Effects of sevuparin on rosette formation and cytoadherence of Plasmodium falciparum infected erythrocytes

In severe falciparum malaria cytoadherence of parasitised red blood cells (PRBCs) to vascular endothelium (causing sequestration) and to uninfected red cells (causing rosette formation) contribute to microcirculatory flow obstruction in vital organs. Heparin can reverse the underlying ligand-receptor interactions, but may increase the bleeding risks. As a heparin-derived polysaccharide, sevuparin has been designed to retain anti-adhesive properties, while the antithrombin-binding domains have been eliminated, substantially diminishing its anticoagulant activity. Sevuparin has been evaluated recently in patients with uncomplicated falciparum malaria, and is currently investigated in a clinical trial for sickle cell disease. The effects of sevuparin on rosette formation and cytoadherence of Plasmodium falciparum isolates from Thailand were investigated. Trophozoite stages of P. falciparum-infected RBCs (Pf-iRBCs) were cultured from 49 patients with malaria. Pf-iRBCs were treated with sevuparin at 37°C and assessed in rosetting and in cytoadhesion assays with human dermal microvascular endothelial cells (HDMECs) under static and flow conditions. The proportion of Pf-iRBCs forming rosettes ranged from 6.5% to 26.0% (median = 12.2%). Rosetting was dose dependently disrupted by sevuparin (50% disruption by 250 μg/mL). Overall 57% of P. falciparum isolates bound to HDMECs under static conditions; median (interquartile range) Pf-iRBC binding was 8.5 (3.0–38.0) Pf-iRBCs/1000 HDMECs. Sevuparin in concentrations ≥ 100 μg/mL inhibited cytoadherence. Sevuparin disrupts P. falciparum rosette formation in a dose dependent manner and inhibits cytoadherence to endothelial cells. The data support assessment of sevuparin as an adjunctive treatment to the standard therapy in severe falciparum malaria.

Introduction 50 samples were cultivated successfully. The three isolates which could not be assessed had very low parasitaemia, grew slowly and developed gametocytes. Of 50 samples 47 (97%) developed to trophozoites and rosetting assays were performed and 49 (98%) samples from cryopreserved parasites grew and static binding assays were performed.

