On the Cytoadhesion of Plasmodium vivax–Infected Erythrocytes

Abstract

It has long been recognized that the directly attributable morbidity and mortality differ for the 2 most prevalent malarial parasite species, Plasmodium falciparum and Plasmodium vivax. The higher multiplicative potential of P. falciparum–infected erythrocytes (Pf-iEs) no doubt contributes to their increased virulence. However, it is the withdrawal of mature Pf-iEs (parasites older than 24 h) from the peripheral circulation to that of the internal organs—a phenomenon known as sequestration [1]—that is considered to be the key pathogenic event. P. falciparum is characterized by almost total sequestration, such that few if any mature Pf-iE forms are observed in peripheral blood samples during infection. Given that mature P. vivax–infected erythrocytes (Pv-iEs) are frequently observed in peripheral blood samples, it was concluded that sequestration did not occur with this parasite. As a consequence, the paradigm was formulated that sequestration of Pf-iEs in specific organs is the principal initial cause of pathology and that, when sequestration occurs in the brain or placenta, the likelihood of cerebral malaria and pregnancy-associated malaria increases. Indeed, most forms of severe malaria and nearly all mortality have been almost exclusively recorded for falciparum cases.

P. vivax, the most prevalent malarial species outside sub-Saharan Africa, imposes a substantial global public health burden [2], with recent estimates of 130–435 million infected persons per year among the 2.6 billion people at risk. Of equal importance was the observation that many types of severe malaria—long considered to be specific to P. falciparum—also commonly occur in P. vivax–infected persons. For instance, infection with P. vivax during pregnancy was found to be associated with a substantial reduction in birth weight [3]. Furthermore, in some areas of endemicity progression of vivax malaria to clinically severe forms, including cerebral malaria and acute respiratory distress syndrome, was found to occur as frequently as for falciparum infections, with similar levels of fatality [4–7]. This raises the possibility that pathological processes linked to cytoadhesion might also operate in P. vivax.

In P. falciparum infection, cytoadhesion of Pf-iEs to endothelial cells is mediated by interactions between members of the P. falciparum erythrocyte membrane protein 1 (PfEMP-1) family, polymorphic proteins encoded by the var multigene family [8, 9], and defined host receptors on endothelial cells. Of the 10 or so receptors identified to date [10, 11], 3 have been extensively investigated: CD36, intercellular adhesion molecule 1 (ICAM-1), and chondroitin sulfate A (CSA). The last has been specifically associated with the binding of Pf-iEs to the placenta [12].

We wished to establish whether P. vivax parasites are able to cytoadhere under static or flow conditions to endothelial cells and placental cryosections and whether the receptors for PfiEs were also implicated. Furthermore, we assessed the involvement of VIR proteins as a potential Pv-iE ligand.

Methods

Ethical approval. Informed consent was sought and granted from all patients attending the Tropical Medicine Foundation of Amazonas (FMT-AM), Amazonas, in northern Brazil. The procedures were approved by the Ethics Committee Board of the FMT-AM (process 2758/2008-FMT-AM; approval no. 1943).

Parasite isolation and enrichment. Once microscopic diagnosis of uncomplicated vivax or falciparum malaria was made and before the treatment was initiated, 5–10 mL of blood were collected into citrate-coated Vacutainer tubes (BD). Parasitemia levels rarely exceeded 5000 parasites/µL of blood. The blood was immediately processed to obtain enriched Pv-iEs. On average, a total of ∼1 × 10⁶ Pv-iEs could be obtained, allowing only a limited number of cytoadhesion assays to be conducted. Patients who had received antimalarial treatment 3–4 weeks before the test were excluded. Immediately after collection, the red blood cells containing trophozoites and schizonts were separated from the younger forms on a 45% Percoll (Amersham) gradient, as described elsewhere [13] with minor modifications. Briefly, after plasma separation by centrifugation, blood pellets were washed 3 times and then resuspended in RPMI 1640 medium (Sigma) to a final hematocrit of 10%. Five milliliters of this suspension was overlaid on a 5-mL 45% Percoll solution (2.25 mL of Percoll, 0.5 mL of RPMI 1640 [×10], and 2.25 mL of distilled water) in a 15-mL tube. After centrifugation, floating mature iEs were collected and resuspended in RPMI 1640. Ex vivo, Pf-iEs were enriched in Percoll gradient at 60%, as described elsewhere [14]. Giemsa-stained thick smears (before enrichment) and thin smears (after enrichment) were examined to determine the Plasmodium species and the percentage of mature stages, respectively. A sensitive nested polymerase chain reaction (PCR) assay was applied to the samples to confirm the diagnosis [15].

