Characterization of a Plasmodium falciparum Macrophage-Migration Inhibitory Factor Homologue

Abstract

Background. Macrophage-migration inhibitory factor (MIF), one of the first cytokines described, has a broad range of proinflammatory properties. The genome sequencing project of Plasmodium falciparum identified a parasite homologue of MIF. The protein is expressed during the asexual blood stages of the parasite life cycle that cause malarial disease. The identification of a parasite homologue of MIF raised the question of whether it affects monocyte function in a manner similar to its human counterpart.

Methods. Recombinant P. falciparum MIF (PfMIF) was generated and used in vitro to assess its influence on monocyte function. Antibodies generated against PfMIF were used to determine the expression profile and localization of the protein in blood-stage parasites. Antibody responses to PfMIF were determined in Kenyan children with acute malaria and in control subjects.

Results. PfMIF protein was expressed in asexual blood-stage parasites, localized to the Maurer's cleft. In vitro treatment of monocytes with PfMIF inhibited random migration and reduced the surface expression of Toll-like receptor (TLR) 2, TLR4, and CD86.

Conclusions. These results indicate that PfMIF is released during blood-stage malaria and potentially modulates the function of monocytes during acute P. falciparum infection.

More than 40% of the world's population is currently at risk of exposure to malaria, and an estimated 1.5–3 million deaths are attributed to Plasmodium infection per year [1, 2]. A majority of malaria cases occur in sub-Saharan Africa, with most deaths occurring in children <5 years of age. P. falciparum infection, especially severe malaria, is known to be associated with an acute inflammatory response. This is characterized by elevated levels of proinflammatory cytokines, such as tumor necrosis factor (TNF)-α, interferon-g, and interleukin (IL)-6 [3]. This inflammatory response may lead to an increase in the cytoadherence of infected red blood cells (iRBCs) to vascular endothelium, resulting in more-severe malarial disease [4]. A putative protein identified during the sequencing of the P. falciparum genome showed sequence homology to the proinflammatory cytokine macrophage-migration inhibitory factor (MIF) [5]. Microarray studies have suggested that P. falciparum MIF (PfMIF) mRNA is transcribed in late ring and early trophozoite stages of the asexual blood cycle of the parasite [6]. The identification of a MIF homologue in P. falciparum suggested a potential mechanism that might contribute to the proinflammatory cytokine profile observed during infection. Human MIF (huMIF) was one of the first cytokines identified and has a wide range of biological activities, including the induction of TNF-α, nitric oxide, IL-6, and IL-8 secretion; the up-regulation of Toll-like receptor (TLR) 4 and intercellular adhesion molecule (ICAM)-1 expression; and the suppression of the effects of glucocorticoids [7, 8]. MIF has been directly implicated in a wide range of infectious and immune-mediated diseases, including sepsis, rheumatoid arthritis, and diabetes [7]. Interestingly, homologues of MIF have been identified in filarial nematodes and in one tick species and show remarkable similarity to mammalian MIF in both crystal structure and in vitro biological activity [9, 10]. These homologues are thought to play an important role in parasite immune-evasion strategies.

The role played by PfMIF during the course of P. falciparum infection has not been determined. Therefore, we generated recombinant PfMIF expressed in bacteria and investigated the expression patterns and localization of PfMIF during the asexual blood-stage cycle of the parasite. We also examined the ability of recombinant PfMIF to modulate monocyte function.

Patients, Materials, and Methods

Patients. Blood samples were collected from children living in the Ngerenya area of Kilifi District, who were under active surveillance for malaria as detailed elsewhere [11]. We analyzed plasma collected from 117 children during the cross-sectional survey conducted during the low transmission season in October 2003. All children were examined clinically, and venous blood samples were collected for whole-blood counts and to determine the presence of malaria parasites. Children who were negative for P. falciparum blood-stage parasites by microscopy were included in the study. In August 2004 and January 2005, blood samples were collected from children with mild, uncomplicated malaria (fever of >37.5°C in association with a positive blood film for P. falciparum parasites and no alternative explanation on careful clinical examination) attending the outpatient clinic at Kilifi District Hospital and from children admitted to the wards with severe malaria. All 80 subjects included in this study were invited to provide a convalescent blood sample 14 days after discharge from the hospital; 35 convalescent samples were collected. The study was approved by the Kenya Medical Research Institute/National Ethical Review Committee and the Oxford Tropical Research Ethical Committee. Written, informed consent was obtained from the parents or guardians of the participating children.

