Correlation between IL-28 polymorphism and spontaneous clearance in HCV patients: systematic review and meta-analysis

Hepatitis C virus (HCV) is a serious global health issue. Nearly 20% of HCV patients spontaneously clear the virus. While some studies have shown an association of spontaneous clearance (SC) of the virus with interleukin (IL) 28B single-nucleotide polymorphisms (SNPs), others did not show such a relationship. Thus, the purpose of the present study was to investigate the association of IL28B polymorphisms (12979860 SNP) with SC of HCV infection. Upon initial screening of the databases, a total of 545 articles were retrieved, of which 22 studies that met predefined eligibility criteria were entered into the meta-analysis. Odds ratios (ORs) with confidence intervals (95% CI), heterogeneity, publication bias, and sensitivity analysis were assessed. According to the meta-analysis results, a significant association was observed between the rs12979860 SNP and SC of HCV infection. The results indicated that the ORs of SC from hepatitis C virus infection were 2.75 times higher in those with cytokine gene polymorphisms (95% CI, 2.23 to 3.38). Moreover, it was found that the prevalence of rs12979860 CC was 0.33 with 95 CI 0.28-0.38 in genotype 1 and was 0.40 with 95 CI 0.34-0.47 in other genotypes. Our meta-analysis results suggest that IL28B rs12979860 CC is a strong predictor for SC of hepatitis C infection in PEG IFN-a/RBV-treated patients.


Introduction
Hepatitis C virus (HCV) is a serious public health problem worldwide, affecting >170 million people globally, and around 3-4 million people are newly infected each year [1][2][3]. Although HCV causes liver-related morbidity and mortality and >70% of patients develop chronic HCV infection, 20% of patients show spontaneous clearance (SC) of the viral infection [4,5]. While the pathogenesis of HCV has not been completely elucidated, studies have shown that several viral and host factors are associated with differences in viral persistence or clearance. Also, HCV patients respond differently to treatment, indicating an important role of the host genetic context [6][7][8]. Several studies have shown that SC of virus and response to treatment might be associated with interleukin (IL) 28B gene polymorphisms. There are three main single-nucleotide polymorphisms (SNPs) in the IL28B gene: rs12979860, rs809917, and rs12980275. Several studies have shown that there is a correlation between these SNPs and SC, treatment, and prediction of response [4,9,10]. Because these SNPs are located in the promoter and regulatory region of the IL28B gene on chromosome 19, they can affect IL28B cytokine gene expression and production. In addition, SC of HCV infections may be affected by these SNPs [7,11].
The antiviral properties of the IL28B protein are mediated through stimulation of the Janus kinase signal transducer and activator of transcription (JAK-STAT) protein, which controls interferon (IFN)-stimulated genes (ISGs) [12]. Many publications have reported the influence of IL28B polymorphisms on the clearance of HCV. In one study conducted on an Asian population, patients with the CC genotype at the rs12079860 SNPshowed spontaneous clearance [13]. Furthermore, patients with the rs12979860 SNPwere associated with IFN-free treatment [14][15][16]. It has also been shown that the CC genotype at rs12979860 of IL28B is associated with sustained virologic response (SVR) in Asian, Caucasian, European, African-American, and Hispanic populations [4,17]. To date, various meta-analyses have been published on the correlation between gene polymorphisms of IL28B and SVR [17][18][19][20][21], but few meta-analyses have been performed on the IL28B polymorphisms that are involved in SC. Furthermore, the studies have reached inconsistent conclusions, and while some studies reported a significant association [19,, some others did not show an association [43,44]. Accordingly, the objective of this study was to clarify the association of IL28B polymorphisms with pegylated interferon (IFN) α and ribavirin (RBV) (PEG-IFN/RBV) treatment response and SC.

Methods
The present systematic review and meta-analysis was performed based on the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement. (http:// www. prisma statement.org/PRISMAStatement/ PIS-MAStatement.aspx).

Search strategy and data extraction
We identified relevant publications through a comprehensive and systematic search of the databases PubMed, MEDLINE/ PubMed, Web of Science Scopus, and Google Scholar, using the keywords "rs 12979860", "IL -28 gene", "Interferon lambda-3", "Polymorphism", "HCV", "Spontaneous clearance", "spontaneous resolution", "spontaneous recovery (SR)", and "chronic hepatitis C". "AND" and "OR" Boolean operators were used for combining search keywords in the above-mentioned databases. The included papers were first screened based on their title, abstract, and keywords to select the eligible ones. Next, the full texts of the selected papers were evaluated by two independent reviewers. In cases of disagreement, the reviewers discussed the articles until a consensus was reached. Through searching various databases, 545 articles were identified. Out of these, 515 were removed because they were irrelevant or duplicate articles. The remaining 30 studies were assessed according to the title and abstract, and two publications did not qualify for assessing the full text. Next, the full text of the 28 eligible articles was evaluated, and six of these were removed because of irrelevant and incomplete results. Finally, 22 studies were included in the meta-analysis (Fig. 1). Full or entire searches were limited to English-language articles published between April 2011 and November 2019. The PRISMA flow diagram (http:// prisma-state ment. org/ PRISM AStat ement/ FlowD iagram. aspx) illustrates the process of selection of included studies (Fig. 1).

