Interferon-β and interferon-λ signaling is not affected by interferon-induced refractoriness to interferon-α in vivo

Authors


  • Potential conflict of interest: Nothing to report.

  • This work was supported by the Swiss National Science Foundation (grants 320000_116106 and 320030_130243) and the Swiss Cancer League/Oncosuisse (grants OCS-02192-02-2008 and KLS-02522-02-2010).

Abstract

Therapy of chronic hepatitis C with pegylated interferon α (pegIFN-α) and ribavirin achieves sustained virological responses in approximately half of the patients. Nonresponse to treatment is associated with constitutively increased expression of IFN-stimulated genes in the liver already before therapy. This activation of the endogenous IFN system could prevent cells from responding to therapeutically injected (peg)IFN-α, because prolonged stimulation of cells with IFN-α induces desensitization of the IFN signal transduction pathway. Whether all types of IFNs induce refractoriness in the liver is presently unknown. We therefore treated mice with multiple injections and different combinations of IFN-α, IFN-β, IFN-γ, and IFN-λ. Pretreatment of mice with IFN-α, IFN-β, and IFN-λ induced a strong expression of the negative regulator ubiquitin-specific peptidase 18 in the liver and gut. As a result, IFN-α signaling was significantly reduced when mice where reinjected 16 hours after the first injection. Surprisingly, both IFN-β and IFN-λ could activate the Janus kinase–signal transducer and activator of transcription (STAT) pathway and the expression of IFN-stimulated genes despite high levels of ubiquitin-specific peptidase 18. IFN-λ treatment of human liver biopsies ex vivo resulted in strong and maintained phosphorylation of STAT1, whereas IFN-α–induced STAT1 activation was transient. Conclusion: Contrary to the action of IFN-α, IFN-β, and IFN-λ signaling in the liver does not become refractory during repeated stimulation of the IFN signal transduction pathway. The sustained efficacy of IFN-β and IFN-λ could be an important advantage for the treatment patients who are nonresponders to pegIFN-α, through a preactivated endogenous IFN system. (HEPATOLOGY 2011;)

The interferons (IFNs) are a group of cytokines that induce an antiviral state. They are currently classified into three groups: type I, type II, and type III IFNs.1, 2 The largest group comprises the type I IFNs including all members of the IFN-α, IFN-β, IFN-ε, IFN-κ, IFN-ω, and IFN-ν families.3 Humans have 12 different IFN-αs and a single IFN-β. Type I IFNs are induced in response to viral infections. All type I IFNs bind to the same IFN-α/IFN-β receptor (IFNAR) that consists of two major subunits: IFNAR1 (the a subunit in the older literature)4 and IFNAR2c (the βL subunit in older literature).5, 6 The different IFN-α and IFN-β members have substantial differences in their specific antiviral activities. However, the molecular basis of these differences is not yet known.

There is only one class II IFN: IFN-γ, which is produced by T lymphocytes when they are stimulated with antigens or mitogens. IFN-γ binds to a distinct receptor, the IFN-γ receptor (IFNGR) that consists of the two subunits IFNGR1 (previously, the α chain)7 and IFNGR2 (previously, the β chain or accessory factor).8, 9

The recently described type III IFNs IFN-λ2, IFN-λ3, and IFN-λ1 are also known as interleukin-28A (IL-28A), IL-28B, and IL-29, respectively. Similar to type I IFNs, they are also induced by viral infections.10 They signal through the IFN-λ receptor consisting of the IL-10R2 chain that is shared with the IL-10 receptor, and a unique IFN-λ receptor chain.11, 12 Unlike IFNAR, the IFN-λ receptor is not expressed ubiquitously, but is mainly restricted to epithelial cells.2 IFN-λ receptors are present in human hepatocytes.13 In the mouse liver, the IFN-λ receptor is expressed at very low levels, and systemic application of IFN-λ had very little effects in the liver compared to other tissues such as intestine, heart, lung, and skin.2, 14

All IFNs signal through the Janus kinase–signal transducer and activator of transcription (Jak-STAT) pathway to regulate the expression of their target genes in the nucleus. IFN-γ predominantly stimulates STAT1 and induces a homodimeric transcription factor complex, whereas members of the IFN-α, IFN-β, and IFN-λ families strongly activate STAT1 and STAT2 and induce the heterotrimeric transcription factor complex interferon-stimulated gene factor 3 (ISGF3). The different IFN subtypes induce overlapping but distinct sets of target genes.15