P. vivax slide preparation and immunofluorescence assays. Immunofluorescence assays (IFAs) were performed in 8-well slides containing mature stages of P. vivax. Immediately after enrichment on Percoll, Pv-iEs were washed and resuspended in 10% fetal calf serum (FCS) (Nutricell) and then deposited on IFA slides (50 µL/well), fixed in acetone for 10 min, airdried, and stored at −20°C until use. Ten micrograms per milliliter of each monoclonal antibody (3F8 or K23) against P. vivax merozoite surface protein 1 (PvMSP1) [16] was diluted in phosphate-buffered saline (PBS) and applied to slides for 30 min at 37C. After washing with PBS, slides were incubated with 10 µg/mL fluorescein isothiocyanate–conjugated sheep anti–mouse immunoglobulin G (Sigma) and 100 µg/mL 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI) (Molecular Probes) for 30 min at 37°C and then washed several times. Positive monoclonal antibody recognition was detected with the aid of an immunofluorescence microscope (Nikon). For IFAs conducted using VIR antisera, Pv-iEs were left in suspension (liquid phase). Briefly, parasites were fixed in 2% paraformaldehyde and washed and diluted in FCS, and then mouse VIR-A4 and VIR-E5 antisera were added at a final dilution of 1:20 and incubated for 60 min. After 2 washes with FCS, cells were incubated with 100 µg/mL Alexa Fluor 488–conjugated goat anti–mouse IgG (Molecular Probes) and 500 µg/mL DAPI for 30 min at 37°C and then washed twice with FCS. In both assays, positive recognition by monoclonal antibody, VIR-A4, VIR-E5, or MSP-1₁₉ antisera was detected with the aid of an immunofluorescence microscope.

Polyclonal VIR antisera were generated by injecting mice with VIR-A4 or VIR-E5 glutathione S-transferase (GST) fusion proteins belonging to the A or E vir subfamilies, respectively [17, 18] and raised after 2 immunizations (21-day interval) with each recombinant protein (5 µg/animal/dose) emulsified in complete or incomplete Freund adjuvant. No Pv-iE–positive labeling was visualized by means of GST antisera.

Selection of monophenotypic cultured P. falciparum parasites and cells. The P. falciparum lines FCR3 [19] and S20 [20] were cultured in candle jars. Briefly, Pf-iEs were cultivated in fresh type O⁺ human erythrocytes (Blood Center, Universidade Estadual de Campinas) suspended at a final hematocrit of 4% in complete medium (RPMI 1640 [pH 7.2]) (Sigma) and supplemented with 10% homologous human plasma.

The following cell lines were used in this study: Saimiri brain endothelial cells (SBECs) and human lung endothelial cells (HLECs) [21, 22], adapted from cultured primary explants, and CHO-ICAM, CHO-CD36 and CHO-K1 cells [23]. Selection of FCR3 parasites to CSA (FCR3^CSA) was performed by panning (5 rounds) of mature-stage iEs on endothelial cells [21, 24] in the presence of soluble CSA (100 µg/mL; Sigma).

Mature S20 trophozoites were selected on CHO-ICAM cells (5 rounds) and then by a further 2 rounds on plastic plates coated with recombinant ICAM-1 [25]. The selected S20^ICAM parasite line bound strongly to CHO-ICAM cells but poorly to nontransfected CHO-K1 cells (data not shown).

