CD36 deficiency protects against malarial anaemia in children by reducing Plasmodium falciparum-infected red blood cell adherence to vascular endothelium


  • Jaffu Chilongola,

    1.  Kilimanjaro Clinical Research Institute, Kilimanjaro Christian Medical College, Moshi, Tanzania
    2.  Department of Immunology and Molecular Biology, Faculty of Medicine, Kilimanjaro Christian Medical College, Tumaini University, Moshi, Tanzania
    Search for more papers by this author
  • Sakurani Balthazary,

    1.  Department of Physiology, Biochemistry, Pharmacology and Toxicology, Faculty of Veterinary Medicine, Sokoine University of Agriculture, Morogoro, Tanzania
    Search for more papers by this author
  • Maximillian Mpina,

    1.  Ifakara Health Institute, Bagamoyo Research and Training Centre, Bagamoyo, Tanzania
    Search for more papers by this author
  • Martin Mhando,

    1.  Department of Physiology, Biochemistry, Pharmacology and Toxicology, Faculty of Veterinary Medicine, Sokoine University of Agriculture, Morogoro, Tanzania
    Search for more papers by this author
  • Erasto Mbugi

    1.  Department of Biochemistry, School of Medicine, Muhimbili University of Health and Allied Sciences (MUHAS), Dar es Salaam, Tanzania
    2.  Wageningen University, Cell Biology and Immunology Group, AH Wageningen, The Netherlands
    Search for more papers by this author

Corresponding Author Jaffu Chilongola, Kilimanjaro Clinical Research Institute, KCMC, Moshi, Tanzania. E-mail:


Objective  CD36 is a receptor that occurs on the surface of activated immune cells, vascular endothelial cells and participates in phagocytosis and lipid metabolism. CD36 is known to be the major endothelial receptor molecule for field isolates of Plasmodium falciparum. A T1264G mutation in exon X of the gene leads to deficiency of CD36. This study aimed at determining associations between CD36 deficiency, P. falciparum in vitro adherence on purified CD36 and anaemia among children in an endemic area.

Methods  Genotypes were determined by nested polymerase chain reaction of isolated DNA from filter blood spots followed by Restriction Fragment Length Polymorphism (RFLP). Plasmodium falciparum adherence assays were performed on immobilized purified CD36.

Results  The data indicate that CD36 is an important cytoadherence receptor that mediates adherence to most P. falciparum field isolates. Our findings also suggest that mutations causing CD36 deficiency may confer significant protection against malarial anaemia (MA) in children (χ2 = 8.58, P < 0.01).

Conclusion  That the protective role that CD36 deficiency may confer against MA in children seems to be mediated through reduced cytoadherence of infected red blood cells to vascular endothelium.

La déficience en CD36 protège contre l’anémie de la malaria chez les enfants en réduisant l’adhésion des globules rouges infectés par Plasmodium falciparumà l’endothélium vasculaire

Objectif:  CD36 est un récepteur apparaissant à la surface de cellules immunitaires activées, des cellules endothéliales vasculaires et participe à la phagocytose et au métabolisme lipidique. CD36 est connu pour être la principale molécule endothéliale jouant le rôle de récepteur pour des isolats de terrain de P. falciparum. Une mutation T1264G dans l’exon X du gène conduit à la déficience en CD36. Cette étude vise à déterminer les associations entre la déficience en CD36, l’adhésion in vitro de P. falciparum sur CD36 purifié et l’anémie chez les enfants dans une zone d’endémie.

Méthodes:  Les génotypes ont été déterminés par ‘nested’ PCR de l’ADN isoléà partir de gouttes de sang récoltées sur papier filtre suivi par une RFLP (Polymorphisme de la taille des fragments de restriction). Les tests sur l’adhésion de P. falciparum ont été effectués sur du CD36 purifié et immobilisé.

Résultats:  Les données indiquent que CD36 est un important récepteur de cytoadhérence médiant l’adhésion de la plupart des isolats de terrain de P. falciparum. Nos résultats suggèrent également que les mutations responsables de la déficience en CD36 peuvent conférer une protection significative contre l’anémie malarique chez les enfants (X= 8,58; p < 0,01).

