythrocyte membrane, the transport of other knob-associated proteins does not appear to be affected in the transgenic parasites.
Unlike membrane proteins that are transported either in vesicles or as part of endomembrane systems, soluble parasite proteins destined for the erythrocyte cytoplasm are first secreted into the parasitophorous vacuole and, in a second step, translocated across the vacuolar membrane (Ansorge et al., 1996; Wickham et al., 2001). What are the signal sequences within such proteins that determine that some proteins remain vacuolar resident proteins whereas others are translocated across the vacuolar membrane into the erythrocyte cytosol? Data presented by Luisa Hiller (from Kasturi Haldar's Laboratory, Northwestern University, Chicago, USA) and by Leann Tilley (La Trobe University, Melbourne, Australia) using transgenic parasites transfected with various reporter constructs indicated that translocation of the green fluorescent protein (GFP) across the vacuolar membrane requires additional sequences from parasite proteins found in the cytoplasm of the infected cell. Petra Burghaus (Philipps University, Marburg, Germany) obtained different data using luciferase either from Renilla or from Photinus as a reporter protein. No additional parasite-specific sequences were required to target these proteins across the vacuolar membrane. The solution to this apparent discrepancy necessitates a molecular analysis of the protein machinery within the vacuole that is involved either in the retention of vacuolar resident proteins or, alternatively, in the forward transport of proteins destined for a location beyond the vacuolar membrane. An initial characterization of the vacuolar proteome, aimed at identifying auxiliary proteins involved in signal sequence recognition and protein translocation, has led to the detection of two chaperones that could possibly play a role in the unfolding of proteins destined for a translocation across the vacuolar membrane (J. Nyalwidhe and K. Lingelbach, unpubl. obs.).
In addition to morphological and structural changes to the infected erythrocyte, there are profound changes in the physiological properties of the erythrocyte membrane. It has long been known that as the intracellular parasite enters the metabolically active trophozoite stage there is an increase in the permeability of the erythrocyte membrane to a wide range of low-molecular-weight organic solutes and ions (Kirk, 2001). The increase is attributable to the appearance, in the membrane, of ‘New Permeability Pathways’ (NPP), thought to be one or more types of channel. There is evidence that these pathways play a key role in the uptake into the infected cell of a number of essential nutrients required by the parasite, as well as providing the major route of entry into the infected cell of a number of anti-plasmodial agents. Despite their likely significance, however, their nature, origin and identity remain to be established. There has been a recent profusion of studies using variants of the ‘patch clamp’ technique to investigate ion channel activity in the infected erythrocyte membrane, and Sanjay Desai (NIH, Bethesda, USA) and Serge Thomas (Station Biologique, Roscoff, France) both presented data of this type at the meeting. Although there are debates on the proper techniques for obtaining acceptable recordings, there is broad agreement that the infected erythrocyte membrane has a much greater ionic conductance than that of uninfected erythrocytes and that this results from the activity of ion channels that are not seen under normal conditions in uninfected erythrocytes. However, it remains controversial how many different channel types are involved, whether the channels are of host or parasite origin, and which (if any) of the channels giving rise to the electrical conductances measured in the patch clamp experiments are actually those that mediate the enhanced uptake of organic nutrients and drugs into the infected erythrocyte (Staines et al., 2004).
Parasite physiology, biochemistry and drug resistance
A number of presentations at the meeting dealt with the physiology of the intraerythrocytic parasite itself. Richard Allen (ANU, Canberra, Australia) described the mechanisms involved in the uptake of K+ and maintenance of the membrane potential in the parasite, whereas Pat Bray (Liverpool School of Tropical Medicine, UK) presented a characterization of the uptake of the phospholipid precursor, choline, by the parasite. In the original annotation of the P. falciparum genome, there was a relative paucity of membrane transport proteins. However, Rowena Martin (ANU, Canberra, Australia) presented the results of a detailed bioinformatic analysis of transport proteins encoded within the parasite genome, in which the number of such proteins that have been tentatively identified has been extended quite significantly.
