Both authors contributed equally to the manuscript.
Effects of hydroxyurea on malaria, parasite growth and adhesion in experimental models
Version of Record online: 18 AUG 2006
Volume 28, Issue 12, pages 675–680, December 2006
How to Cite
PINO, P., TAOUFIQ, Z., BRUN, M., TEFIT, M., FRANETICH, J. F., CICERON, L., KRISHNAMOORTHY, R. and MAZIER, D. (2006), Effects of hydroxyurea on malaria, parasite growth and adhesion in experimental models. Parasite Immunology, 28: 675–680. doi: 10.1111/j.1365-3024.2006.00907.x
Financial support: This work was in part supported by VIH-PAL. Paco Pino received a scholarship from VIH-PAL; Zacharie Taoufiq was financially supported by the Ministère de l’Education Nationale de la Recherche et de la Technologie.
- Issue online: 18 AUG 2006
- Version of Record online: 18 AUG 2006
- Received: 20 December 2005 Accepted for publication: 21 June 2006
- endothelial cell;
- falciparum malaria;
We recently raised concern over using hydroxyurea (HU) in the treatment of sickle cell disease in areas endemic for malaria, becauseit up-regulates the endothelial surface expression of ICAM-1, a major receptor for Plasmodium falciparum-infected erythrocytes in the brain. Using human in vitro models of cerebral malaria, we evaluated the interaction of HU with parasites and demonstrated that HU pretreatment increased the number of infected red blood cells adhering to the endothelium, but did not increase endothelial apoptosis. Moreover, using an experimental cerebral malaria model, HU pretreatment was found to prevent significantly mice from developing neurological syndrome by inhibiting parasite growth, opening potential therapeutic avenues.
sickle cell disease
- S HbF
human endothelial cells
red blood cells
parasitized red blood cells
- Pf AA-RBC
Plasmodium falciparum-infected normal erythrocytes
- Pf SS-RBC
Plasmodium falciparum-infected sickle erythrocytes
intercellular adhesion molecule 1
Sickle cell disease (SCD) is an inherited blood disorder characterized by the presence of abnormal haemoglobin S (HbS). The clinical hallmark of SCD is the periodic occurrence of vaso-occlusive crisis and appears to be due to, among other aetiologies, enhanced adherence of sickle red blood cells (SS-RBC) to the endothelium (1). Oral hydroxyurea (HU), a specific inhibitor of nucleotide reductase, has become a prototypic therapeutic agent for the treatment of adult and paediatric SCD patients in western centres with structured management (2–4). Despite the fact that larger scale clinical trials are still to be performed in order to study side-effects, HU has been proved to be highly efficient in decreasing the frequency of the pain crises and vaso-occlusive strokes in SCD patients. Given the clinical efficacy of this unique therapeutic option for SCD, the temptation is high to extend these benefits to as many patients as possible, especially in Africa, where the incidence of SCD is the highest.
The clinical efficacy of HU is explained by its ability to enhance foetal haemoglobin expression, in that way decreasing HbS polymerization (5). However, recent works suggested that hydroxyurea could also prevent vascular complications by modulating adhesion molecules on various types of cells, such as leucocytes. While decreasing the expression and synthesis of endothelin-1 (ET-1), a vasoconstrictor peptide found at high level in SCD patients, unexpectedly, HU was found to up-regulate ICAM-1 expression by endothelial cells in culture (6). ICAM-1 is one of the major cell surface receptors for adhesion of Plasmodium falciparum-infected red blood cells (PRBC) to endothelium (7). PRBC adhesion to endothelium and the ensuing endothelial damage is believed to be critical in the pathogenesis of severe malaria disease, including cerebral malaria, one of the most life-threatening complications associated with an infection by P. falciparum (8,9). We recently demonstrated that PRBC contact induces apoptosis in primary human endothelial cells (HEC) in vitro. An up-regulation of pro-inflammatory and pro-apoptotic genes (Fas, Fas L, iNOS … ) and the activation of caspase 8, 9 and 3 were observed (10).
Plasmodium falciparum malaria and sickle cell disease endemic zones overlap, especially in central Africa. Even though the sickle cell trait confers a natural resistance to severe complications resulting from malaria infections, (11,12), the administration of HU to SCD patients in malaria-endemic areas still raises concern.
