Introduction

Hepatitis C virus (HCV) infection is the leading cause of chronic liver disease worldwide with an estimated number of 170 million chronically infected individuals [1]. The clinical course is highly variable, but up to 30% of the patients eventually present with liver cirrhosis carrying an annual risk of 1–6% for the development of hepatocellular carcinoma [2]. Numerous viral and host-related factors have been shown to accelerate progression of liver disease such as older age at infection, male sex, alcohol consumption, overweight, hepatic iron status, and co-infection with the hepatitis B or human immunodeficiency virus [3, 4]. Antiviral treatment with polyethylene-glycol-conjugated IFN-α (Peg-IFN-α) in combination with ribavirin aims to prevent progression of the disease by eradication of the virus; it is currently considered the gold standard of care [5].

There is increasing evidence that host immunologic and genetic factors play an important role for disease susceptibility, hepatic inflammation, progression to fibrosis and cirrhosis, risk of hepatocellular carcinoma, and response to therapy [6–10]. Recently, numerous investigators focused on polymorphisms in cytokine, chemokine, and receptor genes which were previously shown to be implicated in HCV pathogenesis. IL-18 is a pleiotropic cytokine with a wide range of proinflammatory biological effects [11–13]. It is a unique member of the IL-1 cytokine family and synthesized as a 24-kDa proform which is activated to the bioactive 18-kDa mature IL-18, initially described as an interferon (IFN)-γ inducing factor [14]. IFN-γ production is synergistically enhanced in conjunction with IL-12. The main sources of IL-18 are monocytes, macrophages, Kupffer cells, and intestinal epithelial cells. Chief immunomodulatory functions of IL-18 comprise stimulation of the differentiation of naïve T-cells to T-helper 1 cells, enhancement of Fas ligand, and perforin-mediated T-cell and NK-cell cytotoxicity as well as modulation of immunoglobulin secretion by B cells. Proinflammatory activity is mediated by the production of nitric oxide (NO), prostaglandins, and inflammatory cytokines (e.g., TNFα, IL-1β, IL-13), as well as the recruitment of monocytes and macrophages by upregulation of chemokines (IL-8, MIP-1α, MIP-1β, MCP-1) [11–13, 15]. In animals, inhibition of IL-18 by antibodies or an IL-18bp-Fc fusion product abrogated liver damage [14, 16, 17]. In patients with chronic hepatitis C, IL-18 serum levels are increased as compared to healthy controls and positively correlated with the serum activity of alanin-aminotransferase (ALAT) [18, 19].

Giadraitis and colleagues described three genetic alterations in the promoter and two variants in the 5′-untranslated region of IL18 [20]. Evidence for a functional relevance was given in their study for the −137 and the −607 IL18 variants. The −137G>C transversion changes the H4TF1 nuclear factor binding site to a binding site for an unknown factor found in the GM-CSF promoter. The −607C>A promoter variant leads to a disruption of a cAMP-responsive element protein binding site. In the meantime, associations of IL18 variants with different chronic inflammatory conditions were reported comprising hepatitis B [21], atopic asthma bronchiale [22], cardiovascular diseases [23], sarcoidosis in a Japanese population [24], Alzheimer’s disease [25, 26], Graves’ disease [27], and juvenile idiopathic arthritis [28].

We investigated the frequencies of the two abovementioned IL18 promoter alterations in HCV patients and control subjects to determine whether these variants might represent susceptibility factors for hepatitis C in Europeans. Further, we investigated whether −137G>C and −607C>A are correlated with HCV patients’ characteristics and, in particular, with treatment response.

Methods

Patients

In the present study, we genotyped 1,548 European individuals, comprising 757 patients with chronic hepatitis C infection and 791 healthy controls. Patients were recruited from the Department of Hepatology and Gastroenterology, Campus Virchow-Klinikum, Charité in Berlin. The study was approved by the local ethics committee, and all patients gave their written consent. Controls consisted of 791 healthy individuals from the same geographical origin.

Patients’ diagnosis was based on elevation of liver enzymes for at least 6 months with corresponding detection of serum HCV-RNA. All patients proved to be negative for hepatitis B surface antigen and antibodies to human immunodeficiency virus.

