Human candidate gene polymorphisms and risk of severe malaria in children in Kilifi, Kenya: a case-control association study

Abstract

Background: Human genetic factors are important determinants of malaria risk. We investigated associations between multiple candidate polymorphisms-many related to the structure or function of red blood cells-and risk for severe Plasmodium falciparum malaria and its specific phenotypes, including cerebral malaria, severe malaria anaemia, and respiratory distress.

Methods: We did a case-control study in Kilifi County, Kenya. We recruited as cases children presenting with severe malaria to the high-dependency ward of Kilifi County Hospital. We included as controls infants born in the local community between Aug 1, 2006, and Sept 30, 2010, who were part of a genetics study. We tested for associations between a range of candidate malaria-protective genes and risk for severe malaria and its specific phenotypes. We used a permutation approach to account for multiple comparisons between polymorphisms and severe malaria. We judged p values less than 0·005 significant for the primary analysis of the association between candidate genes and severe malaria.

Findings: Between June 11, 1995, and June 12, 2008, 2244 children with severe malaria were recruited to the study, and 3949 infants were included as controls. Overall, 263 (12%) of 2244 children with severe malaria died in hospital, including 196 (16%) of 1233 with cerebral malaria. We investigated 121 polymorphisms in 70 candidate severe malaria-associated genes. We found significant associations between risk for severe malaria overall and polymorphisms in 15 genes or locations, of which most were related to red blood cells: ABO, ATP2B4, ARL14, CD40LG, FREM3, INPP4B, G6PD, HBA (both HBA1 and HBA2), HBB, IL10, LPHN2 (also known as ADGRL2), LOC727982, RPS6KL1, CAND1, and GNAS. Combined, these genetic associations accounted for 5·2% of the variance in risk for developing severe malaria among individuals in the general population. We confirmed established associations between severe malaria and sickle-cell trait (odds ratio [OR] 0·15, 95% CI 0·11-0·20; p=2·61 × 10^-58), blood group O (0·74, 0·66-0·82; p=6·26 × 10^-8), and -α^3·7-thalassaemia (0·83, 0·76-0·90; p=2·06 × 10^-6). We also found strong associations between overall risk of severe malaria and polymorphisms in both ATP2B4 (OR 0·76, 95% CI 0·63-0·92; p=0·001) and FREM3 (0·64, 0·53-0·79; p=3·18 × 10^-14). The association with FREM3 could be accounted for by linkage disequilibrium with a complex structural mutation within the glycophorin gene region (comprising GYPA, GYPB, and GYPE) that encodes for the rare Dantu blood group antigen. Heterozygosity for Dantu was associated with risk for severe malaria (OR 0·57, 95% CI 0·49-0·68; p=3·22 × 10^-11), as was homozygosity (0·26, 0·11-0·62; p=0·002).

Interpretation: Both ATP2B4 and the Dantu blood group antigen are associated with the structure and function of red blood cells. ATP2B4 codes for plasma membrane calcium-transporting ATPase 4 (the major calcium pump on red blood cells) and the glycophorins are ligands for parasites to invade red blood cells. Future work should aim at uncovering the mechanisms by which these polymorphisms can result in severe malaria protection and investigate the implications of these associations for wider health.

Funding: Wellcome Trust, UK Medical Research Council, European Union, and Foundation for the National Institutes of Health as part of the Bill & Melinda Gates Grand Challenges in Global Health Initiative.

Figure 1

Venn diagram showing the distribution of the three major phenotypes of severe malaria Numbers indicate the cases for each phenotype or co-expressed phenotypes, and percentages are the case-fatality. Other severe malaria cases are those without any of the three phenotypes shown in the main diagram.

Figure 2

Distribution of p values for association tests with severe malaria phenotypes The minimum p value (−log₁₀[p]) for every polymorphism across the models is shown. We adjusted p values for gender, ethnic group, and rs334 genotype. The dashed vertical line represents the significance threshold of p<0·005, which was determined by permutation testing. The polymorphisms tested are listed in chromosomal order (appendix pp 2–4).