Role of CYP2E1 and NQO1 polymorphisms in oxidative stress derived cancer in Thais with and without dyslipidemia

The study was approved by the Committee on Human Rights related to Research Involving Human Subjects, Faculty of Medicine, Ramathibodi Hospital, Mahidol University (no. MURA2008/809/S_1-2Aug₁₀). We enrolled 1380 employees at the state enterprise Electricity Generating Authority of Thailand (EGAT) after written informed consent to participate in the study was obtained from each participant. Participants lived and worked in Bangkok, and had the possibility of unintended exposed to various kinds and doses of environmental chemicals. Participants completed a self-administered questionnaire, underwent a physical examination, and provided a fasting blood sample. We collected 20 mL of blood from each participant by venipuncture into ethylenediaminetetraacetic acid (EDTA) -containing and heparinized tubes that were immediately centrifuged at 2000g. Buffy coat, erythrocytes, and plasma were separated and stored at −20°C until genotyping and biochemical measurements were conducted.

Whole blood (0.1 mL) was added to distilled water (1.9 mL) together with 3 mL of precipitating solution (100 mL containing 1.67 g glacial metaphosphoric acid, 0.2 g disodium EDTA, and 30 g sodium chloride). After standing for 5 min the mixture was filtered and filtrate (0.5 mL) was added to 0.3 M phosphate buffer, pH 6.4 (2 mL). Finally, we added 1 mM 5,5′-dithiobis-(2-nitrobenzoic acid) in 1% sodium citrate (0.25 mL), mixed the solutions well, and within 4 min the absorbance of the mixture was read at 412 nm. We compared the absorbances with appropriate blanks without blood [19].

Whole blood (0.1 mL) was added to distilled water (1.9 mL) to make a 1:20 hemolysate. Then, prepared solutions of 1 M Tris-HCl, 5 mM EDTA, pH 8.0 (100 μL), 0.1 M GSH (20 μL), 10 U/mL glutathione reductase (100 μL), and 2 mM NADPH (100 μL) ] were added to 10 μL of the hemolysate, together with 660 μL of distilled water and the mixture reaction preincubated at 37°C for 10 minutes. After the preincubation, the enzymatic reaction was initiated by addition of 10 L of freshly prepared 7 mM t-buty1 hydroperoxide. The decrease in the optical density at 340 nm was measured against a blank for 1 min [20].

Components of extraction solution (3.5 mL of cold distilled water, 1 mL of ethanol, 0.6 mL of chloroform) were added to 0.5 mL of hemolysate, then mixed for 1 min. After centriftigation at 3000 rpm for 10 minutes at 4°C, supernatant was collected to determine superoxide dismutase (SOD) activity. Seven serial dilutions were prepared for each sample. The final volume (3 mL) was composed of the following: 200 μL of 0.1 M EDTA containing 1.5 mg sodium cyanide, 100 μL of 1.5 mM nitroblue tetrazolium (NBT), 50 μL of 0.12 mM riboflavin, portion of sample (10, 20, 40, 60, 80, 200, or 500 μL) and 0.067 M potassium phosphate buffer, pH 7.8 to make a final volume. All tubes were illuminated in a light box for 12 min (18 W fluorescence) and the optical density of the mixtures was measured at 560 nm. The resulting inhibition of NBT reduction versus the amount of SOD extract was plotted on a linear scale. The amount of SOD that gave half of this maximum (1 unit) was determined from the plot [21].

Erythrocytes were diluted at the ratio of 1:500 in 50 mM phosphate buffer pH 7.0 (2.0 mL) to make a hemolysate. Then 30 mMH₂0₂ (1.0 mL) was added in a continuous stream from a pipette to promote mixing and to start the reaction. The decrease in the optical density of the mixtures was measured at 240 nm for 30 s following a previously described method [22].

Malondialdehyde (MDA) was determined using HPLC method with fluorescence detection [23]. The coefficient of variation was 4% within runs and 3% between days. The detection limit was 0.25 μmο1/L and the method exhibited a linear response for MDA in range 1.5 to 15 μmο1/L, the calibration curve presented a high correlation coefficient (r² > 0.9), P = 0.001; n =10).

Serum cholesterol and triglyceride were analyzed using routine biochemical procedures at Ramathibodi Hospital. The classifications of lipid profiles were based on the criteria of the National Cholesterol Education Program [24]. Normolipidemia was defined as a triglyceride concentration <150 mg/dL and normal cholesterol concentration <200mg/dL. Dyslipidemias were defined as hypercholesterolemia (triglyceride <150 mg/dL, cholesterol >200 mg/dL), hypertriglyceridemia (triglyceride >150mg/dL, cholesterol <200mg/dL), and combined hyperlipidemia (triglyceride >150 mg/dL, cholesterol >200 mg/dL).

Genomic DNA was extracted from lymphocytes using a modified salting-out procedure [25] and frozen at −20°C until analysis. We conducted a TaqMan assay including a forward target-specific polymerase chain reaction (PCR) primer, a reverse primer, and TaqMan MGB probes labeled with a special dyes: FAM and VIC (Applied Biosystems, Waltham, MA, USA). The reaction mixture consisted of TaqMan Universal Master Mix (1×) and TaqMan-MGB probes for CYP2E1 *5B (rs3813867, rs2031920), CYP2E1 *6 (rs6413432), and NQO1 (rsl800566) in a total volume of 10 μL. The real-time PCR reaction protocol was 10 min at 95°C, 40 cycles of 15 s at 92°C, and 1 min at 60°C, using a 7500 Real-Time PCR System (Applied Biosystems). Data for specific probes and primers are available in the SNP500 database of the U.S. National Cancer Institute at http://snp500cancer.nci.nih.gov/ [26].

