SEARCH

SEARCH BY CITATION

Keywords:

  • cirrhosis;
  • epidermal growth factor;
  • genetic polymorphism;
  • hepatitis B virus

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

Because epidermal growth factor (EGF) up-regulation is characteristic of the cirrhotic liver, we hypothesised that the EGF rs4444903 A > G functional polymorphism might be associated with a worse disease course in patients with chronic HBV infection. To verify this hypothesis, 170 HBV-positive patients (125 males) with a median age of 52 years were studied. Sixty-two of these patients were followed longitudinally for a median time of 21 years. Genotyping for the EGF rs4444903 A > G polymorphism was performed by the polymerase chain reaction-based restriction fragment length polymorphism assay. In the cross-sectional study, the EGF rs4444903 A > G polymorphism genotypic frequencies significantly differed between transplant patients (A/A = 20·4%, A/G = 52·3%, G/G = 27·3%) and HBsAg+ carriers (active and inactive: A/A = 35·7%, A/G = 47·6%, G/G = 16·7%, P = 0·036 for the linear trend). In the longitudinal study, the EGF rs4444903 A > G polymorphism was found to be an independent predictor of cirrhosis development (O.R. 7·73, 95% C.I. 1·21–49·5, P = 0·007). Three groups of patients were identified: A/A female homozygotes (n = 9), A/A male homozygotes (n = 13) and carriers of the G allele of either gender (n = 40). Cirrhosis did not occur among A/A females (n = 0/9), seldom occurred among A/A males (n = 2/13) and reached the highest frequency among G/* patients (n = 13/40, P = 0·026). In conclusion, the EGF rs4444903 A > G polymorphism appears to be associated with an unfavourable disease course of chronic HBV infection and cirrhosis development. This effect might be modulated, at least in part, by the gender of the patient.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

Chronic hepatitis B virus (HBV) infection remains a major public health issue due to its worldwide distribution and the approximately 2 billion infected individuals [1]. The clinical spectrum of HBV infection is wide, ranging from subclinical infection to end-stage liver disease and hepatocellular carcinoma (HCC), with a natural history lasting several decades [2–4]. On the basis of their clinical, viral and histological features, persistently HBV-infected patients can be categorised either as active or inactive carriers. Inactive carriers are hepatitis B e antigen (HBeAg)-negative, have low or undetectable HBV-DNA levels, normal serum alanine aminotransferase (ALT) values and minimal or no damage (inflammation, fibrosis) on liver biopsy. Approximately 0·5–1% of HBV carriers clear HBsAg each year, and most of the carriers seroconvert to anti-HBs [5]. The prognosis of subjects who have cleared HBsAg is generally good, although those who are older or have progressed to cirrhosis before HBsAg clearance remain at risk of developing HCC [6]. Active carriers may be either HBeAg-positive or negative but present high levels of HBV-DNA, abnormal ALT levels, necro-inflammatory activity and/or significant fibrosis on liver biopsy [2]. Fifteen to 40% of HBV active carriers may reach end-stage liver disease and/or HCC. In patients with chronic hepatitis B, several factors have been associated with a progressive course, including older age, longer duration of infection, male gender, HBV genotype C, high HBV-DNA levels and habitual alcohol consumption [2]. Co-infection or super-infection by hepatitis C virus (HCV) or hepatitis D virus may also favour progression [7].

The epidermal growth factor (EGF) rs4444903 polymorphism involving an A > G transition at position 61 of the 5’-untranslated region of the gene is associated with the occurrence of several types of cancer, including malignant melanoma, oesophageal cancer, gastric cancer, breast cancer and colorectal cancer [8–14]. Carriers of the G allele, especially G/G homozygotes, may have an increased risk of developing HCC as shown in European and Asian series [15–17]. The association with HCC, however, could not be confirmed by Qi et al. [18].

EGF up-regulation is characteristic of the cirrhotic liver [19]. However, it has not been clarified whether the EGF rs4444903 A > G polymorphism affects the outcome of chronic HBV infection. The aim of the present study was to ascertain the relationship between the EGF rs4444903 A > G polymorphism and the progression to more advanced stages of chronic HBV-related liver disease.

