G6PD gene variants and its association with malaria in a Sri Lankan population

Glucose-6-phosphate dehydrogenase (G6PD) is an important enzyme present in all cells that catalyzes the first step of the hexose-monophosphate pathway producing NADPH. Cells require NADPH for many reactions in biosynthetic pathways including maintenance of the effective redox potential that helps to protect red cell membranes against oxidative stress and injury [1]. Deficiencies in G6PD activity give rise to the most common enzymopathy known affecting over 400 million people worldwide. Of the various clinical manifestations that occur due to G6PD deficiency, the most common include neonatal jaundice and acute haemolytic anaemia [1]. Certain drugs (e.g. primaquine), infections or foods (particularly fava beans) can trigger this phenomenon. The resulting haemolysis is believed to be the result of the inability of the G6PD-deficient red cells to withstand the oxidative damage produced either directly or indirectly by these triggering agents.

The G6PD enzyme has over 400 different biochemical variants as reviewed by Beutler and Vuilliamy [2]. The coding gene is located on the X chromosome and about 186 genetic variants, which can induce functional enzyme deficiency in the host have been described [3]. Many G6PD gene variants are named after locations from where they were first discovered; for example, Canton, Mahidol, Kaiping and Viangchang (which have been identified as the most prevalent in Asian populations [4]) and G6PD Mediterranean, Orissa, Kalyan-Kerala, Coimbra (more common Indian variants [3,5,6]). Mutations in this gene are of practical relevance to malaria-infected individuals for two main reasons; On one hand G6PD-deficiency is believed to be associated with susceptibility or resistance to malarial disease while on the other hand, certain oxidative anti-malarial drugs (i.e. primaquine and sulphur containing drugs) can trigger haemolytic anaemia [7-9], which is of practical concern to all malaria endemic countries including Sri Lanka.

Although a few studies have been conducted in Sri Lanka to describe the correlation between drug- induced haemolysis and the functional G6PD enzyme deficiency status in mid 1960s [10,11], there is a void due to lack of recent published data, and no information exists on either the genetic variants associated with this condition or its association with malaria in this country. This study focuses on G6PD genetic variants that prevail in a selected population in South-Eastern Sri Lanka (residents of a known endemic area for both Plasmodium vivax and Plasmodium falciparum malaria) and their possible association with susceptibility or protection against uncomplicated malarial disease.

This study was conducted to identify variants within the G6PD gene present in a population residing in a malaria-endemic area in southern Sri Lanka. Although there had been a few studies described in the past in Sri Lanka on G6PD enzyme functional deficiency [10,11], there are no studies, as far as it is known to identify genetic polymorphisms and nucleotide variability of the G6PD gene and no information exists on its association with malaria. The 57 SNPs that were studied in the G6PD gene are known coding variants and/or previously identified through association studies with G6PD deficiency per se or with malaria, and all selected markers made valid Sequenom assays.

Although some G6PD gene variants including G6PD Mahidol variant (rs137852314), Mediterranean variant (rs5630868), Canton variant (rs72554665) and Viangchan variant (rs13785237), G6PD Kaiping (rs72554664) and G6PD Vanua Lava (rs78365220) have been established as those that prevail in Asian populations, none of these SNPs were found to be polymorphic in the Sri Lankan population studied. Similar findings with regard to the Mediterranean variant has been reported previously on a limited number of samples tested from a different location within Sri Lanka [24].

Of the nine markers with MAF > 10%, 4 SNPs i.e. rs915942, rs2230037, rs5986877 and rs7053878 the alternate allele was the minor allele. These studies are in line with data available online on Asian populations [25]. However, it is interesting to note that the allele frequencies of the alternate alleles of rs2230037 and rs5986877 are much higher when compared to a general Asian population as reported previously (Table 4). Moreover, the minor allele for rs766420 in the Sri Lankan population was the ancestral allele (C) whereas the minor allele in the general Asian population was the alternate allele (G). Furthermore, the minor allele frequencies of rs2071429, rs4898389 and rs7879049 (for which the minor allele is the ancestral allele) are higher when compared to the minor allele frequencies of these markers in the general Asian population.

These differences might be due to the fact that the genetic variants of Han-Chinese or Japanese population (which are different from the South-Asian or Sri Lankan populations) may have contributed largely to the frequencies calculated for the general Asian populations [25]. The variants G6PD Punjab, Kalyan-Kerala although identified as common in neighbouring Indian populations were not typed as they could not be designed as SQNM assays.

Of the 57 SNPs genotyped, nine (of which two are in the G6PD gene) had frequencies greater than 10% in this Sri Lankan population. The observations made regarding the males who carried the alternate alleles for rs2071429 and rs2230037 being more susceptible for malaria (Table 2) suggest a gender-biased effect for the mutant type of these SNPs. However homozygous or heterozygous females with the alternate allele of these SNPs did not show any association with malaria disease status.