Conclusion:  Le rôle protecteur que la déficience en CD36 peut conférer contre l’anémie malarique chez les enfants semble être médié par le biais de la cytoadhérence réduite des globules rouges infectés à l’endothélium vasculaire.

La deficiencia de CD36 protege frente a la anemia por malaria en niños, reduciendo la adherencia de eritrocitos infectados por Plasmodium falciparum al endotelio vascular

Objetivo:  CD36 es un receptor presente en la superficie de células inmunes activadas, células del endotelio vascular y participa en la fagocitosis y el metabolismo lipídico. Se conoce que el CD36 es la principal molécula receptora del endotelio para cepas de campo de P. falciparum. La mutación T1264G en el exón X del gen conlleva a una deficiencia de CD36. Este estudio buscaba determinar las asociaciones entre la deficiencia de CD36, la adherencia in vitro de P. falciparum sobre CD36 purificadas y la anemia entre niños de áreas endémicas.

Métodos:  Se determinaron los genotipos mediante PCR anidada de ADN aislado a partir de sangre en papel de filtro, seguida por un análisis de Polimorfismos en la Longitud de Fragmentos de Restricción (RFLP por sus siglas en inglés). Los ensayos de adherencia de P. falciparum se realizaron sobre CD36 purificado e inmovilizado.

Resultados:  Los datos indican que el CD36 es un receptor citoaherente importante, que media la adherencia de la mayoría de cepas de campo de P. falciparum. Nuestros hallazgos también sugieren que las mutaciones que causan una deficiencia de CD36 podrían conferir una protección significativa frente a la anemia por malaria en niños (X= 8.58, p < 0.01).

Conclusión:  El papel protector que la deficiencia de CD36 podría conferir frente a la anemia por malaria en niños parece ser mediada por una citoadherencia reducida de los eritrocitos infectados al endotelio vascular.


A unique characteristic feature of infections with Plasmodium falciparum is the ability of infected red blood cells (IRBCs) to adhere to vascular endothelium by cytoadherence. The result of cytoadherence is accumulation of IRBCs in the deep microvasculature. This phenomenon is called ‘sequestration’ and is associated with disease outcomes (Berendt et al. 1990; Beeson et al. 1999; Amodu et al. 2005). In some instances, parasite populations with a predisposition to adhere to certain receptors are more commonly associated with certain disease outcomes, such as cerebral malaria and placental malaria (Ockenhouse et al. 1992a,b; Ockenhouse 1993), although the precise role of parasite–receptor interactions in determining disease severity remains to be understood. Sequestration plays a major role in the development of severe disease and occurs as a result of both cytoadherence of IRBCs to capillary endothelium (Berendt et al. 1990; Crandall et al. 1994; Ho & White 1999) and the binding of IRBCs to uninfected erythrocytes (rosetting) (Carlson et al. 1990; Rowe et al. 1995). Adherence is mediated via knob-like structures at the surface of IRBCs (Luse & Miller 1971) resulting from the deposition and aggregation of parasite proteins and their interaction with the host cell cytoskeleton underneath the RBC membrane (Culvenor et al. 1987; Pologe et al. 1987; Pasloske et al. 1993). Studies have suggested that P. falciparum erythrocyte membrane protein 1 (PfEMP-1) expresses variant-specific epitopes and mediates adhesion to endothelium. CD36 is among the host ligands that have been found to mediate endothelial binding of IRBCs and has been identified in most field isolates (Oquendo et al. 1989; Biggs et al. 1990; Hasler et al. 1990; Chaiyaroj et al. 1994; Xiao et al. 1996; Ho & White 1999; Trenholme et al. 2000; Yipp et al. 2007; Cojean et al. 2008).

CD36 is one of the most characterized host receptors for P. falciparum IRBCs. CD36 is an 88-kDa glycoprotein expressed on endothelial cells, platelets, macrophages and dendritic cells, and participates in phagocytosis and lipid metabolism. However, it is not expressed on endothelial cells of brain capillaries. The gene encoding CD36 consists of 15 exons, and extends at least 32 kb on the q11.2 band of chromosome 7 in humans. As most P. falciparum field isolates bind CD36, CD36 is considered the major endothelial receptor for sequestration, although not all parasites bind this receptor (Newbold et al. 1997; Rogerson et al. 1999). CD36 exons X–XII occur in numerous polymorphic forms in the Gambia and in Kenya and several of these forms are associated with susceptibility to cerebral malaria. A subsequently described polymorphism of exon X confers protection against severe anaemia in heterozygotes in Kenya by mechanisms not clearly defined (Pain et al. 2001).

