Reduced glutathione and glutathione disulfide in the blood of glucose-6-phosphate dehydrogenase-deficient newborns

GSH was obtained from AMRESCO (Solon, OH, USA). NeoBase non-derivatized MS/MS kits were purchased from PerkinElmer (Boston, MA, USA) and included isotopically labeled glycine (¹⁵N, 2-¹³C-glycinem, 762.2 μmol/l as the working solution) as an internal standard. The DBS extraction solution and the flow solvent were composed of methanol(1000 ml),water (2 ml), and formic acid(0.2 mmol/L).

The endogenous concentration of GSH in whole blood by this method is as low as 200 μmol/L, and as high as 500 μmol/L. The GSH recovery test samples were prepared with the extraction solution. GSH was added to vials containing 0.1 ml of extraction solution to produce GSH concentrations of 0,200,400,600,800,1000,2000,and 4000 μmol/L. All of the samples were dropped onto S&S Grade 903 filter paper using a 100-μl pipette without overlap, dried at room temperature, prepared in dry-spot, and stored at 4 °C until the MS/MS analysis.

The blood samples for newborn G6PD screening, including 12 positive and 623 negative, were collected by puncturing the heel on the third to seventh day of age. The blood was dropped directly onto S&S Grade 903 filter paper without overlapped by another drop of blood and dried at room temperature. The DBS samples were transferred to our newborn screening center and stored for 2 to 4 days at 4 °C. A quantitative spectrophotometric assay of G6PD activity was used for G6PD deficiency screening (Neonatal G6PD KIT, PerkinElmer Life and Analytical Sciences; WallacOy, Turku, Finland).

Thirty seven blood samples were collected from the veins of neonates who tested positive during the G6PD screening (These 37 samples were not all from the 12 babies above, other patients were added). Thirty seven control blood samples from their parents and from 21 neonates control without G6PD deficiency were also intravenous collected. In each case, the blood was drawn into a tube containing sodium citrate as an anticoagulant and mixed immediately. The blood samples were transferred to the laboratory within 3 h and dropped onto S&S Grade 903 filter paper using a 100-μl pipette without overlap and dried at room temperature for 6–24 h until examination by MS/MS. The G6PD activity for the confirmation test was measured immediately when the blood arrived at the laboratory using a method that measures the G6PD/6-phosphogluconate dehydrogenase (6PGD) ratio [18]. The clinical and neonatal samples were obtained from maternity hospitals and the Children’s Hospital of Shanghai. This study approved by the ethics committee of the participating hospitals.

We extracted GSH and GSSG, use with one 3 mm DBS disk per analysis, using 100 μl extraction solution containing isotopically labeled glycine working solution (11:1) per sample in 96-well plate (NUNC),incubating in 45 °C, shaking at 700 rpm for 45 min. The extractions were transfer to another 96-well plate for mass spectrometry analyses.

Mass spectrometry analyses were performed using a Waters MICROMASS Quattro micro™API, (Manchester, UK). The electrospray needle was maintained at +3.5 kV, and the desolvation temperature was set to 350 °C. The desolvation gas flow was 650 L/Hr, the cone potential was 30 V, the collision energy was 3.0 eV, and the dwell time was 50 ms. Nitrogen, the sheath gas, was set at 50 units. The collision gas used was argon. The temperature of the heated capillary was maintained at 120 °C. GSH and GSSG were quantified in multi-reaction-monitor (MRM) mode using positive electrospray ionization mass spectrometry. The strongest MRM signals for GSH and GSSG were selected for quantification. Their fragment ions assumed and generated by collision-induced dissociation of the [GSH + H] ⁺ (Fig. 1)and [GSSG + Na]⁺ions (Fig. 2) were observed in product scans. The settings (precursor ion → fragment ion) for the target analytes were GSH m/z 308.0 → 76 and GSSG m/z 636.8 → 330. The setting for the isotopically labeled internal standard ¹⁵N, 2-¹³C-glycine was m/z 78 → 32 [19, 20]. The samples were delivered using an HPLC pump (Waters 1525 μBinary, MA, USA) equipped with a 20-μl sample loop. The samples were run at 116 μl/min from 0 to 0.23 min, at 20 μl/min from 0.24 to 1.35 min, and at 600 μl/min from 1.36 to 1.7 min. The concentration was calculated using Masslynx software, version 4.0 (Waters, Milford, MA, USA) by combining the intensities of the m/z 308.3 → 76 peak for GSH, the m/z 636.8 → 330 peak for GSSG, and the m/z 78 → 32 peak for the internal standard (¹⁵N, 2-¹³C-glycine) from 0.4 to 1.3 min, as measured as peak areas, for each sample. The serum concentrations were determined using the following formula: GSH = hematocrit × coefficient × Intensity m/z 308.3 → 76/78 → 32 + parameter; GSSG = hematocrit × coefficient × Intensity m/z 636.8 → 330/78 → 32.