The activation of the Jak-STAT pathway is tightly controlled by several negative regulatory mechanisms. Suppressor of cytokine signaling 1 (SOCS1) and SOCS3 are rapidly induced by IFNs and prevent further STAT activation by inhibiting the Jak kinases.16 Likewise, ubiquitin-specific peptidase 18 (USP18) is a classical ISG that provides a strong negative feedback loop at the level of the receptor-kinase complex.17 As a result of the induction of these negative regulators, cultured cells become rapidly unresponsive (refractory) to continuous stimulation with IFNs, a phenomenon that has been known for more than 20 years.18 We have recently shown that refractoriness also occurs in the liver of mice injected with IFN-α.19 Repeated injection of mouse IFN-α (mIFN-α) at regular intervals resulted in constantly elevated serum concentrations, similar to what is observed in patients receiving pegylated IFN-α (pegIFN-α). Within hours after the first injection of mIFN-α, IFN-α signaling in the liver became refractory to further stimulation. Neither SOCS1 nor SOCS3 were instrumental for this long-lasting refractoriness. Instead, USP18 was identified as the key mediator.19

PegIFN-α2 together with ribavirin is the current standard of care for the treatment of chronic hepatitis C (CHC). The treatment achieves a sustained viral clearance in only 50%-60% of patients. The molecular mechanisms underlying treatment failure are still incompletely understood. In recent years, we and others have provided evidence that the endogenous IFN system is already activated in the liver of a substantial number of patients before the therapeutic application of pegIFN-α, and that such a preactivation prevents treatment responses.20-22 It is not known why this preactivation of the endogenous IFN system inhibits the response to therapeutically injected pegIFN-α, but it is conceivable that a constant stimulation of liver cells by endogenous IFNs induces refractoriness to pegIFN-α stimulation.

Comparatively few clinical studies have been performed to assess the efficacy of IFN-β for the treatment of CHC. In treatment-naive Asian patients, 24 weeks of therapy with IFN-β and ribavirin achieved a sustained virological response in 57% of treated patients.23 Interestingly, IFN-β is also effective in some patients who did not respond to previous therapies with IFN-α.24 More recently, pegIFN-λ1 was found to be effective for the treatment of CHC in a phase 1b study with 49 IFN-α–treated patients with relapse and seven treatment-naive patients.25

In the present study, we analyzed the activation patterns of the Jak-STAT signal transduction pathway and the induction of ISGs in different organs after single and repeated subcutaneous injection of IFN-α, IFN-β, and IFN-λ in mice. Unexpectedly, marked refractoriness to repeated stimulation was observed only in case of repeated stimulation with IFN-α. The sustained sensitivity to IFN-β and IFN-λ despite preactivation of the signal transduction pathways with IFN-α provides support for the further clinical exploration of IFN-β and IFN-λ for the treatment of IFN-α nonresponders.

Abbreviations

CHC, chronic hepatitis C; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; hu, human; IFN, interferon; IFNAR, interferon-α/β receptor; IFNGR, interferon-γ receptor; IL, interleukin; ISG, interferon-stimulated gene; ISGF3, interferon-stimulated gene factor 3; ISRE, interferon-stimulated response element; m, murine; PBS, phosphate-buffered saline; pegIFN, pegylated interferon; PKR, protein kinase R; RPL19, ribosomal protein L19; RT-PCR, real-time polymerase chain reaction; SEM, standard error of the mean; SOCS, suppressor of cytokine signaling; STAT, signal transducer and activator of transcription; USP18, ubiquitin-specific peptidase 18.

Materials and Methods

Cell Culture and Reagents.

Huh7 cells were grown at 37°C and 5% CO2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, penicillin, and streptomycin. All cell culture reagents were from Gibco, Basel, Switzerland. Human IFNs used for Huh7 cells treatment were IFN-α-2b (Intron A; Essex Chemie AG, Luzern, Switzerland), IFN-β-1b (Betaferon; Bayer Schering Pharma, Zürich, Switzerland), or IFN-λ2 (Peprotech Inc., Rocky Hill, NJ).

Animals.

