IFN-λ receptor 1 expression is induced in chronic hepatitis C and correlates with the IFN-λ3 genotype and with nonresponsiveness to IFN-α therapies

The strong association of minor alleles of the IFN-λ3 polymorphism with HCV persistence and poor response to peg-IFN-α and ribavirin treatment provides compelling evidence for an important role of the IFN-λ system in the host reaction to HCV. We therefore measured the expression of IFN-λ1, IFN-λ2/3, and IFN-λR1 by quantitative PCR (qPCR) in liver biopsies. Because of the high sequence similarity of 97%, IFN-λ2 and IFN-λ3 mRNA were quantified together using one primer pair that binds to both IFN-λ2 and IFN-λ3. IFN-λ1 mRNA was detectable in 33 and IFN-λ2/3 mRNA in 24 of 53 CHC samples tested (Fig. 1 A). We could not detect IFN-λ1 or IFN-λ2/3 in 22 control biopsies from patients without CHC. Interestingly, IFN-λ1 and IFN-λ2/3 mRNA were detectable in 3 and in 20 of 31 samples from patients with CHB, respectively (Fig. 1 A).

IFN-λR1 mRNA was measured in 53 controls, 122 samples from patients with CHC, and 31 samples from patients with CHB and was detectable in all 206 biopsies (Fig. 1 A). The median expression level of IFN-λR1 mRNA expressed as copies per 40 ng of total RNA was 482.5 in CHC samples, which is significantly higher compared with control samples with a median value of 130.0 (Mann–Whitney test; P < 0.001). IFN-λR1 expression was also strongly induced in CHB samples, with a median expression level of 712 copies per 40 ng of total RNA (Fig. 1 A).

We then used a recently adapted in situ hybridization (ISH) method to identify the cell types with IFN-λR1 expression in human liver biopsies. Most of the specific signals were detected in hepatocytes (Fig. 1 B). We also detected signals near the sinusoids that could not unequivocally be localized in hepatocytes. We conclude that the major source IFN-λR1 receptor mRNA measured by qPCR from liver biopsy extracts are hepatocytes, with possible minor quantitative contributions from nonparenchymal sinusoidal cells.

The induction of IFN-λR1 specifically in patients with CHC is reminiscent of the induction of hundreds of ISGs in a substantial proportion of patients with CHC (Chen et al., 2005; Asselah et al., 2008; Sarasin-Filipowicz et al., 2008). Furthermore, it has been reported that the promoter of IFN-λR1 contains a binding site for the IFN-α signal transducer STAT1 (Yang et al., 2010). We therefore tested whether IFN-λR1 is an ISG in PHHs. Indeed, we found a significant induction of IFN-λR1 both on the transcript and the protein level within hours after IFN-α stimulation (Fig. 2, A and B). The IFN-α–induced up-regulation of IFN-λR1 had important consequences for the responsiveness of PHHs to IFN-λ2. Treatment-naive PHHs showed poor to absent phosphorylation of STAT1 in response to 100 ng/ml IFN-λ2. The response was significantly enhanced after the induction of IFN-λR1 through a pretreatment of PHHs with IFN-α (Fig. 2, C and D).

To further define the IFN-λR1 expression levels required for the activation of the Jak–STAT signaling pathway and the induction of ISGs by IFN-λs, we generated several Huh7-derived clones stably transfected with IFN-λR1 and selected three of them with different receptor expression levels, designated LR clones 1, 2, and 3 (Fig. 3, A and B). The parent Huh7 clone used for these experiments has very low IFN-λR1 expression and consequently showed only weak phosphorylation of STAT1 upon IFN-λ2 treatment and absent or only very weak induction of ISGs (Fig. 3, C and D). LR clones 1, 2, and 3 had stepwise increased expression levels of IFN-λR1 and showed corresponding increases in Jak–STAT activation and ISG induction by IFN-λ2 (Fig. 3, C and D). Of note, a significant increase in responses to IFN-λ2 was only observed in clones with IFN-λR1 expression of >0.005 relative to GAPDH. Interestingly, IFN-λ2–induced ISG expression in LR clone 2 and 3 cells increased over time during a 24-h treatment period (Fig. 3 D). This is in contrast to the IFN-α–induced ISG expression that is typically transient because of the well-known refractoriness of IFN-α signal transduction pathways in Huh7 cells and liver cells (Larner et al., 1986; Sarasin-Filipowicz et al., 2009; Makowska et al., 2011). In accordance with these published results, IFN-α induced only transient phosphorylation of STAT1 (Fig. 3 E). Contrariwise, IFN-λ signaling showed no refractoriness but prolonged stimulation of the Jak–STAT pathway in LR clone 3 cells (Fig. 3 E).