MATERIALS AND METHODS
Culture of human endothelial cells
Primary human lung endothelial cells (HEC) were isolated from human lung after enzymatic digestion and selected using a continuous gradient and immunomagnetic purification technique as described elsewhere (13). Endothelial cells derived from one batch were used for the experiments and were verified for their expression of ICAM-1, CD36, Von Willebrand factor, VCAM-1, CD31, E/P-selectin and CSA. HEC were raised at 37°C with 5% CO2, in M199 medium (Gibco), supplemented with 10 µg/mL of endothelial cell growth supplement (ECGS) (Upstate, Lake Placid, NY) and 10% of foetal calf serum (Biowest, South American origin).
Plasmodium falciparum culture
The P. falciparum 3D7 clone was used for these experiments. The strains were synchronized and maintained in culture according to Trager and Jensen's technique (14) in a suspension of erythrocytes in RPMI (Gibco) supplemented with 8·3 g/L of Hepes (Sigma, St Louis, MO, USA), 2·1 g/L of NaHCO (Merck), 0·1 mg/mL of gentamycin and 2 g/L of dextrose and 0·4% of Albumax II (w/v) (Life Technologies). The 3D7 clone was characterized for adhesion phenotype as previously described (15) and adheres at ICAM-1 and CD36 at 30 and 45%, respectively. For each experiment, parasite cultures were enriched in mature forms by Plasmagel floating (16). Briefly, cells were resuspended in Plasmion® (Fresenius Kabi, France) and incubated during 25 min at 37°C. The upper fraction containing mature trophozoites and schizonts was then collected and washed three times in RPMI. Parasitized red blood cell suspensions were then adjusted to 5% of haematocrit and 50% of parasitaemia.
Sickle erythrocytes (SS-RBC) were infected in vitro by the 3D7 clone and cultivated under the same conditions. PRBC cultivated with normal erythrocytes (AA-RBC) were Plasmagel purified (parasitaemia > 90%) and used to infect SS-RBC. Sickle erythrocytes were then cultivated for two cell cycles; in this way the contamination of the SS-RBC population by AA-RBC was reduced as much as possible. Plasmodium falciparum-infected sickle erythrocytes (Pf SS-RBC) were then purified by Plasmagel before the experiments.
Cytoadherence assays were carried out as previously described (17). Suspensions of PRBC (Pf SS-RBC or Pf AA-RBC) at 50% parasitaemia and 5% haematocrit were incubated on an HEC monolayer for 1 h with gentle continuous shaking at 37°C. After removal of non-adherent erythrocytes with three washes using parasite culture medium, the preparation was fixed for 30 min with 2% glutaraldehyde, rinsed with PBS and stained with Giemsa. Adhesion was then counted by microscopy analysis. HEC were incubated with uninfected erythrocytes as control.
Endothelial ICAM-1 expression
Endothelial ICAM-1 expression was performed as previously described (6). Briefly HEC were treated with hydroxyurea for 24 h and total RNA was prepared using the RNeasy mini kit (Qiagen, Valencia, CA, USA). The reverse transcription reaction was then performed using Superscript II reverse transcriptase from Moloney Murine Leukaemia Virus (Life Technologies Grand Island, NY, USA). ICAM-1 mRNA levels under the different experimental conditions was then quantified by quantitative real time PCR assay (TaqMan 7700TM). A fragment of 102 bp corresponding to the extracellular domain of ICAM-1 was amplified using the following primers: forward hICAM-1 (A) 50-CAATGTGCAAGAAGATAGCCA-30 and reverse hICAM-1 (B) 50-GGGCAAGACCTCAGGTCATGT-30.
The proportion of apoptotic cells was estimated by measuring DNA fragmentation in the cultures. An ELISA technique (Cell Death ELISA Plus, ROCHE) dosing the release of mono- and oligonucleosomes in the cytoplasm of apoptotic cells (Roche) was adapted to our model. Briefly, confluent HEC in 96-well plates (Costar) pretreated or not with HU for 24 h, were co-cultivated with Plasmagel-purified PRBCs (Pf AA-RBCs or Pf SS-RBCs) at a parasitaemia of 50% and an haematocrit of 5% for 4 h. Non-infected AA/SS-RBC were used as controls. Cells were then washed and processed for DNA fragmentation analysis (10).