The use of ion-sensitive dyes to monitor H+ and Ca2+ concentrations in both the parasite cytosol and its acidic digestive vacuole was described in a number of presentations. Oliver Billker (Imperial College, London, UK) presented the results of an elegant study in which Plasmodium berghei parasites were stably transfected with a bioluminescent Ca2+ sensor, thereby providing a sensitive means to detect increases in cytosolic Ca2+ at different points in the life cycle. Using this approach, it was shown that xanthurenic acid (XA), a small mosquito-derived molecule which induces insect-stage parasites to differentiate into male or female gametes, triggers a cytosolic Ca2+ signal that is essential for parasite transmission (Billker et al., 2004). Giancarlo Biagini (Liverpool School of Tropical Medicine, UK) presented evidence that in asexual parasites the digestive vacuole serves as an intracellular Ca2+ store that accumulates Ca2+ through the action of a Ca2+-ATPase, postulated to be PfATP6, a putative sarco/endoplasmic reticululum Ca2+-ATPase (Biagini et al., 2003). As described by Sanjeev Krishna (St George's Hospital Medical School, London, UK), the same protein is inhibited with high potency by artemisinins and it is proposed that this inhibition underlies the anti-plasmodial action of these important anti-malarials (Eckstein-Ludwig et al., 2003)
Although the parasite's digestive vacuole might have a Ca2+ signalling function, its best-understood role is the digestion of haemoglobin, deposited in the vacuole via an endocytotic feeding mechanism. The parasite has a number of cysteine and aspartic proteases and Philip Rosenthal (University of California, San Francisco, USA) reported the results of gene disruption experiments which provide insights into the roles that these enzymes play in the intraerythrocytic cycle. Disruption of the gene-encoding falcipain-2, a cysteine protease, caused a transient perturbation of digestive vacuole cysteine protease activity and morphology, indicative of this enzyme playing a role in haemoglobin hydrolysis.
A challenge in the post-genome era is the development of algorithms that predict protein localization and function. Apicomplexan parasites possess a rudimentary plastid called the ‘apicoplast’ that was acquired by a free-living ancestor of these parasites. Although the apicoplast is essential for parasite survival, its function remains largely unknown. Apicoplast proteins are synthesized in the cytosol of the parasite and transported to the apicoplast by virtue of a bipartite signal sequence consisting of an ER targeting signal and a plastid targeting signal. Geoff McFadden (University of Melbourne, Australia) described an algorithm that identifies such motifs in the plasmodial genome database. The predictive value of this algorithm was tested by extensive mutagenesis of apicoplast targeting signals followed by in vivo analyses that assessed the intracellular distribution of GFP when fused to the mutgenized signals. Using an algorithm finely tuned on the basis of such experimental data, more than 500 putative apicoplast proteins have been identified and await further characterization (Ralph et al., 2004). These include complete sets of enzymes required for fatty acid and isoprenoid synthesis, and enzymes involved in haem synthesis. The presentation clearly demonstrated the potential of computational analyses in predicting the compartmentation of biochemical functions, as a prerequisite for further experimental approaches towards the assessment of potential drug targets.
There was considerable interest at the meeting in mechanisms of drug action and anti-malarial drug resistance. Kevin Saliba (ANU, Canberra, Australia) highlighted the dependence of parasite growth on the vitamin pantothenic acid and presented evidence that pantothenate analogues can inhibit parasite growth via an effect on pantothenate kinase (the first enzyme in the pathway by which pantothenate is converted to Coenzyme A), whereas Katja Becker (Interdisciplinary Research Centre, Giessen, Germany) focused on the potential of the small red-ox active protein plasmoredoxin, and of the parasite's single glutathione S-transferase as drug targets. For all but a few of the anti-malarial drugs currently in use, the parasite has demonstrated the ability to develop high levels of resistance. For some drugs, resistance has arisen through mutations in the target protein (e.g. dihydrofolate reductase-thymidylate synthase) and presentations from Yongyuth Yuthavong (National Centre for Genetic Engineering and Biotechnology, Pathumthani, Thailand) and Worachart Sirawaraporn (Mahidol University, Bangkok, Thailand) dealt with the structural basis of resistance that has arisen through this mechanism, and with