Malaria continues to be a major global public health burden, causing 250 million clinical cases and nearly one million deaths annually, of which 85% are children under the age of five. About 86% of malaria cases and over 91% of malaria deaths worldwide occur in sub-Saharan Africa [1,2]. Cerebral malaria and severe malarial anaemia (SMA) are two major syndromes causing malaria-related mortality in children. The pattern of these two severe forms varies depending on the intensity of transmission: cerebral malaria is more common in older children in areas with lower intensity of transmission, whereas SMA is often seen in children below two years of age in areas with intense transmission [3], such as western Kenya, where Plasmodium falciparum infection is most common [4].

The human host response to P. falciparum can vary. While some infected individuals die of severe malaria, others survive, and still others are infected without becoming severely ill. It has been estimated that host genetic factors account for ~25% of the risk of severe malaria [5]. Malaria has exerted significant pressure on the human genome to select for mutations conferring refractoriness to malaria [6]. This is illustrated by several well studied mutations that protect against different forms of malaria, including sickle cell allele, Duffy mutations, and alpha thalassaemia [7]. An increasing amount of published information has documented a number of genetic associations with malaria susceptibility. The majority of these concern genes coding for molecules involved in physiology of red blood cells or in the host immune response[8]. It is only beginning to understand the human genetic factors that determine resistance to malaria.

Malarial anaemia is characterized by destruction of malaria infected red blood cells and suppression of erythropoiesis [9-11]. Although the mechanisms of erythropoietic suppression associated with pathogenesis of malarial anaemia are poorly understood, studies in murine models of malarial anaemia have demonstrated that interleukin 12 (IL12) significantly boosts erythropoietic responses by enhancement of erythroid progenitor expansion and immune regulation [12-15], and administration of IL12 corrects malarial anaemia [12,16,17]. Moreover, several immuno-epidemiological studies conducted in malaria endemic areas have shown that decreased IL12 levels are associated with an increased risk of SMA in children [18-20]. Taken together, these findings suggest that reductions of IL12 in malaria infection could lead to SMA by suppression of erythropoiesis.

IL12 is a pro-inflammatory cytokine with a pleiotropic effect as a potent immunoregulatory molecule and haematopoietic growth factor in infections with Plasmodium parasites [14]. It is a disulfide-linked heterodimer composed of a 35 kD subunit encoded by IL12A and 40 kD subunit encoded by IL12B and exerts its biological effects through binding to heterodimeric receptors encoded by IL12RB1 and IL12RB2. The region containing IL12B, chromosome 5q31-q33, is linked to P. falciparum parasite density in linkage analysis studies [21,22]. A functional promoter variant in IL12B, IL12Bpro (rs17860508) is associated with protection against cerebral malaria in children [23]. These results suggest that IL12 is an important candidate gene in host resistance to severe malaria.

Given the biological relevance of IL12 in malarial anaemia and the genetic linkage of IL12B in malaria infection and severe disease (cerebral malaria), the potential associations of differences in genes encoding two human IL12 subunits, IL12A and IL12B, and receptors, IL12RB1 and IL12RB2, in the risk of developing SMA has been investigated by retrospectively genotyping samples collected from an infant cohort study, Asembo Bay Cohort Project (ABCP) [24]. This study presents findings of the first significant genetic association within the IL12A and IL12RB1, which are associated with protection against SMA among young children in western Kenya.

In this paper, significant associations between polymorphisms within genes encoding the IL12A and IL12RB1 and protection against SMA are identified. The data shows that a SNP located within the 7^th intron boundary of the IL12RB1 gene (rs429774) is associated with protection against developing SMA. In particular, children possessing two minor alleles (C) had an 82% decreased risk of SMA compared to those with one or two copies of the major allele (T); this remained significant after correction for multiple testing. The decreased risk for developing SMA is also associated with increasing minor allele (C) copy number in a SNP located within the downstream of the IL12A gene (rs2243140). Another two SNPs in IL12A (rs22431348) and IL12RB1 (rs383483) were associated with reduced risk of SA and SAP and with increased risk HDP, respectively. There is no evidence to support any significant association between IL12B genetic variants and SMA or the 3 other clinical phenotypes tested in this study. This suggests that common genetic variations within IL12B did not influence the severity of malaria in this Kenyan population, despite previous reported associations of IL12Bpro (rs17860508) with cerebral malaria in Tanzanian and Malian, but not Kenyan children [23,34], and of several other diseases with variation in IL12Bpro (rs17860508) and IL12B 3' UTR (rs3212227) [35]. A lack of significance in IL12B in this study may be the result of lack of power or small sample size; however, a recent study conducted in western African population also showed that there was no association between these two known cis-regulatory element polymophisms in IL12B and malaria parasitaemia [36].

