USP18-based negative feedback control is induced by type I and type III interferons and specifically inactivates interferon α response

Abstract

Type I interferons (IFN) are cytokines that are rapidly secreted upon microbial infections and regulate all aspects of the immune response. In humans 15 type I IFN subtypes exist, of which IFN α2 and IFN β are used in the clinic for treatment of different pathologies. IFN α2 and IFN β are non redundant in their expression and in their potency to exert specific bioactivities. The more recently identified type III IFNs (3 IFN λ or IL-28/IL-29) bind an unrelated cell-type restricted receptor. Downstream of these two receptor complexes is a shared Jak/Stat pathway. Several mechanisms that contribute to the shut down of the IFN-induced signaling have been described at the molecular level. In particular, it has long been known that type I IFN induces the establishment of a desensitized state. In this work we asked how the IFN-induced desensitization integrates into the network built by the multiple type I IFN subtypes and type III IFNs. We show that priming of cells with either type I IFN or type III IFN interferes with the cell's ability to further respond to all IFN α subtypes. Importantly, primed cells are differentially desensitized in that they retain sensitivity to IFN β. We show that USP18 is necessary and sufficient to induce differential desensitization, by impairing the formation of functional binding sites for IFN α2. Our data highlight a new type of differential between IFNs α and IFN β and underline a cross-talk between type I and type III IFN. This cross-talk could shed light on the reported genetic variation in the IFN λ loci, which has been associated with persistence of hepatitis C virus and patient's response to IFN α2 therapy.

Figure 1. Differential desensitization studies in HLLR1-1.4 cells.

(A) Protocol used to measure desensitization. Unless otherwise indicated, cells were primed with IFN α2 or IFN β (500 pM) or IFN λ1 (50 pM). The priming phase varied between 8 and 24 hr and the resting phase between 16 and 24 hr. Cells were then challenged with IFN for different times depending of the read out. (B) Graphic representation of the EC₅₀ (pM) as determined by the luciferase activity induced by IFN α2, IFN β or IFN λ1 in naïve or primed cells. EC₅₀ were calculated from the non-linear regression fits of the luciferase activity induced by IFN in a concentration range covering 2.4 log. Priming and resting times lasted 24 hr each. Bars represent the 95% confidence limits. (C) Level of OAS-69K mRNA induced by IFN α2 (10 pM), IFN β (10 pM) or IFN λ1 (50 pM) in naïve and primed cells as determined by RT-qPCR. Data are expressed as ratios to GAPDH levels. Priming and resting times lasted 24 hr each. Bars represent the 95% confidence limits (Student's t-test). (D) Dose response induction profile of OAS-69K mRNA in naïve (closed symbols) and IFN α2 primed cells (open symbols) stimulated for 4 hr with different doses of IFN α2 (circles) or IFN β (squares) as determined by RT-qPCR. Priming and resting times lasted 24 hr each. Data are expressed as ratios to GAPDH levels. Bars represent the 95% confidence limits (Student's t-test).

Figure 2. Differential desensitization studies in HLLR1-1.4 cells.

(A) Level of phosphorylation of Stat1, Stat2 and Stat3 induced at 30 min by the indicated doses of IFN in naïve cells and in primed cells. Priming was for 8 hr and resting was for 16 hr. Cell lysates (30 µg) were fractionated on a 7% SDS polyacrylamide gel and immunoblotted with the indicated Abs. (B) Level of tyrosine phosphorylation of immunoprecipitated Tyk2 and Jak1 induced at 30 min by IFN α2 (100 pM), IFN β (100 pM) or IFN λ1 (50 pM) in naïve and primed cells. Priming was for 8 hr and resting was for 16 hr. Lysates (2 mg) were immunoprecipitated with Tyk2 Abs (top) or Jak1 Abs (bottom). The top membrane was incubated with phospho-tyrosine 4G10 mAb (P-Tyr) and the bottom membrane with phospho-Jak1 Abs. Protein content was assessed by re-blotting with Tyk2 or Jak1 specific Abs. (C) Kinetics of Stat1, Stat2 and Stat3 phosphorylation in naïve and primed cells. Cells were stimulated with 50 pM of IFN α2 or IFN β, as indicated. Priming was for 8 hr and resting was for 16 hr. Lysates (30 µg) were fractionated on a 7% SDS polyacrylamide gel and immunoblotted with the indicated Abs. (D) Level of phosphorylation of Stat2 induced at 30 min by 100 pM of IFN in naïve cells and in primed cells. Priming was for 8 hr and the resting phase varied from 24 hr to 72 hr. Cell lysates (30 µg) were fractionated on a 7% SDS polyacrylamide gel and immunoblotted with the indicated Abs.

Figure 3. Differential desensitization of human primary cells.

(A) Human foreskin fibroblasts and (B) human T cells were either left untreated (naïve) or primed for 8 hr. Cells were washed, maintained in medium without IFN for 16 hr and stimulated for 30 min with 10 and 100 pM of the indicated IFN. Cell lysates (30 µg) were analysed with the indicated Abs. (C) Human primary hepatocytes were left untreated (naïve) or primed with 500 pM of IFN α2 or 30 nM of IFN λ1 for 24 hr. Cells were washed, maintained in medium without IFN for 24 hr and stimulated for 30 min with the indicated IFN doses. Cell lysates (50 µg) were analysed with the indicated Abs to evaluate tyrosine phosphorylation and content of Jak1 and Stats. The arrow points to the band corresponding to phosphorylated Jak1. The level of USP18 (bottom panel) was assessed in a 10% SDS PAGE. Of the two USP18 bands (apparent MW of 38 and 35 kDa), the faster migrating one results from proteolytic processing . This latter comigrates with a non specific cross-reacting band detected in naïve cells and indicated by the asterisk (bottom panel).