the possibility of developing new inhibitors that take advantage of the inherent limitations in the mutations that enzymes can undergo. For other drugs, resistance involves proteins other than the target. A number of presentations dealt with the two digestive vacuole membrane proteins that have been implicated in chloroquine resistance – the P. falciparum chloroquine resistance transporter (PfCRT) and P-glycoprotein homologue 1 (Pgh1, encoded by the pfmdr1 gene). Although PfCRT was originally reported as belonging to a previously undescribed family of transporters or channels (Fidock et al., 2000), recent bioinformatics analyses (Martin and Kirk, 2004; Tran and Saier, 2004) place it as a member of the ‘Drug/Metabolite Transporter’ superfamily, members of which include transporters for amino acids, weak bases and divalent organic cations. As described by David Fidock (Albert Einstein College of Medicine, New York, USA), a mutation at position 76, in the first of 10 transmembrane domains in PfCRT, plays a key role both in chloroquine resistance and in the ability of verapamil to reverse that resistance. There was a broad consensus at the meeting that chloroquine resistance is likely to arise as a result of PfCRT transporting the drug away from its primary site of action, in the digestive vacuole. However, the transport of chloroquine by PfCRT is yet to be demonstrated experimentally. Mutations in Pgh1 can enhance (but not, by themselves, induce) chloroquine resistance (Reed et al., 2000) and Rhys Hayward (ANU, Canberra, Australia) showed that at least some such mutations reduce the rate of parasite propagation. Clare Gavigan (Trinity College, Dublin, Ireland) reported that cyclosporins inhibit parasite growth and proposed that they do so via an effect on Pgh1, while Anne-Catrin Uhlemann (St George's Hospital Medical School, London, UK) presented evidence that resistance to another anti-malarial, mefloquine, results from increased copy number of pfmdr1 (Price and Uhlemann et al., 2004). Whether this mefloquine resistance arises as a result of the drug being transported by Pgh1 remains to be established.
In the field, resistance has arisen to all classes of anti-malarial drugs except the artemisinins. Nick White (Mahidol University, Bangkok, Thailand) emphasized that combination therapies can slow or even prevent the emergence of anti-malarial drug resistance; however, the costs are significant.
Genomics technologies and applications for malaria research
The discovery and characterization of biological processes in the parasite and their potential exploitation to develop improved anti-malarial interventions was the impetus behind the completion of the malaria genome. To complement that of P. falciparum, five additional Plasmodium genome projects from related species, including another human malaria parasite species, Plasmodium vivax, have been initiated and are at various stages of completion. Combined with genome sequences of evolutionarily related Apicomplexan parasites, these data provide the foundation of comparative genomics studies – studies that may faciliate the understanding of host–parasite interactions and eventually lead to targeted disruption of those crucial steps in parasite biology. Follow-on technologies such as high-throughput proteomics, DNA microarrays and systems biology, which rely on having an accurate and complete genome, have come in force to assist malaria biologists in understanding Plasmodium. For example, expanding on the recent characterization of large scale protein expression by high-throughput proteomics (Florens et al., 2002; Lasonder et al., 2002), several groups have used similar approaches to identify the subcellular location of proteins from parasite organelles, from infected erythrocyte membranes (Florens et al., 2004) and even from the specialized Maurer's clefts structure in the cytosol of these infected erythrocytes (Sam-Yellowe et al., 2004). The proteomics data have been used to select and test new vaccine candidate antigens in T-cell assays from individuals protected by immunization with radiation-attenuated sporozoites which may represent antigens that are more likely to be protective than antigens currently under clinical development (Doolan et al., 2003).
Proteomics tools have provided a means to study parasite biology at a level of detail that previously was not easily possible. For example, at the meeting Shahid Khan (LUMC, Leiden, the Netherlands) described a proteomics approach that revealed differences in the proteins expressed in the male and female gametes as well as reporting a proteomic analysis of proteins expressed by merozoites from three different species of Plasmodium, each of which has a particular trophism for its hosts erythrocytes.