Discussion
The propensity of P. falciparum-infected red blood cells (Pf-iRBCs) to adhere to vascular endothelium and other RBCs (rosette formation) is thought to be a major factor accounting for the virulence of this parasite [42]. Both of these processes contribute to microvascular obstruction. The potentially lethal syndrome of cerebral malaria is specifically associated with cerebral sequestration [3,43] and with increased rosetting [44], rosetting is not always associated with cerebral malaria [45,46]. In vivo studies of parasite development indicate that the prognosis of severe malaria is related to the sequestered malaria parasite biomass [47]. Severely ill patients require parenteral antimalarial therapy, and any adjunctive treatment also needs to be given parenterally. From previous studies, sevuparin prevents Pf-iRBC adhesion and so has potential as an adjunct treatment of severe malaria [40]. Sevuparin both blocked and reversed sequestration of Pf-iRBC in vivo in rats and monkeys [38]. Sevuparin also inhibited merozoite invasion in vitro [38].
We investigated the effect of sevuparin on rosette formation and cytoadherence of P. falciparum isolates from uncomplicated malaria patients. In the current study, sevuparin concentrations ranged from 62.5-1000 μg/mL in the rosetting experiments and from 50-800 μg/mL in the HDMEC cytoadherence experiments, which were based on in vitro concentration ranges published in the literature [17,24,40,48,49]. Studies in healthy volunteers show that sevuparin in an i.v. dose of 5.3 mg/kg results in mean (SD) maximal plasma concentrations (Cmax) of 116 (19.3) μg/mL. Dosing in clinical studies in patients with falciparum malaria has ranged from 1.5 to 6 mg/kg per dose (unpublished). The concentration range in our in-vitro experiments range from 50-1000 μg/mL and e.g. sevuparin concentrations of 100 μg/mL showed clear inhibition of cytoadhesion, which is within the concentration range obtainable in patients (Leitgeb et al, personal communication). Sevuparin significantly disrupted the number of rosettes formed in a dose dependent manner. Rosette and cytoadherence disruption by heparin, heparan sulfate (HS) and sevuparin, and the binding of these glycosaminoglycans (GAGs) to the DBL1α domain of PfEMP-1 are well described [16-18, 24, 38, 48]. Some of these sulphated polysaccharides not only disrupt rosetting, but also inhibit adhesion of Pf-iRBCs to the endothelial and placental receptors such as chondroitin sulphate A (CSA) and CD36 [50,51]. However, the anti-coagulant effects of heparin are an important drawback for its clinical use in severe malaria, since this is always accompanied by significant thrombocytopenia, and has been associated with an increased bleeding risk [36,37,52].
CD36 binding is a property of almost all P. falciparum isolates obtained freshly from malaria patients. In the current in vitro study, 43% of fresh parasite isolates did not cytoadhere to our HDMEC cell culture expressing CD36 and the level of binding was low, varying between 1-181 infected red blood cells/1000 HDMECs. This is comparable with Udomsangpetch et al, 1996 (over 60% of isolates did not adhere to CD36) [54]. Part of the explanation could be that some VAR genes code for PfEMP1 variants that do not bind to CD36, estimated as 10 out of the 60 different P. falciparum VAR genes [21]. Studies in Africa have found no difference in CD36-binding ability between parasite isolates from severe and uncomplicated malaria patients [45,55]. More than 90% of clinical parasite isolates tested adhere to CD36, whereas, 10% adhere to ICAM-1. For isolates that adhere to both molecules, the degree of adhesion to CD36 is at least 10-fold higher than adhesion to ICAM-1 [54]. In Thailand, two small studies showed a significant positive correlation between CD36-binding and presentation with severe malaria [56,57]. A single human genetics study on CD36 polymorphisms in South East Asia (SEA) reported that CD36 deficiency was protective against cerebral malaria [58]. Not all studies support the importance of CD36 in mediating severe disease; a study by Rogerson et al showed preferential binding to ICAM-1 rather than CD36 in parasite isolates from paediatric severe malaria [55].
Binding of PfEMP-1 to endothelium is promiscuous and mediated by a range of membrane receptors, including Endothelial Protein-C Receptor (EPCR), thrombospondin, intercellular adhesion molecule-1 (ICAM-1), heparan sulphate (HS) and CD36. There is a wide range of literature on the importance of different receptors in mediating severe disease [21,59]. More recently, binding of the N-terminal cysteine-rich inter domain region (CIDRα1) of PfEMP-1 to Endothelial Protein-C Receptor (EPCR) has been associated with severe malaria [13]. So, although the central role of CD36 in parasite cytoadherence is not disputed, additional receptors might play an important role in parasite sequestration in the microcirculation. Detailed studies of PfEMP-1 binding have revealed that the level of binding to its receptors is dependent on PfEMP-1 domain architectures, in particular of the Duffy Binding-like (DBL) and CIDR domains (reviewed in Smith et al.). Most PfEMP-1 (84%) will bind to CD36, and mainly carry the DBLα0-CIDRα2-6 domains. How heparin sulfate-like substance including sevuparin can inhibit or reverse these binding properties, remains to be fully elucidated. Heparan sulfate, and HS-like GAG, expressed on uninfected RBCs are receptors for rosetting [17, 60,61]. Sevuparin might act as a decoy receptor for CD36, thus causing binding to the drug rather than to the endothelial receptor. Heparan sulfate, and thus likely also sevuparin, binds to the DBL1α domain of PfEMP-1, which could in addition directly affect both rosetting and cytoadherence. [17,18]. It could also affect cytoadherence to parasite PfEMP-1-DBL1α receptors beyond CD36. However, a recent study using co-crystallization showed that the CIDR-domain of PfEMP-1 interacts closely with the hydrophobic pocket in CD36 [62], questioning the central role of the DBL1α domain. The interference of HDMEC binding observed in our study could thus also be mediated by the effect of sevuparin on binding to other endothelial receptors.
Pf-iRBCs tethered, rolled, and adhered on human microvascular endothelial cells. Under shear stress conditions, sevuparin significantly inhibited cytoadherence to HDMECs. A previous study reported that Pf-iRBCs interaction was synergistic with multiple adhesion molecules on vascular endothelial cells whereas adhesion of Pf-iRBCs to HDMECs was largely CD36 dependent, while the rolling could be mediated by any of the adhesion molecules [63]. Pf-iRBCs adherence to HDMECs was significantly inhibited by sevuparin under flow conditions. Inhibition of adhesion was most effective with higher sevuparin doses. Previous work has shown that sevuparin reverses sequestration of IRBCs in an animal model [38]. Our current study clearly showed inhibition of binding, but did not address reversal of cytoadherence. Desequestration with sevuparin has been described in a monkey model [38]. Mimicking this in the in-vitro model will require additional studies. In conclusion, sevuparin disrupts P. falciparum rosette formation and cytoadherence to HDMECs in a dose dependent manner and therefore might be a candidate for adjunctive therapy in severe falciparum malaria.