Static cytoadhesion assays. We assessed the ability of Pv-iEs or Pf-iEs obtained from infected patients to adhere to placenta cryosections, HLECs, and SBECs by performing static cytoadhesion assays as described elsewhere [19, 21, 26] with minor modifications. Briefly, HLECs or SBECs were grown to confluence on 8-well culture slides (0.69 cm²/well; BD), and 5 × 10⁴ Percoll-enriched iEs were then added to each well in a total volume of 200 µL of cytoadhesion medium (RPMI 1640 [pH 6.8]), either alone or in the presence of 100 µL of soluble CSA (100 µg/mL) or anti-CD36 (5 µg/mL). Culture slides were incubated for 1 h at 37°C and then extensively washed in cytoadhesion medium. To confirm that the ligands on the surface of Pv-iEs were proteins, Pv-iEs were treated with trypsin (1 µg/mL) for 45 min at 37°C before incubation over endothelial cells. For confirmation of CSA as a receptor, HLECs were previously incubated with chondroitinase ABC (CaseABC) (0.5 U/ mL; Sigma) for 2 h at 37°C. Involvement of VIR antigens in the cytoadhesion of Pv-iEs to HLECs was evaluated in an assay where inhibition of parasite binding was tested with antisera (diluted at 1:5 or 1:10) to VIR-A4 or VIR-E5 GST-fused proteins. Serum samples from mice immunized solely with GST in complete or incomplete Freund adjuvant were used as specificity controls. For these assays, inhibition was determined as a percentage of the negative control and was expressed as the mean value for 3 wells ± standard deviation.

Assays of adhesion to placental trophoblasts were performed as described elsewhere [19, 26, 27] with minor modifications. Placental biopsy samples from 3 human immunodeficiency virus–negative Brazilian women with malaria were collected immediately after delivery, snap-frozen in liquid nitrogen/n-hexane (Merck), and then stored frozen in Tissue-Teck (Thermo) before use. Serial placenta cryosections (5–7 µm) were cut with a cryostome and mounted on individual glass slides. Cryosections ∼1cm² was delimited with a Dako Pen device. Assays of the adhesion of Pv-iEs to placental cryosections were performed as for endothelial cells. Three to 4 placenta cryosections were used for each adhesion assay. After a 1-h incubation at 37°C and following Giemsa staining, the number of infected erythrocytes that adhered to the endothelial cell monolayer or to placenta cryosections was counted under the microscope.

Involvement of CD36 or ICAM-1 host receptors was verified by allowing Pv-iEs to adhere under static conditions to CHOCD36 cells, CHO-ICAM cells, or CHO-745 cells, a cell line that does not express either of these receptors [23]. In these experiments, we performed adhesion assays without using human serum in the medium because it has been shown that human immunoglobulins present in normal serum can mediate binding of Pf-iEs to CHO-745 cells [28].

Flow-based cytoadhesion assays. To assess the resistance of Pv-iEs to shear stress, we performed flow-based cytoadhesion assays according to a modified version of a method that has been described elsewhere [19, 26, 27]. Briefly, HLECs were cultured to confluence in single-well culture microslides (8.6 cm², corresponding to 12.5 times the area of each well in a 8well culture slide), to which 1.5 mL of 5 × 10⁵ Pv-iEs enriched on 45% Percoll gradient was added. After a 1-h incubation at 37°C, microslides were mounted in a flow chamber system (Immunetics), and cytoadhesion medium (RPMI 1640 [pH 6.8]) was flowed through at a wall shear stress of 0.09, 0.36, or 1.44 Pa for 10, 5, and 2.5 min, respectively. After this, the remaining bound Pv-iEs were counted in 20 randomly selected fields through a digital camera attached to a microscope (Moticam 2500; Motic). Results were expressed as the mean number of infected erythrocytes per square millimeter ± standard deviation. In some experiments, microslides were stained with Giemsa after a determined shear stress condition, to visualize bound Pv-iEs. Alternatively, HLECs were incubated with lipopolysaccharide (LPS) (1.0 µg/mL; Sigma) for 4 h at 37°C and then washed before parasite cytoadhesion. Cultured P. falciparum panned isolates (FCR3^CSA and S20^ICAM) wereusedas controls. To ascertain whether P. vivax adhesion to HLECs occurred under flow conditions, 1 × 10⁶ enriched Pv-iEs diluted in cytoadhesion medium were flowed at a wall shear stress of 0.09 or 0.36 Pa for 1 h and then filmed. For both the static and flow-based assay, parasite adhesion was normalized to 1 × 10³ infected erythrocytes/mm², taking into account the area and the number of infected erythrocytes used in each assay (5 × 10⁴ infected erythrocytes in 0.69 cm² or 5 × 10⁵ infected erythrocytes in 8.6 cm²).

