Influence of Hepatitis C Virus and IL28B Genotypes on Liver Stiffness

Objective Liver fibrosis has been associated with hepatitis C virus (HCV) genotype and genetic variation near the interleukin 28B (IL28B) gene, but the relative contribution is unknown. We aimed to investigate the relation between HCV genotypes, IL28B and development of liver stiffness. Patients and Methods This cross-sectional study consists of 369 patients with chronic hepatitis C (CHC). Liver stiffness was evaluated using transient elastograhy (TE). Factors associated with development of liver fibrosis were identified by logistic regression analysis. Results We identified 369 patients with CHC. 235 were male, 297 Caucasians, and 223 had been exposed to HCV through intravenous drug use. The overall median TE value was 7.4 kPa (interquartile range (IQR) 5.7–12.1). HCV replication was enhanced in patients carrying the IL28B CC genotype compared to TT and TC (5.8 vs. 5.4 log10 IU/mL, p = 0.03). Patients infected with HCV genotype 3 had significantly higher TE values (8.2 kPa; IQR, 5.9–14.5) compared to genotype 1 (6.9 kPa; IQR, 5.4–10.9) and 2 (6.7 kPa; IQR, 4.9–8.8) (p = 0.02). Within patients with genotype 3, IL28B CC genotype had the highest TE values (p = 0.04). However, in multivariate logistic regression, using various cut-off values for fibrosis and cirrhosis, only increasing age (odds ratio (OR) 1.09 (95% confidence interval (CI), 1.05–1.14 per year increment)), ALT (OR 1.01 (95% CI, 1.002–1.011), per unit increment) and HCV genotype 3 compared to genotype 1 (OR 2.40 (95% CI, 1.19–4.81), were consistently associated with cirrhosis (TE>17.1 kPa). Conclusions Age, ALT and infection with HCV genotype 3 were associated with cirrhosis assessed by TE. However, IL28B genotype was not an independent predictor of fibrosis in our study.


Introduction
It is estimated that about 170 million people are infected with hepatitis C virus (HCV) world-wide, and that more than 350.000 people die annually from chronic hepatitis C (CHC) related end-stage liver disease [1].
Several genome-wide association studies have revealed that the IL28B singlenucleotide polymorphism (SNP), rs12979860 genotype CC is advantageous in predicting the probability of spontaneous clearance and sustained virologic response following pegylated-interferon and ribavirin treatment in genotype 1 infected patients [11][12][13].
Recent publications of both HCV monoinfected and HIV-HCV coinfected patients, have studied the role of IL28B and the development of liver steatosis and fibrosis. Using transient elastography (TE), Barreiro et al. reported an association of the CC genotype to liver fibrosis in coinfected individuals whereas Marabita et al. did not show any association as assessed by liver biopsy in monoinfected individuals [14,15]. Ydreborg et al. and Rembeck et al. found an association between the CC genotype and more pronounced liver histopathology in HCV genotype 3 infection [16,17]. Noureddin et al. was unable to show a difference in the frequency of fibrosis progression between patients with IL28B CC and non-CC genotypes but the IL28B CC genotype was associated with hepatic necroinflammation, higher alanine aminotransferase (ALT) levels, and worse clinical outcomes [18]. Similarly, two studies of steatosis came to conflicting results [19,20].
We aimed to investigate the relation between HCV genotypes and IL28B and development of liver fibrosis, assessed by TE, in patients infected with HCV.

Study population
This observational, retrospective study consisted of two independent cohorts of patients with CHC, seen consequetively in Department of Infectious Diseases, Odense University Hospital (OUH) or Department of Infectious Diseases, Hvidovre Hospital (HH) with a TE performed between March 2007 and December 2012. Inclusion criteria were HCV antibody positivity and one positive serum HCV RNA followed by a valid TE. All patients had a positive HCV antibody measurement prior to TE. Exclusion criteria were HCV treatment within 6 months prior to TE, any other cause of liver disease, hepatitis B virus and HIV infection. Demographic, clinical and laboratory parameters were collected from medical records by a physician.
According to Danish Legislation, the Research Ethics Committee can grant an exemption from obtaining informed consent for research projects based on biological material under certain circumstances, and for this study such an exemption was granted.
The study was approved by the Danish Data Protection Agency (record no. 2012-231-0031).

