Polymorphisms in the RNASE3 gene are associated with susceptibility to cerebral malaria in Ghanaian children

Abstract

Background: Cerebral malaria (CM) is the most severe outcome of Plasmodium falciparum infection and a major cause of death in children from 2 to 4 years of age. A hospital based study in Ghana showed that P. falciparum induces eosinophilia and found a significantly higher serum level of eosinophil cationic protein (ECP) in CM patients than in uncomplicated malaria (UM) and severe malaria anemia (SA) patients. Single nucleotide polymorphisms (SNPs) have been described in the ECP encoding-gene (RNASE3) of which the c.371G>C polymorphism (rs2073342) results in an arginine to threonine amino acid substitution p.R124T in the polypeptide and abolishes the cytotoxicity of ECP. The present study aimed to investigate the potential association between polymorphisms in RNASE3 and CM.

Methodology/principal findings: The RNASE3 gene and flanking regions were sequenced in 206 Ghanaian children enrolled in a hospital based malaria study. An association study was carried out to assess the significance of five SNPs in CM (n=45) and SA (n=56) cases, respectively. The two severe case groups (CM and SA) were compared with the non-severe control group comprising children suffering from UM (n=105). The 371G allele was significantly associated with CM (p=0.00945, OR=2.29, 95% CI=1.22-4.32) but not with SA. Linkage disequilibrium analysis demonstrated significant linkage between three SNPs and the haplotype combination 371G/*16G/*94A was strongly associated with susceptibility to CM (p=0.000913, OR=4.14, 95% CI=1.79-9.56), thus, defining a risk haplotype. The RNASE3 371GG genotype was found to be under frequency-dependent selection.

Conclusions/significance: The 371G allele of RNASE3 is associated with susceptibility to CM and forms part of a risk associated haplotype GGA defined by the markers: rs2073342 (G-allele), rs2233860 (G-allele) and rs8019343 (A-allele) respectively. Collectively, these results suggest a hitherto unrecognized role for eosinophils in CM pathogenesis.

Figure 1. Location of the studied SNPs in RNASE3, linkage disequilibrium (LD) patterns and haplotype association analysis.

A) A shematic of the region of the RNASE 3 gene (NM_002935.2) and LD plot of the respective SNPs visualised using Haploview v4.2. At the top of the LD plot, the two exons are shown with the intron joining them. Untranslated regions of the exons are indicated by shadings. The LD plot shows pairwise r² values (×100) given in the squares for each comparison between the SNPs. White squares represent r² values equal to 0. Different shades of grey represent r² values between 0 and 1. B) Haplotype associations with susceptibility to cerebral malaria (CM) compared to uncomplicated malaria (UM). Odds ratio (OR) and 95% confidence intervals (CI) were determined using multivariate logistic regression controlling for age and gender. The reference groups in the multivariate logistic regression analyses were those without the respective haplotypes. Haplotype 3 (GGA) was significanlty associated with susceptibility to CM.

Figure 2. Association of c.371G>C genotypes with coma score in the entire patient population studied.

Coma score distribution among individuals with 371 (GG, GC and CC) genotypes is shown. Comparisons that yielded statistical significance are indicated with horizontal lines linking the respective groups at the top of the plots with the p – value stated. P-values were determined by Mann Whitney test and horizontal lines within plots represent the median of the distribution.

Figure 3. RNASE3 haplotypes, c.371G>C genotypes and alleles distribution in malaria endemic (Ghanaian) versus non-malaria (Danish) populations.

A) Distribution of haplotypes defined by the three SNPs (rs2073342, rs2233860 and rs8019343) as from block 1 in Figure 1A in the two populations. B) c.371G>C genotype distribution in all Ghanaian subjects compared to Danes. C) c.371G>C allelic distribution in the two populations.

Figure 4. Tajima's D index .

Sliding window analysis of Tajima's D index across the Ghanaian (A) and Danish (B) RNASE3 sequences. Bar at the top shows the sequence region analyzed (nucleotides 1–1183) and the positions of the SNP. Window sizes used were 100 bp with a 25-bp step size. Windows that gave a significant D value p<0.10), are indicated above the relevant window midpoint by a hash (#).