Whereas large-scale quantitative measurement of transcripts was once the realm of DNA microarrays, advances in proteomics is now permitting a quantitative assessment of protein expression. One method uses an isotopically labelled amino acid to create a visible ‘shift’ in the peak seen in mass spectrometry (Gygi et al., 1999). By comparing the areas under the control and test sample, one can determine relative increases or decreases in protein expression. As the isotopically labelled amino acid must be incorporated at high efficiency into the test sample during protein synthesis, there are some obvious limitations with respect to the Plasmodium life cycle. The most obvious first choice to assess this approach in Plasmodium was in in vitro cultured blood-stage parasites. Nirmalan and colleagues showed that by using 13C6, 15N1-isoleucine in parasite cultures they could readily detect the expected 7 Da shift in the mass spectra peaks (Nirmalan et al., 2004). More importantly, they were able to reproducibly detect quantitative effects in protein expression in parasite cultures treated with either pyrimethamine or tetracycline. Although currently limited to use in in vitro cultures, quantitative proteomics adds to our armamentarium to study parasite biology.
Immunity, susceptibility and pathogenesis of disease
Immunity to malaria is slow to develop, is not sterile (i.e. does not protect completely against infection) and may be short-lived in the absence of exposure. Moreover, some components of the immune response might play a role in pathology. As there is a complex life cycle in the vertebrate host, where there are extracellular as well as intracellular stages in nucleated and non-nucleated cells, it is not surprising that the immune response is diverse and that different effector arms are responsible for elimination of the parasite at different stages. Understanding immunity and susceptibility to malaria is critical for the development of better protective strategies, yet we still have only limited tools to address this, although new ones are now emerging.
Kevin Marsh (KEMRI, Kilifi, Kenya) reviewed the current status of studies of clinical immunity to malaria. In studies of human infections, the mechanisms of parasite elimination are not really known and we are hampered by the lack of assays that indicate clinical immunity. To date there are few examples of strong correlations between antibody or T-cell responses to a single antigen and immune status. In many instances, studies of human responses have yielded conflicting results and there are several reasons for this. Correct study design is essential and matching must be appropriate: too much matching might have ‘matched out’ any protective effects. Diversity in parasite and host genetics is extensive and has not always been taken into account, and responses have generally been measured as total antibody or T-cell proliferation rather than functional assays. For us to study malaria immunity properly, it is important that we measure the right thing. The complex relationship between parasite and morbidity makes definition of protective immunity difficult. It was highlighted that there are relatively few systematic studies of the kinetics and dynamics of the immune response. Many studies use crude measures of antibody or T-cell responses with vaccine candidate molecules as antigens. We need to take into account polymorphisms of parasite antigens and concentrate on assays that, for example, can take into account effector functions of antibodies.
New ways of looking at immune response are now possible using microarray technology. These might provide us with a more global overview of genes which are switched on or off in individuals who are clinically immune or who develop severe malaria, and thus could help define effector pathways involved. Clear patterns and pathways of gene activation, especially related to innate immune response, were demonstrated in studies presented by Chris Ockenhouse (WRAIR, Silver Spring, USA) and by Jürgen Kun (University of Tübingen, Germany). In challenge infections of human volunteers and in children with symptomatic malaria, respectively, clear patterns and pathways of gene activation, especially related to innate immune response are seen. T-cell-specific genes were downregulated, whereas monocyte and macrophage genes were switched on. Fascinatingly, striking changes can already be detected in gene expression well before parasitaemia is symptomatic or is detectable by microscopy. However, this approach is still at a very preliminary stage, and it will be necessary to incorporate this ‘fingerprinting’ into functional assays of protection and responses to specific plasmodial antigens.
A recent comparison by Jim Kazura and colleagues (Case Western Reserve University, USA) in western Kenya, between a functional assay, growth inhibition resulting from antibody to a specific protein domain of the merozoite surface protein 1 (MSP-119) and total antibody to this domain measured by ELISA, supports the value of functional assays, and teams working on measuring immunity to malaria vaccines are developing similar functional assays. These might allow more robust correlations between antibody levels and vaccine efficacy, for example, although confounding factors such as inflammatory mediators with direct anti-parasitic effects in serum require careful exclusion.
