Stickel F, Österreicher CH, Datz C, Ferenci P, Wölfel M, Norgauer W, Kraus MR, Wrba F, Hellerbrand C, Schuppan D. Prediction of Progression to Cirrhosis by a Glutathione S-Transferase P1 Polymorphism in Subjects With Hereditary Hemochromatosis. Arch Intern Med. 2005;165(16):1835-1840. doi:10.1001/archinte.165.16.1835
Copyright 2005 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2005
Oxidative stress plays an important pathogenic role in hereditary hemochromatosis (HHC) and chronic hepatitis C virus infection (CHC). Several enzymes involved in the degradation of reactive oxidants and xenobiotics, such as glutathione S-transferase P1 (GSTP1) and manganese superoxide dismutase (MnSOD), reveal polymorphisms that affect their antioxidant capacity and may therefore modulate the progression to cirrhosis. Our objective was to establish the role of the functional polymorphisms of GSTP1 (codon 105 Ile→Val) and MnSOD (codon 16 of precursor protein Ala→Val) on the evolution of cirrhosis in patients with HHC and CHC.
One hundred seventy-two patients with HHC who were homozygous for the C282Y mutation and 285 patients with CHC underwent liver biopsy and genotyping for the GSTP1 and MnSOD polymorphisms.
In HHC, the GSTP1 Val/Val genotype was more common in patients with than in those without cirrhosis (14.8% vs 2.1%, P = .009), whereas the distribution of MnSOD variants was not different. Logistic regression analysis identified GSTP1 Val/Val genotype, serum ferritin level, male sex, and age as independent predictors for the presence of cirrhosis. The odds ratio for the GSTP1 Val/Val genotype for the development of cirrhosis was 3.85 (95% confidence interval, 1.18-12.62; P = .03). However, in patients with CHC, the GSTP1 and MnSOD genotypes were not associated with cirrhosis.
Cirrhosis is more likely to develop in C282Y homozygotes with the GSTP1 Val/Val genotype than in those with non-Val/Val genotypes, which in part explains the variable phenotypic expression of HHC and highlights the central role of oxidative stress in its pathogenesis.
Liver cirrhosis may develop as a consequence of various chronic liver diseases, including hereditary hemochromatosis (HHC)1 and chronic hepatitis C virus infection (CHC).2 Hereditary hemochromatosis is among the most common genetic diseases, affecting 1 in 200 to 400 individuals of Northern European descent.3 Several mutations within the hemochromatosis gene (HFE) lead to excess iron deposition in the liver and other organs, and among affected individuals, 80% to 100% are homozygous for the C282Y mutation. However, with HFE genotyping becoming widely available, population studies showed a paradox of a high rate of homozygosity for the C282Y mutation in the population vs a relative infrequency of clinical disease.4 For example, Beutler et al5 found 152 C282Y homozygotes among 41 038 healthy subjects attending a health appraisal clinic, but only 1 had signs or symptoms suggestive of HHC. Thus, clinically overt disease does not develop in many individuals who carry the HFE mutation and therefore a susceptibility for overload of iron levels (hereafter referred to as iron overload), and their long-term prognosis remains unclear.
The major cause of chronic hepatitis infection in the western world is CHC, which affects more than 170 million people worldwide and at least 3 million in the United States.6 However, the natural course of CHC is highly variable, and studies indicate that infection progresses to cirrhosis within the first 2 decades of hepatitis C virus (HCV) infection in only 5% to 20% of infected individuals, whereas advanced liver damage may never develop in some patients throughout their lifetime.7,8 Thus, defining the prognosis before therapy is important when considering the substantial adverse effects, high costs, and limited efficacy of current interferon-based therapies.