Parasite culture. Intraerythrocytic-stage P. falciparum parasites derived from the ITG/A4 clone were cultured in vitro following a protocol described elsewhere [12]. Parasites cultures were synchronized using the sorbitol lysis method [13]. Tightly synchronized parasites were sampled throughout the asexual blood stages to perform Western blotting and immunofluorescence microscopy.

Cloning and expression of PfMIF. The PfMIF sequence was amplified by reverse-transcription polymerase chain reaction of P. falciparum total RNA by use of oligo(dT) primers and SuperScript II reverse transcriptase (Invitrogen). The 2 terminal primers used were 5′-GAATTCCATATGCCTTGCTGTGAAGTAATAACAAACG-3′ and 5′-CGCCCTAGGCTAGCCGAAAAGAGAACCAC-3′. The amplified DNA fragment was subcloned into a T7/NT-TOPO expression vector that contained an in-frame N-terminal histidine tag (Invitrogen). The construct containing the complete PfMIF sequence in correct orientation was transformed into the Escherichia coli BL21(DE3) pLysS strain. The transformed cells were cultured for 3 h with an overnight saturated culture used as an inoculum and then were induced at 0.6 OD₆₀₀ with 0.5 mmol/L isopropyl β-D-thiogalactopyranoside for another 4 h. The pelleted cells were lysed using Bugbuster reagent (Novagen), and the crude bacterial extract was purified through an Ni-NTA column (Invitrogen). The peak fractions, as determined by SDS-PAGE, were subjected to ion-exchange chromatography using DEAE Sepharose (Amersham Biosciences). The PfMIF protein eluted between 150 and 250 mmol/L NaCl. The purified protein was then denatured in 6 mol/L urea containing 10 mmol/L β-mercaptoethanol for 20 min, and subsequently the protein was dialyzed against a buffer consisting of 20 mmol/L Tris-HCl, 50 mmol/L NaCl (pH 7.5), and decreasing concentrations of urea, until final dialysis with no urea. Lipopolysaccharide (LPS) was removed using EndoTrap (Profos), and removal was confirmed using the Limulus Amebocyte Lysate assay (Cambrex). The LPS concentration in the recombinant protein was routinely found to be <2 pg/mg of protein.

Antibody production. Polyclonal anti-PfMIF serum was generated by immunization of New Zealand White rabbits with 4 subcutaneous injections of 50 µg of recombinant PfMIF over the course of 35 days. The initial injection was emulsified with Freund's complete adjuvant, with the following 3 injections emulsified with Freund's incomplete adjuvant. Serum was harvested on day 49 after the initial injection. An anti-peptide antibody against a PfMIF-specific peptide sequence (NRSNNSALADQITKC; Sigma-Genosys) was also generated in rabbits. Polyclonal anti-serum cross-reacted with huMIF at a dilution of up 1:200 but was specific for PfMIF at higher dilutions. The anti-peptide antibody did not cross-react with huMIF (figure 1). Rabbit IgG was purified from anti-PfMIF peptide serum by use of Protein G Sepharose (Amersham) following the standard protocol. Purified anti-PfMIF peptide IgG was used in subsequent experiments where indicated.

Figure 1.

Specificity of Plasmodium falciparum macrophage-migration inhibitory factor (PfMIF) antibodies at different dilutions. The figure and legend are available in their entirety in the online edition of the Journal of Infectious Diseases.

Northern blotting. Parasites were sorbitol synchronized, cultures were at 3%–10% parasitemia, and 200–500 µL of packed iRBCs were processed for each RNA sample. Cells were spun directly from culture, and packed cells were resuspended in the appropriate volume of TRIzol (Invitrogen). Samples were then processed as described elsewhere [14].

Immunoblotting. Tightly synchronized parasites were sampled during the ring and trophozoite stages of the life cycle. A small aliquot was removed for immunofluorescence microscopy, and the remaining iRBCs were lysed with 0.01% saponin (Sigma) and thoroughly washed to remove hemoglobin and other RBC proteins [15]. Purified parasite lysates were run on 12% SDS-PAGE gels and transferred to nitrocellulose membranes (Schleicher and Schuell) for immunoblotting with anti-PfMIF rabbit serum or purified IgG where indicated, followed by alkaline phosphatase-conjugated swine anti-rabbit IgG secondary antibody (Dako). Blots were developed using BCIP/NBT substrate solution (Invitrogen). For the detection of PfMIF in the culture supernatant, a 4% hematocrit culture of 20% trophozoites was grown overnight to allow schizont development and rupture. Culture supernatant was then collected and spun to remove any remaining RBC and used in Western blotting as described above.