Flow cytometry. Levels of ICAM-1 expression on HLECs were assessed by flow cytometry. Briefly, 1 × 10⁶ cells either treated with LPS (1.0 µg/mL) or left untreated were harvested, incubated with anti–human CD54 (ICAM-1) (phycoerythrin; BD Bioscience) for 30 min at 4°C, washed, and fixed in 2% formaldehyde. Analysis was performed using a FACScanto flow cytometer (BD), and the mean fluorescence intensity and the percentage of positive cells were analyzed with the aid of FCS Express software (version 3.00.0320; De Novo Software). For each sample, a minimum of 100,000 events were acquired.

Statistical analysis. The statistical significance of adhesion to different cells types at various conditions was determined using the Mann-Whitney U test or the Kruskal-Wallis test. Calculations were performed using BioEstat software (version 3.0; CNPq, Brazil) and Prism software (version 3.02; GraphPad Software). Differences were considered significant at P < .05.

Results

Our investigation of Pf-iE cytoadherence was substantially facilitated by the availability of in vitro–cultured P. falciparum, an avenue not open for P. vivax. We conducted our experiments with Pv-iEs using samples obtained directly from patients presenting with uncomplicated P. vivax malaria in Manaus, Brazil. The limitations imposed by reliance on clinical samples with parasites of diverse genotypes and of varying suitability for cytoadherence assays were compensated for by the collection of a relatively large number of samples ( n = 34). As controls for the cytoadhesion assays, P. falciparum–infected blood was also obtained from 4 patients attending the same hospital. In all cases, a sensitive PCR assay was used to exclude the presence of mixed-species infections. As an inital step, cytoadherence was assessed using HLECs, SBECs, and human placental cryosections. These cell lines and placental cryosections have been characterized and validated for assays of P. falciparum cytoadhesion [19, 21, 22, 26, 27]. In P. falciparum infection, mature forms of the parasite cytoadhere more strongly than do freshly invaded forms. We hypothesized that this could be also the case for P. vivax infection. Given the different densities of mature-stage and early-stage parasites, we subjected P. vivax parasites collected from the patients directly and without prior short-term culturing to a Percoll gradient, enabling the recovery of almost-pure trophozoite-and schizont-stage Pv-iEs (Figure 1). Of note, after enrichment the percentage of MSP-1₁₉–expressing forms observed in IFA varied from 85% to 97%.

Figure 1

Percoll gradient enrichment of Plasmodium vivax maturing forms. A, Giemsa staining of the mature trophozoites and young schizonts obtained after Percoll gradient enrichment, yielding cell suspensions with 85%–97% P. vivax–infected erythrocytes. B, Parasite species and maturity confirmed by immunofluorescence assay using anti–PvMSP-1₁₉ conformational monoclonal antibodies (3F8 and K23) diluted at 1:50. Normal mouse serum samples were used as negative controls. DAPI, 4′,6-diamidino-2-phenylindole dihydrochloride; FITC, fluorescein isothiocyanate.

Open in new tabDownload slide

Pv-iEs were first tested for their ability to cytoadhere to HLECs, SBECs, and placental cryosections under static conditions. For some patients, aliquots of the same Pv-iE suspension were tested after digestion with trypsin or in the presence of soluble CSA, anti-CD36 monoclonal antibody in the cytoadhesion medium, or HLECs pretreated with CaseABC (Figure 2 and Table 1). All the Pv-iEs tested displayed some level of cytoadherence to HLECs and/or SBECs or to placental cryosections. Pv-iE binding to areas not containing endothelial cells was not observed. Pretreatment of Pv-iEs with trypsin generally decreased adhesion to HLECs or SBECs by 19%–85% (mean, 54%), depending on the parasite isolate. Significant inhibition of cytoadherence was also observed in the presence of soluble CSA (range, 16%–81%; mean, 42%). By contrast, when assays were conducted in the presence of anti-CD36, the extent of inhibition of cytoadherence was not statistically significant (Table 1), although only a few isolates could be tested. Moreover, unlike the higher inhibition observed with soluble CSA, HLECs pretreated with CaseABC did not significantly abolish adhesion of the same Pv-iEs (Table 1). Cytoadhesion assays were also conducted in parallel with similarly enriched mature Pf-iEs derived from 4 patients. The levels of cytoadherence observed were ∼10-fold higher than those recorded for Pv-iEs (Table 1).

