Introduction

Hepatitis C virus (HCV) is one of the leading causes of liver cirrhosis and hepatocellular carcinoma (HCC), with nearly 170 million people infected worldwide [1]. A combination therapy with pegylated interferon (IFN)-α (PEG-IFN-α) and ribavirin (RBV) has been used for chronic hepatitis C (C-CH) patients as the standard of care, achieving sustained virological response (SVR) in 42–52 % of genotype 1 patients [2]. Even in the coming era of all oral and IFN-free regimens for the treatment of C-CH patients [3–5], PEG-IFN-α plus RBV therapy could hold promise for elderly patients with advanced fibrosis and high risk of HCC.

Genome-wide association studies, including ours, have demonstrated that single nucleotide polymorphisms (SNPs) upstream of the promoter region within the IFNL3 gene (also known as IL28B), which encodes a type III IFN (IFN-λ3), are strongly associated with the response to PEG-IFN-α plus RBV therapy in C-CH patients [6–9]. Although such significant impact of the IFNL3 genotype on the outcome of the combination therapy is well acknowledged, the biological and clinical roles of IFN-λ3 in chronic HCV infection are still elusive. Furthermore, it is controversial if patients with the IFNL3 major genotype are capable of producing larger amounts of IFN-λ3 than those with the minor genotype.

The IFN-λ family consists of several subtypes, such as IFN-λ1 (IL-29), IFN-λ2 (IL-28A), and IFN-λ3 (IL-28B), which are biologically active for the suppression of HCV replication [10, 11]. On initial exposure to HCV, primary human hepatocytes in vitro produced IFN-λ and subsequently induced antiviral IFN-stimulated genes [12]. It is thus rational to consider that the more IFN-λ family members are produced in the exposed hosts, the more likely they are to protect the hosts from HCV virulence in the primary infection. However, in chronically HCV-infected patients, it has not been proven that such a scenario could be applicable for the outcome of the disease.

To gain insight into the role of IFN-λ3 in chronic HCV infection, we aimed to clarify the factors influencing serum IFN-λ3 levels, including IFNL3 genotype, clinical parameters, and various cytokines and chemokines. For application in clinical practice, we evaluated whether serum IFN-λ3 levels are associated or not associated with the response to PEG-IFN-α plus RBV therapy for C-CH patients.

Materials and methods

Study subjects

One hundred nineteen Japanese patients with C-CH (genotype 1b and high viral load) were enrolled in the study. All patients were negative for hepatitis B virus (HBV) and human immunodeficiency virus (HIV) and did not have any other chronic liver diseases, such as alcoholic, autoimmune, and fatty liver disease. The presence of HCC was ruled out by ultrasonography or computed tomography examinations. The patients had been followed at the National Center for Global Health and Medicine Kohnodai Hospital, the National Hospital Organization Nagasaki Medical Center, Shin-Kokura Hospital, and Musashino Red Cross Hospital. They were treated with PEG-IFN-α2b (subcutaneously once a week; 1.5 μg/kg body weight) or PEG-IFN-α2a (180 μg once a week) plus RBV (600–1,000 mg daily depending on body weight) for 48 weeks according to the guidelines of the Japan Society of Hepatology [13]. Virological response to the combination therapy was defined according to the practical guidelines of the American Association for the Study of Liver Diseases [14]. All patients attained adherence to PEG-IFN-α plus RBV therapy exceeding 80 % of the estimated total dose. Liver biopsy was performed before the start of the therapy. Histological activity and fibrosis were determined according to the METAVIR scoring system [15]. Serum samples were collected from the patients before PEG-IFN-α plus RBV treatment started and were stored at −80 °C. In some patients, the samples were obtained 24 weeks after the cessation of the therapy (at the end of follow-up).

As controls, serum was obtained from 23 healthy subjects without HCV, HBV, and HIV infection (male-to-female ratio, 5:5, mean age ± standard deviation, 45 ± 12 years). In the comparison of serum IFN-λ levels between C-CH patients and patients with other types of liver diseases, 11 patients with chronic HBV infection (three HBeAg-positive patients and eight HBeAg-negative patients) were examined as well. They were not treated with IFN or nucleot(s)ide analogues for HBV infection. In addition, we compared serum IFN-λ3 levels among patients with acute viral hepatitis of various causes, such as acute hepatitis A, acute hepatitis B, or acute hepatitis E, the diagnosis of which was determined by serological examinations at Teine Keijinkai Hospital and Kurume University Hospital. The serum samples were obtained from the patients at the time of active liver inflammation [alanine aminotransferase (ALT) levels more than two times the upper limit of the normal range]. As representatives for noninvasive fibrosis markers, the fibrosis-4 (FIB-4) score and the aspartate aminotransferase (AST) platelet ratio index (APRI) were calculated as reported previously [16, 17].

