Background
The haemoglobin S mutation (
HbS) is very well recognized as the human single gene conferring the strongest advantage against falciparum malaria [
1,
2]. Evidence from
HbS genetic association studies in African countries with malaria endemicity levels, helped to settle the notion that epidemiology of sickle-cell trait results from the selective pressure imposed on the human genome by
Plasmodium infection [
3‐
8]. Multiple biochemical and immune mechanisms have been suggested to explain
HbS protection against
Plasmodium falciparum pathogenesis and morbidity. It is proposed that such mechanisms act concurrently to improve parasitaemia control and protect from uncomplicated and severe malaria [
9].
In Angola, sickle-cell disease is a critical public health concern and is considered a main contributor to the high mortality rate in children [
10,
11]. However, there are few reports on
HbS prevalence in Angolan populations [
12‐
14]. Malaria is one of the leading causes of morbidity and mortality in Angola, mainly among preschool children, and it is estimated that
P. falciparum causes 90 % of all malaria infections [
11,
15‐
17].
This study aimed to analyse the role of HbS mutation in malaria protection in Angolan children, according to disease severity and degree of infection. The results of this stratified analysis suggest that HbS carriers exhibiting severe malaria syndromes are significantly protected against hyperparasitaemia.
Methods
Place of study and ethical permission
Luanda is an endemic-malaria province with high level of transmission [
18]. Hospital Pediátrico David Bernardino (HPDB) is a tertiary national reference paediatric hospital in Luanda, Angola. Ethical permission for this study was granted by the Ethics Committee of the HPDB in Luanda that was appointed by the Angolan Ministry of Health.
Subjects
A total of 749 children, living in Luanda and ranging from 6 months to 13 years of age, were enrolled in the present study. Cases and controls were selected among attendance to the HPDB. The sample collection was carried out from February 2005 to May 2007 and comprised 288 severe malaria children (130 with cerebral malaria and 158 with severe malaria but not cerebral malaria), 142 patients with uncomplicated malaria and 319 uninfected controls. Samples were also collected from mothers of severe malaria children and comprised 226 mother–child pairs. Mothers were enrolled only after written, informed consent and children were enrolled only after written, informed consent from their parents or guardians.
Phenotypic and inclusion criteria
Malaria was diagnosed on the basis of a positive asexual parasitaemia detected by a single reader on a Giemsa-stained thick film. For parasitaemia quantification 100 high-power microscopic fields were observed. The number of asexual parasites and white blood cells (WBCs) were counted in each field until the number of WBCs reached 200 and the parasite density was calculated from this value [
19]. To confirm and select for
P. falciparum infection, children with mixed infections, as ascertained by malaria species-specific nested-PCR in peripheral blood DNA [
20], were excluded from the study. Triage and clinical examinations of patients with severe disease included a clinical history collected upon admission to the paediatrics urgency ward, followed by detailed clinical examination as described below.
Clinical criteria
The diagnosis of cerebral malaria followed the criteria of the WHO definition, strictly for research purposes, valuing unarousable coma, which was defined by the inability to localize painful stimuli in absence of other causes of encephalopathy [
21]. For the purposes of this study, whenever coma has been preceded by a seizure the assessment of coma was made 1 h after the end of the seizure. This procedure is meant to exclude transitional post-convulsive coma.
The cerebral malaria patients had to meet all the following criteria: (1) coma score <3 in the Blantyre Scale, for children younger than 60 months or score <7 on the Glasgow Scale, for children aged equal or more than 60 months; (2) absence of diagnostic criteria for any other possible cause of encephalopathy, including hypoglycaemia (venous blood glucose below 40 mg/dL); (3) exclusion of meningitis-cerebrospinal fluid without pleocytosis (up to 8 lymphocytes/cu mm) nor hypoglycorrachia (determination of glucose in cerebrospinal fluid >50 mg/dL or equal to 50 % of blood glucose, never <40 mg/dL).
Severe malaria anaemia (SMA) was defined as haemoglobin <5 g/dL or haematocrit <15 % and hyperparasitaemia (HP) required a value of parasitaemia >100 parasitized erythrocytes per microscopic field (magnification 1000×). The presence of any diagnosis or any other neurological sign, including seizures, installed during the course of the disease excluded the child from the study.
Uncomplicated malaria patients had no clinical manifestations suggestive of complications of malaria, as well as any other pathology that could explain a febrile syndrome.
Controls were uninfected children randomly selected in the vaccination ward of the HPDB. Children in this group had no symptom of disease and were negative in a PCR test for Plasmodium DNA performed in peripheral blood.
Blood DNA preparation
DNA was extracted from peripheral blood using the Chemagen Magnetic Bead Technology in an automated nucleic acid isolation station. DNA preparations were quantified using PicoGreen reagents (Invitrogen®) according to the supplier instructions. The dbSNP, rs334, that defines the HbS mutation, was genotyped with the Mass Array system to design multiplex reactions for PCR and iPlex primer extension (Sequenom) and the MALDI-TOF based Mass Array genotype platform (Sequenom). Genotyping rate was higher than 90 %.
Phenotypic data analysis
The statistical analysis for sample characterization was performed with SPSS version 15.0. The data that were not normally distributed were analysed using non-parametric methods (Mann–Whitney U and Kruskal–Wallis) and Pearson Chi square was used to study the association between qualitative variables. A threshold for statistical significance was P < 0.05; odds ratios (OR) and 95 % confidence interval were calculated to measure the magnitude of association.
