Background
Glucose-6-phosphate dehydrogenase (G6PD) is a house-keeping enzyme that acts as a rate limiting enzyme in the pentose phosphate pathway. It generates NADPH which is required for downstream anti-oxidative activity in glutathione (GSH) cycle. G6PD deficiency is the most common enzymopathy in the world, affecting more than 400 million people; most are living in malaria endemic countries. G6PD deficiency is caused by a functional mutation of the enzyme, making it unstable and less able to form dimers or tetramers thus decreasing enzyme activity [
1]. Clinical manifestations of G6PD deficiency vary widely from mostly asymptomatic to chronic anaemia (non-spherocytic haemolytic anaemia). The deficiency is defined according to the 1967 World Health Organization (WHO) enzyme classification. Class I as the most severe and chronically anaemia; class II as severe and class III as intermediate; and class IV as normal [
2].
G6PD deficiency is an X-linked disorder and thus deficient males and homozygous deficient women will exhibit the full extent of the deficiency whereas G6PD heterozygous women will have a range of G6PD activities from deficient to very normal G6PD activity. This is due to random X-inactivation that occurs very early on in the epiblast during gastrulation in individual cells [
3‐
6] creating genetic mosaicisms among G6PD heterozygous women with the same gene mutation [
4]. These mosaicisms give rise to variable G6PD activities and are thus very difficult to diagnose with the currently available qualitative diagnostic tests. Current qualitative tests can only detect < 30% G6PD activities. Previous studies have shown that the heterozygous women with G6PD activities between 30 and 80% of normal are the most difficult to diagnose since they may give ‘normal’ results, and yet, they can still haemolyse upon exposure to oxidative agents such as drugs (anti-malarial primaquine), food (fava beans) and infections [
7‐
9]. Single-dose tafenoquine, another 8-aminoquinoline drug, has recently been approved by FDA (United States Food and Drug Administration) as another radical cure for malaria, aside from primaquine [
10]. Because of the long half-life of tafenoquine, screening for G6PD must be done prior to giving the drug to malaria patient. To avoid the potential harmful side effect of the drug to G6PD deficient individuals, those with G6PD activity ≤ 70% of normal should not receive this drug and this includes G6PD heterozygous women [
11].
G6PD quantitative test is the only test capable of diagnosing G6PD heterozygous women [
12]. However, the severity of haemolysis in relation to the quantitative result when exposed to oxidative stress is unknown in these women. Oxidative stress is an imbalanced condition between the level of oxidant and antioxidant. Normally, cells have the abilities to maintain the redox equilibrium of oxidant and antioxidant level [
13‐
15]. However, the imbalance of redox equilibrium may occur in certain conditions, such as impairment in antioxidant systems leading to oxidative stress [
16,
17]. Oxidative stress in red blood cells (RBC) occur when RBC are exposed to endogenous and exogenous oxidative agents in the circulation [
18]. High exposure of oxidative agents to RBC will lead to the haemoglobin oxidation to methaemoglobin and thus condense into Heinz body formations. The Heinz bodies will precipitate in the RBC membrane and become attached to cytoskeletal proteins, such as spectrin, and membrane proteins, such as Band 3. This process eventually leads to RBC membrane disruption and haemolysis or removal in the spleen [
19,
20].
In this study, an in vitro RBC model was employed using CuClfor G6PD deficiency as described by Baird et al. [
7] that were exposed to high concentration of oxidant(300 mM H
2O
2) to determine the quantitative cut-off for haemolysis. The extent of oxidative stress and degree of protection from G6PD will be measured using assays such as MDA and GSH, respectively. This result was validated in ex vivo RBC from G6PD heterozygous women having the same genotype. Knowing this cut-off value likely to result in haemolysis enables the treatment of some G6PD heterozygous women who would be likely to tolerate primaquine or tafenoquine therapy for malaria who otherwise might have been excluded because of their heterozygosity.