Figure 2

Adhesion of Plasmodium vivax–infected erythrocytes (Pv-iEs) to lung and brain endothelial cells or placental cryosections. Shown are representative photomicrographs of (left) and cytoadhesion assays for (right) Pv-iEs binding to human lung endothelial cells (HLECs) (A), Saimiri brain endothelial cells (SBECs) (B), and placental cryosections (C) stained with Giemsa stain and visualized using a Nikon microscope at ×100 magnification. Arrows indicate Pv-iEs bound to host cells. Representative results for Pv-iEs obtained from different patients and used for cytoadhesion assays conducted on untreated cells, on Pv-iEs pretreated with trypsin, or in the presence of soluble chondroitin sulfate A (CSA) or anti-CD36 antibody are shown. In all cases, parasites (5 × 10⁴) were left for 1 h at 37°C in 8-well culture slides (0.69 cm²/well) and then extensively washed, and adhered Pv-iEs were counted. Data are the mean number of bound Pv-iEs per square millimeter, normalized to an input of 1 × 10³ Pv-iEs/mm²; error bars indicate standard deviations. * .05> P > .01 and ** 0.01 > P > .001 (Kruskal-Wallis test).

Open in new tabDownload slide

Table 1

Cytoadhesion Assays for Plasmodium vivax–Infected Erythrocytes (Pv-iEs) and Plasmodium falciparum–Infected Erythrocytes (Pf-iEs) Collected from 24 Patients

Open in new tabDownload slide

Having demonstrated that Pv-iEs could cytoadhere to endothelial cells under static conditions, it was important to evaluate whether the observed cytoadherence was biologically relevant and whether it could be maintained under the flow conditions that parasites are subjected to in the bloodstream. Shear stress in postcapillary venules is close to 0.08 Pa [29]. Thus, we enriched Pv-iEs from 7 individuals and conducted flow-based cytoadhesion assays on HLECs at shear stress conditions varying from 0.09 to 1.44 Pa. The parasites that remained attached after being subjected to increasing flow rates for defined periods were enumerated by microscopic examination after Giemsa staining (Figure 3 and Table 2). Given that HLECs express CSA and ICAM-1 on their surface, we performed parallel assays using P. falciparum FCR3 and the S20 iEs preselected on CSA (FCR3^CSA) and ICAM-1 (S20^ICAM). As for Pf-iEs (FCR3^CSA), even at a relatively high shear stress (1.44 Pa) a substantial proportion (30%) of the cytoadherent Pv-iEs could not be detached (Table 2). Furthermore, stimulation of the HLECs with LPS significantly strengthened the cytoadherence, because 56% of the cytoadherent Pv-iEs could not be detached at 1.44 Pa (Figure 3). Of note, LPS treatment augmented ICAM-1 expression levels by 3-fold (data not shown). The strength of Pv-iE adhesion was similar to that observed for P. falciparum FCR3^CSA to unstimulated HLECs (Table 2). The behavior of the cytoadherent Pv-iEs to HLECs under flow conditions was recorded in real time and can be seen in Videos 1 and 2, which are available in the online version of the Journal and which show that cytoadherent Pv-iEs display the rolling characteristics of cytoadherent P. falciparum parasites [30]. Video 1 shows the rolling and binding of a Pv-iE to an HLEC; parasites were diluted in medium and flowed at a wall shear stress of 0.09 Pa. Video 2 shows 2 Pv-iEs bound to HLECs under a flow condition of 0.36 Pa.