Prior to treatment, a liver biopsy was performed in 514/757 (67.9%) of the patients. Hepatic fibrosis (stage) and inflammation (grade) was classified according to the scoring system established by Scheuer [29].

Patients were treated with a combination of standard interferon-α-2a or standard interferon-α-2b plus ribavirin (800–1,200 mg/day) or pegylated interferon-α and ribavirin for 24 weeks in HCV genotype 2/3 infection and 48 weeks in HCV genotype 1 infection, respectively. Data regarding the treatment schedule were unavailable in 282 patients. For the remaining 475 patients, outcome was classified as sustained virological response (HCV-RNA negative at 6 months after end of treatment; n = 176, 37.1%; SR), relapse (HCV-RNA negative at the end of treatment but recurrence of HCV-RNA thereafter; n = 103, 21.7%), viral breakthrough (HCV-RNA recurrence during treatment after initial clearance, n = 23, 4.8%), or nonresponse (failure to clear virus during treatment at week 24 n = 173, 36.4%). For further analysis, patients with relapse, viral breakthrough, and nonresponse were comprised in one group (non-SR).

HCV-RNA Measurements and Genotyping

Virological response was determined by qualitative HCV-RNA assays with a lower detection limit of 30–50 IU/ml (HCV Amplicore 2.0, Roche Diagnostics, Mannheim, Germany or Superquante NGI, Los Angeles, CA, USA). Quantification of HCV-RNA was performed by different standardized, quantitative HCV-RNA assays (Amplicore HCV Monitor 2.0, Roche Diagnostics, Mannheim, Germany, Versant Quantitative HCV, Bayer, Emeryville, CA, USA, or Superquante NGI, Los Angeles, CA, USA). All results for HCV-RNA levels were reported or transformed in IU/ml. HCV genotypes were assessed by a reverse hybridization assay (Ino LiPA HCV II, Innogenetics, Gent, Belgium).

IL18 Gene Promoter Variants

Genetic alterations at position −607C>A (rs1946518) and −137G>C (rs187238) in the IL18 promoter were determined by allele-specific PCR (Table 1) [20]. For each variant, four amplifications were carried out in two reactions. For characterization of the −607C>A variant, PCR reaction was performed by applying a sequence-specific sense primer which yielded a 196-bp product with the common antisense primer. A second sense primer was included to amplify a 301-bp fragment representing the positive control. Accordingly, the −137G>C alteration was determined by amplification of a 261-bp fragment with an allele-specific sense primer and a common reverse primer. A 446-bp fragment—using a common sense and common antisense primer—served as the positive control.

Table 1 Sequences of Oligonucleotides Used for Characterization of IL18 SNPs by Allele-Specific PCR
Full size table

Statistical Analysis

Quantitative variables are presented as mean ± standard deviations or (for skewed data) as median together with minimum and maximum. Hardy–Weinberg equilibrium was assessed by using a Chi2 test with 1 df. Categorical variables including treatment responses were compared by χ 2 statistics. When appropriate, odds ratios (ORs) together with 95% confidence intervals were estimated. Two-sample t tests or 1-way ANOVAs were applied in order to compare continuous variables if the data were approximately normally distributed. For skewed data, the nonparametric Kruskal–Wallis test was used. Furthermore, odds ratios together with 95 % confidence intervals have been calculated in order to determine the association of IL-18 genotypes with treatment response.

SAS software, release 9.01 (SAS Institute Inc., Cary, NC, USA) was used for all statistical calculations. Test results were considered statistically significant when p < 0.05.

Results

Patients’ characteristics are listed in Table 2. Altogether, 757 HCV patients comprising 382 (50.6%) males and 373 (49.4%) females were enrolled in the study. Additionally, 791 healthy subjects from the same geographical area served as controls. No statistically significant deviation from the Hardy–Weinberg equilibrium was observed.

Table 2 Characteristics of 757 Patients with Chronic Hepatitis C Infection
Full size table

In 84.4% of the patients, viral genotype was known. More than 70% of patients with known genotype presented with a viral genotype 1. In 67.9% (514/757) of the patients, a diagnostic liver biopsy was performed, whereas data regarding inflammatory activity were available in 64.1% (485/757) of the patients. In 59.4% (450/757) of the HCV-infected patients, complete information regarding genotype and treatment response was available for further analysis.