The correlation between IFN-λR1 expression and activation level of the Jak–STAT pathway by IFN-λ was not restricted to Huh7 cells and was also found in PHHs (Fig. 3 F). We made use of the observation that IFN-α stimulation induced IFN-λR1 to variable degrees. When these cells were stimulated with IFN-λ2 after incubation with IFN-α for 4 h, the strength of STAT1 phosphorylation correlated with the expression level of IFN-λR1. We conclude that in unstimulated PHHs, the IFN-λR1 expression level is low and IFN-λ–induced activation of the Jak–STAT pathway is weak. Responsiveness to IFN-λ strongly increases after induction of IFN-λR1.

There was considerable variation of IFN-λR1 mRNA expression among the CHC samples (Fig. 1 A). To investigate the influence of IFN-λ3 genotypes on IFN-λR1 expression, we genotyped CHC samples for the rs12979860 and for the rs8099917 single-nucleotide polymorphisms. We found significantly higher IFN-λR1 expression in samples from patients having one or two minor IFN-λ3 alleles (Fig. 4 A). To investigate the mechanisms responsible for the association of the IFN-λ3 genotype with IFN-λR1 expression, we genotyped PHH samples from 30 donors not infected with HCV (Table S1), stimulated them for 4 h with IFN-α, and measured the mRNA expression of IFN-λR1 and two well-characterized ISGs, i.e., STAT1 and RSAD2. In unstimulated PHHs, the expression of all three genes was low, independent of the IFN-λ3 genotype (Fig. 4, B and C). IFN-α treatment resulted in a significant induction of all three genes in all samples. For STAT1 and RSAD2, the magnitude of gene induction was not significantly associated with the IFN-λ3 genotype. Contrariwise, there was a significantly stronger induction of IFN-λR1 in samples having one or two minor IFN-λ3 alleles (Fig. 4, B and C).

To address whether the response to treatment with peg-IFN-α and ribavirin is influenced by the expression level of IFN-λR1 in the liver, we reanalyzed the expression values of liver biopsy samples of patients with CHC for whom we had complete virological response data (Table S2). There were 21 samples from patients who had a sustained virological response defined as the absence of detectable HCV RNA in the serum 6 mo after the end of the treatment. 27 samples were from patients who did not respond to peg-IFN-α and ribavirin and remained HCV RNA positive. The median expression of IFN-λR1 measured as copies per 40 ng of total RNA was 732 in pretreatment biopsy samples from nonresponder patients (Fig. 5). In the responders, the median expression was 514. The difference was statistically significant (Mann–Whitney test, P < 0.05). We conclude that the response to treatment with peg-IFN-α and ribavirin negatively correlates with expression levels of IFN-λR1 in the liver.

To further address the pharmacological response to IFN-α and IFN-λ, we used an ex vivo treatment protocol using liver biopsy pieces. 2–3-mm-long biopsy cylinder pieces of 1-mm diameter were transferred into test tubes and either left untreated or stimulated with IFN-α or IFN-λ for 15 min. The method was previously used and validated in liver biopsies with different fibrosis stages including cirrhosis (Duong et al., 2004). The activation of the Jak–STAT pathway was assessed by immunostaining of fresh-frozen biopsy material with phospho-STAT1–specific antibodies. For each sample, confocal microscope pictures of three random fields were obtained and used to quantify the number of nuclei with phospho-STAT1 signals. The expression of IFN-λR1 was quantified by qPCR. Two liver biopsies from uninfected patients were used as controls. In these samples, IFN-α induced nuclear phospho-STAT1 signals in most cells, predominantly in hepatocytes, whereas IFN-λ was ineffective (Fig. 6, A and B). The same response pattern was observed in samples from patients with CHC and low expression of the IFN-λR1 (Figs. 7 A and 8). The results were strikingly different in biopsies with high expression levels of IFN-λR1. In many of these samples, nuclear phospho-STAT1 signals were detectable already in the PBS samples, and IFN-α rarely elicited further responses. On the contrary, IFN-λ was able to increase nuclear phospho-STAT1 signals in all nine biopsy samples with high IFN-λR1 (Figs. 7 B and 8). Based on these ex vivo data, we conclude that there is an inverse correlation between responsiveness to IFN-α and IFN-λ in human liver of patients with CHC. Thus, we propose that patients with high-level pretreatment expression of ISGs in the liver, who according to previously published work often respond poorly to IFN-α (Chen et al., 2005; Asselah et al., 2008; Sarasin-Filipowicz et al., 2008; Dill et al., 2011), might indeed have better responses to IFN-λ because of significantly increased expression of IFN-λR1.