Human studies will always be correlative, and well-controlled studies in experimental models are valuable in dissecting out important effector mechanisms and providing evidence of causal links between a particular response and protective immunity or pathology. Human studies show the importance of antibody response, but information on the role of T-cell immunology lags behind. In mice, CD8+ T-cell memory is clearly important in protection from the liver stage infection and Fidel Zavala (John Hopkins University, Baltimore, USA) presented evidence that the IL4 receptor is critical to memory development. CD8+ T-cells are also critical to the development of cerebral malaria in mice, as animals lacking these cells, or lacking perforin secreted by CD8 T-cells, are protected from severe disease (Belnoue et al., 2002; Sarah Potter and Nick Hunt, University of Sydney, Australia). For blood-stage immunity, B-cells and CD4 T-cells are important and MSP-1 responses are being studied as a prototype in relations to patterns of antigen presentation and cell activation using TCR transgenic mice to investigate how memory develops (Jean Langhorne, NIMR, London, UK).
It is still not really clear whether MSP-1 is a target of the naturally acquired protective immune response in humans. Although there are studies showing an association between antibodies to MSP-1 that can inhibit merozoite invasion of erythrocytes, it is extremely difficult to demonstrate a causative link. A mouse study presented by Richard Carter (University of Edinburgh, UK) demonstrated that mice infected with Plasmodium chabaudi, using a ‘linkage group selection’ method, using genetic analysis of parasite genetic markers which are reduced or eliminated under selection, has directly shown that MSP-1 is a target of infection-induced immunity.
If T- and B-cell responses are critical, how do we elicit the ‘best’ responses? Michael Good and colleagues (QIMR, Brisbane, Australia) showed that apoptotic depletion of CD4 T-cells and B-cells occurs upon infection of mice, suggesting that apoptosis plays a role in regulating responses. Subpatent infections, however, lead to much lower T- and B-cell apoptosis and stimulate a protective response against homologous and heterologous challenge. The next step is to determine how to extend this to the clinical arena, and how it might relate to suppression of parasitaemia and evolution of immunity, especially in the context of partially effective drugs or vaccines. Innate immune responses contribute to pathogenesis and protection from malaria, and this is an area of active research. Pathogenesis of severe disease could result from a dysregulated immune response. For example, nitric oxide might protect against severe disease and evidence presented by Nick Anstey (Menzies School of Health Research, Darwin, Australia) suggests that when arginine (a substrate for nitric oxide production) is deficient, monocytes instead synthesize toxic superoxide, exacerbating cytokine release and endothelial activation. Diana Hansen (WEHI, Melbourne, Australia) showed how innate immune system Natural Killer Complex loci appear to affect susceptibility of mice to cerebral malaria. More complex is the relationship between lymphotoxin alpha (a TNF-like cytokine) and disease severity (Christian Engwerda, QIMR, Brisbane, Australia). Lymphotoxin alpha knockout mice are resistant to cerebral malaria, and microvascular endothelial cells in the brain are an important source of this cytokine. Lymphotoxin alpha appears to mediate disease pathogenesis in the local tissue microenvironment, rather than systemically. Monocyte and endothelial cell activation also lead to platelet deposition and the shedding of microparticles, tiny fragments of membrane from platelets, leukocytes and endothelial cells, studied by Georges Grau and colleagues (Université de la Méditerranée, Marseille, France). The former might act as a bridge between brain endothelium (often CD36 deficient) and infected erythrocytes, which frequently adhere to CD36. Microparticles might represent a marker of severity and have a role in the pathogenesis of malaria disease.
For over a dozen years we have known that P. falciparum undergoes antigenic variation, which might be important in immune evasion, and about 10 years ago this variation was found to result from expression of different members of the var gene family coding PfEMP-1 (Baruch et al., 1995; Smith et al., 1995; Su et al., 1995), the dominant red cell surface antigen. The mechanisms underlying silencing and switching of var genes are becoming clearer (Deitsch et al., 2001). Chris Newbold and colleagues (Weatherall Institute for Molecular Medicine, Oxford, UK) studied cloned cultured parasites and found that the rates at which var genes are turned on and off appear to vary widely for each var gene. Switching from certain common vars to others does not take place. There might be a regulatory focus in the genome encoding these varied switched rates. This allows for some stratification of the var repertoire and ensures that not all vars are switched on rapidly during early infection, which would exhaust the repertoire. Obvious regulatory elements such as untranslated regions do not seem to be involved but both 5′-UTR and the var intron might regulate transcription. Kirk Deitsch (Weill Cornell Medical College, New York, USA) showed that full-length var transcription is high in ring stages and low in trophozoites, whereas trophozoites produce sterile transcripts initiated within the intron, which may interact with promoter sequences by forming DNA loops, thus promoting chromatin condensation and var silencing. Nuclear organization appears critical both for var gene ectopic crossover events, leading to antigenic adversity, and for the regulation of expression. Possible epigenetic mechanisms include remodelling of chromatin structure mediated by Sir proteins and changes in histone acetylation and methylation, discussed by Artur Scherf (Institute Pasteur, Paris), and the processes might be similar to those in yeasts, where genes that control chromatin structure influence transcription in conjunction with changes in histone modifications. Complementary data suggest there may be ‘activation clusters’ (Till Voss, WEHI, Melbourne, Australia).