The association of genotypes and alleles with treatment response was tested in 475 HCV-infected patients. The −607C>A variant was found in similar frequency in patients and controls (p = 0.4742, Table 3). When evaluating the −137G>C alteration, the CC genotype was relatively rare. Again, patients and controls did not differ significantly in their genotype frequencies (p = 0.2061). Moreover, allele frequencies for both variants were similar in patients and controls (p = 0.2871 or p = 0.1024, respectively; Table 3).

Table 3 −607C>A and −137G>C IL18 Allele and Genotype Frequencies in Patients with Chronic Hepatitis C Infection and Controls
Full size table

Lack of significant differences in genotype frequencies was stated for the −607C>A variant as far as sex (p = 0.7254), route of transmission (p = 0.8800), viral genotype (p = 0.8530), and other parameters were concerned (Table 4). Further analyses for the −137G > C SNP substantiated that genotype frequencies did not differ according to sex (p = 0.6785), transmission route (p = 0.1960), viral genotype (p = 0.8918), and other parameters (Table 5).

Table 4 Clinical and Demographic Characteristics of Patients with Chronic Hepatitis C Infection According to the −607 IL18 Genotype, Respectively
Full size table
Table 5 Clinical and Demographic Characteristics of Patients with Chronic Hepatitis C Infection According to the −137 IL18 Genotype, Respectively
Full size table

Analysis of −137 and −607 IL18 allele frequencies in relation to treatment response revealed significant differences (Tables 6, 7) for both variants. Patients with −607A had a higher chance for sustained virological response than those with the −607C allele (p = 0.0416 OR 1.331; [CI 95%, 1.011–1.1752]). In parallel, −137C was associated with a higher probability for viral clearance (p = 0.0274; CI 95%, 1.037–1.872). Since viral genotype has been identified as one of the strongest predictors for virological response, we performed subgroup analyses. Surprisingly, positive correlations were observed in genotype 1 HCV patients for −607A (p = 0.019; OR = 1.537 [CI 95%, 1.072–2.205]) and −607C (p = 0.0302; OR = 1.524; [CI 95%, 1.037–1.872]) with treatment response, whereas in patients with non-1 genotypes no association was demonstrated (p = 0.8562 and 0.3770, respectively).

Table 6 Treatment Response According to Allele Frequencies in Patients with Chronic Hepatitis C Infection (n = 450)
Full size table
Table 7 Treatment Response According to Genotype Frequencies in Patients with Chronic Hepatitis C infection (n = 450)
Full size table

Similar results were observed when analyzing −607C>A and −137G>C genotype frequencies after stratification for treatment response. There was a trend towards a higher frequency of the AA genotype among patients with a treatment success (p = 0.0869; Table 6) for the −607C>A variant. Concurrently, CC genotypes of the −137G>C variant were more prevalent among patients with a treatment success by tendency (p = 0.0620, Table 7).

Again, in the subgroup of patients with viral genotype 1, –607 and −137 IL18 genotypes turned out to be stronger predicting factors than in the non-1 viral genotype subgroup (Tables 6, 7). Additionally, odds ratios were calculated to further explore the role of genotypes for treatment response. When using a binary outcome (nonresponse or relapsed response versus sustained response), the odds ratio of genotypes CA versus CC was 1.558 [95% CI, 1.011–2.401] for −607 indicating that CA carriers have a higher probability of a sustained response. For −137 genotypes (GC versus GG), the corresponding odds ratio was 1.563 [95 % CI, 1.046–2.337].

Discussion

This is the first study characterizing the impact of functional IL18 promoter variants in a large cohort of patients with hepatitis C on treatment response and various additional clinical and biochemical features. Both the −137 and the −607 IL-18 promoter variants were significantly associated with treatment response in patients infected with HCV genotype 1.