The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Institutional Review Board at the National Center for Global Health and Medicine (approval ID and date, NCGM-G-001379-00, March 14, 2013) and the ethical committee of each institute. Written informed consent was obtained from all patients.

IFNL3 genotyping

The subjects were evaluated for SNPs near the IFNL3 gene (rs8099917) using the Invader Plus assay (Invader Chemistry, Madison, WI, USA) as previously reported [18]. The TT, TG, and GG genotypes were determined accordingly.

Measurement of serum IFN-λ3

Serum levels of IFN-λ3 were evaluated by the newly developed chemiluminescence enzyme immunoassay system as reported previously [19]. The system enables one to quantify serum IFN-λ3 specifically without any overlap from IFN-λ1 and IFN-λ2. The threshold of the assay is 10 pg/mL and its range is 10–1,000 pg/mL.

Simultaneous measurement of multiple chemokines and cytokines

To quantify multiple chemokines and cytokines simultaneously in the limited volume of the samples, we used the BioPlex 3D system (BioPlex Pro Human GI 27Plex; Bio-Rad, Hercules, CA, USA) for the study. In this system, 27 chemokines and cytokines were measurable, such as basic fibroblast growth factor, eotaxin, granulocyte colony stimulating factor, granulocyte–macrophage colony stimulating factor, IL-1β, IL-1 receptor antagonist, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8, IL-9, IL-10, IL-12, IL-13, IL-15, IL-17, IFN-γ, IFN-γ-inducible protein 10 (IP-10), monocyte chemotactic protein 1, macrophage inflammatory protein (MIP)-1α, MIP-1β, platelet-derived growth factor BB (PDGF-BB), regulated on activation, normally T cell expressed, and secreted (RANTES), TNF-α, and vascular endothelial growth factor. The detection range and thresholds are given in Table S1. For the measurement of IP-10, ELISA (R&D Systems, MN, USA) was performed as well.

Statistical analyses

Continuous variables were compared between groups using the Wilcoxon signed-rank test and the Mann–Whitney U test, and categorical data were compared using the χ 2 test or Fisher’s exact test. The correlations between cytokines, chemokines, and clinical markers were evaluated by Spearman’s correlation coefficient. A p value below 0.05 was considered to be significant. Logistic regression was used for multivariate analyses. All statistical analyses were performed with PRISM and SPSS.

Results

Serum IFN-λ3 levels are increased in patients with chronic HCV infection

The clinical backgrounds of C-CH patients are shown in Table 1. First, we compared serum IFN-λ3 levels among patients with C-CH or chronic hepatitis B (B-CH) and uninfected healthy volunteers (HVs). Such levels in the C-CH group were significantly higher than those in the B-CH group or the HV group (Fig. 1a). The levels in the B-CH group were increased, but the significance of this was much less than in the C-CH group (Fig. 1a). When we compared serum IFN-λ3 levels in B-CH patients between HBeAg-positive and HBeAg-negative patients, we found no difference between them (2.5 ± 0.9  pg/mL vs 1.8 ± 1.7 pg/mL, respectively). Next, we compared serum IFN-λ3 levels between patients with the IFNL3 TT genotype and those with the TG/GG (non-TT) genotype in the C-CH group. Although some patients in the TT group showed relatively higher levels of IFN-λ3 than those in the non-TT group, this difference between the TT and non-TT groups did not reach significance (Fig. 1b). Third, we compared serum IFN-λ3 levels before and after the combination therapy in the relevant cases. In patients who successfully eradicated HCV (SVR), serum IFN-λ3 levels were significantly decreased at 24 weeks after the therapy. In contrast, such levels did not change in those patients who failed to eradicate HCV (transient virological response and no virological response groups, respectively) (Fig. 1c). Fourth, we compared serum IFN-λ3 levels among patients with various causes of acute viral hepatitis. Unfortunately, serum samples from acute hepatitis C patients were not available in this study. The IFN-λ3 levels in the acute hepatitis E group were higher than those in the HVs (Fig. 1d). The IFN-λ3 levels in the acute hepatitis B group tended to be higher than those in the HVs; however, statistical analysis was not performed because of the limited number of samples (N = 2). No significant difference was observed between the acute hepatitis A and HV groups. These results indicate that serum IFN-λ3 levels are increased in patients with C-CH or acute hepatitis E.