Methods
Subjects and sample preparations
The RBCstudies were developed from two normal males having normal G6PD activities and normal haematological profiles and two G6PD heterozygous women having normal haematological profiles. Informed consent was obtained from all subjects prior to obtaining 8 mL of venous blood into ACD tubes from two normal males and 2 mL of venous blood into EDTA tubes from two G6PD heterozygous women. Complete blood count was performed for each subject immediately after blood collection using automated haematology analyzer Cell Dyne pocH-100i (Sysmec, USA) and the blood was aliquoted accordingly for CuCl treatment, H2O2 incubation, G6PD activity measurement, flow cytometer analyses, MDA and GSH tests and were stored at 4 °C until used. Ethical clearance for the study was approved by The Ethics Committee of The Faculty of Medicine, Universitas Indonesia (No. 640/UN2.F1/ETIK/2016).
Red blood cells G6PD deficiencymodels using CuCl
Both Cu
+ and Cu
2+ can inhibit G6PD activity in RBC [
21,
22]. Slight modification to the method previously described was made [
7]. Procedures were done as soon as blood was collected where 400 µL of whole blood in ACD was treated with 10 mM CuCl in water (Fluka, Germany) in various final concentrations of 0 mM, 0.2 mM, 0.4 mM, 0.6 mM, 0.8 mM, 1.0 mM, 1.5 mM and 2.0 mM CuCl to mimic 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80% and 100% G6PD activities in heterozygous women, respectively. All samples were incubated in a 37 °C water bath for 24 h without shaking.
G6PD quantitative test
G6PD activity was measured for each sample after treatment with CuCl using quantitative test from Trinity Biotech (Trinity Biotech, Ireland) to determine the enzyme activity. The assay measured the rate of reduction of NADP+ to NADPH spectrophotometrically at 340 nm using ultraviolet spectrophotometer (Shimadzu UV–VIS 800, Japan). The rate of NADPH formation is proportional to G6PD activity. Haemoglobin was measured using HemoCue (Hemocue® Hb 301 System, Sweden). G6PD activity is calculated in relation to haemoglobin level in U/g Hb according to the manufacturer’s manual.
H2O2 oxidative stress treatment
Prior to incubation with H2O2, the RBC previously treated with CuCl, as well as the RBC from heterozygous G6PDsubjects and normal controls, were washed twice with phosphate buffer pH 8.0. Plasma and buffy coat were discarded during the washing steps. After the first wash, complete blood count was performed for all samples to calculate the amount of RBC in each sample. Final volume of 200 µL of RBC suspension from each sample (each sample contained 108 RBC/200 µL suspension) was incubated with H2O2(30%, Merck, Germany) to final concentration of 300 mM at 37 °C for 2 h, to induce oxidative stress in these cells. These would then be measured for glutathione and TBARS assays. Meanwhile, H2O2-untreated RBC from G6PD heterozygous subjects and G6PD normal controls were also incubated at 37 °C for 2 h alongside treated RBC samples.
Malondialdehyde measurement assay
Lipid peroxidation is an indicator of cellular damage caused by oxidative stress and MDA is a good marker of such lipid peroxidation. To measure MDA, OxiSelect TBARS Assay (Cell Biolabs, USA, Cat.# STA-330) was used where thiobarbituric acid reactive substances (TBARS) is a rapid and direct quantitative measurement of MDA in biological samples. MDA forms a 1:2 adduct with TBA and can be measured colorimetrically. The procedure followed the manufacturer’s protocol where an additional step was added to get rid of the interfering haemoglobin and its derivatives by adding
n-Butanol. All samples and MDA standards were transferred to cuvettes including a blank control and read at 532 nm absorbance. All samples and standards were done in duplicates. All MDA standards were plotted into a standard curve and MDA concentration for every sample was measured in pmol/µg total protein where the protein was measured according to Lowry protein quantitation method [
23].
Glutathione assay
Reduced GSH, as the major anti-oxidant in cells, especially in RBC, is normally present around 90–95% of total glutathione in the cells. Intracellular level of GSH is used as an indicator of the overall redox state of the cell. In normal cells, increased level of GSH indicates that there is an oxidative pressure within the cell. The Glutathione Assay kit (Sigma-Aldrich, Germany, Cat.# CS0260) provides the means to measure level of total glutathione (GSSG and GSH) in a biological sample. The kit uses a kinetic assay where the catalytic amounts of GSH cause a continuous reduction of 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB) to TNB and the generating GSSG is recycled by glutathione reductase and NADPH. The rate of reaction is proportional to the concentration of GSH up to 2 µM where the yellow product, TNB is measured spectrophotometrically at 405 nm and level of GSH is measured in nmol/mg of protein in the sample as stated in the manufacturer’s protocol with slight modification, wherein a 405 nm wavelength was used instead of 412 nm.
