Association of Glutathione S-transferases (GSTT1, GSTM1 and GSTP1) Genes Polymorphisms with Nonalcoholic Fatty Liver Disease Susceptibility: A Meta-analysis of Case-control Studies

Abstract

Background: Glutathione S-transferases (GSTs) genes single-nucleotide polymorphisms (SNPs) have been connected with the susceptibility of nonalcoholic fatty liver disease (NAFLD), but with inconsistent results across the current evidences. The present work was schemed to explore the association between GSTs genes polymorphisms and the NAFLD vulnerability via meta-analysis.

Methods: PubMed, Web of Science, Cochrane Library, China National Knowledge Infrastructure (CNKI) and Wanfang were retrieved for eligible literatures previous to March 10, 2021. The odds ratio (OR) of the dichotomic variables and the standardized mean difference of quantitative variables with corresponding 95% confidence intervals (95%CIs) were computed to evaluate the strength of the associations. The quality of included studies were assessed via using Newcastle-Ottawa Scale (NOS).

Results: In total, 7 case-control studies encompassing 804 NAFLD patients and 1362 disease-free controls in this meta-analysis. Ultimately, this analysis included six, five and five studies for GSTM1, GSTT1 and GSTP1 polymorphisms respectively. The pooled data revealed that the GSTs genes single-nucleotide polymorphisms had conspicuous associations with NAFLD susceptibility: for GSTM1, null vs. present, OR=1.46, 95%CI 1.20-1.79, P=0.0002; for GSTT1, null vs. present, OR=1.34, 95%CI 1.06-1.68, P=0.01; for GSTP1, Ile/Val or Val/Val vs. Ile/Ile, OR=1.60, 95%CI 1.23-2.09, P=0.0005.

Conclusion: This work revealed that the GSTM1 null, GSTT1 null and GSTP1-Val genotypes might be related to increased NAFLD susceptibility.

Background

Nonalcoholic fatty liver disease (NAFLD) is gradually considered as the liver disease component of metabolic syndrome, which a risk factor for further development of fatty liver, non-alcoholic steatohepatitis (NASH), fibrosis, cirrhosis and hepatocellular carcinoma [1]. According to the present data, NAFLD affects 10%-30% of the general population in various countries, which it has been viewed as a huge global health burden [2]. Little is known about the latent mechanism involved in the development and pathogenesis of NAFLD, and yet it is a complicated metabolic process in which both environmental and genetic factors are etiology [3]. At present, genome wide association studies (GWAS) have demonstrated that several conspicuous genetic susceptibility genes such as GSTs, KLF6, GCKR, which had been verified the crutial roles in the disease onset and progression of NAFLD. In resent years, epidemiological studies have shown that GSTs are a multifunctional enzyme in relation to the endogenous toxic metabolites, phase II detoxification of xenobiotics and free radicals [4]. Moreover, the GSTs family acts a significant role in antioxidant defence mechanisms via accelerating detoxification of electrophilic xenobiotics and deactivating a range of endogenous byproduct of oxidative stress [5–8]. So far, studies have confirmed that GST enzymes in possession of eight classes of soluble cytoplasmic isoforms, such as α-(A), ζ-(Z), θ-(T), κ-(K), µ-(M), π-(P), σ-(D), and ω-(O) [9]. In the last few years, GSTT1, GSTM1, and GSTP1 have attracted much ttention. Indeed, the GSTT1, GSTM1, and GSTP1 genes encodes the θ, µ, π class of GST enzymes, and they are located on chromosomes 1p13.3, 22q11.2 and 11q13, respectively [10]. It is reported that GSTT1/GSTM1 deletions (GSTT1/GSTM1 null) could inhibit detoxification of GSTT1/GSTM1 substrates that are either toxicant or carcinogen. Double null genotypes of GSTM1 and GSTT1 might give rise to a complete lack of enzymatic activity [11]. Furthermore, the GSTP1 gene polymorphism is the outcome of a single nucleotide substitution of A to G, which leads to valine instead of isoleucine in the binding site of GSTP1 and alters catalytic activity of enzyme [12, 13].