Four- to 6-week-old male C57Bl/6 mice were used for all experiments. The animals were bred in the animal facility of the Department of Biomedicine of the University Hospital Basel under specific pathogen-free conditions. All animal experiments were conducted with the approval of the Animal Care Committee of the Canton Basel-Stadt, Switzerland.

The animals were injected subcutaneously with murine IFNs alpha-4 (mIFN-α), beta (mIFN-β), lambda2 (mIFN-λ), and gamma (mIFN-γ) in sterile phosphate-buffered saline (PBS). Control animals were injected with PBS only. The mIFN-α4 was produced as described,26 and the concentration was measured by mIFN-α enzyme-linked immunosorbent assay (PBL Interferon Source, Piscataway, NJ). Recombinant mIFN-β and mIFN-γ were purchased from Millipore (Axxora Europe, Lausen, Switzerland) and recombinant IFN-λ2 from Peprotech (Peprotech Inc., Rocky Hill, NJ). Specific activities of recombinant IFN-β and IFN-γ were 107 IU/mg and 1.15 × 107 IU/mg, respectively. The animals were euthanized by CO2 narcosis. Samples from the liver, lung, kidney, and small intestine were collected and immediately frozen in liquid nitrogen and stored at −80°C until further processing.

Western Blot Analysis.

Tissue extracts and western blots (protein immunoblots) were done as described.19 Proteins were detected with primary antibodies specific to phospho-STAT1 (Tyr701, catalog No 9171; Cell Signaling Technology, Bioconcept, Allschwil, Switzerland), STAT1 (catalog no. 610186; Transduction Laboratories, BD Biosciences Pharmingen, San Diego, CA), phospho-STAT2 (Tyr 689, catalog no. 07-224; Upstate Biotechnology, Lake Placid, NY), STAT2 (catalog no. sc950; Santa Cruz Biotechnology, LabForce AG, Nunningen, Switzerland), phospho-STAT3 (Tyr705, catalog no. 9131; Cell Signaling Technology), STAT3 (catalog no. sc482; Santa Cruz Biotechnology), and β-actin (Sigma-Aldrich Chemie GmbH, Steinheim, Germany).

Electrophoretic Mobility Shift Assay.

Nuclear extracts from 150-200 mg of liver tissue were prepared as described.27 Electrophoretic mobility shift assays were done as described.28

RNA Isolation and Real-Time Quantitative Reverse-Transcription Polymerase Chain Reaction.

RNA was isolated from Huh7 cells or shock-frozen liver and gut samples using Trizol Reagent (Invitrogen AG, Basel, Switzerland). Isolated RNA was quantified and 1 μg was reverse-transcribed with random hexamers and Moloney murine leukemia virus reverse transcriptase (Promega Biosciences Inc., Wallisellen, Switzerland). Prior to enzyme mix addition, the reaction mixture was incubated for 3 minutes at 70°C and then cooled on ice. Following the addition of the enzyme, reverse transcription was carried out for 1 hour at 37°C and stopped by incubation at 95°C for 5 minutes. Quantitative real-time polymerase chain reaction (RT-PCR) was performed based on SYBR green fluorescence (Applied Biosystems, Foster City, CA). The primers were: 5′-ATC CGC AAG CCT GTG ACT GT-3′ and 5′-TCG GGC CAG GGT GTT TTT-3′ for murine ribosomal protein L19 (mRPL19), 5′-CGG CGG AGA GAG CTT TGC-3′ and 5′-AGC TGA AAC GAC TGG CTC-3′ for mSTAT1, 5′-GTG GTT GTG GAG GGT GAG ATG-3′ and 5′-GGG ATG AGG TCT CCA GCC A-3′ for mSOCS1, 5′-AAG AGC CCG CCG AAA ACT-3′ and 5′-AGC CAC TGA ATG TAG ATG TGA CAA C-3′ for murine protein kinase R (mPKR), and 5′-CGT GCT TGA GAG GGT CAT TTG-3′ and 5′-GGT CGG GAG TCC ACA ACT TC-3′ for mUSP18. For Huh7 cell samples, the primers were: 5′-CTC AGT CCC GAC GTG GAA CT-3′ and 5′-ATC TCT CAA GCG CCA TGC A-3′ for huUSP18 and 5′-GCT CCT CCT GTT CGA CAG TCA-3′ and 5′-ACC TTC CCC ATG GTG TCT GA-3′ for human glyceraldehyde 3-phosphate dehydrogenase (huGAPDH). All reactions were run in duplicate using an ABI 7500 detection system (Applied Biosystems). The ΔCT value for mouse samples was derived by subtracting the threshold cycle (CT) value for mRPL19, which served as an internal control, from the CT values for mSTAT1, mSOCS1, mPKR, and mUSP18. The messenger RNA (mRNA) expression levels of the transcripts were calculated relative to mRPL19 using the formula 2−dCT. For human cell samples, the internal control used was human GAPDH.