As we learn more about var gene regulation, we also begin to understand more about the relationship between var transcription and clinical disease. Clinical studies suggest that common PfEMP-1 variants cause severe disease in young children (Bull et al., 2000). Rare PfEMP-1 variants are less avid binders to host receptors, but might be more diverse, less antigenic, and important in chronic infection. Field studies are beginning to ask whether these common var gene sequences are associated with severe malaria (Hans-Peter Beck, Swiss Tropical Institute, Basle, Switzerland; Anja Jensen, University of Copenhagen, Denmark). Identifying such sequences might assist vaccine development, and PfEMP-1 domains packaged into virus particles appear to be immunogenic and protective in small animals, supporting the development of a PfEMP-1 vaccine (Qijun Chen, Swedish Institute for Infectious Control, Stockholm, Sweden). Studies were presented by Alyssa Barry (University of Oxford, UK) examining the breadth of var repertoires within a community at the genomic level, to examine population structure and the constraints within one community on population diversity.
The strongest case that a unique subset of relatively conserved var genes causes disease has been demonstrated in P. falciparum-infected pregnant women. Parasites that adhere to the placental receptors chondroitin sulphate and hyaluronic acid, and isolates from the placenta, show increased expression by real-time polymerase chain reaction (PCR) of the var2csa gene (Salanti et al., 2003). However, it has been hard to show that var2csa is translated into a PfEMP-1 expressed in pregnant women, a critical step on the path to a pregnancy-specific vaccine (Fried et al., 2004). Furthermore, the relative importance of other placental ligands has not yet been determined, suggesting that we should be cautious in ascribing disease in pregnant women only to var genes.
var genes have not been found in other plasmodia, although some rodent parasites exhibit variant antigen expression on the erythrocyte surface. Another multigene family, vir, comprising several hundred members, has been described for P. vivax (del Portillo et al., 2001). The VIR proteins are expressed on the surface of infected erythrocytes. Multiple genes are expressed during an infection, and elicit an antibody response. The mechanism of switching is not known, but Hernando del Portillo (University of Sao Paolo, Brazil) reported that more than one VIR protein is expressed in a single cell. Screening of the rodent genome revealed that this multigene family is also present in the rodent malarias, Plasmodium yoelii, P. chabaudi and P. berghei. The yir genes are also expressed on the surface of the infected red cell, and Deidre Cunningham (NIMR, London, UK) reported that their expression is changed by an immune response, supporting the idea that these antigens might play a role in immune evasion.
Molecular biology of transmitting malaria parasites
Transmission essentially refers to the journey the parasite makes from one vertebrate host through the mosquito vector to a second vertebrate. This is a highly complex series of developmental and cell–cell interactive steps involving gamete formation from the ingested gametocytes, gamete fertilization and subsequent development of the zygote into a motile and invasive ookinete (Fig. 1). The ookinete traverses the midgut wall of the mosquito, where it forms the oocyst upon encountering the basal lamina at the external surface of the midgut. After a period of vegetative growth and replication, ∼12 000 sporozoites are released that migrate through the haemolymph and penetrate the salivary glands (SGs), where they are stored and released into the host bloodstream upon ingestion of a blood meal. Successful sporozoites will ultimately reside and develop in hepatocytes. This transition occurs against a background of innate immune responses of the mosquito that is stimulated by midgut tissue damage and parasite presence, which help ensure an attritional environment that the parasite must overcome. Several presentations emphasized the value of animal models of malaria and in particular the use of Plasmodium species that infect rodents. Their use is valid as the emerging genome data for three of the rodent malaria parasites have shown that all are highly orthologous in the core genes to P. falciparum (and therefore in all probability to all species of Plasmodium) and that most genes expressed in Plasmodium transmission stages are conserved. Furthermore, they are safe to use experimentally as well as often being relatively easy to manipulate genetically (Gil Carvalho and Ménard, 2004).