Immunofluorescence microscopy. Fixed and permeabilized iRBCs were stained with purified rabbit anti-PfMIF peptide IgG and anti-P. falciparum skeleton-binding protein 1 (PfSBP1) (provided by Prof. Catherine Braun-Breton, Dynamique Moléculaire des Interactions Membranaires, University Montpellier 2, France). PfSBP1 has been shown to localize to the Maurer's cleft [16]. 4′-6-Diamidino-2-phenylindole was used to stain the parasite nuclei (Sigma). FITC-conjugated anti-rabbit IgG (Dako) and Alexa Fluor 546-conjugated anti-mouse IgG (Molecular Probes) were used as secondary antibodies.

Cellular assays. Monocytes were isolated from buffy coats (UK National Blood Service) by use of anti-CD14 magnetic beads (Miltenyi Biotec). Monocytes were used for migration assays or were cultured in 24-well plates at a concentration of 5×10⁵ cells/0.5 mL RPMI 1640 medium (Sigma) supplemented with 2 mmol/L glutamine, 50 µmol/L kanamycin, and 2% pooled human serum with or without exogenous PfMIF. In some experiments, monocytes were incubated with or without PfMIF for 12 h before the addition of LPS from E. coli 0111 (Sigma) or peptidoglycan (PGN) from Staphylococcus aureus (Sigma) for 24 h. After incubation at 37°C in 5% CO₂, supernatant was collected for ELISA. The cells were stained for surface molecules and analyzed by flow cytometry (FACScalibur), and the data were analyzed using FlowJo (version 6 [Mac]; Treestar). The following primary antibodies were used: antiCD54 (ICAM-1; Dako); anti-CD40 and anti-CD86 (Serotec); anti-TLR2 and anti-TLR4 (eBioscience); and anti-HLA-DR (Dako). The secondary antibody used was FITC-conjugated anti-mouse IgG (Dako).

Migration assays. Migration assays were performed using 24-well, 6.5-mm Transwell membranes with a 5.0-µm pore size (Corning). Briefly, 5×10⁴ purified monocytes in 100 µL of medium were added to the upper chamber with or without 100 ng/mL recombinant PfMIF or 100 pg/mL LPS, and 600 µL of medium with or without 100 ng/mL monocyte chemotactic protein (MCP)-1 was added to the lower chamber (R&D Systems). Plates were incubated immediately at 37δC in 5% CO₂ for 2 h. After incubation, the base of the membrane was rinsed with 200 µL of fresh medium twice, and cells that had passed into the lower chamber were counted using a FACS-calibur flow cytometer (BD Biosciences).

ELISA. ELISA for IL-8, IL-12, and TNF-α were performed according to manufacturer's instructions (Pharmingen). We also developed an ELISA method to determine antibodies to PfMIF in patient serum. Briefly, microtiter wells were coated with 15 µg/mL recombinant PfMIF overnight, and the wells were then blocked with 1% bovine serum albumin for 1 h, followed by the addition of patient serum diluted 1:100 in block buffer. Anti-PfMIF peptide IgG was detected using horseradish peroxidase-conjugated rabbit anti-human secondary antibody (Dako). The color was developed using o-phenylenediamine dihydrochloride substrate (Sigma), and optical density was read at 490 nm. Patients' serum samples were assayed in duplicate, and 7 serum samples from nonimmune European adults were used to control for nonspecific binding. Serum that showed binding 2 SDs above the average of the samples from the European control subjects was considered to be positive.

Statistical analysis. All data were analyzed using SPSS (version 12 [PC]; SPSS). Comparisons were done using Pearson's x² tests, Mann-Whitney U tests, and paired t tests.