Table 1 Clinical backgrounds of the patients with chronic hepatitis C virus (HCV) infection
Full size table
Fig. 1
figure 1

Serum interferon-λ3 (IFN-λ3) levels are increased in patients with chronic hepatitis C virus infection or acute hepatitis E virus infection. a Serum IFN-λ3 levels in patients with chronic hepatitis C (C–CH; N = 119), patients with chronic hepatitis B (B-CH; N = 11), and healthy volunteers (HV; N = 23) were quantified by the chemiluminescence enzyme immunoassay (CLEIA) method as described in “Materials and methods.” One asterisk p < 0.05, two asterisks p < 0.01, three asterisks p < 0.0001 by the Mann–Whitney U test. b Serum IFN-λ3 levels in the C-CH group were compared between the patients with the IFNL3 TT (rs8099917) genotype (N = 100) and those with non-TT (TG/GG) genotype (N = 19). c Serum IFN-λ3 levels in C-CH patients were compared before and 24 weeks after the pegylated interferon-α plus ribavirin therapy. SVR sustained virological response (N = 21), TVR transient virological response (N = 10), NVR nonvirological response (N = 5), one asterisk p < 0.05 by Wilcoxon’s signed-rank test. d Serum IFN-λ3 levels in acute hepatitis patients of various causes were quantified by CLEIA as described in “Materials and methods.” All samples were collected from patients whose alanine aminotransferase levels were two times higher than the upper limit of the normal range. HV healthy volunteers (N = 23), A-AH acute hepatitis A patients (N = 34), B-AH acute hepatitis B patients (N = 2), E-AH acute hepatitis E patients (N = 9), two asterisks p < 0.0001 by the Mann–Whitney U test

Full size image

Serum IFN-λ3 levels may be related to liver inflammation or fibrosis in patients with C-CH

To explore the clinical significance of IFN-λ3 in chronic HCV infection, we simultaneously examined 27 chemokines and cytokines in serum by means of the BioPlex system, which allows one to measure multiple factors at high sensitivity in a small volume of samples (10 μL per sample). In comparison with the results for HVs, we found that the levels of some chemokines in the C-CH group were higher than those in the HV group, such as IP-10, MIP-1α, MIP-1β, RANTES, and PDGF-BB (Figs. 2, S1).

Fig. 2
figure 2

The levels of several chemokines are increased in patients with chronic hepatitis C virus infection. Twenty-seven chemokines and cytokines in serum from chronic hepatitis C patients (C-CH) and healthy volunteers (HV) were assayed by means of the BioPlex method. Interferon-γ-inducible protein 10 (IP-10) was measured by ELISA. Representative results for chemokines that showed statistical significance between the groups are shown, such as IP-10, macrophage inflammatory protein 1α (MIP-1α), macrophage inflammatory protein 1β (MIP-1β), regulated on activation, normally T cell expressed, and secreted (RANTES), and platelet-derived growth factor BB (PDGF-BB). * p < 0.005, ** p < 0.001, *** p < 0.0001 by the Mann–Whitney U test

Full size image

Next, we examined whether serum IFN-λ3 levels are correlated or not correlated with clinical parameters or immunological markers in the C-CH group. The IFN-λ3 levels were weakly and positively correlated with AST, ALT, and α-fetoprotein levels and histological activity (Table 2). These results indicate that the increase of serum IFN-λ3 levels in patients with C-CH is related to liver inflammation. The FIB-4 score and the APRI are representatives of noninvasive markers of liver fibrosis. The levels of serum IFN-λ3 were positively correlated with the APRI, but not with the FIB-4 score (Table 2). With regard to the chemokines displaying higher values in the C–CH group, the levels of IP-10 and PDGF-BB were positively correlated with the IFN-λ3 levels (Table 2). Such chemokines are reported to be involved in the early stage of liver fibrosis [20–22]. Thus, serum levels of IFN-λ3 may be related to the fibrotic markers as well. To clarify the mechanisms causing the increase of serum IFN-λ3 levels in B-CH patients, we examined the correlations between serum IFN-λ3 levels and clinical markers and fibrosis indices. Serum IFN-λ3 levels were correlated with the levels of AST (r = 0.64, p = 0.03) and total cholesterol (r = −0.76, p = 0.03), FIB-4 score (r = 0.65, p = 0.03), and APRI (r = 0.76, p = 0.007) (Table S2). In addition, serum IFN-λ3 levels tended to be higher in HBV-positive patients with liver cirrhosis or HCC (3.0 ± 3.1 pg/mL in liver cirrhosis patients and 4.1 ± 4.7 pg/mL in HCC patients, respectively) (Fig. S2). These results show that serum IFN-λ3 levels are related to liver inflammation and fibrosis not only in C-CH patients but also in B-CH patients.