Several studies have appraised the relationship of GSTs genes SNPs and liver-related diseases. It indicated that the null genotypes of GSTM1/GSTT1 and GSTP1-Val (105) genes SNPs were related to the risk of NAFLD [14–20]. However, these studies yielded varying and divergent results. Accordingly, a meta-analysis was carried out to supply a more accurate and synthetic assessment on the relationship of GSTM1, GSTT1, GSTP1 genes polymorphisms and the NAFLD susceptibility.

Materials And Methods

Literature collection

Two independent researchers (Ming Qiao and Yi Zhu) searched the PubMed, Web of Science, Cochrane Library, CNKI and Wanfang databases prior to March 10, 2021. The searching strategy of PubMed was exhibited as follows: (“nonalcoholic fatty liver disease [Mesh]” OR “non-alcoholic fatty liver disease” OR “NAFLD” OR “non-alcoholic steatohepatitis” OR “NASH”) AND (“GSTT1” OR “GSTM1” OR “GSTP1” OR “glutathione S-transferase” [Mesh]) AND (Single Nucleotide Polymorphism[Mesh] OR Variant OR SNP OR Polymorphism OR mutant OR mutation OR variation). No language restriction was set.  

Inclusion and exclusion criteria

Original studies were incorporated into this analysis in the light of the following inclusion criteria: (i) case-control studies; (ii) study investigated the associations of GSTs polymorphisms and NAFLD predisposition; (iii) NAFLD is diagnosed by pathology or ultrasound; (iv) control subjects were disease-free individuals; (v) detailed genotype data can be calculated for ORs and 95% CIs. Correspondingly, reviews, conference abstracts, commentary articles, letters to editor, animal studies, unpublished data, case reports, as well as family-based studies were excluded. 

Methodological quality assessment

The Newcastle-Ottawa Scale (NOS) was performed to evaluate the methodological quality of included studies by two independent investigators (Ming Qiao and Yi Zhu). The NOS is composed of three aspects, namely selection, comparability and exposure. Each study could receive 0 to 9 scores. Nevertheless, studies with≥6 scores were regarded as high-quality studies. Disagreement was resolved by discussion. 

Data extraction

The extracted information contained the name of first author, year of publication, country, ethnicity and gender of enrolled subjects, numbers of NAFLD and control subjects, diagnostic methods of NAFLD, genotyping of enrolled subjects and Hardy-Weinberg equilibrium (HWE) results. Ming Qiao and Yi Zhu extracted the information independently.  

Statistical analysis

The control participants of incorporated studies were estimated via HWE [21]. Summary ORs with 95% CIs were computed to specifically evaluate the relationship of GSTs polymorphisms and the NAFLD susceptibility. Cochran's Q-test and I² test were applied to appraise the between-study heterogeneity [22]. The random-effect model was employed to calculate merged ORs if P<0.1, I² >50%. If not, the fixed-effect model was utilized for data synthesis [23]. Sensitivity analysis was performed as well to assess the stability of all incorporated studis via the leave-one-out method. When the included studies were more than 10, funnel plot was employed to valuated the publication bias [24]. On the contrary, less than 10 studies are not required. All analyses were done using RevMan 5.3 software.

Results

Literature search

Primary search of electronic databases retrieved 41 potentially relevant publications: 17 from PubMed, 20 from Web of Science, 0 from Cochrane library, 1 from CNKI and 3 from Wanfang. No additional records were acquired from other sources. And then 25 studies remained after removing duplicated articles. Subsequently, a total of 9 irrelevant articles were excluded on the basis of titles and abstracts screening. After applying inclusion and exclusion criteria, 7 unrelated articles, 1 conference abstract were removed and 1 full-text article was not available. Ultimately, seven studies went into the process of meta-analysis. Overall, a flowchart summarizing the procedure of literature identification was illustrated in Fig. 1. 

Main characteristics

The studies were performed in Italy [14], Japanese [15, 17], Iran [16,18] and Ukraine [19, 20]. In total, seven studies encompassing 804 case and 1362 control participants were analyzed in current analysis. For GSTM1, GSTT1 and GSTP1 genes SNPs, there were six, five and five studies ultimately incorporated, respectively. They were all case-control designed and published between 2008 and 2020 (Table 1). In addition, studies with≥6 scores were regarded as high-quality studies according to evaluation of methodological quality (Table 2). 

Table 1

 Main characteristics of the included studies.