Population studies on relationships between genetic polymorphisms in the G6PD gene and phenotypic characteristics in uncomplicated malaria are rare. Louicharoen et al. [9] observed that the G6PD Mahidol variant (found in a Thai population) was associated with a significant reduction in P. vivax parasite density both in heterozygous and homozygous females as well as in hemizygous males, whereas no association was observed with P. falciparum parasite density. This study however, looked at the G6PD Mahidol variant only, whereas the current study analyzed several G6PD variants. The degree of association (protection/susceptibility) with the disease status might differ according to the type of mutation and of course from population to population. A similar case–control study done in Afghanistan by Leslie et al. [26] has also shown that the association between G6PD gene mutations and protection against vivax malaria is seen less often in heterozygous females when compared to hemizygous males. The current study confirms the observations made by Leslie et al. [26] that either protection and/or susceptibility to the disease are enhanced by mutations in G6PD gene in hemizygous males rather than in heterozygous females, indicating that mono-allelic status of X-linked markers might have a correlation with the disease status rather than the bi-allelic status of the markers in females. Furthermore, the results of the present study indicate that it is not enhanced in homozygous females either. Random inactivation of one of the two inherited X chromosomes does occur in females. Therefore, although genotyped as homozygote with the ancestral or the alternate allele the actual proportion of females that express the gene activity might be much lower, therefore, the relationship between the marker and the disease status may not be evident.

Bienzle et al. [27] had reported lower parasite densities (as an indication of disease severity) in two groups of G6PD non-deficient male individuals with different variants of G6PD enzyme activity. The current study reveals male cases carrying the ancestral allele of rs915942 had significantly lower parasite densities for P. falciparum when compared to males carrying the alternate allele. However, since G6PD enzyme functional assays were not carried out, it is not possible to comment on any possible associations with the functional status of this enzyme in the study subjects.

Reports on comparisons between G6PD gene mutations and disease status of severe and complicated malaria are abundant. A large-scale study done in Kenya by Ruwende et al. [7] revealed that mutations in G6PD are associated with reduction in the risk of severe malaria for both hemizygous males and heterozygous females. However, in two studies done in Mali and Pakistan there was no such association between heterozygous females and severe malaria which is believed to be due to the control group being individuals with ‘uncomplicated malaria’ compared to the cases of severe and complicated malaria [8,26,28]. In the current study, the study subjects who had one or more malaria infections during the study period in 1992–1993 (group A) were compared against the individuals (group B) who at the point of data collection (1992–1993) have never acquired malaria during the lifetime according to the records maintained in the field site and subjects’ history based on his or her memory. Therefore, there may have been inaccuracies in the categorization of “protected individuals” and “susceptible individuals” that may have reduced the power of evidence towards the relationship between the mutant genetic markers and malaria. Guindo et al. [8] observed that male individuals with mutations in G6PD (G6PD A- mutation at nucleotide position 202) were protected from severe malaria. This protection was found only in males but not in heterozygous females. More recently, a large multi-centre cases-control study of severe malaria involving 12,000 cases and 17,000 controls demonstrated a highly significant association of G6PD-202 with risk of severe malarial anaemia while a small but significant protection with cerebral malaria [19]. This not only supports the notion that balancing selection occurs in G6PD, but that the relationship of G6PD with disease status is very complex. Moreover a recently study of genetic structure with enzyme function showed a possible interaction between G6PD polymorphisms modifying the outcome [18].

It is, therefore, important to note that there is a major impact of natural selection that occur in malaria hyper-endemic areas, such as Africa, where the numbers of G6PD-deficient homozygous females in populations are very rare [8]. However, this is unlikely to have happened in this study population considering the low and seasonal malaria transmission that prevailed in Sri Lanka, from as far back as it is known. This very fact and the genetic variations between the African and Asian populations may account for the differences observed in this study.

Very high linkage disequilibrium was seen between seven of the nine SNPs with MAF > 10%. Complete linkage disequilibrium could be observed between rs915941 and rs915942, and a similar scenario has been observed in other populations as well (e.g. Gujarati Indians and Han Chinese populations). This could be perhaps due to the very close chromosome locations; 153626649, 153626738 respectively [25]. Unusually high levels of linkage disequilibrium among the markers within G6PD gene has also been observed by Saunders et al. [29]. Furthermore, as many of these tested SNPs are in linkage disequilibrium it could be assumed that there might be a combined effect of these host genetic markers towards malaria disease status. When multiple markers in linkage disequilibrium are studied to assess their association with disease traits, haplotype analysis may be more efficient and accurate than separate analyses of the genetic markers. With this premise and with the finding that rs925941 and rs915942 (Block 1) and rs2071429, rs2230037, rs4898389, rs5986877 and rs7879049 (Block 2) fall into the same haplotype blocks are likely to indicate a possible collective effect of these SNPs over malaria infection.

Single Nucleotide Polymorphisms within the G6PD gene are described in this study population from Sri Lanka. This is the first time that such a dense set of G6PD polymorphisms have been investigated and identified in a Sri Lankan population. Furthermore, it was evident that the common mutations previously described in the Asian or South Asian region are absent or rare in the local population. In light of this further studies to investigate the full spectrum of mutations in Sri Lanka population would help not only understand more on the ancestry of this gene in Sri Lanka but also aid the functional significance of these observed polymorphisms. The study also provides evidence for likely associations of these variations either individually and/or collectively with traits of uncomplicated malaria. Further investigations are underway to combine detailed genetic analyses with functional enzyme studies, which are likely to reveal more information.