Laboratory
Participants were diagnosed with HCV infection on the basis of anti-HCV detected in a serum or plasma sample by 3 rd generation enzyme-linked immunosorbent assay and confirmed by recombinant immunoplot assay or HCV RNA [21,22]. All samples were tested for RNA levels in the real-time polymerase chain reaction (PCR) system COBAS AmpliPrep/COBAS TaqMan HCV (CAP/ CTM HCV) [23]; the detection limit was 15 IU/ml.
HCV genotype was determined with genotype specific primers from the 5' noncoding region of the virus by RT-PCR [24] or by sequence analysis of RT-PCR generated fragments using C/E1 and NS5B-specific primers [25,26].

IL28B genotyping
For patients from HH genotyping of the IL28B rs12979860 SNP (hereafter refered to as the IL28B SNP) were performed by LGC Genomics (Herts, United Lingdom) using a KASP competitive allele-specific PCR [27]. For a minority, genotypings were done by Molecular Genetics, Department of Clinical Biochemistry Department, Copenhagen University Hospital. Patients at OUH were tested by a SNP genotyping assay using real-time PCR(Lifetechnologies).

Statistical analysis
Values are expressed as median, IQR and percent. Differences were compared using the Kruskal-Wallis test or the Wilcoxon Signed Rank Test for continuous variables and for categorical variables using X 2 or Fisher's exact test, as appropriate. Logistic regression analyses using TE cut-off defined above were performed to evaluate predictors of fibrosis after adjusting for age, HCV genotype and IL28B genotypes and results were presented as OR with 95% CI. Genotype equilibrium was tested by the method of Hardy and Weinberg [32]. All statistical analyses were performed using Statistical Analysis Systems (SAS version 9.3; SAS Institute, Cary, NC, USA) or IBM SPSS Statistics (version 20; IBM, Armonk, New York, USA).

Patient characteristics
The study included 369 patients; 230 patients from OUH and 139 from HH. The majority of these were male (64%; 95%CI, 59-69%), Caucasians (80%; 95% CI, 76-85%), and had been exposed to HCV through injection drug use (IDU) (60%; 95% CI, 55-65%). The median time of known HCV seropositivity was 26.3 (IQR, 3.2-96.9) months and the median age was 48 (IQR, 39-54) years at the time of TE. For all HH patients ALT was obtained within a 182 days time range relative to the fibroscan. For forty-three of the OUH patients ALT was not provided within 6 months prior to or after the fibroscan. The first obtained HCV RNA was available for all patients. HCV RNA within a 180 days range from the TE was available for 218 patients (135 HH patients and 83 OUH patients). No statistical significant difference was observed between the two RNA values (p50.09).
There was no difference in sex, route of exposure or known duration of infection according to HCV genotype, but patients infected with genotype 3 were younger (p50.001) than patients infected with genotype 1 or 2. Genotype 1 was more frequent among Caucasians (p50.03). The main characteristics of the study population are depicted in Table 1. The patients were either hepatitis C treatment naïve at the time of -or had relapsed at least 6 months prior to -the TE. Fourteen patients had been treated prior to their TE.

Viral load
Homozygous CC carriers had a significantly higher (first obtained) viral load compared to TC and TT carriers (median 5.8 and 5.4 log 10 IU/mL, p50.03).
The HCV genotype 2 and 3 infected homozygous CC carriers had a higher (first obtained) median HCV RNA compared to CT and TT combined (median 5.6 vs. 5.1 log 10 IU/mL for genotype 2 and 5.6 vs. 5.0 log 10 IU/mL for genotype 3). This significant difference was not observed among the genotype 1 infected patients. HCV genotype 1 infected patients had significantly higher viral load compared to genotype 2 and 3 infected at the time of TE (table 1).

IL28B rs12979860
The majority of our study cohort (350 of 369 (95%)) was genotyped for the IL28B SNP. We lacked DNA to genotype the remaining 19 patients. The overall distribution of the alleles was as follows (no. of patients, %): CC (129, 37%), CT (183, 52%), TT (38, 11%). The allele distribution was in Hardy-Weinberg disequilibrium (p,0.05). IL28B genotype differed according to HCV genotype. Genotype CC was less frequent among genotype 1 infected patients compared to genotype 2 and 3, and consequently CT and TT (non-CC) were more frequent (P,0.005) ( Table 1). Because prior studies have shown that IL28B allelic traits behave as an autosomal recessive inheritance [14] and due to the very few cases of TT, the CT and TT genotypes were combined in regression analyses (see below).
No association was found between ALT and IL28B genotypes.
In genotype 3 patients, the IL28B CC genotype was associated with a two-fold increased risk of advanced liver fibrosis in the unadjusted analysis (OR 2.09, 95% CI 1.05-4.85)) compared to the CT/TT genotype. In multivariate analysis adjusted for age and ALT this association remained but lost statistical significance (OR 1.58, 95% CI 0.66-3.75).