Figure 3

Resistance of Plasmodium vivax–infected erythrocyte (Pv-iE) cytoadhesion to flow conditions. Enriched Pv-iEs (5 × 10⁵) were allowed to adhere to human lung endothelial cells (HLECs) during a 1-h incubation at 37°C in single-well microslides (8.6 cm²) mounted in a flow chamber system, and cytoadhesion medium (RPMI 1640 [pH 6.8]) was flowed through at a wall shear stress of 0.09, 0.36, and 1.44 Pa for 10, 5, and 2.5 min, respectively; medium was not flowed through the control chamber. The Pv-iEs that still bound at the end of the flow period were counted in 20 randomly selected fields, and the percentage of binding compared with that in the control chamber were calculated. Data are mean binding percentages for 2 patient isolates (104 and 106) on HLECs previously treated with lipopolysaccharide (LPS) (1.0 µg/mL) or left untreated; error bars indicate standard deviations. * P < .05 for the comparison of adhesion to nonstimulated vs LPS-stimulated HLECs (Mann-Whitney U test).

Open in new tabDownload slide

Table 2

Cytoadhesion Assays for Plasmodium vivax–Infected Erythrocytes (Pv-iEs) Obtained from 7 Patients or Cultured Plasmodium falciparum–Infected Erythrocytes (Pf-iEs) to Human Lung Endothelial Cells (HLECs) under Flow Conditions

Open in new tabDownload slide

We then tested whether CD36 or ICAM-1, both present on the HLEC surface [21, 22], were implicated in the observed Pv-iE cytoadherence. For this, CHO lines transfected with CD36 or ICAM-1 were used. None of the Pv-iEs from the 3 samples tested showed binding to CD36 (Figure 4), because the levels of cytoadherent Pv-iEs were not significantly different from background (binding of Pv-iEs to untransfected CHO-745 cells). By contrast, a 2.1–2.7-fold increase in cytoadhesion to ICAM-1 over background was observed for Pv-iEs from 2 of the 3 samples tested (Figure 4). Low cytoadhesion of Pv-iEs isolated from patient 098 may indicate variations in the binding phenotypes of different isolates.

Figure 4

Adhesion of Plasmodium vivax–infected erythrocytes (Pv-iEs) to specific receptors. Pv-iEs (5 × 10⁴ cells/well) were allowed to adhere to CHO–intercellular adhesion molecule (ICAM), CHO-CD36, or control CHO-745 cells. Data are the mean number of bound Pv-iEs per square millimeter, normalized to an input of 1 × 10³ Pv-iEs/mm²; error bars indicate standard deviations. * P < .001 for the comparison of adhesion to CHO-ICAM cells vs CHO-CD36 or CHO-745 cells (Kruskal-Wallis test).

Open in new tabDownload slide

Finally, we investigated potential parasite ligands involved in Pv-iE cytoadhesion. A multigene family orthologous to the P. falciparum var genes does not occur in the genome of P. vivax [31]. However, a superfamily of surface-expressed variant antigens genes (vir) is present in P. vivax; it has ∼350 members, which can be subdivided into 10 subfamilies and unclustered members [17, 31, 32]. Hence, we tested 2 specific polyclonal antisera, VIR-A4 and VIR-E5, from the A and E subfamilies, respectively [18]. Pv-iEs were specifically recognized by both antisera in IFAs (Figure 5A) but not by the control GST antisera. Significant inhibition of Pv-iE cytoadherence on HLECs was observed when either VIR antisera was included in the assay medium but not in the presence of the GST antisera (Figure 5B).

Figure 5

Recognition and blocking of Plasmodium vivax–infected erythrocyte (Pv-iE) cytoadhesion to human lung endothelial cells (HLECs) by specific VIR antisera. A, Immunofluorescence of a Pv-iE labeled with antisera against VIR-A4 (1:20) or VIR-E5 (1:20). Shown are results for phase contrast, 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI) staining, and mouse anti–VIR-E5 (top) or anti–VIR-A4 (bottom) with Alexa Fluor 488–conjugated goat anti–mouse immunoglobulin G. B, Results of incubation of Pv-iEs ( 5 × 10⁴ cells/well) for 1 h at 37°C alone or in the presence of VIR-A4 or VIR-E5 antisera diluted at 1:5 (isolate 095) or 1: 10 (isolate 096). In both assays, serum samples from mice immunized solely with glutathione S-transferase (GST) in Freund adjuvant were used as a negative control. For these assays, inhibition was determined as the percentage of negative control counts, expressed as mean values for triplicate wells; error bars indicate standard deviations. * P < .05 for the comparison of inhibition with GST antisera (Kruskal-Wallis test).