Ex Vivo Liver Biopsy Treatment.

Freshly obtained liver biopsies were immersed in PBS-diluted human IFN-α-2b (Intron A; Essex Chemie AG, Luzern, Switzerland), IFN-β-1b (Betaferon; Bayer Schering Pharma, Zürich, Switzerland), or IFN-λ2 (Peprotech) and incubated for 10-60 minutes at 37°C. Longer treatment periods are not feasible ex vivo because of tissue degradation at 37°C. The liquid was then removed and the biopsy material immediately frozen in liquid nitrogen. The protocol was approved by the Ethics Committee of the University Hospital of Basel, Switzerland. Written informed consent was obtained from all patients. Biopsies were obtained from patients suffering from hepatitis C virus infection (biopsy 1), graft-versus-host disease after a liver transplant (biopsy 2), nodular regenerative hyperplasia (biopsy 3), and alcoholic steatohepatitis with liver cirrhosis (biopsy 4).

Results

Contrary to IFN-α, IFN-β Signaling Is Not Affected by Prior Stimulation with Type I IFNs.

Repeated or continuous stimulation with IFN-α rapidly induces a refractory state in cultured cells18 and in mouse liver.19 To test if prolonged stimulation with IFN-β and IFN-λ also desensitizes cells to further stimulation, and if pretreatment of cells with IFN-α induces refractoriness to IFN-β or IFN-λ and vice versa, we stimulated Huh7 cells with the different IFNs for 12 hours, and restimulated them after an additional 12-hour resting period (at the 24-hour time-point). As expected, pretreatment of cells with IFN-α induced a refractory state that prevented the activation of the Jak-STAT pathway by IFN-α at the 24-hour time-point (Fig. 1A, lane 8). IFN-α–induced STAT1 phosphorylation was also strongly reduced by pretreatment of cells with IFN-β (Fig. 1A, lane 11). Because IFN-β binds to and signals through the same receptor as IFN-α, we expected the same pattern of refractoriness. However, IFN-β signaling was not attenuated by IFN-α or IFN-β pretreatments. The phospho-STAT1 signals 30 minutes after the first and the second stimulation with IFN-β showed the same intensity (Fig. 1A, lanes 3, 9, and 13). We conclude that pretreatment of Huh7 cells with type I IFNs induces a refractory state that affects IFN-α–induced signaling, but not the response to IFN-β. Consistent with previous reports,29 IFN-λ was less potent than IFN-α in regard to STAT1 activation. However, the signal intensity after restimulation was not decreased in cells pretreated with IFN-λ or IFN-α compared to treatment-naive cells (Fig. 1A, lane 4, 10, and 14).

Figure 1.

Repeated stimulation with IFN-α, IFN-β, and IFN-λ results in different response patterns in Huh7 cells. (A) Huh7 cells were stimulated with 1000 IU/mL human IFN-α, 1000 IU/mL human IFN-β, or 500 ng/mL human IFN-λ using the following protocols: (1) single stimulation for 30 minutes (lanes 2-4), (2) single stimulation for 12 hours followed by a 12-hour resting period in complete growth medium (lanes 5-7), (3) initial stimulation for 12 hours followed by a 12-hour resting period and then restimulation for 30 minutes with the same or different IFN (lanes 8-14). IFN-induced tyrosine phosphorylation of STAT1 was assessed by immunoblotting of whole-cell extracts. Blots were then reprobed for total STAT1, USP18, and β-actin. (B) Quantitative RT-PCR analysis of USP18 mRNA expression in Huh7 cells treated for 8 hours with 1000 IU/mL human IFN-α, 1000 IU/mL human IFN-β or 500 ng/mL human IFN-λ. The data are plotted as the amount of USP18 mRNA relative to GAPDH mRNA (mean ± standard error of the mean [SEM]). (C) Huh7 cells were stimulated with 1000, 5000, or 10000 IU/mL human IFN-α or 20, 100, or 1000 IU/mL human IFN-β using the following protocols: (1) single stimulation for 15 minutes (lanes 2-7), (2) single stimulation for 12 hours followed by a 12-hour resting period in complete growth medium (lanes 8-11), (3) initial stimulation for 12 hours with 5000 or 10000 IU/mL human IFN-α or 20 or 100 IU/mL human IFN-β followed by a 12-hour resting period and then restimulation for 30 minutes with 1000 IU/mL of the same IFN as used for initial stimulation (lanes 12-15). IFN-induced tyrosine phosphorylation of STAT1 was assessed by immunoblotting of whole-cell extracts. Blots were then reprobed for total STAT1 and β-actin.