Besides environmental factors such as diet, alcohol consumption, and coinfection with hepatitis B virus (HBV) and/or human immunodeficiency virus that influence the progression of liver damage in HHC and CHC, the evolution of cirrhosis may be further modulated by certain host factors including sex, age, obesity, and immune status. Moreover, the clinical diversity of HHC could be related to yet unidentified genes involved in iron metabolism,9 whereas the aggravation of histological damage in CHC could partly result from functional polymorphisms of genes involved in antiviral response, fibrogenesis, and/or inflammation.10 In HHC and CHC, oxidative stress derived from reactive oxygen species and lipid peroxides plays a major role in the pathogenesis of liver cirrhosis, because oxidative stress can trigger hepatic fibrogenesis.11,12 Thus, genetic polymorphisms of antioxidant enzymes could modify the capacity to alleviate oxidative stress and thereby influence the progression of liver damage.
Glutathione S-transferases are a family of sulfur-containing enzymes (GSTA, GSTM, GSTT, and GSTP) that inactivate reactive oxygen species and many toxic and carcinogenic xenobiotics through conjugation with glutathione.13 The isoenzyme GSTP1 is expressed in the liver, and its gene reveals a functional polymorphism leading to an amino acid substitution (codon 105 Ile→Val) that results in a markedly decreased capability of GSTP1 to conjugate certain toxins with glutathione.14 Glutathione S-transferase P1 also deactivates 4-hydroxynonenal, a highly reactive product of lipid peroxidation and promotor of fibrogenesis.15 For GSTP1, a significant association between the GSTP1 Val/Val genotype and the development of cryptogenic liver cirrhosis was demonstrated.16
Mitochondria-derived reactive oxygen species are detoxified to hydrogen peroxide and water by the successive action of manganese superoxide dismutase (MnSOD) and glutathione peroxidase, respectively.17 Manganese superoxide dismutase is synthesized with a cleavable target sequence that enables its transport into mitochondria.18 A polymorphism leads to alanine (Ala) or valine (Val) at amino acid position −9 of the target sequence.19 This results in its enhanced import into mitochondria and a 40% higher MnSOD activity in case of Ala homozygosity.20
In an attempt to establish new prognostic markers for the development of severe liver disease, we investigated whether genetic polymorphisms of GSTP1 and MnSOD are associated with cirrhosis in patients with HHC and CHC.
One hundred seventy-two untreated patients homozygous for the HFE mutation C282Y with elevated levels of iron laboratory values or elevated liver enzyme levels (aspartate aminotransferase level, >1.5 times the upper limit of the reference range) at screening underwent diagnostic liver biopsy for suspected HHC. Fifteen patients were excluded because of the presence of additional chronic liver disease (α1-antitrypsin deficiency in 1; autoimmune hepatitis in 1; HBV coinfection in 1; HCV coinfection in 4; and significant alcohol intake [defined as a daily intake of >60 g in men and >20 g in women] in 8). Cutoff levels for chronic alcohol consumption were chosen according to World Health Organization recommendations. For the histological examination, 4-μm sections were stained with hematoxylin-eosin, Masson trichrome, and Perls Prussian blue. Slides were first reviewed by the local pathologist and then reassessed by one of us (F.W.) who was unaware of the clinical data. Fibrosis was staged according to the method of Ishak et al.21
In addition, 325 patients with CHC underwent screening for eligibility. Forty subjects were excluded because of liver-associated comorbidities (autoimmune hepatitis in 2; HBV coinfection in 12; and significant alcohol intake of >60 g in men and >20 g in women in 26). Finally, 285 patients with CHC underwent analysis. Chronic viral hepatitis infection was confirmed by positive HCV RNA findings on an HCV monitor (Cobas Amplicor; Roche, Basel, Switzerland) and elevated alanine aminotransferase levels at least 1.5 times the upper limit of the reference range. Hepatitis C virus genotypes were determined using the Inno-LipQA HCV II test (Innogenetics, Ghent, Belgium). All patients underwent percutaneous liver biopsy, and histological findings were evaluated according to Ishak et al.21
To determine the genotype and allele frequency of both genes in the general population from which all patients were recruited, 160 randomly selected subjects (age range, 18-58 years) among hospital and research staff from the participating centers underwent genotyping and served as a community-based control group. All had alanine aminotransferase levels within the reference range and negative HBV surface antigen and anti-HCV findings.