Results

PfMIF expression in P. falciparum. PfMIF was initially described during the malaria genome project as a hypothetical MIF homologue. Subsequent microarray studies showed that PfMIF mRNA was transcribed during the ring and trophozoite stages of the P. falciparum life cycle [6]. We confirmed gene expression of PfMIF during the ring and trophozoite stages by Northern blot (figure 2A). In addition, we readily detected PfMIF protein expression during both the ring and trophozoites stages of the parasite asexual blood-stage cycle by Western blot (figure 2B). PfMIF was also detected in culture supernatant after schizont rupture (figure 2B). However, PfMIF was not detected in saponin lysis supernatant of iRBCs (containing RBC cytosol but not parasites) or in the culture supernatant before schizont rupture (data not shown). These results indicate that the protein might be released into the circulation after schizont rupture during blood-stage infection. We subsequently used indirect immunofluorescence microscopy to determine the localization of PfMIF in iRBCs during the intraerythrocytic development of the parasite (figure 3). In ring-stage parasites, PfMIF was almost exclusively located within the parasite. However, as the parasites developed into trophozoites, PfMIF was clearly detected in both the parasite and the iRBC cytosol. Here, PfMIF in iRBCs appeared to be associated with distinct vesicles and colocalized with PfSBP1, suggesting an association with the Maurer's cleft (figure 3). This showed that PfMIF moved from the parasite into the iRBC and that the protein was associated with distinct vesicles in the cytosol.

Figure 2.

Expression profile of Plasmodium falciparum macrophage-migration inhibitory factor (PfMIF). A, Northern blot of synchronized blood-stage parasite RNA blotted with PfMIF-specific probes. B, Western blots of synchronized blood-stage parasite lysates and culture supernatants taken from parasites after schizont rupture, blotted with anti-PfMIF pep-tide IgG.

Figure 3.

Localization of Plasmodium falciparum macrophage-migration inhibitory factor (PfMIF) in blood-stage parasites.Tightlysynchronizedparasites at different stages were stained with 4′-6-diamidino-2-phenylindole (DAPI) to identify infected red blood cells (iRBCs), with P. falciparum skeleton-binding protein 1 (PfSBP1) antiserum to indicate the iRBC cytosol, and with anti-PfMIF peptide IgG to determine their localization within iRBCs. The arrows in the bright-field micrographs indicate iRBCs.

Migration and activation of monocytes affected by PfMIF. We analyzed migration, surface-marker expression, and cytokine secretion in monocytes after treatment with different doses of recombinant PfMIF. We observed that random monocyte migration was significantly inhibited by treatment with 100 ng/mL PfMIF (figure 4). In addition, PfMIF significantly reduced the chemotactic migration of monocytes in response to MCP-1. However, MCP-1-induced migration was still above baseline level even in the presence of 500 ng/mL PfMIF (figure 4). The effect of PfMIF on random migration of monocytes was instable and was lost ∼10 days after purification. We therefore confirmed that PfMIF had maintained its effect on monocyte migration before each subsequent experiment.

Figure 4.

In vitro migration of monocytes in response to Plasmodium falciparum macrophage-migration inhibitory factor (PfMIF). Migration of monocytes across a membrane was assessed in response to recombinant PfMIF or lipopolysaccharide (LPS) and compared with that of untreated monocytes (medium) or monocytes migrating toward monocyte chemotactic protein (MCP)-1 (n≥3 for each condition). *P<.01, vs. the control (paired t test).

Monocytes treated with PfMIF did not release significantly different levels of IL-8, TNF-α, or IL-12 within 24 h, compared with that of controls (figure 5). There was an increase in IL-8 secretion with PfMIF treatment, but it was low in comparison with the IL-8 secretion induced by 100 pg/mL LPS, and the difference did not reach statistical significance (figure 5). Preincubation of monocytes with PfMIF had no significant effect on subsequent cytokine release in response to LPS.

Figure 5.

Effects of in vitro Plasmodium falciparum macrophage-migration inhibitory factor (PfMIF) treatment on cytokine release from monocytes. Culture supernatant from monocytes treated with recombinant PfMIF was assessed for interleukin (IL)-8, tumor necrosis factor (TNF)-α, and IL-12 release by ELISA (n≥5 for each condition). *P<.01, vs. the control (paired t test). LPS, lipopolysaccharide.

We subsequently analyzed whether the expression of the surface molecules HLA-DR, CD40, CD86, TLR2, TLR4, and ICAM-1 was altered on incubation with PfMIF. Only TLR2, TLR4, and CD86 surface expression showed a significant reduction in response to PfMIF, whereas the expression of all other surface markers remained unchanged (P<.05, paired t test) (figure 6A). In contrast, treatment of monocytes with the TLR4 ligand LPS resulted in significant up-regulation of CD40, CD86, and ICAM-1, whereas the expression of TLR4 was significantly reduced (P<.05, paired t test) (figure 6B). Preincubation of monocytes with PfMIF before treatment with LPS for 24 h did not change expression levels for any of the surface markers compared with LPS treatment alone, although TLR2 showed a trend toward reduced expression level (for 500 ng/mL PfMIF, P=.09, paired t test). Treatment of monocytes with the TLR2 ligand PGN increased the expression of CD86 only (P<.05, paired t test) (figure 6C). Preincubation with PfMIF before the addition of PGN for 24 h had no effect on the surface expression of any of the markers we analyzed compared with PGN alone. Taken together, these data indicate that PfMIF alters monocyte function but has no effect on LPS-or PGN-mediated activation of monocytes.

