Introduction

Glucose-6-phosphate dehydrogenase (G6PD) deficiency is the most common genetic disorder in humans. It is prevalent throughout Africa, Asia, Southeast Asia and parts of South America, where malaria is endemic [1]. Different mutations in the gene encoding for G6PD cause varying degrees of protein instability that renders the mature red blood cells more susceptible to oxidative damage. Antimalarial drugs such as the 8-aminoquinolines (primaquine, tafenoquine) cause dose-dependent damage to G6PD deficient red blood cells resulting in intravascular haemolysis and removal of damaged erythrocytes by the spleen. The 8-aminoquinoline drugs are essential for the radical cure of Plasmodium vivax infections and for blocking the transmission of Plasmodium falciparum. The degree of haemolysis associated with the use of these drugs has been analyzed in few genetic variants and very limited data exist in most populations on either the hemolytic risk associated with a specific drug regimen in subjects with a specific G6PD variant or on the interaction with other genetic and clinical factors. Mutations in the G6PD gene located on the X chromosome usually result in reduced enzyme stability (and thus reduced activity in older erythrocytes) in all red cells in hemizygous males and homozygous females. There is variable phenotypic expression in heterozygous females depending on X-inactivation patterns [2]. As with other sex-linked disorders, phenotypic screening of male subjects is the preferred approach to assess G6PD deficiency prevalence and allelic frequency in the population.

The physico-chemical properties of the G6PD protein can be characterized by different techniques (typically electrophoretic mobility and K_m) and by its enzymatic capacity to reduce nicotinamide adenine dinucleotide phosphate (NADP) to NADPH in the presence of substrates (usually called enzymatic activity). The spectrophotometric assay [3] is considered the “gold standard” for quantitative assessment of enzymatic activity in red blood cells. Simple qualitative tests, such as the Fluorescent Spot Test (FST) [4], are used for population screening but they usually require some basic equipment, electricity and a functioning cold chain for storage of reagents. Point-of-Care (POC) tests have been used in validation studies [5] but have not shown a high sensitivity for the detection of deficient subjects.

In Southeast Asia, the prevalence of G6PD deficiency varies widely among ethnic groups. In Cambodia and Laos the Viangchan variant is the most common among the non-Chinese populations. Prevalence in Laos is approximately 7% and in Cambodia ranges between 1% and 20% depending on ethnicity [6] [7]. Similarly, in Vietnam, G6PD deficiency prevalence varies widely between ethnic groups from less than 3% in the Kinh [8] to 11% in the S'Tieng [9]. In Thailand, the prevalence of G6PD mutations is as high as 20%, the Viangchan variant predominates in the East and the Mahidol variant predominates in the West [10] [11]. In studies performed in Myanmar, the Mahidol variant is the most common variant (>90% of the affected population) with a prevalence between 4% and 12% in the general Burman population [12]. In two ethnic groups from Myanmar that migrated to southern Thailand, G6PD deficiency is caused mainly by the Mahidol variant and no Viangchan variant was detected [13] [14].

Various ethnic groups live in northwestern Thailand along the Myanmar border where malaria is endemic, and although G6PD deficiency is common, no phenotypic or molecular characterization of G6PD variants affecting this population has been performed. In a geographic area where there will be increasing use of primaquine as a gametocytocide and wide scale primaquine deployment for P. vivax malaria elimination, we have carried out a population survey in order to estimate the frequency and distribution of G6PD deficiency among 504 unrelated male subjects of mainly Karen and Burman ethnicity

Methods

Results

A total of 514 unrelated males who attended the five SMRU clinics along the Thai-Myanmar border were included in the study between September and November 2011. The mean (SD) age was 27.0 (13.0) years and the main ethnic groups were Karen (69.8%), Burman (24.5%), Mon (0.6%), and Yakine (1.4%). A small proportion had mixed ethnicity (3.7%). Five samples were rejected due to poor quality and five samples were not transferred from the clinic to the central haematology laboratory giving a final sample size of 504.

Prevalence and distribution of G6PD deficiency

The overall G6PD deficiency frequency by the FST (Table 1) was 13.7% (69/504); 14.1% (50/354) in Karen and 12.9% (16/124) in Burman subjects. The prevalence of G6PD deficiency varied amongst clinic sites from 9% to 18% (Fig. 1), mirroring in part the ethnic composition of the population attending each site.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Geographic distribution of G6PD deficient variants detected in five SMRU clinics along the Thailand-Myanmar border.