We have previously shown that USP18 is a key mediator of refractoriness to IFN-α in vivo.19 Pretreatment of cells with IFN-α and IFN-β strongly induced USP18 in Huh7 cells (Fig. 1A,B), but despite this, STAT1 phosphorylation induced by the second treatment with IFN-β and IFN-λ was not impaired.

A single stimulation of cells with 1000 IU/mL IFN-α resulted in a slightly weaker activation of STAT1 compared to 1000 IU/mL IFN-β (Fig. 1A, lanes 2 and 3). To exclude that this difference caused the differential response to repeated stimulation, we pretreated cells with higher concentrations of IFN-α and lower concentrations of IFN-β. In these conditions, IFN-α induced STAT1 activation was similar or stronger compared to IFN-β (Fig. 1C, lanes 2 to 7). However, restimulation with 1000 IU/mL IFN-α or IFN-β (Fig. 1C, lanes 12-15) revealed the same pattern of refractoriness as observed after pretreatment with 1000 IU/mL (Fig. 1A). We conclude that the different sensitivity of IFN-α and IFN-β to pretreatment-induced refractoriness is not influenced by the strength of the initial stimulation, but is an inherent characteristic of the IFN species.

IFN Subtypes Elicit Overlapping but Distinct Responses in the Liver and the Gut of Mice.

In order to gain insight into the in vivo responses to type I, II, and III IFNs, we studied the dose-response curve to IFN-α, IFN-β, IFN-λ, or IFN-γ after subcutaneous administration. Mice were sacrificed 1 hour after the injection, and liver and small intestine samples were collected and analyzed for the activation of Jak-STAT pathway components and ISG induction (Fig. 2). IFN-α, IFN-β, and IFN-γ induced a dose-dependent phosphorylation of STAT1, STAT2, and STAT3 and a dose-dependent induction of the classical ISGs SOCS1, STAT1, and USP18 both in the liver and in the gut (Fig. 2 and Supporting Fig. 1). At the lowest dose, IFN-λ already strongly induced STAT phosphorylation and ISG expression in the gut, but had no effect in the liver even at the highest dose. This ineffectiveness in the liver is most likely due to the absence or very low expression of the IFN-λ receptor chain in the mouse liver.2

Figure 2.

Dose-response relationship of IFN-α, -β, -λ and -γ in liver and gut. C57Bl/6 mice were injected subcutaneously with increasing doses of murine IFN-α, -β, -λ and -γ and sacrificed 1 hour after injection. IFN doses were: 300, 1500 and 7500 pg/g body weight IFN-α; 100, 250, and 500 IU/g body weight IFN-β; 50, 100, and 500 ng/g body weight IFN-λ and 20, 100 and 500 IU/g body weight IFN-γ. IFN-induced phosphorylation of STAT1, STAT2 and STAT3 was assessed in the (A) liver and (B) small intestine. IFN-induced expression of SOCS1 was quantified by RT-PCR in liver (C) and small intestine (D). The data are plotted as the amount of SOCS1 mRNA relative to RPL19 mRNA (mean ± SEM).

IFN-β–Induced Signaling in the Liver Is Not Refractory After Pretreatment.