Factors associated with cirrhosis (TE.17.1 kPa)
By univariate logistic regression analysis, increasing age, ALT and HCV genotype 3 were each associated with cirrhosis (TE.17.1 kPa) whereas sex, race, route of exposure, duration of known seropositivity, IL28B genotypes or center were not (table 3). In multivariate analysis, increasing age, ALT and HCV genotype 3 compared to genotype 1 remained statistically significantly associated with cirrhosis (table 3).
After stratification for HCV genotype, there was no association between IL28B CC genotype and cirrhosis for any HCV genotype.
All analyses involving IL28B were repeated separately for Caucasians in order to account for race specific effects but did not alter the results (data not shown).

Discussion
In this cross-sectional study of patients infected with HCV, increasing age and HCV genotype 3 were associated with increased liver stiffness, while the IL28B CC genotype was not. Steatosis associated with HCV genotype 3 infection is a known risk factor for fibrosis [33], and the presence of steatosis in HCV genotype 3 infection is associated with higher HCV RNA viral load compared to other genotypes [34][35][36]. Several studies suggest that HCV replication could be enhanced in patients carrying the IL28B C allele [11,37,38]. It has been hypothesised that patients with HCV genotype 1 infection have ''metabolic steatosis'', and HCV genotype 3 infected have ''viral steatosis'' [34,39]. The mechanisms underlying viral steatosis caused by HCV genotype 3 are not fully elucidated.
Overall steatosis is an important cofactor in accelerating the development of hepatic fibrosis and necroinflammatory activity in patients with CHC [35]. This corresponds well with our findings of an increased risk of fibrosis in patients infected with HCV genotype 3.
For genotype 3, the IL28B CC genotype was associated with a higher TE value in univariate analysis. This association lost significance in multivariate analysis when adjusted for ALT and age. For HCV infected patients with elevated ALT it should be noted that inflammation of the liver is considered a confounding variable in analysis of liver stiffness [40]. ALT was not associated with IL28B genetic variants.
Barreira et al. showed that the IL28B CC genotype was associated with cirrhosis using a TE cut-off of 14.5 kPa [14]. By liver biopsy, Rembeck et al. showed that the CC genotype in HCV genotype 3 mono-infected patients entailed more pronounced portal inflammation and steatosis [16]. Similarly Ydreborg et al. reported that the CC SNP was associated with a higher TE value in HCV genotype 3 infected patients [17]. These mentioned studies differ from our study in several aspects. In the study of Barreiro et al. 86% were infected through IDU compared to 60% in our population, whereas HCV was transmitted through IDU in fewer patients in the study of Ydreborg et al. [14,17]. Due to sample size, we might lack statistical power in our study.
Other studies have assessed liver fibrosis by liver biopsy, whereas we used TE. TE is known to be a valid tool for estimating fibrosis but liver biopsy is still considered the gold standard [30]. A recent study by Degos et al. suggested that TE and liver biopsy performed comparable with regard to a diagnosis of cirrhosis whereas liver biopsy was superior in diagnosing fibrosis [31].
Various transient elastography cut-off values to define fibrosis and cirrhosis have been suggested [29][30][31]. We evaluated two of these. Using either of these, HCV genotype, ALT and age consistently were associated with fibrosis and cirrhosis, whereas the IL28B CC genotype only was not.
Further studies of a possible role of IL28B and other genetic variants in the development of fibrosis and cirrhosis are warranted.
Stengths of our study include the evaluation in two centers, and that we have a comparable number of HCV genotype 1 and 3 infected patients. Some limitations in our study must be acknowledged. First, TE is an indirect marker of liver fibrosis and several factors -most importantly elevated ALT -may overestimate fibrosis when determined by TE. Second, due to the retrospective design it was not possible to obtain information about certain potential confounding factors e.g. alcohol consumption among our patients. Third, it was not possible to estimate the exact duration of infection with HCV, which may explain the lack of an association between length of infection and cirrhosis. Most of our patients were diagnosed with HCV in their forties, but the vast majority has been infected years before detection. Genotype 3 infected patients had a median duration of known infection that was about twice as long as genotype 1 infected patients. However, this did not affect the results in the adjusted analysis. Fourth, the small numbers of carriers of the IL28B TT genotype infected with HCV genotype 3 restricted the comparison of the three IL28B genotypes separately.
In conclusion this two-center study evaluated the association of IL28B genotypes and the severity of liver fibrosis in HCV infected patients. We described a possible association of increased liver stiffness with HCV genotype 3 infection. Genetic variation in IL28B may influence fibrosis development in HCV genotype 3 infected individuals but further studies are required to confirm this.