Figure 6.

Monocyte surface-marker expression after Plasmodium falciparum macrophage-migration inhibitory factor (PfMIF) treatment. Monocytes either were left untreated (white bars) or were treated with 100 (striped bars) or 500 (black bars) ng/mL PfMIF for 12 h before incubation in medium alone (A), lipopolysaccharide (LPS) (B), or peptidoglycan (PGN) (C) for another 24 h, followed by staining for the expression of surface molecules. The mean fluorescence intensity (MFI) for a given marker was divided by the MFI for the control (monocytes alone [white bars in panel A]), to normalize for variations in expression levels of a given marker among individuals (n≥3 for each condition). *P<.05, for PfMIF-treated monocytes vs. the control (paired t test); **P<.05, for LPS-or PGN-treated monocytes vs. the control (paired t test). ICAM, intercellular adhesion molecule; TLR, Toll-like receptor.

PfMIF antibody responses in patients. Antibody responses to PfMIF were examined in samples obtained from Kenyan children with acute malaria, from the same patients during convalescence, and from healthy Kenyan children during the low transmission season. There was a significant difference in median age between the subjects with acute malaria and the healthy Kenyan children, with the latter group being older than the former (for children with acute malaria, median of 29 months and range of 4–138 months; for convalescent children, median of 26 months and range of 6–70 months; for healthy children, median of 57 months and range of 12–107 months). However, within each group, there was no correlation between age and PfMIF IgG levels (Spearman's correlation coefficients: for children with acute malaria, r=0.076 [P=.547]; for convalescent children, r=.173 [P=.328 ]; for healthy children, r=0.16 [P=.221 ]). In the healthy control group, 62 children (53%) showed antibody responses to PfMIF. However, in this responder group, PfMIF antibodies were not associated with age and, therefore, exposure. In contrast, in the acute malaria and convalescent groups, 65 children (81%) and 35 children (100%), respectively, had antibodies against PfMIF. Therefore, there was a larger-than-expected proportion of positive anti-body responses in the acute and convalescent samples than in the control samples from healthy children (P<.001 , Pearson's x² test). When only responding children in each group were considered, the healthy children from whom samples were obtained during the low transmission season had levels of circulating PfMIF antibodies that were significantly lower than those seen during and immediately subsequent to P. falciparum infection (for children with acute malaria, median OD of 0.7215 and range of 0.1–2.10; for convalescent children, median OD of 1.0649 and range of 0.4–2.20; for healthy children, median OD of 0.1610 and range of 0.01–0.86; P<.001 , Mann-Whitney U test) (figure 7). This result is consistent with the antibody profiles seen in response to other malarial antigens, where a rapid decrease in antibody concentration has been observed after acute infection [17].

Figure 7.

Antibody responses to Plasmodium falciparum macrophage-migration inhibitory factor (PfMIF) in Kenyan patients. Samples were taken from children with acute malaria, from convalescent children, and from healthy control children and were assessed for anti-PfMIF IgG levels. The children with acute malaria and the convalescent children had significantly higher anti-PfMIF IgG levels than did the healthy control children. **P<.0005, vs. acute and convalescent (Mann-Whitney U test).

Discussion

We have shown here that PfMIF protein is expressed during the asexual blood stages of the P. falciparum life cycle and that the protein moved from the parasite into the iRBC cytosol and was associated with the Maurer's cleft. Recombinant PfMIF inhibited the random migration of monocytes and reduced the chemotactic response of monocytes to MCP-1. In addition, PfMIF altered the activation of monocytes, as demonstrated by surface-marker expression. Finally, in plasma from children with acute malaria, antibody responses to PfMIF were readily detected. However, antibody levels were much lower during the low transmission season, suggesting that PfMIF-specific anti-bodies are short lived and follow a similar pattern of responses to other parasite antigens previously studied.