The discovery that genetic variants near the IFN-λ3 gene locus are strongly associated with spontaneous clearance of HCV by the host immune system and with response to therapy of CHC with peg-IFN-α and ribavirin was an important milestone in HCV research (Ge et al., 2009; Suppiah et al., 2009; Tanaka et al., 2009; Rauch et al., 2010). IFN-λ (IL28B) genotyping rapidly became a useful tool for clinicians for predicting the likelihood of treatment response of an individual patient, and it became essential for stratifying patients enrolled in clinical treatment trials with peg-IFN-α–containing regimens. The discovery was also important for basic research of host–virus interactions because it identified the previously neglected IFN-λ system as a key component of the host immune answer to HCV. However, the mechanisms that link IFN-λ3 genotypes with spontaneous HCV clearance and with response to peg-IFN-α remained elusive. Important insights came from studies that linked IFN-λ3 genotypes with hepatic ISG expression in CHC (Honda et al., 2010; Urban et al., 2010). Differential activation of the endogenous IFN system in the liver has previously been established as a major determinant of response to peg-IFN-α and ribavirin (Chen et al., 2005; Asselah et al., 2008; Sarasin-Filipowicz et al., 2008). Patients with an activated IFN system are generally poor responders to peg-IFN-α therapy. Therefore, an increased production and secretion of IFN-λ from minor alleles compared with major IFN-λ3 alleles would be an obvious explanation for the association of the minor IFN-λ3 alleles with increased ISG expression in the liver. However, several groups have failed to find increased expression levels of IFN-λ family members in the serum or the liver of patients with minor IFN-λ3 alleles (Honda et al., 2010; Urban et al., 2010; Dill et al., 2011). In the present work, we provide evidence that the link between the IFN-λ3 genotype and hepatic ISG expression is based on differential expression of the IFN-λR1 receptor, and not by differential induction of the IFN-λ ligands. The receptor is detectably expressed in all liver biopsies. The median expression value is significantly higher in samples from patients with CHC, and within the group of CHC samples, there is significantly increased expression in patients carrying the minor IFN-λ3 allele (compared with the major allele) and in patients who were nonresponders to peg-IFN-α and ribavirin therapy (compared with patients who were cured by this treatment). These results are in accordance with previous published results showing an elevated expression of IFNLR1 in patients with CHC infection compared with controls and in Huh7.5 cells infected with HCVcc (cell culture–derived infectious HCV; Dolganiuc et al., 2012; Kohli et al., 2012).

The ligands (IFN-λ1, -2, and -3) were consistently undetectable in liver biopsies from patients not infected with HCV. In CHC patients, IFN-λ1 and -λ2/3 are detected by qPCR in about half of the samples, mostly at low expression levels. In agreement with published work, we did not find a correlation between IFN-λ expression and response to treatment or the IFN-λ3 genotype. We therefore think that we can exclude differential regulation of IFN-λ1 and -λ2/3 as the missing link between the IFN-λ3 genotype and ISG expression. A very interesting alternative explanation has been generated by the recent discovery of a frame shift variant located upstream of the IFN-λ3 gene (Prokunina-Olsson et al., 2013). This variant creates a new open reading frame of 179 aa that has been designated IFN-λ4. Despite considerable sequence diversity between IFN-λ4 and the other members of the IFN-λ family, recombinant IFN-λ4 signals through the IFN-λR to induce an antiviral state against HCV and coronaviruses (Hamming et al., 2013). It is immediately obvious that the exclusive production of IFN-λ4 from the minor IFN-λ3 alleles (more precisely from the ss469415590-ΔG allele) together with the stronger induction of IFN-λR1 in the liver of patients with one or two minor alleles (Fig. 4) provide an attractive explanation for the high level of ISG expression in these patients. For the moment, however, our attempts to detect IFN-λ4 mRNA in liver biopsies failed. Thus, further investigations are needed to clarify whether IFN-λ4 is indeed the cytokine responsible for elevated ISG expression in patients with CHC.