Sample collection
Adult male and female (18-65 years old) presenting with acute uncomplicated P. falciparum malaria were considered for enrollment. Uncomplicated malaria defined as febrile disease with a positive peripheral blood malaria slide. Patients with any sign of severe malaria, determined on the basis of World Health Organization criteria [64] were not eligible.
Five millilitres of citrate anticoagulated blood samples were collected prior to the start of antimalarial drug treatment for microscopy malaria parasite identification in thick and thin blood smears and blood group assessment. Samples were then centrifuged at 2,500 rpm at 4˚C for 5 minutes and plasma was stored at -80˚C. The packed red blood cells were suspended and washed three times with RPMI-1640. Then Pf-iRBCs were cultured and the parasites cryopreserved. All experiments were performed during the first cycle of in vitro parasite growth. For the rosetting assay, all assays were done on fresh samples and 47 of 50 samples could be evaluated. Out of these, 28 cryopreserved isolates showed cytoadherence in the static assay and were further assessed for sevuparin. Out of these 28, 10 strains with most prominent binding were selected for the flow based endothelial cells adhesion assays.

Endothelial cell isolation and culture
Human dermal microvascular endothelial cells (HDMECs) were harvested from discarded neonatal human foreskins as described previously [63]. In brief, foreskins were dissected into 2 to 3 square millimetre (mm 2 ) segments, and digested at 4˚C for 16 hours in supplemented M199 containing 0.5 mg/mL collagenase type 1A (Boerhringer Mannheim Biochemicals, USA). Microvessels were released from digested tissue segments by compressing the tissue segments gently with a spatula. To remove large debris, vessel preparations were passed through a 100 mm nylon mesh (Becton Dickinson, USA). The cells were collected by centrifugation at 2000 rpm for 10 minutes, and the cell pellets were resuspended in endothelial basal medium (EBM-2) with supplements supplied by the manufacturer (Clonetics, USA), and then seeded in gelatin-coated tissue culture dishes or on cover slips. Finally, The HDMECs culture dishes were incubated at 37˚C, 5% CO 2 . The experiments used HDMECs from passage 3 to 6 obtained from 5 different foreskin donors. HDMECs cultures were assessed for the endothelial marker CD31 and the parasite cytoadhesion receptor CD36 by an indirect immunofluorescence assay using FITC-mouse antihuman CD31 (BD Pharmingen, USA; No. 555445) and PE-mouse antihuman CD36 (BD Pharmingen, USA; No. 555455) (S1 Fig). 5 different HDMEC cell lines were used, which could differ in their adhesion properties. However, all 5 lines were selected and tested with reference CD36-positive binding parasite strains (Panned A4 and TM267 on HDMECs). There are not significant differences in the numbers of cytoadherence. (p = 0.239, for A4 and p = 0.974, for TM267).

Preparation of sevuparin
Sevuparin was provided by Dilaforette AB, Sweden. Sevuparin is a compound derived from heparin. Sevuparin was diluted in RPMI-1640 to 2 mg/mL as stock solution and kept at 4˚C until used. Sevuparin was diluted 2-fold for the experiments (range 62.5-1000 μg/mL).