The localization of PfMIF in the cytosol of iRBCs is interesting in view of recent findings that the Maurer's cleft plays an important role in the trafficking of parasite proteins, such as P. falciparum erythrocyte membrane protein 1 (PfEMP1), to the surface of the erythrocytes and that mammalian MIF has chaperone-like properties in vitro [18–20]. Maurer's clefts are parasite-derived vesicular structures that appear in the RBC cytosol during the early trophozoite stage. Cooke et al. [18] recently demonstrated that PfSBP1, which is associated with the Maurer's cleft, is responsible for the final translocation step of PfEMP1 to the iRBC plasma membrane. Recent studies examining the peptide-binding properties of mammalian MIF in vitro and its role in heat-induced protein aggregation have highlighted its potential role as a chaperone-like protein. In addition to its effect on monocyte function, PfMIF may play a role in protein trafficking within in the iRBC in light of its apparent localization to Maurer's clefts and the chaperone-like properties of MIF proteins from other species. Studies of protein trafficking in iRBC that use PfMIF-knockout parasites could address this question.

During the present study, we observed that TLR2 and TLR4 expression on monocytes was moderately but significantly reduced in response to PfMIF. MIF has previously been shown to be required for the expression of TLR4 in mouse macrophages [21]. However, this study addressed only endogenous MIF; the effect of the addition of exogenous MIF on TLR4 expression on macrophages, or any other cell type, has not been examined. It has long been recognized that TLR tolerance can be induced in that the treatment of cells with LPS renders them unresponsive to further stimulation not only through TLR4 but also through other TLRs [22]. We therefore tested whether preincubation with PfMIF for 12 h would alter monocyte activation by the TLR2 ligand PGN or the TLR4 ligand LPS. Although LPS alone increased the surface expression of CD40, CD86, and ICAM-1 but reduced the surface expression of TLR4, PGN alone increased the surface expression of CD86 only. Preincubation with PfMIF at either concentration had no effect on the TLR2-or TLR4-mediated activation of monocytes, suggesting that PfMIF does not play an active part in TLR tolerance, which has been previously described during Plasmodium yoelii infection in mice [23]. A reduction in TLR expression on myeloid cells has also been demonstrated in patients with filarial infections [24, 25]. T cells, B cells, and monocytes from individuals with filarial infection were shown to exhibit significantly less TLR1, TLR2, and TLR4 expression than those from uninfected controls. The mechanism for reduced expression has not been elucidated, but one hypothesis is that the MIF homologue produced and secreted by filarial nematodes may contribute to this phenomenon.

Analysis of the pathophysiological role of host-derived MIF during Plasmodium infection has to date been limited to its role in the development of malarial anemia [26]. This study indicated that MIF-knockout mice infected with P. chabaudi developed less-severe anemia, had better erythroid development, and demonstrated improved survival relative to controls. Following from this observation, it has been shown that huMIF levels in plasma were significantly increased during acute malarial disease in patients from Zambia [26]. This finding is in direct contrast to that of a study by Awandare et al. [27] that showed a significant decrease in circulating huMIF levels in patients with acute malaria from Gabon. Neither of these studies in patients addressed the likely involvement of PfMIF during acute malaria. Our study highlights the important issue that any future study should take into account: the presence of parasite MIF in circulation and, hence, any potential cross-reactivity between PfMIF and huMIF.

In summary, our study suggests that PfMIF is capable of influencing the host immune system. This is most clearly demonstrated by the ability of recombinant PfMIF to modulate the function of monocytes, specifically by inhibiting migration and decreasing CD86, TLR2, and TLR4 expression in vitro. The localization of PfMIF within the iRBC indicates that the protein is likely to be released on schizont rupture, allowing direct interactions with the host immune system.

Acknowledgments

We thank the children and their parents or guardians, for participating in the study. We also greatly appreciate the help of clinical teams at the Kenya Medical Research Institute ward and outpatient clinic, as well as the field-workers of the malaria study. We would also like to thank Prof. Catherine Braun-Breton, for the gift of Plasmodium falciparum skeleton-binding protein 1 antiserum, and Dr. David Muhia, for his technical assistance.

Potential conflicts of interest: none reported.

Presented in part: Molecular Approaches to Malaria, Lorne, Australia, 1–5 February 2004.

Financial support: This study was supported in part by the European Union BioMalPar Network of Excellence. B.C.U. holds a Wellcome Trust Senior Research Fellowship (079082), and T.N.W. holds a Wellcome Trust Senior Clinical Fellowship (076934/2/05/2). S.K. and K.M. are funded by the Wellcome Trust. U.K. acknowledges support from the European Commission, the Alexander von Humboldt Foundation, and Brunel University.

This study is published with the permission of the director of the Kenya Medical Research Institute.