Patients and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

Patients

In an initial, retrospective cross-sectional study (Table 1), 170 consecutive HBV-positive patients (125 males, 45 females) with a median age of 52 years (range, 22–86 years) were included. These patients were classified into the following three groups: a) HBsAg inactive carriers (n = 30) who had serum HBV-DNA ≤2000 UI/ml, transaminases persistently below the upper normal limit in at least three different determinations and no evidence of liver disease; b) active carriers with chronic hepatitis B (n = 96) who had serum HBV-DNA >2000 UI/ml and transaminases persistently above the upper normal limit; and c) HBV-related end-stage liver disease patients (n = 44) who were referred for liver transplantation. Eighteen patients in the last group had HCC confirmed by histological evaluation of the explanted liver. A group of 209 healthy blood donors (164 males, 45 females) with a median age 50 years (range, 23–70 years) served as controls. Seventy-nine of these subjects had evidence of resolved HBV infection (anti-HBs and anti-core seropositivity).

Table 1.  Clinical and demographic characteristics of the studied population (n = 379). Patients were categorised in the following groups: a) inactive HBsAg carriers, b) active carriers with chronic hepatitis B and c) HBsAg-positive liver transplant candidates. Controls were divided according to the presence or absence of anti-HBc. Variables are presented as frequencies (%) or as medians (range).
 Male genderAge in years
n (%)Median (range)
  1. Anti-HBc, hepatitis B core antibody; HbsAg, hepatitis B surface antigen.

Controls, n = 209  
 Anti-HBc–, n = 13098 (75·4)43 (23–64)
 Anti-HBc+, n = 7966 (83·5)62 (37–70)
Patients, n = 170  
 HBsAg+ inactive carriers, n = 3020 (66·7)53 (24–81)
 Active carriers with chronic hepatitis B, n = 9670 (72·9)51 (22–86)
 HBsAg+ transplant candidates, n = 4435 (79·5)55 (36–63)

The retrospective longitudinal study included a cohort of 62 HBsAg-positive patients (42 males, 20 females). The median age at presentation was 30 years (range 14–63). All of these patients have also been included in the cross-sectional study. Thirty-one (50%) patients were HBeAg-positive, and none of the patients were HCV- or hepatitis delta virus (HDV)-positive. These patients underwent periodic clinical and laboratory monitoring for a median time of 21 years (range, 0·6–33 years). All patients underwent liver biopsy during the follow-up period. Twenty-four patients received antiviral treatment during the follow-up: 15 were treated with interferon (IFN) or pegylated IFN and 9 with nucleoside analogues. All study participants were of European descent, negative for HDV and HCV and gave informed written consent. The study was conducted in strict accordance with the principles of the World Medical Association Declaration of Helsinki, and the procedures received approval by a local ethics committee (Independent Ethical Committee of the Udine Teaching Hospital).

EGF genotyping

The EGF rs4444903 A > G polymorphism assessment was performed by the polymerase chain reaction (PCR)-based restriction fragment length polymorphism assay. Genomic DNA was extracted from whole blood samples by the QIAamp DNA blood mini kit (Qiagen, Milan, Italy) according to the manufacturer's instructions. A 260-base pair (bp) product was obtained with the following primers: forward 5’-CATTTGCAAACAGAGGCTCA-3’ and reverse 5’-GTTTAACAGCCCTGCTCTGG-3’. These primers were newly designed with the aid of the NCBI Primer-Blast Tool (http://www.ncbi.nlm.nih.gov/tools/primer-blast/). PCR amplification was carried out in a total volume of 10 µl containing 10 mM Tris-HCl (pH 8·3), 50 mM KCl, Tween-20 0·01%, 0·2 mM deoxyribonucleotides, 2–4 pmol of each primer, 2·0 mM MgCl2 and 0·5 units of hot-start Euro Taq DNA polymerase (Euroclone, Milan, Italy). Samples containing 10 ng of genomic DNA were subjected to 40 cycles of denaturation (at 95°C for 30 s), annealing (at 61°C for 30 s) and elongation (at 72°C for 45 s) using a Techne TC-412 thermal cycler. In a total volume of 20 µl, 10 µl of the amplicons were digested with 1 unit of the Alu-I restriction endonuclease (New England Biolabs, Hitchin, UK) at 37°C overnight. The digested fragments were 94 bp + 91 bp + 60 bp + 15 bp for the A allele and 185 bp + 60 bp + 15 bp for the G allele variant. The fragments were resolved by electrophoresis in 3·5% agarose gel after staining with ethidium bromide. In 20 patients, the genomic region encompassing the EGF rs4444903 A > G polymorphism was sequenced to confirm the results obtained by the restriction fragment length polymorphism assay.