For the genotyping analysis, the patients’ DNA and clinical data were coded, and the identity of the patients remained unknown to those performing the tests until the final analysis. The study protocol was reviewed and approved by the local ethics committees of the participating centers (Erlangen-Nürnberg, Regensburg, and Würzburg in Germany, and Salzburg and Vienna in Austria), and all recruited patients gave informed and written consent.
Serum iron, serum ferritin, and transferrin levels were assessed biochemically. Hepatic iron concentration was measured by atomic absorption spectrophotometry from a portion of the liver biopsy specimen.22 The hepatic iron index was calculated by dividing the hepatic iron concentration by the age of the patient in years and by the molecular weight of iron. Transferrin saturation was calculated as follows: (serum iron level × 70.9)/(serum transferrin level).
For mutation analysis, genomic DNA was isolated from peripheral blood using a commercial DNA extraction kit (Qiagen GmbH, Hilden, Germany). The C282Y mutation of the HFE gene was determined as described previously.23 The GSTP1 and MnSOD polymorphisms were analyzed by polymerase chain reaction and restriction fragment length polymorphism using primer pairs and digestion enzymes as reported by Henrion-Caude et al24 and Degoul et al,25 respectively.
To ensure the reliability of the genotyping methods, 40% of the DNA samples underwent genotyping twice by 2 examiners (F.S. and Jörg Distler) who were unaware of the patients’ status.
Differences in means were assessed by the unpaired, 2-tailed t test and Mann-Whitney test. Categorical data were analyzed by χ2 and Fisher exact tests. Binary logistic regression analysis was calculated to assess the impact of laboratory and clinical data and of the GSTP1 and MnSOD genotypes on presence of cirrhosis. Two-tailed P values less than .05 were considered significant. We used SPSS 10.0 statistical software (SPSS Inc, Chicago, Ill) for all calculations. Testing for deviation from Hardy-Weinberg equilibrium was performed by using genetic statistics (available at: http://ihg.gsf.de/ihg/snps). Sample-size power calculation was performed by using the home page of the Department of Medical Statistics, University of Vienna, Vienna, Austria (available at: http://www.univie.ac.at/medstat).
Demographic and laboratory data of patients with HHC are summarized in Table 1. Those with cirrhosis were older (P<.001), predominantly male (P<.001), and had higher levels of serum ferritin (P<.001), higher transferrin saturation indices (P<.001), and higher hepatic iron concentrations (P = .02).
The clinical characteristics of patients with CHC are displayed in Table 2.
Results from the genotyping analysis are displayed in Table 3. The genotype and allele frequencies of the 2 genes were similar in patients with CHC and HHC and were in accordance with the Hardy-Weinberg equilibrium in all 3 groups. The genotype and allele distributions were the same as in the healthy control subjects.
In patients with HHC, the distribution of GSTP1 genotypes differed significantly (P = .009) between patients with and those without cirrhosis (Table 4). Patients with cirrhosis carried the GSTP1 Val/Val genotype more frequently than did noncirrhotic patients (14.8% vs 2.1%; odds ratio [OR], 7.33; 95% confidence interval [CI], 1.47-36.52; P = .006). Five (12.8%) of the 39 women, but 56 (47.5%) of the 118 men had cirrhosis (P<.001). Only 1 (2.6%) of the 39 women and 10 (8.5%) of the 118 men carried the GSTP1 Val/Val genotype. The single woman (P = .13) and 8 of the 10 men (P = .046) who were GSTP1 Val/Val genotype carriers had histologically confirmed cirrhosis. Otherwise, patients with the GSTP1 Val/Val genotype did not differ significantly from those with a non-Val/Val genotype (mean ± SD age, 53.1 ± 13.6 vs 51.1 ± 13.3 years; median serum ferritin level, 22 790 ng/mL [range, 1320-75 980 ng/mL] [51 209 pmol/L (range, 2966-170 727 pmol/L)] vs 14 700 ng/mL [range, 340-526 000 ng/mL] [33 031 pmol/L (range, 764-1 181 922 pmol/L)]; median transferrin saturation index, 89.99 [range, 70.49-100] vs 85.71 [range, 23-140]; mean ± SD hepatic iron concentration, 11 566 ± 7132 vs 12 249 ± 8612 μg/g of dry weight; and median hepatic iron index, 3.28 [range, 0.80-5.78] vs 3.53 [range, 0.14-14.38]). The distribution of GSTP1 genotypes were not in Hardy-Weinberg equilibrium in patients without cirrhosis. According to the Hardy-Weinberg equilibrium, 8 patients with the Val/Val genotype would have been expected, but only 2 patients were observed (P = .005).