Despite the fact that IFN-λs are not detectable in a substantial percentage of patients with CHC, we think that the IFN-λ system is the most likely candidate for the continuous up-regulation of hundreds of ISGs in preactivated patients. Contrary to IFN-α signaling, IFN-λ signaling does not become refractory during prolonged stimulation (Fig. 3 D; Makowska et al., 2011). Moreover, the strength of IFN-λ–induced activation of the Jak–STAT pathway is regulated to a large degree by the expression of the IFN-λR1 (Figs. 2, 3, 7, and 8). Because IFN-λR1 is an ISG itself, the IFN-λ system can undergo a positive feedback loop that can be maintained as long as there is a minimal amount of IFN-λ ligands to stimulate the expression of IFN-λR1. In this context, our observation that PHHs from non-HCV–infected donors show a differential induction of IFN-λR1 upon IFN-α stimulation according to their IFN-λ3 genotype (Fig. 4, B and C) provides an important missing link between genotype and phenotype. We propose that early during HCV infection the host innate immune system reacts with the induction of type I (IFN-α or IFN-β) and/or III (IFN-λ) IFN. In individuals with one or two minor IFN-λ3 alleles, there is a strong induction of the IFN-λR1. This induction is maintained through a positive feedback loop, and IFN-λR1 remains constitutively induced together with hundreds of other ISGs. In this group of patients, peg-IFN-α is ineffective, most likely because of the permanent induction of the ISG USP18, a strong inhibitor of signaling through the IFN-α receptor (Sarasin-Filipowicz et al., 2009; François-Newton et al., 2011; Dill et al., 2012). On the other side, in patients homozygous for the major IFN-λ3 allele, the induction of IFN-λR1 remains below a critical threshold for the positive feedback loop, and ISG induction fades away. These non-preactivated patients are highly responsive to peg-IFN-α, most likely because IFN-α signaling is not refractory.

Peg-IFN-λ1 is currently undergoing clinical development for the treatment of viral hepatitis. A phase 1b and a phase 2b study demonstrated antiviral activity in patients with CHC and minimal adverse events and hematologic effects (Muir et al., 2010; Muir, A.J., et al. 2012). The 63rd Annual Meeting of the American Association for the Study of Liver Diseases: The Liver Meeting. Abstr. 214). The better tolerability of peg-IFN-λ1 is presumably caused by the restricted tissue distribution of its receptor. Lymphocytes or monocytes are unresponsive to IFN-λ (Dickensheets et al., 2013). The antiviral activity of peg-IFN-λ1 is probably mediated through a direct stimulation of the IFN-λ receptor on hepatocytes. In tissue blots, IFN-λR1 is detectable in the liver (Sheppard et al., 2003), and IFN-λ had antiviral activity and/or activated Jak–STAT signaling in several hepatoma cells lines and PHHs, albeit often with lower efficacy compared with IFN-α (Robek et al., 2005; Doyle et al., 2006; Marcello et al., 2006; Dickensheets et al., 2013). In vivo, the response to IFN-λ in the human liver has not been investigated. Interestingly, in the mouse liver, we and others have found very little IFN-λR1 expression and little to absent Jak–STAT stimulation in the liver in response to systemic application of mouse IFN-λ (Sommereyns et al., 2008; Makowska et al., 2011).

In the present extensive analysis of IFN-λR1 expression in human liver biopsies, we measured rather low expression levels in samples from patients not infected with HCV. Thus, the noninfected human liver has very little response to IFN-λ. This state of relative unresponsiveness could protect the liver from overstimulation by IFN-λ produced during gastrointestinal viral infections and transported via the portal vein into the hepatic circulation.