Rosetting assay
Trophozoite stages of P. falciparum (0.5% parasitaemia and 1% haematocrit) were treated with sevuparin (62.5 μg/mL-1000 μg/mL) at 37˚C for 30 minutes. The cell suspension alone was used as control (i.e., 0 μg/mL). Rosette formation was then examined under light microscopy. The number of rosettes was counted per one hundred infected RBC under high magnification (100X objective lens). A rosette was scored if two or more uninfected red blood cells were bound to a single infected red blood cell. All samples were done in duplicate.

Static endothelial cell adhesion assay
HDMECs at 80% confluence on 13-15 mm Thermanox coverslips (Nunc) were used. Trophozoite stages of P. falciparum (0.5% parasitaemia and 1% haematocrit) were treated with sevuparin (50 μg/mL-800 μg/mL for 1 hour before removing the drug, then cell suspensions were applied to each coverslip and incubated at 37˚C for 1 hour with gentle rotation every 15 minutes. Unbound cells were removed by washing three times. After fixing with 0.5% glutaraldehyde and staining the cells with Giemsa's stain, the coverslips were mounted using DPX mounting medium. The number of infected red blood cells binding to 1000 HDMECs was counted. Untreated parasites at the same stage of development were used as controls [32, 63,66].

Flow based endothelial cell adhesion assay
Five to one hundred thousand HDMECs in 1 ml of supplemented medium were plated on 35mm glass coverslips (Deckglasser, Germany) in 6 well plates. The cells were allowed to adhere for 2 h at 37˚C, after which 3 ml of supplemented medium were added to each well. After 40 to 48 h of culture, confluent monolayers were rinsed in HBSS (Life Technologies) and used without delay in the laminar flow studies. In the adhesion assys, trophozoite stage P. falciparum (at 0.5% parasitaemia and 1% haematocrit) were preincubated with sevuparin at 50, 200 and 800 μg/mL of seveuparin for 1 hour, after which the drug was removed by washing 1 time by centrifucation at 2500 rpm, for 5 minutes and resuspension in RPMI-1640 to an haematocrit of 1%, after which the Pf-iRBCs suspension was used immediately.
The method was modified from that previously reported [63,66,67]. Briefly, monolayers of HDMECs were assembled in a parallel plate flow chamber of 220 μm gap thickness in which a uniform wall shear stress was generated. The flow chamber was mounted on the stage of inverted phase contrast microscope that is kept 37˚C and IRBC interactions with monolayers, observed under 200 x-magnifications. All assays were performed at a wall shear stress of 0.05 Pa. This sheer stress (ɩ) was calculated from the flow rate (Q), the medium viscosity (ɳ), and the chamber diameters (w, width, and h, height): ɩ = 6. ɳ .Q/w.h 2 . A recording video was obtained of Pf-iRBCs binding over 6 separate fields; 6 minutes/field. The Pf-iRBCs suspension was passed over the endothelial cells coated cover glass slip for 6 minutes. Cell free medium was then flushed through the chamber at the same shear stress for 2 minutes to wash away non-adherent cells. Stationary adherent parasites were counted in at least six fields along the slide (both rolling and static) and the results were expressed as the number of adherent Pf-iRBCs/6 minute [63,65,66]. Tethering referred to initial contact between Pf-iRBCs and the endothelial monolayer. A rolling Pf-iRBC was defined as one that displayed a typical end-on-end rolling motion at a velocity of less than 3 mm and expressed as the number of rolling Pf-iRBCs/6 minutes. The Pf-iRBC was considered adherent if it remained stationary for more than 3 seconds. The results were expressed as the number of adherent Pf-iRBCs/6 minutes.

Statistical analysis
All statistical analyses were performed using SPSS 18 for Windows. The data were tested for distributional assumptions by the Kolmogorov-Smirnov test. The Mann-Whitney U-test was used for non-normally distributed data to compare different levels of drug concentration against the controls. Comparisons across more than two groups were made using the Kruskal-Wallis test. Associations were assessed using the Spearman's rank correlation coefficient. 16. Barragan