Table 2 

Quality assessment of included studies based upon the Newcastle-Ottawa Scale (NOS)

GSTM1 gene polymorphism and NAFLD susceptibility

In total of six studies including 700 NAFLD patients and 1317 controls for GSTM1 gene polymorphism. The fixed-effects model was employed for data analysis on account of the heterogeneity in between-study was not remarkable. It revealed that GSTM1 was appreciably connected with the NAFLD vulnerability (null vs. present, OR=1.46, 95%CI: 1.20-1.79, P=0.0002; Fig. 2). 

GSTT1 gene polymorphism and NAFLD susceptibility

Overall, five researches containing 620 NAFLD and 1237 healthy subjects for GSTT1 to perform data analysis. There was no heterogeneity amidst studies for GSTT1 (P=0.72, I²=0%). So the fixed-effects model was performed for data analysis. The pooled data indicated there was a noticeable association between the single nucleotide polymorphism of GSTT1 and the NAFLD susceptibility (null vs. present, OR=1.34, 95%CI: 1.06-1.68, P=0.01; Fig. 3). 

GSTP1 gene polymorphism and NAFLD susceptibility

In total of five studies with 466 cases and 968 controls were used for data pooled. The fixed-effect model was carried out to estimate the association of GSTP1 gene polymorphism and the NAFLD risk by virtue of no heterogeneity among studies (P=0.47, I²=0%). The results indicated a obvious association between GSTP1 gene polymorphism and NAFLD susceptibility (Ile/Val or Val/Val vs. Ile/Ile, OR=1.60, 95%CI 1.23-2.09, P=0.0005; Fig. 4).  

Sensitivity analysis and publication bias

After the omission of an individual study, the recalculated P-value, ORs and 95%CIs did not change substantially. Therefore, the outcomes were considered to be statistically robust and reliable. The funnel plot for assessment of publication bias was not implemented on account of less than 10 researches.

Discussion

Despite the specific pathological mechanism of NAFLD still needs to be explored. Nevertheless, research increasingly revealed that genetic predisposition plays an crucial intrinsic role in the occurrence and development of NAFLD. In addition, SNPs in human might be one of the critical steps to disclose the genetic factor for NAFLD pathogenesis. With further research, GSTs genes as a genetic factor have obtained increasing attention over current years. So far, the present researches have been implemented concerning the relationship of GSTM1, GSTT1 and GSTP1 genes polymorphism and the NAFLD vulnerability with inconsistent conclusions. This inconsistency might be caused by factors like limited sample sizes, confounding factors, as well as clinical heterogeneity of NAFLD. Therefore, we collected the existing evidence and looked into the associations of GSTs genes SNPs and the NAFLD vulnerability via meta-analysis, which could combine data from individual studies, examine and explain the heterogeneity, and increase the statistical power. In conclusion, the merged data suggested a significant correlation between GSTM1, GSTT1 and GSTP1 genes SNPs and the NAFLD vulnerability. Of note, the recalculated P-value, ORs and 95%CIs did not change substantially after the omission of an individual study.

In a word, we performed a meta-analysis of 7 case-control studies that satisfied the inclusion criteria. It should be noted that the current comprehensive analysis was more necessary and meaningful owing to the conclusions of qualified case-control studies are conflicting and contradictory. In this work, it demonstrated that the frequency of GSTM1 null, GSTT1 null and GSTP1-Val allele genotypes in NAFLD patients was remarkably higher than that in healthy subjects. Namely, these genotypes have a significantly increased risk for NAFLD. Furthermore, the outcomes of the meta-analysis were considered to be statistically robust and reliable according to the sensitivity analysis.

GSTs are enzymes in the second-stage detoxification system, which can not only catalyze reduced glutathione sulfhydryl groups, but also neutralize lipid and DNA oxidation products, and have protective effects against endogenous oxidative stress and exogenous toxins [25–26]. Among them, GSTT1, GSTM1 and GSTP1 have garnered considerable attention from various research teams around the world in the recent decade [27]. Several investigations have disclosed that homozygous deletion of GSTM1 and GSTT1 genes (GSTM1 null and GSTT1 null) were connected with lack of relevant GST isoenzyme synthesis and augmented the susceptibility of genetic damage [28–29]. Furthermore, the double null genotypes of GSTT1 and GSTM1 genes could decline the activity of sulfhydryl binding so as to induce insufficient activity of detoxification in the body [30, 31]. GSTP1 gene polymorphism is the outcome of a single nucleotide substitution of A to G, which leads to valine instead of isoleucine in the binding site of GSTP1 and alters catalytic activity of enzyme [32, 33]. Previous reports also suggested that GSTM1/GSTT1 null or GSTP1-Val genotypes were remarkably associated with the vulnerability of hepatis B virus, hepatocellular carcinoma, alcoholic cirrhosis, and NAFLD [34–39]. Moreover, the GSTM1 null genotype was reported to be more common in NAFLD patients than in controls, and GSTP1-Val was proved to be a hazard for NAFLD vulnerability in the Iranian population [40].