Inclusion and exclusion criteria
Studies satisfying the following criteria were considered in our meta-analysis: 1) case-control studies, 2) published literature related to the association of cytokine gene polymorphisms (exposure) and spontaneous recovery (SR) from hepatitis C virus infection (outcome), 3) all cross-sectional, randomized controlled trials, case-series, and cohort studies reporting on SC in HCV patients with at least nine patients in two compared groups, 4) original studies providing the number of subjects in each group and the percentage of SC in each group. Papers with incomplete or non-relevant data were not included in this meta-analysis.

Assessment of studies risk of bias
The risk of bias of studies was evaluated using the Newcastle and Ottawa (NOS) checklist by the two independent reviewers. In the case of disagreement, the two reviewers discussed the study with a third reviewer until reaching consensus. No study was excluded because of the poor quality in conduct as all papers received 6 or more stars from NOS.

Statistical analysis
Meta-analysis was performed on odds ratios for the association of cytokine gene polymorphisms (exposure) and spontaneous recovery (RS) from hepatitis C virus infection (outcome). Because in most of the included studies the reference group for calculation of the OR was the spontaneousrecovery group, the ORs and their 95% CI were reversed in the rest of the studies to make the reference group identical. The numerator of the reported prevalence was rs12979860 CC, and the denominator was rs12979860 CC, CT and TT (favorable homozygous versus heterozygous plus unfavorable homozygous). A random-effects model was then used to combine ORs. The random-effects model was preferred over a fixed-effects model because it could be assumed that at least some of the variation among studies is real. This variation among studies, also called heterogeneity, was assessed using Cochran's Q-test and the I-squared statistic. To identify study characteristics that could account for the possible heterogeneity, subgroup analysis was performed for country and study type, and meta-regression was applied for the mean age and gender proportion. Publication bias was assessed statistically based on Egger's and Begg's tests and visually based on a funnel plot. In order to determine whether one of the studies had a considerable effect on the summary effect, influential analysis was performed. In this procedure, each study is removed in turn, after which the summary effect is calculated, providing the possibility to compare the summary effect with and without that study. All analysis was performed using packages for meta-analysis installed in Stata 14.1. The significance level was set at 5%.

Results
The flowchart of the literature search procedure is illustrated in Fig. 1. According to the search strategy, 545 articles were retrieved from the related databases, and in the study selection step, 22 papers were considered in our meta-analysis (Fig. 1). From the 22 articles, three studies were performed in Iran, three in Italy, two in Australia, two in France, two in the UK, two in Brazil, one in the USA, one in Mexico, one in Germany, one in Spain, one in Egypt, one in Morocco, one in China, and one in Pakistan. All of the included studies were English-language articles. The characteristics of each study included in the meta-analysis are shown in Table 1. Figure 2 shows the results of all of the included studies based on the random-effects model (REM). The summary effect (OR) of 2.75 with 95% CI from 2.23 to 3.38 (P-value < 0.000) shows that the odds of SC from hepatitis C virus infection is 2.75 times higher in those with cytokine gene polymorphisms. In terms of heterogeneity, I2 was 66.5% (P = 0.00), indicating the presence of some heterogeneity among the results of the included studies, and accordingly, the REM was applied to account for this heterogeneity (Fig. 1). To deal with heterogeneity, subgroup analysis was used for country and study type, and meta-regression was applied for the mean age and gender proportion. However, none of these factors could persuasively account for the observed heterogeneity. Regarding publication bias, there is some asymmetry in the funnel plot presented in Fig. 3. In addition, the Egger's test p-value was 0.001, showing the presence of publication bias (Fig. 3). In order to adjust for this, a bias trim-and-fill approach was used, which resulted in an OR of 2.09, which was not significantly different from the original value of 2.74. To assess the effect of individual studies on the summary effect, influential analysis was  Fig. 4. The three horizontal lines indicate the summary effect and its 95% CI. Each vertical line shows the summary effect and its 95% CI after elimination of the study specified by its ID on the x-axis. The result of the influential analysis shows that the elimination of any of the studies did not statistically or clinically change the summary effect. In other words, the summary effect was not sensitive to which individual studies were excluded.