In contrast to GSTP1, MnSOD genotypes showed a similar distribution in cirrhotic and noncirrhotic patients with HHC (Table 4).
Binary logistic regression analysis with forward and backward stepward inclusion of variables identified male sex, GSTP1 Val/Val genotype, greater age, and higher levels of serum ferritin as independent predictors for the histological presence of cirrhosis (Table 5). The GSTP1 Val/Val genotype was associated with an adjusted OR of 3.85 (95% CI, 1.18-12.62; P = .03) with regard to the development of liver cirrhosis in patients with HHC.
Unlike patients with HHC, allele frequencies and genotype distribution of MnSOD and GSTP1 did not show an association with histological severity in patients with CHC (Table 4), even after correction for potential confounders such as age, sex, duration of infection, obesity, and viral genotype.
Increasing evidence suggests that, in addition to sex and ethnicity, unknown genetic factors play a role in determining the susceptibility to and the progression of liver diseases related to iron overload, chronic viral infection, and autoimmunity.26 Obviously, establishing such factors would greatly improve the management of chronic liver diseases, because they help to identify patients at risk and offer opportunities for prevention.
To our knowledge, the GSTP1 Val/Val polymorphism represents the first identified genetic risk factor for an increased likelihood of progression to liver cirrhosis in patients with HHC. The presence of cirrhosis at the initial clinical presentation is vital in the management of HHC for 2 reasons. When HHC is diagnosed and treated adequately before cirrhosis has developed, life expectancy is not significantly reduced, whereas the risk for the evolution of hepatocellular carcinoma is greatly increased in individuals with established cirrhosis, despite phlebotomy treatment.27- 30
In our study, 9 of 11 patients with the Val/Val genotype had cirrhosis, whereas only 2 patients had no cirrhosis. One of these 2 patients was only 33 years of age and cirrhosis might not have developed owing to the relatively short duration of the disease. The second patient, however, was 51 years of age, and histological analysis showed extensive iron deposition but only mild fibrosis. At the time of the biopsy, this patient had a serum ferritin level of 9170 ng/mL (20 605 pmol/L). According to the current guidelines, a liver biopsy would not have been performed in this case, because this is usually recommended only in patients with serum ferritin levels above 10 000 ng/mL (22 470 pmol/L).31 Although there remains a possibility of sampling error in obtaining liver biopsies, this is rather unusual in patients with HHC because of the homogeneous distribution of hepatic iron stores and fibrosis.32 We therefore conclude that the presence of the GSTP1 Val/Val genotype represents a strong determinant for progression to liver cirrhosis but may not inevitably lead to the development of cirrhosis in some patients because of yet unidentified protective mechanisms. This is further supported by the observation that noncirrhotic patients with HHC were not in Hardy-Weinberg equilibrium. The GSTP1 Ile/Ile and Ile/Val genotypes may therefore protect subjects with HHC from the development of cirrhosis.
Iron deposition is age and sex dependent.33 A potentially confounding effect of age and sex on our results was excluded by binary logistic regression analysis, which is considered the most adequate statistical method for the study of the impact of certain genes on disease development and progression.34 Concomitant conditions that may accelerate the progression of liver damage in HHC, such as CHC and/or regular alcohol consumption, were accounted for by excluding patients with these comorbidities.
In HHC, the most important pathomechanism of tissue injury with subsequently enhanced fibrogenesis is oxidative stress resulting from iron overload that may not be counteracted by the physiological antioxidant defense systems.11 Therefore, antioxidant enzymes such as GSTP1 and MnSOD, which can alleviate oxidative pressure, are of particular interest in patients with iron overload. These enzymes protect the cell from reactive oxygen species that arise through normal metabolic processes and at increased rates in iron overload.35 Alterations in the activity of these enzymes can have a major effect on the ability of cells to resist oxidative pressure.36 Therefore, we can speculate that individuals with the GSTP1 Val/Val genotype have less ability to neutralize lipid peroxidation products.