Statistical analyses were conducted using IBM SPSS Statistics for Windows, version 19.0 (Armonk, NY, USA). All biochemical parameters (glutathione peroxidase (GPx), SOD, catalase (CAT), MDA, and reduced glutathione (GSH)) are expressed as geometric means with 95% confidence intervals (95% CIs) and means with standard deviations (SD). Goodness of fit to normal distribution was determined using a Kolmogorov-Smirnov test. Non-normally distributed data was transformed into a log scale and retested for normal distribution before testing at the next step. A one-way analysis of variance and Mann-Whitney U test were conducted to compare difference in the means between 2 groups for normal and non-normally distributed data, respectively. A general linear model was applied to evaluate the significance of differences in the mean of parameters between more than 2 groups. Genotype frequencies were compared and tested for Hardy-Weinberg equilibrium using a Pearson χ² test. Logistic regression analyses were used to determine the association of oxidative stress markers with genes of interest and lipid profiles in the model adjusted for covariates (sex, smoking, and alcohol consumption). In this case, data were presented as an odds ratios (ORs) with 95% CIs. P < 0.05 was considered significant.

The demographic characteristics of 1380 EGAT employees (983 men and 397 women) were subdivided into 4 groups based on lipid profiles as summarized in Table 1.

By comparison with normolipidemia, participants with different types of dyslipidemia (hypertriglyceridemia, hypercholesterolemia, and combined hyperlipidemia) showed significantly higher BMI, triglyceride, HDL-C, and LDL-C levels (all P ≤ 0.024). The frequency of CYP2E1 *5B c1c1 was approximately 71%, c1c2 was approximately 26%, and c2c2 was approximately 3%. The distribution of CYP2E1 *6 DD was 60%, DC was 34%, and CC was 6%. The distribution of NQO1 *2 CC was 32%, CT was 50%, and TT was 18%. The allele frequencies of CYP2E1*5B, CYP2E1*6, and NQ01*2 were c1 0.84, c2 0.16; D 0.78, C 0.22; C 0.57, and Τ 0.43, respectively. All allele frequencies were in Hardy-Weinberg equilibrium in all lipid profile groups (all P > 0.47). Different blood lipid levels were found in participants who carried the wild-type CYP2E1 (c1/c1) and (DD) alleles with P < 0.001. For NQO1, significant difference in blood lipid levels was found in participants who carried the heterozygous genotype (CT) with P < 0.001. The association of genetic variation and lipid profile on oxidative stress markers is presented in Table 2.

By comparison with the wild type, significant associations were found in heterozygous CYP2E1 PstI with an OR 2.53 (P = 0.007) /RsaI with an OR 2.56 (P = 0.006) for GPx and in variant RsaI with an OR 7.22 (P = 0.004) for CAT. Significant associations between hyperlipidemia with MDA were observed (both P ≤ 0.01). The combined effect of genetic variation and lipid profile on oxidative stress status was analyzed.

No significant differences in GSH level, CAT activity, or GPx activity were found between subgroups. However, the level of GSH (25.66 ±8.66 mg/dL to 35.42 ± 1.76 mg/dL) tended to be higher in participants in the hyperlipidemia subgroup and participants bearing any variant alleles, than in participants bearing a wild-type allele with normolipidemia (28.00 mg/dL ± 6.52 to 32.66 ± 6.32 mg/dL).

Participants bearing any variant allele of either gene of interest tended to have a lower mean CAT activity (from 19.50 ± 9.95 kU/g hemoglobin (Hb) [95% CI] [6.5-39.06] to 37.17 ± 4.75 kU/g Hb [27.84-46.51]) than participants bearing the wild-type alleles of either gene of interest with normolipidemia (from 26.27 ± 1.68 kU/g Hb [22.97-29.57] to 37.06 ± 2.89 kU/g Hb [31.38-42.74]). Significant differences in SOD activity were found between the subgroups (Figure 1A-C) in combined hyperlipidemia subgroups and participants bearing any NQO1 alleles with PstI or RsaI variants. Participants with homozygous PstI or RsaI variants had the highest mean SOD activity (1.69 ± 0.51 U/g Hb [0.70-2.69] to 3.45 ± 0.93 U/g Hb [1.62-5.27]) (P = 0.01 and P = 0.03).

Participants in the hyperlipidemia subgroup and participants bearing any variant allele of either gene of interest tended to have a higher mean GPx activity (31.61 ± 11.56 U/g Hb [95% CI] [8.9-54.33] to 52.29 ± 15.98U/gHb [19.91-82.66]) than participants bearing the wild-type alleles of both genes of interest with normolipidemia (25.24 ± 6.67 U/g Hb [12.13-38.35] to 42.40 ± 3.86 U/g Hb [34.82-49.98]) although the differences were not significant.

Figure 1A-C

In the subgroup with combined hyperlipidemia. participants bearing any NQO1 allele with a RsaI variant showed significantly higher levels of MDA than participants bearing any NQO1 with a wild-type RsaI allele with P = 0.021. Participants bearing a heterozygous NQO1 allele with a DraIvariant had a significantly higher level of MDA than participants bearing any NQO1 with a wild-type DraIallele (P = 0.031). Comparison of the subgroups of hyperlipidemia showed participants bearing any variant alleles had a higher mean level of MDA (6.04 ± 2.74 (SE) μιmol/L. [range] [0.65-11.40] to 16.22 ± 5.68 μιmol/L [5.03-27.41]) than participants bearing a wild-type allele with normolipidemia (4.00 ± 3.49 μmol/L [-2.80-10.8] to 11.97 ± 2.55 μmol/L. [6.95-16.98]) (Figure 2A-C).

Figure 2A-C