There is a growing body of evidence that genetic host factors have a significant impact on outcome and resolution of HCV infection—both treatment-related and spontaneously occurring. A strong Th 1 antiviral immune response is considered a prerequisite for HCV elimination [30]. Furthermore, functional variants of involved cytokines were recently shown to be associated with HCV disease pathogenesis. Huang and coworkers demonstrated an association of the −764 IFN-γ alteration with spontaneous recovery of HCV infection and treatment response [16]. In a recent multicenter study, the functional 174C>G IL6 variant was significantly associated with treatment response in acutely and chronically HCV-infected patients with HIV co-infection [14]. It was suggested that higher IL6 expression may favor Jak-STAT3-signaling pathways in the liver of affected subjects, thus, stimulating a strong antiviral response. Genetic alteration in IL10-encoding the main anti-inflammatory cytokine—were found to be associated with HCV clearance in African Americans in contrast to Americans with European ancestry from the same geographical area [31]. Others did not find a significant impact of IL10 variants on severity and clearance of HCV infection [32, 33]. A European study reported a positive association of IL10 receptor (IL10R) and IL22 variants with outcome in HCV-infected patients [34].

The biological role of the proinflammatory IL-18 in hepatic pathology is complex and still not fully characterized. Activated macrophages, Kupffer cells, natural killer (NK) cells, and NK T cells are believed to be the main sources of IL-18 in the liver [35]. IL-18 potentiates liver injury due to the induction of IFN-γ in conjunction with IL-12. Okamura and coworkers demonstrated that anti-IL-18 antibodies prevented liver damage in mice primed with Propionibacterium acnes and subsequently challenged with LPS [14]. In a similar mouse model, anti-IL-18 antibodies diminished T-cell-mediated liver injury in leptin deficient (ob/ob) mice [16]. In IL18 transgenic mice, upregulated IL-18 expression was associated with severe hepatic injury and spontaneous apoptosis of hepatocytes [35]. In contrast, IL-18 displayed features of a protective factor in viral infection, for instance, by inhibition of hepatitis B viral replication. In the livers of HBV transgenic mice, IL-18 stimulated natural killer cells and NK T cells to secrete IFN-γ. This coincided with a rapid and reversible inhibition of viral replication [36]. Others have demonstrated antiviral effects in models of herpes simplex virus and vaccinia virus infection [37, 38]. On the contrary, HIV replication was stimulated by IL-18 in a monocytic cell line [39].

Regarding the role of IL18 variants in viral hepatitis, there is only limited data available. Zhang and coworkers performed analyses on −137 and −607 IL18 variants in 231 Chinese patients with HBV infection and in 300 healthy controls. The frequency of the −137GG genotype was found to be significantly higher in the patient group leading to the assumption of a protective role of this allele in chronic HBV infection. In addition, the −607AA genotype was linked to a lower viral load [40]. Studying a population of 140 Thai HCV patients and 140 matched controls, the −607AA genotype was more frequently observed in HCV patients when compared with matched controls [21]. In a recent publication by An and coworkers, both the −137C and the −607A alleles were associated with spontaneous HCV clearance in African American intravenous drug users, when comparing 91 drug abusers who cleared the virus with 182 patients with persistent viral infection [41]. Bouzgarrou studied the outcome of HCV patients and IL18 serum levels in relation to the −137 and −607 variants [42]. When comparing HCV patients with different stages of disease (no cirrhosis, cirrhosis, and hepatocellular carcinoma (HCC), IL-18 serum levels increased with disease progression. Furthermore, the −607C allele was associated with cirrhosis and HCC.

In our study, no association of IL18 promoter variants with severity of hepatic inflammation and fibrosis was apparent. In parallel, no association of IL18 alterations with transaminase serum levels could be shown. Of note, alleles with lower transcriptional IL18 promoter activity (−607A and −137C) were associated with treatment response indicating a more complex role of IL-18 than previously assumed. IL-18 is not a mere proinflammatory Th1 cytokine but additionally modulates Th2 functions. Following this concept, downregulated IL18 expression enables an optimal Th1 and Th2 balance for viral eradication. The relevance of our finding is corroborated by the comparison of spontaneous hepatitis C viral eradication in IV drug abusers in association with IL18 alterations [41]. In analogy to our study, alleles with low IL18 promoter activity were associated with a higher eradication rate. In hepatitis B patients, the “low IL18” −137C allele was attributed a protective role when comparing 231 HBV patients and 300 normal controls from China [40].

Here, associations of IL18 promotor alleles with treatment response were only observed in patients infected with HCV genotype 1. This patient subgroup represents a particular challenge in clinical practice since the rate of rapid and sustained virological response is lower than in genotype 2/3 HCV patients. Identifying predictive genetic markers bears the potential of a more individualized treatment regimen.