We then analyzed the in vivo patterns of refractoriness using mouse samples obtained after repeated administration of different combinations of type I, II, and III IFNs. Both for the first and the second stimulation we chose doses of the different IFNs that resulted in a similar STAT1 phosphorylation 1 hour after a single injection, as established in previous experiments (Fig. 2), i.e., 300 pg/g body weight IFN-α, 500 IU/g body weight IFN-β, 50 ng/g body weight IFN-λ, and 100 IU/g body weight IFN-γ. At 16 hours after the first injection, the animals were sacrificed or injected again, either with the same IFN or with IFN-α. Mice receiving the second injection were sacrificed 1 hour later for the collection of liver, small intestine, kidney and lung samples. Signaling through the Jak-STAT pathway was analyzed at the level of STAT activation by tyrosine phosphorylation, at the level of binding of activated STAT1 to the IFN-stimulated response element (ISRE) found in promoters of ISGs, and at the level of transcriptional induction of ISGs.

Consistent with our previous findings,19 mice pretreated with IFN-α showed little response to the second injection with IFN-α (Fig. 3 and Supporting Fig. 2). The same attenuation of signals and ISG induction was observed when IFN-β-pretreated mice were injected 16 hours later with IFN-α. IFN-λ pretreatment had no effect on later IFN-α responses, most likely because of the lack of IFN-λ receptors in mouse liver. Likewise, IFN-γ pretreatment did not induce refractoriness to subsequent stimulation with IFN-α (Fig. 3 and Supporting Fig. 2). This can be explained by the observation that IFN-γ treatment did not induce an up-regulation of USP18 (Fig. 3B and Supporting Fig. 1), the key mediator of refractoriness to IFN-α in vivo.19

Figure 3.

IFN-β-induced signaling in the mouse liver is not abrogated after pretreatment. (A, B, C, D) C57Bl/6 mice were injected subcutaneously with different murine IFNs and sacrificed 1 hour or 16 hours later, or injected again 16 hours after the first administration and then sacrificed 1 hour after the second injection. IFN doses were 300 pg/g body weight IFN-α, 500 IU/g body weight IFN-β, 50 ng/g body weight IFN-λ, or 100 IU/g body weight IFN-γ. (A) Liver samples were subjected to immunoblotting for tyrosine phosphorylation of STAT1 and STAT2 as well as STAT1 and STAT2 total protein and β-actin as a loading control. (B, C) Quantitative RT-PCR analysis of USP18 (B) and PKR (C) mRNA expression in the liver. The data are plotted as the amount of USP18 or PKR mRNA relative to RPL19 mRNA (mean ± SEM). (D) Liver nuclear extracts were analyzed in EMSAs using ISRE oligonucleotide probe. Supershift of the ISGF3 band was performed using antibody specific for STAT1 (lane 12). (E) C57Bl/6 mice were injected every 3 hours with 300 pg/g body weight mIFN-α or PBS (control). Liver samples were collected 1 hour after the first, second, third, and fourth injections, as well as after three injections of mIFN-α and subsequent (3 hours later) single injection of mIFN-β (500 IU/g body weight). Whole-cell extracts were subjected to immunoblotting for tyrosine phosphorylation of STAT1 as well as STAT1 total protein and β-actin as a loading control.

Interestingly, IFN-β–pretreated mice showed a strong response to the second injection with IFN-β. Phosphorylation of STATs, DNA binding, and ISG induction were slightly decreased compared to the first injection of IFN-β. However, ISGs such as USP18 were again strongly induced relative to the expression level at time point 16 hours, demonstrating that the mouse liver remains responsive to repeated injections of IFN-β (Fig. 3A-D and Supporting Fig. 2). The same responses to IFN-α and IFN-β were found in kidney and lung (Supporting Fig. 3). We conclude that consistent with our findings in cultured cells (Fig. 1), cells in the liver, kidney, and lung remain responsive to IFN-β in vivo, whereas IFN-induced signaling becomes refractory to IFN-α.

To check if IFN-β could also overcome refractory state induced by a prolonged stimulation with IFN-α, we administered a single injection of IFN-β in mice that were continuously stimulated with IFN-α for an extended period of time. For this continuous stimulation mice were injected every 3 hours with 300 pg/g body weight mIFN-α to ensure constantly elevated IFN-α serum concentrations. In agreement with previously reported data,19 IFN signaling was refractory to the second, third, and fourth injection of IFN-α (Fig. 3E, lanes 5 to 8). In contrast, IFN-β could still induce STAT1 activation in mice that have been previously injected three times with mIFN-α (Fig. 3E, lane 9).