Statistical analysis

Statistical analysis of the data was performed using the BMDP dynamic statistical software package 7·0 (Statistical Solutions, Cork, Ireland). Continuous variables were presented as medians (range), whereas categorical variables were expressed as frequencies (%). The chi-square G-test goodness of fit was employed to verify whether the proportions of the polymorphism were distributed in controls and patients in accordance with Hardy-Weinberg equilibrium. The differences in allelic and genotypic frequencies between the groups was assessed by Pearson's chi-square test (chi-square test for linear trend where appropriate) and the calculation of the odds ratio with 95% confidence intervals. Stepwise logistic regression analysis with a forward approach was used to verify whether the EGF rs4444903 A > G polymorphism was associated with cirrhosis occurrence independently from other confounding covariates.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

EGF rs4444903 A > G polymorphism frequencies in the cross-sectional study

Table 2 illustrates the allelic and genotypic frequencies in HBsAg-positive patients and controls. In both groups, genotypic frequencies had a distribution that was not significantly different from expected values according to the Hardy-Weinberg equilibrium. The EGF rs4444903 A > G polymorphism allelic and genotypic frequencies significantly differed among groups. The carriage of the G allele was increased in HBsAg+ transplant candidates in comparison to HBsAg+ carriers (active plus inactive). This result was confirmed when only transplant candidates without HCC were taken into account (Table 2). When genotypic frequencies were considered, HBsAg+ transplant candidates were found to carry the A/G and G/G genotypes more frequently (P = 0·036 for linear trend) than the HBsAg+ carriers (active plus inactive). Again, this result was confirmed when only transplanted patients without HCC were considered (P = 0·045 for linear trend).

Table 2.  Frequencies and proportions of the EGF rs4444903 A > G polymorphism alleles and genotypes in the controls and different groups of HBsAg-positive patients. The odds ratio and 95% confidence interval for the EGF rs4444903 A > G allelic frequencies in transplant candidates compared to the other groups were also included. Statistical analysis refers to Pearson's chi-square test.
 A/AA/GG/GAG
n (%)n (%)n (%)
  1. *HBsAg+ inactive carriers plus chronic hepatitis HBsAg+. **HBsAg+ transplant candidates without hepatocellular carcinoma. aHBsAg+ transplant candidates versus HBsAg+ inactive carriers plus chronic hepatitis HBsAg+: O.R. 1·686, 95% C.I. 1·036–2·742, P = 0·035. bHBsAg+ transplant candidates without hepatocellular carcinoma versus HBsAg+ inactive carriers plus chronic hepatitis HBsAg+: O.R. 1·854, 95% C.I. 0·975–3·536, P = 0·043. Anti-HBc, hepatitis B core antibody; HbsAg, hepatitis B surface antigen.

Controls, n = 209     
 Anti-HBc-, n = 13044(33·9)61(46·9)25(19·2)0·5730·427
 Anti-HBc+, n = 7929(36·7)38(48·1)12(15·2)0·6080·392
Patients, n = 170     
 HBsAg+ inactive carriers, n = 309(30·0)18(60·0)3(10·0)0·6000·400
 Chronic hepatitis B, n = 9636(37·5)42(43·7)18(18·8)0·5940·406
 HBsAg+ transplant candidates, n = 449(20·4)23(52·3)12(27·3)0·4660·534a
 HBsAg+ carriers*, n = 12645(35·7)60(47·6)21(16·7)0·5950·405
 HBsAg+ transplant candidates**, n = 265(19·2)13(50·0)8(30·8)0·4420·558b

EGF rs4444903 A > G polymorphism frequencies in the longitudinal study

In the longitudinal study, 62 patients (final median age 51 years, range 28–81) were followed-up for a median of 21 years (range, 0·6–33 years). At the end of the follow-up, 27 of the 62 patients converted to the inactive carrier status; 10 lost HBsAg. Fifteen patients had biopsy proven liver cirrhosis, which was already present at the entry into the study in 7 patients; none developed HCC. Among the pertinent demographic and clinical parameters reported in Table 3, cirrhosis at entry was found to be inversely associated with younger age and presence of the A allele of the EGF rs4444903 A > G polymorphism. At the end of the follow-up (Table 4), older age and presence of the EGF rs4444903 A > G polymorphism G/* genotypes and as well as being an HBV active carrier were found to predict the occurrence of cirrhosis. Using stepwise logistic regression analysis, among the variables presented in Table 4, the EGF rs4444903 A > G polymorphism was found to be an independent predictor of cirrhosis (O.R. 7·73, 95% C.I. 1·21–49·5, P = 0·007). A synergistic effect was detected between the gender of the patient and the EGF rs4444903 A > G polymorphism in predicting the occurrence of cirrhosis. Three groups of patients were identified: A/A female homozygotes (n = 9), A/A male homozygotes (n = 13) and carriers of the G allele of either gender (n = 40). Cirrhosis did not occur among A/A females (n = 0/9), seldom occurred among A/A males (n = 2/13) and reached the highest frequency among G/* patients (n = 13/40, P = 0·026 for the linear trend, Fig. 1).