The mechanism by which CD36 confers protection against malarial anaemia (MA) in children is not clearly understood. This study was designed to determine role of cytoadherence in the protection of CD36 deficiency against MA in children below 5 years of age. In particular, we assessed the role played by CD36 deficiency in the development of MA in children. The study specifically explored the frequency of CD36 deficiency among children in malaria endemic areas and associations between CD36 deficiency and status of MA in children and the role played by cytoadherence on the development of MA in children.

Materials and methods

Participants, recruitment and consent

Mothers were recruited among women presenting at Muheza Designated District Hospital for antenatal care in the third trimester or for delivery hospitalization. Mothers of prospective study children initially learned about the study through community meetings. Mothers who qualified for inclusion were requested to read and sign a consent form on behalf of their newborns. Verbal consent was given by those unable to read and write, followed by thumb stamp. Children whose mothers consented for participation, and whose mothers were willing to bring them fortnightly and when they are sick for sample collection were included in the study. Unwillingness to sign the informed consent form and to give samples when required were exclusion criteria for this study. In total, 155 children were recruited. Of these, 149 completed the follow up, and provided 204 samples used for data analysis. Six children were dropped at different time points due to various exclusion criteria.

Definition of categories

A malaria case was defined as a child testing positive for P. falciparum parasites with a fever (body temperature) of >38.5 °C and any of the typical malaria symptoms. MA was defined as a haemoglobin (Hb) concentration of ≤10 g/dl, in the presence of positive thin and/or thick blood smears at any parasitemia level. Non-anaemia (NA) was defined as an Hb concentration of >10 g/dl, in the presence of malaria parasites at any parasitemia level. Hb concentration was determined using a Cell-Dyn 1200® Haematology Analyser (Spectron Corporation, Burlington, WA, USA). In this study, an IRBC was considered to be a binder to CD36 if the number of IRBCs binding to CD36 was at least twice as much as the number of IRBCs binding to the control molecule, bovine serum albumin (BSA), observed under light microscope at 10 × magnification in 20 fields of the binding plate. This study was approved by both the Tanzanian and the Seattle Biomedical Research Institute (Seattle, WA, USA) Ethics Review Boards.

Sample collection

We genotyped 155 children for the CD36 gene at the start of the follow-up. The children were then followed for a period of 12 months from November 2003, for occurrence of malaria-associated anaemia. Only malaria positive blood samples were used in this study. Each collected sample was treated as a separate sample representing a different infection. Samples were collected as passive cases when a child presented with malaria. Every time a sample was collected, Hb concentration was measured, and IRBCs cultured for binding assays. Approximately 1 ml of blood was collected by venipuncture into 10 ml vacutainers (Fisher Inc, USA) containing citrate phosphate dextrose (CPD) as anticoagulant, at a ratio of 1:10, CPD: blood. A blood drop from whole blood was placed on Whatman® filter paper strips (Whatman, USA) and stored at room temperature for genotyping.

Parasites quantification and in vitro culture

Small volumes (5 μl) of blood were used to prepare a thick and thin smear for detection of malaria parasites. Slides were stained with Giemsa stain at pH 7.2. The number of IRBCs with asexual parasites was counted against 2000 total red blood cells (both infected and non-infected) to obtain the percentage of IRBC. The infected blood was then diluted using a mixture of blood group O+ cells and AB sera from malaria non-immune volunteers to 3–5% IRBCs, and cultured in a complete media with RPMI 1640 (Gibco), 10% human sera, Pen-strep (Gibco) and Gentamycin (Gibco) for 24 h at 37 °C, 5% CO2, humidified incubator, to allow them to develop into the mature form before subjected to binding assays to determine their CD36 binding phenotypes.