Repeated Administration of IFN-λ Does Not Induce a Refractory State in the Gut.

To investigate the effects of repeated administration of IFN-λ in vivo we analyzed STAT1 and 2 phosphorylation and ISG induction in small intestine samples from the same animals that were used to study the hepatic response (Fig. 4). Compared to IFN-α and -β, IFN-λ induced a stronger phosphorylation of STAT1 and STAT2 and stronger ISG up-regulation (Fig. 4A-D and Supporting Fig. 4). This finding is in line with recent studies demonstrating a primary role of IFN-λ in antiviral responses of tissues of epithelial origin.30 Remarkably, repeated administration of IFN-λ did not lead to any detectable decrease in the levels of Jak-STAT pathway activation or transcriptional induction. This apparent lack of refractoriness related to signaling through interferon-λ receptor occurred despite an important up-regulation of USP18 mRNA 1 hour after injection, suggesting that the inhibitory effect of USP18 protein may be specific to stimulation of cells with IFN-α.17

Figure 4.

Repeated administration of IFN-λ does not lead to the reduction of the signaling in the gut. (A, B, C, D) C57Bl/6 mice were injected subcutaneously with different murine IFNs and sacrificed 1 hour or 16 hours later, or injected again 16 hours after the first administration and then sacrificed 1 hour after the second injection. IFN doses were 300 pg/g body weight IFN-α, 500 IU/g body weight IFN-β, 50 ng/g body weight IFN-λ or 100 IU/g body weight IFN-γ. (A) Small intestine samples were subjected to immunoblotting for tyrosine phosphorylation of STAT1 and STAT2 as well as STAT1 and STAT2 total protein and β-actin as a loading control. (B, C, D) Quantitative RT-PCR analysis of USP18 (B), SOCS1 (C) and PKR (D) mRNA expression in the small intestine samples. The data are plotted as the amount of USP18, SOCS1, or PKR mRNA relative to RPL19 mRNA (mean ± SEM). (E) C57Bl/6 mice were injected every 3 hours with 300 pg/g body weight mIFN-α or PBS (control). Small intestine samples were collected 1 hour after the first, second, third, and fourth injections, as well as after a single injection of mIFN-λ (50 ng/g body weight) or three injections of mIFN-α and subsequent (3 hours later) single injection of mIFN-λ (50 ng/g body weight). Whole-cell extracts were subjected to immunoblotting for tyrosine phosphorylation of STAT1 as well as STAT1 total protein and β-actin as a loading control.

IFN-λ signaling in the gut was also not affected after prolonged stimulation with IFN-α. Repeated injections of mice with mIFN-α resulted in a stepwise reduction of the phospho-STAT1 signals 1 hour after every injection (Fig. 4E, lanes 5 to 8), whereas IFN-λ still induced a strong STAT1 activation in animals injected repeatedly with IFN-α (Fig. 4E, lane 10).

IFN-λ Induces Long-Lasting STAT1 Activation in Human Liver.

To study the response patterns to different IFNs in human liver, we treated human liver biopsies ex vivo with IFN-α, -β, and -λ. At a dose of 500 ng/mL, IFN-λ elicited comparable STAT1 phosphorylation as stimulation with 1000 IU/mL IFN-α or -β (Fig. 5A). Using these equipotent doses, we next studied the time-course of STAT1 activation. Consistent with our previous results,31 IFN-α induced a transient phosphorylation of STAT1 that returned to baseline within 60 minutes (Fig. 5B). Interestingly, IFN-λ induced STAT1 was still maximal after 60 minutes.

Discussion

Patients with CHC and an induction of the endogenous IFN system in the liver do not respond to therapeutically injected pegIFN-α with further stimulation of STAT1 phosphorylation, STAT1 nuclear translocation, or induction of IFN target genes.20 This apparent refractoriness of IFN signaling could explain why most patients with a preactivated IFN system are nonresponders to the current standard of care with pegIFN-α and ribavirin.20-22 The molecular mechanisms of nonresponse to therapeutically applied pegIFN-α in such preactivated patients have not been conclusively identified. One of the key components responsible for nonresponse could be USP18. USP18 (also known as UBP43) binds to IFNAR2 and inhibits the interaction of Jak1 with the receptor, thereby preventing the activation of STAT1 and STAT2.17 USP18 was found to be up-regulated in pretreatment liver biopsies of nonresponders20, 22 and it was identified as an important regulator of the antiviral activity of interferon against hepatitis C virus infection in vitro.32 Furthermore, USP18 has been identified as a key mediator of refractoriness to repeated or prolonged stimulation with IFN-α in the mouse liver.19