Table 2 Correlation of interferon-λ3 (IFN-λ 3 ) with clinical or immunological parameters in patients with chronic hepatitis C
Full size table

Pretreatment serum IFN-λ3 is not related to SVR to PEG-IFN-α plus RBV therapy in patients with C-CH

Because the IFNL3 genotype is a strong predictor of the efficacy of PEG-IFN-α plus RBV therapy for C-CH, we sought to examine the clinical value of serum IFN-λ3 in patients who underwent the combination therapy. In a comparison of the clinical and immunological factors between the SVR and non-SVR groups, univariate analysis revealed that AST, IFNL3 genotype, fibrosis score, and MIP-1α were associated with the SVR (Table 3). However, serum IFN-λ3 or IP-10 levels were not different between the SVR and non-SVR groups (Table 3). Subsequently, multivariate analysis including such factors of significance (p < 0.05 by univariate analysis) showed that IFNL3 SNPs, fibrosis score, and MIP-1α were involved in the SVR (Table 3). These results suggest that serum IFN-λ3 fails to be a predictive marker for SVR in PEG-IFN-α plus RBV therapy.

Table 3 Factors associated with sustained virological response in patients with chronic hepatitis C who underwent 48 weeks of pegylated inferferon-α plus ribavirin therapy
Full size table

Discussion

In this study, we demonstrated that serum IFN-λ3 levels were higher in patients with C-CH than in uninfected or HBV-positive patients, the levels in whom did not differ regardless of the IFNL3 genotype. Serum IFN-λ3 levels were correlated with clinical and immunological markers of liver inflammation and fibrosis, suggesting that the production of IFN-λ3 may be regulated by not only the presence or absence of HCV but also by the status of liver disease. It is well acknowledged that IFNL3 genotype is a strong predictor of SVR in PEG-IFN-α plus RBV therapy for C-CH [7–9]. However, serum IFN-λ3 fails to be a surrogate marker for IFNL3 genotype in the combination therapy.

On primary HCV infection, IFN-λ is produced by hepatocytes that subsequently induce antiviral IFN-stimulated genes [23]. Parallel reduction of serum IFN-λ3 levels in C-CH patients who attained SVR by PEG-IFN-α plus RBV treatment indicates that the presence of HCV is involved in the production of IFN-λ3. In addition to hepatocytes, dendritic cells or macrophages are capable of producing IFN-λ in response to HCV [24]. For sensing HCV, hepatocytes and BDCA3+ dendritic cells mainly utilize Toll-like receptor 3 and retinoic acid inducible gene I, and plasmacytoid dendritic cells utilize Toll-like receptor 7 [24, 25]. It is yet to be clarified which cells—hepatocytes or dendritic cells—have stronger potential to secrete IFN-λ at the single-cell level. However, it is rational to consider that serum IFN-λ3 levels in patients are determined by the sum of IFN-λ3 sporadically released from both types of cells. Therefore, it is plausible that the amount of IFN-λ released from hepatocytes or dendritic cells is influenced by the environment of the producers, such as inflammation and fibrosis. A positive correlation observed between serum IFN-λ3 levels and AST levels, FIB-4 score, and APRI in B-CH patients may support such a possibility. In this study, serum IFN-λ3 levels in the B-CH group were higher than those in HVs. However, this difference was slim compared with the difference between the C-CH group and HVs, suggesting that the difference in their genome structure, either RNA or DNA virus, may influence IFN-λ3 production by infected cells. Of interest is the finding that serum IFN-λ3 levels were higher in patients with acute hepatitis E than in patients with acute hepatitis A. It is reported that dendritic cells localized in the intestine are capable of producing IFN-λ in response to rotavirus to protect the host from infection [26]. Although both hepatitis E virus and hepatitis A virus are RNA viruses that are transmissible by the enterofecal route, the difference in serum IFN-λ3 levels suggests that there are distinct mechanisms of recognition of hepatitis E virus and hepatitis A virus by the hosts. Further investigation is needed to disclose which pattern recognition receptors are utilized in hepatocytes or immune cells for the recognition of such viruses to produce IFN-λ.