Up to now, this is the first synthetical study on the relationship between GSTs polymorphisms and NAFLD vulnerability. There were several strengths in this study. First, to gather a maximum amount of relevant literature, a comprehensive search strategy was adopted to retrieve eligible studies in both English and Chinese databases. Besides, the methodological quality of studies was evaluated via NOS, which allowed for the judgment of potential risk of bias. According to the NOS, all eligible studies were of high methodological quality. Furthermore, sensitivity analyses were carried out in this study, which guaranteed the reliability of the findings.

Nevertheless, there still existed several drawbacks should be acknowledged. First, only seven studies were included, the statistical power was limited, and subgroup analyses were not carried out because of the limited degree of freedom. Second, the absence of HWE in individual studies may lead to information bias. Third, We ignored the synergistic effect of polymorphism at other sites of NAFLD because only three loci in the GST gene were studied in association with susceptibility to NAFLD. Thus, interactions between these loci and genes may result in concealing or amplifying the actual function of individual loci or genes. Leave aside these drawbacks, this study is the first to provide a more accurate and powerful evidence on the association between GSTM1, GSTT1 and GSTP1 genes polymorphisms and NAFLD vulnerability.

Conclusion

In brief, it revealed that GSTM1 null, GSTT1 null and GSTP1-Val genotypes were appreciably associated with augmented risk of NAFLD vulnerability. Concerning limitations of this study, it is necessary to confirm the present findings by complementary studies with larger sample size.

Abbreviations

GSTs: Glutathione S-transferases; HWE: Hardy-Weinberg equilibrium; NAFLD: Nonalcoholic fatty liver disease; NOS: Newcastle-Ottawa Scale; OR: Odds ratio; 95%CI: 95% confdence interval

Declarations

Acknowledgments

Not applicable. 

Author contributions

Ming Qiao conceived of the idea and wrote the manuscript. The literature retrieval and data analysis by Yi Zhu. The authors read and approved the final manuscript. 

Funding

This article was supported by the Natural Science Foundation of Xinjiang Uygur Autonomous Region, China. 

Availability of data and materials

All the data generated in the present research is contained in this manuscript. 

Ethics approval and consent to participate

Not applicable. 

Consent for publication

Not applicable. 

Competing interests

The authors declare that they have no competing interests. 

Author details

¹Department of pharmacy, The First Affiliated Hospital of Xinjiang Medical University, Urumqi, Xinjiang 830011, China