Association of IL-28 polymorphism and virus genotype
In order to investigate the association of rs12979860 CC and genotype, we compared the prevalence of rs12979860 CC in genotype 1 versus the other genotypes, restricting the data to SC cases. This comparison was performed through a meta-analysis that pooled the prevalence of rs12979860 CC over two subgroups: genotype 1 and other. It was found that there was some association between rs12979860 CC and genotype, as the prevalence of rs12979860 CC was 0.33 with 95 CI 0.28-0.38 in genotype 1, and it was 0.40 with 95 CI 0.34-0.47 in other genotypes.

Discussion
Numerous studies have shown that the analysis of genetic polymorphisms can help to understand the mechanisms of disease and differences in disease manifestations among individuals [45]. HCV is an important cause of both acute and chronic hepatitis. It is estimated that SC of the virus occurs in 20% of patients with acute HCV infection, and liver cirrhosis and hepatocellular carcinoma (HCC) in chronic HCV patients could be an inevitable process [46]. An increasing body of experimental and clinical evidence Fig. 2 Forest plot of odds ratio for the association between cytokine gene polymorphisms and spontaneous recovery (RS) from hepatitis C virus infection. Each square (■) and its horizontal line, represent the OR and its 95% CI respectively,for each individual study (the size of the grey square corresponds to the weight of that study in the metaanalysis). The diamond (◊) and its horizontal diagonal show the combined overall OR and its 95% CI, respectively.
since 2009 and genome-wide association studies (GWAS) have indicated that polymorphisms close to the IL-28B gene, which encodes interferon lambda, are associated with increased rates of SC in subjects with chronic HCV infection [47,48]. However, in some cases, there have been contradictory conclusions, and while some studies have reported significant associations, others did not find any associations [49][50][51]. IL-28B, like other type III IFNs, such as IL-28A and IL-29, shows strong antiviral activity and induces expression of ISGs [52]. In the present work, we performed a systematic review and meta-analysis to examine the effect of a genetic variant (rs12979860) located in a gene (IFN-λ3) involved in host innate immunity on clearance of HCV.
Our meta-analysis result showed that rs12979860 SNP was significantly associated with HCV SC and SVR. The pooled result of all of the included studies indicated a significant association between rs12979860 CC and SVR in HCV patients treated with PEG IFN-a/RBV. The result indicates that the OR of SC from HCV infection is 2.75 times higher in those with cytokine gene polymorphisms (95% CI 2.23 to 3.38). Although the molecular mechanism of the effect of IL-28B gene polymorphism on HCV clearance is unclear, several studies have shown that various factors, including viral, host, epidemiological, and environmental factors, can be associated with HCV clearance [53,54]. Preliminary reports evaluating the association between IL28B SNPs and the clearance of HCV have been reported in large groups of patients who are chronically infected with HCV and have been treated with standard antiviral drugs [55,56]. Thomas et al. [54] and Shi et al. [21] showed that the rs12979860 SNP is strongly associated with higher probability of HCV clearance in patients of African or European ancestry. Their result indicated that there was almost a 3-fold higher rate of virus clearance in patients with the rs12979860 genotype CC than in those with TT or CT. Although T rs8099917 and C rs12979860 were most strongly associated with HCV clearance, they might be affected by the HCV genotype, racial diversity, and population differences. It has been demonstrated that the polymorphism in the IL-28B gene appears to be a weaker pretreatment predictor for antiviral responses in individuals with HCV-2 and HCV-3 than in those with HCV-1 [20]. Hepatitis C treatment could be affected by several viral factors, such as the genotype of the virus, quasispecies, the primary load of the virus, and the growth kinetics of virus [6,57].
Regarding the existence of different HCV genotypes, patients with HCV who are treated with PEG IFN-a/RBV differ in their ability to achieve an SVR, and the SVR rate differs for each genotype, with individuals with genotype 2 having the highest rate of SVR and those with genotypes 1 and 4 having the lowest SVR rate [58]. In our work, virus genotype stratification analysis revealed that there was some association between rs12979860 CC and genotype, as the prevalence of rs12979860 CC was 0.33 with 95 CI 0.28-0.38 for genotype 1, and it was 0.40 with 95 CI 0.34-0.47 for the other genotypes. However, the observed association, although consistent with the available literature, was not statistically significant, as the two CIs overlapped each other. This could be explained by the low precision of the included studies or the small number of studies. Our work may have some limitations that should be taken into consideration when interpreting the results. In the present analysis, only published studies were included, and it is best to include all available individual patient data, such as previously unpublished data. Moreover, other factors besides genetic factors including viral load and the patient's physiological and physical characteristics can also affect the SC of infection in individuals with chronic HCV treated with antiviral therapy. Also, due to the small sample size for the genotypes 2/3 and genotype 4, it may be argued that our conclusion cannot be generalized, and future studies with a larger number of patients per group for genotypes 2/3 and genotype 4 and from different ethnic groups will be needed to confirm the connection with the IL28B polymorphism. Nonetheless, our results suggest that IL28B rs12979860 CC is a robust predictor of SC in patients with chronic HCV.