The regulatory mechanisms of transcription and translation of IFN-λ3 have not been well documented. The IFNL3 SNPs (rs8099917) are located 8.9 kb upstream of the promoter region of the IFNL3 gene [8, 9, 11]. Because of such localization, it is less likely that the genetic variation has some impact on the transcriptional level of IFNL3. With regard to the relationship between the IFNL3 genotype and its transcripts, controversial results have been reported thus far. Some groups reported that IFN-λ3 messenger RNA levels in peripheral blood mononuclear cells were higher in patients with the IFNL3 major genotype than in those with the minor genotype [9]. In contrast, others showed that in hepatocytes such levels were comparable regardless of IFNL3 SNPs. In the search for some genetic factors influencing IFNL3 transcription, Sugiyama et al. [27] reported the existence of variable-length TA repeats in the promoter of the IFNL3 gene. Other investigators showed that a certain structure of the 3′ untranslated region in the IFNL3 gene is involved in the durability/stability of the gene [28]. Nevertheless, the contribution of such factors is not enough to fill in the gap, suggesting that certain other regulatory factors for IFNL3 are still to be revealed.

Reports concerning serum IFN-λ in C-CH patients are limited. Langhans et al. [29] showed that serum levels of IFN-λ, which includes IFN-λ2 and IFN-λ3, were higher in patients with the IFNL3 major genotype than in those with the minor genotype. One of the limitations of their study seems to be the lack of specificity for the measurement of IFN-λ3. Since the homology of IFNL2 (which encodes IFN-λ2) and IFNL3 is quite high, it is difficult to quantify specifically IFN-λ3 by excluding contamination by IFN-λ2. To exclude such a possibility, we used the newly developed chemiluminescence enzyme immunoassay for IFN-λ3, which enables one to quantify IFN-λ3 without any influence from IFN-λ2 in the range from 0 to 1,000 pg/mL. By means of this system, we found that serum levels of IFN-λ3 are not statistically different between patients with the IFNL3 major genotype and those with the minor genotype.

On primary HCV exposure, the significance of IFN-λ family members as an antiviral protein is evident. However, such impact of IFN-λ3 in chronically HCV-infected patients is still elusive. Langhans et al. [29] reported that serum IFN-λ levels in patients who had spontaneously cleared HCV were higher than in patients with chronic HCV infection, implying that a higher level of IFN-λ somewhat contributed to HCV eradication. In this study, we aimed to clarify the significance of IFN-λ3 in patients with chronic HCV infection with different approaches. Firstly, we searched for the factors influencing serum IFN-λ3 quantity by correlation analysis with clinical markers and multiple cytokines/chemokines. We found that AST, ALT, and α-fetoprotein levels and histological activity were positively correlated with serum IFN-λ3 levels. In addition, one of the noninvasive fibrosis markers, APRI, was weakly correlated with serum IFN-λ3 levels. Among the chemokines examined in this study, serum IFN-λ3 was positively correlated with IP-10 and PDGF-BB. IP-10 (CXCL10) is induced in HCV-infected hepatocytes as one of the IFN-stimulated genes, and attracts CXCR3-positive T cells and natural killer cells and subsequently activates inflammation. IP-10 is also reported to be involved in the early stage of liver fibrosis [30, 31]. A similar fibrotic function was reported for PDGF-BB, the level of which is reported to be increased in patients with advanced/fibrosis stages of HBV infection [32, 33]. These reports support the notion that IFN-λ3 is related to liver inflammation and fibrosis. As well as in B-CH patients, a positive correlation was observed between serum IFN-λ3 levels and inflammation (AST levels) and fibrosis markers (FIB-4 score and APRI). Secondly, we examined whether serum IFN-λ3 and chemokines are involved or not involved in the SVR to PEG-IFN-α plus RBV therapy for C-CH patients. We confirmed that IFNL3 genotypes, fibrosis score, and MIP-1α are associated with SVR in this cohort, but failed to do so with IP-10 and serum IFN-λ3. Several studies showed that pretreatment IP-10 levels could be a predictor of SVR in PEG-IFN-α plus RBV therapy for C-CH [34], the significance of which became stronger in combination with IFNL3 genotypes [35, 36]. One of the reasons why the IP-10 levels failed to be significant in this study may be a bias for the enrollment of patients from multiple hospitals and medical centers.

In summary, serum IFN-λ3 levels are increased in patients with chronic HCV infection regardless of the IFNL3 genotype, the level of which is associated with liver inflammation and fibrosis. The biological role and clinical impact of IFN-λ3 in patients with chronic HCV infection need to be investigated further.