Determination of parasite binding phenotype

Parasites (IRBCs) were allowed to develop into the mature form by in vitro culture before being subjected to binding assays to determine their CD36 binding phenotypes. Purified CD36 (20 μl) and BSA at a concentration of 1 μg/ml (Sigma Chemicals Co. St Louis, USA) were placed onto a Petri dish and incubated overnight at 4 °C. Petri dishes were then blocked with 2% BSA for 30 min at 37 °C and washed three times with 1 × PBS (Dulbeco). The dishes were incubated with 5% Hematocrit of washed parasites for 30 min at 37 °C followed by a triple wash with PBS. Parasites on Petri dishes were fixed using 1% (w/v) Formaldehyde solution at room temperature for 1 min, and then stained with 1% Giemsa stain for 1 min. The number of parasites binding onto each molecule in relation to those binding to BSA was determined by observing under light microscope at 10 × magnification in 20 fields of the binding plate.

DNA extraction and PCR amplification of CD36

DNA extraction was performed using the Gentra DNA extraction protocol (Gentra Systems Inc, Minneapolis, USA) according to the manufacturer’s instructions. DNA was used as 10% of the PCR reaction mix. Nested PCR was used throughout this experiment. The reactions were performed in 50 μl reaction tubes on a PTC-100 Programmable Thermal Controller (MJ Research Inc., USA). The primer sequences for the first PCR reaction were F: 5′ ATG CTT GGC TAT TGA GT and R: 5′ TAT CAC AAA TTA TGG TAT GGA CTG and those for the nested PCR were F: 5′ CTA TGC TGT ATT TGA ATC CGA and R: 5′ ATG GAC TGT GCT ACT GAG GTT ATT CGT T. The nested primers were designed using DNAstar® software (Lasergene, Madison, USA). Distilled water was included in the control reactions instead of the isolated DNA. For both PCR reactions, the following PCR cycle was used: a initial denaturation step of 94 °C for 4 min followed by 45 cycles of 94 °C denaturation for 1 min, annealing at 55 °C for 30 s and elongation at 68 °C for 8 min. The first PCR reactions amplified fragments of 415 bp whereas the nested PCR reaction amplified fragments of 212 bp.

Genotyping by RFLP and agarose gel electrophoresis

The nested PCR product (20 μl) from each sample were placed into a 50-μl microtubes followed by 1 μl of restriction enzyme NdeI (BioLabs Inc, New England, USA). The mixture was incubated for 4 h at 37 °C and then was heated at 65 °C for 1 min to stop enzyme activity. The CD36 gene has a wild-type NdeI restriction site, 5′-CA/TATG. This single nucleotide mutation eliminates the inherent restriction site. Thus, NdeI digestion of the wild-type CD36 allele gave two fragments, 148 bp and 64 bp, whereas the homozygous mutant was not cut and thus was a single fragment of 212 bp. The heterozygous allele gave a mixture of the three fragments from the wild-type and the mutant allele, i.e. 212, 148 and 64 bp fragments. Restriction digestion products, PCR products and molecular weight markers were subjected to agarose gel electrophoresis in a 3.5% (w/v) agarose gel (Sigma Chemicals Co) containing 5 μl of 10 mg/ml ethidium bromide (Sigma Chemicals Co). PCR products, restriction digests and molecular weight markers were loaded onto the wells as 1 μl of 6 × loading dye (0.2% bromophenol blue, 0.2% xylene cyanol, 60% glycerol and 60 mm EDTA) in 10 μl of sample, and run in 1 × TE buffer at constant voltage of 120 V for 25–30 min. The DNA marker FX174/HinfI (BioLabs Inc) with fragment size range from 24 to 726 bp was used to determine the various band sizes for the samples.

Statistical analyses

For categorical (nominal) data, chi-square correlation tests were used to compare expected and observed frequencies for (genotypes, binding patterns and MA) parameters using SPSS version 14.0.1 (SPSS Inc., USA) computer software. Unpaired t-test comparisons of means were used to compare mean Hb readings among different genotypes.


Frequencies of CD36 alleles and CD36 binders

Results for CD36 genotyping and IRBC binding profiles are presented in Fig. 1 and Table 1. The gel presented in Fig. 1 shows an agarose gel for the PCR-RFLP of genotyped CD36 gene for the CD36 T1264G mutation. Table 1 summarizes the frequencies of different CD36 genotypes among children. Results show that out of the 204 genotyped children, 176 (86.3%) children had the wild-type allele, 22 (10.8%) children were heterozygous for the studied mutation and 6 (2.9%) children were homozygous for the CD36 mutation studied.