References

1. Dai L, Zhou WJ, Zhong LD, Tang XD, Ji G. Chinese medicine formulas for nonalcoholic fatty liver disease: overview of systematic reviews. World J Clin Cases. 2021;9(1):102–17.
2. Alqahtani SA, Golabi P, Paik JM, Lam B, Younossi ZM. Performance of noninvasive liver fibrosis tests in morbidly obese patients with nonalcoholic fatty liver disease. Obes Surg. 2021; 2021(Suppl 1):1-9.
3. Kolesnikova OV, Potapenko AV, Chupina VI. Influence of resistance on cardiometabolic risk in non-alcoholic fatty liver disease patients combined with subclinical hypothyroidism. Metabolism. 2021;116:154587.
4. White DL, Li D, Nurgalieva Z, Hashem B. Genetic variants of glutathione S-transferase as possible risk factors for hepatocellular carcinoma: a huge systematic review and meta-analysis. Am J Epidemiol. 2008;167(4):377–89.
5. Hashemi M, Eskandari E, Fazaeli A, Bahari A, Ghavami S. Association of genetic polymorphisms of glutathione-S-transferase genes (GSTT1, GSTM1, and GSTP1) and susceptibility to nonalcoholic fatty liver disease in Zahedan, Southeast Iran. DNA Cell Biol. 2012;31(5):672–77.
6. Kasthurina SP, Ramasamy T, Ayyavoo J, Dave DK, Adroja DA. GST M1-T1 null allele frequency patterns in geographically assorted human populations: a phylogenetic approach. Plos One. 2015;10(4):e0118660.
7. Hayes JD, Flanagan JU, Jowsey IR. Glutathione transferases. Annu Rev Pharmacol Toxicol. 2005;45:51–88.
8. Sydorchuk L, Fediv O, Kohaniuk J. Association of glutathione S-transferase gene class T1 (GSTT1) and M1 (GSTM1) with gastroesophageal reflux disease severity and diabetes mellitus. Immunogastroenterology. 2013;2:109–13.
9. Çelįk SK, ArasN, Yildirim Ö, Turan F, Yildirim H, Tamer L. Glutathione S-transferase GSTM1, null genotype may be associated with susceptibility to age-related cataract. Adv Clin Exp Med. 2015;24:113–19.
10. Hayes JD, Strange RC. Glutathione S-transferase polymorphisms and their biological consequences. Pharmacology. 2000;61:154–66.
11. Benabdelkrim M, Djefal O, Berredjem H. GSTM1 and GSTT1 polymorphisms and susceptibility to prostate cancer: a case-control study of the Algerian population. Asian Pac J Can Prev. 2018;19:2853–58.
12. Weich N, Roisman A, Cerliani B, Chertkoff L, Richard SM, Slavutsky I. Gene polymorphism profles of drug-metabolising enzymes GSTM1, GSTT1 and GSTP1 in an Argentinian population. Ann Hum Biol. 2016;12:379–83.
13. Kuleape JA, Tagoe EA, Puplampu P, Bonney EY, Quaye O. Homozygous deletion of both GSTM1 and GSTT1 genes is associated with higher CD4⁺ T cell counts in Ghanaian HIV patients. Plos One. 2018;13(5):e0195954.
14. Luca M, Valentina DA, Consuelo C, Nedovic B, Arzani D, Amore R. A case-control study on the effect of metabolic gene polymorphisms, nutrition, and their interaction on the risk of non-alcoholic fatty liver disease. Genes Nutr. 2014;9(2):383–93.
15. Masaharu H, Kentaro O, Takehiro N. Association between combinations of glutathione-S -transferase, M1, T1 and P1 genotypes and non-alcoholic fatty liver disease. Liver Int. 2008;29(2):164–68.
16. Mohammad H, Ebrahim EN, Aliakbar F, Fazaeli A, Bahari A, Ghavami S. Association of genetic polymorphisms of glutathione-S-transferase genes (GSTT1, GSTM1, and GSTP1) and susceptibility to nonalcoholic fatty liver disease in Zahedan, southeast Iran. DNA Cell Biol. 2012;31(5):672–77.
17. Kentaro O, Masaharu H, Junji S, Morita K, Kajiwara A, Sakata M. Interactive effects of smoking and glutathione S-transferase polymorphisms on the development of non-alcoholic fatty liver disease. Toxicol Lett. 2013; 220(2013):143-49.
18. Singh HO, Lata S, Angadi M, Bapat S, Pawar J, Nema V. Impact of GSTM1, GSTT1 and GSTP1 gene polymorphism and risk of ARV-associated hepatotoxicity in HIVinfected individuals and its modulation. Pharmacogenomics J. 2017;17:53–60.
19. Vasyl P, Olexander V, Iryna P, Ilashchuk T, Sydorchuk L, Prysyazhnyuk P. Glutathione S-transferase T1 and M1 null genotype distribution among non-alcoholic fatty liver disease patients and its association with cytokine and adipokine profiles. Clin Exp Hepat. 2020;6(2):142–49.