Experimental cerebral malaria experiment
Six- to 8-week-old female C57BL/6 N mice were purchased from Charles River Breeding Laboratories (Saint-Aubin les Elbeufs, France). Clone 1·49 L of the Plasmodium berghei ANKA strain was maintained in C57BL/6 mice and induces experimental cerebral malaria, characterized by hemi- or paraplegia, ataxia, deviation of the head, and convulsions between 6 and 8 days post-infection. The parasite infection was induced by i.p. injection of 106 pRBCs. Parasitaemia was determined daily on Giemsa-stained blood smears and was expressed as the per cent parasitaemia.
C57BL/6 mice were treated daily by gavage with HU (200 mg/kg/day). Parasitaemia of individual mice was evaluated on Giemsa-stained thin blood smears every day.
Each experiment was carried out three times independently. Results were analysed for statistical significance using the one-way anova followed by the Tukey multiple comparison test. A value of P < 0·05 was considered significant.
Effects of hydroxyurea on endothelial cells in vitro
To assess the effect of hydroxyurea on PRBC endothelial adherence, we performed co-culture experiments of human endothelial cells and P. falciparum-infected erythrocytes as previously described (10), including P. falciparum-infected sickle erythrocytes (Pf SS-RBC). Briefly HEC were pretreated with various concentrations of HU for 24 h prior to the incubation with infected erythrocytes. Induction of ICAM-1 expression and cytoadherence of red blood cells were assessed, respectively, by quantitative RT-PCR and microscopy analysis. As reported in Brun and colleagues’ study (6), we found that pretreatment of HEC with HU resulted in a dose-dependent enhanced expression of membrane-bound ICAM-1 within endothelial cells (Figure 1a). Indeed HEC pretreated with 250 µm HU showed a 2 times higher ICAM-1 expression than basal level (untreated cells). There was a 3 times increase of ICAM-1 expression in 500 µm HU-treated endothelial cells. In parallel, an increase of the number of adherent Pf AA-RBC and Pf SS-RBC was observed (Figure 1b), respectively, up to 1·8 ± 0·2 fold and 1·4 ± 0·2 fold in 1 mm of HU-treated cells, vs. the number of Pf AA-RBC counted on control non-treated HEC. We also evaluated consequences of those cytoadherence conditions on endothelial cell apoptosis. Genomic DNA degradation was quantified by intracytoplasmic nucleosome release as a late apoptosic marker (Cell Death ELISA Plus, Roche) (10) (Figure 1c). As expected, HEC co-incubated for 4 h either with Pf AA-RBC or Pf SS-RBC showed strong apoptosis induction (3·2 fold and 2·2 fold, respectively) compared to uninfected erythrocyte/endothelial cell co-incubation. We did not observe any significant difference in parasite-induced endothelial apoptosis when cells were pretreated with increased concentrations of HU (0–1 mm). Therefore we can conclude that although HU enhanced endothelial adhesion of Pf RBC through an ICAM-1 induction, it did not cause an increase in cytoadherence-induced apoptosis.
Effects of hydroxyurea in a murine model of cerebral malaria
On the other hand we evaluated HU potential effects in vivo using the experimental cerebral malaria C57BL/6–P. berghei ANKA mouse model (18,19). In this model all untreated mice died between days 6 and 8 at 7–15% of parasitaemia, with neurological signs similar to those observed in human cerebral malaria (ataxia, convulsions and coma). Daily oral administration of HU (starting 3 days prior to the infection, 200 mg/kg/day) resulted in a protection against parasites in 6 out of 7 mice. Indeed parasitaemia was significantly lower in treated animals, between 0·5 and 2%, with almost total arrest of parasite growth as long as treatment was given (Figure 2). In the treated mouse that died at day 9 from CM, parasite growth was not inhibited. Withdrawal of HU on day 10 caused a rapid parasitaemic burst, but the animals did not manifest neurological signs and developed hyperparasitaemia (25–50%).