G6PD cytofluorometric assay
Prior to treatment with H
2O
2, percentage of G6PD deficient cells were analysed cytochemically to see whether they coincided with G6PD activities. The assay involved nitrite which oxidized all oxyhaemoglobin to methaemoglobin. The methaemoglobin was reduced back to oxyhaemoglobin by glucose as substrate and Nile blue as redox catalyst. This reaction depended on NADPH and G6PD normal RBC would have oxyhaemoglobin converted faster than G6PD deficient RBC. The addition of cyanide reacted with methaemoglobin and produced cyan-methaemoglobin, while oxyhaemoglobin remained inactive. Afterward, H
2O
2 introduction generated fluorescence only in oxyhaemoglobin thus distinguishing G6PD normal from G6PD deficient RBC. Fifty µL of packed RBC was mixed with 90 µL of phosphate buffer pH 8.0 for model cells and only 10 µL of packed RBC and 90 µL of phosphate buffer pH 8.0 for G6PD heterozygous subjects prior to applying the protocol from Shah et al. where the assay assessed G6PD activity at the level of individual RBC [
24]. The samples were analysed in BD Acuri C6 + cytometer (BD Biosciences, USA), using setting system 10,000 total events with FSC-H > 30,000 collected. Samples were excited using an Arion laser (488 nm) and fluorescence emission was measured in the FL1 channel (533 ± 30 nm).
DNA extraction and genotyping
DNA of G6PD heterozygous subjects were extracted using Wizard Genomic Purification DNA Kit (Promega, USA, Cat.# A1120) with a small modification, whereas DNA isolated from 300 µL of whole blood was diluted in 50 µL of DNA Rehydration Solution. DNA was PCR amplified and cut at various restriction enzyme sites to detect for various known G6PD variants that are common in Indonesia [
25,
26]. The fragments of PCR/RFLP product were then analysed using agarose gel electrophoresis.
Statistical analysis
To measure the correlation between CuCl concentration and G6PD activities, we used Spearman test. Linear regression test was used to analyse the association between G6PD activities with proportion of normal RBC in RBC models. Wilcoxon test was used to test the significance between RBC models in MDA and G6PD activities, as well as MDA and GSH levels in heterozygous G6PD subjects. Meanwhile, Kruskal–Wallis test was performed to analyse the differences between GSH and G6PD activity in RBC models. The data analysis was completed using RStudio version R i386 3.3.1. (
http://www.rstudio.com).
Discussion
Currently, the readily available qualitative G6PD tests can only detect < 30% G6PD activity, whereas those with activities between 30 and 80% can only be determined using quantitative G6PD test. This covers women who are G6PD heterozygous. However, the need to define a cut-off value within this 30–80% G6PD activity is needed to differentiate between severe from non-severe haemolysis when exposed to oxidant.
The CuCl RBC model was developed to study the effect of oxidative stress on RBC membrane by subjecting the cells to 300 mM of H2O2 and then tried to validate the results with RBC taken from G6PD heterozygous women with the same mutation (G6PD Viangchan) but different G6PD activities where one was deficient and the other was normal according to G6PD quantitative test. Women with 74% G6PD activity would definitely be considered as normal if tested with the readily available G6PD qualitative tests. The cytochemical results positively correlated with G6PD activitiesin the ex vivo model, and thus this model can be used for oxidative stress analyses. The ex vivo RBC showed promising result that may have reflected the expected result when exposed to high concentration of oxidant unlike RBC model that were treated with CuCl prior to exposing them to H2O2.
In the RBC model treated with varying concentrations of CuCl to achieve different percentages of G6PD activities, the higher the enzyme activity, the higher the percentage of normal red cells (Fig.