All analyses were performed with SPSS 19.0 (SPSS Inc. Chicago, IL, USA). Differences of mean were tested by the one-way ANOVA among three groups and by independent t test between two groups. Significance was accepted at a P value of ≤0.05. Linear regression of the GSH/glycine-IS peak area ratio vs. the concentration of GSH added, and the ROC Curve of GSH/GSSG ratios to predict G6PD deficiency were done.

Based on the MS results for the product and precursor scans of the GSH (Fig. 1)and GSSG solutions (Fig. 2), we decided to use an MRM of m/z 308.0 → 76 for GSH, m/z 636.8 → 330 for GSSG, and m/z 78 → 32 for isotopically labeled glycine (¹⁵N, 2-¹³C-glycine) as an internal standard. We also combined the intensity of scans as the peak area for each analyte.

Calibration curves were obtained by linear regression based on a plot of the GSH/glycine internal standard peak-area ratio(x) vs. the concentration(y) of added GSH (Fig. 3). The concentration range used for the added GSH was 200–4000 μmol/L. The slope was 570 μmol/L, the intercept was 200 μmol/L, and the squared correlation coefficient was 0.994, p < 0.0001. We used the formula Y = 762 μmol/L × intensitym/z 308.3 → 76/78 → 32 + 200 μmol/L(hematocrit0.5 is assumed)to calculate the concentrations of GSH in the blood samples. The value used to represent GSH/GSSG was the peak area ratio of these compounds.

The recoveries of GSH added DBSs were determined at nine concentrations and were found to be in the range 90 to 126% for GSH in the concentration range 200–4000 μmol/L (Table 1) when examined on the first day (within 24 h) after sample preparation. However, the GSH recovery gradually decreased when the samples were examined on the second day and thereafter (Table 1).

The confirmed G6PD-deficient newborns had lower blood levels of GSH and lower GSH/GSSG ratios than the neonatal and parental controls (Table 2). All the blood samples for confirming test contained sodium citrate were prepared in DBS and examined on the first day. The parents should carry the allele conferring G6PD deficiency, and some exhibited decreased, albeit only slightly, G6PD activity, led to the GSH concentrations and GSH/GSSG ratios in the parent group were higher than patient group and lower than the normal control group.

The GSH levels and GSH/GSSG ratios in DBS samples are routinely examined in neonatal disease screening by combining the detection of amino acids and acylcarnitines using MS/MS. These DBS samples were prepared in maternity hospitals without the use of sodium citrate and examined at least three days after collection. The activity of G6PD in the DBS samples was examined on the second day after the GSH levels were examined using MS/MS. The activity of G6PD was positively correlated to the GSH/GSSG ratio (R = 0.213, p < 0.001 n = 278), though no significant positive relationship with the GSH content was observed (R = 0.027, p > 0.05, n = 278). However, there were no significant differences in the mean GSH concentrations and the GSH/GSSG ratios between the G6PD deficiency-positive group and the G6PD deficiency-negative group (Table 3).