This study was conducted in a large and clinically well-characterized cohort of children using a tagSNP method based on the following considerations. First, although methods have been developed to predict the functional impact of variants such as nonsynonymous coding SNPs, which introduce amino acid changes in their corresponding proteins [35], these SNPs represent a small subset of all common variants. Second, several genome wide association studies have indicated that variations in non-coding regions rather than nonsynonymous coding SNPs are likely to be causative to disease susceptibility or resistance in most instances [37]. Regulation of the protein products, rather than differences in the structure or function of the protein may be most important for disease risk [38]. Third, many SNPs show correlated genotypes, or LD, suggesting that a subset of all SNPs (known as tagSNPs) can be genotyped for disease association studies [39]. Lastly, several tagSNP methods have been developed to select a subset of SNPs that efficiently describes existing patterns of variation in candidate gene regions for subsequent genotyping and association analysis, without making assumption about the potential functional impact of each SNP [40]. An efficient set of tagSNPs to cover each candidate gene region, rather than simply focusing on a few putatively interesting SNPs have been selected, such that genotype at untyped SNPs can be accurately inferred from genotype at a tagSNP [28].

Previous immunological studies have shown that IL12 exerts its biological function via binding to its receptors. This study has identified associations of polymorphisms not only in IL12 related gene, IL12A, but also in its receptor, IL12RB1 with SMA and HDP with different effects. Two SNPs in the IL12A (rs2243140 and rs2243138) and one SNP in the IL12RB1 (rs429774) were associated with protection against SMA or SA whereas one SNP in the IL12RB1 (rs383483) was associated with susceptibility to HDP. There are a number of studies that examined the effect of IL12 on erythropoiesis and malarial anaemia in mouse models, in which IL12 was found to be extremely effective in correcting malarial anaemia [12-14]. The dramatic effect of IL12 treatment is mediated by enhancement of erythroid progenitor expansion in bone marrow and by immune regulation of IFN-γ, TNF-α and NO [16]. In humans, a decrease in IL12 production is also clearly associated with SMA [18-20]. Although it remains unknown how the noncoding polymorphic SNPs in IL12A and IL12RB1 identified in this study control or regulate the expression of IL12 and its receptors, the results from this study support the assertion that non-coding regions rather than nonsynonymous coding SNPs potentially have effects on disease outcomes [39]. The results from this study also suggest that the effects of IL12A and IL12RB1 on malaria disease outcomes most likely result from the long evolutionary history of P falciparum parasite within the human population.

The IL23 pathway may offer further insights on the genetic results from this study since IL23 and IL12 share the IL12p40 subunit encoded by IL12B and IL23 also binds to the IL12 receptor subunit encoded by IL12RB1. A recent study has shown that children with malarial anaemia had increased circulating IL23 and IL10 and decreased IL12 relative to healthy controls in Kenyan children, suggesting that IL23 and IL12 are counter-regulators in malaria anaemia [20]. The findings from this study that polymorphisms (3 SNPs) in IL12A and IL12RB1, but not in IL12B, have protective effects on SMA and one SNP in IL12RB1 has an opposite effect on HDP are interesting. It is possible that the genetic effects on disease outcomes observed in this study may occur at the phenotypic level of IL12 and IL23 systems, which may be influenced by the interaction of genetic variants in IL12A, IL12RB1 and IL12B. This hypothesis requires testing. Currently, whether the polymorphisms in IL12A and IL12RB1 observed in this study influence the level and expression of IL12, IL12 receptor and IL23 in vitro in relation to severity of malaria anaemia have being examined in Kenyan children.

This study included a large sample size and used longitudinal data for association analysis. However, it will be important to replicate these findings in additional African populations [41]. These efforts are complicated by the patterns of disease seen in regions of differing malaria transmission, but these studies are needed to allow us to further understand the most important pathways in successfully controlling infection by the malaria parasite and to assess the effects of racial differences, transmission intensity, and other environmental factors on the interaction of IL12 related gene and severe malaria anaemia. In addition, studies are needed to examine whether the polymorphisms observed in these genes influence the level and expression of IL12, IL12 receptor and IL23 in vitro in other African populations in relation to malarial anaemia.

This study has shown strong associations between polymorphisms in the genes of IL12A and IL12RB1 and protection from SMA in Kenyan children, suggesting that human genetic variants of IL12 related genes may significantly contribute to the development of anaemia in malaria patients.

The authors declare that they have no competing interests.

LZ was responsible for tagSNP selection, carried out genotyping work and genetic data analysis, and wrote manuscript. DP conducted genotyping and genetic data analysis. JVE and SC performed statistical analysis. SK, FTK and DT implemented Asembo Bay Cohort Project including collection of clinical and epidemiological data and laboratory diagnosis. BN and AAL were the PIs for Asembo Bay Cohort Project and designed ABCP. LS and VU participated in design of the genetic association study and contributed to data interpretation. YPS was the PI for this genetic association study, designed the study and wrote manuscript. All authors read and approved the final manuscript.