Our findings accord with data from a recent study in which an association between the GSTP1 Val/Val genotype and the development of cryptogenic cirrhosis was reported,16 but they conflict with those of a previous study that demonstrated an association between the GSTP1 Ile/Ile genotype and the manifestation of liver fibrosis in children with cystic fibrosis.24 The latter may be due to the fact that oxidative stress plays only a minor role in the pathophysiology of cystic fibrosis, and genetic variants of antioxidant enzymes that modulate oxidative stress are therefore unlikely to influence the progression of liver damage in this setting.
To our knowledge, our study is the first to investigate the role of MnSOD polymorphisms in HHC, but we could not detect an association between cirrhosis and certain genotypes. However, our results support those from a recent study by Stewart et al,37 who included 281 alcoholic patients with severe liver disease and found no association between alcoholic liver cirrhosis and the MnSOD Ala/Ala genotype.
With regard to CHC, previous studies in patients have demonstrated the presence of oxidative stress,38,39 and data from animal experiments suggest that the HCV core antigen alters the hepatic antioxidant status and promotes the generation of lipid peroxides.40 However, according to our data, polymorphisms of MnSOD and GSTP1 do not confer risk for the development of cirrhosis in CHC.
Most liver diseases are polygenic, and the clinical picture is based on the interaction of at least several genes with a known or unknown environmental factor. These disorders do not show a classical mendelian recessive or dominant inheritance attributable to a single gene mutation. Therefore, the identification of genetic factors affecting the natural course of polygenic disorders is considerably more difficult than identification in diseases with a single genetic defect such as Wilson disease. In humans, case-control studies are a valuable tool to study the potential effects of functional genetic variants on the course of the disease by comparing genotype and allele frequencies of tested genes between case and control groups. However, this approach is subject to potential pitfalls, and several prerequisites for the design of association trials in liver disease were recently recommended.41 It is important to define a sound rationale for the candidate genes, a plausible a priori hypothesis based on the functional significance of the studied polymorphisms, and a detailed characterization of cases ensuring comparable exposure to the pathogenic trigger. Furthermore, the study sample should have a homogeneous genetic background to avoid ethnic heterogeneity and exclude confounding factors. Finally, an appropriate statistical analysis with regard to the applied method, a reasonable sample size, and the level of significance is needed. We accounted for all of these requirements in this study. A sample size of 150 or more has recently been defined as a critical threshold for the replication validity of genetic association studies.42
We clearly show that the GSTP1 Val/Val polymorphism contributes to the development of cirrhosis in HHC, thereby partially explaining the clinical paradox of a highrate of homozygosity for C282Y in the population vs the relative infrequency of clinical disease. Moreover, our data emphasize the role of oxidative stress as a major pathogenic insult in HHC. However, the relatively low prevalence of the GSTP1 Val/Val genotype cannot explain the development of cirrhosis in patients with non-Val/Val genotypes. Therefore, the search for additional contributing genetic polymorphisms will have to continue, to establish an even larger panel of genes that determine the risk profile of progression to cirrhosis in HHC.
Correspondence: Felix Stickel, MD, Department of Medicine I, University of Erlangen-Nürnberg, Ulmenweg 18, D-91054 Erlangen, Germany (firstname.lastname@example.org).
Accepted for Publication: January 26, 2005.
Financial Disclosure: None.
Funding/Support: This study was supported by a research fellowship from the Interdisciplinary Centre of Clinical Research of the University of Erlangen-Nürnberg, Erlangen, Germany (Dr Stickel); and in part by grant 01.03.14.1 from the Funds for Research and Training of the University of Erlangen-Nürnberg.
Additional Information: Drs Stickel and Österreicher have performed the study with equal contribution to study design, collection and interpretation of data, and preparation of the manuscript and therefore share premier authorship.
Acknowledgment: We thank Jörg Distler for valuable technical assistance.