IFN-β and IFN-αs bind to the same receptor, IFNAR. It has also been shown that IFN-β induced signaling is enhanced in USP18 deficient mouse embryonic fibroblasts.17 However, IFN-β signaling in the mouse liver was largely unaffected in mice pretreated with IFN-α or IFN-β despite high expression levels of USP18 (Fig. 4). There is experimental evidence that IFN-β has a higher affinity to IFNAR2 compared to IFN-αs.33 It is also conceivable that such different affinity bindings might induce different conformational changes to the receptor molecules and thereby differential sensitivities to inhibition by USP18. However, solid experimental evidence supporting such a model is still lacking.

The absence or low-level expression of the IFN-λ receptor in the mouse liver prevented us from studying the induction of refractoriness of the type III IFN signaling system in the liver. We therefore used the gut as a model tissue in vivo (Fig. 4). Sensitivity to IFN-λ remained unaffected by repeated injections of IFN-λ or IFN-α, again despite high expression levels of USP18. In the case of IFN-λ, the use of a different and structurally unrelated receptor that is not bound by USP18 is a likely explanation.

IFN-λ has been shown to activate the Jak-STAT pathway in human hepatoma cells and to inhibit hepatitis C virus replication.13, 29 Using liver biopsies treated ex vivo, we provide evidence that IFN-λ is also active in the human liver. Moreover, contrary to the transient STAT1 phosphorylation signal detected in IFN-α–stimulated samples, IFN-λ induced STAT1 phosphorylation was maintained (Fig. 5).

Figure 5.

IFN-λ induces STAT1 activation in the human liver biopsies. (A) Freshly obtained human liver biopsy sample was incubated for 20 minutes at 37°C with human IFN-α (1000 IU/mL), -β (1000 IU/mL), -λ (500 ng/mL) or PBS as a control. Whole-cell extracts were subjected to immunoblotting for tyrosine phosphorylation of STAT1, STAT1 total protein, and β-actin. (B) Freshly obtained human liver biopsy samples were incubated for 10, 30, or 60 minutes at 37°C with human IFN-α (1000 IU/mL) or IFN-λ (500 ng/mL). Whole-cell extracts were subjected to immunoblotting for tyrosine phosphorylation of STAT1, STAT1 total protein, and β-actin.

It is widely assumed that the constant high serum concentrations achieved with pegIFN-α provide the decisive advantage over nonpegylated forms of recombinant IFN-α, because the permanent stimulation of the IFN signal transduction pathway will induce an uninterrupted antiviral activity in the infected hepatocytes. However, there is no experimental evidence to support this hypothesis. On the contrary, in previous work we observed a long-lasting refractoriness to IFN-α in mice that were repeatedly injected in short intervals with mIFN-α in order to maintain high serum concentrations over a prolonged period of time. The present observation that IFN-λ signaling is unaffected by IFN induced up-regulation of USP18 could provide an alternative explanation for the increased efficacy of pegIFN-α. Similar to IFN-αs, IFN-λs can be induced by type I IFNs.34 If refractoriness to IFN-α would be restricted to hepatocytes, pegIFN-α could still stimulate dendritic cells or macrophages to secrete IFN-λ and thereby indirectly sustain the expression of antiviral genes in hepatocytes. Indeed, we have previously found that pegIFN-α injections could still activate STAT1 in nonparenchymal, sinusoidal cells in patients with a preactivated hepatic IFN system whereas no further increase in phospho-STAT1 signals were found in hepatocytes.20

In conclusion, contrary to IFN-α, both IFN-β and IFN-λ continue to induce signaling through the Jak-STAT pathway in the setting of repeated or prolonged stimulation with type I or type III IFNs in vivo. We propose that pegylated IFN-β and pegylated IFN-λ could be promising treatment options specifically for patients with a preactivated hepatic IFN system who have little chances to be cured by the current standard of care with pegIFN-α and ribavirin, and for patients who are known nonresponders to previous therapies with (peg)IFN-α–based regimens.

Ancillary