The map indicates the geographic location of the five SMRU clinics where G6PD phenotypes and genotypes were investigated. The white and grey pie charts indicate the proportion of samples taken in each clinic that were either G6PD normal (white) or G6PD deficient (grey). The second-layer pie charts indicate the genetic variants causing phenotypic deficiency at the corresponding site. In two clinics, MLA and MKT, only Mahidol variant was found.

https://doi.org/10.1371/journal.pone.0116063.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Number of subjects according to clinic site, G6PD phenotype and ethnic group.

https://doi.org/10.1371/journal.pone.0116063.t001

Quantitative assessment of G6PD activity

Quantification of G6PD activity by spectrophotometry was performed on a subset of the deficient subjects including 21 samples with Mahidol variant, one with Viangchan, one with Seattle and one with Chinese-4 mutations (Table 2). The activity in the 21 Mahidol samples had a normal distribution with a mean (SD) of 0.059 (0.031) IU/gHb. When compared to the Mahidol variant, Viangchan showed an activity in the same range while Chinese-4 and Seattle variants showed activity almost 10 times higher; although the quantitative assay was performed only in single subjects for the non-Mahidol mutations, all these mutations had a residual activity lower than 10% and they all were correctly diagnosed by the fluorescent spot test.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Enzymatic activity assessed by the spectrophotometric assay in G6PD normal males and hemizygous males for 4 different G6PD variants.

https://doi.org/10.1371/journal.pone.0116063.t002

Discussion

G6PD deficiency is common in this malaria endemic region along the northwestern border of Thailand. The Mahidol variant (Gly163Ser) which is the most common G6PD mutation in Myanmar [12], [13] and western Thailand [10] accounted for the majority (88%) of G6PD deficiency. G6PD Mahidol was discovered in 1972 [20] and a molecular characterization was reported in 1989 [21]. Recently a lower capacity of protein folding has been suggested as the cause of enzymatic deficiency [22]. The present study shows that in the population living along the Thailand-Myanmar border, G6PD Mahidol enzymatic activity in hemizygous males is below 10% of the activity detected in G6PD normal males of the same population. Although the quantification was performed in one third of the deficient subjects, these data seem to contradict the historical classification of G6PD Mahidol as a WHO Class III or “mild deficient variant”. The inclusion in the Class III came probably from data reported in the paper by Panich and colleagues [20] where a mutation with residual activity ranging from 5 to 32% (mean 11%) was found in 22 Thai male subjects. The mutation was called “Mahidol” but no analysis of DNA was performed at that time and therefore it is impossible to say whether the subjects analyzed were all carriers of the molecular Mahidol variant or possibly Viangchan, the most common mutation among Thais [10]. In the present study, the phenotypic analyses have been performed using the WHO recommended protocol on samples with PCR confirmed Mahidol mutation; our estimation of residual enzymatic activity is consistent with other published data on samples characterized at the DNA level [13] and it is well below the 10% of normal. This illustrates the limitation of a classification system often based on single observation and non-standardized laboratory methodology and considered “no longer useful” for discrimination of Class II and Class III variants [23].

Chinese-4 variant (Gly131Val, also called Quin Yuan) found in the present study at three different sites and in subjects of different ethnic groups (Karen, Burman and Yakine) has been described before in China and in the Thai population residing in the western regions of Thailand [14] at a population frequency of approximately 1%. Viangchan variant, the most common G6PD mutation in Laos, Vietnam, and Cambodia was found here in a subject of Burman ethnicity. In the present study the analysis of the phenotypically deficient subjects has led to the detection of other 3 uncommon mutations, some of which have never been described in the area. Açores variant (Ile199Val) has been described only in Portugal so far [24]. Seattle variant (Asp282His, also called Lodi/Modena, FerraraII, Seattle-like) has been found repeatedly and characterized in Italy [25] and found in the Canary Islands [26] and in Greece [27]. The Mediterranean variant has been found here and in many distant populations ([28], [29], [30], [31]), and it is a striking example of convergent evolution resulting from malaria selective pressure [32].

Not surprisingly, the molecular heterogeneity of the G6PD gene found in this area is comparable to most of other places where this has been investigated (Cambodia, [5]; West Africa, [33] and [34]; China, [35]). Now that genotyping is relatively easy, G6PD population surveys are often undertaken without phenotypic testing and a panel of mutations are screened; this tends to underestimate the overall prevalence of deficiency in the population level by missing the less common variants; in trials where whole gene sequencing has been performed, even without phenotypic tests, additional heterogeneity has been found at the gene level [36].

In conclusion, the population from Myanmar living in the study area has a high frequency of G6PD deficiency ranging from 9% to 18%. Several G6PD variants were found among subjects of Burman ethnic origin while the Karen ethnic group seems to have a more homogeneous genetic background. The fluorescent spot test correctly diagnosed all the hemizygous males in the study subjects, confirming its reliability in this setting. Quantitative characterization of enzymatic activity in deficient subjects with Mahidol, Viangchan, Seattle and Chinese-4 variants showed a residual enzymatic activity <10% for all the variants. This level of deficiency would be expected to be associated with clinically significant haemolysis during treatment with primaquine with the standard 14 days 30 mg base/day adult dose regimen for radical cure of Plasmodium vivax.

Acknowledgments

We would like to thank all the study participants, clinical staff and administrative staff for their cooperation, hard work and dedication. In particular we would like to acknowledge the efforts of Htun Htun Win, Kyi Khin, Yin Mon, Naw Dah, and Sourideth for their contributions as the leaders of this study at the clinics.