Table 3.  Associations in the longitudinal study (n = 62) at the entry in follow-up between the presence of cirrhosis and the main demographic and clinical characteristics of the population. The statistical analysis was performed by Pearson's chi-square test or the chi-square test for linear trend where appropriate.
 Cirrhosis at entry in the studyP
YesNo
n = 8n = 54
  1. HbeAg, hepatitis B e antigen.

Age >30 years, n = 318(100·0%)23(42·6%)0·002
Male gender, n = 425(62·5%)37(68·5%)0·734
HBeAg positivity, n = 312(25·0%)29(53·7%)0·130
Alcohol consumption for at least 5 years, n = 71(12·5%)6(11·1%)0·908
EGF rs4444903 A > G genotypes   
 A/A, n = 220(0·0%)22(40·7%)0·009
 A/G5(62·5%)26(48·2%)
 G/G3(37·5%)6(11·1%)
Table 4.  Associations in the longitudinal study (n = 62) at the end of follow-up between the presence of cirrhosis and the main demographic and clinical characteristics of the population. The statistical analysis was performed by Pearson's chi-square test or the chi-square test for linear trend where appropriate.
 Cirrhosis at the end of the studyP
YesNo
n = 15n = 47
  1. HbsAg, hepatitis B surface antigen; HbeAg, hepatitis B e antigen.

Age >50 years, n = 3111(73·3%)20(42·6%)0·038
Male gender, n = 4211(73·3%)31(66·0%)0·595
HBeAg positivity at entry, n = 315(33·3%)26(55·3%)0·138
HBeAg loss during the follow-up, n = 305(33·3%)25(53·2%)0·180
Alcohol consumption for at least 5 years, n = 72(13·3%)5(10·6%)0·774
EGF rs4444903 G/* genotypes, n = 4013(86·7%)27(57·4%)0·039
Inactive carrier   
 No14(93·3%)21(44·7%)0·002
 Yes without HBsAg loss1(6·7%)16(34·0%)
 Yes with HBsAg loss0(0·0%)10(21·3%)
image

Figure 1. Percentage of patients with or without cirrhosis during the follow-up plotted in relation to the interaction between gender and EGF rs4444903 A > G genotypes. Three groups were identified: females carrying the A/A genotype (n = 9), males carrying the A/A genotype (n = 13), and all other patients (either gender carrying the A/G or G/G genotypes) (n = 40).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

Several studies have searched for possible associations in both Caucasian and Asian ethnic groups between cytokine polymorphisms and the natural history of HBV infection. In particular, polymorphism of cytokines such as interleukin-6 (IL-6), Il-10, Il-18, IFN-gamma and tumour necrosis factor-alpha (TNF) have been related to HBV clearance and/or to an unfavourable disease course of chronic HBV hepatitis [20–24]. Moreover, polymorphisms of genes coding for growth factors, such as those polymorphisms of transforming growth factor-alpha (TGF-α), have been suggested to have a role in the clearance of HBV and the occurrence of HCC [25].

EGF is synthesized as a large precursor molecule that is proteolytically cleaved to generate the 53-amino acid mature form [26]. Cell culture studies have shown that EGF induces hepatic proliferation and that during liver regeneration EGF RNA levels rapidly increase [27,28]. EGF competes for the TGF-α receptors (TGFRs) and shows strong sequential and functional homology with TGF-α. EGFR is a specific high-affinity and low-capacity receptor expressed on the surface of the responsive cells and possesses several biological functions related to its capability to stimulate proliferation and differentiation of epidermal and epithelial tissues [27,29]. Experimental data and clinical studies have demonstrated that EGF over-expression leads to increased susceptibility to cancer of epithelial and mesenchymal origin, including HCC [30,31].

The EGF gene is located in chromosome 4q25-27. EGF rs4444903 A > G is a functional polymorphism because the presence of the G allele is characterised by a two-fold longer messenger RNA half-life that can induce enhanced EGF production and secretion [15]. Large differences have been detected in the allelic and genotypic distributions of this polymorphism among different ethnicities. In Europeans, the frequencies of the G allele and of the G/G genotype average 0·40 and 0·20, respectively, and higher frequencies (0·70 and 0·50) have been described in Far East populations [16,32]. In the present study that involved an Italian population, the frequencies of alleles and genotypes observed in control subjects were very close to those previously reported in Italians and similar to those described in other European populations [10,32,33].