Figure 4. Analysis of the type I IFN receptor in naïve and primed HLLR1-1.4 cells.

(A) Surface level of IFNAR1 and IFNAR2 in naïve cells and in IFN β or IFN λ1-primed cells as determined by FACS. Cells were primed for 24 hr, washed and maintained in medium without IFN for 24 hr. Cells were then stained with AA3 mAb (IFNAR1) or CD118 mAb (IFNAR2) followed by biotinylated rat anti-mouse Ab and streptavidin-PE. (B) Level of phosphorylation of Stat1 in naïve and primed cells stimulated for 30 min with 100 pM of IFN α2, IFN β or IFN α2-HEQ. Lysates (30 µg) were immunoblotted with the indicated antibodies. Priming was for 8 hr followed by 16 hr resting in medium without IFN. (C and D) Binding of ¹²⁵I labelled IFN α2 (C) or IFN α2-HEQ (D) at 37°C for 1 hr to naïve (closed circles), IFN β-primed cells (triangles) or IFN λ1-primed cells (open circles). Cells were primed for 8 hr and maintained without IFN for 16 hr.

Figure 5. USP18 is necessary for differential desensitization.

(A) Stat1 phosphorylation induced in HLLR1-1.4 cells stimulated for 30 min with IFN α2 (100 pM), IFN β (100 pM) or IFN γ (1 ng/ml) in naïve cells and in cells primed with either IFN β (500 pM) or IFN γ (10 ng/ml). Cells were primed for 8 hr and maintained without IFN for 16 hr. (B) Stat3 phosphorylation induced in HLLR1-1.4 stimulated for 30 min with with IFN α2 (100 pM), IFN β (100 pM) or hIL-6 (10 ng/ml) in naïve cells and in cells primed with IFN β (500 pM) or hIL-6 (100 ng/ml). Cells were primed for 8 hr and maintained without IFN for 16 hr. Lysates (30 µg) were immunoblotted with the indicated Abs. (C) Level of USP18 mRNA in HLLR1-1.4 cells stimulated for 6 hr with IFN α2, IFN β (500 pM), IFN λ1 (50 pM), IFN γ (1 ng/ml) or hIL-6 (100 ng/ml) as determined by qRT-PCR. Each sample was run in triplicate. Transcripts were normalized to the level of 18S transcripts. The ratios between treated and untreated samples in each subset are shown, taking as 1 the ratio in untreated samples. (D) Kinetic profile of USP18 induction in HLLR1-1.4 cells stimulated with 100 pM of IFN β or IFN λ1 for the indicated times. Cell lysates (30 µg) were immunoblotted with the indicated Abs. The asterisk points to a nonspecific band. (E) USP18 is necessary for differential desensitization. HLLR1-1.4 cells were transfected with a control pool of siRNA (Control siRNA) or a pool of four USP18 targeting siRNA (USP18 siRNA). Twenty four hr after transfection, cells were either left untreated (naïve) or primed for 8 hr with the indicated IFN. After 16 hr of resting, cells were stimulated for 30 min with 100 pM of IFN α2 or IFN β. Cell lysates (30 µg) were analysed with the indicated antibodies. The asterisk in the bottom panel points to a band cross-reacting with anti-USP18 Abs (see also USP18 blot in Fig. 3C). Individual USP18 targeting siRNA were also used with similar results (data not shown).

Figure 6. USP18 is sufficient to induce differential desensitization.

(A) Level of Stat2 and Stat1 phosphorylation induced by 30 min stimulation with IFN α2 or IFN β in naïve and primed HLLR1-1.4 cells and in clone HU13 stably expressing USP18. Level of USP18 in naïve and primed HLLR1-1.4 cells (endogenous USP18) and in HU13 cells (ectopic USP18). Level of ISG15, a typical ISG, in naïve and primed HLLR1-1.4 and in HU13 cells. Loading was evaluated by measuring AKT. Lysates (30 µg) were immunoblotted with the indicated Abs. (B) Kinetics of Tyk2, Stat1 and Stat2 phosphorylation in the USP18-expressing clone HU13 and in the parental HLLR1-1.4 cells. Cells were stimulated as indicated with 100 pM of IFN α2 or IFN β. Lysates (30 µg) were immunoblotted with the indicated Abs. (C) Kinetics of Tyk2 and Stat1/2 phosphorylation in parental HLLR1-1.4 cells and USP18-expressing HU13 cells. Cells were stimulated as indicated with 30 pM of IFN λ1. (D) Luciferase activity induced by IFN α2 (closed circles) or IFN β (open circles) in HP1 control clone and in HU13 clone constitutively expressing USP18. (E) Ratio of the EC₅₀ values determined for luciferase activity on the control clone HP1 and clone HU13. Cells were stimulated with the indicated IFN subtypes for 6 hr. Bars represent support limits of the ratio from 95% confidence intervals of the individual EC₅₀.

Figure 7. Cells expressing USP18 are defective in IFN α2 binding.

(A) Cell surface levels of IFNAR1 (left) and IFNAR2 (right) in parental HLLR1-1.4 cells, USP18-expressing clone HU13 and control clone HP1was determined by FACS. Binding of ¹²⁵I labelled IFN α2 (B), or IFN α2-HEQ (C) at 37°C for 1 hr on HLLR1-1.4 cells (closed circles) and clone HU13 (triangles).