Both mature male and female gametocytes circulating in the bloodstream are in a state of cell cycle arrest (G0). Proteomic studies on purified gender-specific gametocyte populations from P. berghei revealed that the gender-specific populations are indicative of their forthcoming developmental roles once they are activated in the mosquito midgut (for review, see Janse and Waters, 2004). Females are in a state or readiness for fertilization and have a full complement of ribosomes and many RNA binding proteins that might influence the stocks of specific mRNA species that are cytoplasmic but not yet available for translation (notably that encoding P28 and ookinete surface proteins; Paton et al., 1993). They alone have a unique repertoire of secreted proteins containing LCCL proteins that are anticipated to play a role in cell–cell interactions shown to be unique to gametocytes and subsequent stages (Claudianos et al., 2002; Pradel et al., 2004). Male gametocytes have degraded ribosomes and, instead, have stockpiled proteins, notably structural proteins associated with motility and DNA replication, and a greater proportion of unique proteins compared with the female gametocytes (Shahid Khan, LUMC, Leiden, the Netherlands).
Once transferred to the milieu of the mosquito midgut (or its experimental equivalent which for P. berghei permits the development of fully infectious ookinetes in vitro), gametocytes are reactivated and emerge from the erythrocyte in response to a drop in temperature and undergo final gametogenesis in response to a second environmental cue, XA, a mosquito eye pigment precursor and metabolite of trytophan that is also present in the midgut. Female gametogenesis is discrete when viewed in the light microscope, with few fundamental alterations being observed. However, male gametogenesis is one of the more spectacular events of the Plasmodium life cycle. Within 10 min, the male genome is replicated threefold giving rise to eight nuclei that are packaged into motile flagelles that detach in a process known as exflagellation from the residual cell body (the exflagellation centre). The process appears violent and energetic, and uninfected bystander erythrocytes can become attached to an individual male gamete and are tossed around like a rag-doll. Clearly gametogenesis would be expected to involve a signal transduction pathway, one stimulated by XA. Mobilization of intracellular calcium stocks is also known to be important to gametogenesis (Janse and Waters, 2004). All of these observations were tied together by the work reported by Oliver Billker (Imperial College, London, UK) showing that internal calcium mobilization in gametocytes is dependent on XA stimulation and that the process of female gamete maturation and male gamete formation is dependent on the presence of a single calcium-dependent protein kinase shared by both genders (Billker et al., 2004). Thus, the first signalling pathway in Plasmodium has been defined and, although the remaining details await definition and the possibility of different branches downstream must be investigated, a significant landmark has been achieved.
Once formed, the job of the gametes is fertilization. It is well established that this is probably a receptor ligand-mediated event that is susceptible to interruption by specific antibodies. In particular, IgG antibodies that recognize a conserved cysteine-rich, GPI-anchored protein P48/45 (which contains two so-called 6-cys domains) and IgM antibodies (with complement) that are specific for another member of the same family, P230 (that contains six 6-cys domains and forms a protein complex with P48/45 and is thereby retained at the parasite surface) are effective in blocking fertilization and reducing transmission of P. falciparum (Kaslow, 2002). Furthermore, even though fluorescence and proteome studies show that P48/45 is present in both genders, a specific role in male gamete fertility was demonstrated by gene disruption studies using P. berghei (van Dijk et al., 2001). Earlier work had shown that of the 10 members of the P48/45 family (all predicted to be expressed at the parasite surface) a further four were exclusively transcribed in gametocytes. Milly van Dijk (LUMC, Leiden, the Netherlands) reported that gene disruption studies in this family have revealed a female-specific ligand and confirmed that the P48/45–P230 complex exists and controls male gamete fertility.