Sequestration of P. falciparum-infected erythrocytes within the brain capillary network is believed to be critical in severe malaria pathogenesis. Cytoadherence appears to be essential but not sufficient, since not all P. falciparum infections lead to malaria complications (20). PRBC contact via adhesion molecules induces endothelium damage and may then play a key role in CM, raising the importance of modulation of those receptors. Although HU up-regulates the expression of membrane-bound ICAM-1, increasing the adhesion of both Pf AA-RBC and Pf SS-RBC in vitro, these events did not lead to increased endothelial cell apoptosis (Figure 1). This absence of correlation between the degree of cytoadherence and the capacity of infected erythrocytes to induce apoptosis is in concordance with what we observed with different isolates from patients living in Gabon (Touréet al. submitted). It is unclear how PRBC adhesion to HEC induces apoptosis, but cross-linking of adhesion molecules at the endothelial surface can trigger intracellular specific signalling pathways leading to a variety of cellular responses, including apoptosis. Indeed binding of the CD36 ligand thrombospondin-1 to HEC is sufficient to transmit an apoptotic signal mediated by p59/fyn and caspases (21). According to these data, we can hypothesize that the activation of HEC-specific surface molecules (still to be identified) by PfEMP1 or other P. falciparum surface ligands may be responsible for the observed apoptosis.
In separate experiments performed in assays as previously described (22), HU was found to inhibit the growth of P. falciparum with an IC50 of 473 µm at 96 h. Holland et al. (23) have already tested HU against P. falciparum, but the incubation time with the drug was only 24 h, and the resulting IC50 was markedly higher (792 µm). HU needs longer periods to be efficient. This weak activity does not make HU a new anti-plasmodial drug, since, for comparison, the IC50 of choroquine is at the nanomolar range. Nevertheless, in serum from treated patients, HU concentrations can reach 500–1000 µm, at which a significant anti-malarial activity is observed (24,25).
It was also demonstrated that foetal haemoglobin (HbF) can ‘trap’Plasmodium parasites independently of the spleen clearance mechanism. Erythrocytes containing HbF are invaded more quickly but growth of the parasites is diminished (26). Indeed in vitro experiments showed that recombinant P. falciparum plasmepsin II could digest HbA twice as well as HbF (27). Gavage treatment of HU in our CM mouse model might have provided a double protection via (a): HU enhanced foetal haemoglobin production and (b) HU direct anti-plasmodial activity; maintaining low parasitaemia (< 2%). As a matter of fact, when HU treatment was withdrawn parasite growth rapidly increased (Figure 2). Mice were then highly protected from developing P. berghei ANKA-induced cerebral symptoms. This is in concordance with what was observed by Shear et al. (27), where transgenic mice expressing human HbF were protected from another rodent cerebral malaria strain P. yoelii 17XL.
In conclusion, we showed that though hydroxyurea would up-regulate the endothelial cell surface molecule ICAM-1 and would increase the cytoadherence phenomenon, it would not necessarily induce more endothelium apoptosis and subsequently worsen the disease. Moreover in the in vivo model, HU prevents mice from developing cerebral symptoms. This may be due to static effects on parasite growth, but the point is that HU could, at significant SCD treatment concentrations, maintain low Plasmodium parasitaemia in patients as well. HU cannot be considered as an anti-malarial drug, but man should consider it as this molecule has the ability to enhance foetal haemoglobin (HbF) production and influence indirectly parasite growth. Indeed it has been demonstrated that Plasmodium parasites invade more quickly erythrocytes that contain HbF but cannot digest it properly to complete their development. This ‘suicidal invasion’ or ‘HbF trap’ concept has already been well studied (5,27–29).
Taken together, these data allow us to speculate that administration of oral HU would not be detrimental in areas where malaria and sickle cell disease are endemic.
We thank Eliane Giboyau for technical assistance and VIH-PAL and PAL+ for financial support. Zacharie Taoufiq was financially supported by the Ministère de l’Education Nationale, de la Recherche et de la Technologie.
- 32001: CD002202.& Hydroxyurea for sickle cell disease. Cochrane Database Syst Rev