Logistic regression was performed with parasitaemia as a dependent variable dichotomized as <100 and ≥100 parasitized erythrocyte/microscopic field using diagnosis and age group in bivariate analysis.
Genetic analysis
Hardy–Weinberg equilibrium requirements (
P > 0.05) were met in uninfected controls. Case–control association analysis was performed with the logistic regression model implemented in the SNPassoc package for R software (version 2.7.0) using allelic and genotypic frequencies where dichotomized disease outcomes were analysed as variables dependent on presence of allele T or presence of genotype AT. Genotypic analysis used the dominant and additive genetic models which are different from the actual underlying mode of inheritance of the minor allele [
22]. The significance level of the likelihood ratio test
P < 0.05 was considered as suggestive evidence for association. TRANSMIT software that allows TDT testing when phase is unknown was used [
23] to analyse
HbS transmission in mother–child pairs. Quantitative trait analysis (QTL) of parasite densities was performed with the program Plink (version 1.06) that calculates the level of significance, either by using the asymptotic model, or the empirical model [
24]. QTL analysis considered the additive linear model where the parasite density is analysed as a continuous variable dependent on the
HbS genotype status (AA, AT and TT).
Discussion
The mechanism of
HbS protection against
P. falciparum morbidity remains controversial [
25‐
27]. Studying a sample of Angolan children with different malaria syndromes collected evidence showed that
HbS conferred specific protection against hyperparasitaemia, specifically in the context of severe malaria. In this study population, a significant number of malaria patients (42 %) exhibited concomitantly different malaria clinical complications. These overlapping malaria syndromes corroborate the notion that severe malaria cases often represent clinical syndromes rather than clinical isolated entities [
28]. In this context, studying
HbS protection against one clinical condition faces possible confounding by co-occurrence of other infection parameters. Strikingly, by contrasting allelic and genotypic
HbS frequencies in severe malaria patients stratified either for parasitaemia values or by other causes of clinical severe malaria (namely, cerebral malaria and/or severe malaria anaemia), it was found that protection against severe malaria was at a large extent attributable to resistance to hyperparasitaemia. Although with limitations due to the reduced number of
HbS homozygotes the analysis supports the notion that control of parasite density was quantitatively dependent on the number of T alleles.
Furthermore,
HbS showed to control quantitatively parasitaemia in clinical malaria both in presence or absence of uncomplicated malaria cases. This result suggests that
HbS contributes to resistance to expansion of blood stage parasite irrespective of malaria clinical presentation. Indeed, these results sustain the debate around the effect of
HbS protection against hyperparasitaemia and its mechanisms. An experimental cerebral malaria (ECM) study in mice expressing sickle haemoglobin concluded that the protective effect exerted against lethal ECM was irrespective of parasite load [
25]. Meanwhile, a human study evidenced that median parasite densities were significantly higher in
APOE ε4 children for
Plasmodium spp. densities compared to non-
APOE ε4 children and concluded a suggestive epistatic interaction between
APOE and
HbS genes such that sickle cell trait only had an effect on parasite density in
APOE ε4 children [
27].
The predominant occurrence of severe malaria in children under 5 years of age in the study sample is in line with what is reported in other areas of stable malaria [
29,
30]. It has been argued, that lower frequency of severe malaria as well as decreased parasitaemia at both extremes of the age distribution could reflect, on one hand, the protective role of maternal immunity in children under 18 months and, on the other hand, the acquisition of immunity with age in children older than 9 years [
31,
32]. Nevertheless, it cannot be ruled out that uneven representation of hyperparasitaemia and severe malaria across the distribution age may have an impact on
HbS protection effects here identified.
The
HbS allele frequency in the uninfected control group was 15.1 %, similar to that of the HapMap Yoruba (Nigeria) sample where the frequency is 12.5 %, though
HbS frequency may exceed 20 % in other African populations [
33]. Case–control analysis demonstrated a stronger protective effect of the
HbS allele against severe malaria syndromes (OR = 0.15) as compared to uncomplicated malaria (OR = 0.50). Studies in Kenya reported similar genetic effects on protection against uncomplicated malaria (OR = 0.50) and against severe malaria (OR = 0.17) [
6]. This strong protection against severe malaria conferred by the heterozygous HbAS was also clearly confirmed with a case–control study of 2591 severe falciparum malaria children enrolled at a tertiary referral center in Ghana (OR = 0.08) [
34]. Another study that followed a cohort of 1070 children in Ghana showed protection against high parasitaemia in uncomplicated malaria [
35]. It is worth mentioning that similarly to this Angolan sample, the protective effect of the
HbS allele was evaluated in case–control supplemented with TDT [
36] and was not diluted in a genome-wide association study of severe malaria, despite the ethnic diversity of the Gambian population [
2].
The results of the present study reinforce the notion that parasitaemia levels are to be taken into account on evaluation of
HbS protection in severe malaria. Nevertheless, the mechanism of
HbS protection against hyperparasitaemia was not addressed in this study leaving open the possibility that
HbS may partially accelerate the process of immunity acquisition against
P. falciparum [
37].
Authors’ contributions
MRS, MJT, AC, and CPG conceived and designed the experiments; MRS, MJT and JC performed the experiments; MRS, MJT, JC, and CPG analysed the data; AC and CPG contributed reagents/materials/analysis tools; MRS, CPG and RF wrote the manuscript. All authors read and approved the final manuscript.