Figure 1.

 Agarose gel showing restriction fragments for CD36 alleles. Lanes 1, 20, 21 and 38: molecular marker with the following band size (bp) (top–bottom): 726, 713, 553, 500, (427, 417, 413 together), 311, 249 200, 151, 140, 118, 100, 82, 66, (48, 42, 40 together), and 24 bp. Lanes 2, 4, 6, 8, 10, 12, 14, 16, 18, 22, 24, 26, 28, 30, 32 and 34: undigested nested PCR product. Lanes 3, 5, 9, 11, 13, 15, 17 19, 23, 25, 27, 29: wild-type CD36 allele with two bands of 148 and 64 bp. Lanes 7 and 31: heterozygous for CD36 allele with three bands of 212, 148 and 64 bp. Lane 35: homozygous mutant with one band of 212 bp. Lanes 36 and 37: control samples (distilled water) digested and undigested, respectively.

Table 1.   Frequencies of CD36 genotypes and binders in each category
GenotypeGenotype frequency (%)CD36 binders (%)
  1. Data presented in column A: Numbers indicate the number of children that fall in each genotype (Wild-type, Heterozygous and Homozygous mutant). Numbers in parentheses show the percentage of respective genotypes out of 204 typed children. Column B: Data in column B show the number of IRBC samples binding to CD36. Numbers in parentheses indicate the respective percentage of binder IRBC samples in each genotype.

Wild-type176 (86.3)124 (70.5)
Heterozygous22 (10.8)20 (9.8)
Homozygous mutant6 (2.9)0 (0.0)

The same table shows results for proportion of binder IRBCs in isolated from each genotype. It is shown in Table 1 that 124 out of 176 (70.5%) IRBC samples from the wild-type genotype were CD36 binders. Twenty samples out of 22 (9.8%) and none of the 6 (0.0%) IRBC samples from the heterozygous and homozygous mutant genotypes were CD36 binders, respectively. Statistical analyses using chi-square and Fisher’s exact tests revealed an association between binding phenotype and occurrence of MA (χ2 = 19.5. P < 0.01) such that the non-binding IRBC phenotype was least associated with occurrence of MA in children. This was clearly shown by the absence of any binder IRBC sample from the homozygous mutant genotype, from which none of the children had MA at any time. Figure 2 presents overall binding results for all IRBC samples tested. Out of 204 samples, 144 (70.6%) of IRBC samples were binders while 60 (29.4%) were non-binders. From our data, we show here that most (70.6%) of the samples, from all genotypes, were CD36 binders.

Figure 2.

 Overall results for CD36- IRBC binding phenotypes. IRBC samples were either binders or non-binders if the number of CD36 binding IRBC were twice as much (or more) as the number of IRBC binding to BSA, tested on the same plate.

Occurrence of MA cases among CD36 binding phenotypes

Results for associations between occurrence of MA and binding phenotype are presented in Fig. 3. Children were categorized based on whether they had MA or not. This categorization was based on the cut-off point of 10 g/dl. It is shown in Fig. 3 that 20 IRBC samples out of 144 IRBCs that bound to CD36 (16.12%) were from anaemic children. Out of the 60 non-binder IRBC samples, only 2 (3.4%) were from children with MA. These data show that 96.6% of children who donated IRBCs samples which were non-binders to CD36 did not present with MA. The non-binder IRBC phenotype was found to be significantly associated with protection against MA (χ2 = 8.58, P < 0.05). Figure 4 presents mean Hb levels in which homozygous mutants had the highest Hb concentration of 11.46 g/dl. Conversely, children with the wild-type CD36 gene allele had the lowest Hb concentration of 6.1 g/dl, whereas heterozygous children had a mean Hb concentration of 8.36 g/dl. All the means were statistically different from one another by unpaired t-tests (P < 0.01).

Figure 3.