After the vegetative growth period, development and release of the sporozoite from the oocyst into the haemolymph, this motile form must invade the lumen of the SG. There is increasing evidence that this invasion process results in a qualitative difference in the sporozoites. SG sporozoites are more infectious than those isolated from haemolymph upon mechanical passage and studies have now shown that SG invasion triggers the expression of an additional repertoire of genes that includes protein kinases, four proteins with signal peptides and proteins involved in translation regulation (Matuschewski et al., 2002). What is not clear is whether there is a concomitant downregulation of certain genes in SG sporozoites that would pinpoint those that might be exclusively involved in sporozoite colonization of the SG. UIS4, one of these proteins that has a signal peptide, has been investigated further by gene disruption in P. yoelii. UIS4 was found to be essential for the complete development of the parasite in exoerythrocytic stage (Stefan Kappe, BRI, Seattle, USA).
Whole-genome sequencing revealed the complete catalogue of P. falciparum genes (Gardner et al., 2002) and the majority of those in the rodent malaria parasite genomes (Carlton et al., 2002) but the question of whether or not they are expressed is still being addressed. Whole-genome transcriptomes have been developed for many stages of P. falciparum but the liver stage remains a mystery. However, Affymetrix-format oligo-based arrays developed to the published P. yoelii sequence (Le Roch et al., 2003) will facilitate the exploitation of the ability of the rodent parasites to develop as exoerythrocytic forms in hepatocyte cultures. Many hundreds of genes that are expressed in these stages have now been identified and, through comparison with the hybridization patterns generated by other parasite forms, significant numbers of exoerythrocytic stage-specific transcripts are now known (Lawrence Bergmann, Drexel University College of Medicine, Philadelphia, USA).
What controls stage-specific transcription is still an open question. One interesting feature of the annotation of the whole P. falciparum genome is that the number of readily identifiable transcription factors is very small. Cascades of transcription (transcripts to go) have been demonstrated in asexual blood-stage development and the same probably happens at all points of vegetative growth in the Plasmodium life cycle (Bozdech et al., 2003; Le Roch et al., 2003; Waters, 2003). In the absence of a range of clearly definable transcription factors, one alternative hypothesis would be that epigenetic chromatin accessibility would determine gene expression. Histone modification is well known to play a central role in determining the accessibility of genes, largely through phosphorylation, methylation and acetylation of the core histones that form the nucelosomes around which DNA is wrapped. Modification of the histones leads to relaxation of nucleosome structure and greater accessibility to the transcriptional apparatus (Berger, 2002). Therefore, identification of modified histones is an indication of active chromatin and is referred to as the histone code. Studies on P. falciparum to date have demonstrated that there are modified histones in Plasmodium chromatin and future studies using chromatin immunoprecipitation with antibodies that recognize the highly conserved modified histones are expected to show that modified histones are localized to active areas of the genome (Henk Stunnenberg, University of Nijmegen, the Netherlands).
The journey of the parasite from gametocyte to liver-stage hepatocyte is fraught with danger. The host immune system can be active against circulating gametocytes and sporozoites as well as attacking the infected hepatocyte as a result of MHC II-associated antigen presentation. However, the greatest biological bottlenecks that result in severe reduction of parasite numbers result from interaction of the parasite with the mosquito innate immunity (MII) mechanisms (Dimopoulos et al., 2002). The line between success and failure can be fine for the parasite. A susceptible strain of A. gambiae kills 80% of the ookinetes and the resistant mosquito line 100% of ookinetes largely through melanization of the parasite as it traverses or emerges at the midgut epithelium. Thanks to recent technological breakthroughs, notably the completion and annotation of the A. gambiae genome sequence (Holt et al., 2002) and the application of RNAi-based technologies to specific gene knock-down (Blandin et al., 2002), it is possible to analyse swiftly the effect of interesting mosquito genes on parasite transmission. Bioinformatic analysis of the genome identified a thioester-containing protein (TEP) that has the hallmarks of an effector molecule that could be involved in parasite killing. The alleles of TEP were quite different in susceptible and resistant mosquitoes and knock-down of the TEPR allele conferred parasite sensitivity. Pattern recognition molecules are also implicated in MII surveillance (Blandin et al., 2004).
In summary, genomes and genetic manipulation methodologies and model parasites have all been combined in recent years to enhance greatly our understanding of Plasmodium transmission. This exciting trend looks set to continue for many years yet and offers real promise that new therapies and therapeutics will be produced for efficacy testing.