IFN-λR1 expression is not increased in all CHC patients, but rather predominantly in patients with an induced IFN system in the liver who usually are carriers of the IFN-λ3 minor allele. This differential expression can be explained by the fact that IFN-λR1 is an ISG. The association of IFN-λR1 expression with the IFN-λ3 minor allele has important consequences for the use of peg-IFN-λ1 for the treatment of CHC. Collectively with the results from our ex vivo stimulation experiments (Figs. 7 and 8), the findings provide a rationale for the preferential use of peg-IFN-λ1 in patients with an induced hepatic IFN system who are usually responding poorly to the current standard of care treatment regiments that contain peg-IFN-α. On the negative side, peg-IFN-λ1 could have only limited efficiency in patients with an uninduced endogenous IFN system because of a below-threshold expression of the IFN-λ receptor. A quantification of IFN-λR1 mRNA in liver biopsies could therefore help to select the most promising peg-IFN (peg-IFN-α vs. peg-IFN-λ) to be included in the treatment regimens of individual patients.

Human IFN-α (Roferon-A) was a gift from Roche. IFN-λ2 was purchased from PeproTech, Inc. The following antibodies were used for Western blotting and immunofluorescence staining: phospho-STAT1 and phospho-STAT1 Alexa Fluor 555 conjugated (Cell Signaling Technology), STAT1 (BD), β-actin (Sigma-Aldrich), and IFN-λR1 (ProSci Inc.).

Huh7 cells were cultured in DMEM supplemented with 10% FBS and penicillin/streptomycin. The coding sequence of IFN-λR1 was amplified from a commercially available IMAGE human cDNA clone (catalog number EHA10001-99865534; Thermo Fisher Scientific) using the primers 5′-TTTTCTAGACGGCAGGAAGGCCATGGC-3′ and 5′-TCTAGACCTGGCCATGTAATGCCCCAAT-3′ and then cloned into a pcDNA3.V5-His vector with XbaI restriction sites. The expression plasmid was transfected into Huh7 cells using FuGENE HD (Roche). Cells were then selected with DMEM culture medium supplemented with 10% FBS and 1 mg/ml G418 (EMD Millipore) to generate stable IFN-λR1–expressing clones.

PHHs were isolated from liver resections from patients at the Strasbourg University Hospitals with approval from the Institutional Review Board and isolated and cultured as described previously (Krieger et al., 2010). The characteristics of the PHH samples are summarized in Table S1.

Liver biopsies from CHC and non-CHC patients were obtained in the outpatient clinic of the Division of Gastroenterology and Hepatology, University Hospital Basel. Parts of the biopsy material that were not needed for routine histopathology were used for research purposes after obtaining written informed consent. The use of biopsy material for this project was approved by the Ethikkommission Nordwest- und Zentralschweiz, Basel, Switzerland, protocol number M989/99. Grading and staging of CHC was defined according to the METAVIR classification. Serum HCV RNA was quantified using the COBAS AmpliPrep/COBAS TaqMan HCV Test and the COBAS Amplicor Monitor from Roche. Response to treatment in CHC patients was defined as described previously (Dill et al., 2011). Diagnosis of non-CHC patients was based on clinical, laboratory, and histopathological assessment according to routine clinical practice. CHC and non-CHC patients’ characteristics are summarized in Tables S2 and S3, respectively.

Protein extracts and immunoblots were performed as described previously (Duong et al., 2004).

RNA isolation from Huh7 cells was performed with TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. DNA digestion was performed with RQ1 RNase-Free DNase (Promega). RNA isolation from PHHs was performed with NucleoSpin RNA II (MACHEREY-NAGEL) according to the manufacturer’s instructions. 1 µg RNA was reverse transcribed with random hexamers and Moloney murine leukemia virus RT (Promega). Real-time qPCR was performed with SYBR green fluorescence (Applied Biosystems) and an ABI 7500 detection system (Applied Biosystems). Primers used for real-time qPCR are listed in Table S4. The IL28A gene shares 97% sequence similarity with the IL28B gene; therefore, we designed primers that specifically recognize both genes. The specificity of the PCR was assessed on 3% agarose gel. For in vitro experiments, gene expression was normalized to human GAPDH expression using the ΔCt method. For analysis in bulk tissues, gene expression was quantified using cloned cDNA (a gift from R. Hartmann, Aarhus University, Aarhus, Denmark) as standard, and results are shown as copies per 40 ng of total RNA. Complete gene expression data from PHH samples are listed in Table S1. Complete gene expression data from CHC patients are listed in Table S2.