Open in new tabDownload slide

Discussion

Our observations provide the first evidence, to our knowledge, that mature Pv-iEs are capable of cytoadhering to endothelial cells and placental cryosections. Two receptors used by P. falciparum for binding to endothelial cells, ICAM-1 and CSA, were also implicated in the cytoadhesion of P. vivax parasites, at least to some extent. However, CaseABC treatment suggested that cytoadhesion to CSA, a highly negative charged molecule, might be due to charge interaction. Indeed, it has been shown that binding to CSA is highly dependent on ionic strength [33]. The fraction of Pv-iEs that cytoadhered was up to 10-fold lower than that for Pf-iEs. Importantly, cytoadhesion of Pv-iEs, once established, is as strong as that of CSA-selected Pf-iEs, as demonstrated by the flow assays (Table 2). Because the parasites we assayed were directly obtained from patients, our data suggest that only a minor fraction of P. vivax would have the capacity to cytoadhere in vivo. Partial retention of P. vivax from the peripheral circulation would explain why the mature forms of this parasite are generally found in peripheral blood. This is consistent with the observation of partial and differential accumulation of Pv-iEs in organs of P. vivax–infected squirrel monkeys [34], as well as with observations of partial depletion of mature P. vivax in the peripheral blood of humans [35, 36]. Our observation of Pv-iE cytoadhesion to placental tissue from sections of the placenta obtained after delivery (data not shown) suggests that the phenomenon can occur in vivo. However, the histopathological studies of postmortem tissues that would be required to indicate whether this phenomenon extends to other organs in infected individuals are limited.

For all isolates for which trypsin treatment was tested, the only partially abrogated Pv-iE cytoadhesion suggests that trypsin-resistant ligands commonly occur on the Pv-iE surface. It would be important to establish whether other receptors are implicated in Pv-iE cytoadhesion and explore the precise roles played by ICAM-1 and CSA. Finally, detailed investigations of the role played by VIR proteins as cytoadhesive ligands, for which we present indirect evidence, are complicated because (1) multiple members of the VIR protein family are expressed on the infected red blood cell surface [17], unlike the clonally expressed var genes of P. falciparum [37], and (2) the high number of vir genes present in the genome (346 vir genes for the Sal line of P. vivax [31], as opposed to 59 var genes for the 3D7 line of P. falciparum) [38]. Nevertheless, our data call for further investigation of the role played by VIR proteins in PviE cytoadhesion.

Although infections with P. vivax are less life-threatening than those with P. falciparum, morbidity in P. vivax infection is associated with anemia and a pronounced cytokine-mediated inflammatory response [4]. Differential accumulation of a proportion of parasites to some organs, such as the lungs or placenta, might be targeted by the inflammatory response to this organ, leading to a more severe clinical presentation [4, 39, 40].

In conclusion, our observations add a new aspect to the pathophysiology of a major (yet mostly neglected) human pathogen, which could lead to novel therapeutic approaches to alleviate the increasingly recognized health burden globally imposed by this distinctly not-so-benign parasite.

Acknowledgments

We especially thank Dr Jürg Gysin (Institut Pasteur, Guadeloupe, France) and Dr Artur Scherf (Institut Pasteur, Paris, France) for providing endothelial cells, P. falciparum FCR3 parasites, and CHO-CD36 cells. We thank Dr Cristina Vicente and Dr Claudio C. Werneck for providing the megapixel digital camera and for discussion. We are grateful to Anne Charlotte Grüner for her critical comments. L.R. and G.S. are currently part of an official collaboration between the Singapore Immunology Network, Agency for Science, Technology, and Research, and the Institut National de la Santé et de la Recherche Médicale (Laboratoire International Associé). F.T.M.C. dedicates this study to Zeci T. M. Costa (in memoriam).

References

Author notes