2) as predicted [
7]. However, there was a significant drop as seen in Table
1, from 80% G6PD activity with 89% normal cells to 60% G6PD activity with 43% normal cells (p > 0.05, was considered insignificant). It had been known that intracellular free copper was probably associated with macromolecule structures, such as DNA, enzymes and protein [
27]. The over dosage of copper is toxic to cells because it interferes with glycolytic enzymes, such as hexokinase and phosphofructokinase 1, as well as other enzymes, such as phosphoglyceric kinase, 6-phosphogluconate dehydrogenase, catalase and glutathione peroxidase [
21,
22,
28].
Excess copper also induces peroxidative damage in cell membranes leading to lipid bilayer destruction [
29], which might explain the discrepancy of normal cells between the 80% and 60% G6PD activity. Therefore, between 60 and 80% of G6PD enzyme activity, the drop in normal RBC is significant and thus may represent the targeted cut-off. The result of the cytochemical analysis of CuCl RBC model above was validated in cytochemical analysis of
ex vivoRBC. The ex vivo RBC showed subject with 33% G6PD activity had only 39% normal cells compared to subject with 74% G6PD activity with 97% normal cells (Fig.
3a) which might indicate that those below 60% may have already lower number of normal cells compared to those > 70% as shown in RBC model (Table
1).
MDA represents the level of RBC membrane damage caused by oxidative stress, therefore, the lower the G6PD activity, one would expect the higher the damage and thus higher MDA level. While ex vivo RBC showed higher MDA level in G6PD deficient cells compared to normal (p = 0.5, was considered insignificant), the CuCl RBC model showed the opposite, i.e. lower G6PD activity and lower MDA level [
22,
30]. This could be explained as before, that CuCl already caused oxidative stress prior to H
2O
2 treatment of the RBC. During the CuCl procedure, the membrane had gone through peroxidative damage already, leaving behind less membrane lipid bilayer prior to H
2O
2 treatment. Thus, fewer MDA level was detected in the CuCl-treated RBC model. This was skipped in ex vivo RBC because there was not CuCl treatment prior to H
2O
2 exposure.
Glutathione cycle depended on NADPH which was generated by G6PD in RBC [
31,
32]. In G6PD deficient RBC where NADPH was reduced, the total GSH was also reduced compared to normal RBC. The total GSH would be very low so that it would not overcome oxidative stress either from CuCl or H
2O
2. On the contrary, in normal G6PD RBC, in the presence of oxidative stress, total GSH would be induced because there was enough NADPH produced by G6PD. These results also indicated an inversely proportional correlation between CuCl and GSH level in CuCl RBC model. Rafter showed that leukocytes incubated with high concentration of Cu
2+ decreased the reduced GSH level [
33]. Previous reports by Smith et al. and Kachur et al. showed that the thiol group in GSH can be oxidized directly by copper (Cu
2+) which led to the decrease of GSH level intracellularly and thus interfere with glutathione peroxidase function in neutralizing H
2O
2 to water [
34,
35].
Total GSH measurements from both CuCl RBC model and ex vivo RBC showed that G6PD normal samples were higher than G6PD deficient samples (Figs.
3c and
4b). Figure
4b showed that in RBC with 80 and 100% G6PD activity there was a significant increase of total GSH to almost double and 4 times, respectively, compared toRBC with less than 60% G6PD activity. This suggested that cells with less than 60% G6PD activity did not have enough NADPH to significantly increase total GSH level. Therefore, this might suggest a cut-off (at 60% G6PD activity) in which one can use to differentiate one that is dangerously haemolytic from one that is safe when exposed to oxidant. In ex vivo RBC, there was a 30% difference of total GSH level between subject with 33% G6PD activity and subject with 74% G6PD activity, and the same for both H
2O
2 treated and untreated samples which suggested that the extra 30% GSH level was protective toward oxidative damage compared to those at lower G6PD activity as shown in MDA result. Figure
3b showed that the lower G6PD activity the higher MDA level as a marker for membrane damage. These findings were supported by other studies which explained that GSH level in normal RBC was higher than G6PD deficient-RBC [
36‐
38].
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.