Freshly obtained liver biopsies from CHC patients were incubated with PBS, 1,000 U/ml IFN-α, or 100 ng/ml IFN-λ2 for 15 min at 37°C. Specimens were then embedded in Tissue-Tek OCT (Sakura) and frozen in liquid nitrogen. Cryosections of 8-µm thickness were prepared and used for immunostaining analysis.

Liver biopsies sections were fixed in freshly prepared 4% formaldehyde for 15 min. After a wash with PBS, they were permeabilized in cold methanol. Background was then removed with Background Buster (Innovex Biosciences, Inc.). Phosphorylated STAT1 was detected using phospho-STAT1–Alexa Fluor 555–conjugated antibody. Sections were mounted with Mount FluorCare DAPI (Carl Roth GmbH).

Adherent cells were incubated with cold methanol for 10 min, washed with TBS-Tween, and incubated in blocking solution (5.5% FBS in TBS-Tween) for 60 min at room temperature. Cells were then washed and incubated with antibodies against phosphorylated STAT1 or IFN-λR1.

For the present study, we adapted a highly sensitive and specific ISH system (QuantiGene ViewRNA; Affymetrix) using a custom probe set targeting IFN-λR1 (IL28RA; Affymetrix; GenBank accession no. NM_170743). In brief, 10-µm sections from OCT-embedded liver biopsies were prepared, mounted onto Superfrost Plus Gold glass slides (Thermo Fischer Scientific), and kept at −80°C until use. Upon fixation, washing, and dehydration in ethanol, sections were digested in Protease QF (Affymetrix) and then hybridized with custom-designed QuantiGene ViewRNA probes against target RNA sequences. Bound probes were preamplified and subsequently amplified according to the manufacturer’s instructions. Probe oligonucleotide conjugated to alkaline phosphatase type 1 was added, followed by Fast Red substrate. Finally, slides were counterstained with Meyer’s hematoxylin and embedded with DAPI-containing aqueous mounting medium (Roti-Mount FluorCare DAPI; Roth AG).

Images were acquired using a laser-scanning confocal microscope (LSM710; Carl Zeiss) associated with Zen2 software (Carl Zeiss). For each liver biopsy section, we acquired three random fields. We avoided areas with high lymphocyte concentration and any other areas where the quality of DAPI staining would not allow reliable identification of nuclei.

Genomic DNA was isolated from liver biopsies using TRIzol reagent (Invitrogen) and DNeasy Blood & Tissue kit (QIAGEN) according to the manufacturers’ instructions. Genomic DNA from donors listed in Table S1 was obtained from nonparenchymal liver cells using the DNeasy Blood & Tissue kit. rs12979860 and rs8099917 genotyping was performed as described previously (Dill et al., 2011).

Prism4 (GraphPad Software) was used for statistical analysis.

Table S1, included as a separate PDF file, shows PHH donors’ characteristics. Table S2, included as a separate PDF file, shows CHC patients’ characteristics. Table S3, included as a separate PDF file, shows gene expression levels in non-HCV patients. Table S4, included as a separate PDF file, shows primer sequences for qPCR.

We thank Prof. P. Bachellier (Service de Chirurgie Générale, Hépatique, Endocrinienne et Transplantation, Hôpitals Universitaires de Strasbourg, France) for providing liver resections for the isolation of PHHs.

This work was supported by Swiss National Science Foundation (SNF) grants 320030_130243 and 310030B_147089 to M.H. Heim and by EU INTERREG-IV-Rhin Supérieur-FEDER-Hepato-Regio-Net 2012, Agence Nationale de Recherche sur le Sida (2012/239, 2012/318, 2013/108), and Laboratoire d’excellence LabEx HEPSYS (Investissement d’Avenir; ANR-10-LAB-28) grants to T.F. Baumert.

The authors declare no competing financial interests.