The present study is the first to report that the EGF rs4444903 G allele associates with the severity of HBV infection; this allele is maximally represented among patients that are end-stage liver disease candidates for liver transplantation and is less common among chronic HBV active and inactive carriers. Although the EGF rs4444903 polymorphism has been repeatedly assessed in relation to HCC occurrence in patients with chronic HBV infection or in mixed cohorts, no study has been devoted to specifically ascertain whether this polymorphism could be implicated in the progression of liver fibrosis in patients with HBV chronic infection [15–18]. In the study of Li et al., no difference was detected in allelic and genotypic frequencies between control subjects and patients with liver cirrhosis, whereas the latter frequencies differed from those pertaining to patients affected by HCC [16]. In addition, Qi et al. enrolled a cohort of chronic HBV patients without HCC that were considered to be free of significant liver disease. This cohort was compared to a group of HBV patients with HCC that were affected by liver cirrhosis in approximately a half of the cases. A healthy group of subjects was used as the control group. The authors did not find a significant difference in EGF rs4444903 polymorphism allelic and genotypic frequencies between HBV patients with or without HCC. Moreover, the frequencies of chronic HBV-infected patients without HCC did not differ substantially from those of the healthy controls [18]. Finally, in patients with HCC, no significant association was detected between EGF rs4444903 polymorphism allelic frequencies and the severity of liver damage determined by the Child-Pugh score (score A versus score B + C). Therefore, the data from these two studies argue against a possible role of the EGF rs4444903 A > G polymorphism in an unfavourable disease course of chronic hepatitis B. However, the population used in the aforementioned studies was of Chinese origin whereas the present study includes a Caucasian population. The G allele and the G/G genotype of the EGF rs4444903 A > G polymorphism are largely more frequent in the Chinese population than those reported in the Caucasian population. Additionally, the results of our study are consistent with the report of Tanabe et al. showing an association between the carriage of the G allele and the occurrence of HCC in two Caucasian populations [15]. Thus, the data from Europe and the U.S. differ from what occurs in the Far East, where the prevalence of HBV infection is high and vertical transmission is common. HCC develops almost exclusively in patients with advanced cirrhosis who are listed for liver transplant [4]. More importantly, no study has previously ascertained longitudinally the relationship between the EGF rs4444903 A > G polymorphism and disease progression in chronic HBV infection. In the present report, the EGF rs4444903 A > G polymorphism assessed longitudinally was found to predict a worse disease course of chronic hepatitis B. The carriage of at least one G allele was associated with cirrhosis occurrence after adjusting for the main confounding variables. Interestingly, the effect exerted by the possession of the G allele seems to appear early in the course of the disease and to favour faster disease progression similar to what is observed in patients with chronic hepatitis C [34].

Because EGF rs4444903 A > G is a functional polymorphism, genotypes that directly influence the phenotype of the patient can lead to differences in EGF serum levels as demonstrated by Tanabe et al. and by Qi et al. [15,18]. These authors found a relationship between the carriage of the G allele and the presence of higher EGF serum levels in patients with liver cirrhosis and HCC. Therefore, it is plausible that the higher EGF levels could facilitate the onset of cirrhosis. Because the EGF serum levels could be influenced by confounding variables such as age and ethnicity, medications, diet and co-morbidities, EGF determination in serum might be of limited value for investigational purposes.

The protective effect of carrying the A/A genotype was more evident in females in comparison to males. An interaction has been observed between EGF and sex-specific steroid signalling. In experimental studies, this mechanism appeared to act through Src, a non-receptor protein tyrosine kinase, activation, because EGFR, oestrogen receptor and Src were found to form a complex upon oestrogen stimulation [35]. In clinical studies, the association between EGF polymorphism and disease course was modulated by the patient's gender. For example, the carriage of the G allele was associated with ARDS only in males, and hepatic metastases from colorectal cancer occurred mostly in males carrying the EGF rs4444903 G/G genotype [36,37].

Some limitations may be attributed to the present study. First, although the combined assessment of a cross-sectional and of a longitudinal cohort of patients may reinforce the results, the total number of patients was rather limited. This limited number gives a reduced power to the analysis. Second, the present data refer to the Caucasian population. Because of the strong differences in the allelic and genotypic frequencies among different ethnicities, these data cannot be generally extrapolated. Third, serum levels of EGF were not determined because sera were not available. However, the knowledge of these phenotypic data probably would not add substantial information to the genotypic assessment.

In conclusion, carriers of the G allele with the EGF rs4444903 A > G polymorphism that have chronic HBV infection appear to be at risk of a worse disease course with a higher likelihood of cirrhosis development. This association might be modulated in part by gender.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

This work was supported in part by the Regione Piemonte Ricerca Sanitaria Finalizzata Program.

Disclosure

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

The Authors have nothing to disclose concerning this paper.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References