20. Prysyazhnyuka VP, Rossokhab ZI, Gorovenko NG. Variation in particular biochemical indicators, cytokine and adipokine profiles of the blood, and the structural and functional parameters of the liver in patients with nonalcoholic fatty liver disease and different genotypes by the polymorphic locus A313G of the GSTP1 Gene. Cytol Genet. 2017;51(6):455–61.
21. Ryckman K, Williams SM. Calculation and use of the Hardy-Weinberg model in association studies. Curr Protoc Hum Genet. 2008; Chapter 1:Unit 1.18.
22. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327:557–60.
23. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21(11):1539–58.
24. Ioannidis JP, Trikalinos TA. The appropriateness of asymmetry tests for publication bias in meta-analyses: a large survey. CMAJ. 2007;176(8):1091–6.
25. Khan AJ, Choudhuri G, Husain Q, Parmar D. Polymorphism in glutathione-S-transferases: a risk factor in alcoholic liver cirrhosis. Drug Alcohol Depen. 2009;101(3):183–90.
26. Kassogue Y, Dehbi H, Quachouh M. Association of glutathione S-transferase (GSTM1 and GSTT1) genes with chronic myeloid leukemia. Springerplus. 2015;4(1):1–5.
27. Kaltoum AB, Sellama N, Hind D, Yaya K, Asma Q. Association of glutathione s-transferase genes (M1 and T1) with the risk of acute myeloid leukemia in a moroccan population. Middle East J Cancer. 2016;8(1):7–12.
28. Ghorbel R, Salah G, Ghorbel R, Mahmoud AB. Do GSTM1 and GSTT1 polymorphisms infuence the risk of developing mitochondrial diseases in a Tunisian population. Environ Sci Pollut Res. 2018;25(23):1–9.
29. Ellithy HN, Yousri S, Shahin GH. Relation between glutathione S-transferase genes (GSTM1, GSTT1, and GSTP1) polymorphisms and clinical manifestations of sickle cell disease in Egyptian patients. Hematology. 2015;20(10):598–606.
30. Kassab A, Msolly A, Lakhdar R, Gharbi O, Miled A. Polymorphisms of glutathione-S-transferases M1, T1, P1 and susceptibility to colorectal cancer in a sample of the Tunisian population. Med Oncol. 2013;31:1–6.
31. Polimanti R, Carboni C, Baesso I. Genetic variability of glutathione S-transferase enzymes in human populations: functional inter-ethnic diferences in detoxifcation systems. Gene. 2013;512(1):102–7.
32. Wu K, Wang X, Xie Z, Liu Z, Lu Y. Glutathione s-transferase p1 gene polymorphism and bladder cancer susceptibility: an updated analysis. Mol Bio Rep. 2013;40(1):687–95.
33. Zarebska A, Jastrzebski Z, Ahmetov II, Zmijewski P, Cieszczyk P, Leonska A. GSTP1 c.313A>G polymorphism in Russian and Polish athletes. Physiol Genomics. 2017;49(3):127–31.
34. Li T, Meng Q, Zou Z. Correlation between promoter methylation of glutathione-S-transferase P1 and oxidative stress in acute-on-chronic hepatitis B liver failure. J viral hepat. 2011;18(7):226–31.
35. Goncharova IA, Rachkovskii MI, Beloborodova EV. Liver cirrhosis pathogenetics: polymorphism of glutation s-transferase genes. Mol Bio. 2010;44(3):431–38.
36. Sharma M, Gupta S, Singh K, Mehndiratta M, Gautam A, Kalra OP. Association of glutathione-S-transferase with patients of type 2 diabetes mellitus with and without nephropathy. Diabetes Metab Syndr Clin Res Rev. 2016;10:194–7.
37. Rong SL, Zhou XD, Wang ZK, Wang XL, Li B. Glutathione S-Transferase M1 and T1 polymorphisms and hypertension risk: an updated meta-analysis. J Hum Hypertens. 2018;33:454–65.
38. Chirilă DN, Chirilă MD, Turdeanu NA, Pop TR. The glutathione S-transferases (GSTS) gene polymorphisms in hepatocellular, pancreatic and gallbladder cancers. Hum Vet Med. 2016;8:34–40.
39. Ma J, Zhu SL, Liu Y, Huang XY, Su DK. GSTP1 polymorphism predicts treatment outcome and toxicities for breast cancer. Oncotarget. 2017;8(42):72939–49.
40. Tamandani D, Mohammad H, Elnaz B, Bahari A, Valizadeh J, Torkamanzehi A. Lack of association of GSTT1 and GSTP1 genes methylation and their expression profiles with risk of NAFLD in a sample of Iranian patients. Clin Res Hepatol Gas. 2011;35(5):387–92.