 Binding phenotypes of IRBCs isolated from different CD36 genotypes to CD36. The y-axis shows the number of anaemic cases from both binders and non-binder IRBCs to CD36. The binding phenotype was found to be statistically associated with occurrence of MA among children using the Fisher’s exact test (P < 0.01). MA, malarial anaemia, NA, non-anaemia.

Figure 4.

 Mean haemoglobin (Hb) levels among different genotypes of CD36. Bars represent the mean (mean ± SE) Hb level (g/dl) for each CD36 genotype. All means were statistically different from one another (P < 0.01).

Discussion and conclusions

The absence of data that intimately defines, at the molecular level, the host–parasite interface during infection with malaria parasites leaves a major gap in our understanding of the critical phenomena that lead to severe malaria, particularly in children. The only available information provides statistical associations between polymorphisms that occur in genes encoding host proteins used by malaria parasites as ligands of cytoadherence and severe malaria. Consistent with previous observations in Kenya and The Gambia (Newbold et al. 1997), results from this study show CD36 to be the major ligand for cytoadherence of field isolates of P. falciparum. In line with previous findings elsewhere (Yamamoto et al. 1990; Greenwalt et al. 1992; Lipsky et al. 1994; Ikeda 1999; Yasunaga et al. 2007), we have shown in our study that the prevalence of CD36 deficiency in Muheza, Tanzania is (2.9%).

A small proportion of IRBCs from CD36 deficient children bound to the purified CD36 protein. Binding to CD36 has been shown to be highest when IRBC used were from children with the wild-type CD36 allele. This association between CD36 genotypes and parasite adherence to CD36 in our study was found to be statistically significant. The association between IRBC binding phenotype and occurrence of MA provides a direct explanation that failure of adherence of IRBC to vascular endothelial cells may be responsible for reduced occurrence of MA. The findings in this study suggest that protection against MA in CD36 deficient children is most likely to be a result of changes (absence/reduction) in IRBC adherence to vascular endothelial CD36 receptors.

As CD36 deficient children do not express the CD36 receptor, it is likely that CD36 deficiency alters the ability of IRBCs to bind to CD36. A reduction/absence of in vivo IRBC adherence to CD36 may result into a reduction in the number of sequestered, agglutinating and rosetting IRBC and non-infected RBC, which would result to resistance to reduction of number of circulating erythrocytes due to parasite-induced hemolysis. This phenomenon is reflected as normal Hb level, and thus absence of anaemia. This is the first study that has linked polymorphisms in the CD36 gene, genotype-specific, IRBC CD36 binding profiles and MA.

Apart from acting as a receptor for P. falciparum-infected IRBC, CD36 serves as an important molecule in modulating host immunity. CD36 is expressed on endothelial cells, platelets, macrophages and dendritic cells and participates in phagocytosis and lipid metabolism (Rouabhia et al. 1994; Urban et al. 2001; Corcoran et al. 2002; Drover & Abumrad 2005; Alexander et al. 2006; Nassir et al. 2007; Harasim et al. 2008) all of which are crucial processes of life. This may partly explain the mechanisms by which a polymorphism that protects against a severe malarial syndrome (MA) is kept at a relatively low and stable frequency in the study population and elsewhere. The results of this study focus on the contribution of CD36 polymorphisms to the development of MA in children. The deficiency of CD36 on immune cells is most likely to interfere with immune processes, including phagocytosis of ageing RBCs, which may have significant implications in terms of Hb levels, aberrant lipid metabolism, predisposition to atherosclerosis insulin intolerance and many other fatal conditions (Furuhashi et al. 2003; Glintborg et al. 2008; Harasim et al. 2008).

A delicate equilibrium is therefore likely to exist between the protective role of CD36 deficiency and reduced CD36 adherence against MA and its predisposition to other equally fatal diseases. This observation calls for future studies to better explain how CD36 deficiency may influence host immunity, and ways in which such deficiency modulates the clinical outcomes of fatal syndromes.


The authors acknowledge the support of the MOMS Clinical Research Laboratories at Muheza Designated District Hospital, Muheza Tanzania for providing space for this work. The Seattle Biomedical Research Institute (USA) is acknowledged for providing financial and technical assistance. The authors also thank the mothers who by consent, decided to participate in our study in virtue of its outcome and future use.